Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Prefrontal cortical GABAergic regulation of cognition : implications for schizophrenia & other psychiatric… Auger, Meagan 2018

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata

Download

Media
24-ubc_2018_september_auger_meagan.pdf [ 3.75MB ]
Metadata
JSON: 24-1.0367926.json
JSON-LD: 24-1.0367926-ld.json
RDF/XML (Pretty): 24-1.0367926-rdf.xml
RDF/JSON: 24-1.0367926-rdf.json
Turtle: 24-1.0367926-turtle.txt
N-Triples: 24-1.0367926-rdf-ntriples.txt
Original Record: 24-1.0367926-source.json
Full Text
24-1.0367926-fulltext.txt
Citation
24-1.0367926.ris

Full Text

PREFRONTAL CORTICAL GABAERGIC REGULATION OF COGNITION: IMPLICATIONS FOR SCHIZOPHRENIA & OTHER PSYCHIATRIC DISORDERS  by  Meagan L. Auger M. Sc., Neuroscience-McGill University, 2012 B.Sc. (Honours), Microbiology- University of Victoria, 2008     A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in The Faculty of Graduate and Postdoctoral Studies (Neuroscience)   THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver) May 2018 © Meagan L. Auger, 2018ii  The following individuals certify that they have read, and recommend to the Faculty of Graduate and Postdoctoral Studies for acceptance, the dissertation entitled:  Prefrontal cortical GABAergic regulation of cognition: Implications for schizophrenia & other psychiatric disorders  submitted by                      Meagan Auger                         in partial fulfillment of the requirements for the degree of                   Doctor of Philosophy____________                                                                   in _________Neuroscience____________   Examining Committee:  Dr. Stan Floresco, Psychology_____________________________________________________  Supervisor  Dr. Anthony Phillips, Psychiatry___________________________________________________  Supervisory Committee Member   Dr. Jeremy Seamans, Psychiatry____________________________________________________  Supervisory Committee Member  Dr. Jason Snyder, Psychology_____________________________________________________  Examining Committee Member   Dr. Cheryl Wellington, Pathology__________________________________________________  Examining Committee Member  Additional Supervisory Committee Members:  Dr. Liisa Galea, Psychology_______________________________________________________  Supervisory Committee Member    iii  Abstract Deficient GABA signalling in the frontal lobes has been posited as a pathophysiological mechanism underlying symptoms and cognitive impairments in schizophrenia and other psychiatric disorders. Yet, there has been a lack of basic research assessing how decreased prefrontal cortex (PFC) GABAergic transmission impacts cognition. The experiments described here were aimed at elucidating how PFC GABA signalling regulates working memory and attention, two core cognitive processes altered in psychiatric conditions. In the first experiment, pharmacological reduction of PFC GABAA receptor transmission led to delay-independent deficits in working memory, suggesting that PFC GABA signalling may be particularly important for working memory encoding, while PFC NMDA glutamatergic transmission appears to be necessary for working memory maintenance. Given that attention strongly influences encoding, the next experiment identified separable attentional processes modulated by PFC GABAergic transmission. In addition to disrupting attention, PFC GABA dysfunction may contribute to working memory deficits by impairing filtering of distracting information. In the third study, PFC GABAergic regulation of resistance to proactive interference from past information was examined in a massed-trials variant of the reference/working radial maze task. While PFC GABAA antagonism did not increase proactive interference effects, strong impairments in working and reference memory were found across the test session. PFC inactivation did not affect performance, indicating that disinhibition of the PFC may interfere with activity of other circuitry responsible for mnemonic or cognitive functions. To investigate this, expression of c-Fos, a marker of neuronal activation, was throughout the brain following PFC GABAA antagonism in the final studies. Reduced PFC GABA function was associated with widespread increases in neuronal activation in PFC efferent regions in animals at rest. iv  Intriguingly, enhanced neuronal activation following PFC disinhibition was only observed in the hippocampus and rhomboid thalamic nucleus of animals trained on the task, suggesting that plasticity in the PFC-thalamic-hippocampal circuit associated with learning may alter the effects of diminished PFC GABA function on neuronal activation. Collectively, the work described identifies component aspects of cognition affected by deficiencies of PFC GABA, and suggests that diminished or dysfunctional PFC GABA signalling could play a role in cognitive deficits observed in neuropsychiatric disorders.   v   Lay Summary Schizophrenia has been recognized both clinically and in the public as a disorder with the identifying features of hallucinations or delusions, perhaps hearing voices or having illusions of persecution from unlikely sources. Although these symptoms may be dramatic, it is increasingly understood that deficits in cognition, including impaired memory and attention, are major predictors of patient outcomes. Furthermore, while positive symptoms are treatable with antipsychotics, currently available treatments do not ameliorate cognitive dysfunction in schizophrenia. Thus, identification of brain mechanisms that underlie these aspects of cognition is critical for developing new, effective treatments. The experiments described attempt to elucidate how deficiencies in an inhibitory neurochemical act within regions of the cerebral cortex to contribute to abnormal cognition. In particular, they focus on working memory, the ability to maintain and manipulate information in the mind, which is known to be affected in schizophrenia and other psychiatric disorders.   vi   Preface Experiment 1 (Chapter 3) has previously been published in the manuscript Auger ML & Floresco SB (2017) Prefrontal cortical GABAergic and NMDA glutamatergic regulation of delayed responding. Neuropharmacology 113:10-20. I conducted all experiments, performed data analysis, made figures and wrote the manuscript. Dr. Stan Floresco performed data analysis, edited and helped prepare the manuscript and was responsible for the experimental design. Experiment 2 (Chapter 4) has previously been published in the manuscript Auger ML, Meccia J & Floresco SB (2017) Regulation of sustained attention, false alarm responding and implementation of conditional rules by prefrontal GABAA transmission: comparison with NMDA transmission. Psychopharmacology ePub ahead of print. I trained animals, performed surgeries, administered all pharmacological agents, analyzed data, made figures and wrote the manuscript. Juliet Meccia participated in animal training and data collection. Dr. Stan Floresco analyzed data, edited and helped prepare the manuscript and was responsible for the experimental design. Experiment 3 (Chapter 5) has previously been published in the manuscript Auger ML & Floresco SB (2014) Prefrontal cortical GABA modulation of spatial reference and working memory. International Journal of Neuropsychopharmacology 18:pyu013. I conducted all experiments, made figures and wrote the manuscript. Dr. Stan Floresco analyzed the data, edited and helped prepare the manuscript and was responsible for the experimental design. Experiment 4 (Chapter 6) was submitted as a manuscript as Auger ML, Meccia J, Galea LAS & Floresco SB.  I trained animals, administered pharmacological agents, performed and quantified the result of immunohistochemistry experiments, made the figures, wrote the manuscript and vii  contributed to experimental design. Juliet Meccia participated in training of animals and data collection. Dr. Liisa Galea contributed to the experimental design and provided space and equipment for the experiments. Dr. Stan Floresco performed statistical analysis, edited and helped prepare manuscript and contributed to the experimental design.  All experimental protocols were approved by the Animal Care Committee (ACC), University of British Columbia, and conducted in compliance with guidelines provided by the Canadian Council on Animal Care (CCAC).  ACC certificate numbers: A10-0197 or A14-0210   viii  Table of Contents  Abstract.........................................................................................................................................iii Lay Summary.................................................................................................................................v Preface............................................................................................................................................vi Table of Contents....................................................................................................................... viii List of Tables................................................................................................................................ xv List of Figures............................................................................................................................. xvi List of Abbreviations............................................................................................................... xviii Acknowledgements .................................................................................................................... xx   Chapter 1: General Introduction.................................................................................................1 1.1 GABAergic transmission in the PFC.................................................................................2 1.1.1 Synthesis and release of GABA....................................................................................2 1.1.2 GABA receptors............................................................................................................3 1.1.3 GABA reuptake.............................................................................................................5 1.1.4 Cortical GABAergic neurons are diverse......................................................................5 1.2 Dysfunctional Prefrontal GABA Signalling is a Feature of Schizophrenia & Other Psychiatric Disorders................................................................................................................8 1.2.1 Schizophrenia................................................................................................................8 1.2.1.1. Pathological findings. ...........................................................................................9 1.2.1.2. Electrophysiological findings.  ...........................................................................11 1.2.1.3. In vivo imaging findings.  ...................................................................................13 1.2.1.4. Environmental risk factors.  ...............................................................................15  1.2.1.5. Genetic risk factors.  ...........................................................................................16 1.2.1.6. PFC NMDA receptor hypofunction.  ..................................................................18 1.2.1.7. Oxidative Stress.  ................................................................................................19 ix  1.2.2 Depression...................................................................................................................21 1.2.3 Autism spectrum disorder............................................................................................22 1.2.4 Age-related cognitive decline......................................................................................23 1.2.5 Common risk factors and common cognitive impairments: a role for deficient PFC GABA signalling across psychiatric disorders? ..................................................................24 1.3 Preclinical Studies Assessing the Effects of Decreased PFC GABAergic Transmission............................................................................................................................25 1.3.1 Behavioral Flexibility..................................................................................................27 1.3.2 Attention......................................................................................................................28 1.3.3 Decision-making..........................................................................................................30 1.3.4 Speed-of-processing....................................................................................................31 1.3.5 Positive symptomatology of schizophrenia.................................................................32 1.3.6 Negative symptomatology of schizophrenia...............................................................34 1.3.7 PFC GABAergic regulation of working memory and the present work.....................35   Chapter 2: Prefrontal cortical GABAergic and NMDA glutamatergic regulation of delayed responding. ..................................................................................................................................39 2.1 Introduction.......................................................................................................................39 2.2 Materials and Methods.....................................................................................................43 2.2.1 Subjects.......................................................................................................................43 2.2.2 Behavioral procedures.................................................................................................43 2.2.3 Delayed non-match to position task............................................................................44 2.2.4 Surgery........................................................................................................................45 2.2.5 Drugs & microinfusion procedures.............................................................................46 2.2.6 Histology.....................................................................................................................47 2.2.7 Data Analysis..............................................................................................................49    x  2.3 Results................................................................................................................................49 2.3.1 PFC Inactivation..........................................................................................................49 2.3.2 Pharmacological Reduction of PFC GABAA transmission.........................................51 2.3.3 Non-selective PFC NMDA glutamate receptor antagonism.......................................54 2.3.4 PFC GluN2B subunit-specific NMDA receptor antagonism......................................56 2.4 Discussion...........................................................................................................................56 2.4.1. Prefrontal GABA signalling and delayed-response working memory.......................59 2.4.2. Prefrontal NMDA receptor signalling and delayed-response working memory........61 2.4.3. Implications for working memory functions regulated by the PFC...........................64 2.4.4. Implications for schizophrenia...................................................................................65 2.4.5. Conclusions................................................................................................................66   Chapter 3: Regulation of sustained attention, false alarm responding and implementation of conditional rules by prefrontal GABA transmission: comparison with NMDA transmission..................................................................................................................................67 3.1 Introduction.......................................................................................................................67 3.2 Methods..............................................................................................................................70 3.2.1 Subjects.......................................................................................................................70 3.2.2 Behavioral procedures.................................................................................................71 3.2.3 Sustained attention task...............................................................................................72 3.2.4 Conditional visual discrimination...............................................................................74  3.2.5 Conditional auditory discrimination...........................................................................74 3.2.6 Surgical procedures.....................................................................................................75 3.2.7 Drugs & microinfusion procedures.............................................................................76 3.2.8 Histology.....................................................................................................................78 3.2.9 Data analysis...............................................................................................................78   xi  3.3 Results................................................................................................................................80 3.3.1 PFC GABAA receptor antagonism and sustained attention.........................................80   3.3.2 PFC NMDA receptor antagonism and sustained attention...........................................82    3.3.3 Systemic NMDA receptor antagonism and sustained attention...................................85  3.3.4 PFC GABAA receptor antagonism and conditional visual discrimination...................87  3.3.5 PFC GABAA receptor antagonism and conditional auditory discrimination...............90 3.4 Discussion...........................................................................................................................92 3.4.1 PFC GABA hypofunction disrupts vigilance and increases false alarm responding...92 3.4.2 Diminished PFC NMDA receptor signalling produces subtle vigilance deficits…......96 3.4.3 Diminished PFC GABA signalling disrupts performance of conditional discrimination.......................................................................................................................98 3.4.4 Deficient PFC GABA transmission leads to schizophrenia-like cognitive impairments........................................................................................................................100 3.4.5 Conclusions................................................................................................................103  Chapter 4: Prefrontal cortical GABA modulation of spatial reference and working memory.......................................................................................................................................104 4.1 Introduction.....................................................................................................................104 4.2 Methods............................................................................................................................107 4.2.1 Subjects......................................................................................................................107 4.2.2. Behavioral procedures..............................................................................................107  4.2.3 Reference/working memory task...............................................................................108 4.2.4 8-arm foraging task....................................................................................................109 4.2.5 Spatial discrimination................................................................................................109  4.2.6 Surgical procedures....................................................................................................109 4.2.7 Drugs & microinfusion procedures............................................................................110 4.2.8 Histology....................................................................................................................111 4.2.9 Data analysis..............................................................................................................111  xii  4.3 Results..............................................................................................................................113 4.3.1 Reduced prefrontal GABAA receptor transmission and reference/working memory performance........................................................................................................................113  4.3.2 Reduced Prefrontal GABAA transmission and performance of an 8-arm foraging task......................................................................................................................................115 4.3.3 PFC inactivation and reference/working memory performance...............................118 4.3.4 Reduction of prefrontal GABAA transmission and spatial discrimination................120 4.4 Discussion.........................................................................................................................122 4.4.1 Prefrontal GABA signalling and reference/working memory....................................123 4.4.2 PFC GABA transmission and choice latencies...........................................................126 4.4.3 Implications for schizophrenia...................................................................................127 4.4.4 Conclusions................................................................................................................129   Chapter 5: Disinhibition of the prefrontal cortex leads to brain-wide increases in neuronal activation that are modified by spatial learning.....................................................................130 5.1 Introduction.....................................................................................................................130 5.2 Methods............................................................................................................................133 5.2.1 Subjects.....................................................................................................................133 5.2.2 Surgical procedures...................................................................................................133 5.2.3 Experimental groups & timeline...............................................................................134  5.2.4 Behavioral procedures...............................................................................................135 5.2.5 Drugs & microinfusion procedures...........................................................................136 5.2.6 Tissue processing......................................................................................................136 5.2.7 Histology...................................................................................................................136 5.2.8 c-Fos immunohistochemistry....................................................................................138 5.2.9 c-Fos quantification...................................................................................................138 5.2.10 Data Analysis..........................................................................................................138  xiii  5.3 Results..............................................................................................................................141 5.3.1 Effects of PFC GABAA antagonism on neuronal activation under baseline conditions...........................................................................................................................144 5.3.2 PFC GABA antagonism and radial maze performance.............................................145 5.3.3 Neuronal activation following training and testing on the RM/WM task.................147 5.3.3.1 Frontal lobe regions. ...............................................................................................147 5.3.3.2 Striatum. .................................................................................................................149  5.3.3.3 Thalamus. ...............................................................................................................151 5.3.3.4 Temporal lobe regions. ..........................................................................................153  5.4 Discussion.........................................................................................................................157 5.4.1 Baseline effects of PFC disinhibition.........................................................................158 5.4.2 Effects of PFC disinhibition following spatial learning or performance....................160 5.4.3 Effects of training and testing on the RM/WM radial maze in saline-treated animals................................................................................................................................162 5.4.4 Neuronal activation in limbic structures implicated in anxiety and fear...................163 5.4.5 PFC imbalances in excitation-inhibition impact activity throughout the brain.........164 5.4.6 Implications for schizophrenia and other psychiatric disorders................................166 5.4.7 Conclusions...............................................................................................................168    Chapter 6: General Discussion.................................................................................................169 6.1 Overview of findings.......................................................................................................169 6.2 Implications for cognition mediated by the frontal lobes............................................172 6.2.1 PFC GABAergic regulation of working memory processes.....................................172 6.2.2 PFC GABAergic regulation of attention & response inhibition...............................178 6.2.3 PFC GABAergic transmission & interference resistance.........................................180 6.2.4 PFC GABAergic transmission & processing speed..................................................182 6.2.5 Contrasting of deficient PFC GABAergic transmission & other PFC alterations....183 6.2.6 Consequences of PFC disinhibition in cortical and subcortical circuitry.................186 xiv  6.3 Experimental considerations, limitations & future directions.....................................190 6.4 Relevance to schizophrenia and other neuropsychiatric disorders..............................195 6.4.1 Schizophrenia............................................................................................................195 6.4.2 Other neuropsychiatric disorders...............................................................................199 6.4.2.1. Depression. .......................................................................................................199  6.4.2.2. Autism spectrum disorders. ..............................................................................200 6.4.2.3. Epilepsy & traumatic brain injury-related disinhibition.  ..................................201     6.4.3 Cortical GABAergic deficits & age-related cognitive impairment...........................203 6.4.4 PFC GABA hypofunction as a cross-diagnostic mechanism underlying cognitive impairment..........................................................................................................................204 6.5 Conclusions......................................................................................................................206   References...................................................................................................................................207 xv  List of Tables Table 3.1. Latencies for signal and non-signal trials and total omissions for each experiment in Chapter 3. ......................................................................................................................................82 Table 5.1. Results of ANOVAs conducted on neuronal activation data in baseline animals in individual brain regions with treatment as between-group factor. ................................................142    xvi  List of Figures Figure 2.1. Locations of all acceptable placements within the medial PFC region for DNMTP experiments. ..................................................................................................................................47 Figure 2.2. PFC inactivation impairs DNMTP performance. ......................................................49  Figure 2.3. PFC GABAA receptor antagonism leads to delay-independent impairments in DNMTP performance. ..................................................................................................................................51  Figure 2.4. PFC NMDA receptor antagonism also impairs DNMTP performance......................54  Figure 2.5. Prefrontal NR2B subunit-specific NMDA receptor antagonism does not disrupt DNMTP task performance.............................................................................................................56 Figure 3.1. Location of all acceptable placements within the medial PFC for each experiment in Chapter 3…………........................................................................................................................78  Figure 3.2. PFC GABAA receptor antagonism decreases vigilance at all stimulus durations and increases false alarm responding on the SAT................................................................................80 Figure 3.3. PFC NMDA receptor antagonism decrease vigilance at short durations without influencing hits or false alarm responses.......................................................................................83 Figure 3.4. Systemic NMDA receptor antagonism decreases vigilance without affecting hits or false alarm responses.....................................................................................................................85 Figure 3.5.  Intra-PFC administration of a higher dose of GABAA antagonist disrupts the ability to perform visual conditional discriminations. .............................................................................88  Figure 3.6. Intra-PFC administration of the higher dose of GABAA antagonist disrupts performance of conditional auditory discriminations....................................................................90  Figure 4.1. Location of all acceptable infusion placements within the medial PFC for experiments in Chapter 4..................................................................................................................................111 Figure 4.2. Prefrontal GABAA antagonism disrupts spatial reference and working memory performance on a radial maze task...............................................................................................113  Figure 4.3.  Prefrontal GABAA receptor antagonism disrupts short-term memory on an 8-arm foraging task.................................................................................................................................115   Figure 4.4. PFC inactivation does not affect spatial reference/working memory performance.................................................................................................................................118  Figure 4.5.  Prefrontal GABAA receptor antagonism and performance of a 2-arm spatial discrimination..............................................................................................................................120 Figure 5.1. Location of all acceptable placements in neuronal activation experiments..................................................................................................................................136  Figure 5.2. Regions of interest in neuronal activation experiments............................................138 Figure 5.3. Behavioral data for animals tested in the RM/WM radial maze................................141 xvii  Figure 5.4. Neuronal activation data for untrained animals (baseline).......................................144 Figure 5.5. Frontal neuronal activation data in animals trained and tested on the RM/WM radial maze.............................................................................................................................................147 Figure 5.6. Striatal neuronal activation data in animals trained and tested on the RM/WM radial maze.............................................................................................................................................149 Figure 5.7. Thalamic neuronal activation data in animals trained and tested on the RM/WM radial maze.............................................................................................................................................151 Figure 5.8. Temporal lobe neuronal activation data in animals trained and tested on the RM/WM radial maze...................................................................................................................................153    xviii  List of Abbreviations 3- or 5-CSRTT- 3 or 5 choice serial reaction time task ACC-anterior cingulate cortex AM-anteromedial thalamus Amg-amygdala ANOVA-analysis of variance ATC-average time per choice BDNF-brain-derived neurotrophic factor BIC-bicuculline BLA-basolateral amygdala CANTAB-Cambridge Neuropsychological Test Automated Battery CM-centromedial nucleus of thalamus CeA-central amygdala Cing-cingulate Cl-chloride CPT-continuous performance test CS+ conditioned stimulus CS- unconditioned stimulus DNMTP-delayed non-match to position EEG-electroencephalography Ent-entorhinal cortex FDG-fluorodeoxyglucose FGF-fibroblast growth factor GABA-gamma-aminobutryic acid GAD-glutamic acid decarboxylase GAT-GABA transporter  IEG-immediate early gene IL-infralimbic cortex ITI-inter-trial interval M1-primary motor cortex MD-mediodorsal thalamus xix  MEG-magnetoencephalography MRI-magnetic resonance imagery MRS-magnetic resonance spectroscopy n.s.- not significantly different NAc-nucleus accumbens OFC-orbitofrontal cortex Peri-perirhinal cortex PET-positron emission tomography PFC-prefrontal cortex PPI-prepulse inhibition PrL-prelimbic cortex PV-parvalbumin PVT-paraventricular nucleus of thalamus Re-nucleus reuniens Rh-rhomboid nucleus of thalamus RM-reference memory S1-primary somatosensory cortex SAL-saline SAT-sustained attention task SST-somatostatin STR-striatum Sub-subiculum TTI-time to initiate VIP-vasoactive intestinal protein VGAT-vesicular GABA transporter WM-working memory      xx  Acknowledgements To begin, I would like to thank my supervisor, Dr. Stan Floresco, for his guidance over the course of this degree. I couldn’t have had a supervisor who was more fun and supportive, and who always has an open door. I am grateful for all of the laughs, parties, advice, insight and enthusiasm for the fine art of behavioral analysis along the way. I am also grateful for my committee members, Dr. Liisa Galea, Dr. Anthony Phillips and Dr. Jeremy Seamans. I thank Liisa for her involvement in the c-Fos project and also for all of her support and encouragement over the years. I also thank Tony for his collaboration, and for his guidance towards reading more about oscillations and epilepsy during preparation for my Comprehensive Exam, which has been valuable for me. Jeremy provided important critiques of the project that led me to develop my thinking and improved the way I speak about my work. I would also like to thank Dr. Todd Woodward, who served as an examiner on my Comprehensive Exam. The papers he suggested I read in preparation for the exam changed the way I think about cognitive dysfunction in schizophrenia, and provided important points of comparison for the present research. Next, I would like to thank my lab mates and other students in the program who made this journey more enjoyable. In our lab, I thank Courtney Bryce, Debra Bercovici, Nicole Jenni, Patrick Piantadosi, Colin Stopper and Josh Larkin and Ryan Tomm, fellow grad students in the lab for all of the support and the adventures along the way. Shelly Su, Wendy Adams, Gemma Floresco, Einar Einarsson, and Mieke van Holstein were excellent post-doc mentors to me. I am grateful to Maric Tse for his training and all of the technical pointers over the years. I was lucky to have excellent undergrad assistants throughout my degree, including Juliet Meccia, Nicholas Chan, and Natasha Lieuson, and to also profit off of the other wonderful undergrads in our lab who trained animals and helped out in other ways from time to time, including Olivia Li and xxi  Jade Adalbert. In particular, I thank Juliet, I don’t think I would have made it through this degree with (as much of) my sanity without your assistance and friendship. I thank the staff of the animal facility, Lucille Hoover, Alice Chan and Anne Cheng, for all of their hard work and care over the years. I would also like to thank my friends and family for all of their support over the years. In particular, I thank Stephanie Meitz, who provided me with a place to live when I came here from Montreal, even though that place sometimes involved her daughter, Oneida, leaving half-eaten carrots in my bed. I thank my cat, Misha, for all of the cuddles and for being as reliable an alarm clock as any. Lastly, I thank my partner, Spencer Davis, for everything; for always showing interest in what I have to say, for encouraging me in everything I do, and for making life outside the lab wonderful.  1  Chapter 1: General Introduction γ-aminobutyric acid (GABA) is the main inhibitory neurotransmitter in the mammalian central nervous system. It is released chiefly by interneurons, locally projecting cells with diverse morphological, physiological and molecular specialization (Ascoli et al. 2008), though certain long-distance projections are also known to consist of neurons that release GABA (Lee et al. 2014; Tamamaki and Tomioka 2010). Despite the immense diversity of cell types that release GABA, GABAergic neurons share the common property of regulating the excitability and output of other neurons (Roux and Buzsáki 2015). Recent work suggests that rather than simply serving the function of dampening or balancing out levels of excitatory transmission, GABAergic interneurons play a critical role in shaping neural activity at both the microcircuit and network level (Tremblay et al. 2016). In line with this, GABA release from interneurons has been implicated in the generation of oscillatory activity that is thought to give rise to cognitive functions, including working memory and cognitive control (Buzsáki and Wang 2012). In psychiatric disorders like schizophrenia, both disruptions in oscillations and cognition are thought to stem from dysfunction of interneurons and GABAergic neurotransmission, particularly in the prefrontal cortex (PFC), a region which is a key player in many cognitive functions (Gonzalez-Burgos and Lewis 2012; McNally and McCarley 2016; Senkowski and Gallinat 2015). Yet, there has been a lack of preclinical work that identifies the precise manner in which GABAergic neurotransmission in the PFC regulates separable aspects of cognition. The experiments described in this Thesis sought to elucidate how PFC GABAergic transmission is involved in regulating working memory and attention, two core cognitive processes that are known to be affected in schizophrenia and other conditions associated with deficient PFC GABA function.  2  1.1 GABAergic transmission in the PFC The majority of inhibition in the mammalian central nervous system, including in PFC areas, is the result of GABA release from interneurons. The following section provides a brief primer on GABAergic neurotransmission and interneuron diversity. Much of the information holds true throughout the brain, but specific information for PFC, such as selective protein expression, is given where available. 1.1.1 Synthesis and release of GABA GABA is synthesized from L-glutamic acid by the enzyme glutamic acid decarboxylase (GAD). In both rodents and primates, there are two isoforms that differ in molecular weight and are transcribed from two separate genes (Erlander et al. 1991). These two isoforms are co-expressed in the majority of GABAergic cells, though different cell types may exhibit differential levels of expression of the isoforms (Esclapez et al. 1994). The 65 kilo-Dalton variant, GAD65, is localized in synaptic terminals, and is thought to be responsible for the majority of GABAergic synaptic transmission (Erlander et al. 1991; Feldblum et al. 1993). GAD67 is localized throughout the cytoplasm (Esclapez et al. 1994) and is thought to produce GABA for more general metabolic functions, in addition to synthesizing GABA that functions a trophic factor during development (Chattopadhyaya et al. 2007) or in regulating redox reactions during oxidative stress (Lamigeon et al. 2001). However, deletion of the GAD65 gene produces unchanged brain GABA levels (Asada et al. 1996), whereas loss of GAD67 is associated with a 90% decrease in brain GABA concentrations (Asada et al. 1997). These results suggest that GAD67 may account for the majority of GABA synthesis in the brain, or may be able to partially compensate for losses in GAD65 function (Wu et al. 2007). Interestingly, mice that develop without GAD65 are viable but have increased anxiety (Kash et al. 1999) and increased 3  susceptibility to seizures (Kash et al. 1997). In addition to dramatically decreased GABA levels, global GAD67 knock-out mice die after birth due to cleft palate malformation, though with no obvious alterations in brain structure (Asada et al. 1997). More recent studies have examined region- or cell-type specific knock-out of GAD67 and have found evidence for decreased GABA concentrations and impaired inhibitory neurotransmission (Brown et al. 2015; Fujihara et al. 2015; Obata et al. 2008).  The expression of the GAD enzymes is activity-dependent, likely providing a homeostatic feedback mechanism for balancing changes in excitatory neurotransmission (Lau and Murthy 2012). . Vesicular GABA transporter (VGAT) is responsible for transporting GABA into synaptic vesicles. VGAT knock-out mice show perinatal lethality with hunched posture and cleft palate (Wojcik et al. 2006). 1.1.2 GABA Receptors There are two main families of receptors for GABA. GABAA receptors are ionotropic receptors that are responsible for the majority of fast inhibition in the brain. They are heteropentameric fast-acting channels that mediate synaptic inhibition by permitting influx of negatively charged chloride ions (Cl-) (Olsen and Sieghart 2009; Sigel and Steinmann 2012). Nineteen different subunits for the GABAA receptor (α1-6, β1-3, γ1-3, δ, ε, θ, π and ρ1-3) have been identified, with the majority of GABAA receptors composed of 2 α, 2β and 1 γ or δ subunit. Throughout neocortex, αl, β2 and γ2 subunits form the most frequent GABAA receptors and are highly expressed in all cortical layers (Pirker et al. 2000; Sieghart and Sperk 2002). On the other hand, α2 subunits are more strongly expressed in outer cortical layers, whereas α3 and 5 subunits are localized predominately in deep layers (Fritschy and Mohler 1995; Wisden et al. 1992). The subunit composition of GABAA receptors affects physiological and pharmacological properties 4  of the receptor (Rudolph et al. 1999; Vicini 1991). Thus, differential spatiotemporal expression of GABAA receptor subunits may give rise to differences in microcircuit function in different cortical layers, or in the maturation of physiological function during development (Fritschy et al. 1994; Hashimoto et al. 2009). Further, certain GABAA subunits are more common in either synaptic (α1-3, γ2) or extra-synaptic sites (α4-6, δ), suggesting involvement in phasic versus tonic inhibition, and also that subunit composition may a play a role in sub-cellular targeting of the receptor (Belelli et al. 2009; Brickley and Mody 2012).  GABAA receptors have several sites where binding of endogenous and exogenous neuromodulators, such as neurosteroids (Belelli and Lambert 2005), endocannabinoids (Sigel et al. 2011), benzodiazepines (Sigel and Lüscher 2011), barbiturates (Rho et al. 1996), and ethanol, can occur. Common GABAA receptor antagonists used experimentally can be competitive, directly displacing GABA from its binding site, non-competitive agents that affect functioning through the neuromodulatory sites, or blockers of the open channel, such as picrotoxin (Akaike et al. 1985; Inoue and Akaike 1988). The experiments outlined in this Thesis employ bicuculline, a competitive GABAA receptor antagonist (Akaike et al. 1985; Johnston 2013), to pharmacologically reduce PFC GABAergic transmission. Bicuculline is a broad-spectrum antagonist, reducing GABAergic transmission at most GABAA receptor types (Johnston 2013). Additionally, the GABAA agonist, muscimol, is used to induce suppression of neural activity in inactivation experiments, along with GABAB agonist, baclofen. GABAB receptors are G-protein coupled metabotropic receptors that are responsible for more prolonged and lasting responses to GABA release (Kaupmann et al. 1997). GABAB receptors are a heterodimer of GABAB1 and GABAB2 subunits (Kuner et al. 1999). Binding of GABA to GABAB receptors leads to G-protein activation which initiates intracellular signalling 5  cascades which may mediate hyperpolarization by opening of potassium channels or inactivation of voltage gated calcium channels (Doze et al. 1995; Gähwiler and Brown 1985). As such, GABAB receptors are often insensitive to modulators that affect GABAA function. Further, GABAB receptors are more commonly localized in presynaptically than GABAA receptors, where they play a central role in feedback regulation of GABA release (Kabashima et al. 1997; Misgeld et al. 1995) and in suppression of release of other neurotransmitters, including serotonin, norepinephrine (Bowery et al. 1980) and dopamine (Reimann 1983). 1.1.3 GABA reuptake After release into the synaptic cleft, GABA is removed from the synapse by the GABA transporter (GAT). GAT also mediates non-synaptic release of GABA in certain conditions. Following reuptake, GABA may be reloaded into vesicles by VGAT or degraded as part of the TCA cycle, which also is involved in synthesis of glutamate/glutamic acid which then serves as a precursor to synthesize more GABA (Rowley et al. 2012). 1.1.4 Cortical GABAergic neurons are diverse GABAergic interneurons comprise approximately 20-30% of cells in prefrontal cortex (Hendry et al. 1987; Tamamaki et al. 2003). These GABAergic neurons include a vast number of cell types that differ based on their connectivity, morphology and molecular factors, including differential protein expression (Ascoli et al. 2008; DeFelipe et al. 2013). However, within cortical regions including PFC, GABAergic interneurons may be grouped into three main categories based on non-overlapping characteristics that account for the majority of interneurons (Rudy et al. 2011; Xu et al. 2010). The first express the calcium binding protein, parvalbumin (PV) and have fast-spiking properties.  The second group express the neuropeptide somatostatin (SST) and predominately target dendrites and include Martinotti cells, as well as interneurons 6  with other morphologies (McGarry et al. 2010). The third are a varied class of neurons that express the serotonin 3A receptor and include neurons that express the neuropeptide VIP, as well as non-VIP expressing cells that include neuragliaform cells (Rudy et al. 2011; Ascoli et al. 2008).  Each of these classes may have further morphological, molecular or physiological specializations (Ascoli et al. 2008). For instance, PV expressing cells occur with at least two distinct morphologies that target different subcellular compartments and have different physiological properties (Povysheva et al. 2013), with PV basket cells innervating the soma and proximal dendrites and forming dense interconnections with other PV basket cells and PV chandelier cells selectively innervating axon initial segments (Williams et al. 1992). Likewise, varieties of SST-expressing interneurons also have distinct morphology and spiking properties (McGarry et al. 2010).    The morphological, anatomical, molecular and physiological specializations give rise to unique functional roles for each interneuron type in each class. For instance, interneurons that target the somata or axon initial segments, such as PV interneurons, are well placed to control the overall output of neurons, while SST interneurons that target dendrites may play a more precise role in regulating whether particular excitatory inputs lead to spiking. Interneurons that target other interneurons, such as VIP-positive cells, may exert disinhibitory effects over circuitry, by decreasing the level of inhibitory input from other interneurons onto pyramidal cells (Pi et al. 2013). These unique specializations influence whether and how specific interneurons types participate in the myriad of functions known to be regulated by GABAergic transmission, such as feed-forward or feed-back inhibition, gain control of circuitry, shaping of receptive fields and the timing of neuronal spiking (Roux and Buzsáki 2015). Interneurons are also known to be critically involved in synchronization of oscillatory activity that plays a role in cortical 7  computations by gating and modulating information flow. The activity of particular populations of interneurons shape these oscillations in different manners; for instance, fast-spiking PV basket cells are ideally suited to sustain faster oscillations such as those that occur within the gamma range (Buzsáki and Wang 2012). Though GABAergic cells comprise a minority of cortical neurons, certain GABAergic cell types, such as PV basket cells, display extensive arborisation, meaning that they may exert profound and far-ranging inhibitory effects (Kubota and Kawaguchi 2000; Wouterlood et al. 1995).  When both the strength and diversity of interneuronal influence over neural activity are considered, it is clear that insight into GABA function in the PFC is necessary to understand information processing in this region and also how PFC neural activity is involved in the generation of behaviors. In light of this prominent role in regulating PFC activity, is also not surprising that deficits in PFC GABA function are associated with neuropsychiatric disease, which will be the topic of the next section of this chapter. Although recent work has focused on identifying unique functional roles of each interneuron type in PFC function (Kamigaki and Dan 2017; Kim et al. 2016a), since certain aspects of GABA signalling, like synthesis enzymes and transporters, are common to most or all GABAergic cells and known to be altered in pathological states, it is also of interest to consider how more global impairments in GABA function impact upon circuit function and behavior. The work described in this Thesis investigated how PFC GABAergic neurotransmission as a whole regulates basic aspects of cognition, including working memory and attention, and how deficits in PFC GABA function affect neural activation in other brain regions.    8  1.2 Dysfunctional Prefrontal GABA Signalling is a Feature of Schizophrenia & Other Psychiatric Disorders Disrupted or diminished GABAergic transmission within PFC has been associated with several neuropsychiatric conditions including schizophrenia (Benes 2015; Lewis et al. 2012; Tse et al. 2015b), depression (Levinson et al. 2010), and autism (Dickinson et al. 2016), and is also observed over the course of aging (McQuail et al. 2015). The following section details pathological, genetic, physiological, functional imaging and animal model work that implicate deficits in PFC GABAergic functioning in these conditions. 1.2.1 Schizophrenia Schizophrenia is a severe and chronic psychiatric condition affecting 1% of the population. In addition to positive symptoms, which include hallucinations and delusions, and negative symptoms, such as avolition, anhedonia and social withdrawal, patients also present with a stable and lasting cognitive impairment, which can include disruptions in working memory (Lee & Park, 2005), attention (Demeter et al. 2013; Finkelstein et al. 1997), decision-making (Hutton et al., 2002; Heerey et al., 2008), and behavioral flexibility (Pantelis et al. 1999). While positive symptoms are known to correlate with increased mesolimbic dopaminergic transmission, considerably less is known about the etiology of negative symptoms and cognitive impairments. Furthermore, degree of cognitive impairment is a strong predictor of patients outcomes in schizophrenia (Green and Nuechterlein 1999), and unlike positive symptoms, cognitive dysfunction is not readily treated by currently available neuroleptics (Keefe et al. 2007). Thus, the study of cognitive impairment in schizophrenia is vital to our understanding of the disorder and in the development of new treatments. 9  1.2.1.1. Pathological findings.  In recent years, one of the main theories regarding the origin of cognitive dysfunction in schizophrenia has posited that decreases in prefrontal cortical GABAergic transmission may underlie many of these impairments (Benes 1995; Coyle 2004; Gonzalez-Burgos and Lewis 2012; Tse et al. 2015b).  The central support for this theory has been that changes in markers of GABA function within the PFC have been amongst the most reproducible pathological findings observed in schizophrenia. Studies examining both mRNA and protein expression of GAD67, the rate-limiting enzyme involved in production of GABA, have consistently revealed decreases in expression in post-mortem PFC samples from schizophrenia patients (Akbarian et al. 1995b; Curley et al. 2011; Guidotti et al. 2000; Hashimoto et al. 2008a; Hashimoto et al. 2008b; Lewis et al. 2008; Thompson et al. 2009; Volk et al. 2000). On the other hand, GAD65 expression is unaffected (Guidotti et al. 2000).  Importantly, reduced expression of GAD67 is not the result of decreased interneuron number, but a consequence of reduced cellular expression of GAD67 in a proportion of PFC interneurons (Akbarian et al. 1995b; Hashimoto et al. 2003; Volk et al. 2000). Decreased GAD67 expression is particularly observable in PV interneurons (Hashimoto et al. 2003), which are believed to play a key role in pathophysiology of the disorder. Expression of PV, which is important to function of this interneuron type through calcium-binding functions, is also reduced in these interneurons (Curley et al., 2011). However, SST (Fung et al. 2010; Morris et al. 2008) and calbindin, a calcium-binding protein frequently expressed in SST neurons (Beasley et al. 2002) are also decreased, with a significant correlation between SST and GAD67 levels in PFC in schizophrenia, suggesting further alterations in function in other interneuron types, including PFC SST interneurons in particular.  10  Decreases in GAD67 and other interneuron markers may be the result of a preceding decrease in excitatory input onto GABAergic interneurons, given that expression of GAD67, PV and SST are activity-dependent (Lau and Murthy 2012; Marty and Onténiente 1997; Philpot et al. 1997) and the numbers of excitatory post-synaptic sites on PV interneurons were decreased in PFC of schizophrenia patients (Chung et al. 2016). PFC GABAergic interneurons in schizophrenia also show decreased expression of the immediate early gene zif268 (Kimoto et al. 2014) which is a marker of neural activity and transcription factor that regulates expression of GAD67 (Szabó et al. 1996). Expression of the vesicular GABA transporter is also decreased (Hoftman et al. 2015), which could indicate deficient loading of GABA into synaptic vesicles and less GABA release overall. Taken together, these findings suggest decreased interneuron activity in the PFC in schizophrenia and corresponding decreases in GABA synthesis and release. In addition to changes in presynaptic GABA function, changes in post-synaptic receptor composition are also observed in PFC regions in schizophrenia. Post-mortem studies of PFC from schizophrenia patients revealed an overall increase in binding of radioactive-labeled muscimol, suggesting a slight but significant increase in numbers of GABAA receptors (Benes et al., 1996).  However, decreased expression of α1, α5 and β2 GABAA subunit mRNA is observed in layers III and IV, and increased expression of α2 is observed in layer II (Akbarian et al. 1995a; Hashimoto et al. 2009). Given that the majority of GABAA receptors in adult human brain consist of α1 and β2 subunits (Hashimoto et al. 2009) these changes may mean that a significant portion of the receptors in PFC of schizophrenia patients are altered, and may have corresponding changes in physiology and function. Further, the overall pattern of changes has been likened to a developmental disturbance in receptor expression, as α2 subunits are more 11  common in development and more highly expressed in patients, while α1 receptors, which are more prevalent in adulthood are reduced (Beneyto et al. 2011; Hashimoto et al. 2009). The normal developmental change in subunit composition in PFC is accompanied by a corresponding maturation in the physiology of receptors, with post-adolescent receptors expressing α1 subunit showing much faster deactivation kinetics in comparison to α2 receptors expressed prior to adolescence (Hashimoto et al. 2009; Hollrigel and Soltesz 1997), and is likely to be disturbed in schizophrenia. 1.2.1.2. Electrophysiological findings.  Electrophysiological experiments in patients showing abnormalities in oscillatory activity have also indirectly pointed towards dysfunctional PFC GABA signalling in schizophrenia. Cortical functions, including sensory processing and higher-order cognitive processes, are a product of co-ordinated patterns of activity groups of neurons. This activity often takes the form oscillations, which can be measured electrophysiologically, including in humans through the use of electro- or magneto-encephalography (EEG or MEG). Intact GABAergic interneuron function is thought to be critical for synchronization of oscillatory activity, with different populations of interneurons differentially regulating oscillations in particular frequency ranges (Chen et al. 2017; Kuki et al. 2015). In recent years, a key role for PV interneurons, which are fast-spiking and show rapid deactivation kinetics, has been shown for fast-paced PFC gamma oscillations (Buzsáki and Wang 2012). Stimulation of PV, but not SST, interneurons results in induction of gamma rhythms (Cardin et al. 2009; Sohal et al. 2009). In both control and schizophrenia patients, the amplitude of gamma oscillations and working memory performance were correlated with resting dorsolateral PFC GABA levels, suggesting that PFC GABA function influences gamma oscillatory strength and cognitive performance. Furthermore, both amplitude and performance 12  were decreased in schizophrenia in this study (Chen et al. 2014). A number of other studies have demonstrated reduced or perturbed gamma oscillations in schizophrenia patients, often with increased amplitude of resting or baseline gamma, with a failure to induce gamma oscillations while engaged in cognitive tasks (Basar-Eroglu et al. 2007). Selective reductions in gamma oscillatory power, but not other oscillatory activity, during cognitive control tasks have been observed in schizophrenia patients with corresponding impairments in task performance, again suggesting that gamma range activity is a feature of successful cognitive engagement  (Cho et al. 2006; Minzenberg et al. 2010). Furthermore, abnormal gamma oscillations are observed at the onset of schizophrenia symptomatology (Minzenberg et al. 2010), suggesting that they may be a stable trait associated with schizophrenia. In addition to gamma, PFC oscillations in other frequency bands are implicated in cognitive functioning and also altered in schizophrenia. In particular, slower oscillations, such as theta, are thought to play a role in top-down control and can modulate lower frequency oscillations, including gamma. Haenschel et al. (2009) found a reduction in alpha, beta, and theta band activity during encoding of working memory in schizophrenia patients. While they also found evidence for decreased gamma activity across working memory maintenance and decreased theta and gamma in retrieval, only the changes in slower oscillations during encoding predicted working memory performance in this study. In the n-back working memory task, successful performance with increasing cognitive load is associated with increased amplitude of theta oscillations, which was deficient in schizophrenia (Schmiedt et al. 2005). Further, decreased coupling of theta and gamma oscillations is observed in schizophrenia patients performing the n-back, supporting the idea that co-ordination of activity in several frequency bands is important for working memory performance (Barr et al. 2017). 13  When considered with the knowledge that oscillatory activity is known to be dependent upon intact interneuron function, these observations of suppressed or abnormal PFC oscillations across frequency bands in schizophrenia patients support the idea that reduced PFC GABA interneuron activity contributes to the disorder, especially when viewed in conjunction with post-mortem pathological studies of the PFC in schizophrenia. Further, abnormal oscillatory activity in schizophrenia patients is often observed particularly during cognitive engagement, suggesting that disrupted oscillations and underlying interneuron dysfunction in the PFC may contribute to their cognitive impairments. 1.2.1.3. In vivo imaging findings.  Levels of GABA in the PFC may be measured in vivo by proton magnetic resonance spectroscopy (MRS), which assesses total tissue levels of GABA or other neurochemicals. In vivo assessments of GABA and metabolite levels in schizophrenia have produced contradictory results, with findings that show increased (Kegeles et al. 2012; Ongür et al. 2010), decreased (Chiu et al. 2017; Marsman et al. 2014; Rowland et al. 2013; Rowland et al. 2016a) or unchanged (Chen et al. 2014; Goto et al. 2010; Tayoshi et al. 2010) GABA levels in PFC regions in schizophrenia. However, there are a number of methodological issues and inconsistencies between these studies, including small sample sizes, discrepant subject age ranges, illness duration and differences in medication status, and differences in the quality of the spectral data (Egerton et al. 2017). In particular, some medications administered to patients, including both antipsychotics and anticonvulsants, are known to affect GABA levels (Ongür et al. 2010). Furthermore, some of the studies reporting no change have observed correlations between GABA levels and working memory performance, suggesting that variations in GABA levels in patients may be related to the level of cognitive impairment. (Chen et al. 2014; Rowland et al. 2016b). The largest MRS study of schizophrenia patients to date (Rowland et al. 2016a) 14  found evidence that tissue GABA levels were decreased in medial frontal and cingulate regions selectively in elderly schizophrenia patients, a finding that had also been observed in an earlier MRS study (Rowland et al. 2013). Importantly, Rowland’s more recent study used scanning parameters that suppress noise from macromolecules, which can constitute up to 50% of the GABA signal in MRS (Aufhaus et al. 2013), removing a significant confound of earlier studies. These results are particularly intriguing since aging is associated with its own subtle decreases in PFC GABA levels (Grachev et al. 2001), and may reflect an additive effect with schizophrenia. While the outcomes of MRS studies are somewhat unclear with respect to cortical GABA concentrations in schizophrenia, it is important to note that this method only provides a measure of total tissue, not synaptic, GABA concentrations, and thus does not directly assess whether PFC GABAergic transmission itself is affected in schizophrenia. While MRS work provided information regarding total concentrations of GABA in the PFC in schizophrenia, more recent work with PET imaging using radioligands that bind to the benzodiazepine site of the GABAA receptor has begun to provide indirect evidence of perturbed PFC GABAergic transmission in patients. One recent study using this approach (Frankle et al. 2015) in conjunction with a GAT blocker found that patients had elevated levels of binding at rest, suggestive of increases in GABAA receptors similar to post-mortem work (Benes 1995).  However, while both controls and medicated schizophrenia patients had an upregulation of ligand-binding after administration of an antagonist of the GABA transporter, unmedicated schizophrenia patients showed no such upregulation (Frankle et al. 2015). Increased benzodiazepine-ligand binding following GAT antagonist administration has been observed in healthy populations before, and is thought to result from increased synaptic GABA (Frankle et al. 2009). The findings of this study suggest that PFC GABAergic neurotransmission is disrupted 15  in schizophrenia, particular in response to challenges or stressors, and represent an exciting avenue forward for in vivo quantification of GABA function in patient subpopulations. 1.2.1.4. Environmental risk factors.  Strikingly, many animal models of the schizophrenia that manipulate parameters related to risk factors for the disorder have also repeatedly found evidence of disrupted PFC GABA function. Environmental insults during the pre- or peri-natal period are associated with increased risk for developing schizophrenia (Mcdonald & Murray, 2000) and can include maternal infection, stress or malnutrition, and birth complications (Cannon et al. 2002) . Many of these factors result in disrupted GABA function that emerges after adolescence when administered to laboratory animals, mirroring the developmental trajectory observed in schizophrenia (Meyer and Feldon 2010). Prenatal infection is often modeled by activating the maternal immune system using the viral mimetic polyI:C or bacterial lipopolysaccharide during pregnancy. Maternal immune activation leads to decreased GAD67  and PV expression, impaired PV interneuron activity and altered gamma oscillations, and is also accompanied by schizophrenia-like cognitive impairment (Canetta et al. 2016; Meyer et al. 2008; Richetto et al. 2014). One mechanism by which these early developmental insults are thought to impact brain development is by disrupting maturation of hippocampal projections to prefrontal cortex which occurs during the early postnatal period and is important for establishing network activity in the developing PFC (Brockmann et al. 2011). In line with this, the neonatal ventral hippocampal lesion has also been shown to lead to PFC interneuron dysfunctions later in life (François et al. 2009; Tseng et al. 2008). Since maturation of PV and other interneurons occurs in an activity-dependent fashion (Chattopadhyaya et al. 2004), they may be particularly sensitive changes in input or network activity that occur as the result of adverse events during key periods of development. 16  During adolescence and early adulthood, a stressful life event or exposure to drugs are factors associated with instatement of positive symptoms that typically lead to diagnosis of schizophrenia. In particular, cannabis use is thought to be a factor that can lead to the onset of symptoms in predisposed individuals (Henquet et al. 2008). Repeated treatment with the cannabinoid THC during adolescence produces decreased GAD67 and PV expression, and several schizophrenia-related behavioral changes (Zamberletti et al. 2014). A more recent electrophysiological study found that a similar regimen produced disinhibition of PFC neurons and abnormal gamma oscillations and was reversed by pharmacologically augmenting GABAergic transmission (Renard et al. 2017). In light of these findings, it is possible that the deleterious effects of cannabis observed in individuals at risk for schizophrenia may also be mediated in part by disruptions in PFC GABAergic signalling.   1.2.1.5. Genetic risk factors.  In addition to environmental factors, schizophrenia has a substantial genetic component, with heritability estimate of 60-80% (Giusti-Rodríguez and Sullivan 2013). Genetic studies of schizophrenia have revealed an extremely complex, polygenic inheritance with hundreds of genes implicated, each contributing a small enhancement of risk for schizophrenia. Rare copy number variants of particular chromosomal loci, including 22q11.2, 1q21.1, 2p16.3 15q13.3, 16p11.2 have also been implicated in the etiology of the disorder, with larger effects size but less frequent prevalence (Birnbaum and Weinberger 2017; Xu et al. 2008). One possible route by which these diverse genetic factors may lead to schizophrenia is by a convergence on pathophysiologically-relevant pathways. In line with this idea, several genetic risk factors for schizophrenia are known to be involved in interneuron development or adult GABAergic inhibitory function.  17  Rare copy number variants associated with schizophrenia often involve deletion or duplication of chromosomal segments containing multiple genes, several of which may be related to neural function. Therefore, they can often increase risk for psychiatric disorder more profoundly than single genes. The 22q11 microdeletion causes a ~30-fold increase in schizophrenia risk (Xu et al. 2008), and is one of the most studied animal models of the disorder. Mice with this microdeletion show altered cortical interneuron migration (Meechan et al. 2012) and reduced PV expression (Steullet et al. 2017). The Dgcr8 gene is known to account for some of the psychiatric phenotypes associated with 22q11 microdeletion, with conditional knock-out of Dgcr8 in pyramidal cells leading to fewer inhibitory synapses and reduced PV cell number (Hsu et al. 2012).  The 22q11 loci also contains gene for proline dehydrogenase; decreased expression of this enzyme results in depletion of GABA from interneurons through accumulation of cytosolic proline that acts as a GABA mimetic (Crabtree et al. 2016).  An animal model of the 15q13.3 microdeletion has also found decreased interneuron activity at baseline with accompanying deficits in attention (Nilsson et al. 2016). This chromosomal region contains the gene for α7 nicotinic acetylcholine receptor, which is highly localized on interneurons. Deletion of this receptor in a mouse model led to decreases in PV and GAD67 that emerged in late adolescence (Lin et al. 2014), partially accounting for decreased inhibition in 15q13.3 mice. Polymorphism in genes for neurotrophins, such as brain-derived neurotrophic factor (BDNF), and fibroblast growth factors (FGF), have been associated with several psychiatric diseases, including schizophrenia (Di Re et al. 2017; Gratacòs et al. 2007; Neves-Pereira et al. 2005; O'Donovan et al. 2009; Verbeek et al. 2012). Along with its receptor, trkB, BDNF regulates synapse formation and maintenance in development and adulthood (Kuipers and Bramham 2006; Luikart and Parada 2006). Reduced expression of the trkB is observed in 18  dorsolateral PFC of schizophrenia patients (Hashimoto et al. 2005; Weickert et al. 2005), while mice with reduced trkB have reduced PV, SST and GAD67 (Hashimoto et al. 2005), suggesting that reduced BDNF signalling may contribute to altered presynaptic GABA function in schizophrenia patients. Similarly, mice with deletion of another neurotrophin that has been associated with schizophrenia, FGF-14, have decreased expression of GAD67 in PFC (Alshammari et al. 2016). Neuregulin and its receptor erbB4 are proteins that direct migration of interneurons during development (Fazzari et al. 2010), and have been associated with schizophrenia risk in several studies (Jagannath et al. 2018). Deletion of the erbB4 receptor from PV neurons results in deficient GABA release in the PFC, and corresponding PFC disinhibition (Wen et al. 2010).  Finally, genetic polymorphism and reduced protein expression of dysbindin-1, a protein involved in trafficking of synaptic proteins, have repeatedly been observed in schizophrenia (Larimore et al. 2011; Riley et al. 2009; Tang et al. 2009) and dysbindin deletion leads to PFC disinhibition in mice (Jentsch et al. 2009) . Intriguingly, this PFC disinhibition was recently shown to result from decreased exocytosis of BDNF that consequently resulted in less inhibitory input onto pyramidal cells (Yuan et al. 2016). Such findings highlight the notion that dysregulation of proteins involved in neurodevelopment and synaptogenesis may converge on altered GABAergic transmission. 1.2.1.6. PFC NMDA receptor hypofunction.  Schizophrenia is also associated with disruptions in NMDA receptor transmission that are observed within PFC.  This hypothesis initially emerged from the observation that NMDA antagonists induce cognitive and behavioral changes relevant to both cognitive dysfunction and positive and negative symptoms of schizophrenia in control populations (Adler et al. 1999; Krystal et al. 1994) and cause exacerbation or instatement of symptoms in schizophrenia patients (Lahti et al. 1995). More 19  recently, it has been supported by genetic studies that show association of genes for NMDA receptor subunits with schizophrenia (Weickert et al. 2013; Zhao et al. 2006). Finally, post-mortem studies pathological studies of the PFC in schizophrenia show altered expression of NMDA receptor subunits (Weickert et al. 2013).  Intriguingly, systemic administration of NMDA antagonists have been observed to produce cortical disinhibition through selective effects on interneurons. Reductions in NMDA receptor activation on interneurons may decrease interneuron activity, in turn leading to pyramidal cell disinhibition and glutamate efflux (Homayoun and Moghaddam 2007). PV interneuron-specific deletion of the GluN1 NMDA receptor subunit, which leads to a loss of NMDA receptor expression, causes schizophrenia-like phenotypes, including deficits in short-term memory and altered gamma oscillations (Belforte et al. 2010). Acute or repeated exposure to ketamine decreases GAD67 and PV expression in animal models (Behrens et al., 2007; Amitai et al., 2012; Thomases et al., 2013). Together, these findings suggest that deficits in PFC NMDA receptor function observed in schizophrenia may contribute to cognitive impairments via disruptions in interneuron function. On the other hand, NMDA receptors are also localized on pyramidal cells, and these receptors have been shown to support PFC delay period activity (Wang et al. 2013). Thus, studies that contrast the effects of NMDA and GABA hypofunction in PFC are important to contribute the relative contribution of each to cognitive dysfunctions observed in schizophrenia and other psychiatric disorders. 1.2.1.7. Oxidative Stress.  Oxidative stress appears to be one potential pathway in which genetic and environmental risk factors for schizophrenia may converge to alter interneuron function in schizophrenia. Since PV neurons are highly active and energetically demanding, they are particularly vulnerable to oxidative stress from reactive oxygen species generated by high 20  levels of mitochondrial activity that supports their energetic need (Kann et al. 2014). Increased expression of a marker of cellular stress was observed in anterior cingulate in mouse models of the 22q11.2 1q21.1 15q13.3 copy number variants, decreased NMDA receptor function, neonatal ventral hippocampal lesion, and a two-hit environmental model consisting of prenatal immune activation and adolescent stress (Steullet et al. 2017). These increases in oxidative stress were accompanied by a decrease in PV expression in interneurons and a loss of the perineuronal net associated with PV interneurons, which regulates their plasticity and is also protective against cellular stress (Steullet et al. 2017). At present it is not clear how these diverse factors converge upon enhanced oxidative stress; nevertheless it appears that a dysfunction in PV interneurons is a common consequence of these insults. Furthermore, increased oxidative damage has been suggested in PFC of schizophrenia patients through MRS studies that show altered balance of NAD+/NADH (Kim et al. 2017). Collectively, these findings show that oxidative stress could represent a major pathophysiological pathway in schizophrenia, and one that leads directly to PFC GABAergic interneuron dysfunction. When viewed together, the above-mentioned findings provide strong support for the idea of a deficit in PFC GABA function in schizophrenia, drawing upon pathological, electrophysiological, and in vivo imaging work. Notably, many risk factors for schizophrenia, whether genetic or environmental, also lead to GABA dysfunction in the PFC when administered in animal models, further supporting this idea. These findings raise the possibility that deficient PFC GABA hypofunction may be a common final pathway for the diverse risk factors associated with schizophrenia, perhaps via enhanced oxidative stress. While it seems deficiencies in PFC GABA signalling are likely to be present in at least some schizophrenia patients, and though decreased PFC GABA has been correlated with cognitive impairment in a number of studies, 21  preclinical studies that directly probe PFC GABAergic regulation of separable aspects of cognition are still crucial to establish a causal role for PFC GABA in these cognitive impairments.  1.2.2 Depression Depression is one of the most commonly occurring psychiatric disorders, with notable symptoms including low mood, anhedonia and changes in energy levels. The symptoms of depression are accompanied by cognitive impairments, particularly a decrease in cognitive flexibility (Airaksinen et al. 2004), and also deficits in attention (Koetsier et al. 2002), working memory (Pelosi et al. 2000), and decision-making (Chamberlain and Sahakian 2006). In recent years, it has been observed that patients with depression have deficiencies in PFC GABA signalling that appear to result primarily from decreased SST interneuron function. The idea that PFC GABAergic signalling could be impacted in depression first came from MRS studies of PFC regions which have shown a relatively stable decrease in PFC total GABA concentrations across studies (Gabbay et al. 2012; Hasler et al. 2007; Price et al. 2009; Sanacora et al. 2006; Sanacora et al. 2004).   Rajkowska et al. (2007) found decreased size and density of GABAergic interneurons in the dorsolateral PFC of people who suffered major depressive disorder. Karolewicz et al. (2010) then found that PFC expression of GAD67 also is decreased in depression, a finding which has been confirmed by recent work (Scifo et al. 2018). Multiple studies have found evidence for decreases in SST expression in major depressive disorder (Fee et al. 2017; Povysheva et al. 2013; Sibille et al. 2011; Tripp et al. 2011). Post-synaptic GABAergic alterations in the form of reduced mRNA expression of α1, 3 and 4 and δ subunits of the GABAA receptor have also been observed in depressed suicide completers (Merali et al. 2004).  22  Both depressive disorders and schizophrenia are associated with cognitive impairments in overlapping domains, with abnormal PFC GABA function representing a potential common pathophysiological mechanism between the two conditions. However, the disorders may differ in the stability and course of cognitive impairment, and perhaps underlying changes in PFC GABA function. In schizophrenia, cognitive impairments are stable, regardless of symptom fluctuations (Hoff et al. 2005), and are observed in the prodromal period (Addington et al. 2017; Corigliano et al. 2014), suggesting that they are a stable trait associated with disorder. On the other hand, whether cognitive impairments are present between depressive episodes has been a matter of debate. Some studies find evidence that cognitive impairments are associated with the depressive state (Lee et al. 2012; Liu et al. 2002), while others have found evidence for a more stable cognitive impairment that is present even in remitted patients (Hammar et al. 2003; Paelecke-Habermann et al. 2005). Intriguingly, risk of relapse and severity of depression has been associated with the degree of cognitive impairment (Majer et al. 2004). If PFC GABA dysfunction is indeed related to these cognitive impairments, PFC GABA levels may be an important predictor of patient outcomes and therapeutic target across disorders, making studies that identify the precise consequences of dysfunctional PFC GABA signalling vital.  1.2.3 Autism spectrum disorder  Cortical imbalances in excitatory-inhibitory transmission have also been proposed as a pathophysiological mechanism in autism spectrum disorder (Foss-Feig et al. 2017; Rubenstein and Merzenich 2003), with recent work emphasizing deficits in PFC GABA function. Genetic studies of autism patients have found polymorphisms in GABA-related genes, including several subunits of the GABAA receptor (Kang and Barnes 2013). In vivo imaging has found evidence for reduced binding of radioactively-labelled muscimol and benzodiazepine ligands, suggesting 23  reduced GABAA receptor density in the anterior cingulate of patients with autism (Oblak et al. 2009).  A recent study examined PFC inhibition by comparing levels of GABA to glutamate and GABA metabolites using MRS, which the authors referred to as the ‘Inhibitory Index’ (Ajram et al. 2017). While this study found no changes in this parameter at baseline, men with autism and showed an increase in the Inhibitory Index following challenge with riluzole, a drug that is an allosteric positive modulator of the GABAA receptor (He et al. 2002c), which was not observed in controls. There also appears to be a selective reduction in numbers of PV neurons, but not other interneurons, in the PFC of people with autism spectrum disorder (Hashemi et al. 2017), and particularly the chandelier cell subtype (Ariza et al. 2018). Autism is a disorder with impaired communication, social deficits and repetitive behavior as distinguishing features. Neuropsychological studies of people with autism have revealed that these changes are accompanied by cognitive impairments that consist of impaired response inhibition, working memory (Luna et al. 2007), and behavioral flexibility (Ozonoff et al. 2004). Both schizophrenia and autism are viewed as neurodevelopmental disorders, and share many of the same risk factors. For instance, adverse events during the prenatal period are also associated with increased likelihood of developing autism, and many of the copy number variations that increase risk for schizophrenia are also associated with autism spectrum disorder (Sebat et al. 2009). While the two disorders differ substantially in terms of disease course and symptomatology, common cognitive impairments and PFC GABA alterations again highlight the possibility of a shared PFC GABAergic deficit contributing to cognitive impairment. 1.2.4 Age-related cognitive decline Aging is associated with the development of mild cognitive impairments in working memory (Holtzer et al. 2004), attention and cognitive flexibility (Ashendorf and McCaffrey 24  2008). A number of GABAergic alterations in the PFC are also observed over the course of age, suggesting that age-related alterations in PFC GABA function may be a contributor to these cognitive impairments. MRS studies show that GABA levels decrease, particularly in dorsolateral PFC, starting from middle age (Grachev et al. 2001). Numbers of interneurons and inhibitory synapses are lost over the course of aging in PFC in rodents and primates together with an overall loss of neurons (Peters et al. 2008; Soghomonian et al. 2010; Stranahan et al. 2012). However, expression of GABA transporter was decreased while GAD67 was increased, suggestive of increased synaptic GABA levels (Bañuelos et al. 2014). In aged rats, this was associated with an increase in overall inhibitory tone, but loss of activity-dependent inhibition and deficit in exploratory behavior (Bories et al. 2013). Electrophysiological work combined with transcranial magnetic stimulation has suggested that aged subjects had a loss of GABAA receptor transmission that was compensated by increased GABAB receptor signalling (Noda et al. 2017). Intriguingly, in a rat model of aging, GABAB receptor expression predicted behavioral flexibility but was inversely correlated with working memory performance (Beas et al. 2017; Beas et al. 2013), suggesting that the degree of GABAB receptor compensation may impact presentation of cognitive impairments. 1.2.5 Common risk factors and common cognitive impairments: a role for deficient PFC GABA signalling across psychiatric disorders? The work presented in this section has provided evidence for dysfunctional or deficient PFC GABAergic transmission in several neuropsychiatric disorders, including schizophrenia, depression and autism. These disorders appear to share etiological factors, including common genetic polymorphisms, such as copy number variations or genes implicated in development, and environmental risk factors, which appear to converge upon a PFC GABAergic impairment 25  (Steullet et al. 2017). Furthermore, these disorders share overlapping patterns of cognitive dysfunction, which appear to correlate with deficiencies in PFC GABA. Likewise, cognitive decline associated with non-pathological aging is also associated with dysfunctional PFC GABAergic transmission, further supporting the notion that abnormal PFC GABA signalling leads to cognitive impairments. In spite of this, there has been a surprising lack of preclinical work that directly assesses the impacts of deficits in PFC GABA transmission. As such, the specific cognitive functions impacted by diminished PFC GABA function in schizophrenia or other disorders are not entirely clear. The next section will discuss basic research to date that has examined how PFC GABA signalling regulates cognitive functions relevant to psychiatric disease, and has provided insight into how deficits in PFC GABA function may impact cognition in these conditions.  1.3 Preclinical Studies Assessing the Effects of Decreased PFC GABAergic Transmission In recent years, our group and other have attempted to address the paucity of work that examines the specific cognitive and behavioral alterations associated with decreases in PFC GABA function by examining the effects of pharmacologically reducing PFC GABA transmission via infusion of GABAA receptor antagonists into the medial PFC of rats.  In the following literature review on the effects of PFC GABAA antagonism and in the work described in this Thesis, infusions targeted the prelimbic (PrL) medial PFC unless specified otherwise. While rats have a much less developed PFC in comparison to primates, this region is thought to have some functional homology with dorsolateral PFC in humans, which is a key node in the orchestration of higher-order cognition, including working memory, decision-making and cognitive control (Uylings et al. 2003; Wise 2008). A key feature of both the past and present 26  studies on the effects of PFC GABAA antagonism has been an attempt to define the specific qualitative nature of the impairments that result from deficient PFC GABA function.  The experiments described in this Thesis will also compare the effects of pharmacological reduction of PFC GABAergic transmission in rats with other PFC manipulations, such as inactivation or NMDA receptor antagonism, in order to provide further insight into the nature of how PFC GABA function is involved in the regulation of cognition and behavior. A major methodological consideration of the present work is the selection of doses of bicuculline that are sufficient to perturb PFC GABAergic function and dependent components of behavior, but not induce seizure, as the agent can be used for this purpose at sufficient doses. The doses of bicuculline (12.5-50 ng) employed in the present studies and in past works from our group (Auger and Floresco, 2015; Auger et al. 2017a; Auger et al. 2017b; Enomoto et al. 2011; Piantadosi and Floresco, 2014; Piantadosi et al. 2016) and others (Paine et al. 2014; 2016; 2017) are considerably lower than those that induce ictal states in more seizure-prone areas like the hippocampus or amygdala (150-1500 ng, intra-cortical administration) (Rodrigues et al. 2004; Turski et al. 1985). To date, our group has infused these doses of bicuculline into the medial PFC of >300 rats and observed exceedingly few instances of seizure-like behavior (<1 in 50 animals), that occurred predominately in animals that were co-administered other drugs. In instances when seizure-like activity was observed, the animals were excluded from the analyses, similar to Pehrson et al. (2013). In general, it appears that administration of 1-3 infusions of 12.5-50 ng bicuculline is well-tolerated by animals, and does not produce any permanent change in behavior, as no differences between behavior on saline-treatment days occurring before or after bicuculline are observed (Auger and Floresco, unpublished observations).  27  1.3.1 Behavioral Flexibility A first study by our group revealed that intra-PFC administration of small doses of the competitive GABAA receptor antagonist, bicuculline (12.5-50 ng per hemisphere) causes changes relevant to both cognitive impairments and positive symptomatology associated with schizophrenia (Enomoto et al. 2011). One of the most remarkable findings of this study was that PFC GABAA antagonism led to impaired behavioral flexibility assayed with an operant set-shifting task. The task requires subjects to acquire performance of an initial visual cue discrimination on a first day of testing, but shift to a novel response strategy on a second test day. While PFC GABAA antagonism did not affect learning of the initial visual cue discrimination when antagonist was administered prior to the first test day or retrieval of this rule when administered on the second day, shifting to follow the new rule was impaired regardless of when PFC GABA antagonism occurred. However, more detailed analysis revealed that the types of errors made while shifting to the new rule differed depending on whether PFC GABAA receptors were antagonized on the first or second day. Animals receiving infusions of antagonist on the first day made more perseverative errors, meaning that they tended to follow the old rule of pressing the lever indicated by the cue light. Increased perseverative responding is also observed following PFC inactivation (Floresco et al. 2008), highlighting a key role for intact PFC function, and in particular, PFC GABAergic signalling, in the ability to disengage from past strategies. On the other hand, when the PFC infusion of GABAA antagonist was administered on the second day, subjects tended to make never-reinforced errors by pressing the lever in a manner that was not consistent with either of the two strategies that were reinforced, indicating difficulties in acquiring the new rule. Collectively the findings indicate that PFC GABAergic 28  signalling plays a role in both updating of task contingencies and formulating appropriate courses of action. These findings in the attentional set-shifting task mirror changes in behavioral flexibility in schizophrenia patients in that selective impairments in rule-shifting, but not initial discrimination learning, are frequently observed. Further, in the Wisconsin Card Sort task, which is highly dependent on set-shifting ability, patients show both perseverative and non-perseverative errors paralleling the pattern of results we observed in the operant set-shifting paradigm (Li 2004). Taken together, these findings from Enomoto’s study suggested that deficient PFC GABAergic transmission may play a central role in decreased cognitive flexibility observed in schizophrenia and raised the question of how PFC GABAergic transmission regulates other cognitive functions mediated by the frontal lobes. 1.3.2 Attention At around the same time, Paine & colleagues (2011) conducted a study assessing how intra-PFC administration of comparable doses of bicuculline (6.125-25 ng per hemisphere) impacts performance of the 5-choice serial reaction time task (5-CSRTT), which is often viewed as a gold-standard test of attention in rodents. In this task, animals must await appearance of a stimulus light above one of the five apertures of an operant chamber and make a nosepoke response in the illuminated chamber. Premature responses occurring before the appearance of the stimulus light are viewed as a measure of motor impulsivity. PFC GABAA antagonism led to decreased accuracy in the 5-CSRTT, without increasing motor impulsivity. On the other hand, this study found that activation of PrL GABAA receptors using the agonist, muscimol, led only to increased premature responses. Importantly, Paine (2011) also showed that intra-PFC administration of bicuculline led to increased neuronal activation in the medial PFC as indexed 29  by the immediate early gene, c-Fos, confirming that PFC GABAA antagonism produces disinhibition of PFC neural activity.   Soon after, Bita Moghaddam’s group conducted a study assessing how anterior cingulate administration of another competitive GABAA antagonist, SR95531, impacted performance of a 3-choice variant of the serial reaction time task (3-CSRTT) and also found decreased accuracy with unaffected premature responding (Pehrson et al., 2013). In this study, the effects of PFC GABAA antagonism were compared with PFC NMDA receptor antagonism, which caused subtle decreases in accuracy with increased premature responses. In this way, Pehrson et al. showed that though PFC GABAA or NMDA receptor antagonism may both cause some degree of disinhibition of PFC neural activity, their effects are dissociable, an idea that will be further interrogated in this Thesis. Pezze et al. (2014) further confirmed the findings of Paine et al.’s initial study by observing the effects of the GABAA channel blocker picrotoxin in PFC on 5-CSRTT and again found decreased accuracy without increases in motor impulsivity. The study also examined the effects of suppressing PFC neural activity through GABAA agonism using muscimol, which again increased motor impulsivity, this time with decreased accuracy that emerged from administration of higher doses of muscimol than that used in Paine’s 2011 study. Pezze et al. also found that PFC administration of picrotoxin did not affect the startle response or sensorimotor gating assessed using the prepulse inhibition paradigm. Although Japha and Koch (1999) showed that a smaller dose of picrotoxin did reduce prepulse inhibition in Sprague Dawley rats, Pezze’s findings in hooded rats suggest that the cognitive effects of PFC GABAA antagonism may be dissociated from changes in more basic sensorimotor processes in some instances.  30  Attentional impairments have consistently been observed in schizophrenia patients, particularly in performance of variants of the Continuous Performance Test (CPT), a widely used assessment of attention in humans (Finkelstein et al. 1997; Hsu et al. 2015; Liu et al. 2002). The CPT differs from SRTTs in that it includes trials that both include presentation of a stimulus or no stimulus presentation, requiring subjects to correctly identify lack of stimuli. The inclusion of both signal and no-signal trials allows for the dissociation of whether errors emerge from deficits in detecting the stimuli or false-positive responses in the absence of stimuli. Since SRTTs do not include non-signal trials, one issue that remains concerning the findings on the effects of PFC GABAA antagonism is whether the decreases in accuracy observed emerge from deficient signal detection or false alarm responses in non-signal trials. Such information would provide an important clarification to the nature of regulation of attentional processes by PFC GABA signalling, and will be the focus of Chapter 3 of this Thesis. 1.3.3 Decision-making Later work from our group and the Paine laboratory investigated how deficits in PFC GABA function impact cost-benefit decision making, using antagonism of PFC GABAA receptors with bicuculline. Paine et al. (2015b) showed the rats receiving PFC administration of bicuculline chose less advantageously in a rodent gambling task. Our group later showed that rats had impaired performance in a probabilistic discounting task following PFC GABAA antagonism. They also were less likely to expend more effort to obtain larger rewards in an effort-based decision-making task (Piantadosi et al. 2016).  All of the decision-making tasks mentioned are dependent on the ability to differentiate between larger and smaller rewards. Animals that receive PFC infusions of GABAA antagonists are able to discriminate between rewarded and non-rewarded options in an operant box, and also 31  do not have altered breakpoints in a progressive ratio task (Piantadosi et al. 2016), indicating intact basic processing of reward. However, they show a slight but significant decrease in their choice of larger versus smaller rewards in a series of operant control tasks (Piantadosi et al. 2016). Given the relatively subtle change in the probabilistic discounting task following PFC GABAA antagonism, it is likely that this result mostly stemmed from the impairment in the ability to discriminate between larger and smaller rewards. On the other hand, the impairments in both the rodent gambling task and effort-discounting task were more substantial, raising the possibility other deficits may have been at play. Taken together, these findings in relatively complex cognitive tests draw attention to the fact that the deficits in cognition observed following PFC GABAA antagonism may be the result of impairments in several underlying processes. The experiments described in this Thesis are concerned with understanding how decreased PFC GABA function impacts working memory and attention; two basic cognitive functions that would impact performance of these decision-making tasks by affecting the tracking of changing task contingencies or outcomes and the ability to stay on task. 1.3.4 Speed-of-processing During the time leading to the experiments described in this Thesis, one consistent feature of the findings relating to pharmacologically reduced PFC GABAA transmission was observation of increased latencies to respond in cognitive tasks and also slight but significant increases in trial omissions that emerged when time limits were placed on responding. In light of this, it has been tempting to speculate that increased latencies to respond following PFC GABAA antagonism may be related to decreased processing speed. This hypothesis is particularly interesting in regards to schizophrenia, because the decrease in performance in many cognitive tasks administered to patients is thought to emerge predominately from decreased speed-of-32  processing (Rodríguez-Sánchez et al. 2007). For instance, the digit symbol substitution task, which has amongst the largest impairment effect sizes in schizophrenia patients, first and foremost requires timely processing of stimuli that are presented rapidly (Dickinson et al. 2007). However, increases in latency that results from deficiencies in PFC GABA function may also result from disruptions in attention, motivation to obtain reward or other factors. The operant delayed-response working memory task employed in Chapter 2 may be particularly useful in discriminating between these possibilities, as it assesses latencies at distinct phases in the task, including trial initiation, initiation of the choice period and the choice itself. 1.3.5 Positive Symptomatology Aspects of the positive symptomatology of schizophrenia, in particular hallucinations, are known to arise from excessive mesolimbic dopamine transmission (Breier et al. 1997; Howes and Kapur 2009). Thus, while there are no direct methods for assessing the presence of hallucinations in animal models, assessment of mesolimbic dopamine function has been used as a proxy measure corresponding to vulnerability to hallucinate. In particular, behavioral responses to the dopamine-releasing drug, amphetamine, are used as an index of positive symptomatology in animal models, especially since schizophrenia patients are known to have elevated responses to amphetamines, which may also cause the onset or reinstatement of hallucinations (Lieberman et al. 1987). In our group’s initial study on the effects of PFC GABAA antagonism, PFC administration of GABAA antagonist increased both baseline and amphetamine-locomotion, suggesting increased mesolimbic dopamine activity and recapitulating enhanced behavioral responses to dopamine-enhancing agents observed in patients (Enomoto et al. 2011). To assess dopamine function more directly, recordings of ventral tegmental area dopamine neurons were conducted. These experiments revealed increased dopamine neuron burst firing, but unchanged 33  numbers of active dopamine cells, an outcome consistent with increased phasic, not tonic dopamine release (Enomoto et al. 2011). Likewise, positron emission tomography findings in patients also indicate increased synaptic dopamine concentrations in the striatum that are thought to reflect a phasic dopamine signal (Breier et al. 1997). Thus, in addition to mirroring enhanced sensitivity to dopamine agonists observed in patients, the data suggested that decreased PFC GABA function leads to changes in dopamine neuron activity that could contribute to positive symptomatology in schizophrenia. Importantly, though positive symptoms are believed to result from increased presynaptic dopamine function, the factors leading to this sensitization of mesolimbic dopamine function are unclear and the results of Enomoto’s study suggest that a deficit in cortical GABA function may be one of the underlying factors that contribute to this change. One prominent theory on the neuropsychological origins of positive symptoms in schizophrenia is that they emerge from abnormalities in salience attribution of stimuli. That is, stimuli or thoughts that should otherwise be irrelevant are attributed with increased importance or salience, while decreased salience is attributed to stimuli that convey important information. Our group found evidence for similar alterations in salience attribution in an auditory fear-conditioning paradigm (Piantadosi and Floresco 2014). Animals were first trained to lever press for sucrose reward, and then underwent a conditioning session in which four conditioned stimuli (CS+) consisting of a particular light and tone combination were delivered with a mild foot shock and four unconditioned stimuli with no averse consequence (CS-) were played. During a test session conducted in extinction, the two tones were played while access to the lever was given. Suppression of lever pressing was used as an index of conditioned fear. Remarkably, PFC GABAA antagonism prior to the test session resulted in increased expression of fear to the CS-, 34  which had never been associated with an averse outcome, and decreased conditioned fear to the CS+. This type of stimulus generalization has been described in schizophrenia patients tested in fear-conditioning assays in the laboratory (Jensen et al. 2008).  1.3.6 Negative symptomatology of schizophrenia  Similar to cognitive impairment associated with schizophrenia, the origins of negative symptoms of schizophrenia, which include decreased motivation, affective blunting, social withdrawal and anhedonia, are unknown. Our group observed that PFC GABAA antagonism caused animals to be less willing to expend more effort to obtain a larger reward in an effortful decision-making task (Piantadosi et al. 2016). Schizophrenia patients show decreased motivation to expend both cognitive and physical effort, and these changes in effort-discounting have been related to negative symptoms (Culbreth et al. 2016; Fervaha et al. 2013; Hartmann et al. 2015). Patients with major depressive disorder also show similar changes in effort discounting tasks (Treadway et al. 2012). Importantly, PFC GABAA antagonism spares basic aspects of processing reward, such as discrimination between rewarded and non-rewarded options and progressive ratio breakpoint, similar to spared basic reward processing in schizophrenia patients (Heerey et al. 2008). Thus, our findings in the effortful decision-making task raise the possibility that deficient PFC GABAergic transmission may also contribute to decreased motivation observed in schizophrenia patients. These findings are also particularly relevant to depression, where decreased motivation and anhedonia are thought to play a prominent role in the disorder (Pizzagalli, 2014).  Social withdrawal is another key feature of the negative symptoms of schizophrenia and one that has been shown to be a strong predictor of patient psychosocial function (Puig et al. 2008). Disinhibition of PFC using GABAA antagonism led to decreased social interactions in a 35  social preference test, and also a decrease in social memory (Paine et al. 2017). Given that reduced social behavior and altered social cognition is a cornerstone of autism spectrum disorder and are also observed to some degree in depression patients (Winograd-Gurvich et al. 2006), these findings also implicate cortical imbalances in excitation and inhibition in these aspects of social functioning.  1.3.7 PFC GABAergic regulation of working memory and the present work As indicated by the work discussed above, knowledge of the nature of PFC GABAergic regulation of cognition has grown substantially in recent years, with roles for PFC GABA signalling identified in behavioral flexibility (Enomoto et al. 2011), attention (Paine et al. 2011; Pehrson et al. 2013; Pezze et al. 2014), decision-making (Piantadosi et al. 2016), associative learning (Piantadosi and Floresco, 2014), social cognition (Paine et al. 2017) , and also behavioral changes relevant to psychiatric symptomatology. Yet, several questions concerning how PFC GABAergic transmission regulates cognition remain, particularly concerning PFC GABAergic regulation of working memory processes. For instance, though PFC GABAA receptor antagonism impaired performance of delayed-alternation working memory in primates (Sawaguchi et al. 1988), PFC administration of GABAA antagonist in rodents during the delay of a delayed-response radial maze task had no impact on working memory performance (Enomoto et al. 2011). The delayed-response radial maze task employed consisted of a single trial in which responding was self-paced. In contrast, many cognitive assays administered to human subjects consist of many back-to-back trials with responding restricted to a short time frame (e.g. Park and Holzman, 1992). As such, it is possible that PFC GABAergic transmission contributes to working memory selectively in such circumstances, when cognitive load is increased. Deficiencies in PFC GABA may also be more disruptive during other task phases, including 36  encoding and maintenance of working memory. Yet another possibility is that PFC GABA signalling may play a role in interference resistance; for instance, when past information interferes with current working memory processes, a phenomenon known as proactive interference.  With respect to attentional processes that may regulate working memory encoding, though a role for PFC GABA signalling in regulation of attention has been clearly established in several studies employing SRTs, the component processes involved are unknown. It may be that decreased PFC GABAergic transmission impact signal detection or formation of representations of stimuli, and/or they may contribute to aberrant responding to off-target stimuli, i.e. false alarm responses. Furthermore, speed-of-processing is an aspect of cognition that could impact performance of a variety of cognitive tasks, and it is not presently clear whether increased latencies following PFC GABA hypofunction result from processing speed deficits or other causes. In sum, several important questions regarding PFC GABAergic regulation of basic aspects of cognition remain. In recent years, chemogenetic and optogenetic techniques have increasingly been used to manipulate interneurons in more genetically and temporally precise ways, and have also added to the body of knowledge concerning how PFC GABA signalling modulates cognition. For instance, these studies have provided insight into the differential roles of PFC PV and SST neurons in regulating network activity and behaviors (Carlén et al. 2012; Kim et al. 2016a; Sohal et al. 2009). In spite of this, and as discussed above, knowledge of the broad ways in which PFC GABA activity regulates cognition is still lacking in key respects. This line of research remains important in light of the fact that alterations in PFC GABAergic function in schizophrenia and other disorders, such as deficits in GABA synthesis and release, may be present in a multitude of interneuron types. As such, studies that provide insight into the ways that broad deficits in PFC 37  GABAergic transmission impact upon cognition are timely. Further, these studies lay a foundation that may provide an important point of comparison as more precise techniques are increasingly used, and as their limitations are discovered. With these issues in mind, the experiments conducted in this Thesis sought to clarify whether and how PFC GABAergic transmission is involved in the regulation of core cognitive processes, including working memory, attention, speed-of-processing and interference resistance. One major outcome of this work is the notion that PFC disinhibition that emerges from deficient PFC GABA transmission may have impacts for brain-wide circuits, an idea we attempted to address my examining how PFC GABA hypofunction impacts neuronal activation in these circuits. In Chapter 2, we examined PFC GABAergic regulation of delayed-response working memory, using an operant delayed non-match to position task that bears similarity to delayed-response tasks administered to human subjects. We also confirmed that performance of this task is dependent on the PFC, and examined PFC NMDA gluatamatergic regulation of this working memory task. In Chapter 3, we examined whether attentional impairments arising from deficient PFC GABAergic transmission were the result of decreased ability to recognize signals or an increase in aberrant, false-alarm responses using a sustained attention task (SAT). Since the SAT employed was dependent on the ability to perform conditional discriminations, how PFC GABAA antagonism affects performance of visual and auditory conditional discrimination tasks was also assessed. 38  In Chapter 4, the question of how PFC GABA signalling contributes to interference resistance, specifically resistance to proactive interference, was investigated using a massed-trials variant of a traditional working/reference memory radial maze task, and compared to the effects of PFC inactivation.  The results of Chapter 4 prompted us to examine how PFC GABAA receptor antagonism affects neuronal activation of downstream circuitry in Chapter 5.  To this end, expression of c-Fos, a commonly used marker of neuronal activation, was assessed following PFC GABAA receptor antagonism was assessed in animals at rest, and also in animals trained and/or tested on the radial maze task employed in Chapter 4.   39  Chapter 2: Prefrontal cortical GABAergic and NMDA glutamatergic regulation of delayed responding  2.1  Introduction Working memory may be viewed as a collection of cognitive operations that facilitate the temporary representation, storage and manipulation of trial unique information that aids in guiding prospective behavior (Baddeley 1986; Goldman-Rakic 1991). Impairments in working memory are a cardinal feature of psychiatric illnesses including schizophrenia, and may underlie other abnormalities in cognition observed in the disorder (Goldman-Rakic 1994). These deficits can be observed prior to presentation of psychotic symptoms (Reichenberg et al. 2010), are stable throughout the course of the disorder, and can be observed in first degree relatives (Conklin et al. 2000), indicating that working memory impairments may be a putative schizophrenia endophenotype.  Furthermore, working memory and other cognitive impairments are strong predictors of patient outcomes (Green 1996) and no currently available treatment can consistently ameliorate these problems (Keefe et al. 2007).  As such, understanding the underpinnings of working memory deficits in schizophrenia is an important avenue in psychiatric research. Working memory is dependent on precise and co-ordinated oscillatory activity within the frontal lobes, particularly within the gamma range (30-100 Hz) (Howard et al. 2003). These oscillations are generated by glutamatergic pyramidal cell firing and regulated by GABAergic interneurons (Buzsáki and Wang 2012). In particular, fast-spiking parvalbumin (PV) GABA interneurons are thought to be necessary for initiating and maintaining gamma oscillatory activity (Cardin et al. 2009, Sohal et al. 2009), while NMDA glutamate receptors localized on PV cells have been shown to be important for both gamma rhythm induction and working 40  memory performance (Carlén et al. 2012). Schizophrenia patients have abnormal patterns of activity within the gamma range, particularly when engaged during tasks that require working memory and cognitive control (Chen et al. 2014; Cho et al. 2006; Haenschel et al. 2009; Minzenberg et al. 2010). This raises the possibility that abnormalities in GABA or NMDA glutamate function within the frontal lobes may play a central role in the pathophysiology of schizophrenia, and particularly in working memory impairment in schizophrenia. Indeed, prominent theories addressing the origin of schizophrenia have proposed both deficiencies in GABA and NMDA receptor glutamate signalling within the prefrontal cortex (Benes 2015; Coyle 2004; Lewis et al. 2012; Tse et al. 2015b). Decreases in the expression of markers of GABA function within the PFC are amongst the most reliable pathologies observed in post-mortem brains of individuals with schizophrenia. Reduced mRNA and protein levels of the GABA synthesizing enzyme, GAD67, are detected within multiple regions of the frontal lobes (Akbarian et al. 1995b; Curley et al. 2011; Guidotti et al. 2000; Hashimoto et al. 2008a; Thompson et al. 2009). These findings are complemented by recent in vivo work revealing disrupted PFC GABA transmission in medication-naive schizophrenia patients, using a benzodiazepine-specific radioligand (Frankle et al. 2015).  Finally, many different approaches to modelling schizophrenia in animals reduce PFC GAD67 expression and/or dysregulation of inhibitory transmission. These include pharmacological (Amitai et al. 2012; Morshedi and Meredith 2007; Thomases et al. 2013), genetic (Ji et al. 2009; Shen et al. 2008) and neurodevelopmental (François et al. 2009; Richetto et al. 2014; Tseng et al. 2008) models of the disorder.   On the other hand, the hypothesis that NMDA glutamate receptor hypofunction may be involved in schizophrenia arose from observations that NMDA receptor antagonists induce 41  phenotypes relevant to positive, negative and cognitive symptoms of schizophrenia in controls, and reinstate or exacerbate these symptoms in patients (Adler et al. 1999; Krystal et al. 1994). As such, NMDA receptor antagonists have been used to model aspects of schizophrenia in both rodents and non-human primates (for review see (Bondi et al. 2012). Recent work has revealed genetic association between NMDA receptor and other glutamate-related genes and schizophrenia (Zhao et al. 2006), as well as decreased expression of the GluN1 subunit of the NMDA receptor (Weickert et al. 2013), further supporting the idea that the NMDA receptor contributes to the pathophysiology of schizophrenia.  The above-mentioned findings suggest that deficits in GABAergic and/or NMDA glutamatergic transmission may contribute to working memory impairments observed in schizophrenia. This being said, preclinical studies directly assessing how diminished GABA and NMDA signalling within the PFC can affect working memory can provide additional insight into this issue. With respect to GABA transmission, recent work from our group and others have investigated how pharmacological reductions in GABAA receptor transmission within the rat medial PFC affect various cognitive functions.  Antagonism of PFC GABAA receptors alters attention (Paine et al. 2011; Pehrson et al. 2013; Pezze et al. 2014), cognitive flexibility (Enomoto et al. 2011), response latencies (Enomoto et al. 2011), cost/benefit decision-making (Paine et al. 2015b; Piantadosi et al. 2016), and salience attribution within the context of discriminative fear (Piantadosi and Floresco 2014).  Similarly, optogenetic silencing of PFC GABAergic interneurons induces similar deficits in attention and cognitive flexibility that are associated with perturbations in gamma oscillatory activity (Cho et al. 2015; Kim et al. 2016b).  Moreover, these effects often show qualitative similarities to cognitive abnormalities observed in schizophrenia.  In comparison, PFC GABAA antagonism did not impair performance of a 42  delayed-response version of the radial maze when administered prior to the choice phase of the task, although response latencies were increased, suggestive of speed-of-processing deficits (Enomoto et al., 2011). Since the delayed-response radial maze consists of only a single test trial in which responding is self-paced, it remains unclear whether deficiencies in GABA may impact upon working memory in conditions in which demands on speed-of-processing and attention are higher; for instance, when multiple trials must be completed in quick succession. Numerous studies have shown that systemic administration of NMDA antagonists impairs PFC-dependent working memory function (Chrobak et al. 2008; Cole et al. 1993; Smith et al. 2011), yet less work has examined how NMDA receptors specifically within the PFC regulate these functions.  Intra-PFC infusion of NMDA antagonists in rodents disrupts other PFC-dependent functions such as attention (Murphy et al. 2005; Pehrson et al. 2013) and behavioral flexibility (Stefani and Moghaddam 2005). In non-human primates, PFC NMDA receptors contribute to working memory related delay-period firing, as iontophoretic application of either non-specific or GluN2B-preferring NMDA receptor antagonists disrupts this activity (Wang et al., 2013). Similarly, antagonism of PFC GluN2B NMDA receptors in rats decreased working memory capacity measured by an odour span task (Davies et al. 2013a). However, at present, there is a lack of research addressing how antagonism of PFC NMDA receptors affect components of working memory that are taxed during delayed responding. With these issues in mind, the goal of the present study was to clarify the contribution of PFC GABA and NMDA receptor activity to working memory performance assessed with an operant delayed non-match to position (DNMTP) task. This task resembles delayed-response tasks used with human subjects in that they both require maintenance of a memory across a short delay, consist of a sample, delay and choice phase, and are often comprised of multiple trials 43  delivered in quick succession. Schizophrenia patients display reliable, delay-independent deficits in delayed-response tests of spatial working memory (Lee and Park 2005; Park and Holzman 1992).  In the present study we first used reversible inactivation of the medial PFC in rats to confirm this region mediates this form of delayed responding.  Subsequent studies investigated how pharmacological reductions of PFC GABAA and NMDA receptor activity (using both broad-spectrum and selective GluN2B antagonists) may affect delayed responding.   2.2  Materials and Methods 2.2.1 Animals Male Long Evans rats (250-300 g) were purchased from Charles River Laboratories (Montreal, Canada) and were initially group-housed upon arrival from the supplier. After 1 week, animals were single-housed and food-restricted to 85-90% of their free-feeding weight prior to beginning the behavioral training described below. Experiments were conducted in accordance with the Canadian Council on Animal Care and University of British Columbia Animal Care Committee. 2.2.2 Behavioral procedures  All behavioral tests were conducted during the animals’ light cycle.  All testing was conducted in standard operant boxes (Med Associates). Each chamber was equipped with a pellet dispenser and magazine with an infrared photobeam located within the magazine, two retractable response levers positioned to the left and right of the magazine, one stimulus light positioned above each lever and a house light. Initial training in the operant boxes initially consisted of training rats to retrieve 45 mg sweetened food reward pellets (BioServ, Frenchtown NJ) from the 44  magazine delivered on a VI 30 schedule. They were then trained over 2-4 days to press the right and left levers on a FR1 schedule for food reward at least 60 times within a 30 min period. This was followed by 3-6 days of retractable lever training. Over these 90 trial sessions (20 s intertrial interval), a pellet was dropped in the food magazine. A nosepoke response caused one of the two levers to extend (randomized in pairs), and rats were required to press it within 20 s of its insertion (otherwise scored as an omission).  Once a rat made <10 omissions, it moved to different phases of training on the main task.  2.2.3 Delayed Non-Match to Position Task The delayed non-match to position (DNMTP) task consisted of a sample phase and a choice phase, separated by a variable delay (1-24s).  The basic procedure of a daily session was as follows; during the sample phase, the houselight was illuminated, one of two levers was inserted into the chamber (randomized in pairs) that coincided with illumination of the stimulus light above that lever.  Rats were required to press the lever to initiate the delay period. During the delay, all lights were extinguished and levers retreated. At the end of the delay, the houselight was again illuminated and the rat was required to nosepoke in the central food receptacle to initiate the choice phase. This nosepoke procedure was employed to reduce the tendencies for rats to use mediating strategies, such as placing themselves close to the correct lever during the delay period.  During the choice phase, both stimulus lights were illuminated, both levers were inserted into the chamber and the rat was required to press the lever opposite that presented during the sample phase to lever to obtain a single food pellet reward.  A daily session consisted of 100 trials, and the intertrial interval was 5 sec.  Failure to press the lever during the sample or choice phases or initiate a nosepoke at the start of the choice phase within 20 sec resulted in the chamber reverting to the intertrial state (all lights extinguished and levers 45  retracted) and the trial was scored as an omission, although there were no additional consequences for failing to respond.  Rats were trained 5-7 days/week. During the initial phase of DNMTP training, animals received 100 trials using a 1s delay and received food reinforcement both for making the nosepoke response at the end of the delay period and for making the correct choice. Once a rat made >85% correct choices and made fewer than 10 omissions on this version of the task for 2 consecutive days, it was trained on a similar version of the task, but where only correct choices were reinforced.  When a criterion of >85% correct for 2 days was achieved for this stage, longer delays were incrementally added (1 and 4 s, 50 trials each; 1, 4 and 8 s, 33-34 trials each, etc) until a rat reached the final stage of training on the full task, consisting of 20 trials at each delay (1, 4, 8, 16, and 24s). In order for a rat to advance to the next stage of training, it was required to display criterion performance at 4 of the 5 delays for 2 consecutive days (1s =85%, 4s = 75%, 8s= 65%. 16s =60%). Each animal learned at its own pace. Once an individual rat displayed criterion performance on the full task for 2 days as described above, it received surgery.  Rats required between 30-45 training sessions (including all versions of the task) to display criterion performance on the final phase of the task.  Following surgical recovery, a rat was retrained on the full task until again reaching criterion performance, after which, it received a mock infusion and drug infusions tests commenced.  2.2.4 Surgery Once animals achieved a criterion performance, they were given ad libitum food for 2-3 days and then subjected to surgery.  Anaesthesia was achieved with ketamine/xylazine (100/7 mg/kg) and bilateral 23-gauge stainless-steel guide cannulae were implanted to target the prelimbic region of the medial PFC using the following stereotaxic coordinates (flat skull: AP = 46  +3.4 mm; ML = +/- 0.7 mm from bregma; DV = -3.0 mm from dura).  Rats were then given ~7 days to recover from surgery before resuming behavioral training. They were given ad libitum food for 3-4 days post-surgery, after which time food restriction to 85-90% of their free feeding weight was resumed.  2.2.5 Drugs & microinfusion procedures Inactivation of the PFC was achieved using a cocktail of the GABA agonists baclofen and muscimol (100 ng each in 0.5 µl). Antagonism of PFC GABAA receptors was achieved using the competitive antagonist bicuculline methobromide (12.5 or 50 ng in 0.5 µl). These doses were chosen because they have been shown to be effective at altering forms of cognition when infused into the PFC, but are below the threshold for inducing seizures when infused in this region (Enomoto et al., 2011; Paine et al., 2011). NMDA receptor antagonism was achieved using the non-competitive antagonist MK-801 (3-6 g in 0.5 µl). The 3 g dose was chosen as it has been shown to alter cognition within this brain region in the past (Stefani and Moghaddam, 2005; Pehrson et al., 2013), the higher dose was administered as this initial dose did not reveal an effect. GluN2B-specific NMDA receptor antagonism was achieved with the selective antagonist Ro-25-6981 (2.5 g in 0.5 µl). This dose was chosen as it has been demonstrated to alter cognition in previous studies, with this dose maintaining selectivity for GluN2B vs GluN2A-containg NMDA receptors (Davies et al., 2013).  All drugs except Ro-25-6981 were dissolved in 0.9% saline; Ro-25-6981 was dissolved in saline containing 20% DMSO.  A within-subjects design was used for all experiments, in that all animals received infusions of vehicle and at least one dose of the drug, with separate groups of animals used to test the effects of each drug. For the experiment examining the effect of intra-PFC bicuculline, 47  separate groups of animals were used for each dose in order to minimize the likelihood of kindling epileptiform activity after repeated exposures to bicuculline.  For the experiment using MK-801, rats initially received counterbalanced infusions of vehicle or the 3 g dose.  These same rats were then retrained and then received another series of counterbalanced infusions of either vehicle or a 6 g dose of the drug. One or two days before their first microinfusion test day, rats received a mock infusion procedure prior to training, during which obdurators were removed from the guide cannulae, and replaced with stainless steel injectors for 2 min, without an infusion. They then received the first of two (vehicle and drug) counterbalanced microinfusion test days, i.e. a proportion of the animals received vehicle infusion on the first test day, while the remaining received drug.  Following the first test session, animals were re-trained daily in the task until they re-achieved criterion performance for at least 2 consecutive days.  Once criterion was re-achieved, rats received a second counterbalanced infusion of saline or drug.  Bilateral infusions were made through 30-gauge injectors extending 0.8 mm below the guide cannulae. Saline or drug were infused at a rate of 0.5 µl/75 s.  Following infusions, the injectors were left in place for 1 min to allow for diffusion. Rats were then placed back in their home cages for 10 min, after which, testing commenced.  2.2.6 Histology After the final test session, rats were sacrificed using CO2.  Brains were removed and fixed in a 4% formalin solution for at least 24 h before being frozen and sliced into 50 m sections and mounted onto gelatin-coated slides. The section were then stained with Cresyl Violet. Figure 2.1 shows the location of all acceptable infusion placements. 48   Figure 2.1. Locations of all acceptable placements within the medial PFC region for DNMTP experiments. Different symbols indicate the location of cannulae placements in each of the four different experiments conducted.   49  2.2.7 Data Analysis Accuracy, as defined by the percent correct trials in each session, was the primary measure of interest. The main analysis consisted of a comparison of the % correct responses at each delay length, factoring out trial omissions. For experiments using a single drug dose, these data were analyzed with a two-way repeated measures ANOVA with treatment and delay length as the within subject factors. For the intra-PFC bicuculline experiment where two doses were used in separate groups of animals, the data were analyzed with a mixed ANOVA with dose as a between-subjects factor and delay and treatment as within-subjects factors. For the intra-PFC MK-801 experiment, performance data were analyzed with a three-way repeated measures ANOVA, with treatment (drug or vehicle), dose and delay as three within-subjects factors.  The effect of each dose of MK 801 was compared to the most recent vehicle infusion test day. In each of these analyses, the main effect of delay was significant and will not be reported further. We also analyzed the number of trial omissions, and latencies to initiate to sample, nosepoke, and delay phases of the task. Number of omissions were analyzed using one-way ANOVAs. Latency data was analyzed using a two-way repeated measures ANOVA with treatment and phase (sample, nosepoke, and delay) as within-subject factors.    2.3   Results 2.3.1 PFC Inactivation Previous studies have revealed that permanent lesions of the medial PFC lead to substantial delay-independent impairments in working memory assessed with delayed-response tasks (Chudasama and Muir 1997; Mair et al. 1998).  To confirm that this form of delayed 50     Figure 2.2. PFC inactivation impairs DNMTP performance. All data in this and following figures are expressed as the mean +/- SEM. A, Inactivation of the medial PFC with the GABA agonists baclofen and muscimol led to a delay-independent decrease in accuracy in the DNMTP task. B, PFC inactivation had no significant effects on the number of omissions made during the test session. C, PFC inactivation did not affect response latencies in any of the phases of the trials. p<0.05 vs saline.            51  responding was dependent on intact neural activity in the medial PFC, an initial experiment examined how reversible inactivation of the medial PFC affects performance of the DNMTP task in a cohort of well-trained rats (n=8).  Analysis of these data revealed that reversible inactivation of the PFC led to a significant, delay-independent reduction in accuracy, as indicated by a significant main effect of treatment (F1,7=22.26, p<0.01) but no treatment x delay interaction (F4,28=0.27, not significant (n.s.); Figure 2.2A).  PFC inactivation had no effect on the total number of omissions (F1,7=0.80, n.s.; Figure 2.2B) or response latencies during any phase of the task (main effect of treatment F1,7=1.39, n.s.; treatment x phase interaction: F2,14=0.58, n.s.; Figure 2.2C).  These findings confirm that accurate performance of this DNMTP task is dependent on neural activity within the medial PFC. 2.3.2 Pharmacological Reduction of PFC GABAA transmission We next examined the impact of pharmacologically reducing PFC GABAA receptor transmission on DNMTP performance (Figure 2.3).  Intra-PFC administration of the GABAA antagonist, bicuculline induced a delay-independent reduction in working memory accuracy.  An overall analysis of the data using dose as a between-subjects factor and treatment (drug vs. saline) and delay as within-subjects factors yielded a significant main effect of treatment (F1,24=9.85, p<0.01) and a significant main effect of dose (F1,24=7.36, p<0.01), although the treatment x dose or the three way interaction did not approach statistical significance (both Fs <1.0, n.s.).  To further explore the dose-dependent effects of bicuculline on accuracy, separate, two-way repeated-measures ANOVAs were conducted on the data obtained from each dosing group. These exploratory analyses revealed that rats receiving the 50 ng dose of bicuculline (n=13) displayed seemingly delay-independent impairments in accuracy as evidenced by a significant main effect of treatment (F1,12=13.08, p<0.01) but no treatment x delay interactions  52   Figure 2.3. PFC GABAA receptor antagonism leads to delay-independent impairments in DNMTP performance. A, Infusion of bicuculline (50 ng) reduced DNMTP accuracy in a delay-independent manner. B, The 50 ng dose also increased trial omissions and (C) increased response latencies during the sample and nosepoke phase of each trial, but not during the choice phase. D, Intra-PFC infusion of a lower dose of bicuculline (12.5 ng) had no significant effects on accuracy, (E) trial omissions or (F) response latencies.   53  (F4,48=0.38, n.s.)  As displayed in Figure 2.3A, at the 1 s delay, rats displayed a subtle decrement in performance following this dose of bicuculline relative to their performance after control treatments, and this effect persisted across the other delay conditions.  In comparison, rats that receive the 12.5 ng dose (n=13) did not display a significant impairment in performance relative to saline treatments (main effect of treatment: F1,12=2.03, n.s.; treatment x delay interaction: F4,48=2.06, n.s.). Together, these analyses indicate that reducing PFC GABA transmission impairs working memory accuracy in a relatively delay-independent manner, and that this effect was more prominent in rats treated with the higher dose. With respect to other performance measures, intra-PFC administration of both doses of bicuculline increased trial omissions (main effect of treatment: F1,24=4.83, p<0.05; main effect of dose: F1,24=0.29, n.s.; treatment x dose interaction: F1,24=0.06, n.s; Figure 2.3B,E).  Further exploration of this effect revealed that the majority of these omissions occurred during the sample phase (saline = 0.7 +/- 0.5, bicuculline = 3.8 +/-1.5) and to a lesser extent, during the choice phase (saline = 0.2 +/- 0.1, bicuculline = 0.9 +/-0.3). Omissions during the nosepoke phase were extremely rare after both treatments.  Analysis of latency data obtained from rats treated with the 50 ng dose of bicuculline yielded a significant main effect of treatment (F1,12=6.171, p<0.05) and notably, a significant of treatment x phase interaction (F2,24=7.978, p<0.01).  Simple main effects analysis further unveiled that this dose retarded response latencies during the sample phase (p<0.05) and also slightly delayed the nosepoke response to initiate the choice phase (p<0.05), but did not affect latencies during the choice phase (Figure 2.3C).  In contrast, treatment with the lower dose of bicuculline did not impact upon latencies during any phase (all Fs < 1.3, n.s.; Figure 2.3F).  Taken together, these results indicate that in addition to disrupting accuracy, the intra-PFC infusions of the 50 ng dose of bicuculline increased response 54  latencies during the sample and end of the delay phases, but not during the actual choice phase.  The lack of effect during the choice phase render it unlikely that impairments in accuracy were due to animals experiencing longer delays between sample phases and their actual choice due to slower choice times. 2.3.3 Non-selective PFC NMDA glutamate receptor antagonism We next examined the contribution of PFC NMDA receptors to DNMTP performance. Administration of the non-competitive NMDA receptor antagonist, MK-801 (3 or 6 ug, n=9) led to dose-dependent effects of working memory accuracy. A three-way ANOVA of the accuracy data yielded a significant treatment x dose interaction: (F1,8=14.51, p<0.05).  Simple main effects analyses comparing performance after each drug dose vs the respective vehicle infusion further showed that the 6 g dose impaired accuracy significantly relative to saline infusions (p<0.01; Figure 2.4A).  In contrast, the 3 g dose did not affect performance (Figure 2.4C).  Furthermore, performance did not differ between the two vehicle infusion test days.  In this experiment, the treatment x delay interaction was not significant (F4,36=1.96, p>0.05), yet, visual inspection of the data clearly shows that treatment with the 6 g dose did not appear to impair performance at the 1 s delay, but did cause a decrement in accuracy when delays were 4-8 s.  These impressions were confirmed by exploratory, Bonferroni-corrected t-tests (1 s delay, t9=1.08; n.s.; 4 s delay, t9=2.73, p<0.05; 8 s delay, t9=5.30, p<0.01).    In contrast to the effects on accuracy, infusions of either dose of MK-801 did not significantly alter trial omissions (all Fs <1.1, n.s.; Figure 2.4B,E) or response latencies (all Fs<1.2, n.s. Figure 2.4 C,F)  relative to saline treatments.  From these results, we conclude that reducing glutamatergic NMDA receptor signalling within the medial PFC also impairs working memory functions required for accurate DNMTP performance.  However, unlike the effects of  55   Figure 2.4. PFC NMDA receptor antagonism also impairs DNMTP performance. A, Infusion of MK-801 (6 g) reduced DNMTP accuracy, with this effect being more pronounced after 4-8 s delays relative to 1 s delays. B, These treatments did not affect trial omissions or (C) response latencies. D, Infusion of a lower dose of MK-801 (3 g) had no significant effects on accuracy, (E) omissions or (F) latencies.    56  reducing PFC GABA transmission, reduced NMDA receptor activity appears to induce more delay-dependent disruptions in accuracy.   2.3.4 PFC GluN2B subunit-specific NMDA receptor antagonism  GluN2B subunit-containing NMDA receptors in the dorsolateral PFC of primates have been proposed to contribute to delay-period activity and working memory maintenance (Wang et al., 2013).  In light of these findings we explored the contribution of GluN2B containing NMDA receptors by investigating how prefrontal blockade of these receptors affects delayed responding on this DNMTP task. To this end, we administered the NR2B-specific NMDA antagonist, Ro-25-6981 into the medial PFC of a separate group of well-trained rats (n=10).  In contrast to the effects of non-selective NMDA receptor blockade with MK-801, infusions of Ro-25-6981 did not impair accuracy, and if anything, these treatments caused a slight, non-significant improvement in performance (main effect of treatment: F1,9=3.26, n.s.; treatment x delay interaction: F4,36=0.482, n.s.; Figure 2.5A). No effects of intra-PFC Ro-25-6981 were observed on the omission rate or any aspects of response latencies (all Fs<1.0, n.s.; Figure 2.5 B,C).  Thus, even though blockade of NMDA receptors impairs delayed responding (Figure 2.4), the results of this study suggest that this aspect of working memory functioning does not appear to be critically dependent on NMDA receptors containing GluN2B subunits.    2.4  Discussion The key findings of the present study are that both GABAA and NMDA receptor activity within the rat PFC modulate delayed-responding. PFC GABAA receptor antagonism and inactivation of this region via infusion of high doses of GABA agonists led to delay-independent  57       Figure 2.5. Prefrontal NR2B subunit-specific NMDA receptor antagonism does not disrupt DNMTP task performance. A, Intra-PFC infusion of the GluN2B subunit-specific NMDA receptor antagonist Ro-25-6981 (2.5 g) did not significantly affect accuracy on the DNMTP task. B, Intra-PFC Ro-25-6981 treatment did not affect trial omissions or (C) response latencies.   58  impairments in working memory accuracy in the DNMTP task, suggestive of disruptions in implementing information to guide accurate responding, as opposed to mnemonic aspects of task performance. In addition, reducing PFC GABA activity increased response latencies and omissions, attributable potentially to disruptions in attention or other task-related cognitive processes. On the other hand, non-selective antagonism of PFC NMDA receptors impaired performance in a more delay-dependent manner, without affecting omissions or latencies, suggesting that these manipulations may have disrupted processing of mnemonic information used to guide behavior. These later effects did not appear to be mediated by PFC GluN2B NMDA receptors, as GluN2B-specific antagonism had no deleterious impact upon performance. Taken together, these findings indicate that although both GABAergic and NMDA glutamatergic transmission are necessary for accurate delayed responding, they may contribute to distinct aspects of working memory processes. The DNMTP task is a well-validated test of working memory in rodents, with translational relevance to human delayed-response tasks (Dudchenko et al. 2013). Early work investigating the role of the PFC in the task revealed that relatively large, permanent lesions of the medial PFC led to delay-independent impairments in DNMTP accuracy and suggested that multiple aspects of task performance are strongly dependent upon intact PFC function (Chudasama and Muir, 1997; Mair et al., 1998). In the present study, we again confirmed that neural activity within the PFC is essential for efficient task performance, as more circumscribed reversible inactivation of this region impaired DNMTP performance in a manner similar to that induced by permanent lesions. Although delay-independent deficits are typically viewed reflect impairments in non-mnemonic aspect of the task, here they may reflect the fact that the PFC is necessary for a variety of aspects of task performance regardless of delay, such as working 59  memory encoding and retrieval, response inhibition and appropriate action selection.  In this regard, our subsequent pharmacological experiments suggest that different aspects of working memory may be differentially regulated by GABAergic or NMDA glutamatergic transmission. 2.4.1. Prefrontal GABA signalling and delayed-response working memory Previous work by our group and others has revealed a critical role for GABA signalling in regulating cognitive processes mediated by the frontal lobes, including attention (Paine et al., 2011; Pehrson et al., 2013), cost/benefit decision-making (Paine et al., 2015; Piantadosi et al., 2016) and behavioral flexibility (Enomoto et al., 2011).  The extent and manner to which PFC GABA signalling regulates working memory has been less clear.  For example, PFC GABAA antagonism did not impair efficient search behavior on a PFC-dependent delayed-response version of the radial maze task, although these treatments did increase choice latencies (Enomoto et al., 2011).  Note that in our previous study, delayed responding on the radial maze task utilized a relatively long delay (30 min) and choice behavior was self-paced. Thus, a main goal of the present study was to further explore how PFC GABA signalling regulates working memory under task conditions similar to those used with human patients: with short delays and many trials employed, placing higher demands on attention and speed-of-processing.  We observed that PFC GABA antagonism induced delay-independent deficits in working memory accuracy, in keeping with similar results obtained from primates (Sawaguchi et al., 1989).  Note that under these conditions, PFC GABA signalling would have been attenuated during the sample, delay and choice phases of each trial.  This is to be contrasted with our previous study discussed above, where similar manipulations immediately prior the choice phase did not affect delayed responding on a radial maze (Enomoto et al., 2011).  When these findings are viewed together, they suggest that normal GABA signalling within the PFC may be more 60  important for regulating mechanisms underlying encoding or maintenance of trial-unique information that is subsequently used to guide accurate responding, as opposed to retrieval of this information after a delay.  If this is indeed the case, it follows that, by perturbing the initial encoding of information, intra-PFC bicuculline would be expected to disrupt subsequent accuracy even at the shortest delay, resulting in delay-independent deficits in performance.  This in turn may be related to the involvement of PFC GABA transmission in attentional control mechanisms (Rao et al., 2000; Paine et al., 2011; Pehrson et al., 2013; Kim et al., 2016b) which are an integral component of working memory functions (Allen et al. 2014; Baddeley 2012). This latter idea finds support from the observation that infusions of the 50 ng bicuculline slowed response latencies during the sample phase, which may reflect a deficit in attending to task-relevant information during this phase of a trial.  Likewise, optogenetic silencing of PFC parvalbumin GABAergic interneurons also impaired attentional performance by increasing error rates and trial omissions, similar to what was observed here (Kim et al., 2016b).  In addition to disrupting working memory accuracy, reduced PFC GABA transmission increased response latencies during the sample and end-of-delay phases of the task, and also caused a corresponding increase in trial omissions.  This observation is in keeping with previous studies showing that these manipulations increase speed-of-processing times on a variety of tasks assessing multiple domains of cognition including cost/benefit decision making (Piantadosi et al., 2016) and cognitive flexibility (Enomoto et al., 2011).  This being said, it may be argued that these effects may reflect impairments in motivational or motoric processes.  Yet, even though intra-PFC bicuculline rendered animals slower to initiate during the sample phase and to make a nosepoke response to initiate the choice phase, these treatments did not increase choice phase latencies (i.e. a response that could result in reward delivery).  Furthermore, previous studies 61  have shown that pharmacological reduction of PFC GABA signalling does not alter selection of rewarded vs. non-rewarded option during simple spatial discrimination tasks conducted in operant chambers (Piantadosi et al., 2016).  These findings suggest that the ability to perform simple left/right discriminations necessary for the performance of the DNMTP task is not disrupted by these manipulations. Moreover, intra-PFC infusions of bicuculline did not alter motivated instrumental responding for food delivered on a progressive ratio schedule (Piantadosi et al., 2016).  In light of these observations, we find it unlikely that the impairments in working memory observed here following reduction of PFC GABAergic transmission are driven by non-specific impairments in motoric, motivational or spatial discrimination abilities. Instead, we propose that intact PFC GABA transmission facilitates direction of attentional resources enabling accurate encoding of information within working memory. 2.4.2. Prefrontal NMDA receptor signalling and working memory Systemic administration of MK-801 and other NMDA receptor antagonists such as phencyclidine and ketamine induce delay-independent impairments in delayed responding, with accompanying disruptions in motoric and reward processes that confound evaluation of the contribution of NMDA receptors to mnemonic aspects of the task (Smith et al., 2011). Here, we found that non-competitive antagonism of PFC NMDA receptors impaired DNMTP performance in a dose-dependent manner, without affecting response latencies or trial omissions.  When comparing these effects to those induced by GABA-A receptor antagonism, it is notable that the 50 ng dose of bicuculline induced a decrement in performance that was apparent even at the shortest delay.  In comparison, infusions of MK-801 induced their most pronounced effects at the 4 and 8 sec delays, whereas performance the 1, 16 and 24 s delays was relatively unaffected.  Thus, even though the statistical analyses of these data did not yield a significant treatment x 62  delay interaction, our impression is that PFC NMDA antagonism induced an impairment that was more dependent on the length of the delay. Lack of an effect at the longer delays may be the result of a floor effect, as performance of this cohort during these delays was only marginally better than chance following saline infusions. Thus, these results suggest that PFC NMDA receptor activity may play a more prominent role in facilitating maintenance of the mnemonic information across a delay, rather than the initial encoding of this information. This idea is supported by recent findings from non-human primates, demonstrating that NMDA receptor activity is necessary for delay-period firing during performance of an oculomotor delayed-response task (Wang et al., 2013). In this way, the results of this study indicate that working memory impairments induced by diminished NMDA receptor signalling appear to be distinct from the effects of both PFC inactivation and diminished PFC GABA function and are likely to driven at least in part by perturbations in maintenance of information within working memory systems that include the medial PFC. Although non-specific antagonism of NMDA receptors impaired DNMTP accuracy, intra-PFC administration of the GluN2B subunit-specific NMDA antagonist with Ro 25-6981 did not impair any aspect of performance on this task.  In contrast to this lack of effect, intra-PFC infusions of this dose of Ro 25-6981 has previously been shown to decrease working memory capacity in an odour span task (Davies et al., 2013) and these receptors have also been implicated in facilitating set-shifting functions mediated by the medial PFC (Dalton et al. 2011). Furthermore, local iontophoretic application of this compound disrupts delay period activity of PFC neurons, suggesting that this NMDA receptor subtype facilitates maintenance of information within working memory (Wang et al., 2013).  It is possible that a higher dose of this compound may have yielded an effect. We chose not to test a higher dose of this drug to 63  minimize the chance of losing selectivity for the GluN2B receptor, which could complicate interpretation of the data.  With this in mind, the lack of effect of intra-PFC GluN2B antagonism reported here is in line with studies assessing the effect of systemic administration of these types of antagonists in rats performing operant delayed match-to-position tasks.  In these studies, accuracy following systemic GluN2B blockade was either unimpaired (Higgins et al. 2005), or was only reduced by doses that substantially increased omissions and latencies (Smith et al., 2011), suggestive of non-specific effects on task performance.  One potential explanation for why GluN2B antagonism impaired certain aspects of working memory in some studies but not others may be attributable to task-related differences.  For example, operant-based delayed response tasks used with rats require maintenance of a single item of information on a particular trial, and during the choice phase, the subject is only presented with two options. In comparison, both the odour span task used with rats and the oculomotor delayed-response task used with primates require maintenance of numerous items within working memory over trials, with multiple incorrect responses possible in the choice phase.  These factors may place higher demands on cognitive control compared to the relatively simpler operant-based tasks.  As such, it is plausible that situations that tax working memory functions more heavily, such as those requiring greater control in the face of numerous distracting stimuli and/or integration of multiple items of information, may recruit GluN2B-containing NMDA receptors within the PFC to facilitate these functions.  Alternatively, GluN2B receptors may play a greater role in facilitating other executive functions mediated by the PFC, such as behavioral flexibility (Dalton et al., 2011).  However, other types of NMDA receptors may regulate more basic functions entailing maintenance of mnemonic information to guide responding after a delay.  64  2.4.3. Implications for working memory functions regulated by the PFC GABAA receptors are the primary mediators of fast synaptic inhibition within the frontal lobes. GABA release from fast-spiking parvalbumin cells in particular is known to be important for generating gamma oscillations that underlie working memory processes (Sohal et al. 2009; Carlén et al. 2012).  Thus, disinhibition of neural activity following antagonism of PFC GABAA receptors may result in disorganized and hyperactive output from PFC pyramidal cells, disrupted gamma oscillations and disruptions in other task-related patterns of firing that facilitate accurate encoding and maintenance of information within working memory. Indeed, neuronal activity within the PFC as indexed by both immediate early gene expression (Paine et al., 2011) and rodent PET imaging (Parthoens et al., 2015) is substantially increased following infusion of this dose of bicuculline within the PFC.  Paradoxically, systemic administration of non-competitive NMDA antagonists such as MK-801 also leads to disinhibited PFC neuronal activity and glutamate efflux within the PFC (Homayoun and Moghaddam, 2007). This disinhibition potentially arises via preferential blockade of NMDA receptors on GABAergic interneurons, that reduces their activity and is associated with a corresponding disinhibition of pyramidal cell firing and dysregulation of gamma oscillatory activity (Homayoun and Moghaddam, 2007; Carlén et al., 2012). Intriguingly, although both PFC GABAA and NMDA receptor antagonism may lead to disinhibitory increase in PFC activity, the present findings suggest that cognitive and behavioral consequences of attenuated activity at these receptors may differ. In this regard, it is likely that local infusions of MK-801 used in the present study would not only have blocked these receptors on GABAergic interneurons, but would also have affected receptors on pyramidal cells as well.  Thus, diminished NMDA receptor activity within the PFC appears to predominately affect the maintenance of information within working memory.  In comparison, 65  diminished activation of GABAA receptors may affect working memory encoding, potentially related to disruption of attentional functions that facilitate task performance.  2.4.4. Implications for schizophrenia Working memory impairments are considered to be a core aspect of cognitive dysfunction in schizophrenia. Meta-analyses have revealed that these impairments appear across different sensory modalities and do not worsen with increasing delay (Park and Holzman, 1992; Lee and Park, 2005). Moreover, visuospatial working memory processing is strongly affected (Lee and Park, 2005). In light of the delay-independent nature of these deficits, impairments in encoding have been proposed to be a key contributing factor to these types of deficits observed in the disorder (Hartman et al. 2003b). However, even when encoding is optimized by increasing stimulus strength or duration, working memory deficits in patients persist, indicating that working memory maintenance is also likely affected (Tek et al. 2002). Analyses of oscillatory activity in schizophrenic patients performing delayed-response tasks reveal that processing is abnormal during encoding, maintenance and retrieval (Haenschel et al., 2009).  With this in mind, deficiencies in both prefrontal GABA and NMDA glutamate signalling within the PFC have been proposed to contribute to cognitive deficits in schizophrenia (Gonzalez-Burgos and Lewis 2012; Krystal et al. 2003; Tse et al. 2015b) but how these deficiencies contribute to specific aspects of working memory impairment in schizophrenia has been unclear. The present results indicate that alterations in both PFC GABA and NMDA receptor signalling do indeed cause deficits in delayed-response working memory that resemble those observed in schizophrenia, and point to the specific aspects of the task in which they are implicated. Insufficient GABA signalling appears to reproduce the delay-independent deficits observed in schizophrenia, with similar accompanying attentional impairments and slowing of responding. 66  Likewise, reduced NMDA receptor function may specifically lead to impairments in working memory maintenance that appear in addition to encoding deficits. Taken together, these results support the notion that both diminished GABAergic or NMDA glutamatergic transmission may play a role in schizophrenia-related working memory impairment, and raise the possibility that an interplay of these abnormalities may contribute to frontal lobe dysfunction observed in the disorder.  Perturbations in glutamatergic “signals” encoding task-relevant information, combined with aberrant disinhibition in activity (i.e.; “noise”) via reduced GABAergic transmission may lead to a “cortical cacophony” that impairs attention for important stimuli while simultaneously perturbing maintenance and manipulation of information used to guide behavior (Floresco 2013). 2.4.5. Conclusions To summarize, the present study reveals that both GABAergic and NMDA glutamatergic transmission within the PFC play an integral role in facilitating working memory performance dependent upon the frontal lobes. Diminished GABAergic or NMDA glutamatergic tone leads to different patterns of impairments in the DNMTP task, suggesting that these receptors may regulate distinct aspects of the working memory process. Intriguingly, the NMDA receptors implicated in delayed-response working memory do not appear to contain GluN2B subunits in rodents. These data build upon previous work suggesting important roles for PFC GABA and NMDA glutamate signalling in modulating cognition, and clarify their roles in working memory functions underlying delayed responding. Further, they support the notion that disinhibition of frontal lobe activity, caused by either GABA or NMDA hypofunction, may be a primary contributing factor to deficits in working memory associated with schizophrenia.   67  Chapter 3: Regulation of sustained attention, false alarm responding and implementation of conditional rules by prefrontal GABAA transmission: comparison with NMDA transmission  3.1 Introduction Attention encompasses a collection of cognitive operations that influence the detection of sensory stimuli and whether they are selected for higher-level processing. These operations include selective attention, divided attention and sustained attention, or vigilance. Vigilance refers to the ability to remain in a state of readiness to detect and respond to incoming stimuli that appear rarely and unexpectedly (Parasuraman et al. 1998). Impairments in attentional processes, including vigilance, are a core cognitive feature of many psychiatric disorders, including attention deficit hyperactive disorder (Huang-Pollock et al. 2012), schizophrenia (Cornblatt and Malhotra 2001; Liu et al. 2002) and bipolar disorder (Hegerl et al. 2010; Marotta et al. 2015). Furthermore, the ability to direct attention is often a necessary component process for more complex cognitive operations, such as working memory (Gazzaley and Nobre 2012), and may, therefore, partially account for deficits in higher order cognition observed in these disorders.  Vigilance and other attentional processes are thought to be mediated by networks of brain regions that include the prefrontal and parietal cortices(Petersen and Posner 2012). Human patients with frontal lobe damage show impaired performance of the Continuous Performance Test (CPT) and other tests of vigilance (Alexander et al. 2005; Rueckert and Grafman 1996). These impairments often worsen in later trials of the session, suggesting a specific deficit in vigilance (Rueckert and Grafman 1996).  Furthermore, brain imaging studies have consistently shown that the right PFC is recruited during performance of tasks requiring high levels of 68  vigilance (Berman and Weinberger 1990; Cohen et al. 1992; Coull et al. 1998; Lewin et al. 1996; Pardo et al. 1991). In rodents, both lesions and reversible inactivation of the PFC impair attentional performance measured using the 5-choice serial reaction time task (5-CSRTT) and also increase premature responding(Chudasama et al. 2003; Maddux and Holland 2011; Paine et al. 2011; Pezze et al. 2014). Intact NMDA receptor signalling within the PFC also appears to facilitate attention, as blockade of these receptors in the anterior cingulate or infralimbic cortices disrupt performance of the 3- and 5- CSRTTs, respectively (Pehrson et al. 2013; Murphy et al. 2005).  Serial reaction time tasks have become popular for assessing attentional performance in rodents, in part because they permit monitoring of numerous behavioral variables, including measures of accuracy, premature and perseverative responding, latencies and omissions (Bari et al. 2008; Lustig et al. 2013). However, with respect to attentional accuracy, these tasks do not differentiate between whether the errors that occur are attributable to  deficits in signal detection (i.e. failing to see a signal light) or to erroneous responses that occurs when no signal is present (i.e. false alarms). Sarter and colleagues developed and validated a sustained attention task (SAT) conducted in operant chambers that bears resemblance to CPTs administered to human patients, in that it includes both signal and non-signal trials, allowing for the dissociation of deficits in signal detection and false alarm responses (McGaughy and Sarter 1995). Diminished detection of a visual stimulus during signal trials reflects impaired signal detection, whereas responding as if the signal light was present during a non-signal trial constitutes a false alarm. Further work from the Sarter group showed that vigilance, as assayed with the SAT, is dependent upon intact function of the PFC (Miner et al. 1997) and cholinergic projections from the basal forebrain (McGaughy et al. 1996; McGaughy and Sarter 1998; Turchi et al. 1996; Turchi and 69  Sarter 1997). Perturbations in PFC cholinergic signalling, induced using 192 IgG-saporin lesions, are associated with deficits in signal detection, while false alarm responding is unaffected (Gill et al. 2000; McGaughy and Sarter 1998). In addition to NMDA glutamatergic and cholinergic transmission, recent work by our group and others has shown that intact PFC GABAergic transmission also serves a critical role in several cognitive processes mediated by the frontal lobes. These include discriminative fear conditioning (Piantadosi and Floresco, 2014), decision-making (Paine et al. 2015; Piantadosi et al. 2016), behavioral flexibility (Enomoto et al. 2011) and, of particular relevance to the present study, attention.  Studies from a number of groups have shown that antagonism of PFC GABAA receptors leads to attentional impairments in both the 3-CSRTT and the 5-CSRTT, while premature responding is unaffected (Paine et al. 2011; Pehrson et al. 2013; Pezze et al. 2014). However, what remains unclear from these findings is whether these attentional deficits arise from impaired signal detection and/or increased false alarm responding. Thus, the primary goal of this study was to clarify the specific contribution of PFC GABAA signalling in mediating attentional accuracy, utilizing the well-validated SAT described by Sarter and colleagues (McGaughy and Sarter 1995).  As noted above, antagonism of NMDA receptor in the PFC also disrupts performance of the 3- and 5-CSRTT (Murphy et al. 2005; Pehrson et al. 2013), but whether these effects are due to deficient signal detection or an elevated false alarm rate is not currently known.  The effects of NMDA receptor blockade in the PFC using non-competitive antagonists are complex, as these treatments attenuate excitatory glutamatergic signalling on PFC pyramidal neurons, but can also cause disinhibition of PFC networks (Homayoun and Moghaddam 2007). This may occur through actions on inhibitory GABAergic interneurons; reductions in excitatory NMDA receptor 70  activity onto interneurons that decrease their output could result in pyramidal cell disinhibition and increased glutamate efflux (Homayoun and Moghaddam 2007). In contrast, PFC GABAA receptor antagonism likely exerts disinhibitory effects through direct action on pyramidal cells (Lodge 2011). Interestingly, the results of Chapter indicate that even though PFC NMDA or GABAA receptor antagonism may both disinhibit PFC neural network activity, they produce distinct types of deficits on a test of working memory in rats. Thus, we also examined how intra-PFC or systemic antagonism of NMDA receptors impacts performance of the SAT, in order to compare how these how PFC NMDA and GABAA receptors may differentially influence attentional processing.  In addition to assessing signal detection and vigilance, successful performance of the SAT requires animals to make conditional discriminations, wherein a correct response entails making different actions based on the presence or absence of a visual stimulus.  In light of this consideration, we also examine how antagonism of PFC GABAA receptors affects performance of visual conditional discriminations, using a modified version of the SAT that minimized the demands on attention, as well as a conditional discriminations using different auditory stimuli.  3.2 Methods 3.2.1 Animals Male Long Evans purchased from Charles River (Montreal, Canada) were used in five experiments. Rats were group-housed upon arrival from the supplier. After 1 week of acclimatizing to the colony, animals from Experiments 1-3 were single-housed and food-restricted to 85–90% of their free-feeding weight prior to beginning training in the SAT 71  described below. Because animals in Experiments 4 and 5 were trained on tasks that required considerably fewer training sessions than the SAT, these animals underwent the surgical procedures described below first and were allowed to recover for approximately one week before commencing training in the visual and auditory discrimination tasks, respectively. Experiments were conducted in accordance with the Canadian Council on Animal Care and University of British Columbia Animal Care Committee. 3.2.2 Behavioral procedures All training and testing occurred during the animals’ light cycle. Testing was conducted in standard operant boxes (Med Associates). Each chamber was equipped with a house light, pellet dispenser and magazine with an infrared photobeam located within the magazine, two retractable response levers positioned to the left and right of the magazine. Three stimulus lights were situated above each lever and the magazine, although only the central light was used in these experiments. Initial training in the operant boxes consisted of training rats to press either the left or right lever (one lever per day) to retrieve 45 mg sweetened food reward pellets (BioServ, Frenchtown NJ) on an Fixed Ratio 1 (FR1) schedule. Rats underwent FR1 training on both levers for 2-4 days, until they pressed at least 60 times during a 30 minute session. After this, rats were given retractable lever training. During the retractable lever training task, the houselight was illuminated every 40 s to signal the start of the trial, after which the left or right lever (randomized in pairs) was extended into the chamber. Animals had 10 s to press on the extended lever in order to receive 1 food reward pellet. Sessions consisted of 100 trials per day. Once animals had reached a criterion of 10 or fewer trial omissions, they were advanced to training in one of the three tasks below.  72  3.2.3 The sustained attention task Rats were first trained to discriminate between signal and non-signal events. During signal trials, the stimulus light located above the magazine was illuminated for 1 s. Two seconds after the light was presented, both levers were extended into the chamber and the rat was required to press on a designated lever (left or right) to obtain food reward. The lever to be pressed on signal trials was kept constant throughout training and testing, and was counterbalanced across animals. During non-signal trials, the stimulus light was not illuminated and when the levers were extended into the chamber, animals were required to press the other, non-signal lever to obtain food reward. Animals had 4 s to make a response once levers were extended into the chamber, or the trial was recorded as an omission. During this initial training, animals were given up to 3 correction trials following an error on both signal or non-signal trials. Following a third consecutive error within the trial, the animal was given a forced choice trial. Initial training consisted of 80 signal and non-signal trials each, with a variable intertrial interval (ITI) of 6-12 s. All trials during this phase of training occurred with the houselight off. Animals were required to meet a criterion of >65% correct on both signal and non-signal trials for three consecutive days before advancing to the next stage of training. During the next stage, animals were trained to respond to briefer signal lights and were given no correction trials. Non-signal trials were identical to the prior phase of training. In the signal trials, after the variable ITI (6-12 s), the signal light was illuminated for 500, 250 and 50 ms.  After 2s, both levers were extended into the chamber and the animal had 4s to make a response. Animals received 81 non-signal trials, with 27 trials at each different duration of signal light presented in random order throughout the session. All trials during this phase of training 73  occurred with houselight off.  Animals were required to meet a criterion of >65% correct on both 500 ms signal trials and non-signal trials before advancing to the next stage of training.   The final phase of training was identical to the previous, except that houselight was illuminated throughout the session to increase attentional demands. Animals were required to meet a criterion of >65% correct on 500 ms signal trials and non-signal trials before they were subjected to surgery. After surgery, they were retrained to a criterion of >70% prior to drug testing. Two separate cohorts (Cohort 1 and 2) were trained in the SAT. Cohort 1 was initially trained with stimulus durations of 500, 50 and 25 ms as has been described previously by Sarter and colleagues (McGaughy et al. 1996; McGaughy and Sarter 1995; Turchi et al. 1995). However, after considerable training on the task, performance of this cohort remained consistently below chance performance at both the 50 and 25 ms stimulus durations. Therefore, training continued using stimulus durations of 500, 250 and 50 ms. Cohort 2 was trained only using the 500, 250 and 50 ms durations throughout training.  Rats in Cohort 1 and 2 required 47 +/- 5 and 32 +/- 3 days of training on all phases of the task before initially achieving criterion performance on the final stage of SAT training. When performance of the cohorts after control treatments were compared, a main effect of cohort (F1,16=6.26, p<0.05) and significant cohort x duration interaction (F2,32=5.65 p<0.05) were found for the vigilance index, with rats in Cohort 1 showing slightly better performance compared to those in Cohort 2, particularly at the longest signal duration (data not shown). However, there were no significant interactions of cohort and treatment following both intra-PFC bicuculline (all Fs<1.32, p>0.25) and MK-801 treatment (all Fs <1.31, p>0.25), indicating that these treatments had comparable effects in both cohorts 74  irrespective of their baseline levels of performance.  Thus, the data from both cohorts was pooled for the final analyses. 3.2.4 Conditional visual discrimination  The visual discrimination task was identical to sustained attention task in all parameters, except that the signal light was illuminated for 2.5 s during all signal trials, while the houselight was off throughout the test session, and the levers were extended into the chamber immediately after the signal light had extinguished. After retractable lever training, animals went through a first phase of training in which they received up to three correction trials followed by a forced choice following errors. The lever used to indicate presence of a signal or non-signal trial were kept constant in each animal and counterbalanced across animals. Once animals had reached a criterion of >65% in the first phase of training, they were advanced to the final phase in which no correction trials were given. Animals were trained to a criterion of >70% in the final phase before infusions commenced.  Rats required 9 +/- 1 days of training to initially achieve this level of performance. 3.2.5 Conditional auditory discrimination  The auditory discrimination task required subjects to differentiate between two sound stimuli (a 3 kHz pure tone and white noise). Prior to training on the full task, rats were trained to press retractable levers as in the other experiments.  For the conditional auditory discrimination stage of training, trials occurred every 40s, and entailed illumination of the houselight that coincided with presentation of one of the two auditory stimuli. Levers were inserted into the chamber 3 s after initiation of the auditory stimulus, and the tone/white noise continued until rats either pressed one of the levers or 10 s elapsed.  If no response was made within 10s of insertion 75  of the levers, they were retracted, the houselight was extinguished and the trial was scored as an omission.  In order to receive a food reinforcement, rats were required to press the lever associated with a particular stimulus (e.g. 3 kHz = left lever, white noise = right lever). The levers associated with each stimulus were kept constant throughout training for each animal and were counterbalanced across animals. Rats were given 10 initial days of training that began with 20 forced-choice trials where only the correct lever associated with a particular tone was inserted, followed by 20 free choice trials where both levers were inserted. Rats were then trained on the final task for ~4 days that consisted of 40 free choice trials. Animals were required to reach a criterion of >80% prior to drug testing. 3.2.6 Surgical procedures Rats were implanted with bilateral guide cannulae in the prelimbic region of the medial PFC.  This region was selected as it is the same region used in other studies demonstrating impairments in attention and other forms of cognition following infusion of GABA antagonists (Enomoto et al. 2011; Paine et al. 2011; Piantadosi and Floresco 2014; Piantadosi et al. 2016). For animals in SAT experiments, the animals were given 1-2 days of ad libitum food after reaching criterion performance. Animals in conditional discrimination experiments received surgery prior to behavioral training, and therefore were not on food restriction prior to surgery. Anaesthesia was induced using ketamine/xylazine (50 and 10 mg/kg respectively) and maintained using isoflurane vapour. Animals were given Anafen (5 mg/kg, subcutaneous) as a pre-surgical analgesic, and received the same dose of Anafen for at least two days following surgery. Bilateral 23 gauge stainless steel cannulae were implanted into the medial PFC using the following co-ordinates (flat skull AP = +3.2 mm; ML = ±0.7 mm from bregma; DV = −2.8 mm from dura).  Animals were given approximately one week to recover after surgery, with ad 76  libitum food during the first 3-4 days. At this point, food restriction to 85-90% of the animal’s free feeding weight was either resumed in the case of Experiments 1-3 or started in Experiments 4 and 5.  3.2.7 Drugs & microinfusion procedures Antagonism of PFC GABAA receptors was achieved using the competitive antagonist bicuculline methobromide (12.5 or 50 ng in 0.4 μl per hemisphere). These doses were chosen because they have been shown to be effective at altering forms of cognition when infused into the PFC, but are below the threshold for inducing seizures when infused in this region (Enomoto et al., 2011, Paine et al., 2011). We chose to use a lower infusion volume than previously reported to reduce infusion site damage for animals in SAT experiments, which received 5 infusions each. These same infusion parameters were kept constant for all animals receiving bicuculline infusions, regardless of experiment. Antagonism of PFC NMDA receptors was achieved using MK-801 (6 µg in 0.5 µl per hemisphere). We chose this dose as we determined it led to working memory impairments in Chapter 2. Here, we maintained the same infusion volume as previously reported due to the high concentration of the solution. Following all intracranial infusion tests, a subset of rats were retrained to criterion performance on the SAT task, after which they received counterbalanced intraperitoneal injections of two doses of MK-801 (0.1-0.3 mg/kg) or vehicle. These doses have previously been shown to induce attentional impairments in 3- and 5-choice serials reaction time tasks (Paine et al. 2007; Pehrson et al. 2013).  After reaching criterion performance on their respective task, rats received a mock infusion procedure prior to training, during which obdurators were removed from the guide 77  cannulae, and replaced with stainless steel injectors for 2 min, without an infusion. They then received their first microinfusion test day. All intra-PFC infusion test were administered bilaterally using 30-gauge injectors extending 0.8 mm below the guide cannulae.  Following infusions, the injectors were left in place for 1 min to allow for diffusion. Rats were then placed back in their home cages for 10 min, after which, testing commenced. The first series of tests involved infusions of saline or bicuculline (12.5 or 50 ng) into the medial PFC at rate of 0.4 µL/125 s. Following each test session, animals were trained daily in the task until they re-achieved criterion performance for at least 2 consecutive days. Once criterion was re-achieved, rats received counterbalanced infusions of saline or bicuculline until both doses of bicuculline and saline had been infused.  The same animals in the first experiment were used in the second and third experiment. Once animals had received infusions of both doses of bicuculline and saline, they were given a 1 week minimum washout period and then were retrained to criterion performance. Upon re-achieving criterion, they received an infusion of either saline or MK-801 (6 µg in 0.5 uL) at a rate of 0.5 µL/75 s. They were the re-trained to criterion and received the second counter-balanced infusion. Only animals in Cohort 2 were used for systemic MK-801 experiments. After PFC infusions of MK-801, animals were given a one week washout period and re-trained to criterion. At this point, they received the first of 3 i.p. injections (saline, 0.1 mg/kg and 0.3 mg/kg MK-801).    78  3.2.8 Histology After the final test session, rats were sacrificed using CO2. Brains were removed and fixed in a 4% formalin solution for at least 24 h before being frozen and sliced into 50 μm sections and mounted onto gelatin-coated slides. The section were then stained with Cresyl Violet. Figure 3.1 shows the location of all acceptable infusion placements. 3.2.9 Data analysis For the experiments employing the SAT, the vigilance index was the primary measure of interest. The vigilance index is an overall measure of task performance that combines the relative numbers of hits and correct rejections and is not confounded by the number of omissions. This was calculated using the formula (h - f) / 2(h + f) - (h + f)2, where h is the proportional number of hits, and f is the proportion of false alarms (1-relative number of correct rejections), as in McGaughy & Sarter (1995). The main analysis consisted of a comparison of the vigilance index at each stimulus duration. These data were analyzed with a two-way repeated measures ANOVA with drug dose and stimulus duration as the within-subject factors. We also analyzed the proportion of hits per completed signal trials (% hits) using a two-way repeated measures ANOVA with dose and signal duration as within-subjects factors, and the relative number of correct rejections in the non-signal trials (% correct rejections) using a one-way ANOVA. In each of these analyses, the main effect of stimulus duration was significant for both the vigilance index and the % hits, and will not be reported further. The number of trial omissions and latencies to make a response in both signal and non-signal trials were also analyzed. Number of omissions were analyzed using one-way ANOVAs. Latency data was analyzed using a two-way repeated measures ANOVA with dose and trial type (signal and non-signal) as within-subject factors. 79     Figure 3.1. Location of all acceptable placements within the medial PFC for each experiment in Chapter 3. Different symbols indicate the location of cannulae placements in each of the four different experiments conducted.  80  For the conditional visual discrimination task, the primary measure of interest was overall % correct of completed trials which was analyzed using a one-way ANOVA. The % hits (correct responses on signal trials), % correct rejections (correct responses on non-signal trials) and omissions were analyzed in separate one-way ANOVAs. Latency data was analyzed using a two-way repeated measures ANOVA with dose and trial type as within-subjects factors. For the conditional auditory discrimination task, the primary measure of interest was overall % correct of completed trials, which was analyzed using a one-way ANOVA. The number of omissions and the latency to make a response in each trial were analyzed using one-way ANOVAs. For SAT experiments with intra-PFC infusions, the average of both saline infusion days were used in both analyses, as the two saline infusions were found not to be significantly different when the vigilance index was analyzed using a two-way ANOVA with infusion day and duration as within-subjects factors (F1,16=0.34, p>0.55).  Lastly, all analyses involving multiple comparisons used Dunnett’s test.  3.3 Results 3.3.1 PFC GABAA receptor antagonism and sustained attention  Two separate cohorts of animals (total n=18) were well-trained on the SAT prior to receiving counterbalanced infusions of the GABAA receptor antagonist, bicuculline (12.5-50 ng). Antagonism of PFC GABAA receptors using both doses of bicuculline produced a significant decrease in the vigilance index at all stimulus durations (Figure 3.2A), as indicated by significant main effect of treatment (F2,34=6.14, p<0.01 and Dunnett’s, p<0.05) and no significant treatment  81         Figure 3.2. PFC GABAA receptor antagonism decreases vigilance at all stimulus durations and increases false alarm responding on the SAT. A, Antagonism of PFC GABAA receptors decreased the vigilance index at all stimulus durations and after administration of both doses of bicuculline. B, When signal and non-signal trials were analyzed separately, the total hits were unaffected.   C, Intra-PFC bicuculline infusions reduced the proportion of correct rejections decreased, indicative of an increase in false alarm responses. For this and all other figures, symbols/bars represent the mean ±SEM, p<0.05 vs saline.   82  x duration interaction (F4,68=1.43, p>0.2). These decrements in the vigilance index appeared to arise predominately from increased false alarm responding and not deficits in signal detection, as bicuculline treatment induced a subtle, but statistically-reliable reduction in the number of correct rejections at both doses (main effect of treatment: F2,34=3.43, p<0.05 and Dunnett’s, p<0.05; Figure 3.2C). In contrast, even though bicuculline treatments tended to reduce the number of hits at the 500 and 250 ms signal durations, which may have contributed to the reduction in the vigilance index, no statistically significant differences in the proportion of hits on signal trials were observed across treatments (main effect of treatment: F2,34=0.46, p>0.6; treatment x duration interaction: F4,68=1.98, p>0.1.; Figure 3.2B). Antagonism of PFC GABAA receptors did not significantly affect latency to respond in either signal or non-signal trials at either dose (all F values < 3.60, p>0.05) or the total number of omissions (F2,34=1.43, p>0.25; see Table 3.1).   Thus, reducing PFC GABAA transmission impaired attentional performance, primarily by increasing the likelihood that animals would respond on non-signal trials as if they had detected a visual stimulus.  3.3.2 PFC NMDA receptor antagonism and sustained attention  Similar to PFC GABAA antagonism, antagonism of PFC NMDA receptors has been reported to impair performance on the 3- and 5-CSRTT (Murphy et al. 2005; Pehrson et al. 2013), but it is unclear if these impairments result from decreased signal detection or increased false alarms. Therefore, in order to compare the effects of PFC GABAA and NMDA receptor antagonism, we administered counterbalanced intra-PFC infusions of the NMDA antagonist MK-801 and saline in the same group of rats used in Experiment 1 (n=17). One animal from this group was excluded from this second round of infusions due to a block in its cannulae. 83  Table 3.1. Latencies for signal and non-signal trials and total omissions for each experiment in Chapter 3.  Response latency, signal trials (s) Response latency, non-signal trials (s) Trial Omissions Experiment 1   Saline   Bicuculline, 12.5 ng   Bicuculline, 50 ng   0.45 ± 0.02 0.50 ± 0.03 0.62 ± 0.09  0.51 ±0.03 0.55 ± 0.03 0.59 ± 0.08   9 ± 4 10 ± 4 14 ± 55 Experiment 2   Saline   MK-801, 6 µg  0.46 ± 0.04 0.43 ± 0.03  0.52 ± 0.04 0.53 ±0.05   9 ± 4 1 ± 1 Experiment 3   Saline, i.p.   MK-801, 0.1 mg/kg   MK-801, 0.3 mg/kg  0.50 ± 0.05 0.51 ± 0.06 0.70 ± 0.1*  0.55 ± 0.05 0.53 ± 0.06 0.81 ± 0.1 *  12 ± 7 7 ±5 29 ± 11  Experiment 4   Saline   Bicuculline, 12.5 ng   Bicuculline, 50 ng  0.55± 0.5 0.53 ± 0.07 0.72 ± 0.07  1.0 ± 0.1 1.3 ± 0.2 1.1 ± 0.1  12 ± 8 20 ± 10 22± 11  Experiment 5   Saline   Bicuculline 12.5 ng   Bicuculline, 50 ng  0.55 ± 0.05 0.78 ± 0.1 1.5 ± 0.5 *   0 ± 0 0 ± 0 3± 2  * denotes p< 0.05 vs saline. 84         Figure 3.3. PFC NMDA receptor antagonism decrease vigilance at short durations without influencing hits or false alarm responses. A, Antagonism of PFC NMDA receptors resulted in a decrease in the vigilance index at the shortest stimulus duration. B,C, When signal and non-signal trials were analyzed separately, no effects on hits or correct rejections were observed.   85  In contrast to the effects of reducing PFC GABA transmission, antagonism of PFC NMDA receptors (6 µg) did not cause an overall change in the vigilance index (main effect of treatment: F1,16=0.16, p>0.65; Figure 3.3A), although a significant treatment x stimulus interaction of was observed (F2,32=3.54, p<0.05).  Infusions of MK-801 produce a subtle deficit (p<0.05) in discriminating between signal and non-signal events at the shortest stimulus duration. Yet, separate analyses of the proportion of hits and correct rejections showed that these treatments did not affect signal detection at any stimulus duration (both Fs <2.27, p>0.1; Figure 3.3B) nor did they alter the correct rejection rate (F1,16=2.98, p>0.1; Figure 3.3C). The average latencies to respond in signal and non-signal trial were unaffected by PFC NMDA receptor antagonism (all Fs <1.99,p>0.15), as were the total number of omissions in either trial type (both Fs <3.71, p>0.05; Table 3.1).   3.3.3 Systemic NMDA receptor antagonism and sustained attention  Systemic administration of certain NMDA receptors antagonists including MK-801 (Paine et al. 2007) and CPP (Quarta et al. 2007) produce robust deficits in 3- and 5-CSRTT, whereas other antagonists such as ketamine produce non-specific effects in SAT performance (Nelson et al. 2002). In light of the subtle effects observed following intra-PFC MK-801 treatments, we injected MK-801 (0.1-0.3 mg/kg, i.p. injection) to confirm that performance on this task was sensitive to systemic treatment with this compound. These experiments were completed using only the animals from Cohort 2 (n=10). Injections of the higher dose of MK-801 (0.3 mg/kg) decreased the vigilance index at all stimulus durations in comparison to the lower dose of MK-801 and i.p. injections of saline, as indicated by a significant main effect of treatment (F2,18=3.80, p<0.05 and Dunnett’s, p<0.05) but no treatment x duration interaction (F4,36=0.71, p>0.55; Figure 3.4A).  This effect on the  86        Figure 3.4. Systemic NMDA receptor antagonism decreases vigilance without affecting hits or false alarm responses. A, Systemic NMDA receptor antagonism resulted in a decrease in the vigilance index at all stimulus durations, following administration of the higher, but not lower, dose of antagonist. B,C, When signal and non-signal trials were analyzed separately, no significant effects on hits or correct rejections were observed.   87  vigilance index appears to be driven primarily by disruption in signal detection, as visual inspection of Figure 3.4B indicates a reduction in the proportion of hits following treatment with the higher dose. However, analyses of these data failed to yield a significant difference between treatment conditions on this measure (both Fs<1.84, p>0.1). Likewise, the proportion of correct rejections also did not differ significantly across treatments (F2,18=1.96, p>0.15). The 0.3 mg/kg dose of MK-801 also increased latencies in both signal and non-signal trials, as indicated by a main effect of treatment (F2,18=5.08, p<0.05) and no significant interaction of treatment and trial type (F2,18=1.57, p>0.2 ; Table 3.1). Administration of the higher dose led to a >2-fold increase in omissions, however this effect was not statistically reliable (F2,18=2.26, p>0.1).  Taken together, the results of Experiment 2 and 3 reveal that non-competitive antagonism of NMDA receptors can disrupt attentional performance, yet the contribution of these receptors in the PFC to these processes appears to be relatively minor.  3.3.4 PFC GABAA receptor antagonism and conditional visual discrimination  In addition to requiring vigilance to discriminate between signal and non-signal trials, the SAT also requires animals to perform a conditional discrimination (e.g.; if the light was on, press the left lever, if not, press the right lever). In light of this, another experiment investigating how reducing PFC GABAA transmission affects performance of these types of discriminations was conducted to ascertain whether the impairments in performance on the SAT induced by this treatment were attributable primarily to disruptions in attentional-related processes or the ability to apply conditional discrimination rules. In so doing, a separate cohort of rats (n=9) was trained in a visual conditional discrimination task that was identical to the SAT, except that the attentional demands of the task were reduced by increasing the duration of signal light to 2.5 s, 88  turning off the houselight, and removing the delay between signal presentation and lever extension.   Under control conditions, rats tested on the conditional visual discrimination displayed better performance when compared to rats tested on the SAT.  A direct comparison revealed that after saline infusions, rats tested on the conditional discrimination task made a greater proportion of hits compared to those made by rats on the SAT during 500 ms signal trials (t25=2.71, p<0.05; Figure 3.5A, left).  Similarly, the correct rejection rate was also significantly higher in rats tested on the conditional discrimination relative to those tested on the SAT (t25=2.54, p<0.05; Figure 3.5A, right).  This indicates that these changes to the SAT task were effective in reducing attentional load. In well-trained rats, treatment with the 12.5 ng bicuculline did not affect accuracy, whereas the 50 ng dose did cause a noticeable reduction in the percentage of correct trials on the conditional visual discrimination (Figure 3.5B). Yet, the overall analysis of these data failed to detect a statistically significant difference across treatment conditions (Figure 3.5B; F2,16=1.79, p>0.15).  However, when signal and non-signal trials were analyzed separately, a significant decrease in correct rejections was found following the 50 ng, but importantly, not the 12.5 ng dose of bicuculline (F2,16=5.48, p<0.05 and Dunnett’s, p<0.05; Figure 3.5C ). In comparison, the relative number of hits was unaffected (F2,16=0.48, p>0.6.). There were no significant effects of bicuculline treatment on latencies or the number of trial omissions (all Fs<3.11, p>0.05).  Thus, intra-PFC infusions of a higher dose of bicuculline that impaired performance on the SAT also impaired performance of a visual conditional discrimination and in both instances, this was driven by an increase in erroneous responses during non-signal trials. However, the lower, 12.5   89   Figure 3.5.  Intra-PFC administration of a higher dose of GABAA antagonist disrupts the ability to perform visual conditional discriminations. A, Accuracy of animals receiving intra-PFC saline infusions in the SAT compared to the conditional visual discrimination task. Saline-infused animals performed significantly better in the visual discrimination task, indicating that latter task had a reduced attentional load relative to the SAT. B, Accuracy on the conditional visual discrimination task was unaffected by PFC GABAA receptor antagonism, although the 50 ng dose did induce a noticeable reduction on this measure. C, However, when signal and non-signal trials were analyzed separately, the number of hits was unaffected, while correct rejections decreased after administration of the higher dose of antagonist.  90  ng dose, which also impaired SAT performance, did not impair accuracy on this task that placed fewer demands on attentional processes.  3.3.5 PFC GABAA receptor antagonism and conditional auditory discrimination The results of the last experiment indicate that a higher dose of intra-PFC bicuculline disrupted performance of a visual conditional discrimination. In this task, animals were required  to make a particular response based on whether a visual cue was either presented or not, and the higher dose of bicuculline preferentially increased errors during non-signal trials. We were interested in determining whether the involvement of PFC GABAA transmission in conditional discriminations also extends to situations when animals must discriminate between two distinct stimuli and within a different sensory modality.  To address this, a separate group of rats (n=7) were trained in an auditory discrimination task requiring discrimination between two sound stimuli (3 kHz tone and white noise). Analysis of the performance data from this experiment revealed a significant main effect treatment on accuracy of conditional auditory discriminations (F2,12=12.03, p=0.001).  Multiple comparisons with Dunnett’s test further revealed that only the 50 ng, but not the 12.5 ng dose, caused a significant impairment in accuracy relative to saline treatment (p<0.05; Figure 3.6).  In this experiment, the latency to make a response was increased by the higher dose of bicuculline, but not the lower dose (F2,12=4.84, p<0.05 and Dunnett’s, p<0.05; Table 3.1). However, these treatments did not affect the number of omissions made during the test sessions (F2,12=2.32, p>0.1; Table 3.1).  Thus, disinhibiting PFC activity with a higher dose of bicuculline impaired the ability to effectively apply a conditional rule that required discrimination between two distinct auditory stimuli. 91        Figure 3.6. Intra-PFC administration of the higher dose of GABAA antagonist disrupts performance of conditional auditory discriminations. Administration of the higher, but not lower dose of bicuculline reduced accuracy during performance of a conditional auditory discrimination task.   92  3.4 Discussion The primary finding of this study is that PFC GABAA activity plays a key role in facilitating sustained attention, as reducing this activity increased erroneous responses to non-signal events. PFC GABAA receptor antagonism reduced the vigilance index regardless of stimulus duration and increased false alarm responding. Deficits induced by the lower dose of GABAA antagonist reflected a purely attentional effect, as this dose did not affect performance of conditional discriminations using either a visual or auditory stimuli.  However, a higher dose also impaired these types of discriminations, suggesting an additional deficit in implementation of conditional rules.  In contrast, non-competitive antagonism of PFC NMDA receptors produced only a minimal deficit in vigilance at the shortest stimulus duration, without directly affecting signal detection or false alarm responding. Systemic administration of the same NMDA antagonist at a high dose resulted in an overall reduction in vigilance and non-specific increases in latencies. Taken together, the present findings further elucidate the role of PFC GABAA and NMDA receptors in attentional processes and may provide important insight into understanding pathophysiological mechanisms underlying attentional impairments induced by deficient PFC GABAA and NMDA receptor signalling observed in other tasks or in pathological conditions, such as schizophrenia. 3.4.1 PFC GABA hypofunction disrupts vigilance and increases false alarm responding Reducing PFC GABAA transmission decreased vigilance irrespective of stimulus duration, with both low and high doses of antagonist producing this effect. When signal and non-signal trials were analyzed separately, the detection of target stimuli on signal trials did not differ significantly across treatment conditions, while during non-signal trials, false alarm responding was increased by intra PFC-bicuculline. These findings expand on previous studies indicating 93  that antagonism of PFC GABAA receptors disrupts 3- and 5-CSRTT performance, without enhancing motor impulsivity (Paine et al. 2011; Pehrson et al. 2013; Pezze et al. 2014).  In a study using bicuculline (Paine et al. 2011), 5-CSRTT performance was impaired by intra-PFC infusions of 25 ng, but not 12.5 or 6.25 ng of the compound.  The lack of a dose-response effect observed here suggests that attention assessed in this manner may be particular sensitive to disruption of PFC GABA transmission, and that the dose range we used was not broad enough to detect a null effect.  The present results indicate that impairments in attentional performance arising from decreases in PFC GABA function do not emerge from deficits in signal detection, i.e. an inability to register the occurrence of a brief visual stimulus. Rather, deficiencies in PFC GABA signalling disrupt attentional performance by impairing the ability to distinguish between signal and non-signal events, thereby increasing the tendency to respond as if a stimulus is present when it is not.  This in turn may reflect an increased sensitivity to distraction, either via a reduction in the ability to distinguish non-signal events against the background of the illuminated houselight, or by impairing inhibition of erroneous responses during non-signal trials (McGaughy and Sarter, 1995).  Intact PFC function is known to be necessary for attentional processes, as PFC inactivation and/or lesions disrupt both 5-CSRTT and SAT performance (Chudasama et al. 2003; Maddux and Holland 2011; Miner et al. 1997; Paine et al. 2011; Pezze et al. 2014). Interestingly, while PFC lesions disrupt SAT performance on both signal and non-signal trials (Miner et al., 1997), perturbations in PFC cholinergic signalling selectively disrupt detection during signal trials, but do not impact false alarm responses (Floresco and Jentsch 2011; Gill et al. 2000; McGaughy and Sarter 1998). In this way, disruptions in PFC cholinergic signalling appear to produce effects that are opposite to those observed following deficient PFC GABA signalling. 94  As such, it may be that cholinergic signalling is a primary regulator of signal detection, while PFC GABA signalling plays a complementary role in filtering external and internal sources of noise. Furthermore, it appears that even subtle perturbations in GABA signalling, such as those observed following infusion of the 12.5 ng dose of bicuculline, are sufficient to interfere with either the filtering of background stimuli or inhibition of erroneous responses. In Chapter 2, we found that PFC GABAA receptor antagonism causes delay-independent impairments in working memory, suggesting that intact GABAA transmission may play a more prominent role in encoding, rather than maintenance or retrieval of information within working memory. Given that the ability to maintain attention towards target stimuli is a major factor in successful encoding, one of the goals of this study was to understand how PFC GABA transmission contributes to the attentional processes that enable accurate encoding of information within working memory. Although it was previously known that diminished PFC GABA signalling disrupts attention in serial reaction time tasks (Paine et al. 2011; Pehrson et al. 2013; Pezze et al. 2014), the exact nature of this impairment was unknown. The observation that reducing PFC GABAA transmission does not affect signal detection but impairs identification of non-signal events clarifies the specific attentional component that may contribute to deficient encoding following decreased PFC GABA function.  It is important to note that even though the SAT used here imposed a brief delay (2 s) between presentation of the signal and extension of the choice levers, only false alarm responses were affected by antagonism of PFC GABAA receptors. A more global deficit in the process of encoding information would be expected to affect representations during both non-signal and signal trials, resulting in both decreases in hits and increases in false alarms.  Thus, disruptions in PFC GABA transmission do not appear to affect encoding directly. Rather, they seem to interfere primarily with the attentional regulation 95  of encoding, by impairing the gating of external or internal noise, thereby weakening the encoded representation. Therefore, the present data, in combination with previous findings, suggest that one of the key functions of PFC GABA signalling is filtering of irrelevant or distracting information, an ability that is crucial to a wide variety of cognitive processes. In the present study, pharmacological reductions in PFC GABA transmission did not reliably increase latencies or omissions during performance of the SAT, which stands in contrast to previous studies examining the effects of PFC GABAA receptor antagonism on cognition, including our results in Chapter 2 (Enomoto et al. 2011; Piantadosi et al. 2016). In comparison to the tasks used in these previous studies, the SAT used here requires persistent orientation towards the front panel of the chamber while awaiting visual stimuli that appear unpredictably. The SAT also uses shorter ITIs and delays between phases of each trial. As such, it appears that for tasks requiring continuous engagement, reductions in PFC GABA signalling do not disrupt speed-of-processing, but that in situations with longer delays, subjects can take longer to orient towards stimuli, leading to increases in latencies and trial omissions.  In support of this idea, PFC GABAA receptor antagonism increased latencies in the sample and pre-choice phases of a delayed response task, which were preceded by the ITI or a delay, but not in the actual choice phase, in which subjects were already oriented toward the front panel in the DNMTP task. Likewise, similar infusions of bicuculline that disrupted attentional accuracy did not alter latencies to make a correct choice on a 5-CSRTT (Paine et al., 2011). Furthermore, in the present study, latency effects were observed in the auditory discrimination task, which had longer ITIs, but not the SAT or visual discrimination task which have shorter ones.  Although the effects of reducing PFC GABA transmission on speed-of-processing may vary across different tasks, it is important to emphasize that the lack of effects on latencies and omissions on the SAT indicate 96  that impairments induced by PFC GABAA antagonism on this task do not result from non-specific disruptions in reward or motoric processes. 3.4.2 Diminished PFC NMDA receptor signalling produces subtle vigilance deficits In contrast to the effects of PFC GABAA receptor antagonism, blockade of NMDA receptors in this region had almost no effect on vigilance, producing only a minor impairment in discriminating between signal and non-signal events at the shortest signal duration. Furthermore, this impairment in vigilance could not be attributed exclusively to either diminished signal detection or increased false alarm responding. These results are somewhat surprising in light of the fact that administration of MK-801 into the anterior cingulate cortex disrupted performance of the 3-CSRTT (Pehrson et al. 2013), while the competitive NMDA receptor antagonist R-CPP induced deficits in 5-CSRTT when administered in the infralimbic cortex (Murphy et al. 2005). The dose of MK-801 employed in this study was larger than that employed by Pehrson et al. (2013), and was sufficient to induce deficits in working memory following intra-PFC administration in the DNMTP task, suggesting that the lack of effect observed here was not due to insufficient dosing.  However, in the present study, our infusions were clustered within the prelimbic cortex, perhaps indicating that NMDA receptors within the cingulate or infralimbic regions are more crucial to attention. Indeed, Murphy and colleagues (2005) found that prelimbic administration of NMDA antagonists produced small impairments in 5-CSRTT performance in comparison to infusions targeting the infralimbic cortex. Finally, this disparity in the attentional effects of PFC NMDA receptor antagonists may reflect differences in the tasks employed. In the 5-CSRTT, subjects are required to detect a stimulus light that may appear in one of five locations, while in the SAT, subjects are only required to attend to one location for presence of the stimulus. Thus, it may also be that PFC NMDA receptors are recruited preferentially in 97  situations in which attention is divided between multiple locations. Furthermore, performance of the 5-CSRTT requires inhibition of impulsive responding, while the SAT does not test this ability, suggesting that PFC NMDA receptors may have a selective role in modulating impulsivity, particularly in more ventral regions (Murphy et al. 2005).  The relatively muted effects of intra-PFC MK-801 on SAT performance contrast with pronounced impairments in delayed-responding induced the same dose administered in Chapter 2.  In this regard, it is interesting to compare how reducing PFC GABA vs NMDA receptor signalling differentially affects attention and working memory.  As described above, intra-PFC antagonism of GABAA receptors impaired both attention and induced delay-independent impairments on a delayed non-matching to position task. In comparison, infusions of MK-801 impaired working memory in a delay-dependent manner, suggesting these treatments altered processes related to maintenance and/or retrieval of short-term information from working memory.  These findings, in combination with the relative subtle effect of intra-PFC NMDA receptor antagonism on attentional performance, highlight the complementary yet dissociable roles that GABAA and NMDA receptor-mediated transmission play in regulating attentional, cognitive and executive functions mediated by the frontal lobes.  In light of the subtle effects found following intra-PFC administration of MK-801, it was of interest to examine how systemic administrations of the same drug affected SAT performance. Decreased accuracy on the 3- and 5-CSRTT has been observed following systemic NMDA receptor antagonism at the doses employed in this study (Paine et al. 2007; Pehrson et al. 2013). However, it is also not clear whether these impairments reflect increased false alarm responding or deficient signal detection. We found that a dose of 0.3 mg/kg produced a significant main effect on the vigilance index, whereas the 0.1 mg/kg had no effect on any aspect of task 98  performance. When signal and non-signal trials were analyzed separately, the higher dose had no significant effects on either signal detection or false alarms. Furthermore, administration of the higher dose led increases in latencies and omissions (although this latter effect failed to achieve statistical significance). Given these considerations, our impression is that impairments in SAT performance induced by systemic MK-801 were at least partially driven by non-specific motoric disruptions rather than specific deficits in signal detection or increased vulnerability to distraction. This notion is in keeping with the findings of Nelson and colleagues (2002), who examined the effect of systemic ketamine on SAT performance. In that study, the only significant effect of the highest dose of ketamine on SAT performance was a marked increase in omissions, also indicative of non-specific motor effects.   3.4.3 Diminished PFC GABA signalling disrupts performance of conditional discriminations In addition to placing high demands on vigilance, performance of the SAT is dependent upon the ability to follow rules that indicate that one response should be executed in the event of appearance of the signal, and while another response should be executed in the event of a non-signal trial. To examine how reducing PFC GABA transmission might affect the ability to follow these types of conditional rules, we conducted a control experiment.  Here animals were trained on a visual conditional discrimination task that was identical to the SAT in every respect except that the signal was presented for an extended duration, and there was no background illumination throughout the session or delay following presentation of the signal.  Thus, although the two tasks were nearly identical, these subtle changes collectively reduced the attentional demands of this task relative to the SAT (see Figure 3.5A). Intra-PFC infusion of the lower dose of GABAA receptor antagonist had no effect on the performance of the visual conditional discrimination. This observation, combined with the finding that this dose was effective at impairing SAT 99  performance, indicates that the increases in false alarm responding following this dose emerged because of the increased attentional load of the SAT. Thus, more subtle perturbations in PFC GABA function appear to selectively impair processes related to attention.  This idea is supported by the findings that intra-PFC infusions of 25 ng bicuculline also impaired attentional accuracy on the 5-CSRTT, which does not require implantation of conditional rules, but merely required animals to approach a location where a light has been illuminated (Paine et al., 2011).  On the other hand, the higher dose of bicuculline had no significant effect on overall accuracy for the visual discrimination task, but when signal and non-signal trials were analyzed separately, this dose led to increased false alarm responses, similar to what was observed in the SAT.  Thus, it is possible that even when attentional demands are low, more intense disruptions in GABA function impair vigilance.  Alternatively, the deficits on the visual conditional discrimination induced by the 50 ng dose of bicuculline may reflect an impairment in the implementation of conditional rules, which in this instance, biased responding towards the signal lever on non-signal trials. Both the SAT and conditional visual discrimination task involve distinguishing between trials where a single visual stimulus is either present or absent. In light of the above-mentioned findings, we examined whether PFC disinhibition impaired ability to perform conditional discriminations also applies to the ability to distinguish between two different stimuli, and whether this impairment extends to other sensory modalities.  We did so by assessing how PFC GABAA receptor antagonism affects performance of an auditory conditional discrimination. Similar to the result observed in the visual discrimination experiment, antagonism of PFC GABAA receptors impaired the ability to perform auditory conditional discriminations in a dose-dependent manner. In this way, moderate-to-strong reductions in PFC GABAA receptor 100  signalling not only disrupt attentional processes, but can also interfere with implementation of conditional rules, regardless of sensory modality or whether discrimination between two stimuli or presence vs. absence of stimuli is required. Thus, while subtle deficits in PFC GABA may interfere with input selection or allocation of attentional resources, greater reductions in PFC GABA signalling may additionally interfere with determining which rules to adhere to in the task. This deficit in rule selection appears to apply predominately to complex or conditional rules, as we have previously shown that simple left/right spatial discriminations performed in operant chambers are unaffected by intra-PFC administration of 50 ng bicuculline (Piantadosi et al. 2016). As the inability to follow conditional rules is likely to have impacts on performance of more complicated cognitive tasks, this deficit reflects a core feature of cognitive dysfunction associated with insufficient PFC GABA signalling. 3.4.4 Deficient PFC GABA transmission leads to schizophrenia-like cognitive impairments Hypofunction of the PFC GABA system is thought to be one pathophysiological mechanism that contributes to cognitive and other abnormalities observed in schizophrenia. Both in vivo imaging work (Frankle et al. 2015) and post-mortem studies examining markers of GABA function (Akbarian et al. 1995; Curley et al. 2011; Volk et al. 2000) point to decreased GABA synthesis and release in the frontal lobes of individuals diagnosed with the disorder. Although a deficiency in GABA transmission within the PFC has been proposed to contribute to cognitive dysfunction associated with schizophrenia (Gonzalez-Burgos and Lewis 2012; Uhlhaas and Singer 2010), it is unclear which domains of cognition may be affected by these perturbations. In recent years, our group and others have addressed this question by pharmacologically reducing PFC GABA transmission and identifying the distinct changes in cognition and behavior caused by diminished PFC GABAergic transmission (Enomoto et al., 101  2011; Paine et al., 2011; 2015; 2017; Pehrson et al., 2013; Piantadosi et al., 2014; 2016; Tse et al. 2015).  Although the lower dose of GABAA receptor antagonist used in this study produced an effect that was more specific to attention, the higher dose employed has consistently led to cognitive and behavioral deficits that mirror those observed in schizophrenia. For instance, in addition to the attentional deficits observed here, intra-PFC administration of this dose produces delay-independent deficits in delayed-response working memory, impaired set-shifting (Enomoto et al. 2011), and maladaptive decision-making (Paine et al. 2015; Piantadosi et al. 2016). In each of these instances, these impairments bear striking, qualitative similarity to the cognitive impairments described in schizophrenia patients (Fervaha et al. 2015; Lee and Park 2005; Pantelis et al. 1999). Furthermore, PFC GABAA receptor antagonism also results in deficits in social interaction (Paine et al. 2017), increased behavioral sensitivity to amphetamine (Enomoto et al. 2011) and impaired salience attribution (Piantadosi and Floresco, 2014), reflecting negative and positive symptoms in patients, respectively.   Attention deficits have been attributed to schizophrenia since the first descriptions of the disorder, and appear to be a core feature of the cognitive dysfunction associated with this condition (McCleery et al. 2015; Nuechterlein et al. 2004). Impairments in performance of the CPT are observed regardless of medication status (Finkelstein et al. 1997), appear in remitted patients (Nuechterlein et al. 1992), and are present to a lesser degree in first-degree relatives of patients (Chen and Faraone 2000; Finkelstein et al. 1997), making them a stable trait associated with schizophrenia. Deficits in CPT performance in patients are characterized by an inability to distinguish signal and non-signal events, and include both deficits in signal detection as well as increased false alarm responding (Nuechterlein et al. 2015). Furthermore, a translational version of the SAT administered to schizophrenia patients and controls showed an overall decrease in 102  task accuracy that was driven by increased false alarm responses, and also impairments in signal detection that were more predominate in conditions of increased distraction (Demeter et al. 2013). Intriguingly, though patients showed larger impairments in conditions of increased distraction, the impairment was still present in a low distraction condition, suggesting that attentional deficits may be present in schizophrenia even in situations with relatively low attentional demands, as with impairments observed after administration of the high dose of bicuculline in the present study. Taken together, these findings seem to indicate that patients with schizophrenia show increased vulnerability to distraction or are perhaps are unable to override inappropriate responses, as indexed by increases in false alarm responses, similar to what was observed following pharmacological reduction in PFC GABA signalling. This combination of findings provides evidence supporting the notion that PFC GABA hypofunction may be a major contributing factor to these aspects of the attentional deficits observed in schizophrenia. The impairments in performance of conditional discriminations observed induce by more robust reductions in PFC GABA transmission also bear a notable resemblance to deficits observed in schizophrenia patients. Visuospatial paired-associates learning, a component of the CANTAB test battery, is dependent on implementation of a conditional rule, and is impaired in first-episode schizophrenia patients. Notably, this is one aspect of cognitive function that has been shown to be predictive of patient outcomes (Wood et al. 2002). Coding tasks, such as the digit symbol substitution test of the Weschler intelligence scale, also require application of conditional stimulus-response rules, and have amongst the largest impairment effect sizes for cognitive tasks administered to schizophrenia patients (Dickinson et al. 2007). Both the cognitive aspects and the positive and negative symptoms of schizophrenia have been proposed to result from an underlying dysfunction in processing and responding to contextual information 103  (Hemsley 2005; Servan-Schreiber et al. 1996). The present results suggest that another aspect of cognition that is sensitive to deficiencies in PFC GABA transmission is the ability to implement conditional rules that are necessary for responding to unique contexts. When taken together with previous studies investigating the contribution of PFC GABA signalling to different domains of cognition, the present results suggest that deficits in PFC GABA function not only lead to behavioral changes that may be relevant to schizophrenia, they are in many instances sufficient to reproduce core features of the disorder.  3.4.5 Conclusions The present study adds to a growing body of literature examining how PFC GABAergic transmission contributes to cognition mediated by the frontal lobes. Here, we show that deficient PFC GABAergic transmission disrupts attentional processes, including vigilance, by increasing false alarm responding, rather than by interfering with the detection of targets themselves. Depending on the extent of the deficit, the ability to follow complex or conditional rules also is impacted. Taken together, these results provide additional insight into both the importance of inhibitory transmission within the PFC to different aspects of cognition, as well as the underlying pathophysiology that drives cognitive impairments observed in psychiatric disorders that are associated with diminished GABA function.   104  Chapter 4: Prefrontal cortical GABA modulation of spatial reference and working memory  4.1 Introduction Schizophrenia is associated with deficits in multiple domains of cognition, including working memory, attention, reasoning/problem solving and a generally slower speed of processing (Goldberg and Weinberger 1988; Hartman et al. 2003a; Heerey et al. 2008; Hutton et al. 2002; Keefe et al. 2007; Marder et al. 2004; Nuechterlein et al. 2004). These impairments often present prior to the onset of the psychotic symptoms, persist throughout the course of disease and are highly predictive of long-term patient outcome (Green 1996; Reichenberg et al. 2010). Furthermore, while currently available medications may treat ‘positive’ symptoms of schizophrenia (hallucinations, delusions), they are relatively ineffective at treating cognitive symptoms.  Thus, understanding the neural basis of cognitive deficits associated with schizophrenia is critical in order to develop better treatments for this disorder. Among the numerous neural and cellular perturbations that have been proposed to underlie aspects of schizophrenia symptomology, several lines of evidence suggest that reduced prefrontal GABA transmission may be a key contributor to the cognitive impairment associated with the disease.  Alterations in the expression of markers of GABA function are amongst the most highly reproducible pathologies observed in post-mortem samples of brains of schizophrenia patients.  mRNA and protein levels of the GABA synthesis enzyme glutamic acid decarboxylase-67 are consistently found to be reduced in the prefrontal cortex (PFC) of schizophrenics (Akbarian et al., 1995; Guidotti et al., 2000; Volk et al., 2000; Hashimoto et al., 2008; Thompson et al., 2009; Curley et al., 2011), potentially leading to decreased GABA synthesis and release.  Optimal prefrontal GABA transmission is thought to be critical for 105  generating neural oscillations, particularly in the gamma range, that underlie cognitive functions such as working memory (Howard et al., 2003; Gonzales-Burgos & Lewis, 2008).  This cortical network activity has been shown to be suppressed in schizophrenic patients engaged in cognitive tasks (Cho et al., 2006; Minzenberg et al., 2010).  Furthermore, several pharmacological and neurodevelopmental animal models of schizophrenia are also associated with disruptions in prefrontal GABA functioning (Amitai et al., 2012; Richetto et al., 2013; Thomases et al., 2013). Although a link between prefrontal GABA hypofunction and cognitive deficits associated with schizophrenia has been proposed, until recently there has been surprisingly little preclinical research investigating the contribution of the GABA transmission to cognitive functions governed by the frontal lobes.  Early reports found that GABAA receptor antagonism within the monkey dorsolateral PFC impaired delayed alternation (Sawaguchi et al.,1989) and disturbed task-related interneuron and pyramidal cell activity (Rao et al., 2000).  In keeping with these earlier findings, work by our group has shown that pharmacological reduction in prefrontal GABAA transmission recapitulates several cognitive and behavioral aspects of schizophrenia in the rat.  Infusion of the GABAA receptor antagonist bicuculline into the prelimbic medial PFC resulted in impaired cognitive flexibility, enhanced striatal phasic dopamine activity and speed-of-processing deficits (Enomoto et al., 2011).  Subsequent studies further revealed that intra-PFC infusion of low doses of GABAA receptor antagonist also induced schizophrenia-like deficits in attention assessed with a 3 or 5-choice serial reaction-time task (Paine et al., 2011; Pehrson et al., 2013). In contrast to the above mentioned findings, antagonism of PFC GABAA receptors did not impair working memory accuracy on a delayed response variant of the radial-arm maze task, although these manipulations did increase response latencies (Floresco et al., 1997; Enomoto et 106  al., 2011).  This was a surprising finding, considering that working memory assessed in this manner is markedly impaired by inactivation of the medial PFC (Seamans et al., 1995; Floresco et al. 1997, 1999).  In that study, it was proposed that the lack of effect of PFC GABAA antagonism on working memory may be related the incorporation of a relatively long delay between sample and test phases, and that responding during the test phase was self-paced. In comparison, tasks used to assess working memory capacity in patients often use shorter delay intervals and provide subjects with multiple trials in quick succession, requiring them to overcome proactive interference and discard information from previous trials (Park & Holzman, 1992; Fleming et al., 1997; Pukrop et al. 2003). Recent studies with schizophrenic subjects have employed back-translational approaches, using virtual maze tasks similar to those used with rodents to assess cognition.  In one notable study (Spieker et al., 2012), schizophrenic patients displayed marked reference and working memory deficits on a virtual version of the classic spatially-cued radial maze task initially described by Olton and Papas (1979).  These findings highlight the translational relevance of using these types of tasks in combination with preclinical animal models to probe the potential mechanisms underlying cognitive impairment in schizophrenia.  As such, a primary goal of the present study was to assess the importance of GABAergic transmission in the rat medial PFC in regulating and reference and working memory assessed with a similar radial-maze task.  In so doing, we used a “massed trials” variant of this task where rats were given multiple trials within a daily session.  This was employed to increase cognitive load and induce proactive interference, as it requires subjects to disambiguate whether an arm was entered on the current or previous trial, which may increase increases errors on later trials (Roberts & Dale, 1980).  A similar procedure has also been used in combination with a reference/working memory task using a 12-107  arm maze (Meck and Williams, 1999).  Our expectation was that reducing PFC GABA transmission would exert a greater effect on performance on latter trials relative to the first.  Subsequent experiments employed reversible inactivation of the medial PFC to assess its contribution to reference/working memory performance, and the effects of PFC GABAA antagonism on simpler tasks assessing short-term memory and spatial discrimination abilities.  4.2 Methods 4.2.1 Subjects Long Evans rats (250-300 g) were purchased from Charles River Laboratories (Montreal, Canada) and were initially group-housed upon arrival from the supplier. After 1 week, animals were single-housed and food-restricted to 85-90% of their free-feeding weight prior to beginning behavioral training or surgical preparations. Experiments were conducted in accordance with the Canadian Council on Animal Care and University of British Columbia Animal Care Committee.  4.2.2. Behavioral procedures  All behavioral tests were conducted during the animals’ light cycle.  All testing was conducted on an 8-arm radial maze. The radial maze used in these experiments had a 40 cm diameter octagonal center platform connected to eight equally spaced arms (50 × 9 cm) with a food cup placed at the end of each arm. The maze was placed in a (270 X 330 cm) room with numerous extramaze cues on the walls. On the first 2 days of testing, rats were familiarized to the maze by being placed in the center and allowed to explore for 10 min with no food available, after which they were returned to their home cages and given approximately 20 food reward pellets (BioServ, Frenchtown, NJ) that were used as reinforcement during training. The arms and center of the maze were wiped 108  down after each rat was removed from the maze. For all tasks, an arm entry was recorded when a rat moved down the entire length of an arm and reached the food cup at the end of the arm. The latencies to enter the first arm and to complete the trial were also recorded. 4.2.3 Reference/Working Memory Task This task was initially described by Olton and Papas (1979) and assessed both long-term (or “reference”) and short-term (or “working”) memory (RM and WM, respectively).  Although the term working memory has been used to describe a variety of cognitive operations related to storage and manipulation of information (Baddeley, 2012), here the term “working memory” refers to processes related primarily to short-term memory, to maintain consistency with other studies using similar tasks.  Rats were required to retrieve reward pellets placed in 4 of the 8 arms of the maze.  The particular arms that were baited remained constant over testing for each rat, but the specific patterns were counterbalanced across rats.  Rats were allowed a maximum of 5 min to retrieve the four pellets.  Errors were divided into RM (i.e.; entering an arm that was never baited) and WM errors (re-entering previously baited arm during a trial).  All entries to arms that were never baited were scored as RM errors, irrespective of whether a rat had already visited that arm in the current or earlier trials of the session.  Initially, rats were given one trial per day.  Once an individual rat reached a criterion of 1 or fewer errors for 3 consecutive days, subsequent sessions consisted of 5 trials per day, separated by an interval of ~ 1 min.  Daily sessions continued until a rat again made 1 or fewer errors on the first trial of the session for 3 consecutive days, after which they underwent surgery.  Following surgery, animals were re-trained, receiving 5 trials/day until they again achieved criterion performance for 3 consecutive days.  109  4.2.4 8-arm Foraging Task  Training was similar to that described above, with the primary exception that rats were required to retrieve food placed on all 8 arms.  Over the course of training, rats were discouraged from using a serial selection strategy by distracting them if they chose more than 4 adjacent arms consecutively. By the end of training, none of the rats were using this type of strategy. As with the RM/WM task, rats were trained using 1 trial/day to a criterion of  ≤ 1 error for 3 consecutive days, and then given 5 trials/day until achieving criterion performance during the first trial.  4.2.5 Spatial Discrimination  For this experiment, animals were implanted with cannulae prior to behavioral training.  In this task, 5 of 8 arms of the maze were blocked using opaque Plexiglass inserts to create a T-shaped configuration, with a stem arm and two choice arms.  For each rat, the orientation of the stem and choice arms relative to the testing room, and the location of a baited arm (left or right) remained constant across trials and training days (counterbalanced across rats).  At the start of each trial, a rat was placed in the stem and required to traverse towards the baited arm, during which the latency to enter an arm was recorded.  Error were scored upon entry into the unbaited arm.   Rats initially received 10 trials/day until making ≤ 2 errors (~ 2 days) and then received 20 trials/day until they reached criterion of ≤ 3 errors/day for 3 consecutive days.  4.2.6 Surgical procedures Upon achieving criterion performance on the RM/WM or 8-Arm Foraging task, rats were fed ad libitum for 2-3 days and then subjected to surgery.  Rats were anesthetized with ketamine/xylazine (100/7 mg/kg) and implanted with bilateral 23-gauge stainless-steel guide cannulae that targeted the prelimbic region of the medial PFC using the following stereotaxic coordinates (flat skull: AP = +3.4 mm; ML = +/- 0.7 mm from bregma; DV = -3.0 mm from dura; 110  Paxinos and Watson, 1998).  Rats were then given ~7 days to recover from surgery.   4.2.7 Drugs & microinfusion procedures Reductions PFC GABAA transmission was achieved using the receptor antagonist bicuculline methobromide (12.5 or 50 ng in 0.5 µl). These doses have been shown to be effective at altering forms of cognition when infused into the PFC, but are below the threshold for inducing seizures (Enomoto et al., 2011; Paine et al., 2011).  Similarly, in the present study, no signs of seizures, such as wet dog shakes, were observed. Previous neurophysiological studies have shown that intra-PFC infusions of 50 ng bicuculline can enhance neural activity in downstream regions for at least 2 hours (Enomoto et al., 2011), which was well within the timeframe required for rats to complete a test session.  Inactivation of the PFC was achieved using a cocktail of the GABAA receptor agonists, baclofen and muscimol (100 ng each in 0.5 µl). All drugs were dissolved in 0.9% saline. Bilateral infusions were made through 30-gauge injectors extending 0.8 mm below the guide cannulae. Saline, bicuculline, or baclofen/muscimol were infused at a rate of 0.5 μl/75 s. Following infusions, the injectors were left in place for 1 min to allow for diffusion. Rats were then placed back in their home cages for 10 min, after which, testing commenced. On test days, rats remained on the maze until all baited arms were entered. A within-subjects design was used for all experiments.  On the first infusion test day, a proportion of the rats in each group received saline infusions, and the remaining rats received drug (bicuculline or baclofen/muscimol).  After the first test day, rats were retrained until they again achieved criterion performance, then received a second counterbalanced test day. One or two days before their first microinfusion test day, rats received a mock infusion procedure, during which obdurators were removed from the guide cannulae, and replaced with stainless 111  steel injectors for 2 min, without an infusion. Rats subsequently received daily training sessions on their respective task until they again achieved criterion performance. On the day after criterion performance was re-established, rats received a second counterbalanced infusion of saline, or drug, and this continued until rats had received all designated infusion treatments. 4.2.8 Histology After testing, rats were sacrificed in a CO2 chamber. Brains were removed and fixed in a 4% formalin solution for at least 24 h. Brains were frozen, sliced in 50 μm sections, mounted, and stained with Cresyl Violet. The positions of all acceptable placements for each experiment are presented in Figure 4.1. 4.2.9 Data Analysis For the radial maze tasks, the main dependent variable of interest was errors committed during test days. For these analyses, we compared errors made during the first trial of the session to the average number of errors made on trials 2-5, to assess whether treatments disproportionately affected errors made during the latter trials, when rats would be more susceptible to proactive interference from previous trials. These data were analyzed with two-way repeated measures ANOVA, with treatment and trial (1 vs 2-5) as two within-subjects factors. For the RM/WM task, we also analyzed the total number of RM and WM errors across the 5 trials with a similar ANOVA model, with treatment and error type as two within-subjects factors. For the latency data, we analyzed the time to enter the first arm on each trial (time to initiate) and the average time per subsequent choice with separate two-way ANOVAs, with treatment and trial as within-subjects factors. Lastly, for the spatial discrimination, error data were collated into 5 blocks of 4 trials. These data were analyzed with two-way ANOVAs, with  112   Figure 4.1. Location of all acceptable infusion placements within the medial PFC in experiments in Chapter 4. Different symbols denote the locations of infusions for rats for each of the four specific experiments.    113  treatment and trial block as factors. Multiple comparisons were made with Tukey’s post hoc test when appropriate.  4.3 Results 4.3.1 Reduced prefrontal GABAA receptor transmission and RM/WM performance  Pharmacological reduction in PFC GABAA transmission via infusion of low doses of bicuculline markedly impaired performance (Figure 4.2A).  Analysis of the error data yielded a significant main effect of treatment (F2,14=6.51, p<0.05) and a treatment x trial interaction (F2,14=4.34, p<0.05).  Simple main effects analyses revealed that, contrary to our expectation, infusion of both the 12.5 and 50 ng dose increased errors (p<0.05) relative to saline treatments during the first trial of the session.  However, during subsequent trials, only the 50 ng dose impaired performance (p<0.05).  Interestingly, this analysis also revealed that under control conditions, the number of errors made during the first trial of the session did not differ from the average number of errors made over trials 2-5 (p>0.40, n.s.).  This indicates that any potential proactive interference that may have been induced by this massed trials procedure was insufficient to increase erroneous responding on subsequent trials.  Likewise, the number of errors committed on trials 1 vs 2-5 did not different after infusions of either dose of bicuculline (both p>0.50, n.s.).  Two rats not included in these analyses had placements ventral to the medial PFC, and did not display marked alterations in performance after bicuculline vs saline. A separate analysis of RM and WM errors across the 5 trials revealed a main effect of treatment (F2,14=4.32, p<0.05) and a treatment x error type interaction (F2,14=3.87, p<0.05; Figure 4.2B).  Breakdown of the interaction revealed that RM and WM errors were increased by   114   Figure 4.2. Prefrontal GABAA antagonism blockade disrupts spatial RM/WM performance on a radial maze task. For this and all other figures, values displayed are the mean +/- SEM. A, Infusion of both a 50 and 12.5 ng dose of bicuculline increased the number of errors on Trial 1, while only the 50 ng dose increased the average number of errors on Trials 2-5. B, Both doses of bicuculline increased the total number of RM and WM errors committed over the 5 trials of a test session.  However, the 50 ng dose caused significantly more reference memory errors than the 12.5 ng dose. C, Intra-PFC infusions of 50 ng bicuculline increased the time to initiate a trial, specifically in Trials 1 and 2. The 50 ng dose increased the average time per subsequent choice across all trials.  p<0.05 vs saline, +p<0.05 50 ng vs 12.5 ng.   115  both doses of bicuculline (p<0.05), with the 50 ng dose causing significantly (p<0.05) more RM errors when compared to the 12.5 ng dose.  Thus, antagonism of PFC GABAA receptors disrupted both RM and WM performance, with the 50 ng dose being having a more pronounced deleterious effect on RM and overall performance during the latter trials.  Reducing prefrontal GABAA transmission also caused a pronounced and dose-dependent increase in the time to initiate a trial (Figure 4.2C) and the average time per subsequent choice (Figure 4.2D).  Analysis of the time to initiate data yielded a main effect of treatment (F2,14=4.43, p<0.05), and a treatment x trial interaction (F8,56=2.46, p<0.05).  Subsequent analysis confirmed that only the 50 ng dose was effective at increasing initiation latencies, with this effect being statistically significance during the first 2 trials (p<0.05).  For the average time per subsequent choice, the analysis only yielded a main effect of treatment (F2,14=8.50, p<0.01) but no interaction (F8,56=1.25, n.s.).  Again, only the 50 ng dose was effective at increasing choice latencies, although the effects of this dose during the first trial were quite variable (range 26-92 s).   4.3.2 Reduced Prefrontal GABAA transmission and performance of an 8-arm foraging task The first experiment revealed that intra-PFC GABAA antagonism impaired both RM and WM, with this effect manifesting during the first trial of the test session and persisting over subsequent trials.  Notably, under control conditions, rats did not commit more errors during latter trials relative to the first, suggesting that the cognitive load of this procedure did not induce sufficient proactive interference to degrade performance.  This may be related to the fact that once rats were well-trained on this task, they made only 4-6 choices per trial, whereas in  116   Figure 4.3.  Prefrontal GABAA receptor antagonism disrupts short-term memory on an 8-arm foraging task.  A, Intra-PFC infusion of 50 ng bicuculline increased the total number of errors committed across the 5 trials of a test session. B, These treatments the number of errors made during for Trial 1 and subsequent errors during Trials 2-5.  Under control conditions, rats made significantly more errors during Trials 2-5 relative to the first trial, indicative of an increase proactive interference during the latter trials.  C, Infusion of bicuculline increased the time to initiate during the first trial of a test session. D, These treatments increased the average time per subsequent choice during Trials 1-3. p<0.05 vs saline, +p<0.05 Trials 1 vs 2-5 for saline condition.   117  previous studies using this procedure used a 12-arm maze so that well-trained rats made 8-12 choices per trial (Meck and Williams, 1999).  Thus, in the second experiment, we trained a separate group of rats on a short-term memory version of the radial maze task, wherein subjects had to visit all 8-arms of the maze in a non-repetitive manner.  The expectation was that increasing the number of items to be remembered across each trial would make rats more susceptible to proactive interference and lead to a greater number of errors during latter trials.  This in turn would enable us to better evaluate how PFC GABA transmission may modulate cognitive performance when levels of proactive interference were higher.  In this experiment, we only tested the effect of the 50 ng dose of bicuculline, as this dose was more effective at impairing performance in the last experiment.    Similar to the first experiment, intra-PFC bicuculline produced a profound impairment in performance of this task that emerged on the first trial.  Analysis of the error data revealed a significant main effect of treatment on the total number of errors (F1,5=33.40, p<0.01; Figure 4.3A).  There was no significant treatment x trial interaction (F1,5=0.001, n.s.; Figure 4.3B), nor was there a significant main effect of trial (F1-5=0.52, n.s.).  However, a direct comparison of the number of errors committed after control treatments on trials 1 vs. 2-5 revealed that rats made significantly more errors in later trials in comparison to the first (F1,5=10.00; p<0.05).  This indicates that the use of a massed trials procedure with this task was sufficient to induce proactive interference.  However, this effect was occluded by the substantial increase in errors after bicuculline treatments.  One rat not included in these analyses had placements in the medial orbitofrontal cortex and also exhibited impairments on this task.  Reducing prefrontal GABAA transmission again increased the time taken to initiate each subsequent choice (Figure 4.3C and 4.3D).  A repeated measures ANOVA revealed a significant 118  treatment x trial interaction on the time to initiate a trial (F4,20=2.98, p<0.05).  Simple main effects analysis revealed that time to initiate was increased only on the first trial (p<0.01).  For average subsequent choice time, a main effect of treatment and treatment x trial interaction were observed (F1,5=18.13, p<0.01; F4,20=19.15, p<0.001) in that bicuculline increased choice latencies on trials 1-3 (p<0.05).  Again, the effects of bicuculline on this measure during the first trial were highly variable (range 20-73 s).  However, unlike what was observed in Experiment 1, here we observed a significant negative correlation between choice latencies and errors during the first trial (r=-0.75, p<0.05).  Thus, even though reducing PFC GABA signalling impaired accuracy and increased choice latencies on two separate radial maze tasks, the relationship between these two effects was not consistent across experiments. Collectively, the results of this experiment demonstrate that reducing PFC GABA transmission also impairs short-term memory assessed with a radial maze procedure, but this effect did not appear to be driven by an increased susceptibility to proactive interference. 4.3.3 PFC inactivation and RM/WM performance Reducing PFC GABAA transmission markedly impaired both RM and WM assessed with two different radial-arm maze tasks.  These findings raise the question regarding the normal contribution of the PFC in facilitating search behavior guided by reference/working memory.  Previous studies have shown that reversible inactivation of the medial PFC does not affect continuous foraging on a radial-arm maze task similar to the one used in the second experiment (Seamans et al., 1995).  In comparison, large aspiration lesions of the medial PFC destroying multiple subregions induce relatively transient deficits in acquiring or re-learning RM/WM versions of this task (Kolb et al., 1983; Kesner et al, 1987).  What was unclear is how acute and more circumscribed suppression of neural activity within the medial PFC would affect RM/WM  119   Figure 4.4. PFC inactivation does not affect spatial RM/WM performance.  A, Inactivation of the medial PFC with infusion of the GABA agonists baclofen and muscimol had no effect on the number of errors on Trial 1 or the average number of errors on Trials 2-5. B, PFC inactivation did not alter on the total number of RM or WM errors relative to saline.  C, Inactivation of the PFC increased the latency to initiate a trial during the first trial of a test session, but did not affect the average time per subsequent choice, (D).     120  performance using a massed-trials version of this task.  Thus, we assessed the effects of reversible inactivation of the medial PFC on performance of the same task used in the first experiment in a separate group of well-trained rats.   In stark contrast to the effects of prefrontal GABAA receptor antagonism, PFC inactivation had no effect on performance of the task (total errors main effect of treatment; F1,7=1.10, n.s.; treatment x trial interaction; F1,7=1.58, n.s.; Figure 4.4A).  No difference in the number of errors made over trials 2-5 in comparison to trial 1 was observed for animals receiving infusions of either saline or inactivation, indicating lack of the proactive interference effect (F1,7=0.35, n.s.).  Separate analysis of RM and WM errors across the 5 trials revealed no main effect of treatment on errors across trials (F1,7=2.16, n.s.), no main effect of error type (F1,7=0.30, n.s.) and no treatment  x error type interaction (F1,7=3.18, n.s.). Thus, PFC inactivation had no overall effect on RM or WM performance (Figure 4.4B).   PFC inactivation produced no main effect on the time taken to initiate a trial (F1,7=3.01, n.s.), but did reveal a significant treatment x trial interaction on the time to initiate (F4,28=5.39, p<0.01).  Subsequent analysis revealed that PFC inactivation caused a modest, but significant (p<0.05) increase in the time taken to initiate at the start of the first trial (Figure 4.4C).  PFC inactivation did not affect the average time taken per subsequent choice (main effect of treatment (F1,7=2.36, n.s.); treatment x trial interaction (F4,28=0.54, n.s; Figure 4.4D). Thus, PFC inactivation had little impact on speed of processing times. 4.3.4 Reduction of prefrontal GABAA transmission and spatial discrimination The impairments induced by prefrontal GABAA antagonism prompted us to explore whether these effects reflected a more general disruption in the ability to discriminate between different maze arms.  In so doing, rats were well-trained on a simple left/right discrimination  121    Figure 4.5.  Prefrontal GABAA receptor antagonism and performance of a 2-arm spatial discrimination. A, Intra-PFC infusion of 50 ng bicuculline did not alter the total number of errors committed across the 20 trials of a spatial discrimination test session. B, These treatments did cause a slight increase in errors made during the first 4 trials, but this effect disappear during the latter trials of a test session.  C, Choice latencies were also increased during the first 4 trials.    122  task, whereby one of two arms of the maze were consistently baited, and rats merely had to discriminate between them and enter one of them to obtain food.  In this experiment, we again only tested the 50 ng dose of bicuculline.   Infusions of bicuculline had no overall effect on accuracy in the spatial discrimination task across 20 trials, as the analysis revealed no significant main effect of treatment (F1,6=0.36, n.s; Figure 4.5A).  However, a significant treatment x trial block interaction was observed (F4,24=4.57, p<0.01).  This effect was driven by a slight increase in errors committed during the first four trials of the session, with rats making ~1 more error after bicuculline treatment relative to their performance on saline.  However, this effect was transient, as no differences in accuracy were observed during the latter trials (Figure 4.5B).  Data from 3 rats excluded from analysis for inaccurate placements showed that they did not appear to exhibit impairments on either the early or later trials.   Bicuculline infusion did not cause an overall change in the amount of time required to complete a session (Fig 4.5C; F1,6=2.39, n.s.).  However, a significant interaction between treatment and trial block was revealed (F4,24=8.001, p<0.001). This interaction was driven by an increase in response times in the first block of trials, but not on subsequent trials (Figure 4.5D).  Thus, aside from a slight and transient increase in errors committed at the start of the test session, reducing PFC GABA transmission did not impair performance of a relatively simple spatial discrimination.     4.4 Discussion The primary findings of the present study are that reducing GABAA receptor activity in the medial PFC results in a profound impairment in search behavior guided by RM or WM.  The 123  effects on search accuracy were accompanied by increases in choice latencies.  However, decreasing PFC GABAA transmission did not markedly impair performance of a 2-arm spatial discrimination, suggesting that impairments observed in the other experiments were unlikely to be the result of non-specific deficits in motor, motivational or spatial discrimination processes.  Thus, these results suggest that intact PFC GABA transmission is essential for using spatial information to guide search behavior in complex environments, and disinhibition of the medial PFC can lead to pronounced impairments in these functions.  4.4.1 Prefrontal GABA signalling and reference/working memory A key question we attempted to address in this study is whether prefrontal GABA transmission facilitates memory performance when information about previous choices may interfere with ongoing behavior (i.e.; conditions that incur relatively high levels proactive interference).  This hypothesis was based in part on the idea that if PFC GABA aids in filtering out irrelevant information, then memory functioning under conditions that provide interference effects from previous choices may be more susceptible to disruption by reduced PFC GABA signalling.  To this end, we used a variation of the radial arm maze procedure where multiple trials were presented with a short inter-trial delay, requiring rats to discard information from previous trials and focus on information relevant to the current trial.  Previous studies using similar procedures have shown that well-trained rats typically make more errors during later trials relative to the first trial of a daily session (Roberts and Dale, 1980).  However, with the RM/WM task used in the first experiment, this effect was not observed under control conditions.  Moreover, PFC GABA antagonism did not preferentially increase errors during the latter trials, but instead caused a considerable increase in errors apparent on the first trial of a session.  The lack of a proactive interference effect prompted another experiment where we increased 124  cognitive load by using a short-term memory version of the task with eight arms baited.  Here, we did observe an increase in errors in later trials relative to the first under control conditions.  Yet, reducing prefrontal GABA signalling did not appear to exacerbate this effect, as these treatments again increased errors during the first trial of the session that persisted over subsequent trials.  Thus, although intact prefrontal GABA transmission may be essential for efficient search guided by spatial RM or WM, reducing this activity does not appear to increase sensitivity to proactive interference. In the first experiment, disinhibition of the PFC increased both RM and WM errors.  Indeed, animals were equally likely to enter an arm that had never been baited as they were to re-enter arms visited previously during a trial.  The neural substrates that underlie learning of this task have been studied in considerable detail.  Pre-training lesions of the hippocampus and medial thalamus impair both RM and WM (Olton and Papas, 1979; Pothuizen et al., 2004; Stokes and Best, 1990).  In comparison, lesions to the dorsal striatum or nucleus accumbens shell selectively impair RM and WM, respectively (Packard and White, 1990; Jongen-Rêlo et al., 2003).  The observation that reducing PFC GABA activity in well-trained rats non-selectively disrupted both RM and WM performance suggest that these deficits may not reflect an impairment in mnemonic processing per se.  Rather, we interpret these findings to indicate that disinhibition of the medial PFC disrupts the utilization of information processed by short and long-term memory systems to implement an efficient search strategy.  In contrast to the effects of PFC GABA antagonism on the eight-arm maze tasks, similar treatments did not impair overall performance on a simpler two-arm spatial discrimination.  Bicuculline infusions did cause a slight reduction in accuracy early in the test session (~1 more error), but this effect quickly dissipated as trials progressed.  These treatments may have caused 125  a transient disorientation that interfered with approach towards the baited arm on the first few trials.  The relatively muted nature of this effect contrasts with the pronounced impairment in performance during the more complex maze tasks that persisted over the entire course of the test sessions.  Importantly, because impairments on the two-arm spatial discrimination task did not persist across the session, it is unlikely that the deficits in performance observed in the other experiments are attributable to a general disruption in the ability to discriminate between arms or motivational processes. One of the more remarkable features of the present findings is that although disinhibition of the PFC impaired radial maze performance, suppression of PFC neural activity with GABA agonists had no effect on search accuracy guided by RM and WM.  Similarly, previous studies have shown that medial PFC inactivation does not disrupt performance of a short-term memory version of the eight-arm maze task similar to the one used here, although these manipulations do impair performance when a delay is inserted between choices (Seamans et al., 1995).  As such, the present findings demonstrate that memory processes not normally dependent on the PFC may nevertheless be disrupted by disinhibitory increases in PFC activity.  The prelimbic PFC sends excitatory projections the dorsomedial striatum, the nucleus accumbens, midline thalamic nuclei and also indirect input to the hippocampus, via the entorhinal cortex (Sesack et al., 1989; Brog et al., 1993; Sesack & Pickel, 1992; Prasad & Chudasama, 2013).  These regions have been shown to play an essential role in enabling efficient search behavior guided by spatial RM and/or WM (Becker et al. 1980; Floresco et al. 1997; Jongen-Rêlo et al. 2003; Olton and Papas 1979; Pothuizen et al. 2004; Schacter et al. 1989; Seamans et al. 1995; Stokes and Best 1990). Given these anatomical considerations, it is possible that impairments in search behavior induced by PFC disinhibition may be driven by aberrant patterns of excitatory outflow to downstream 126  structures.  This in turn may interfere with normal patterns of neural activity in these regions that regulate memory processes used to guide behavior.  As such, the present findings highlight that impairments in cognition arising from a disinhibited or “noisy” PFC may not always reflect local perturbations of task-related neural activity, but also result from a cacophonous influence that PFC projections exert over mnemonic-related activity in downstream brain regions.   4.4.2 PFC GABA transmission and choice latencies In addition to disrupting efficient search behavior, antagonism of prefrontal GABAA receptors consistently increased the latency to initiate choices.  This observation mirrors previous studies from our laboratory, in which pharmacological reduction in PFC GABA signalling increased choice latencies on a delayed-response variant of the radial maze task (Enomoto et al., 2011).  However, in the same study, we observed that similar treatments actually increased spontaneous locomotion measured in activity chambers (Enomoto et al., 2011).  Thus, is unlikely that increased choice latencies reflect reduced locomotion function.  Unfortunately, we were unable to quantify changes in activity patterns induced by bicuculline infusions. Anecdotally, following these treatments, rats were active within the center of the maze, but were much slower to move from the centre and initiate their first choice.  Furthermore, upon entering an arm, they spent a disproportionate amount of time on it following reward retrieval before initiating another choice. Thus, we interpret these effects as indicating that rats were more hesitant to initiate specific choices, rather than having a reduced level of activity overall.  It is notable that the effect of intra-PFC bicuculline on choice latencies was most prominent during the earlier trials, whereas the increase in errors persisted across all trials of the test session. This suggests that the pharmacokinetics of the effects of these treatments on these two measures (accuracy vs choice latencies) differ considerably.  Furthermore, these findings, in addition to the fact that increased 127  choice latencies during the first trial showed opposing relationships to the number of errors, suggest that slower responding induced by PFC disinhibition may not be related to impairments in performance accuracy.    On the other hand, inactivation of the PFC, which did not affect performance accuracy, caused a relatively modest increase in the latency to initiate responding on the first trial, and did not affect subsequent choice latencies.   In comparison, similar treatments have been reported to increase choice latencies and impair performance during tests of working memory capacity (Davies et al. 2013b).  Thus it appears that the effects of PFC inactivation on speed of responding are dependent on whether this region actually contributes to task performance. 4.4.3 Implications for schizophrenia Converging evidence from post-mortem and electrophysiological studies of schizophrenic brains, as well as findings from animal models of the disorder suggest schizophrenia is associated with disruptions in GABAergic transmission in the frontal lobes (Akbarian et al. 1995b; Amitai et al. 2012; Gonzalez-Burgos and Lewis 2012; Richetto et al. 2014; Thomases et al. 2013). These pathophysiological alterations may contribute to cognitive impairments associated with the disease.  A primary goal of the present study was to examine how reductions in prefrontal GABA signalling may affect cognitive and mnemonic functions that are impaired in schizophrenia to provide translational evidence that may support this assertion.  In this regard, the findings of a recent study by Spieker and colleagues (2012) are particularly striking when viewed in light of the present findings.  In that study, the authors employed a back-translational approach, wherein schizophrenic patients performed a virtual version of the RM/WM radial arm maze task nearly identical to the one used in the present study.  When compared to healthy controls, patients showed reduced efficiency and accuracy of search, with 128  increased number of both RM and WM error and time required to complete each trial.  In a similar vein, schizophrenic patients performing the CANTAB spatial working memory search task show an increase tendency to return to locations which had already been reinforced during a previous searching sequence (Pantelis et al. 1997).  Notably, these deficits were only observed when subjects had to search amongst 4-8 locations, but not fewer locations.  The constellation of effects across these two studies were nearly identical to what we observed following pharmacological reductions in prefrontal GABA signalling.  The remarkable qualitative similarities between these two sets of findings suggests that diminished PFC GABA neurotransmission may be a key contributing factor to the impairments in RM and WM observed in schizophrenia.  Deficits in goal-directed navigation in schizophrenia have also been observed used other paradigms, such as virtual versions of the Morris water maze task and way-finding in other virtual environments (Hanlon et al. 2012; Zawadzki et al. 2013). Previous work suggests that spatial deficits in schizophrenia are a result of deficient ‘allocentric’ navigation and formation of cognitive maps based on external cues, rather than ‘egocentric’ strategies used to navigate to a target.  With respect to the present study, prefrontal GABA antagonism produced profound deficits in more complicated 4 or 8 choice versions of the radial arm maze task, where the subject must rely on external cues to navigate through the maze environment. While impairments in spatial search observed in schizophrenia patients have typically been thought to reflect deficits in hippocampal functioning, the present findings suggest that deficient prefrontal GABA signalling may also contribute to these types of abnormalities.    129  4.4.4 Conclusions To summarize, the present findings highlight how optimal GABA signaling within the PFC is essential for efficient performance of cognitive tasks guided by spatial RM and WM.  Decreasing prefrontal GABAA transmission impaired search behavior on a radial maze in a manner that was nearly identical to those observed in schizophrenia patients.  These data complement a growing preclinical literature implicating intact PFC GABA signalling in the mediation of numerous domains of cognition, and that disinhibition of the PFC can disrupt cognitive and affective dysfunction that in many instances is qualitatively similar to what is observed in schizophrenia (Japha and Koch, 1999; Enomoto et al, 2011; Paine et al., 2011; Piantadosi and Floresco, 2014).  In turn, these data provide additional support for the idea that dysfunction in GABA transmission in the frontal lobes may contribute to cognitive dysfunction in schizophrenia, and further suggest that treatments designed to normalize inhibitory signalling may provide some therapeutic benefits for ameliorating these symptoms. Importantly, pharmacological reduction of PFC GABA signalling impaired performance of cognitive tasks that are not normally mediated by the PFC, suggesting that a hyperactivity of the frontal lobes may interfere with cognitive/mnemonic processes mediated by other circuits. Deficient PFC inhibition may lead to increased PFC projection neuron activity, in turn altering activity patterns in PFC terminal regions associated with task performance. Given that the PFC is one of the most densely connected structures in the mammalian brain, there are many potential regions in which altered activity following PFC disinhibition could disrupt performance of the RM/WM task. The experiments described in Chapter 5 will examine whether PFC GABAA receptor antagonism alters neuronal activation in PFC terminal regions, both in animals at rest and after training and testing in the RM/WM task. 130  Chapter 5: Disinhibition of the prefrontal cortex leads to brain-wide increases in neuronal activation that are modified by spatial learning  5.1 Introduction GABAergic interneurons are a population of cells with immense diversity in their molecular, morphological and neurophysiological specializations (Tremblay et al. 2016). However, they share the common feature of regulating excitability and output of other neurons, in turn enabling co-ordination of activity at the microcircuit or network level. In the prefrontal cortex (PFC), interneuron activity and GABA release play a key role in establishing neuronal oscillations that are thought to be a mechanism that enables information processing necessary for many higher-order cognitive functions. For instance, fast-spiking interneurons that express the calcium-binding protein parvalbumin (PV) are necessary for oscillatory activity within the gamma range (30-80 Hz) (Cardin et al. 2009; Sohal et al. 2009).  Gamma oscillations occur when working memory and cognitive control mechanisms are engaged, and abnormal gamma is associated with impairments in these aspects of cognition; as is observed in schizophrenia (Chen et al. 2014; Haenschel et al. 2009; Minzenberg et al. 2010). Deficiencies in prefrontal GABAergic transmission have been observed in several pathological and non-pathological conditions, including schizophrenia (Benes 1995; Gonzalez-Burgos and Lewis 2012; Tse et al. 2015b), depression (Lener et al. 2017; Luscher et al. 2011) and aging (McQuail et al. 2015). In schizophrenia, one of the most reliable pathologies found in post-mortem brain is a decrease in mRNA and protein expression of the GABA synthesis enzyme, GAD67 in prefrontal regions (Akbarian et al. 1995b; Curley et al. 2011; Guidotti et al. 2000; Volk et al. 2000). PET imaging studies using a benzodiazepine site ligand revealed disrupted PFC GABAergic transmission in schizophrenia patients (Frankle et al. 2015), while the 131  largest in vivo MRS study of schizophrenia to date found evidence for a decrease in PFC GABA levels in aged individuals (Rowland et al. 2016a). Collectively, these findings suggest that certain symptoms of schizophrenia may be driven by decreased synthesis and release of GABA, especially within the PFC.  Strikingly, many animal models of the disorder, whether neurodevelopmental (François et al. 2009; Tseng et al. 2008), pharmacological (Amitai et al. 2012; Behrens et al. 2007; Morshedi and Meredith 2007) or genetic (Ji et al. 2009; Lee et al. 2013; Shen et al. 2008), also seem to converge on disrupting PFC GABA function. Since changes in GABA function in schizophrenia are detected particularly within PFC regions that regulate higher-order cognition, they are well-placed to contribute to cognitive impairments associated with the disorder.  In recent years, our group and others have investigated how PFC GABAergic transmission regulates cognitive processes mediated by the frontal lobes by assessing how different domains of cognition are affected by pharmacological targeting GABA transmission in the medial PFC of rats. These studies have revealed that reducing PFC GABA signalling produces a myriad of cognitive deficits with relevance to schizophrenia, including impairments in attention (Paine et al. 2011; Pehrson et al. 2013; Pezze et al. 2014), working memory, and cognitive flexibility (Enomoto et al. 2011). Intra-PFC administration of GABAA antagonists also increased phasic dopamine activity and locomotor response to stimulants (Enomoto et al. 2011), led to aberrant salience attribution (Piantadosi and Floresco, 2014), and impaired social behavior and reward-related decision making (Paine et al. 2017)). These latter findings suggest that deficiencies in PFC GABA signalling may also contribute to positive and negative symptoms of schizophrenia, respectively.  132   The experiments of the previous chapter revealed that PFC GABAA receptor antagonism produces a robust impairment in the classical reference/working memory (RM/WM) variant of the radial-maze task. This was associated with increases in both RM and WM errors, mirroring the impairments observed in schizophrenia patients performing a virtual version of this task (Spieker et al. 2012). Another notable finding of this study that even though disinhibition of the PFC induces robust impairments on this task, PFC inactivation did not affect performance. Thus, it appears that manipulations that disinhibit PFC activity can interfere with certain cognitive certain mnemonic or cognitive functions that not normally mediated by the PFC. One possible explanation for these observations is that decreased PFC GABA function may lead to aberrant increases in activity of PFC projection neurons that consequently alters patterns of activation of neurons of downstream regions. This in turn may disrupt normal functioning of these downstream circuits that mediate PFC-independent forms of cognition. In this regard, PFC GABAA receptor antagonism is known to elevate expression of the immediate early gene (IEG) c-Fos in medial PFC (Paine et al. 2011), but how local disinhibition of the PFC may impact neuronal activity in PFC efferent regions is not known. To address this question, we investigated how PFC GABAA antagonism impacts neuronal activation throughout the brain, using c-Fos expression as an index of activation. Given that PFC GABAA antagonism disrupts performance of the non-PFC dependent RM/WM radial maze task, we were particularly interested in how PFC GABAA antagonism affected neuronal activation in animals that were trained and/or tested on this task. To this end, c-Fos expression following PFC GABAA antagonism was measured in behaviorally-naive animals, those that were trained on the RM/WM task, and rats that were trained and performed the behavior on the test day, enabling assessment of how PFC disinhibition alters terminal region neuronal activation following plasticity or 133  behavioral activation associated with spatial memory processes.  The PFC is one of the most densely connected structures within the brain, with many efferent projections that enable top-down control over circuits (Sesack et al. 1989; Vertes 2002; 2004). Therefore, we hypothesized local disinhibition of the PFC would alter neuronal activation in multiple regions receiving inputs from PFC.  We targeted our analyses on regions known to be involved in spatial memory processes, including the striatum (Colombo et al. 1989; Floresco et al. 1997; Packard and White 1990; Schacter et al. 1989), thalamus (Aggleton and Nelson 2015; Harvey et al. 2017; Stokes and Best 1988; 1990), hippocampus (Becker et al. 1980; Duva et al. 1997; Floresco et al. 1997) and parahippocampal cortices (Otto et al. 1997; Pouzet et al. 1999).   5.2 Methods 5.2.1 Subjects Adult male Long Evans rats (Charles River; 275-300 g) were used in all experiments. Rats were acclimatized to the colony for ~1 week before undergoing surgery. After recovery, they were restricted to 85-90% of their free-feeding weight. Food restriction for cage controls commenced on the same day as animals that received surgery. Experiments were conducted in accordance with the Canadian Council on Animal Care and University of British Columbia Animal Care Committee. 5.2.2 Surgical Procedures Subjects were implanted with bilateral guide cannulae targeting the prelimbic (PrL) PFC (flat skull co-ordinates AP: +3.2 mm, M/L: 0.8 mm and DV -2.8 mm).  Anaesthesia was induced using ketamine/xylazine (50 and 10 mg/kg respectively) and maintained using isoflurane vapour. Animals were given Anafen (5 mg/kg, s.c.) as a pre-surgery analgesic, and received the same 134  dose of Anafen for at least two days following surgery. Animals were given ~10 days to recover from surgery before beginning maze training. They were given ad libitum access to food for 3-4 days following surgery, and then were food-restricted for a week prior to beginning maze training. 5.2.3 Experimental Groups & Timeline  The study consisted of 7 experimental groups that differed based on drug treatment, maze training (i.e., behavioral history) and testing following infusion of drug or vehicle. Six of the groups received infusions of drug or saline prior to sacrifice, while the ‘Cage Control’ group did not receive surgery, infusions, or behavioral training/testing, and was included to assess the impact of the surgical and infusion procedures on c-Fos expression. The ‘Saline (SAL) Only’ and ‘Bicuculline (BIC) Only’ groups received only infusions of vehicle or drug respectively (no behavioral training), and were included to assess the effects of intra-PFC GABAA antagonism in baseline conditions. Animals in the ‘SAL Training’ and ‘BIC Training’ groups underwent maze training until reaching criterion performance. On the following day, they received an infusion of vehicle or drug, but were not tested on the maze on the day of sacrifice (i.e. no behavioral activation). These groups were included to assess whether plasticity associated with maze learning had an impact on c-Fos expression, particularly following prefrontal disinhibition. Finally, ‘SAL Test’ and ‘BIC Test’ groups were trained on the maze, received an infusion of vehicle or drug on test days, and performed the task prior to sacrifice. The ‘Test’ groups were included to assess how performance of the maze task impacts c-Fos expression and how this may interact with PFC disinhibition. All animals that received infusions were sacrificed 90 minutes after infusion/behavior. 135  Following food restriction, animals in behavioral training groups underwent maze training until achieving criterion performance. Animals in cage control groups were treated identically, (i.e. they were food restricted, handled and weighed daily) but they received no exposure to the maze. When animals in behavioral groups reached criterion performance, they received an infusion and were sacrificed later that day. Animals in untrained conditions received mock infusions and were sacrificed on the same days, so that animals in behavioral and non-behavioral groups underwent an equal duration of food restriction.  5.2.4 Behavioral procedures All training and testing was conducted during the animals’ light cycle. Testing was conducted on the same 8-arm radial maze employed in the last chapter. On the first day of training, rats were habituated to the maze for ten minutes with no food present. The next day, they underwent a second habituation session with sucrose-based food pellets (BioServ, Frenchtown NJ) scattered across the maze. Animals were habituated once per day until they readily ate the pellets off of the maze. Once habituated, they commenced formal training with one trial per day. In each training session, the same 4 arms of the maze were baited for each animal, with different patterns of baiting across animals. The patterns of baiting contained no more than two consecutive arms that baited. Errors were scored when a rat either (1) made an entry into an arm that was never baited, i.e. RM error or (2) re-entered an arm they entered earlier in the trial, i.e. WM error. Rats were trained until they reached a criterion of approximately 1 or fewer total errors per day. At this point, they received a mock infusion. If the rat achieved criterion performance following the mock infusion, it received the test infusion the next day. If it did not, it was retrained until it displayed criterion performance following a mock infusion. 136  5.2.5 Drugs & microinfusion procedures Antagonism of PFC GABAA receptors was achieved using bicuculline methobromide (BIC). A dose of 50 ng BIC was chosen as it has been shown to produce robust deficits in performance of the RM/WM maze task and of other cognitive tasks (Enomoto et al., 2011; Piantadosi et al. 2016).  Bilateral infusions were made using 30-gauge injectors extending 0.8 mm below the guide cannulae. Saline or drug were infused at a rate of 0.5 ul/75 s. Following infusions, the injectors were left in place for 1 min to allow for diffusion. Rats were then placed back in their home cages. Animals that were tested on the RM/WM maze task were placed on the maze 10 minutes after the infusion.  5.2.6 Tissue processing Approximately 90 minutes after infusion and/or behavioral testing, animals were overdosed with chloral hydrate and transcardially perfused with 30 mL 0.9% saline following by 60 mL of 4% paraformaldehyde. Following extraction, brains were post-fixed in paraformaldehyde overnight and then transferred to 30% sucrose for cryoprotection and storage until slicing. Coronal 40 um sections were collected in series of 10 along the rostral-caudal axis of the brain from the olfactory bulb to the cerebellum and stored at -20 C in antifreeze solution consisting of ethylene glycol, glycerol, and 0.1 M PBS. 5.2.7 Histology To verify cannulae placements, sections containing PFC were mounted onto gelatin coated slides and stained using Cresyl Violet. Acceptable infusion placements are depicted in Figure 5.1.    137   Figure 5.1. Location of all acceptable placements in neuronal activation experiments. Shaded regions indicate the extent of the infusions, with darker regions reflecting areas with most infusion overlap and lighter regions indicating fewer infusions. Symbols are used to denote the group of individual subjects. Filled in symbols indicate BIC-treated animals, while open symbols indicated SAL treated animals of Behav Only (●/○), Trained (▲/∆) and Tested (■/□) groups.    138  5.2.8 c-Fos Immunohistochemistry The tissue was pre-washed at 4°C overnight in 0.1 M PBS in order to remove antifreeze solution. Peroxidase activity was neutralized using a 30 min room temperature incubation in 0.3 % hydrogen peroxide, followed by 3 X 10 min washes in 0.1 M PBS. Sections were incubated at 4 C overnight in 1:1000  rabbit c-Fos antibody (SantaCruz Antibodies) diluted in 0.04% Triton-X and 3% normal goat serum in 0.1 M PBS. Following washing, tissue was incubated at 4 C overnight in goat anti-rabbit biotinylated IgG (Vector Labs). The sections were then washed, incubated in ABC solution for 1 hr at room temperature, and washed again before being developed using DAB solution (Vector Labs). Sections were then mounted onto Superfrost microscope slides, dehydrated and cleared in xylene and coverslipped using Permount. 5.2.9 c-Fos Quantification The % immunoreactive (IR) area was quantified as an estimate of cell number as in Yagi et al. (2016). Sections were examined and photographed at 40X magnification using a Nikon E600 microscope. Digitized images were analyzed using ImageJ. The average optical density of 6 non-IR regions was used as a measure of background. IR area was then considered to be parts of the image with an optical density of 1.5 times the background. Brain regions of interest were traced, and the total area and IR area of the tracing recorded. Quantification of IR regions was typically carried out on at least 3-4 sections, and the reported % IR area reported is an average of these measurements. Regions of interest with abbreviations are depicted in Figure 5.2.  5.2.10 Data Analysis Behavioral data from animals that performed the task on infusion test days (SAL Test and BIC Test) were analyzed with a mixed ANOVA with group as a between-subjects factor and error  139   Figure 5.2. Regions of interest for neuronal activation experiments. OFC: lateral orbitofrontal cortex; Cing: cingulate cortex; PrL: prelimbic medial PFC; IL: infralimbic medial PFC; Core: nucleus accumbens core; Shell: nucleus accumben shell; M1: primary motor cortex; S1: primary somatosensory cortex; BLA: basolateral amygdala; CeA: central nucleus of amygdala; PV: paraventricular nucleus of thalamus; CM: centromedial nucleus of thalamus; MD: mediodorsal nucleus of thalamus; AM: anteromedial nucleus of thalamus; Rh: rhomboid nucleus of thalamus; Re: nucleus reuniens of thalamus; DG: dentate gyrus of hippocampus; CA3/1: cornu ammonis region 3/1 of hippocampus; vSub: ventral subiculum; Peri: perirhinal cortex; Ent: entorhinal cortex.  140  type (RM or WM) or latency type (time to initiate, TTI; or average time per choice, ATC) as repeated measures.   The c-Fos data were subjected to a multi-tier analysis.  The initial analysis compared cFos expression in 23 different brain regions from animals with no behavioral training (baseline - Cage Controls, SAL only and BIC only groups), to clarify how PFC disinhibition may increase c-Fos expression in the absence of any other behavioral manipulation. These data were analyzed with a mixed ANOVA with treatment group as a between-subjects factor and brain region as a repeated measure, and multiple comparisons were conducted using Dunnett’s test with the SAL only group as a control.   The more comprehensive analyses compared baseline groups with those that were tested and/or trained on the task.  These analyses consisted of a series of factorial ANOVAs, comparing differences in c-Fos expression within related groups of brain regions, to reduce the number of comparisons made.  These regions included frontal cortical, striatal, thalamic, amygdalar, hippocampal and para-hippocampal regions. Each of these analyses were comprised of three-way ANOVAs, with drug treatment (SAL/Control or BIC) and behavioral history (No training, training, training + testing) as between-subjects factors and brain region as a within-subjects factors. The results of these ANOVAs were further decomposed with additional ANOVAs and Tukey’s tests as necessary. In these analyses, the primary comparisons of interests were main effects of treatment, treatment x history, treatment x region and the three-way interaction. For the hippocampal regions, dorsal versus ventral position was included as an additional within-subjects factor, with the four-way interaction also being of interest.   Data from 1-2 brain regions were missing in a small number of animals (n= 6, total data points missing = 9/920 total points). These missing values were imputed using expectation 141  maximization algorithm in Systat (version 13.00.05, Systat Software Inc), a method of imputation which does not alter the means of the groups. Inclusion of the imputed data points did not alter the outcome of any analyses when compared to those conducted with the data removed, except for analyses of thalamic nuclei, where a main effect of behavior and region x drug interaction that were at trend levels of significance without these imputed data achieved statistical significance with the imputed data included. These effects did not appear to be driven by differences in the Rh and AM nuclei, where the values were missing, as ANOVAs within both structures had the same outcome in datasets with both missing and imputed values. Rather, because imputation enabled inclusion of data from other nuclei for 3 additional subjects in the analysis, these effects approached significance. Furthermore, both the analyses from the dataset with values imputed or missing contained a significant three-way interaction within the thalamus. Thus, we chose to report the analyses containing values imputed using expectation maximization in order to keep a consistent subject number through all analyses.  5.3 Results In the present study, we assessed neuronal activation in a total of 46 animals, 25 of which were trained on the RM/WM task. On test days, trained animals were divided into groups that received only an infusion but were exposed to the maze on that day (‘SAL Train’; n=7, or ‘BIC Train’; n=6), or those that both received an infusion and performed the maze behavior on the test day following an infusion of vehicle or drug (‘SAL Test’; n=6; and ‘BIC Test’ n=6). The untrained groups consisted of animals receiving no infusion (‘Cage Controls’, n=7), or infusions of vehicle (‘SAL Only’; n=6) or drug (‘BIC Only’; n=8). 142     PrL ILCing OFC M1 S1NAc Core ShellDorsal STR PV Rh AM Re MD CM BLACeASub EntPeri DG CA3CA10510152025Cage ControlSAL OnlyBIC OnlyFrontal Lobe Striatum Thalamus Amg HPCPara-HPCregions  Figure 5.3. Neuronal activation data for untrained animals (baseline). Animals receiving intra-PFC BIC infusions had increased c-Fos expression in most structures, in comparison to Cage Controls or intra-PFC SAL.   One notable exception was in HPC subregions, which did not display an increase in c-Fos expression following PFC disinhibition under baseline conditions. p<0.05 vs SAL.  143  Table 5.1. Results of ANOVAs conducted on neuronal activation data in baseline animals in individual brain regions with treatment as between-group factor. Region, cont’d F ratio  Region, cont’d F ratio, cont’d Frontal Lobe PrL  IL  Cing  OFC  M1  S1   Striatum Core  Shell  STR   Thalamus PV  Rh  AM  Re  MD  CM   20.44,*** 10.01,** 9.05, ** 3.12, n.s 12.78,*** 12.22,***   6.67,* 6.58,** 7.20,**   4.31,* 0.02, n.s. 0.95, n.s. 3.80,* 9.56,** 2.85, n.s   Amg  BLA  CeA   Para-HPC regions Sub  Ent  Peri   HPC DG  CA3  CA1  8.05,** 2.37, n.s    5.56,* 9.96,** 6.49,**   0.54, n.s. 0.78, n.s. 0.23, n.s  *, p<0.05; **, p<0.01; ***, p<0.001   144  5.3.1 Effects of PFC GABAA antagonism on neuronal activation under baseline conditions We first analyzed the effects of intra-PFC SAL or BIC infusion on neuronal activation in untrained animals (baseline- Fig. 3). This analysis compared c-Fos expression in 23 brain regions in animals receiving no infusion (‘Cage Controls’, n=7), or infusions of vehicle (‘SAL Only’; n=6) or drug (‘BIC Only’; n=8). PFC administration of BIC led to increased c-Fos expression throughout the brain (main effect of drug treatment: F2,18=25.89, p<0.001). Overall, the levels of c-Fos expression varied significantly across regions (main effect of region: F22,396.=7.01, p<0.001), and  the extent to which BIC treatment enhanced c-Fos expression was different across regions (significant interaction of drug treatment and region: F44,396=2.93, p<0.001). Subsequent partitioning of the interaction term with one-way ANOVAs and Dunnett’s tests with treatment group as a between-subjects factor and ‘SAL Only’ as the control group revealed that c-Fos expression was greater in BIC Only in comparison to SAL Only in most studied regions, as displayed in Fig. 3 and Table 1.  One notable exception to this was within the hippocampus proper, where no differences between groups were found in any subregion (DG, CA3 or CA1).  We chose to combine the data of dorsal and ventral HPC to simplify the analysis, because, though a preliminary analysis revealed that cFos expression was slightly increased in ventral relative to dorsal subregions (F1,18=5.42, p<0.05), dorsal-ventral position did not interact with treatment (both Fs<0.352) and the lack of increased activation following BIC treatment was observed through all HPC subregions.  Finally, throughout the brain, there were no differences in c-Fos expression between SAL Only and Cage Controls in any of the regions assessed, as indicated by a separate ANOVA that compared only these two treatment groups (no effect of treatment (F1,11=0.44, n.s.) or treatment x brain region interaction (F22,242=0.76, n.s.)).  In light of the fact the SAL Only and Cage Control groups did not differ , data from these two groups were 145  combined into a single control group designated ‘Control- No Train’ that was used for subsequent analyses that included data from those trained on the radial maze. Taken together, these analyses indicated that PFC GABAA antagonism induces broad increases in neuronal activation in numerous downstream brain regions. 5.3.2 PFC GABA antagonism and RM/WM radial maze performance Previous studies in our laboratory have shown that intra-PFC GABAA antagonism disrupts performance of the RM/WM radial maze task using a within-subjects design and a task  variant that required animals to perform 5 back-to-back trials of the RM/WM task.  In that study, animals made significantly more RM and WM errors on both the first and subsequent trials of the test session. In the present study, a subset of the animals were trained on a standard, single trial variant of the RM/WM task, and on test days, some of these rats that received either SAL or BIC infusions prior being tested on the maze and subsequent to sacrifice. Analysis of the behavioral data revealed that intra-PFC BIC treatment significantly impaired performance (Figure 5.4A; F1,10=9.98, p=0.01). No significant treatment x error type interaction was observed (F1,10=0.78, n.s.), indicating that intra-PFC BIC induced a comparable increase of both RM and WM errors. Taken together with the findings from our previous study, the present results indicate that PFC GABAA antagonism disrupts both RM and WM processes, and that this impairment is observed using both within- and between-subjects designs and using single or multiple daily trials.  Latency data were analyzed with a similar mixed ANOVA , with treatment as a between-subjects factor and task phase (time to initiate the trial, TTI and average time per subsequent choice, ATC) as a within subjects factor (Figure 5.4B). As was observed in Chapter 4, PFC GABAA antagonism significantly increased both types of choice latencies.  Analysis of the data  146       Figure 5.4. Behavioral data for animals tested in the RM/WM radial maze. A, Subjects receiving intra-PFC BIC infusions made significantly more RM and WM errors in compared to SAL-treated animals. B, Subjects receiving intra-PFC BIC infusions were both slower to initiate the trial and to make subsequent choice. p<0.05, main effect of BIC treatment   147  yielded a significant main effects of treatment (F1,10=6.56, p<0.05) but no treatment x phase interaction (F1,10=1.27, n.s.).  5.3.3 Neuronal activation following training and testing on the RM/WM task Reducing PFC GABA transmission disrupted performance of the RM/WM task, which, notably, is not normally dependent on PFC activity, as inactivation of the PFC does not affect performance on this task.  In light of this consideration, one of the main objectives of this study was to assess how pharmacological reduction of PFC GABA transmission influenced neuronal activation in animals that were trained and tested on the RM/WM task.  Thus, we compared c-Fos expression in groups of animals receiving intra-PFC infusions of BIC or vehicle/control groups that were (1) untrained (‘BIC Only’ and ‘Control-No Training’), (2) trained on the RM/WM task (‘BIC Train’ and ‘SAL Train’) and (3) trained on the RM/WM task had performed the task on test day (‘BIC Test’ and SAL Test’). The data were analyzed using factorial ANOVAs with intra-PFC BIC treatment and behavioral group as between-subjects factors, and brain region as a within-subjects factor. To simplify the analysis, rather than analyzing all brain regions together, data were groups into related brain regions and analyzed with separate ANOVAs. 5.3.3.1 Frontal lobe regions. Neuronal activation data from prefrontal regions that included the prelimbic (PrL), infralimbic (Il), lateral orbitofrontal cortex (OFC), cingulate (Cing), primary motor (M1) and primary somatosensory (S1) were analyzed together using a factorial ANOVA (Figure 5.5). PFC GABAA antagonism significantly increased c-Fos expression in all areas assessed, as indicated by a main effect of treatment (F1,40=70.70, p<0.001). The levels of c-Fos  expression varied between regions (F5,200=7.65, p<0.001), and more importantly, the analysis revealed that the extent to which intra-PFC BIC treatment increased c-Fos expression  148   Figure 5.5. Frontoparietal neuronal activation data in animals trained and tested on the RM/WM radial maze. A, Intra-PFC BIC treatment increased c-Fos expression throughout frontoparietal regions. B-G, representative micrographs of the PrL in all groups. B-D, SAL treatment groups. E-G, BIC treatment groups. B,E, Behav Only groups. C,F, Trained groups. D,G, Tested groups. p<0.05, main effect of BIC treatment.   149  was also different depending on region (treatment x region interaction; F5,200=2.67, p<0.05). Partitioning this interaction revealed that BIC treatment increased c-Fos expression in all brain regions relative to the control groups (Control- No Train, Sal Train and Sal Test; all p<0.01). Moreover,BIC-induced activation was higher in medial PFC regions near the infusion site (PrL, IL, Cing), relative to the other frontal regions (p<0.01).  On the other hand, c-Fos expression in the control groups did not differ across regions. Notably, there was no significant treatment x history, or treatment x region x history interactions (all Fs<1.1, n.s.).  These latter findings indicate that the magnitude of neuronal activation induced by BIC treatment did not differ between rats that had no maze exposure and those that were either trained and/or tested on the maze prior to BIC treatment.    5.3.3.2 Striatum. Analysis of the neuronal activation data from the core and shell of the nucleus accumbens (NAc) and dorsal striatum (dSTR) (Figure 5.6) revealed c-Fos expression was significantly increased throughout these regions following PFC GABAA antagonism (main effect of treatment; F1,40=27.14, p<0.001). The analysis also yielded a significant main effect of region (F2,80=28.46, p<0.001), and treatment x region interaction (F2,80=16.14, p<0.001). Simple main effects analyses revealed that intra-PFC BIC significantly increased c-Fos expression in all regions relative to the control groups (p<0.001), but this effect was significantly greater in the NAc shell (p<0.01) compared to the other two regions. Moreover, there was no main effects or interactions with the behavioral history factor (all Fs<1.72, n.s.). Taken together, this analysis indicated that PFC disinhibition increases neuronal activation throughout the STR, with this effect being most pronounced in the NAc shell. Furthermore, the magnitude of neuronal activation induced by effects of PFC disinhibition was not modulated by previous behavioral history. 150    Figure 5.6. Striatal neuronal activation data in animals trained and tested on the RM/WM radial maze. A, Intra-PFC BIC treatment significantly increased neuronal activation throughout the STR, though the increase was most prominent in the NAc Shell. B,C, Representative micrographs of the NAc of (B) SAL and (C) BIC-treated animals with no behavioral training or testing. p<0.05, main effect of BIC treatment, #, p<0.05, BIC treatment x region interaction.   151  5.3.3.3 Thalamus. Neuronal activation data from the paraventricular (PV), anteromedial (AM), nucleus reuniens (Re), rhomboid (Rh), mediodorsal (MD) and centromedial (CM) nuclei  of the thalamus were analyzed in a factorial ANOVA (Figure 5.7). The analysis revealed PFC GABAA antagonism significantly increased c-Fos expression across these thalamic regions (F1,40=12.31, p=0.001). Similar to PFC and STR, the levels of c-Fos expression varied in different thalamic regions (F5,200=27.45, p<0.001), with this effect accompanied by a treatment x region interaction (F5,200=2.32, p<0.05).  Of particular interest, the analysis also produced a three-way interaction of treatment, region and history (F10,200=3.34, p<0.001).  This was subsequently partitioned with a series of two-way ANOVAs, analysing data from each thalamic region separately, with treatment and history as between-subjects factors.  For the PV, the ANOVA revealed intra-PFC BIC treatment increased c-Fos expression in all behavioral groups (main effect of drug treatment: F1,40=10.83, p<0.01). However, there was also a main effect of behavioral history (F2,40=4.35, p<0.05) and in particular, the effects of behavioral training/testing were different in SAL vs. BIC treatment groups (treatment x behavioral history interaction: F2,40=3.48, p<0.05).  Subsequent analyses further revealed that for the control groups, higher expression of c-Fos was observed in the PV of rats that had performed the task on test day relative to the other groups (F2,23=13.05, p<0.001, and Tukey’s, p<0.05). In comparison, for the BIC-treated groups, behavioral treatment had no significant effect on c-Fos expression within the PV nucleus (F2,17=1.91, n.s.), although the levels of expression was numerically greater in rats with maze experience compared to the BIC-No Training group.  Furthermore, even though BIC treatment significantly increased c-Fos expression within the PV of all behavioral groups, the level of c-Fos expression in the PV was not different between SAL and BIC treated animals who performed the task on the test day (F1,9=0.37, n.s.). Thus, testing on  152    Figure 5.7. Thalamic neuronal activation data in animals trained and tested on the RM/WM radial maze. A, Intra-PFC BIC treatment significantly increased c-Fos expression throughout thalamic regions. In the PV, testing on the maze increased c-Fos expression in SAL-treated animals. No increase in c-Fos expression was observed in the Rh of BIC-treated animals that were untrained. B-E, Representative micrographs of the thalamus of (B) Behav Only, (C) SAL Test, (D) BIC Only and (E) BIC test groups. p<0.05, main effect of BIC treatment. + p<0.05 interaction of treatment and behavioral group. 153  the maze increased neuronal activation in the PV, as did PFC GABAA antagonism. However, there did not appear to be an additive effect of these two treatments. In the Rh nucleus, both PFC GABAA antagonism (F1,40=6.54, p<0.05) and behavioral history (F2,40=4.80, p<0.05) affected c-Fos expression, although no significant interaction between the two factors was observed (F2,40=2.03, n.s.). The main effect of behavior reflected the fact that across both treatment groups, animals with maze experience displayed higher levels of  c-Fos expression compared to the No-Train groups (Tukey’s, p<0.05). Although no significant effect of drug was found when each of the behavioral groups were analyzed separately (all Fs< 3.02, n.s.), both the trained and tested groups showing more c-Fos expression relative to the untrained groups following PFC GABAA antagonism. These analyses revealed that i) maze exposure alone increased activity in the Rh and ii) PFC GABAA antagonism specifically increased neuronal activation in the Rh of animals that had been trained on the RM/WM task. For the remaining thalamic nuclei (AM, Re, MD and CM), individual factorial ANOVAs revealed that only PFC GABAA antagonism impacted levels of c-Fos expression (main effect of treatment; all Fs>5.35, p<0.01), with no effects of behavior or interactions of drug and behavior (all Fs<2.2, n.s.). Thus, PFC GABAA antagonism increased neuronal activation in these regions to a similar degree in all groups, irrespective of behavioral history. 5.3.3.4. Temporal lobe regions. Separate factorial ANOVAs were performed for amygdalar (Amg), parahippocampal structures and hippocampus proper (HPC) regions (Figure 5.8).  The Amg regions included the basolateral Amg (BLA) and central Amg (CeA). The analysis revealed a main effect of drug treatment (F1,40=23.91, p<0.01), as well as three-way interaction of drug, behavioral history and region (F2,40=15.16, p<0.05).  PFC disinhibition increased c-Fos expression within the BLA (F1,40=23.60, p<0.001) in a manner that was 154    Figure 5.8. Temporal lobe neuronal activation data in animals trained and tested on the RM/WM radial maze. Intra-PFC BIC treatment increased c-Fos expression in the Amg (A) and other temporal lobe regions (B), including Ent and Peri cortices and Sub. C, Intra-PFC BIC treatment selectively increased c-Fos expression in the HPC animals trained on the maze, but not animals at rest. D, In CA3 of HPC, a significant increase in activation in trained versus untrained animals following BIC treatment was observed only in the dorsal, but not ventral, aspect. No differences were observed between dorsal and ventral DG or CA1 (not shown). Representative micrographs of the dorsal HPC of (E) No Training, (F) SAL Test, (G) BIC Only and (H) BIC test groups. p<0.05, main effect of BIC treatment; + p<0.05 interaction of treatment and behavioral group, with trained animals receiving BIC infusions showing higher levels of activation than untrained animals in all HPC subregions. 155  independent of behavioral history (treatment x behavioral history interaction; F1,40=1.31, n.s.). Analysis of c-Fos expression in the CeA revealed only a main effect of BIC treatment (F1,40=10.41, p<0.01, all other Fs<2.52, n.s.). Moreover, BIC treatment increased levels of neuronal activation to a larger degree in the BLA when compared to the CeA (F1,40=9.02, p<0.01), whereas in in SAL treated animals, no observable differences in neuronal activation relative to behavioral history were observed (F2,23=0.3, n.s.). When taken together, these analyses reveal that BIC treatment increased neuronal activation in both Amg subregions, but to a greater degree in the BLA vs CeA. For the parahippocampal cortical structures, which included the entorhinal (Ent) and perirhinal (Peri) cortices and the ventral subiculum (vSub), only a main effect of BIC treatment was found (F1,40=35.40, p<0.001, all other Fs<1.35, n.s.). Thus, PFC GABAA antagonism increased neuronal activation throughout these structures, regardless of behavioral history. Furthermore, all parahippocampal structures tended to have similar levels of c-Fos expression and did not vary in the extent to which PFC GABA antagonism increased their activation. In contrast to the relatively straightforward effects in the Amg and parahippocampal regions, the effects of PFC disinhibition on neuronal activation in the HPC (including DG, CA3 and CA1 subfields) were more complex.  Note that in our initial experiment, reducing PFC GABAA transmission did not increase c-Fos expression in rats that remained in their home cages for the duration of the experiment (Figure 5.3).  However, when the analyses incorporated data from rats with maze experience, this produced significant main effects of BIC treatment (F1,40=24.71, p<0.001), behavior (F2,40=8.82, p=0.001) and region (F2.80=31.64, p<0.001).  There was also a treatment x region interaction (F2,80=3.28, p<0.05), which reflected higher levels of c-Fos expression in the DG of BIC treated rats relative to the CA3 or CA1.  Of particular interest, 156  this analysis also yielded a significant treatment x behavioral history interaction (F2,40=5.24, p=0.01).  The three-way interaction of drug, behavior and subregion was not significant (F4,80=0.69, n.s.). There was a significant four-way interaction of drug, behavior, subregion and dorsal versus ventral position (F4,80=3.50, p<0.05). The treatment x history interaction was partitioned using separate one-way ANOVAs with behavioral history as a between-subjects factor for data obtained from SAL/Control and BIC treated animals. For SAL/Control rats, the results of the ANOVA only approached trend levels of significance (F2,34=2.98, p=0.07). Notably, c-Fos expression in these animals was highest in the DG and CA3 in rats that performed the task on test day, compared to the other two groups. Thus, in the absence of drug treatment, cognitive engagement may have led to a subtle increase in HPC activation.    In contrast, behavioral history had a major influence on how BIC-treatment influence HPC neuronal activation (F2,34=13.64, p<0.001). Specifically, c-Fos levels were higher in BIC-treated rats that were trained and/or tested on the maze compared those that did not receive any training (p<0.01).  Additional partitioning of the treatment x history interaction confirmed that BIC treatment did not increase c-Fos expression in rats that did not receive training (F1,19=1.38, n.s.). However, for rats in the train and test history conditions, BIC-treatment did increase hippocampal neuronal activation relative to SAL (both Fs>7.97, both p<0.05).   Taken together, these analyses indicate that reducing PFC GABA activity by itself is insufficient to increase neuronal activation in the HPC.  In contrast, it appears that neural plasticity associated with learning a spatial memory task alters PFC output pathways so that disinhibitory increases in frontal lobe activity causes aberrant increases in HPC activation.  157   There tended to be higher activation in ventral HPC in comparison to dorsal HPC, though, this effect was not significant (F1,40=3.2, p=0.081). However, there was a significant four-way interaction of dorsal versus ventral position, drug, behavior, and HPC subregion. (F4,80=3.50, p<0.05). Partitioning of this interaction revealed no differences between the patterns of effects in dorsal versus ventral DG or CA1 (both Fs<1.88, p>0.05), with higher activation observed in trained groups that received BIC. In CA3 of BIC-treated animals, the selective increase in activation in trained animals was observed only in dorsal (Figure 5.8D; F2,40=3.61, p<0.05), not ventral CA3 (F2,40=1.964, p>0.15). However, visual inspection of the ventral CA3 data suggests a similar pattern of effects in comparison to other HPC subregions. Furthermore, in the analyses completed with the whole HPC, there was a significant drug x behavior interaction, but no interaction of drug, behavior and subregion suggesting that neuronal activation following BIC treatment was higher in trained versus untrained animals throughout the HPC.  5.4 Discussion The main finding of the present study is that PFC disinhibition increases neuronal activation in PFC efferent regions, including throughout the STR, thalamus, Amg and other cortical regions. Remarkably, increases in neuronal activation in the HPC, a region that does not receive direct PFC input (Sesack et al. 1989; Vertes 2004), were only observed in animals that had been trained on the RM/WM maze task. A similar pattern of activation was observed in the Rh thalamic nucleus, suggesting that plasticity within the PFC-thalamic-HPC pathway may have potentiated the effects of PFC disinhibition.  Alternatively, plasticity within the HPC or in cortical inputs to HPC may also have contributed to this induction of activation in trained animals. Thus, our findings raise the possibility that experience-dependent plasticity within 158  circuits, in this case induced by spatial learning, may affect the extent to which local imbalances in excitation and inhibition within the PFC affect the activity and function of other brain regions. 5.4.1 Baseline effects of PFC disinhibition In the present study, infusions of GABAA antagonist were predominately localized within the anterior PrL PFC (Figure 5.1), with these treatments causing a marked increase in neuronal activation near the infusion area, as has been reported by other groups (Paine et al. 2011). Comparable levels of activation was observed through adjacent IL and Cing cortical regions. Likewise, increased activation was also observed in more distant cortical regions, including the Ent and Peri cortices. Furthermore, cortical regions that do not receive direct inputs from the PrL, including M1 and S1 (Sesack et al. 1989; Vertes 2004), also showed increases in activation, perhaps via inputs from Cing (Sesack et al. 1989) or relay thalamic nuclei (Haque et al. 2010). However, the magnitude of increase in sensorimotor regions tended to be less than cortical regions that were directly connected to the PrL PFC. Thus, local PFC disinhibition appears to have widespread effects on cortical activity patterns, in some cases even having polysynaptic effects at baseline, though monosynaptic effects tended to be strongest.  PFC GABAA antagonism increased neuronal activation throughout the STR in untrained animals. However, this effect of BIC treatment was greater in the NAc shell in comparison to the core or dSTR. The NAc shell receives direct inputs from more ventral aspects of the medial PFC, x including the PrL and IL cortices (Sesack et al. 1989) where the majority of BIC infusions occurred. On the other hand, the NAc core and dorsal STR tend to receive inputs from more dorsomedial portions of the PFC, i.e. the Cing (Sesack et al. 1989). Disynaptic input via the CM thalamic nucleus, which receives dense innervation from the PrL (Vertes 2004), and projects widely to the dorsal STR (Vertes et al. 2012) may also represent a potential pathway mediating 159  dorsal STR neuronal activation. Thus, similar to cortical regions, PFC disinhibition induced increased activation in regions where input from the PrL PFC may have been polysynaptic, but the magnitude of neuronal activation was larger in regions receiving direct PrL inputs. PFC GABA antagonism also increased neuronal activation in several other subcortical structures, including the Amy, and certain thalamic nuclei.  However, the Rh thalamic nucleus and the HPC notably did not show increases in activation in BIC-treated animals under baseline conditions (no behavioral training). A recent study that employed [18F] flurodeoxyglucose (FDG) PET to assess neuronal activation also found brain-wide increases in glucose uptake following unilateral PFC administration of the same dose of BIC in anaesthetized rats (Parthoens et al. 2015). In that study, analyses of individual brain areas revealed that most regions receiving direct PrL inputs had increased levels of activity following intra-PFC BIC treatment, in keeping with the present findings. However, the previous PET study also reported increased HPC activation in animals at rest. Although both IEG expression and FDG PET are indirect methods of assessing neuronal activation, the specific aspects of neural activity that these techniques measure differs. FDG PET capitalizes on metabolic or vascular changes that are most reflective of synaptic potentials and input strength, rather than action potentials (Magistretti and Allaman 2015). On the other hand, expression of IEGs like c-Fos is induced to mediate transcription of genes necessary for mounting neuronal responses to sensory stimuli, drugs or changes in synaptic inputs. Furthermore, induction of c-Fos is Ca2+-dependent (Lerea et al. 1992), occurring in response to inputs that are repeated or summated (Worley et al. 1993), or when learning or plasticity have occurred (Grimm et al. 1997; Kemp et al. 2013; Morrow et al. 1999). When taken in combination with the observation the c-Fos expression is low at rest (Herdegen and Leah 1998), the present findings suggest that increases in c-Fos expression most likely occurred in 160  regions with the most substantial changes in activity patterns following PFC GABAA antagonism. Indeed, the threshold for induction of c-Fos expression is higher than that of another IEG, zif268 (Worley et al. 1993), and the same may also be true of the FDG signal. Nevertheless, the observation that PFC GABA antagonism did not increase c-Fos expression in all brain regions at rest reflects that structures vary in their responsiveness to increased excitatory outflow from the PFC that results from disinhibition of this region. In general, regions that received direct PrL inputs, tended to have a larger induction of c-Fos expression, although some thalamic nuclei, including the Rh, AM and CM, which receive substantial PrL input (Vertes 2002; 2004), did not show significant changes in activation under baseline conditions. Therefore, additional factors including numbers of PFC projection neurons targeting the area, degree of axonal arborisation of PFC efferents, number of polysynaptic PFC inputs into the region or local inhibitory tone may also regulate changes in activation in response to altered PFC inputs. 5.4.2 Effects of PFC disinhibition following spatial learning or performance In most of the brain regions examined, the effects of PFC disinhibition were the same in untrained versus trained or tested groups. This was true even in regions known to be involved in spatial learning and memory or goal-directed behavior, such as anterior and MD thalamus (Stokes and Best 1988; 1990), striatum (Floresco et al. 1997) and parahippocampal structures (Otto et al. 1997) that may have been recruited during performance of the maze behavior. Thus, in most instances, PFC disinhibition caused aberrant increases in regions that play a key role in facilitating performance of the spatial RM/WM task, but previous experience with the maze task or engaging in search behavior on test day did not alter levels of neuronal activation. The increases in neuronal activation following PFC GABAA antagonism were often substantial, perhaps obscuring some plasticity-related effects. 161  The most notable exception to the observations described above was in the HPC, which does not receive direct input from PFC regions, and only receives disynaptic inputs via the ventral midline thalamus and Ent (Prasad and Chudasama 2013).  In this region, neuronal activation was not altered by PFC GABAA antagonism in animals that did not have any experience with the spatial task. However, PFC disinhibition did increase HPC activation in rats trained on the RM/WM task, regardless of whether they performed the behavior on test day.  A parsimonious explanation for this effect is that spatial learning may have strengthened inputs to the HPC that potentiated the effects of PFC disinhibition, leading to enhanced HPC activation. The dorsal HPC receives PFC input via lateral entorhinal cortex, while ventral HPC receives this input via the midline thalamus, including the Re, Rh and PV nuclei (Cassel et al. 2013; Prasad and Chudasama 2013). We observed equivalent increases in neuronal activation following BIC treatment in both the dorsal and ventral HPC, suggesting that strengthening of either pathway could have contributed to elevated HPC activity. Intriguingly, the Rh nucleus of the thalamus displayed a similar pattern of activation to the HPC, with PFC disinhibition failing to increase c-Fos expression in animals at rest, but increases in c-Fos expression observed in trained and tested BIC-treated animals. Therefore, it seems at least some of the increase in neuronal activation observed in trained animals receiving BIC infusions resulted from an enhancement of input from the corticothalamic-HPC pathway, though the findings do not rule out plasticity within the entorhinal cortical pathway. Alternatively, local plasticity within the HPC that occurred as a result of training may also have led to increased c-Fos expression following maze learning in BIC-treated animals. Studies of learning and memory within the HPC and other regions have often defined neural assemblies that represent a given memory as the set of neurons showing an induction of c-Fos or 162  other IEGs following recall of the memory (Reijmers et al. 2007; Trouche et al. 2016; Zhou et al. 2009). Under basal conditions, it is possible that only a portion of the assembly expresses c-Fos, but increased input, as can occur following PFC disinhibition, may induce expression throughout the assembly via increased number or strength of synapses. It is also possible that c-Fos expression spread to connected cells outside the assembly. This aberrant HPC activation may underlie disruptions in radial-maze performance induced by PFC disinhibition, given the critical role this region plays in mediating search guided by spatial RM and WM (Becker et al. 1980; Olton and Papas 1979).  In this regard, even though the present experiments were unable to identify whether PFC disinhibition caused aberrant HPC activation via increased activity in thalamic, entorhinal and/or local HPC circuitry, either of these options would represent novel forms of plasticity following spatial learning. As such, it will be of considerable interest to investigate these possibilities in the future. 5.4.3 Effects of training and testing on the RM/WM radial maze in SAL-treated animals In most brain regions observed, training and testing on the RM/WM maze task had no effect on c-Fos expression in animals receiving control infusion treatments. This is somewhat surprising, in light of the fact that several of the regions examined, including STR and HPC, have been implicated in performance of identical or similar spatial tasks (Colombo et al. 1989; Floresco et al. 1997; Packard and White 1990; Schacter et al. 1989). Furthermore, inductions of c-Fos or other IEGs have been observed in the thalamus, HPC and prefrontal regions following performance of RM/WM (He et al. 2002b) and WM-only (Vann et al. 2000a; Vann et al. 2000b) variants of the radial maze task. However, past work has shown hippocampal IEG expression often reaches peak levels during learning of similar spatial tasks, and returns to levels near baseline in well-trained animals (He et al. 2002a; He et al. 2002b; Vann et al. 2000b). On the 163  other hand, tasks requiring more complex hippocampal operations, such as pattern separation or completion, continue to induce high levels of neuronal activation in hippocampal subfields in well-trained animals (Yagi et al. 2016). Given the relatively simple nature of the task employed in the present study, the extended training rats received may have refined the ensemble of cells recruited to execute the task efficiently, leading to relatively low demands on the HPC. In this way, c-Fos induction in the HPC of SAL-treated animals may have been subtle and beyond the limit of detection of our methods of quantification. This may also have occurred if performance of the radial maze led to a peak in c-Fos expression before or after the 90 minute time point of the present study. Indeed, trends towards increased IEG expression in some HPC subfields were observed in trained or tested control animals, suggesting that the lack of a significant effect may merely be attributable to insufficient statistical power and/or different time course of c-Fos expression.  However, when taken together with the observation that BIC-treatment increased neuronal activation only in the HPC of trained rats, it is likely that learning of the maze task induced some experience-dependent modification of the HPC circuitry. 5.4.4 Neuronal activation in limbic structures implicated in anxiety and fear The primary focus of this study was to explore how PFC disinhibition affects neuronal activation in brain regions known to facilitate performance of the RM/WM task. However, we also examined neuronal activation in structures implicated in anxiety and conditioned fear, in light of the fact that PFC GABAA antagonism has been shown to produce both an increase anxiety-like behavior (Bi et al. 2013) and impair the ability to discriminate between aversive and neutral stimuli in a fear conditioning paradigm (Piantadosi and Floresco 2014). Surprisingly, the PV nucleus of the thalamus, which has been implicated in maintenance and retrieval of fear memory (Do Monte et al. 2016) showed a selective induction of neuronal activation in groups 164  that were tested on the RM/WM maze task, an effect that was observed in both SAL and BIC-treated animals. The PV has also been implicated in feeding, indicating PV neuronal activation may have reflected consumption of the food reward, which only occurred for tested animals (Matzeu et al. 2014).  Neuronal activation was increased in Amg regions following PFC GABAA antagonism in all behavioral groups. In the Amg, these increases were more prominent in the BLA when compared to the CeA.  Similarly, Jones et al. (2011) showed that a combination of stress and intra-PFC BIC infusion increased neuronal activation to a larger degree in the BLA in comparison to the CeA, although no data from behaviorally-naive animals were included in that study. Taken together with these past findings, the present results suggest that alterations in anxiety or fear-conditioning that result from deficient PFC GABA signalling may result in part from altered activity within the Amg and PV nucleus, in addition to local disinhibition of activity within the PFC. 5.4.5 PFC imbalances in excitation and inhibition impact neuronal activation throughout the brain Maintenance of an appropriate balance between excitatory and inhibitory transmission in different brain regions is critical for their efficient neural functioning and execution of the behaviors they mediate. While transient deviations from this balance are known to occur during development or when learning and plasticity occur (Letzkus et al. 2015), prolonged disturbances in the excitation-inhibition (E-I) balance are associated with pathological states, including schizophrenia and autism. In the present study, antagonism of PFC GABAA receptors was employed to diminish PFC GABAergic transmission, leading to increased pyramidal cell activity (Lodge 2011), and thus transiently disrupting the PFC E-I balance. Disturbances in E-I induced 165  by PFC GABA antagonism might be expected to disturb local oscillations and tuning of individual neurons (Rao et al. 2000) or affect the gating of irrelevant activity or noise within the PFC, leading to deficits in PFC-mediated cognition. However, the discovery that PFC GABAA antagonism also affects non-PFC dependent functions in Chapter 4, such as spatial reference memory, raised the possibility that increased output from PFC projection neurons could be altering activity and function of regions that receive PFC input.  Likewise, though deviations from the E-I balance are often conceptualized as having impacts at the microcircuit level (Murray et al. 2014), they would also be expected to impact long-range synchrony of circuits. The main finding of the present study is that local disinhibition of the PFC induces widespread effects on the levels of neuronal activation within PFC efferent regions. The increase in activation induced by PFC GABAA antagonism in many of these terminal regions is reflected not only by a measure of the strength of the input into the area (Parthoens et al. 2015), but also induction of c-Fos, an IEG known to alter neuronal responses on a prolonged time scale. In this way, increased signal outflow by glutamatergic PFC projection neurons appeared to have profound effects on activity patterns within many structures that receive either monosynaptic or disynaptic PFC innervation, lending support to the idea that a local disinhibition of the PFC can disrupt the function of other circuits. Importantly, the extent of the change in activation was also modulated by plasticity induced by learning in the relevant circuitry. Notably, increases in neuronal activation following PFC disinhibition were observed both in task-related regions, which in this case include HPC, STR and thalamus, and task-unrelated regions, such as the Amg and sensorimotor cortices (Vann et al. 2000a). Therefore, in addition to altering activity in regions needed to learn and perform the maze behavior, PFC disinhibition may also disrupt task performance by inducing aberrant activations in task-unrelated networks. 166  As noted previously, the medial PFC is not required for performance of this variant of the radial maze, raising the possibility that PFC activity that was elevated above normal in the maze context could have also contributed to RM/WM impairments.  Importantly, it is not yet clear how PFC disinhibition would affect E-I balance within individual PFC terminal regions, as the identity of the neurons expressing c-Fos was not assessed in this study. If the majority of activated neurons within a particular region are inhibitory interneurons that target principle projection cells, PFC disinhibition may result in net inhibition of some structures, while other structures where the main targets are glutamatergic may be disinhibited. Finally, while alterations in on-task and off-task circuit activity represent ways that PFC disinhibition may disrupt PFC-independent behaviors, the present findings indicate that reduced inhibitory transmission within the frontal lobes may impair PFC-dependent behaviors via perturbations of co-ordinated activity of PFC-cortical-subcortical networks.  5.4.6 Implications for schizophrenia and other psychiatric disorders One of the pathophysiological mechanisms posited to underlie schizophrenia is a disruption in cortical E-I balance. Deficits in PFC GABA function are amongst the key alterations thought to participate in this altered E-I balance, a notion supported by neuroanatomical (Akbarian et al. 1995b), physiological (Chen et al. 2014; Minzenberg et al. 2010) and functional imaging studies (Frankle et al. 2015) that appear to primarily reflect decreased synthesis and release of GABA in the PFC in schizophrenia. Here, and in past work, we and others have employed antagonism of PFC GABAA receptors to provide insight into how PFC GABAergic transmission, and disinhibitory increases in the E-I balance induced by reducing GABA function, affect behavior and underlying neural activity (Enomoto et al. 2011; Paine et al. 2011; Tse et al. 2015a). Although the change in the E-I balance induced by PFC 167  GABAA antagonism is acute, it is sufficient to reproduce many features of schizophrenia including qualitatively similar cognitive deficits in attention (Paine et al. 2011), working memory, decision-making (Paine et al. 2015b; Piantadosi et al. 2016) and cognitive flexibility, and behaviors reflective of positive (Enomoto et al. 2011; Piantadosi and Floresco 2014) and negative symptoms (Paine et al. 2017; Piantadosi et al. 2016).  The present findings add to this literature to show that the cortical disinhibition induced by a deficient PFC GABA signalling also has the potential to alter the activity and function of circuits throughout the brain in a manner that may be relevant to schizophrenia. A substantial body of recent work has examined circuit-level alterations in schizophrenia (Hunt et al. 2017; Li et al. 2017; Yoon et al. 2013) or models of the disorder (Floresco et al. 2009; Hartung et al. 2016), and the present work suggests that deficiencies in inhibitory transmission within PFC could be an important contributor to many of these effects. While some of this literature has focussed on circuits that mediate cognitive functions known to be impaired in schizophrenia, heightened activity in off-task networks that may disrupt cognitive performance has also been observed (Anticevic et al. 2013), similar to the aberrant activations in regions that do not contribute to RM/WM performance observed here. It is also important to note that decreased PFC GABA function in schizophrenia may take place along with deficiencies in GABA within other structures, including other cortical regions (Hashimoto et al. 2008b), STR and thalamus (Thompson et al. 2009). The growing literature using pharmacological approaches to investigate the functional role PFC GABAA transmission  provide insight in to how decreased GABAergic transmission specifically within the PFC contributes to cognitive and behavioral alterations observed in schizophrenia, and their underlying circuit-level causes. Importantly, these findings suggest that even deficiencies in 168  GABAergic transmission that are restricted to the PFC may interfere with function of other circuits implicated in cognitive and behavioral abnormalities present in schizophrenia. 5.4.7 Conclusions In conclusion, the present study has revealed that deficiencies in PFC GABAergic transmission can have substantial consequences on neuronal activation in both near and distant PFC projection neuron targets. Intriguingly, learning of a spatial RM/WM radial maze task strengthened these effects in HPC and thalamic regions, revealing that increased inputs from thalamic and entorhinal cortical pathways or local remodelling of circuitry may represent novel forms of plasticity associated with learning of these types of spatial tasks. In turn, these forms of experience-dependent plasticity appear to modulate how PFC terminal regions respond to increase in PFC output. Given that dysfunctional PFC GABA signalling is observed in schizophrenia and other psychiatric disorders, these findings indicate that deficiencies in PFC GABA also have the potential to contribute to the circuit-level alterations that have been observed in these disorders. Future studies that investigate how disinhibiting distinct PFC circuits impacts behavior will provide further insight into how diminished PFC GABA function plays a role in these psychiatric conditions.     169  Chapter 6: General Discussion The work described in the past chapters of this Thesis has interrogated how pharmacological reduction of PFC GABA signalling impacts tests of working memory, attention, resistance to proactive interference and speed-of-processing in rats. Further, we also assessed changes in neuronal activation throughout the brain that result from this manipulation, both in animals at rest and following spatial learning. In the following discussion, the implications of these findings will be discussed, revealing the key component processes of cognition affected by deficient PFC GABA function, the consequences of PFC disinhibition on neural activity throughout the brain, and providing insight into how altered PFC GABA signalling may play a role in neuropsychiatric disorders.  6.1 Overview of findings In Chapter 2, PFC GABAergic regulation of delayed-response working memory was assessed. We first confirmed that the operant delayed non-match to position (DNMTP) working memory task was dependent on PFC function by showing that reversible, pharmacological inactivation of the PFC yields delay-independent impairments in DNMTP task performance. We then showed that antagonism of PFC GABAA receptors also leads to delay-independent impairments in the DNMTP task, with accompanying subtle increases in trial omissions and latencies.  When taken in combination with the past finding that PFC GABAA antagonism had no impact on memory retrieval when administered during the delay in a delayed-response radial maze task, the observation of delay-independent impairments following PFC GABAA antagonism indicates that intact PFC GABA function may be of particular importance for encoding of information in working memory. Alternatively, it may play a role in other cognitive 170  processes necessary for performance of the DNMTP working memory task irrespective of delay, such as attention or cognitive control. In contrast, PFC NMDA receptor antagonism produced a more delay-dependent deficit that reflects a selective role for PFC NMDA receptors in working memory maintenance. Although the components of working memory tested with delayed-response tasks appear to be dependent on PFC NMDA receptors, GluN2B-subunit specific NMDA receptor antagonism did not affect any aspect of task performance. Thus, our findings indicate separable, yet complementary roles for PFC GABAergic and NMDA glutamatergic transmission in regulating basic aspects of working memory, including formation of mental representations of stimuli and their maintenance over a brief delay. Given that attentional processes contribute to working memory encoding, we next chose to examine the specific manner in which PFC GABAergic transmission regulates attention in Chapter 3. In so doing, we employed a sustained attention task containing signal and non-signal trials, allowing us to differentiate from impaired signal detection or increased false alarm responding. Here, it was found that PFC GABAA receptor antagonism disrupts sustained attention, predominately by enhancing false alarm responses, rather than by disrupting detection of stimuli. Further, PFC administration of the higher dose of antagonist were associated with an impairment in performance of both visual and auditory conditional discriminations. Thus, stronger perturbations in PFC GABAergic function have prominent impacts on conditional discrimination, an ability that is necessary for the performance of many cognitive tasks. In contrast, PFC NMDA receptor antagonism produced only a subtle deficit in the ability to discriminate between signals and non-signals at very brief signal durations. Our findings suggest that attentional performance is very sensitive to deficiencies in PFC GABA signalling, and that 171  an inability to distinguish between target versus irrelevant signals may be one of the main ways that deficient PFC GABA signalling impacts attention and also working memory encoding. In Chapter 4, we addressed whether PFC GABA signalling also plays a role in resistance to distracting or interfering information during engagement of short- and long-term spatial memory processes. We chose to examine proactive interference, a phenomenon in which past and now irrelevant trial-unique information interferes with current memory performance, as a potential source of distraction or interference. In so doing, we employed a massed-trials variant of the traditional reference/working memory (RM/WM) maze. Although it did not enhance proactive interference effects, PFC GABAA antagonism substantially impaired performance of both reference and working memory aspects of the radial maze task from the first trial. Importantly, PFC inactivation had no effect on the same task, suggesting that PFC disinhibition may interfere with mnemonic and other cognitive processes that are not dependent on the PFC. Furthermore, PFC GABAA antagonism did not affect performance of a simple spatial discrimination task conducted on the maze, ruling out the possibility that the impairments in the RM/WM maze task were due to disruptions in processing of rewards, motoric disruptions, or a basic inability to discriminate between arms of the maze. Taken together, the findings of Chapter 4 suggested that a loss of PFC GABAergic signalling can also impair cognitive function by disrupting activity of other brain regions.  In Chapter 5, we sought to address whether disinhibition of the PFC alters neuronal activation in other brain regions by measuring induction of the immediate early gene, c-Fos, following PFC GABAA antagonism. It was found that PFC GABAA antagonism enhanced neuronal activation in many first-order PFC terminal regions, providing a basis for how PFC GABAA antagonism may disrupt the performance of non-PFC dependent tasks. Neuronal 172  activation was increased in many regions receiving PFC innervation following PFC GABAA antagonism in animals at rest, with the exception of the hippocampus and rhomboid thalamic nucleus. In most cases, training on a single-trial variant of the same RM/WM maze task used in Chapter 4 had no further impacts on neuronal activation. However, increased neuronal activation following PFC GABA antagonism was observed in the hippocampus and rhomboid thalamic nucleus selectively in cohorts that had been trained on the RM/WM task. These results make clear that plasticity within thalamic-hippocampal circuits may modify the extent to which cortical excitation-inhibition imbalances impact activity and function of this circuit.  6.2 Implications for cognition mediated by the frontal lobes 6.2.1 PFC GABAergic regulation of working memory processes The experiments in Chapters 2 and 4 of this Thesis were prompted in part by the observation that PFC GABA antagonism failed to disrupt performance of a spatial delayed-response working memory task conducted on a radial maze (Enomoto et al. 2011). This was a surprising finding given that PFC function is known to be necessary for performance of this variant of the radial maze (Seamans et al. 1995), and that GABAergic signalling in the PFC plays a critical role in generating oscillatory activity which is thought to be a substrate of short-term memory processes necessary for execution of these types of tasks. Further, studies in primates also showed that PFC GABAA receptor antagonism leads to impairments in a spatial delayed-alternation working memory task (Sawaguchi et al. 1988), indicating that PFC GABA function is necessary for working memory in certain conditions. As such, a goal of Chapters 2 and 4 was to identify the circumstances in which PFC GABA signalling does play a critical role in working memory. With respect to the delayed-response radial maze task employed by Enomoto et al. 173  (2011) previously, PFC GABAA receptors were subject to antagonism only before retrieval of the memory. Though it was surprising that PFC GABAA antagonism did not affect retrieval given that PFC activity is known to be particularly critical for this aspect of working memory (Seamans et al. 1995), these findings raised the possibility that intact PFC GABA function may be most critical during either the encoding or maintenance of the memory. Further, the delayed-response radial maze consists of only a single trial in which responding is self-paced.  Many working memory tasks administered to human and animal subjects consist of multiple trials given in short succession in which responding must occur in a restricted time frame (Park and Holzman 1992), increasing the cognitive load associated with the task. Therefore, the Experiments in Chapter 2 and 4 were designed with a goal of observing the effects of deficient PFC GABAergic transmission throughout the encoding, maintenance and retrieval phases of working memory, and in conditions of increased cognitive load generated by multiple back-to-back trials with short ITIs.  When taken with the findings of Enomoto’s delayed-response radial maze study (2011), the findings of this Thesis indicate that PFC GABA function is particularly essential during the encoding period of working memory. An equivalent impairment in working memory function following PFC GABAA antagonism was observed irrespective of delay in Chapter 2, meaning that the cognitive function(s) affected contribute(s) to working memory at all delay lengths, which could include WM encoding or retrieval, or non-working memory functions necessary for performance, such as cognitive control or attentional processes. Given that it was shown that spatial working memory retrieval was not impacted by reducing PFC GABAergic transmission (Enomoto et al. 2011), at least some of the working memory impairment resulting from PFC GABA hypofunction is likely to result from disruptions during the encoding period. This 174  assertion is supported by the findings in Chapter 3, which show that attentional performance, a factor known to affect encoding (Gazzaley and Nobre 2012), is extremely sensitive to disruptions in PFC GABA function. Notably, a selective increase in false alarm responses in non-signal trials, but no deficits in signal detection were observed in the SAT employed in Chapter 3.  Given that this task imposed a 2s delay, meaning that trial-unique information was encoded and briefly maintained in both trial types, the observation that only false alarm responses were affected could suggest that PFC GABA function is more important in suppression of competing activity or filtering of noise rather than directly contributing to encoding. On the other hand, if diminished PFC GABA function leads to an overall bias to respond as if signals are present when they are not, this may have obscured more subtle deficits in signal detection. In this way, intact PFC GABA signalling appears to play a role in the selection of stimuli to be encoded, either by the suppression of activity from competing inputs or by a more subtle contribution to formation of representations of stimuli. Nevertheless, these processes would both be vital during encoding, highlighting this as a crucial period for intact PFC GABA function. The idea that intact PFC function, and in particular PFC GABA function, is essential to encoding is also supported by findings in animal models and in humans. Our findings are in agreement with recent optogenetic studies, which have shown that activity of somatostatin (SST) GABAergic interneurons regulates encoding in a delayed-response T-maze task (Abbas et al. 2017). Further, human EEG work with delayed-response working memory showed that oscillatory activity during encoding strongly predicted working memory performance (Chen et al. 2014; Haenschel et al. 2009; Kang et al. 2018). Since interneurons are known to play a key role in organization of oscillatory activity, decreased PFC GABA function during encoding would result in a disruption of activity in this critical period, which subsequently affects 175  behavioral responses. Indeed, PFC GABA levels were related to the strength of gamma oscillations in encoding, and both were predictive of subsequent working memory performance, further suggesting that PFC GABA function is related to encoding success (Chen et al. 2014).  With respect to whether PFC GABAergic plays a role in working memory specifically in situations that require handling of increased cognitive load, the results are less clear. Multiple back-to-back trials in short succession were a feature of the behavioral tasks used in Chapter 2 through 4, and would have resulted in increased cognitive load relative to the delayed-response radial maze. In the operant DNMTP task, we observed impaired working memory concomitantly with slightly increased latencies and trial omissions. When viewed with increased response latencies observed following PFC GABA antagonism in other studies (Enomoto et al. 2011; Paine et al. 2011), it seemed possible that slower processing times resulting from deficient PFC GABA function could potentially affect performance when responding was required in a short time frame, and that increased cognitive load could have exacerbated this issue. Then, remarkably, PFC GABA antagonism did not produce any increases in latencies or trial omissions in the SAT administered in Chapter 3, which contains even more trials presented with a short, variable length ITI, requiring animals to remain oriented toward the front panel of the operant chamber. Thus, in this condition of very large cognitive load, animals with reduced PFC GABA function remained as able to respond in a timely fashion as controls, ruling out the possibility that cognitive load was responsible for increased latencies to respond. Further, since slow responding did not occur during the actual choice period in either the DNMTP or SAT task, it is unlikely to be a factor that directly affects choice behavior. Next, we examined whether deficiencies in PFC GABAergic transmission augment proactive interference, or when past-trial unique information disrupts current memory processes. PFC GABAA antagonism resulted in an 176  impairment from the first trial, indicating that intact PFC GABA function is not particularly necessary in mitigating interference generated by the presence of multiple back-to-back trials. At present, the collective findings do not point to a selective role for PFC GABA function in adjusting to larger cognitive load. Still, it is possible that PFC GABA function is preferentially recruited in other conditions associated with increased cognitive load; for instance, when multiple items must be remembered in a single trial. Though our findings support a clear role for PFC GABA signalling in working memory encoding and its attentional regulation, they do not rule out the possibility that PFC GABAergic transmission in some way contributes to working memory maintenance. Weakened encoding of information would lead to delay-independent impairments as was observed; however, it is possible that additional impairments in maintenance of memories may have been present. Further, we only measured working memory maintenance over a maximum of 24 s, as performance at this longest delay was near chance. However, in other circumstances, such as T- and radial mazes, rats have been shown to accurately maintain information for several minutes, so it is possible that interneurons are recruited over the course of more extended working memory maintenance. Kim et al. showed that continuous optogenetic activation of SST interneurons impaired T-maze working memory performance across a 10 s delay, but not over a shorter period (Kim et al. 2016a). Given that the continued activation would be likely to alter SST activity from normal patterns, this finding could be interpreted to suggest that intact SST interneuron activity is particular important for maintenance of working memories as the delay lengthens. While this finding contrasts more recent work showing that SST interneuron inactivation predominately affects encoding (Abbas et al. 2017), these discrepancies may reflect task-phase specific roles for SST interneurons in working memory that are supported by 177  divergent patterns of activity. During encoding, high levels of SST interneuron activity may serve as a gating function in which only the most strongly activated pyramidal cells would become part of the working memory signal. Later in the delay, activity of selective SST interneurons may result in lateral inhibition of pyramidal cells not involved in maintenance of the working memory, thereby decreasing working memory interference. As such, both the present work and findings from recent optogenetic studies prompt further examination of whether PFC GABA signalling plays a role in working memory maintenance, particularly in resistance to interference as the delay progresses.    In contrast to optogenetic work manipulating SST interneuron activity, enhancement or suppression of PV interneuron activity has had no effect on any phase of a working memory in two studies using the T-maze (Abbas et al. 2017; Kim et al. 2016a). On the other hand, intact activity of PV interneurons is necessary for attentional performance in serial reaction time tasks (Kim et al. 2016b). In line with the fact that PFC GABA was manipulated in a more global fashion in the present studies, we saw impairments that were indicative of both impaired encoding and attentional deficits. Further, it may be that encoding is strengthened in maze tasks relative to operant tasks, as the animal has to traverse to the sample arm which takes more time and physical effort relative to a lever press. Thus, in the operant task, PV interneurons may still play a crucial role by supporting attentional processes that enable the relatively fast schedule of responding required. Such discrepancies also highlight the impacts that changes in task parameters may have upon recruitment of different aspects of the PFC GABA system. Nevertheless, the present results are largely consistent with the alterations reported in recent studies employing interneuron type-specific manipulations, and in some respects can be viewed as the sum of decreased PV and SST interneuron function. 178  6.2.2 PFC GABAergic regulation of attention & response inhibition The experiments in Chapter 3 highlight a key role for PFC GABA function in attentional functions mediated by the frontal lobes. Intra-PFC administration of even a very low dose of GABAA antagonist caused significant attentional impairments and increased false alarm responding in the operant SAT. Further, these results suggest that the ability to inhibit inappropriate responding is one of the key aspects of cognition that is compromised by deficient PFC GABA signalling. When PFC GABAergic transmission was more substantially blocked with the higher dose of antagonist, this manifested as an inability to follow conditional rules even when attentional demands were very low. However, the deficit in sustained attention observed in our studies cannot be explained exclusively by an inability to perform conditional discriminations, since administration of a lower dose of antagonist which disrupted SAT performance did not impact performance of auditory or visual conditional discriminations. It is also important to note that PFC GABA antagonism at an intermediate dose (25 ng) disrupts the 5-CSRTT of attention (Paine et al. 2011), which is not dependent on performing conditional discriminations, highlighting that impaired attention is at least one component of the core features affected by deficient PFC GABA transmission. Furthermore, our findings are in agreement with recent optogenetic work showing that suppression of PV interneuron activity disrupts performance of a 3-CSRTT (Kim et al. 2016b) and a significant body of pharmacological work with both competitive and non-competitive antagonists, showing that reduction of PFC GABAergic transmission leads to deficits in attention in SRTs (Paine et al. 2011; Pehrson et al. 2013; Pezze et al. 2014). The core deficit observed following reduced PFC GABAergic transmission in the SAT and conditional discrimination may be likened to an impairment in response inhibition. In both 179  tasks, an increase in aberrant responses in inappropriate task epochs was observed following PFC GABAA antagonism. However, decreased response inhibition does not appear to be the result of an increase in motor impulsivity, since no changes in premature responding have been observed in the 5-CSRTT following PFC GABAA antagonism (Paine et al., 2011; Pezze et al., 2014). In fact, reduced PFC GABAergic transmission was associated with slowing of responding in many circumstances. Though it appears that PFC disinhibition leads to an increase in aberrant responding distinct from increased motor impulsivity, the conditions that give rise to these aberrant responses are not entirely clear. One possibility is that intact interneuron function is particularly important in mediating interference resistance, which is also supported by the observation that PFC GABA hypofunction led to slowing of responding selectively in instances when animals had the opportunity to engage in off-task activity, such as after a delay, where sources of environmental interference would be higher.   The ability to perform conditional discriminations, which was perturbed following more severe disruptions of PFC GABA function, may be viewed as a core element of cognition in that many cognitive tasks, or indeed tasks in everyday life, are dependent upon this ability. Both humans and animals are regularly presented with situations that requires them to make one response in case of one event, and a different one in case of another. Furthermore, since performance of conditional discriminations is an ability that is required in many tasks used to assess cognition, past studies that have found impairments in behavioral tasks following PFC GABA antagonism should be reassessed in light of whether they require this ability. For instance, in addition to tracking changing task contingencies, optimal performance on probabilistic or effort based cost-benefit decision-making tasks may be dependent on coding these contingencies as a conditional discrimination since it is most advantageous to choose the 180  high reward option when reward costs are low, but the low reward option when costs are high. For instance, in the probabilistic discounting task used by our group, one lever in an operant chamber delivers a smaller, one pellet reward 100% of the time, while the other lever will deliver a larger, four pellet reward some of the time. Typically, the probability of reward delivery on the larger reward is increased or decreased throughout the test session in blocks of trials. When the probability of reward delivery is greater than 25%, it is most advantageous to choose the uncertain lever with larger reward, as this leads to accumulation of more rewards. When the probability of reward delivery on the large reward lever is decreased below 25%, more rewards would be obtained by selecting the safer lever. Thus, optimal performance is contingent upon being able to make a conditional discrimination where the large reward lever is selected when the probability of reward delivery on this lever is high, but the smaller lever is selected when probability of reward delivery on the larger lever decreases, i.e. when the cost of selection of the larger lever begins to outweigh the benefits. Animals receiving PFC administrations of GABAA antagonists also have difficulty in discriminating between larger vs. smaller, but not rewarded vs. non-rewarded options. Thus, more profound deficiencies in PFC GABA may lead to an overall inability to act appropriately when distinct actions are required in order to respond to different task requirements.   6.2.3 PFC GABAergic transmission & interference resistance Both working memory and attentional performance depend on the ability to both filter internal sources of noise and environmental distractors, also known as interference resistance. In particular, GABAergic cells in the PFC may facilitate this process by lateral inhibition of competing activity and also through temporal organization of activity that plays a role in gating of inputs; for instance, during oscillations. Our results indicated that this organization of PFC 181  neural activity by GABA signalling is particularly important during working memory encoding. Here, a deficit in PFC GABAergic transmission results in a weakened representation of target information that subsequently leads to sub-optimal performance when choices dependent on that information are required. However, resistance to interference is also vital over the course of working memory maintenance, and even across a test session with many trials. Therefore, a major question is whether and how PFC GABA signalling plays a role in mitigating interference at these different time points. In tasks that engage memory and contain multiple trials with unique information, execution of the correct response on the current trial may be disrupted by interference from past and now irrelevant information, a phenomenon known as proactive interference (Roberts & Dale, 1981). The experiments in Chapter 4 took a first step towards understanding if PFC GABA function is involved in mediating interference resistance at other stages, specifically resistance that results from accumulation of information over time.  We hypothesized that, since PFC GABAergic cells appear to contribute to the integrity of memory through silencing of irrelevant input, one of the sources of this input may be past memories that interfere with current memory processes. However, our results indicated that deficits in PFC GABAergic signalling result in impaired working and reference memory performance that is present from the first trial and is not augmented with successive trials. Thus, subjects did not appear to have increased difficulty differentiating between choices that they had made on the present versus previous trials, and therefore did not appear to be any more sensitive to the effects of proactive interference in comparison to controls. It is possible that disinhibition of the PFC using GABAA antagonists induces an initial, transient disorientation that may have obscured a proactive interference effect. Indeed, some animals did show a subtle impairment in spatial discrimination on the T-maze that 182  was present only in the first few trials following the infusion. It may also be that PFC GABAergic signalling is more important for resisting retroactive interference, in which new information interferes with the maintenance of an earlier working memory. One further issue is that the memory processes investigated in Chapter 4 were shown not to be dependent on PFC. In contrast, a recent study showed that the PFC is actively involved in reducing proactive interference in a spatial reversal learning task that is dependent upon both PFC and hippocampus (Guise and Shapiro 2017). Specifically, intact PFC activity was necessary for separation of competing representations in hippocampus that protected against the effects of proactive interference. Thus, it is possible that disinhibited PFC activity induced by reduced PFC GABAergic transmission could lead to increased interference in this task, an interesting possibility to investigate. 6.2.4 PFC GABAergic transmission & processing speed Paradoxically, impaired PFC GABA signalling has been observed to cause hyperlocomotion and increased responsivity to stimulants, yet slowing of responding in both maze and operant environments. In operant-based tasks, slowing of responding occurred after long delays; when animals were engaged at the front panel of the operant chamber, such as during the choice phase of the DNMTP task or during the SAT, increased latency to respond was not observed. Because increased in latencies were not observed when animals were engaged and making choices, our results suggest that deficits in PFC GABAergic transmission more selectively influence orientation towards relevant stimuli, rather than speed-of-processing itself. The failure to respond to stimuli in a timely fashion induced by deficient PFC GABA function also led to slight increases in trial omissions in circumstances in which time limits on responding were placed. On the radial maze, responding was self-paced and PFC GABA antagonism led to 183  increased latencies in reference/working memory, foraging and simple discrimination tasks. The maze is a larger and more complex environment, raising the possibility that the presence of more environmental distractors diverted attention from the task at hand. Interestingly, an MRS study of young and aged schizophrenia patients and controls found that PFC GABA levels were correlated specifically with working memory performance and not speed-of-processing (Rowland et al. 2016a), mirroring these results. In sum, our findings concerning increased latencies following decreased PFC GABA function are reflective of impairments in orientation to stimuli or the ability to ignore environmental distractors, not speed-of-processing deficits. 6.2.5 Contrasting of deficient PFC GABAergic transmission & other PFC alterations The findings in this Thesis reveal that intact PFC GABA function is necessary for many cognitive processes dependent on the frontal lobes. However, the effects of deficient PFC GABAergic transmission are distinct from the effects of lesions or inactivation of the PFC and also decreased PFC NMDA receptor function, and suggest divergent consequences for losses of PFC function in comparison to disruptions in PFC excitation-inhibition balance. In many cases, in addition to disruptions in functions known to be mediated by PFC, further deleterious changes in task performance were associated with PFC GABA antagonism in comparison to PFC inactivation. For instance, in the DNMTP task, while PFC inactivation resulted in a similar delay-independent deficit, PFC GABA antagonism also increased omissions and sample phase latencies. In the reference/working memory radial maze task, deficits in performance were only observed following decreased PFC GABA function, while PFC inactivation had no effects. These results were in keeping with results from studies of pre- and post-training lesions of the medial PFC on the same radial maze task (Becker et al. 1980; Kesner et al. 1987), and also studies showing that inactivation of medial PFC have no effects on the random foraging aspect 184  of the task (Seamans et al. 1995). We also observed that PFC administration of a higher dose of GABAA antagonist led to impairment in the performance of conditional discriminations.  This also does not depend on the PFC (Shaw et al. 2013; Van Holstein and Floresco 2017), again suggesting that disinhibition of the PFC leads to consequences that are not observed following PFC inactivation. On the other hand, while past work showed that both decreased PFC GABA function and PFC inactivation disrupt attentional performance, only PFC inactivation leads to increased impulsive responding (Paine et al. 2011; Pezze et al. 2014). The lack of motor impulsivity and additional impairments observed after PFC disinhibition in comparison to loss of PFC function may both be explained by the widespread increases in neuronal activation in prefrontal regions and throughout the brain observed in our c-Fos studies. Intact PFC function normally is thought to provide top-down control or a brake on motor activity, particularly in striatal regions. Pharmacological disconnection of PFC projections to NAc removes this inhibition and can lead to impulsive responding (Feja and Koch 2015). An overactivity of this projection induced by PFC disinhibition could lead to increased latencies to respond by interrupting ongoing motor programs, as was observed. In terms of aberrant responding, alterations of activity in both on-task and off-task regions may have driven inappropriate responses, an idea that will be discussed in further detail in the next section of this chapter. Our experiments in Chapters 2 and 3 provide insight into complementary, yet separable roles for PFC GABA and NMDA glutamate function in regulation of cognitive processes mediated by the frontal lobes. We chose not to examine PFC NMDA receptor function in Chapters 4 and 5, as we found that even systemic administration of NMDA antagonists did not impact performance of the RM/WM radial maze task (unpublished observations). PFC NMDA receptor activity appears to be particularly important for maintenance of information during 185  working memory performance, as PFC NMDA receptor antagonism did not impair working memory accuracy at the shortest delay. In line with this, PFC NMDA receptors may not be needed for attentional regulation of encoding in most instances, as PFC NMDA receptor antagonism only impaired the ability to discriminate between signal and non-signal trials at very brief stimulus durations. Therefore, it is possible that PFC NMDA receptors are involved in attentional processes relevant to encoding particularly when stimuli are very brief. Past work indicated that NMDA receptor antagonism in both dorsal and ventral PFC subregions disrupts performance of serial reaction time tasks, while targeting of more ventral regions additionally caused an increase in motor impulsivity (Murphy et al. 2005). Taken together, these findings suggest that NMDA receptors within the PFC are preferentially recruited when demands on attention are high, either because of weak stimulus intensity, the need to attend multiple stimuli locations, or when inhibition of impulsive responding is required. This pattern of results gives the impression that PFC NMDA receptors are particularly important in the generation of cortical signals and their integrity in instances in which high levels of cognitive control are required. NMDA receptors are known to be localized on both PFC pyramidal cells and interneurons. Systemic administration of NMDA antagonists leads to cortical hyperactivity that is thought to emerge from a decrease in interneuron activity (Homayoun and Moghaddam 2007). Furthermore, NMDA receptor signalling on interneurons has been shown to be necessary for generation of gamma rhythms associated with working memory performance in mice (Carlén et al. 2012). Mice with conditional deletion of the GluN1 subunit of the NMDA receptor in PV neurons had WM accuracy of 65% of delay of 1-40 s, whereas controls in this experiment showed markedly better performance at a 1 s delay with an accuracy of 85%, suggesting a delay-independent deficit (Carlén et al. 2012). While these results appear to contradict with our 186  findings, genetic deletion of the GluN1 subunit would have decreased NMDA receptor expression throughout the lifespan, amounting to a long-term decrease in excitatory input that consequently led to interneuron hypofunction overall. In this way, the results would support our finding that intact interneuron function is necessary for encoding, because mice with GluN1 deletion had impaired performance even at the shortest delay, similar to a direct block of PFC GABAergic transmission. On the other hand, NMDA receptors on pyramidal cells of PFC are known to be involved in delay period activity during working memory, so the delay-dependent effects on working memory observed in the present work track well with the idea that pyramidal cell NMDA receptor activity contributes to working memory maintenance (Wang et al. 2013). Our results suggest that the timing and duration of PFC NMDA receptor hypofunction may affect whether only maintenance or both maintenance and encoding of working memory are affected. 6.2.6 Consequences of PFC disinhibition in cortical and subcortical circuitry As the primary cortical inhibitory transmitter, release of GABA serves several functions in regulating neural activity within PFC. The manipulation employed in these studies, a competitive block of PFC GABAA receptors, would be expected to dampen inhibitory input at all GABAergic synaptic sites in proximity to the infusion, in turn leading to a non-selective increase in pyramidal cell activity. Infusion of 100 ng of bicuculline in PFC regions led to profound neural spiking, and we expect that administration of 12.5 or 50 ng would produce disinhibition, which was supported by an increase in PFC neural activity measured by c-Fos (Paine et al. 2011) and PET imaging (Parthoens et al. 2015).  The increase in neural activation in our c-Fos study was present throughout the medial wall of the PFC. This aberrant increase in activity could account for a loss in the precision of PFC representations necessary for the execution of 187  cognitive tasks dependent on PFC. Indeed, iontophoretic studies using bicuculline in primate dorsolateral PFC showed that blocking of GABA leads to a loss of tuning of PFC neurons to their preferred locations in the delay period of an oculomotor delayed-response task that was associated with impaired working memory (Rao et al. 2000). While it is controversial whether rodents exhibit the same PFC neuron delay-period tuning observed in primates (Bolkan et al. 2017), these findings indicate that disruption of appropriate patterns of activity within the PFC by reducing GABA function leads to a loss of working memory accuracy. Our results then further add to these studies, as they indicate that disinhibition of the PFC via GABAA antagonism also alters neuronal activation throughout cortical and subcortical PrL PFC projection targets. These findings are in agreement with optogenetic studies that have employed stimulation of PFC pyramidal cells, and observed increased IEG expression in downstream structures, including the MD (Benn et al. 2016) and NAc core and shell (Kumar et al. 2013; Lobo et al. 2013). Given that both the top-down projections from the PFC and the regions themselves are known to influence goal-directed behavior at all stages, from discriminating stimuli to planning and action selection through to processing of action outcomes, the alteration in activity could disrupt behavioral performance at many levels. At the simple level of stimulus discrimination, distinct populations of PFC neurons have been shown to be activated by appetitive versus aversive stimuli, with exposure to cocaine activating more NAc-projecting cells and shock leading to activation of more neurons that project to the lateral habenula (Ye et al. 2016). Given that these results suggest that different stimuli are encoded by different ensembles in PFC, an uncontrolled and widespread activation of projection neurons induced by deficiencies in PFC GABA could potentially lead to inappropriate responding by activation of projections associated with processing stimuli with other valences or contextual associations. In 188  fact, past work using a fear-conditioning paradigm showed that PFC GABAA antagonism led to increased fear of a neutral CS-, but decreased fear to the aversive CS+, suggesting that PFC disinhibition does lead to confounded responses to stimuli of different valences (Piantadosi and Floresco 2014). One region in which changes in neuronal activation driven by PFC disinhibition would be expected to lead to profound consequences for cognition and goal-directed behavior would be the NAc, which is involved in integrating input from cortical, limbic, thalamic areas, and ventral tegmental area dopamine neurons, and then serving as an interface between these regions and motor areas. Studies examining how PFC activity impacts gating of inputs in NAc showed that high frequency stimulation of the PFC leads to a suppression of HPC and thalamic-evoked responses in the NAc. Recent preliminary work from our group has found that PFC GABAA antagonism leads to reduced HPC and thalamic, but increased BLA-evoked responses in the NAc (Tse et al. 2015a). These results may suggest heightened emotional responses in line with the increased anxiety or fear that have been observed after PFC disinhibition, but reduced transmission of spatial or other information that is used to guide cognition. Our group has also previously shown that PFC GABAA antagonism increased VTA dopamine neuron firing, leading to enhanced phasic dopamine release in NAc that is associated with increased locomotor activity, further suggesting that PFC disinhibition can lead to profound alterations in function within NAc that lead to altered behavior. Though disruption of PFC GABA signalling by infusion of bicuculline would be expected to lead to a disinhibitory increase in the overall activity of PFC neurons, disrupting cognitive functions mediated by PFC, such as working memory or attention, our findings also suggest that altered activity in regions receiving PFC innervation after PFC disinhibition may 189  contribute to disruption in these functions. For instance, ventral midline thalamus (Hallock et al. 2016), the mediodorsal thalamus (Parnaudeau et al. 2013), and striatal regions (Akhlaghpour et al. 2016) are known to have roles in delayed-response working memory, with recent work supporting the idea that these regions interact with PFC during working memory performance (Bolkan et al. 2017; Hallock et al. 2016). Both increased excitatory outflow from PFC and disinhibition of activity within PFC could disrupt such interactions. The idea that variations in PFC GABA levels may influence interactions between regions is supported by a recent study in humans indicating that GABA in PFC measured by MRS predicted PFC-amygdala functional connectivity (Delli Pizzi et al. 2017). While in this study, they hypothesized that an excess of PFC GABA function could lead to changes in emotional behavior by inhibiting activity of top-down projections to amygdala, it is also possible that an excess of increased activity of this projection from deficient PFC inhibition that in turn leads to inappropriate amygdalar activation would be detrimental and disrupt amygdala-related behaviors, such as fear, anxiety and emotional memory. Indeed, disinhibition of the PFC using bicuculline at the same dose employed here led to an increase in anxiety-related behavior in several assays in rats (Bi et al. 2013).  One particularly intriguing aspect of the work described is that extended training on the reference/working memory radial maze task led to changes in neuronal activation induced by PFC disinhibition. PFC disinhibition led to an increase in neuronal activation in the HPC only in animals that had been trained on the radial maze task. The HPC does not receive direct innervation from the PFC, but there are neurons within the ventral midline thalamus that receive input from PFC and project to HPC, providing a disynaptic pathway from PFC (Prasad and Chudasama 2013). Remarkably, one of these nuclei, the Rh, showed the same pattern of results, 190  with no changes in neuronal activation induced by PFC disinhibition in untrained animals, but a significant increase in activation after training.  Thus, our findings emphasize plasticity in this thalamic-hippocampal pathway over the course of spatial learning.   6.3 Experimental considerations, limitations & future directions The present studies have established roles for PFC GABA function in modulating key cognitive processes mediated by the frontal lobes, providing clarification of the role of this transmitter system to working memory and attention. Nevertheless, a number of limitations exist, particularly pertaining to the use of bicuculline as an agent to perturb PFC GABA function. In sufficient doses, perturbation of PFC GABA signalling with bicuculline could lead to seizing. The doses of bicuculline employed in this and past work (12.5-50 ng in 0.5 uL per hemisphere) are equivalent to 54-216 mM. In slice electrophysiology preparations, 10 uM baths of bicuculline cause profound disinhibition and epileptiform-like discharges (Albowitz and Kuhnt 1993; Chervin et al. 1988). These methods of applying bicuculline differ in that intra-cortical infusions involve administration of a fixed amount that may spread away from the infusion site and be metabolized, while a bath application of bicuculline would remain at constant concentration for prolonged periods, and also induce disinhibition of any intact inputs. However, given that infusion concentrations of bicuculline greatly exceed that necessary to induce epileptiform-like discharges in slice preparations, it is likely that some degree of epileptiform activity would be induced in the infusion area directly after application. A major limitation of the present study is that no recordings were conducted in the infusion area following application of bicuculline at these doses, so it is uncertain whether and for how long such activity was induced. In vivo recordings in PFC following infusion of 100 ng of bicuculline have been conducted in 191  anaesthetized animals and showed enhanced spike wave discharge activity for 30 minutes after the infusion (Lodge 2011). Thus, we expect that the amount and/or duration of such activity would be reduced in the present experiments, particularly following administration of the 12.5 ng dose.   The doses we selected in the present studies (≤50 ng) were chosen as they rarely lead to behavioral signs of seizing when administered in rats (Bi et al. 2013; Enomoto et al. 2011; Jones et al. 2011; Paine et al. 2011; Paine et al. 2017; Parthoens et al. 2015; Piantadosi and Floresco 2014; Piantadosi et al. 2016) following PFC administration, whereas infusions of higher doses of bicuculline (100 ng) were shown to induce seizing in a three of eight animals in pilots by our group. Furthermore, behavioral signs of seizing were never observed following 12.5 ng of bicuculline. Animals displaying behavioral signs of seizing at the 50 ng dose (< 1 in 50) were removed from the analyses. While none of the animals included our analyses appeared to seize, it is possible that epileptiform activity leading to seizing without prominent behavioral signs was present in some or all animals. However, our impression is that occurrence of this type of seizing was not widespread. Disruption of ongoing cognition by seizures would likely result in failures to respond that would manifest as omissions, at least some of the time. In particular, the SAT required responding within the relatively short time frame of 4 s after lever extension, yet omissions were not significantly increased in that task, suggesting that performance of most trials was unhindered by the presence of seizure-like activity. Nevertheless, it is important to question to what degree the cognitive impairments described result from deficits in PFC GABA function alone, versus more profound disruptions in task performance that would be induced by epileptiform-like events. Notably, the behavioral impairments we observed following PFC GABAA antagonism were relatively subtle, in the range of an approximate 10% decrease in 192  overall accuracy in the operant tasks, with comparable levels of omissions relative to controls. Furthermore, some tasks, such as spatial discriminations, are unaffected by PFC GABAA antagonism, further suggesting seizing is not widespread.  Lastly, altered cognition was still observed in the SAT and RAM tasks following administration of the lower dose of GABAA antagonist, supporting the notion that even very subtle perturbations in PFC GABA function that were well below the threshold of seizing resulted in cognitive impairment. Importantly, disorders associated with altered GABA function, including schizophrenia and autism, have increased risk of seizures and comorbidity for epilepsy (Cascella et al. 2009; Jacob 2016), suggesting that impairments in PFC GABAergic transmission sufficient to lead to psychiatric symptoms can cause seizing in a subset of individuals. Likewise, increased spike wave discharges are observed in animal models of schizophrenia and autism (Missig et al. 2018), indicating that induction of disinhibition via other methods also produces similar epileptiform-like activity in some instances. Thus, it is possible that even the use of more modern tools to perturb PFC GABA function, such as chemogenetics or optogenetics, would have similar limitations. Throughout the experiments described here and in past work from our group, PFC infusions of bicuculline were targeted at the prelimbic medial PFC; however, infusions with areas that extended into anterior cingulate or infralimbic cortex were not excluded, as the primary goal of these experiments was not to ascertain differences between medial PFC subregions. However, the results of Chapter 5 support this approach, as they show the PFC GABAA antagonism leads to profound activations throughout the medial wall of the PFC, suggesting that the manipulation was unlikely to be restricted to a single PFC subregion in any case. Nevertheless, it may be of interest to determine how more targeted disinhibition impacts 193  upon cognition in behavior or patterns of neural activation; for instance if there are differences in optogenetic stimulation of pyramidal cells in medial PFC subregions. Though manipulation employed may be viewed as causing an overall decrease in the GABAergic inhibition of post-synaptic cells, only GABAA, and not GABAB, receptors were targeted. GABAergic inhibition via GABAB receptors is complex, in some cases causing decreased GABA release or inhibiting release of other neurotransmitters (Bowery et al. 1980; Reimann 1983). In this regard, the selective targeting of the GABAA receptor was chosen in order to provide a simple view of the consequences of reducing the majority of fast GABAergic transmission in PFC. Nevertheless, in intact systems, an overall decrease in GABAergic tone would reduce the GABA concentrations available to bind to both receptors. Future studies that compare GABAB receptor regulation of cognition, and perhaps the additive effects of GABAA and GABAB antagonism, would be of great interest given the complex actions of the GABAB receptor and the relative lack of knowledge of how it regulates cognition. It is also important to note that some GABAA receptors are insensitive to the effects of bicuculline, and is not clear how the lack of actions at these receptors would have impacted the present results (Johnston 2013). Our c-Fos studies identified changes in neuronal activity induced by PFC GABA hypofunction in cortical and subcortical targets, and also identified changes in neuronal activation induced by plasticity associated with spatial learning. However, such changes cannot be interpreted in regards to altered excitation-inhibition balance in downstream regions as increased activity of descending projections from PFC may have resulted in activation of either inhibitory or excitatory neurons, thus leading to either inhibition or disinhibition of PFC efferent regions. Electrophysiological studies that directly examine how activity is altered within 194  individual regions following PFC disinhibition will be of great interest. Further, since there were widespread increases in activity, we were unable to identify whether increases in activation were the direct result of PFC projection neuron activity or other disynaptic effects. In this regard, experiments that combine measurement of IEG expression with anterograde tracers from PFC and/or cellular markers that identify the cell-types affected in terminal regions may be particularly informative. In light of this, the present work should be viewed as being a foundation that informs further interrogation of the circuits affected by PFC disinhibition. For instance, studies examining projection-specific disinhibition of PFC-NAc core versus PFC-shell would be of interest to determine how PFC GABA hypofunction may impact goal-directed behavior. Likewise, interrogation of the PFC-ventral midline thalamus-hippocampal circuit would be intriguing in light of the plasticity-related effects observed. Given that the increases in activity were widespread, an important question is the extent to which impairments observed were caused by disruptions in task-related regions, in comparison to induction of aberrant activity in task-unrelated areas, further emphasizing the need for studies that examine disinhibition of specific descending PFC projections.  Finally, we did not examine the impacts of PFC administration of the lower dose of GABAA antagonist on neuronal activation. In light of the observation that the behavioral impairments observed after this dose were more subtle than the higher dose, it would be of interest to determine if the lower dose caused increased neuronal activation in PFC efferent regions. A lack of increased activity in other regions following more subtle PFC GABA hypofunction would highlight loss of specificity of representations in the PFC as a key component of many of the cognitive impairments observed following decreased PFC GABAergic transmission. 195   6.4 Relevance to schizophrenia and other neuropsychiatric disorders 6.4.1 Schizophrenia Schizophrenia is a complex psychiatric disorder with unique emotional and behavioral changes that are accompanied by a stable and lasting impairment in cognition that is not treatable by currently available pharmaceuticals. As a first step towards identifying better treatments for cognition in schizophrenia, the Measurement and Treatment Research to Improve Cognition in Schizophrenia (MATRICS) initiative of the National Institute of Mental Health conducted a review of factor analysis studies of cognition in schizophrenia and found that seven main domains of cognition are impacted: working memory, attention, speed-of-processing, verbal learning and memory, visual learning and memory, reasoning and problem-solving, and social cognition (Nuechterlein et al. 2004). All of these, with the exception of verbal learning and memory, may be assessed in animals. Furthermore, a suggested battery of behavioral tests assessing each of these domains in animals has been generated in order to facilitate translational pharmacological work (Young et al. 2009).  A substantial body of past work has indicated that deficiencies in PFC GABAergic transmission may underlie cognitive dysfunction observed in schizophrenia, as well as the positive and negative symptoms. With the translational behavioral tests proposed by MATRICS as a framework, our group and others have attempted to assess which aspects of schizophrenia-related cognition are mediated by these deficits in PFC GABA function. The work presented in this Thesis provides further clarification of how PFC GABA hypofunction can contribute to schizophrenia-related cognitive impairments, specifically in the domains of working memory, 196  attention and processing speed. Our results also speak to underlying impairments in conditional discriminations and/or response inhibition. Meta-analyses of delayed-response working memory studies of schizophrenia patients have revealed a consistent delay-independent impairment in working memory performance with a concomitant slowing of reaction times (Lee and Park 2005). These impairments are present regardless of the sensory modality of the stimuli, and also in both spatial and non-spatial visual tasks. Here, we employed a spatial delayed-response visual working memory task, and found that antagonism of PFC GABAergic transmission leads to the same delay-independent impairments, with concomitant slowing of response times, mirroring the results of Park & Holzman in their initial studies (1992) of spatial delayed-response working memory in schizophrenia patients, and many subsequent studies which have replicated the finding (Lee & Park, 2005). Furthermore, both our findings regarding decreased PFC GABAergic and findings in schizophrenia patients highlight encoding as a critical period in which working memory performance is affected (Haenschel et al. 2009; Kang et al. 2018; Minzenberg et al. 2010). We next found a deficit in sustained attention following PFC GABAA antagonism using the operant SAT, which bears resemblance to the Continuous Performance Test administered to human subjects. Several studies of schizophrenia patients have used the CPT to assess attention and have found an overall decrease in the ability to discriminate between signal and non-signal events, which stems from both decreased signal detection and increased false alarm rates (Cornblatt and Malhotra 2001; Finkelstein et al. 1997; Liu et al. 2002). In a translational version of the SAT administered to schizophrenia patients, false alarm rates were observed even in conditions of low distraction, while the impairments in signal detection emerged only when a visual distractor was added to the SAT (Lustig et al. 2013). While PFC GABA antagonism also 197  impaired discrimination of signal and non-signal events, it did so primarily by increasing the false alarm rate, suggesting a role for PFC GABA signalling in suppression of off-target responding, more so than signal detection, in schizophrenia. Deficits in signal detection may be mediated by other neurotransmitters in PFC, such as acetylcholine (Turchi and Sarter 1997), or by co-ordinated activity in other brain regions, such as thalamus (Wright et al. 2015) or posterior parietal cortex (Yang et al. 2017), which have been implicated in visual attention. Third, we observed that disruption of PFC GABA signalling led to disruptions in response inhibition in the form of an inability to follow conditional rules. Evidence of a similar impairment in schizophrenia patients has been observed during testing in several behavioral paradigms, including the visuospatial paired associates learning test of the CANTAB battery, and could contribute to a greater impairment in the ability to process contextual information (Hemsley 2005; Servan-Schreiber et al. 1996). Although increased impulsivity is considered to be one aspect of deficient response inhibition in schizophrenia (Fortgang et al. 2016) and though pharmacological inhibition of GAD enzymes also leads to increased motor impulsivity (Paine et al. 2015a), it appears that a net decrease in PFC GABAergic transmission does not contribute to increases in premature responding. Finally, when tested on a virtual version of the radial maze task, schizophrenia patients made both more reference and working memory errors (Spieker et al. 2012), just as PFC GABAA antagonism induced an increase in both these error types in animals performing the actual maze task. The presence of multiple back-to-back trials did not lead to increased proactive interference in rats with reduced PFC GABA function. Likewise, work with schizophrenia patients has shown that they are protected from the effects of proactive interference (Kaller et al. 2014). Rather, 198  patients have difficulty in dealing with distraction in the form of retroactive interference (Torres et al. 2001). Although we found that deficits in PFC GABA function were associated with slowing of responding in some instances, the absence of latency effects during actual choice periods suggests that decreased PFC GABAergic transmission is insufficient to account for speed-of-processing deficits in schizophrenia patients. It may be that alterations in function of other neurotransmitters, or perhaps in the functional connectivity of larger-scale networks, play a larger role in mediating deficient processing speed. Prominent abnormalities in white matter tracts connecting brain regions have been observed in schizophrenia, and the resulting dysconnectivity would be expected to have an impact on speed-of-processing (Kochunov et al. 2016).  Interneuron dysfunction in schizophrenia is thought to arise from alterations in PV, SST and perhaps other interneuronal subtypes. In the present work, we probed PFC GABA function in a way that would be expected to reduce the majority of fast GABAergic transmission in the PFC, blocking a substantial part of the activity of most interneurons. Past studies have shown that manipulations of particular interneuron types account for some, but not all aspects of schizophrenia. For instance, ablation of PFC PV neurons led to impairments in spatial short-term memory and cognitive flexibility, but did not alter social interaction or responsivity to stimulants, markers of negative and positive symptoms, respectively (Murray et al. 2015). On the other hand, the relatively broad manipulation of GABA function in the PFC employed here has been capable of recapitulating most cognitive deficits observed in schizophrenia, and features of positive and negative symptomatology (Paine et al. 2017; Piantadosi and Floresco 2014; Piantadosi et al. 2016; Tse et al. 2015b), supporting the idea that schizophrenia results from 199  alterations in multiple classes of interneurons. Likewise, pharmacological inhibition of GAD enzymes is also not sufficient to recapitulate schizophrenia-related cognitive impairment (Paine et al. 2015a), suggesting that a more profound alteration in PFC GABA function is necessary to produce aspects of altered cognition related to schizophrenia.  6.4.2 Other neuropsychiatric disorders Although the constellation of behavioral effects observed after bears striking similarity to symptoms and cognitive impairments described in schizophrenia, dysfunctional PFC GABA signalling has been observed in several other neuropsychiatric disorders, including depression and autism, and has also been observed in aging.  Therefore many of the findings described here have important implications for understanding the pathophysiology of GABA-related disorders as a whole. 6.4.2.1. Depression.  Meta-analysis has revealed that currently depressed patients have moderate alterations in attention, executive function, spatial working memory and paired associates learning (Rock et al. 2014). Some degree of impairment in attention, executive function and working memory are present in remitted patients as well (Rock et al. 2014), and also in people with high familial risk for mood disorders (Papmeyer et al. 2015), suggesting that cognitive dysfunction observed in depression is may be related to the biological underpinnings of the disorder, not just current symptomatology. In delayed-response tasks, patients with depression exhibit delay-independent impairments (Moffoot et al. 1994), and they are also consistently impaired on several other working memory tests, such as the n-back (Harvey et al. 2004; Rose and Ebmeier 2006), and backward digit span of the WAIS intelligence scale (Egeland et al. 2003).  Delay-independent impairments in working memory are consistent with the effects of PFC GABAA antagonism and also impaired encoding observed following SST 200  interneuron suppression, and are particularly interesting given that dysfunction of SST interneurons is strongly implicated in depression (Fee et al. 2017). In terms of attention, depression patients have repeatedly been found to be impaired in the CPT (Hsu et al. 2015), including both decreased hits and increase false alarms (Koetsier et al. 2002). The cognitive deficits associated with depression have been likened to an overall deficit in executive function, with accompanying deficits in maintenance of working memories (Snyder 2013). Though not the subject of this Thesis, shifting attentional sets, which is one component of executive function, is thought to be one of the most impaired aspects of cognition in major depressive disorder (Austin et al. 2001). Our group previously found a marked impairment in an operant set-shifting test following PFC GABAA antagonism (Enomoto et al. 2011). In this respect, cognitive deficits in PFC GABA function in animals appear to be similar to those observed in depressed patients, as delay-independent working memory impairments, decreased vigilance and evidence of executive dysfunction are present in both conditions. Hyperactivation of the PFC was observed in depressed patients performing the n-back task (Matsuo et al. 2007), and across several other tests assessing working memory and executive function (Papmeyer et al. 2015), suggesting that disturbed excitation-inhibition balance in this region, which could result from deficient PFC GABAergic transmission, may be associated with these impairments. 6.4.2.2. Autism & other neurodevelopmental disorders.   Imbalances in excitation and inhibition are also proposed to underlie cognitive impairments observed in autism. Throughout development and into adulthood, people with autism showed preserved processing speed but impaired accuracy in oculomotor delayed-response tasks (Luna et al. 2007). Although performance on shorter delays improved for autism patients over the course of development, a delay-independent impairment was maintained in adulthood (Luna et al. 2007). Further, these 201  impairments in delayed-responding are observed in first-degree relatives, suggesting they are related to genetic risk for autism (Koczat et al. 2002). People with autism also show perseverative responses in tests of behavioral flexibility, highlighting another element of executive function that is perturbed in the disorder (Ozonoff et al. 2004; Ozonoff and Jensen 1999). Thus, our findings that deficient PFC GABA signalling leads to disrupted working memory and cognitive flexibility, but not slower processing-speed, reflect what is observed in autism. In contrast, impaired attention, which was one of the functions most affected by loss of PFC GABA function, does not appear to be a facet of cognitive impairment associated with autism spectrum disorder. Children with autism displayed decreased hit rates in the CPT only when the reinforcement was social in comparison to more tangible rewards, including food or money. False alarm rate was not affected in this study (Garretson et al. 1990). (Pascualvaca et al. 1998; Yasuda et al. 2014) also found no significant effects in the CPT. These results are somewhat surprising given that we showed that even subtle perturbations in PFC GABA function were associated with attentional impairment. Furthermore, PV interneurons are thought to play a key role in attention (Kim et al. 2016b), and are believed to be preferentially affected in the PFC in autism. However, recent work has also suggested that the PFC PV chandelier cells are predominately affected (Ariza et al. 2018), thus it is possible that because the basket cells are less compromised, attentional processes regulated by PFC GABA are generally spared in autism. The lack of attentional affects in autism disorder point to a more restricted deficit in PFC GABA function relative to the one induced in the present work. 6.4.2.3. Epilepsy & traumatic brain injury-related disinhibition.  Importantly, disinhibition of PFC areas may occur in many other neurological and psychiatric conditions in 202  addition to the ones discussed in detail here, and may account for cognitive impairments observed in these conditions. Perhaps the most obvious of all functions of GABAergic transmission is to balance excitation, thereby preventing run-away activity. Thus, of all neuropsychiatric disorders, epilepsy and seizures have been most directly related to disrupted GABA function and/or excitation-inhibition imbalance. Notably, several epilepsy-related conditions occur with disruptions in cognitive functioning that are observed outside the ictal state, including deficits in working memory, decision-making, behavioral flexibility and attention (Hermann et al. 2010; Loughman et al. 2014; MacAllister et al. 2014; Stretton and Thompson 2012). In particular, patients with frontal lobe epilepsy appear to be more impaired in tests of executive function in comparison to temporal lobe epilepsy patients (Risse 2006), suggesting that persistent disinhibition of frontal regions may result in cognitive impairments similar to those observed here.  Traumatic brain injury (TBI) is one major cause underlying development of epilepsy disorders (Lucke-Wold et al. 2015), and similar disruptions in E-I balance are induced following TBI in animals (Nichols et al. 2015). Furthermore,  models of the disorder show reduced expression of markers of GABA function in frontal regions, including GAD65/67 and VGAT (Gu et al. 2017), as well as decreased GABAA  receptor binding and α1 subunit expression (López-Picón et al. 2016). In humans, working memory appears to be one of the most impacted cognitive functions following TBI (Dunning et al. 2016), while attentional deficits are also very commonly found (Mathias and Wheaton 2007). Given that the present results strongly implicate PFC GABA function as a key regulator of these cognitive functions, it is possible that deficits in inhibition induced by TBI may partially account for working memory and attentional impairments in patients with these injuries. 203  6.4.3 Cortical GABAergic deficits & age-related cognitive impairment   Deficits in cortical GABA function also emerge over the course of healthy aging.  Aged humans, primates and rats tend to show delay-dependent working memory deficits in delayed-response tasks (Beas et al. 2013; Rapp and Amaral 1989), though some work in humans has suggested impairment at relatively short delays (5 s) (Lyons-Warren et al. 2004). Electrophysiological studies of aged primate showed suppressed delay-period activity, suggesting impairments in maintenance, rather than activity associated with encoding cues (Wang et al. 2011). On the other hand, aging is associated with attentional impairments, including in the CPT (Fortenbaugh et al. 2015), that are described as an increase in distractibility. Intriguingly, work on this matter has shown that older adults often have better implicit memory because of increased encoding of off-target items (Rowe et al. 2006). Human imaging work showed a marked inability to suppress distraction-related activity in cortical regions that was related to working memory performance in aged individuals (Gazzaley et al. 2005).  These results are strikingly similar to what was observed following PFC GABAA antagonism, in that both PFC GABAA antagonism and aging are associated with deficient attentional regulation of encoding, and that these impairments may be related to aberrant cortical activity patterns. Older adults also show increased vulnerability to both proactive (Bowles and Salthouse 2003) and retroactive interference (Campbell et al. 2012) of working memory processes, and also are impaired in dealing with larger working memory loads (Mattay et al. 2006), emphasizing the need to investigate PFC GABAergic regulation of these aspects of working memory in future work.   204  6.4.4 PFC GABA hypofunction as a cross-diagnostic mechanism underlying cognitive impairment  Aging and the psychiatric disorders mentioned above share a common deficit in basic cognitive processes, including working memory, attention, and behavioral flexibility. The findings in this Thesis highlight a role for PFC GABAergic transmission in regulation of these processes and suggest that impairments in PFC GABA function can indeed reproduce many cognitive features of these disorders. While each of the conditions described share dysfunctional PFC GABA signalling as a feature, the nature of the GABA deficit and the interneuron types affected appear to be different, and indeed may be different in individual patients, perhaps partially accounting for differences in the presentation of cognitive impairment. More global deficits in PFC GABA function, such as those induced by PFC GABAA antagonism, appear to lead to a schizophrenia-like constellation of behavioral changes, including changes relevant to the positive and negative symptoms of the disorder, and in line with observations that multiple interneuron types are thought to function abnormally in schizophrenia. Aging is accompanied by a more subtle deficit in PFC GABA function, and correspondingly, more subtle cognitive and behavioral changes are observed. On the other hand, depression and autism spectrum disorders are thought to result from more selective impairments in SST and PV interneurons, respectively, that may explain why some PFC GABA-related impairments, but not others, are observed. Additionally, each of these conditions may be accompanied by other molecular, anatomical and neurochemical alterations that may interact with PFC GABA function to produce divergent symptomatology.  Nevertheless, the approach employed in this Thesis to assess deficits in PFC GABA function differs from these neuropsychiatric disorders discussed in important ways. 205  Schizophrenia appears to result from marked alterations in presynaptic interneuron function as well as changes in GABA receptor subunits, while PFC bicuculline treatment would be expected to produce an overall decrease in the PFC GABA input to post-synaptic neurons. PFC GABA alterations in autism and depression are thought to be somewhat more restricted to PV and SST interneurons, respectively, while altered PFC GABAergic transmission in aging is associated with changes in GABAB receptor function. Thus, the changes in PFC GABA function induced by administration of GABAA antagonist may either be too broad, or failing to take into account other compensatory alterations related to deficient PFC GABAergic transmission. While IEG (Paine et al. 2011), functional imaging (Parthoens et al. 2015) and physiological studies (Lodge 2011) clearly point to a state of PFC disinhibition induced by administration of GABAA antagonists, it is also not clear whether PFC is truly disinhibited in schizophrenia, other psychiatric disorders, or aging. In depression, MRI studies have indicated an overall hyperactivity of the anterior cingulate cortex, which may hinder cognitive flexibility and lead to rumination (Papmeyer et al. 2015), but PFC hypofunction has also been observed (Schecklmann et al. 2011). Importantly, MRI often reflects synaptic input rather than local spiking, so how these findings track with actual unit activity in the PFC in depression is not presently clear. In schizophrenia, diminished PFC GABAergic signalling appears to result from an initial decrease in glutamatergic signalling (Chung et al. 2016), perhaps via a preceding disruption in hippocampal development earlier in life, and decreased inhibition may reset the excitation-inhibition balance. Likewise, though aging is associated with interneuron loss and decreased inhibitory synapses, excitatory neurons are also lost and tonic levels of GABA are increased, potentially restoring excitation-inhibition balance somewhat. Nevertheless, these changes in 206  components that regulate excitation-inhibition balance could produce a decrease in the ability of the PFC GABAergic system to adaptively respond to in the face of cognitive challenge.  While pharmacological reduction of PFC GABAA receptor transmission cannot be viewed as an animal model of schizophrenia or any other condition, the present experiments and past work have been successful in identifying which aspects of cognition are affected by deficits in PFC GABA function. These studies have also provided some insights into the qualitative nature of these changes; for instance, by identifying component processes of working memory and attention affected by PFC GABA hypofunction that may be relevant to psychopathology.  These findings also highlight modulation of PFC GABA signalling as a potential route to ameliorate cognitive dysfunction observed across disparate disorders. While broadly-acting GABAergic drugs such as benzodiazepines are unlikely to be helpful in this regard because of sedative effects, there are many other routes to modulation of GABA signalling in the PFC; for instance, via modulation of PFC dopamine function (Gorelova et al. 2002; Seamans et al. 2001). Further, PFC GABAA antagonism may have some predictive validity as a system for screening compounds that potentially reverse deficiencies in PFC GABA transmission as a treatment for cognitive deficits observed in schizophrenia or other disorders.  6.5 Conclusions The findings described in this Thesis provide insight into the ways that PFC GABAergic transmission regulates core aspects of cognition, highlighting roles for PFC GABA in working memory encoding, attentional processes, and inhibition of inappropriate responses. Further, they reveal that deficits in PFC GABA signalling have deleterious effects on circuits throughout the 207  brain, thereby disrupting mnemonic and cognitive functions not necessarily dependent on the PFC. Strikingly, the constellation of behavioral and cognitive changes observed following pharmacological reduction of PFC signalling bears close resemblance to what is observed in schizophrenia, suggesting that dysfunctional PFC GABAergic transmission may play a key role in the pathophysiology of the disorder. Furthermore, since deficits in some of these cognitive domains are also observed in other conditions with reduced PFC GABA function, deficient PFC GABAergic transmission may be a common mechanism leading to deficient working memory and attention across neuropsychiatric conditions.     208  References Abbas AI, Sundiang MJM, Myrhe E, Henoch B, Morton MP, Bolkan SS, Harris AZ, Kellendonk C, Gordon JA (2017) The role of prefrontal interneuron types in working memory. Society for Neuroscience, Neuroscience Meeting Planner, Washington, DC. Addington J, Liu L, Perkins DO, Carrion RE, Keefe RS, Woods SW (2017) The Role of Cognition and Social Functioning as Predictors in the Transition to Psychosis for Youth With Attenuated Psychotic Symptoms. Schizophr Bull 43: 57-63. Adler CM, Malhotra AK, Elman I, Goldberg T, Egan M, Pickar D, Breier A (1999) Comparison of ketamine-induced thought disorder in healthy volunteers and thought disorder in schizophrenia. Am J Psychiatry 156: 1646-9. Aggleton JP, Nelson AJ (2015) Why do lesions in the rodent anterior thalamic nuclei cause such severe spatial deficits? Neurosci Biobehav Rev 54: 131-44. Airaksinen E, Larsson M, Lundberg I, Forsell Y (2004) Cognitive functions in depressive disorders: evidence from a population-based study. Psychol Med 34: 83-91. Ajram LA, Horder J, Mendez MA, Galanopoulos A, Brennan LP, Wichers RH, Robertson DM, Murphy CM, Zinkstok J, Ivin G, Heasman M, Meek D, Tricklebank MD, Barker GJ, Lythgoe DJ, Edden RAE, Williams SC, Murphy DGM, McAlonan GM (2017) Shifting brain inhibitory balance and connectivity of the prefrontal cortex of adults with autism spectrum disorder. Transl Psychiatry 7: e1137. Akaike N, Hattori K, Oomura Y, Carpenter DO (1985) Bicuculline and picrotoxin block gamma-aminobutyric acid-gated Cl- conductance by different mechanisms. Experientia 41: 70-1. Akbarian S, Huntsman MM, Kim JJ, Tafazzoli A, Potkin SG, Bunney WE, Jones EG (1995a) GABAA receptor subunit gene expression in human prefrontal cortex: comparison of schizophrenics and controls. Cereb Cortex 5: 550-60. Akbarian S, Kim JJ, Potkin SG, Hagman JO, Tafazzoli A, Bunney WE, Jones EG (1995b) Gene expression for glutamic acid decarboxylase is reduced without loss of neurons in prefrontal cortex of schizophrenics. Arch Gen Psychiatry 52: 258-66. Akhlaghpour H, Wiskerke J, Choi JY, Taliaferro JP, Au J, Witten IB (2016) Dissociated sequential activity and stimulus encoding in the dorsomedial striatum during spatial working memory. Elife 5. Albowitz B, Kuhnt U (1993) Spread of epileptiform potentials in the neocortical slice: recordings with voltage-sensitive dyes. Brain Res 631: 329-33. Alexander MP, Stuss DT, Shallice T, Picton TW, Gillingham S (2005) Impaired concentration due to frontal lobe damage from two distinct lesion sites. Neurology 65: 572-9. Allen RJ, Baddeley AD, Hitch GJ (2014) Evidence for two attentional components in visual working memory. J Exp Psychol Learn Mem Cogn 40: 1499-509. Alshammari TK, Alshammari MA, Nenov MN, Hoxha E, Cambiaghi M, Marcinno A, James TF, Singh P, Labate D, Li J, Meltzer HY, Sacchetti B, Tempia F, Laezza F (2016) Genetic 209  deletion of fibroblast growth factor 14 recapitulates phenotypic alterations underlying cognitive impairment associated with schizophrenia. Transl Psychiatry 6: e806. Amitai N, Kuczenski R, Behrens MM, Markou A (2012) Repeated phencyclidine administration alters glutamate release and decreases GABA markers in the prefrontal cortex of rats. Neuropharmacology 62: 1422-31. Anticevic A, Repovs G, Barch DM (2013) Working memory encoding and maintenance deficits in schizophrenia: neural evidence for activation and deactivation abnormalities. Schizophr Bull 39: 168-78. Ariza J, Rogers H, Hashemi E, Noctor SC, Martínez-Cerdeño V (2018) The Number of Chandelier and Basket Cells Are Differentially Decreased in Prefrontal Cortex in Autism. Cereb Cortex 28: 411-420. Asada H, Kawamura Y, Maruyama K, Kume H, Ding R, Ji FY, Kanbara N, Kuzume H, Sanbo M, Yagi T, Obata K (1996) Mice lacking the 65 kDa isoform of glutamic acid decarboxylase (GAD65) maintain normal levels of GAD67 and GABA in their brains but are susceptible to seizures. Biochem Biophys Res Commun 229: 891-5. Asada H, Kawamura Y, Maruyama K, Kume H, Ding RG, Kanbara N, Kuzume H, Sanbo M, Yagi T, Obata K (1997) Cleft palate and decreased brain gamma-aminobutyric acid in mice lacking the 67-kDa isoform of glutamic acid decarboxylase. Proc Natl Acad Sci U S A 94: 6496-9. Ascoli GA, Alonso-Nanclares L, Anderson SA, Barrionuevo G, Benavides-Piccione R, Burkhalter A, Buzsáki G, Cauli B, Defelipe J, Fairén A, Feldmeyer D, Fishell G, Fregnac Y, Freund TF, Gardner D, Gardner EP, Goldberg JH, Helmstaedter M, Hestrin S, Karube F, Kisvárday ZF, Lambolez B, Lewis DA, Marin O, Markram H, Muñoz A, Packer A, Petersen CC, Rockland KS, Rossier J, Rudy B, Somogyi P, Staiger JF, Tamas G, Thomson AM, Toledo-Rodriguez M, Wang Y, West DC, Yuste R, Group PIN (2008) Petilla terminology: nomenclature of features of GABAergic interneurons of the cerebral cortex. Nat Rev Neurosci 9: 557-68. Ashendorf L, McCaffrey RJ (2008) Exploring age-related decline on the Wisconsin Card Sorting Test. Clin Neuropsychol 22: 262-72. Aufhaus E, Weber-Fahr W, Sack M, Tunc-Skarka N, Oberthuer G, Hoerst M, Meyer-Lindenberg A, Boettcher U, Ende G (2013) Absence of changes in GABA concentrations with age and gender in the human anterior cingulate cortex: a MEGA-PRESS study with symmetric editing pulse frequencies for macromolecule suppression. Magn Reson Med 69: 317-20. Auger ML, Meccia J, Floresco SB (2017) Regulation of sustained attention, false alarm responding and implementation of conditional rules by prefrontal GABAA transmission: comparison with NMDA transmission. Psychopharmacology (Berl). Austin MP, Mitchell P, Goodwin GM (2001) Cognitive deficits in depression: possible implications for functional neuropathology. Br J Psychiatry 178: 200-6. Baddeley A (1986) Working memory. Oxford University Press, New York. 210  Baddeley A (2012) Working memory: theories, models, and controversies. Annu Rev Psychol 63: 1-29. Bari A, Dalley JW, Robbins TW (2008) The application of the 5-choice serial reaction time task for the assessment of visual attentional processes and impulse control in rats. Nat Protoc 3: 759-67. Barr MS, Rajji TK, Zomorrodi R, Radhu N, George TP, Blumberger DM, Daskalakis ZJ (2017) Impaired theta-gamma coupling during working memory performance in schizophrenia. Schizophr Res 189: 104-110. Basar-Eroglu C, Brand A, Hildebrandt H, Karolina Kedzior K, Mathes B, Schmiedt C (2007) Working memory related gamma oscillations in schizophrenia patients. Int J Psychophysiol 64: 39-45. Bañuelos C, Beas BS, McQuail JA, Gilbert RJ, Frazier CJ, Setlow B, Bizon JL (2014) Prefrontal cortical GABAergic dysfunction contributes to age-related working memory impairment. J Neurosci 34: 3457-66. Beas BS, McQuail JA, Ban Uelos C, Setlow B, Bizon JL (2017) Prefrontal cortical GABAergic signaling and impaired behavioral flexibility in aged F344 rats. Neuroscience 345: 274-286. Beas BS, Setlow B, Bizon JL (2013) Distinct manifestations of executive dysfunction in aged rats. Neurobiol Aging 34: 2164-74. Beasley CL, Zhang ZJ, Patten I, Reynolds GP (2002) Selective deficits in prefrontal cortical GABAergic neurons in schizophrenia defined by the presence of calcium-binding proteins. Biol Psychiatry 52: 708-15. Becker JT, Walker JA, Olton DS (1980) Neuroanatomical bases of spatial memory. Brain Res 200: 307-20. Behrens MM, Ali SS, Dao DN, Lucero J, Shekhtman G, Quick KL, Dugan LL (2007) Ketamine-induced loss of phenotype of fast-spiking interneurons is mediated by NADPH-oxidase. Science 318: 1645-7. Belelli D, Harrison NL, Maguire J, Macdonald RL, Walker MC, Cope DW (2009) Extrasynaptic GABAA receptors: form, pharmacology, and function. J Neurosci 29: 12757-63. Belelli D, Lambert JJ (2005) Neurosteroids: endogenous regulators of the GABA(A) receptor. Nat Rev Neurosci 6: 565-75. Belforte JE, Zsiros V, Sklar ER, Jiang Z, Yu G, Li Y, Quinlan EM, Nakazawa K (2010) Postnatal NMDA receptor ablation in corticolimbic interneurons confers schizophrenia-like phenotypes. Nat Neurosci 13: 76-83. Benes FM (1995) Altered glutamatergic and GABAergic mechanisms in the cingulate cortex of the schizophrenic brain. Arch Gen Psychiatry 52: 1015-8; discussion 1019-24. Benes FM (2015) The GABA system in schizophrenia: cells, molecules and microcircuitry. Schizophr Res 167: 1-3. 211  Beneyto M, Abbott A, Hashimoto T, Lewis DA (2011) Lamina-specific alterations in cortical GABA(A) receptor subunit expression in schizophrenia. Cereb Cortex 21: 999-1011. Benn A, Barker GR, Stuart SA, Roloff EV, Teschemacher AG, Warburton EC, Robinson ES (2016) Optogenetic Stimulation of Prefrontal Glutamatergic Neurons Enhances Recognition Memory. J Neurosci 36: 4930-9. Berman KF, Weinberger DR (1990) Lateralisation of cortical function during cognitive tasks: regional cerebral blood flow studies of normal individuals and patients with schizophrenia. J Neurol Neurosurg Psychiatry 53: 150-60. Bi LL, Wang J, Luo ZY, Chen SP, Geng F, Chen YH, Li SJ, Yuan CH, Lin S, Gao TM (2013) Enhanced excitability in the infralimbic cortex produces anxiety-like behaviors. Neuropharmacology 72: 148-56. Birnbaum R, Weinberger DR (2017) Genetic insights into the neurodevelopmental origins of schizophrenia. Nat Rev Neurosci 18: 727-740. Bolkan SS, Stujenske JM, Parnaudeau S, Spellman TJ, Rauffenbart C, Abbas AI, Harris AZ, Gordon JA, Kellendonk C (2017) Thalamic projections sustain prefrontal activity during working memory maintenance. Nat Neurosci 20: 987-996. Bondi C, Matthews M, Moghaddam B (2012) Glutamatergic animal models of schizophrenia. Curr Pharm Des 18: 1593-604. Bories C, Husson Z, Guitton MJ, De Koninck Y (2013) Differential balance of prefrontal synaptic activity in successful versus unsuccessful cognitive aging. J Neurosci 33: 1344-56. Bowery NG, Hill DR, Hudson AL, Doble A, Middlemiss DN, Shaw J, Turnbull M (1980) (-)Baclofen decreases neurotransmitter release in the mammalian CNS by an action at a novel GABA receptor. Nature 283: 92-4. Bowles RP, Salthouse TA (2003) Assessing the age-related effects of proactive interference on working memory tasks using the Rasch model. Psychol Aging 18: 608-15. Breier A, Su TP, Saunders R, Carson RE, Kolachana BS, de Bartolomeis A, Weinberger DR, Weisenfeld N, Malhotra AK, Eckelman WC, Pickar D (1997) Schizophrenia is associated with elevated amphetamine-induced synaptic dopamine concentrations: evidence from a novel positron emission tomography method. Proc Natl Acad Sci U S A 94: 2569-74. Brickley SG, Mody I (2012) Extrasynaptic GABA(A) receptors: their function in the CNS and implications for disease. Neuron 73: 23-34. Brockmann MD, Pöschel B, Cichon N, Hanganu-Opatz IL (2011) Coupled oscillations mediate directed interactions between prefrontal cortex and hippocampus of the neonatal rat. Neuron 71: 332-47. Brown JA, Ramikie TS, Schmidt MJ, Báldi R, Garbett K, Everheart MG, Warren LE, Gellért L, Horváth S, Patel S, Mirnics K (2015) Inhibition of parvalbumin-expressing interneurons results in complex behavioral changes. Mol Psychiatry 20: 1499-507. 212  Buzsáki G, Wang XJ (2012) Mechanisms of gamma oscillations. Annu Rev Neurosci 35: 203-25. Campbell KL, Grady CL, Ng C, Hasher L (2012) Age differences in the frontoparietal cognitive control network: implications for distractibility. Neuropsychologia 50: 2212-23. Canetta S, Bolkan S, Padilla-Coreano N, Song LJ, Sahn R, Harrison NL, Gordon JA, Brown A, Kellendonk C (2016) Maternal immune activation leads to selective functional deficits in offspring parvalbumin interneurons. Mol Psychiatry 21: 956-68. Cannon TD, van Erp TG, Rosso IM, Huttunen M, Lönnqvist J, Pirkola T, Salonen O, Valanne L, Poutanen VP, Standertskjöld-Nordenstam CG (2002) Fetal hypoxia and structural brain abnormalities in schizophrenic patients, their siblings, and controls. Arch Gen Psychiatry 59: 35-41. Cardin JA, Carlén M, Meletis K, Knoblich U, Zhang F, Deisseroth K, Tsai LH, Moore CI (2009) Driving fast-spiking cells induces gamma rhythm and controls sensory responses. Nature 459: 663-7. Carlén M, Meletis K, Siegle JH, Cardin JA, Futai K, Vierling-Claassen D, Rühlmann C, Jones SR, Deisseroth K, Sheng M, Moore CI, Tsai LH (2012) A critical role for NMDA receptors in parvalbumin interneurons for gamma rhythm induction and behavior. Mol Psychiatry 17: 537-48. Cascella NG, Schretlen DJ, Sawa A (2009) Schizophrenia and epilepsy: is there a shared susceptibility? Neurosci Res 63: 227-35. Cassel JC, Pereira de Vasconcelos A, Loureiro M, Cholvin T, Dalrymple-Alford JC, Vertes RP (2013) The reuniens and rhomboid nuclei: neuroanatomy, electrophysiological characteristics and behavioral implications. Prog Neurobiol 111: 34-52. Chamberlain SR, Sahakian BJ (2006) The neuropsychology of mood disorders. Curr Psychiatry Rep 8: 458-63. Chattopadhyaya B, Di Cristo G, Higashiyama H, Knott GW, Kuhlman SJ, Welker E, Huang ZJ (2004) Experience and activity-dependent maturation of perisomatic GABAergic innervation in primary visual cortex during a postnatal critical period. J Neurosci 24: 9598-611. Chattopadhyaya B, Di Cristo G, Wu CZ, Knott G, Kuhlman S, Fu Y, Palmiter RD, Huang ZJ (2007) GAD67-mediated GABA synthesis and signaling regulate inhibitory synaptic innervation in the visual cortex. Neuron 54: 889-903. Chen CM, Stanford AD, Mao X, Abi-Dargham A, Shungu DC, Lisanby SH, Schroeder CE, Kegeles LS (2014) GABA level, gamma oscillation, and working memory performance in schizophrenia. Neuroimage Clin 4: 531-9. Chen G, Zhang Y, Li X, Zhao X, Ye Q, Lin Y, Tao HW, Rasch MJ, Zhang X (2017) Distinct Inhibitory Circuits Orchestrate Cortical beta and gamma Band Oscillations. Neuron 96: 1403-1418.e6. 213  Chen WJ, Faraone SV (2000) Sustained attention deficits as markers of genetic susceptibility to schizophrenia. Am J Med Genet 97: 52-7. Chervin RD, Pierce PA, Connors BW (1988) Periodicity and directionality in the propagation of epileptiform discharges across neocortex. J Neurophysiol 60: 1695-713. Chiu PW, Lui SSY, Hung KSY, Chan RCK, Chan Q, Sham PC, Cheung EFC, Mak HKF (2017) In vivo gamma-aminobutyric acid and glutamate levels in people with first-episode schizophrenia: A proton magnetic resonance spectroscopy study. Schizophr Res. Cho KK, Hoch R, Lee AT, Patel T, Rubenstein JL, Sohal VS (2015) Gamma rhythms link prefrontal interneuron dysfunction with cognitive inflexibility in Dlx5/6(+/-) mice. Neuron 85: 1332-43. Cho RY, Konecky RO, Carter CS (2006) Impairments in frontal cortical gamma synchrony and cognitive control in schizophrenia. Proc Natl Acad Sci U S A 103: 19878-83. Chrobak JJ, Hinman JR, Sabolek HR (2008) Revealing past memories: proactive interference and ketamine-induced memory deficits. J Neurosci 28: 4512-20. Chudasama Y, Muir JL (1997) A behavioural analysis of the delayed non-matching to position task: the effects of scopolamine, lesions of the fornix and of the prelimbic region on mediating behaviours by rats. Psychopharmacology (Berl) 134: 73-82. Chudasama Y, Passetti F, Rhodes SE, Lopian D, Desai A, Robbins TW (2003) Dissociable aspects of performance on the 5-choice serial reaction time task following lesions of the dorsal anterior cingulate, infralimbic and orbitofrontal cortex in the rat: differential effects on selectivity, impulsivity and compulsivity. Behav Brain Res 146: 105-19. Chung DW, Fish KN, Lewis DA (2016) Pathological Basis for Deficient Excitatory Drive to Cortical Parvalbumin Interneurons in Schizophrenia. Am J Psychiatry 173: 1131-1139. Cohen RM, Semple WE, Gross M, King AC, Nordahl TE (1992) Metabolic brain pattern of sustained auditory discrimination. Exp Brain Res 92: 165-72. Cole BJ, Klewer M, Jones GH, Stephens DN (1993) Contrasting effects of the competitive NMDA antagonist CPP and the non-competitive NMDA antagonist MK 801 on performance of an operant delayed matching to position task in rats. Psychopharmacology (Berl) 111: 465-71. Colombo PJ, Davis HP, Volpe BT (1989) Allocentric spatial and tactile memory impairments in rats with dorsal caudate lesions are affected by preoperative behavioral training. Behav Neurosci 103: 1242-50. Conklin HM, Curtis CE, Katsanis J, Iacono WG (2000) Verbal working memory impairment in schizophrenia patients and their first-degree relatives: evidence from the digit span task. Am J Psychiatry 157: 275-7. Corigliano V, De Carolis A, Trovini G, Dehning J, Di Pietro S, Curto M, Donato N, De Pisa E, Girardi P, Comparelli A (2014) Neurocognition in schizophrenia: from prodrome to multi-episode illness. Psychiatry Res 220: 129-34. 214  Cornblatt BA, Malhotra AK (2001) Impaired attention as an endophenotype for molecular genetic studies of schizophrenia. Am J Med Genet 105: 11-5. Coull JT, Frackowiak RS, Frith CD (1998) Monitoring for target objects: activation of right frontal and parietal cortices with increasing time on task. Neuropsychologia 36: 1325-34. Coyle JT (2004) The GABA-glutamate connection in schizophrenia: which is the proximate cause? Biochem Pharmacol 68: 1507-14. Crabtree GW, Park AJ, Gordon JA, Gogos JA (2016) Cytosolic Accumulation of L-Proline Disrupts GABA-Ergic Transmission through GAD Blockade. Cell Rep 17: 570-582. Culbreth A, Westbrook A, Barch D (2016) Negative symptoms are associated with an increased subjective cost of cognitive effort. J Abnorm Psychol 125: 528-536. Curley AA, Arion D, Volk DW, Asafu-Adjei JK, Sampson AR, Fish KN, Lewis DA (2011) Cortical deficits of glutamic acid decarboxylase 67 expression in schizophrenia: clinical, protein, and cell type-specific features. Am J Psychiatry 168: 921-9. Dalton GL, Ma LM, Phillips AG, Floresco SB (2011) Blockade of NMDA GluN2B receptors selectively impairs behavioral flexibility but not initial discrimination learning. Psychopharmacology (Berl) 216: 525-35. Davies DA, Greba Q, Howland JG (2013a) GluN2B-containing NMDA receptors and AMPA receptors in medial prefrontal cortex are necessary for odor span in rats. Front Behav Neurosci 7: 183. Davies DA, Molder JJ, Greba Q, Howland JG (2013b) Inactivation of medial prefrontal cortex or acute stress impairs odor span in rats. Learn Mem 20: 665-9. DeFelipe J, López-Cruz PL, Benavides-Piccione R, Bielza C, Larrañaga P, Anderson S, Burkhalter A, Cauli B, Fairén A, Feldmeyer D, Fishell G, Fitzpatrick D, Freund TF, González-Burgos G, Hestrin S, Hill S, Hof PR, Huang J, Jones EG, Kawaguchi Y, Kisvárday Z, Kubota Y, Lewis DA, Marín O, Markram H, McBain CJ, Meyer HS, Monyer H, Nelson SB, Rockland K, Rossier J, Rubenstein JL, Rudy B, Scanziani M, Shepherd GM, Sherwood CC, Staiger JF, Tamás G, Thomson A, Wang Y, Yuste R, Ascoli GA (2013) New insights into the classification and nomenclature of cortical GABAergic interneurons. Nat Rev Neurosci 14: 202-16. Delli Pizzi S, Chiacchiaretta P, Mantini D, Bubbico G, Edden RA, Onofrj M, Ferretti A, Bonanni L (2017) GABA content within medial prefrontal cortex predicts the variability of fronto-limbic effective connectivity. Brain Struct Funct 222: 3217-3229. Demeter E, Guthrie SK, Taylor SF, Sarter M, Lustig C (2013) Increased distractor vulnerability but preserved vigilance in patients with schizophrenia: evidence from a translational Sustained Attention Task. Schizophr Res 144: 136-41. Di Re J, Wadsworth PA, Laezza F (2017) Intracellular Fibroblast Growth Factor 14: Emerging Risk Factor for Brain Disorders. Front Cell Neurosci 11: 103. 215  Dickinson A, Jones M, Milne E (2016) Measuring neural excitation and inhibition in autism: Different approaches, different findings and different interpretations. Brain Res 1648: 277-289. Dickinson D, Ramsey ME, Gold JM (2007) Overlooking the obvious: a meta-analytic comparison of digit symbol coding tasks and other cognitive measures in schizophrenia. Arch Gen Psychiatry 64: 532-42. Do Monte FH, Quirk GJ, Li B, Penzo MA (2016) Retrieving fear memories, as time goes by…. Mol Psychiatry 21: 1027-36. Doze VA, Cohen GA, Madison DV (1995) Calcium channel involvement in GABAB receptor-mediated inhibition of GABA release in area CA1 of the rat hippocampus. J Neurophysiol 74: 43-53. Dudchenko PA, Talpos J, Young J, Baxter MG (2013) Animal models of working memory: a review of tasks that might be used in screening drug treatments for the memory impairments found in schizophrenia. Neurosci Biobehav Rev 37: 2111-24. Dunning DL, Westgate B, Adlam AR (2016) A meta-analysis of working memory impairments in survivors of moderate-to-severe traumatic brain injury. Neuropsychology 30: 811-819. Duva CA, Floresco SB, Wunderlich GR, Lao TL, Pinel JP, Phillips AG (1997) Disruption of spatial but not object-recognition memory by neurotoxic lesions of the dorsal hippocampus in rats. Behav Neurosci 111: 1184-96. Egeland J, Sundet K, Rund BR, Asbjørnsen A, Hugdahl K, Landrø NI, Lund A, Roness A, Stordal KI (2003) Sensitivity and specificity of memory dysfunction in schizophrenia: a comparison with major depression. J Clin Exp Neuropsychol 25: 79-93. Egerton A, Modinos G, Ferrera D, McGuire P (2017) Neuroimaging studies of GABA in schizophrenia: a systematic review with meta-analysis. Transl Psychiatry 7: e1147. Enomoto T, Tse MT, Floresco SB (2011) Reducing prefrontal gamma-aminobutyric acid activity induces cognitive, behavioral, and dopaminergic abnormalities that resemble schizophrenia. Biol Psychiatry 69: 432-41. Erlander MG, Tillakaratne NJ, Feldblum S, Patel N, Tobin AJ (1991) Two genes encode distinct glutamate decarboxylases. Neuron 7: 91-100. Esclapez M, Tillakaratne NJ, Kaufman DL, Tobin AJ, Houser CR (1994) Comparative localization of two forms of glutamic acid decarboxylase and their mRNAs in rat brain supports the concept of functional differences between the forms. J Neurosci 14: 1834-55. Fazzari P, Paternain AV, Valiente M, Pla R, Luján R, Lloyd K, Lerma J, Marín O, Rico B (2010) Control of cortical GABA circuitry development by Nrg1 and ErbB4 signalling. Nature 464: 1376-80. Fee C, Banasr M, Sibille E (2017) Somatostatin-Positive Gamma-Aminobutyric Acid Interneuron Deficits in Depression: Cortical Microcircuit and Therapeutic Perspectives. Biol Psychiatry 82: 549-559. 216  Feja M, Koch M (2015) Frontostriatal systems comprising connections between ventral medial prefrontal cortex and nucleus accumbens subregions differentially regulate motor impulse control in rats. Psychopharmacology (Berl) 232: 1291-302. Feldblum S, Erlander MG, Tobin AJ (1993) Different distributions of GAD65 and GAD67 mRNAs suggest that the two glutamate decarboxylases play distinctive functional roles. J Neurosci Res 34: 689-706. Fervaha G, Graff-Guerrero A, Zakzanis KK, Foussias G, Agid O, Remington G (2013) Incentive motivation deficits in schizophrenia reflect effort computation impairments during cost-benefit decision-making. J Psychiatr Res 47: 1590-6. Fervaha G, Duncan M, Foussias G, Agid O, Faulkner GE, Remington G (2015) Effort-based decision making as an objective paradigm for the assessment of motivational deficits in schizophrenia. Schizophr Res 168: 483-90. Finkelstein JR, Cannon TD, Gur RE, Gur RC, Moberg P (1997) Attentional dysfunctions in neuroleptic-naive and neuroleptic-withdrawn schizophrenic patients and their siblings. J Abnorm Psychol 106: 203-12. Floresco SB (2013) Prefrontal NMDA receptors and cognition: working 2B remembered. Neuron 77: 603-5. Floresco SB, Block AE, Tse MT (2008) Inactivation of the medial prefrontal cortex of the rat impairs strategy set-shifting, but not reversal learning, using a novel, automated procedure. Behav Brain Res 190: 85-96. Floresco SB, Jentsch JD (2011) Pharmacological enhancement of memory and executive functioning in laboratory animals. Neuropsychopharmacology 36: 227-50. Floresco SB, Seamans JK, Phillips AG (1997) Selective roles for hippocampal, prefrontal cortical, and ventral striatal circuits in radial-arm maze tasks with or without a delay. J Neurosci 17: 1880-90. Floresco SB, Zhang Y, Enomoto T (2009) Neural circuits subserving behavioral flexibility and their relevance to schizophrenia. Behav Brain Res 204: 396-409. Fortenbaugh FC, DeGutis J, Germine L, Wilmer JB, Grosso M, Russo K, Esterman M (2015) Sustained Attention Across the Life Span in a Sample of 10,000: Dissociating Ability and Strategy. Psychol Sci 26: 1497-510. Fortgang RG, Hultman CM, van Erp TG, Cannon TD (2016) Multidimensional assessment of impulsivity in schizophrenia, bipolar disorder, and major depressive disorder: testing for shared endophenotypes. Psychol Med 46: 1497-507. Foss-Feig JH, Adkinson BD, Ji JL, Yang G, Srihari VH, McPartland JC, Krystal JH, Murray JD, Anticevic A (2017) Searching for Cross-Diagnostic Convergence: Neural Mechanisms Governing Excitation and Inhibition Balance in Schizophrenia and Autism Spectrum Disorders. Biol Psychiatry 81: 848-861. 217  Frankle WG, Cho RY, Narendran R, Mason NS, Vora S, Litschge M, Price JC, Lewis DA, Mathis CA (2009) Tiagabine increases [11C]flumazenil binding in cortical brain regions in healthy control subjects. Neuropsychopharmacology 34: 624-33. Frankle WG, Cho RY, Prasad KM, Mason NS, Paris J, Himes ML, Walker C, Lewis DA, Narendran R (2015) In vivo measurement of GABA transmission in healthy subjects and schizophrenia patients. Am J Psychiatry 172: 1148-59. François J, Ferrandon A, Koning E, Angst MJ, Sandner G, Nehlig A (2009) Selective reorganization of GABAergic transmission in neonatal ventral hippocampal-lesioned rats. Int J Neuropsychopharmacol 12: 1097-110. Fritschy JM, Mohler H (1995) GABAA-receptor heterogeneity in the adult rat brain: differential regional and cellular distribution of seven major subunits. J Comp Neurol 359: 154-94. Fritschy JM, Paysan J, Enna A, Mohler H (1994) Switch in the expression of rat GABAA-receptor subtypes during postnatal development: an immunohistochemical study. J Neurosci 14: 5302-24. Fujihara K, Miwa H, Kakizaki T, Kaneko R, Mikuni M, Tanahira C, Tamamaki N, Yanagawa Y (2015) Glutamate Decarboxylase 67 Deficiency in a Subset of GABAergic Neurons Induces Schizophrenia-Related Phenotypes. Neuropsychopharmacology 40: 2475-86. Fung SJ, Webster MJ, Sivagnanasundaram S, Duncan C, Elashoff M, Weickert CS (2010) Expression of interneuron markers in the dorsolateral prefrontal cortex of the developing human and in schizophrenia. Am J Psychiatry 167: 1479-88. Gabbay V, Mao X, Klein RG, Ely BA, Babb JS, Panzer AM, Alonso CM, Shungu DC (2012) Anterior cingulate cortex γ-aminobutyric acid in depressed adolescents: relationship to anhedonia. Arch Gen Psychiatry 69: 139-49. Garretson HB, Fein D, Waterhouse L (1990) Sustained attention in children with autism. J Autism Dev Disord 20: 101-14. Gazzaley A, Cooney JW, Rissman J, D'Esposito M (2005) Top-down suppression deficit underlies working memory impairment in normal aging. Nat Neurosci 8: 1298-300. Gazzaley A, Nobre AC (2012) Top-down modulation: bridging selective attention and working memory. Trends Cogn Sci 16: 129-35. Gill TM, Sarter M, Givens B (2000) Sustained visual attention performance-associated prefrontal neuronal activity: evidence for cholinergic modulation. J Neurosci 20: 4745-57. Giusti-Rodríguez P, Sullivan PF (2013) The genomics of schizophrenia: update and implications. J Clin Invest 123: 4557-63. Goldberg TE, Weinberger DR (1988) Probing prefrontal function in schizophrenia with neuropsychological paradigms. Schizophr Bull 14: 179-83. Goldman-Rakic PS (1991) Prefrontal cortical dysfunction in schizophrenia: the relevance of working memory. In: Carroll B (ed) Psychopathology and the brain. Raven Press, New York, pp 1-23 218  Goldman-Rakic PS (1994) Working memory dysfunction in schizophrenia. In: Salloway SP, Malloy PF, Duffy JD (eds) The Frontal lobes and neuropsychiatric illness. American Psychiatric Publishing, Washington, DC, pp 71-81 Gonzalez-Burgos G, Lewis DA (2012) NMDA receptor hypofunction, parvalbumin-positive neurons, and cortical gamma oscillations in schizophrenia. Schizophr Bull 38: 950-7. Gorelova N, Seamans JK, Yang CR (2002) Mechanisms of dopamine activation of fast-spiking interneurons that exert inhibition in rat prefrontal cortex. J Neurophysiol 88: 3150-66. Goto N, Yoshimura R, Kakeda S, Moriya J, Hori H, Hayashi K, Ikenouchi-Sugita A, Nakano-Umene W, Katsuki A, Nishimura J, Korogi Y, Nakamura J (2010) No alterations of brain GABA after 6 months of treatment with atypical antipsychotic drugs in early-stage first-episode schizophrenia. Prog Neuropsychopharmacol Biol Psychiatry 34: 1480-3. Grachev ID, Swarnkar A, Szeverenyi NM, Ramachandran TS, Apkarian AV (2001) Aging alters the multichemical networking profile of the human brain: an in vivo (1)H-MRS study of young versus middle-aged subjects. J Neurochem 77: 292-303. Gratacòs M, González JR, Mercader JM, de Cid R, Urretavizcaya M, Estivill X (2007) Brain-derived neurotrophic factor Val66Met and psychiatric disorders: meta-analysis of case-control studies confirm association to substance-related disorders, eating disorders, and schizophrenia. Biol Psychiatry 61: 911-22. Green MF (1996) What are the functional consequences of neurocognitive deficits in schizophrenia? Am J Psychiatry 153: 321-30. Green MF, Nuechterlein KH (1999) Should schizophrenia be treated as a neurocognitive disorder? Schizophr Bull 25: 309-19. Grimm R, Schicknick H, Riede I, Gundelfinger ED, Herdegen T, Zuschratter W, Tischmeyer W (1997) Suppression of c-fos induction in rat brain impairs retention of a brightness discrimination reaction. Learn Mem 3: 402-13. Gu F, Parada I, Shen F, Li J, Bacci A, Graber K, Taghavi RM, Scalise K, Schwartzkroin P, Wenzel J, Prince DA (2017) Structural alterations in fast-spiking GABAergic interneurons in a model of posttraumatic neocortical epileptogenesis. Neurobiol Dis 108: 100-114. Guidotti A, Auta J, Davis JM, Di-Giorgi-Gerevini V, Dwivedi Y, Grayson DR, Impagnatiello F, Pandey G, Pesold C, Sharma R, Uzunov D, Costa E, DiGiorgi Gerevini V (2000) Decrease in reelin and glutamic acid decarboxylase67 (GAD67) expression in schizophrenia and bipolar disorder: a postmortem brain study. Arch Gen Psychiatry 57: 1061-9. Guise KG, Shapiro ML (2017) Medial Prefrontal Cortex Reduces Memory Interference by Modifying Hippocampal Encoding. Neuron 94: 183-192.e8. Gähwiler BH, Brown DA (1985) GABAB-receptor-activated K+ current in voltage-clamped CA3 pyramidal cells in hippocampal cultures. Proc Natl Acad Sci U S A 82: 1558-62. 219  Haenschel C, Bittner RA, Waltz J, Haertling F, Wibral M, Singer W, Linden DE, Rodriguez E (2009) Cortical oscillatory activity is critical for working memory as revealed by deficits in early-onset schizophrenia. J Neurosci 29: 9481-9. Hallock HL, Wang A, Griffin AL (2016) Ventral Midline Thalamus Is Critical for Hippocampal-Prefrontal Synchrony and Spatial Working Memory. J Neurosci 36: 8372-89. Hammar A, Lund A, Hugdahl K (2003) Long-lasting cognitive impairment in unipolar major depression: a 6-month follow-up study. Psychiatry Res 118: 189-96. Hanlon FM, Houck JM, Klimaj SD, Caprihan A, Mayer AR, Weisend MP, Bustillo JR, Hamilton DA, Tesche CD (2012) Frontotemporal anatomical connectivity and working-relational memory performance predict everyday functioning in schizophrenia. Psychophysiology 49: 1340-52. Haque T, Yamamoto S, Masuda Y, Kato T, Sato F, Uchino K, Oka A, Nakamura M, Takeda R, Ono T, Kogo M, Yoshida A (2010) Thalamic afferent and efferent connectivity to cerebral cortical areas with direct projections to identified subgroups of trigeminal premotoneurons in the rat. Brain Res 1346: 69-82. Hartman M, Steketee MC, Silva S, Lanning K, Andersson C (2003a) Wisconsin Card Sorting Test performance in schizophrenia: the role of working memory. Schizophr Res 63: 201-17. Hartman M, Steketee MC, Silva S, Lanning K, McCann H (2003b) Working memory and schizophrenia: evidence for slowed encoding. Schizophr Res 59: 99-113. Hartmann MN, Hager OM, Reimann AV, Chumbley JR, Kirschner M, Seifritz E, Tobler PN, Kaiser S (2015) Apathy but not diminished expression in schizophrenia is associated with discounting of monetary rewards by physical effort. Schizophr Bull 41: 503-12. Hartung H, Cichon N, De Feo V, Riemann S, Schildt S, Lindemann C, Mulert C, Gogos JA, Hanganu-Opatz IL (2016) From Shortage to Surge: A Developmental Switch in Hippocampal-Prefrontal Coupling in a Gene-Environment Model of Neuropsychiatric Disorders. Cereb Cortex. Harvey PO, Le Bastard G, Pochon JB, Levy R, Allilaire JF, Dubois B, Fossati P (2004) Executive functions and updating of the contents of working memory in unipolar depression. J Psychiatr Res 38: 567-76. Harvey RE, Thompson SM, Sanchez LM, Yoder RM, Clark BJ (2017) Post-training Inactivation of the Anterior Thalamic Nuclei Impairs Spatial Performance on the Radial Arm Maze. Front Neurosci 11: 94. Hashemi E, Ariza J, Rogers H, Noctor SC, Martínez-Cerdeño V (2017) The Number of Parvalbumin-Expressing Interneurons Is Decreased in the Medial Prefrontal Cortex in Autism. Cereb Cortex 27: 1931-1943. Hashimoto T, Arion D, Unger T, Maldonado-Avilés JG, Morris HM, Volk DW, Mirnics K, Lewis DA (2008a) Alterations in GABA-related transcriptome in the dorsolateral prefrontal cortex of subjects with schizophrenia. Mol Psychiatry 13: 147-61. 220  Hashimoto T, Bazmi HH, Mirnics K, Wu Q, Sampson AR, Lewis DA (2008b) Conserved regional patterns of GABA-related transcript expression in the neocortex of subjects with schizophrenia. Am J Psychiatry 165: 479-89. Hashimoto T, Bergen SE, Nguyen QL, Xu B, Monteggia LM, Pierri JN, Sun Z, Sampson AR, Lewis DA (2005) Relationship of brain-derived neurotrophic factor and its receptor TrkB to altered inhibitory prefrontal circuitry in schizophrenia. J Neurosci 25: 372-83. Hashimoto T, Nguyen QL, Rotaru D, Keenan T, Arion D, Beneyto M, Gonzalez-Burgos G, Lewis DA (2009) Protracted developmental trajectories of GABAA receptor alpha1 and alpha2 subunit expression in primate prefrontal cortex. Biol Psychiatry 65: 1015-23. Hashimoto T, Volk DW, Eggan SM, Mirnics K, Pierri JN, Sun Z, Sampson AR, Lewis DA (2003) Gene expression deficits in a subclass of GABA neurons in the prefrontal cortex of subjects with schizophrenia. J Neurosci 23: 6315-26. Hasler G, van der Veen JW, Tumonis T, Meyers N, Shen J, Drevets WC (2007) Reduced prefrontal glutamate/glutamine and gamma-aminobutyric acid levels in major depression determined using proton magnetic resonance spectroscopy. Arch Gen Psychiatry 64: 193-200. He J, Yamada K, Nabeshima T (2002a) A role of Fos expression in the CA3 region of the hippocampus in spatial memory formation in rats. Neuropsychopharmacology 26: 259-68. He J, Yamada K, Nakajima A, Kamei H, Nabeshima T (2002b) Learning and memory in two different reward tasks in a radial arm maze in rats. Behav Brain Res 134: 139-48. He Y, Benz A, Fu T, Wang M, Covey DF, Zorumski CF, Mennerick S (2002c) Neuroprotective agent riluzole potentiates postsynaptic GABA(A) receptor function. Neuropharmacology 42: 199-209. Heerey EA, Bell-Warren KR, Gold JM (2008) Decision-making impairments in the context of intact reward sensitivity in schizophrenia. Biol Psychiatry 64: 62-9. Hegerl U, Himmerich H, Engmann B, Hensch T (2010) Mania and attention-deficit/hyperactivity disorder: common symptomatology, common pathophysiology and common treatment? Curr Opin Psychiatry 23: 1-7. Hemsley DR (2005) The schizophrenic experience: taken out of context? Schizophr Bull 31: 43-53. Hendry SH, Schwark HD, Jones EG, Yan J (1987) Numbers and proportions of GABA-immunoreactive neurons in different areas of monkey cerebral cortex. J Neurosci 7: 1503-19. Henquet C, Di Forti M, Morrison P, Kuepper R, Murray RM (2008) Gene-environment interplay between cannabis and psychosis. Schizophr Bull 34: 1111-21. Herdegen T, Leah JD (1998) Inducible and constitutive transcription factors in the mammalian nervous system: control of gene expression by Jun, Fos and Krox, and CREB/ATF proteins. Brain Res Brain Res Rev 28: 370-490. 221  Hermann B, Meador KJ, Gaillard WD, Cramer JA (2010) Cognition across the lifespan: antiepileptic drugs, epilepsy, or both? Epilepsy Behav 17: 1-5. Higgins GA, Ballard TM, Enderlin M, Haman M, Kemp JA (2005) Evidence for improved performance in cognitive tasks following selective NR2B NMDA receptor antagonist pre-treatment in the rat. Psychopharmacology (Berl) 179: 85-98. Hoff AL, Svetina C, Shields G, Stewart J, DeLisi LE (2005) Ten year longitudinal study of neuropsychological functioning subsequent to a first episode of schizophrenia. Schizophr Res 78: 27-34. Hoftman GD, Volk DW, Bazmi HH, Li S, Sampson AR, Lewis DA (2015) Altered cortical expression of GABA-related genes in schizophrenia: illness progression vs developmental disturbance. Schizophr Bull 41: 180-91. Hollrigel GS, Soltesz I (1997) Slow kinetics of miniature IPSCs during early postnatal development in granule cells of the dentate gyrus. J Neurosci 17: 5119-28. Holtzer R, Stern Y, Rakitin BC (2004) Age-related differences in executive control of working memory. Mem Cognit 32: 1333-45. Homayoun H, Moghaddam B (2007) NMDA receptor hypofunction produces opposite effects on prefrontal cortex interneurons and pyramidal neurons. J Neurosci 27: 11496-500. Howard MW, Rizzuto DS, Caplan JB, Madsen JR, Lisman J, Aschenbrenner-Scheibe R, Schulze-Bonhage A, Kahana MJ (2003) Gamma oscillations correlate with working memory load in humans. Cereb Cortex 13: 1369-74. Howes OD, Kapur S (2009) The dopamine hypothesis of schizophrenia: version III--the final common pathway. Schizophr Bull 35: 549-62. Hsu R, Schofield CM, Dela Cruz CG, Jones-Davis DM, Blelloch R, Ullian EM (2012) Loss of microRNAs in pyramidal neurons leads to specific changes in inhibitory synaptic transmission in the prefrontal cortex. Mol Cell Neurosci 50: 283-92. Hsu SE, Chen KC, Lee LT, Tsai HC, Lee IH, Chen PS, Yang YK (2015) Comparison of cognitive deficits among drug-naive patients with schizophrenia and major depressive disorder. J Affect Disord 175: 133-8. Huang-Pollock CL, Karalunas SL, Tam H, Moore AN (2012) Evaluating vigilance deficits in ADHD: a meta-analysis of CPT performance. J Abnorm Psychol 121: 360-71. Hunt MJ, Kopell NJ, Traub RD, Whittington MA (2017) Aberrant Network Activity in Schizophrenia. Trends Neurosci 40: 371-382. Hutton SB, Murphy FC, Joyce EM, Rogers RD, Cuthbert I, Barnes TR, McKenna PJ, Sahakian BJ, Robbins TW (2002) Decision making deficits in patients with first-episode and chronic schizophrenia. Schizophr Res 55: 249-57. Inoue M, Akaike N (1988) Blockade of gamma-aminobutyric acid-gated chloride current in frog sensory neurons by picrotoxin. Neurosci Res 5: 380-94. 222  Jacob J (2016) Cortical interneuron dysfunction in epilepsy associated with autism spectrum disorders. Epilepsia 57: 182-93. Jagannath V, Gerstenberg M, Correll CU, Walitza S, Grünblatt E (2018) A systematic meta-analysis of the association of Neuregulin 1 (NRG1), D-amino acid oxidase (DAO), and DAO activator (DAOA)/G72 polymorphisms with schizophrenia. J Neural Transm (Vienna) 125: 89-102. Jensen J, Willeit M, Zipursky RB, Savina I, Smith AJ, Menon M, Crawley AP, Kapur S (2008) The formation of abnormal associations in schizophrenia: neural and behavioral evidence. Neuropsychopharmacology 33: 473-9. Jentsch JD, Trantham-Davidson H, Jairl C, Tinsley M, Cannon TD, Lavin A (2009) Dysbindin modulates prefrontal cortical glutamatergic circuits and working memory function in mice. Neuropsychopharmacology 34: 2601-8. Ji Y, Yang F, Papaleo F, Wang HX, Gao WJ, Weinberger DR, Lu B (2009) Role of dysbindin in dopamine receptor trafficking and cortical GABA function. Proc Natl Acad Sci U S A 106: 19593-8. Johnston GA (2013) Advantages of an antagonist: bicuculline and other GABA antagonists. Br J Pharmacol 169: 328-36. Jones KR, Myers B, Herman JP (2011) Stimulation of the prelimbic cortex differentially modulates neuroendocrine responses to psychogenic and systemic stressors. Physiol Behav 104: 266-71. Jongen-Rêlo AL, Kaufmann S, Feldon J (2003) A differential involvement of the shell and core subterritories of the nucleus accumbens of rats in memory processes. Behav Neurosci 117: 150-68. Kabashima N, Shibuya I, Ibrahim N, Ueta Y, Yamashita H (1997) Inhibition of spontaneous EPSCs and IPSCs by presynaptic GABAB receptors on rat supraoptic magnocellular neurons. J Physiol 504 ( Pt 1): 113-26. Kaller CP, Loosli SV, Rahm B, Gössel A, Schieting S, Hornig T, Hennig J, Tebartz van Elst L, Weiller C, Katzev M (2014) Working memory in schizophrenia: behavioral and neural evidence for reduced susceptibility to item-specific proactive interference. Biol Psychiatry 76: 486-94. Kamigaki T, Dan Y (2017) Delay activity of specific prefrontal interneuron subtypes modulates memory-guided behavior. Nat Neurosci 20: 854-863. Kang JQ, Barnes G (2013) A common susceptibility factor of both autism and epilepsy: functional deficiency of GABA A receptors. J Autism Dev Disord 43: 68-79. Kang SS, MacDonald AW, Chafee MV, Im CH, Bernat EM, Davenport ND, Sponheim SR (2018) Abnormal cortical neural synchrony during working memory in schizophrenia. Clin Neurophysiol 129: 210-221. 223  Kann O, Papageorgiou IE, Draguhn A (2014) Highly energized inhibitory interneurons are a central element for information processing in cortical networks. J Cereb Blood Flow Metab 34: 1270-82. Karolewicz B, Maciag D, O'Dwyer G, Stockmeier CA, Feyissa AM, Rajkowska G (2010) Reduced level of glutamic acid decarboxylase-67 kDa in the prefrontal cortex in major depression. Int J Neuropsychopharmacol 13: 411-20. Kash SF, Johnson RS, Tecott LH, Noebels JL, Mayfield RD, Hanahan D, Baekkeskov S (1997) Epilepsy in mice deficient in the 65-kDa isoform of glutamic acid decarboxylase. Proc Natl Acad Sci U S A 94: 14060-5. Kash SF, Tecott LH, Hodge C, Baekkeskov S (1999) Increased anxiety and altered responses to anxiolytics in mice deficient in the 65-kDa isoform of glutamic acid decarboxylase. Proc Natl Acad Sci U S A 96: 1698-703. Kaupmann K, Huggel K, Heid J, Flor PJ, Bischoff S, Mickel SJ, McMaster G, Angst C, Bittiger H, Froestl W, Bettler B (1997) Expression cloning of GABA(B) receptors uncovers similarity to metabotropic glutamate receptors. Nature 386: 239-46. Keefe RS, Bilder RM, Davis SM, Harvey PD, Palmer BW, Gold JM, Meltzer HY, Green MF, Capuano G, Stroup TS, McEvoy JP, Swartz MS, Rosenheck RA, Perkins DO, Davis CE, Hsiao JK, Lieberman JA, Investigators C, Group NW (2007) Neurocognitive effects of antipsychotic medications in patients with chronic schizophrenia in the CATIE Trial. Arch Gen Psychiatry 64: 633-47. Kegeles LS, Mao X, Stanford AD, Girgis R, Ojeil N, Xu X, Gil R, Slifstein M, Abi-Dargham A, Lisanby SH, Shungu DC (2012) Elevated prefrontal cortex γ-aminobutyric acid and glutamate-glutamine levels in schizophrenia measured in vivo with proton magnetic resonance spectroscopy. Arch Gen Psychiatry 69: 449-59. Kemp A, Tischmeyer W, Manahan-Vaughan D (2013) Learning-facilitated long-term depression requires activation of the immediate early gene, c-fos, and is transcription dependent. Behav Brain Res 254: 83-91. Kesner RP, DiMattia BV, Crutcher KA (1987) Evidence for neocortical involvement in reference memory. Behav Neural Biol 47: 40-53. Kim D, Jeong H, Lee J, Ghim JW, Her ES, Lee SH, Jung MW (2016a) Distinct Roles of Parvalbumin- and Somatostatin-Expressing Interneurons in Working Memory. Neuron 92: 902-915. Kim H, Ährlund-Richter S, Wang X, Deisseroth K, Carlén M (2016b) Prefrontal Parvalbumin Neurons in Control of Attention. Cell 164: 208-218. Kim SY, Cohen BM, Chen X, Lukas SE, Shinn AK, Yuksel AC, Li T, Du F, Öngür D (2017) Redox Dysregulation in Schizophrenia Revealed by in vivo NAD+/NADH Measurement. Schizophr Bull 43: 197-204. Kimoto S, Bazmi HH, Lewis DA (2014) Lower expression of glutamic acid decarboxylase 67 in the prefrontal cortex in schizophrenia: contribution of altered regulation by Zif268. Am J Psychiatry 171: 969-78. 224  Kochunov P, Rowland LM, Fieremans E, Veraart J, Jahanshad N, Eskandar G, Du X, Muellerklein F, Savransky A, Shukla D, Sampath H, Thompson PM, Hong LE (2016) Diffusion-weighted imaging uncovers likely sources of processing-speed deficits in schizophrenia. Proc Natl Acad Sci U S A 113: 13504-13509. Koczat DL, Rogers SJ, Pennington BF, Ross RG (2002) Eye movement abnormality suggestive of a spatial working memory deficit is present in parents of autistic probands. J Autism Dev Disord 32: 513-8. Koetsier GC, Volkers AC, Tulen JH, Passchier J, van den Broek WW, Bruijn JA (2002) CPT performance in major depressive disorder before and after treatment with imipramine or fluvoxamine. J Psychiatr Res 36: 391-7. Krystal JH, D'Souza DC, Mathalon D, Perry E, Belger A, Hoffman R (2003) NMDA receptor antagonist effects, cortical glutamatergic function, and schizophrenia: toward a paradigm shift in medication development. Psychopharmacology (Berl) 169: 215-33. Krystal JH, Karper LP, Seibyl JP, Freeman GK, Delaney R, Bremner JD, Heninger GR, Bowers MB, Charney DS (1994) Subanesthetic effects of the noncompetitive NMDA antagonist, ketamine, in humans. Psychotomimetic, perceptual, cognitive, and neuroendocrine responses. Arch Gen Psychiatry 51: 199-214. Kubota Y, Kawaguchi Y (2000) Dependence of GABAergic synaptic areas on the interneuron type and target size. J Neurosci 20: 375-86. Kuipers SD, Bramham CR (2006) Brain-derived neurotrophic factor mechanisms and function in adult synaptic plasticity: new insights and implications for therapy. Curr Opin Drug Discov Devel 9: 580-6. Kuki T, Fujihara K, Miwa H, Tamamaki N, Yanagawa Y, Mushiake H (2015) Contribution of parvalbumin and somatostatin-expressing GABAergic neurons to slow oscillations and the balance in beta-gamma oscillations across cortical layers. Front Neural Circuits 9: 6. Kumar S, Black SJ, Hultman R, Szabo ST, DeMaio KD, Du J, Katz BM, Feng G, Covington HE, Dzirasa K (2013) Cortical control of affective networks. J Neurosci 33: 1116-29. Kuner R, Köhr G, Grünewald S, Eisenhardt G, Bach A, Kornau HC (1999) Role of heteromer formation in GABAB receptor function. Science 283: 74-7. Lahti AC, Koffel B, LaPorte D, Tamminga CA (1995) Subanesthetic doses of ketamine stimulate psychosis in schizophrenia. Neuropsychopharmacology 13: 9-19. Lamigeon C, Bellier JP, Sacchettoni S, Rujano M, Jacquemont B (2001) Enhanced neuronal protection from oxidative stress by coculture with glutamic acid decarboxylase-expressing astrocytes. J Neurochem 77: 598-606. Larimore J, Tornieri K, Ryder PV, Gokhale A, Zlatic SA, Craige B, Lee JD, Talbot K, Pare JF, Smith Y, Faundez V (2011) The schizophrenia susceptibility factor dysbindin and its associated complex sort cargoes from cell bodies to the synapse. Mol Biol Cell 22: 4854-67. 225  Lau CG, Murthy VN (2012) Activity-dependent regulation of inhibition via GAD67. J Neurosci 32: 8521-31. Lee AT, Vogt D, Rubenstein JL, Sohal VS (2014) A class of GABAergic neurons in the prefrontal cortex sends long-range projections to the nucleus accumbens and elicits acute avoidance behavior. J Neurosci 34: 11519-25. Lee FH, Zai CC, Cordes SP, Roder JC, Wong AH (2013) Abnormal interneuron development in disrupted-in-schizophrenia-1 L100P mutant mice. Mol Brain 6: 20. Lee J, Park S (2005) Working memory impairments in schizophrenia: a meta-analysis. J Abnorm Psychol 114: 599-611. Lee RS, Hermens DF, Porter MA, Redoblado-Hodge MA (2012) A meta-analysis of cognitive deficits in first-episode Major Depressive Disorder. J Affect Disord 140: 113-24. Lener MS, Niciu MJ, Ballard ED, Park M, Park LT, Nugent AC, Zarate CA (2017) Glutamate and Gamma-Aminobutyric Acid Systems in the Pathophysiology of Major Depression and Antidepressant Response to Ketamine. Biol Psychiatry 81: 886-897. Lerea LS, Butler LS, McNamara JO (1992) NMDA and non-NMDA receptor-mediated increase of c-fos mRNA in dentate gyrus neurons involves calcium influx via different routes. J Neurosci 12: 2973-81. Letzkus JJ, Wolff SB, Lüthi A (2015) Disinhibition, a Circuit Mechanism for Associative Learning and Memory. Neuron 88: 264-76. Levinson AJ, Fitzgerald PB, Favalli G, Blumberger DM, Daigle M, Daskalakis ZJ (2010) Evidence of cortical inhibitory deficits in major depressive disorder. Biol Psychiatry 67: 458-64. Lewin JS, Friedman L, Wu D, Miller DA, Thompson LA, Klein SK, Wise AL, Hedera P, Buckley P, Meltzer H, Friedland RP, Duerk JL (1996) Cortical localization of human sustained attention: detection with functional MR using a visual vigilance paradigm. J Comput Assist Tomogr 20: 695-701. Lewis DA, Curley AA, Glausier JR, Volk DW (2012) Cortical parvalbumin interneurons and cognitive dysfunction in schizophrenia. Trends Neurosci 35: 57-67. Lewis DA, Hashimoto T, Morris HM (2008) Cell and receptor type-specific alterations in markers of GABA neurotransmission in the prefrontal cortex of subjects with schizophrenia. Neurotox Res 14: 237-48. Li CS (2004) Do schizophrenia patients make more perseverative than non-perseverative errors on the Wisconsin Card Sorting Test? A meta-analytic study. Psychiatry Res 129: 179-90. Li P, Fan TT, Zhao RJ, Han Y, Shi L, Sun HQ, Chen SJ, Shi J, Lin X, Lu L (2017) Altered Brain Network Connectivity as a Potential Endophenotype of Schizophrenia. Sci Rep 7: 5483. Lieberman JA, Kane JM, Alvir J (1987) Provocative tests with psychostimulant drugs in schizophrenia. Psychopharmacology (Berl) 91: 415-33. 226  Lin H, Hsu FC, Baumann BH, Coulter DA, Anderson SA, Lynch DR (2014) Cortical parvalbumin GABAergic deficits with α7 nicotinic acetylcholine receptor deletion: implications for schizophrenia. Mol Cell Neurosci 61: 163-75. Liu SK, Chiu CH, Chang CJ, Hwang TJ, Hwu HG, Chen WJ (2002) Deficits in sustained attention in schizophrenia and affective disorders: stable versus state-dependent markers. Am J Psychiatry 159: 975-82. Lobo MK, Zaman S, Damez-Werno DM, Koo JW, Bagot RC, DiNieri JA, Nugent A, Finkel E, Chaudhury D, Chandra R, Riberio E, Rabkin J, Mouzon E, Cachope R, Cheer JF, Han MH, Dietz DM, Self DW, Hurd YL, Vialou V, Nestler EJ (2013) ΔFosB induction in striatal medium spiny neuron subtypes in response to chronic pharmacological, emotional, and optogenetic stimuli. J Neurosci 33: 18381-95. Lodge DJ (2011) The medial prefrontal and orbitofrontal cortices differentially regulate dopamine system function. Neuropsychopharmacology 36: 1227-36. López-Picón F, Snellman A, Shatillo O, Lehtiniemi P, Grönroos TJ, Marjamäki P, Trigg W, Jones PA, Solin O, Pitkänen A, Haaparanta-Solin M (2016) Ex Vivo Tracing of NMDA and GABA-A Receptors in Rat Brain After Traumatic Brain Injury Using 18F-GE-179 and 18F-GE-194 Autoradiography. J Nucl Med 57: 1442-7. Loughman A, Bowden SC, D'Souza W (2014) Cognitive functioning in idiopathic generalised epilepsies: a systematic review and meta-analysis. Neurosci Biobehav Rev 43: 20-34. Lucke-Wold BP, Nguyen L, Turner RC, Logsdon AF, Chen YW, Smith KE, Huber JD, Matsumoto R, Rosen CL, Tucker ES, Richter E (2015) Traumatic brain injury and epilepsy: Underlying mechanisms leading to seizure. Seizure 33: 13-23. Luikart BW, Parada LF (2006) Receptor tyrosine kinase B-mediated excitatory synaptogenesis. Prog Brain Res 157: 15-24. Luna B, Doll SK, Hegedus SJ, Minshew NJ, Sweeney JA (2007) Maturation of executive function in autism. Biol Psychiatry 61: 474-81. Luscher B, Shen Q, Sahir N (2011) The GABAergic deficit hypothesis of major depressive disorder. Mol Psychiatry 16: 383-406. Lustig C, Kozak R, Sarter M, Young JW, Robbins TW (2013) CNTRICS final animal model task selection: control of attention. Neurosci Biobehav Rev 37: 2099-110. Lyons-Warren A, Lillie R, Hershey T (2004) Short- and long-term spatial delayed response performance across the lifespan. Dev Neuropsychol 26: 661-78. MacAllister WS, Vasserman M, Rosenthal J, Sherman E (2014) Attention and executive functions in children with epilepsy: what, why, and what to do. Appl Neuropsychol Child 3: 215-25. Maddux JM, Holland PC (2011) Effects of dorsal or ventral medial prefrontal cortical lesions on five-choice serial reaction time performance in rats. Behav Brain Res 221: 63-74. 227  Magistretti PJ, Allaman I (2015) A cellular perspective on brain energy metabolism and functional imaging. Neuron 86: 883-901. Mair RG, Burk JA, Porter MC (1998) Lesions of the frontal cortex, hippocampus, and intralaminar thalamic nuclei have distinct effects on remembering in rats. Behav Neurosci 112: 772-92. Majer M, Ising M, Künzel H, Binder EB, Holsboer F, Modell S, Zihl J (2004) Impaired divided attention predicts delayed response and risk to relapse in subjects with depressive disorders. Psychol Med 34: 1453-63. Marder SR, Fenton W, Youens K (2004) Schizophrenia, IX: Cognition in schizophrenia--the MATRICS initiative. Am J Psychiatry 161: 25. Marotta A, Chiaie RD, Spagna A, Bernabei L, Sciarretta M, Roca J, Biondi M, Casagrande M (2015) Impaired conflict resolution and vigilance in euthymic bipolar disorder. Psychiatry Res 229: 490-6. Marsman A, Mandl RC, Klomp DW, Bohlken MM, Boer VO, Andreychenko A, Cahn W, Kahn RS, Luijten PR, Hulshoff Pol HE (2014) GABA and glutamate in schizophrenia: a 7 T ¹H-MRS study. Neuroimage Clin 6: 398-407. Marty S, Onténiente B (1997) The expression pattern of somatostatin and calretinin by postnatal hippocampal interneurons is regulated by activity-dependent and -independent determinants. Neuroscience 80: 79-88. Mathias JL, Wheaton P (2007) Changes in attention and information-processing speed following severe traumatic brain injury: a meta-analytic review. Neuropsychology 21: 212-23. Matsuo K, Glahn DC, Peluso MA, Hatch JP, Monkul ES, Najt P, Sanches M, Zamarripa F, Li J, Lancaster JL, Fox PT, Gao JH, Soares JC (2007) Prefrontal hyperactivation during working memory task in untreated individuals with major depressive disorder. Mol Psychiatry 12: 158-66. Mattay VS, Fera F, Tessitore A, Hariri AR, Berman KF, Das S, Meyer-Lindenberg A, Goldberg TE, Callicott JH, Weinberger DR (2006) Neurophysiological correlates of age-related changes in working memory capacity. Neurosci Lett 392: 32-7. Matzeu A, Zamora-Martinez ER, Martin-Fardon R (2014) The paraventricular nucleus of the thalamus is recruited by both natural rewards and drugs of abuse: recent evidence of a pivotal role for orexin/hypocretin signaling in this thalamic nucleus in drug-seeking behavior. Front Behav Neurosci 8: 117. McCleery A, Green MF, Hellemann GS, Baade LE, Gold JM, Keefe RS, Kern RS, Mesholam-Gately RI, Seidman LJ, Subotnik KL, Ventura J, Nuechterlein KH (2015) Latent structure of cognition in schizophrenia: a confirmatory factor analysis of the MATRICS Consensus Cognitive Battery (MCCB). Psychol Med 45: 2657-66. McGarry LM, Packer AM, Fino E, Nikolenko V, Sippy T, Yuste R (2010) Quantitative classification of somatostatin-positive neocortical interneurons identifies three interneuron subtypes. Front Neural Circuits 4: 12. 228  McGaughy J, Kaiser T, Sarter M (1996) Behavioral vigilance following infusions of 192 IgG-saporin into the basal forebrain: selectivity of the behavioral impairment and relation to cortical AChE-positive fiber density. Behav Neurosci 110: 247-65. McGaughy J, Sarter M (1995) Behavioral vigilance in rats: task validation and effects of age, amphetamine, and benzodiazepine receptor ligands. Psychopharmacology (Berl) 117: 340-57. McGaughy J, Sarter M (1998) Sustained attention performance in rats with intracortical infusions of 192 IgG-saporin-induced cortical cholinergic deafferentation: effects of physostigmine and FG 7142. Behav Neurosci 112: 1519-25. McNally JM, McCarley RW (2016) Gamma band oscillations: a key to understanding schizophrenia symptoms and neural circuit abnormalities. Curr Opin Psychiatry 29: 202-10. McQuail JA, Frazier CJ, Bizon JL (2015) Molecular aspects of age-related cognitive decline: the role of GABA signaling. Trends Mol Med 21: 450-60. Meechan DW, Tucker ES, Maynard TM, LaMantia AS (2012) Cxcr4 regulation of interneuron migration is disrupted in 22q11.2 deletion syndrome. Proc Natl Acad Sci U S A 109: 18601-6. Merali Z, Du L, Hrdina P, Palkovits M, Faludi G, Poulter MO, Anisman H (2004) Dysregulation in the suicide brain: mRNA expression of corticotropin-releasing hormone receptors and GABA(A) receptor subunits in frontal cortical brain region. J Neurosci 24: 1478-85. Meyer U, Feldon J (2010) Epidemiology-driven neurodevelopmental animal models of schizophrenia. Prog Neurobiol 90: 285-326. Meyer U, Nyffeler M, Yee BK, Knuesel I, Feldon J (2008) Adult brain and behavioral pathological markers of prenatal immune challenge during early/middle and late fetal development in mice. Brain Behav Immun 22: 469-86. Miner LA, Ostrander M, Sarter M (1997) Effects of ibotenic acid-induced loss of neurons in the medial prefrontal cortex of rats on behavioral vigilance: evidence for executive dysfunction. J Psychopharmacol 11: 169-78. Minzenberg MJ, Firl AJ, Yoon JH, Gomes GC, Reinking C, Carter CS (2010) Gamma oscillatory power is impaired during cognitive control independent of medication status in first-episode schizophrenia. Neuropsychopharmacology 35: 2590-9. Misgeld U, Bijak M, Jarolimek W (1995) A physiological role for GABAB receptors and the effects of baclofen in the mammalian central nervous system. Prog Neurobiol 46: 423-62. Missig G, Mokler EL, Robbins JO, Alexander AJ, McDougle CJ, Carlezon WA (2018) Perinatal Immune Activation Produces Persistent Sleep Alterations and Epileptiform Activity in Male Mice. Neuropsychopharmacology 43: 482-491. Moffoot AP, O'Carroll RE, Bennie J, Carroll S, Dick H, Ebmeier KP, Goodwin GM (1994) Diurnal variation of mood and neuropsychological function in major depression with melancholia. J Affect Disord 32: 257-69. 229  Morris HM, Hashimoto T, Lewis DA (2008) Alterations in somatostatin mRNA expression in the dorsolateral prefrontal cortex of subjects with schizophrenia or schizoaffective disorder. Cereb Cortex 18: 1575-87. Morrow BA, Elsworth JD, Inglis FM, Roth RH (1999) An antisense oligonucleotide reverses the footshock-induced expression of fos in the rat medial prefrontal cortex and the subsequent expression of conditioned fear-induced immobility. J Neurosci 19: 5666-73. Morshedi MM, Meredith GE (2007) Differential laminar effects of amphetamine on prefrontal parvalbumin interneurons. Neuroscience 149: 617-24. Murphy ER, Dalley JW, Robbins TW (2005) Local glutamate receptor antagonism in the rat prefrontal cortex disrupts response inhibition in a visuospatial attentional task. Psychopharmacology (Berl) 179: 99-107. Murray AJ, Woloszynowska-Fraser MU, Ansel-Bollepalli L, Cole KL, Foggetti A, Crouch B, Riedel G, Wulff P (2015) Parvalbumin-positive interneurons of the prefrontal cortex support working memory and cognitive flexibility. Sci Rep 5: 16778. Murray JD, Anticevic A, Gancsos M, Ichinose M, Corlett PR, Krystal JH, Wang XJ (2014) Linking microcircuit dysfunction to cognitive impairment: effects of disinhibition associated with schizophrenia in a cortical working memory model. Cereb Cortex 24: 859-72. Nelson CL, Burk JA, Bruno JP, Sarter M (2002) Effects of acute and repeated systemic administration of ketamine on prefrontal acetylcholine release and sustained attention performance in rats. Psychopharmacology (Berl) 161: 168-79. Neves-Pereira M, Cheung JK, Pasdar A, Zhang F, Breen G, Yates P, Sinclair M, Crombie C, Walker N, St Clair DM (2005) BDNF gene is a risk factor for schizophrenia in a Scottish population. Mol Psychiatry 10: 208-12. Nichols J, Perez R, Wu C, Adelson PD, Anderson T (2015) Traumatic brain injury induces rapid enhancement of cortical excitability in juvenile rats. CNS Neurosci Ther 21: 193-203. Nilsson SR, Celada P, Fejgin K, Thelin J, Nielsen J, Santana N, Heath CJ, Larsen PH, Nielsen V, Kent BA, Saksida LM, Stensbøl TB, Robbins TW, Bastlund JF, Bussey TJ, Artigas F, Didriksen M (2016) A mouse model of the 15q13.3 microdeletion syndrome shows prefrontal neurophysiological dysfunctions and attentional impairment. Psychopharmacology (Berl) 233: 2151-63. Noda Y, Zomorrodi R, Backhouse F, Cash RFH, Barr MS, Rajji TK, Chen R, Daskalakis ZJ, Blumberger DM (2017) Reduced Prefrontal Short-Latency Afferent Inhibition in Older Adults and Its Relation to Executive Function: A TMS-EEG Study. Front Aging Neurosci 9: 119. Nuechterlein KH, Barch DM, Gold JM, Goldberg TE, Green MF, Heaton RK (2004) Identification of separable cognitive factors in schizophrenia. Schizophr Res 72: 29-39. Nuechterlein KH, Green MF, Calkins ME, Greenwood TA, Gur RE, Gur RC, Lazzeroni LC, Light GA, Radant AD, Seidman LJ, Siever LJ, Silverman JM, Sprock J, Stone WS, Sugar CA, Swerdlow NR, Tsuang DW, Tsuang MT, Turetsky BI, Braff DL (2015) 230  Attention/vigilance in schizophrenia: performance results from a large multi-site study of the Consortium on the Genetics of Schizophrenia (COGS). Schizophr Res 163: 38-46. O'Donovan MC, Norton N, Williams H, Peirce T, Moskvina V, Nikolov I, Hamshere M, Carroll L, Georgieva L, Dwyer S, Holmans P, Marchini JL, Spencer CC, Howie B, Leung HT, Giegling I, Hartmann AM, Möller HJ, Morris DW, Shi Y, Feng G, Hoffmann P, Propping P, Vasilescu C, Maier W, Rietschel M, Zammit S, Schumacher J, Quinn EM, Schulze TG, Iwata N, Ikeda M, Darvasi A, Shifman S, He L, Duan J, Sanders AR, Levinson DF, Adolfsson R, Osby U, Terenius L, Jönsson EG, Cichon S, Nöthen MM, Gill M, Corvin AP, Rujescu D, Gejman PV, Kirov G, Craddock N, Williams NM, Owen MJ, Collaboration MGoS (2009) Analysis of 10 independent samples provides evidence for association between schizophrenia and a SNP flanking fibroblast growth factor receptor 2. Mol Psychiatry 14: 30-6. Obata K, Hirono M, Kume N, Kawaguchi Y, Itohara S, Yanagawa Y (2008) GABA and synaptic inhibition of mouse cerebellum lacking glutamate decarboxylase 67. Biochem Biophys Res Commun 370: 429-33. Oblak A, Gibbs TT, Blatt GJ (2009) Decreased GABAA receptors and benzodiazepine binding sites in the anterior cingulate cortex in autism. Autism Res 2: 205-19. Olsen RW, Sieghart W (2009) GABA A receptors: subtypes provide diversity of function and pharmacology. Neuropharmacology 56: 141-8. Olton DS, Papas BC (1979) Spatial memory and hippocampal function. Neuropsychologia 17: 669-82. Ongür D, Prescot AP, McCarthy J, Cohen BM, Renshaw PF (2010) Elevated gamma-aminobutyric acid levels in chronic schizophrenia. Biol Psychiatry 68: 667-70. Otto T, Wolf D, Walsh TJ (1997) Combined lesions of perirhinal and entorhinal cortex impair rats' performance in two versions of the spatially guided radial-arm maze. Neurobiol Learn Mem 68: 21-31. Ozonoff S, Cook I, Coon H, Dawson G, Joseph RM, Klin A, McMahon WM, Minshew N, Munson JA, Pennington BF, Rogers SJ, Spence MA, Tager-Flusberg H, Volkmar FR, Wrathall D (2004) Performance on Cambridge Neuropsychological Test Automated Battery subtests sensitive to frontal lobe function in people with autistic disorder: evidence from the Collaborative Programs of Excellence in Autism network. J Autism Dev Disord 34: 139-50. Ozonoff S, Jensen J (1999) Brief report: specific executive function profiles in three neurodevelopmental disorders. J Autism Dev Disord 29: 171-7. Packard MG, White NM (1990) Lesions of the caudate nucleus selectively impair "reference memory" acquisition in the radial maze. Behav Neural Biol 53: 39-50. Paelecke-Habermann Y, Pohl J, Leplow B (2005) Attention and executive functions in remitted major depression patients. J Affect Disord 89: 125-35. Paine TA, Cooke EK, Lowes DC (2015a) Effects of chronic inhibition of GABA synthesis on attention and impulse control. Pharmacol Biochem Behav 135: 97-104. 231  Paine TA, O'Hara A, Plaut B, Lowes DC (2015b) Effects of disrupting medial prefrontal cortex GABA transmission on decision-making in a rodent gambling task. Psychopharmacology (Berl) 232: 1755-65. Paine TA, Slipp LE, Carlezon WA (2011) Schizophrenia-like attentional deficits following blockade of prefrontal cortex GABAA receptors. Neuropsychopharmacology 36: 1703-13. Paine TA, Swedlow N, Swetschinski L (2017) Decreasing GABA function within the medial prefrontal cortex or basolateral amygdala decreases sociability. Behav Brain Res 317: 542-552. Paine TA, Tomasiewicz HC, Zhang K, Carlezon WA (2007) Sensitivity of the five-choice serial reaction time task to the effects of various psychotropic drugs in Sprague-Dawley rats. Biol Psychiatry 62: 687-93. Pantelis C, Barber FZ, Barnes TR, Nelson HE, Owen AM, Robbins TW (1999) Comparison of set-shifting ability in patients with chronic schizophrenia and frontal lobe damage. Schizophr Res 37: 251-70. Pantelis C, Barnes TR, Nelson HE, Tanner S, Weatherley L, Owen AM, Robbins TW (1997) Frontal-striatal cognitive deficits in patients with chronic schizophrenia. Brain 120 ( Pt 10): 1823-43. Papmeyer M, Sussmann JE, Hall J, McKirdy J, Peel A, Macdonald A, Lawrie SM, Whalley HC, McIntosh AM (2015) Neurocognition in individuals at high familial risk of mood disorders with or without subsequent onset of depression. Psychol Med 45: 3317-27. Parasuraman R, Warm JS, See JE (1998) Brain systems of vigilance. In: Suraman R (ed) The Attentive Brain. MIT Press, Cambridge, MA Pardo JV, Fox PT, Raichle ME (1991) Localization of a human system for sustained attention by positron emission tomography. Nature 349: 61-4. Park S, Holzman PS (1992) Schizophrenics show spatial working memory deficits. Arch Gen Psychiatry 49: 975-82. Parnaudeau S, O'Neill PK, Bolkan SS, Ward RD, Abbas AI, Roth BL, Balsam PD, Gordon JA, Kellendonk C (2013) Inhibition of mediodorsal thalamus disrupts thalamofrontal connectivity and cognition. Neuron 77: 1151-62. Parthoens J, Servaes S, Verhaeghe J, Stroobants S, Staelens S (2015) Prelimbic Cortical Injections of a GABA Agonist and Antagonist: In Vivo Quantification of the Effect in the Rat Brain Using [(18)F] FDG MicroPET. Mol Imaging Biol 17: 856-64. Pascualvaca DM, Fantie BD, Papageorgiou M, Mirsky AF (1998) Attentional capacities in children with autism: is there a general deficit in shifting focus? J Autism Dev Disord 28: 467-78. Pehrson AL, Bondi CO, Totah NK, Moghaddam B (2013) The influence of NMDA and GABA(A) receptors and glutamic acid decarboxylase (GAD) activity on attention. Psychopharmacology (Berl) 225: 31-9. 232  Pelosi L, Slade T, Blumhardt LD, Sharma VK (2000) Working memory dysfunction in major depression: an event-related potential study. Clin Neurophysiol 111: 1531-43. Peters A, Sethares C, Luebke JI (2008) Synapses are lost during aging in the primate prefrontal cortex. Neuroscience 152: 970-81. Petersen SE, Posner MI (2012) The attention system of the human brain: 20 years after. Annu Rev Neurosci 35: 73-89. Pezze M, McGarrity S, Mason R, Fone KC, Bast T (2014) Too little and too much: hypoactivation and disinhibition of medial prefrontal cortex cause attentional deficits. J Neurosci 34: 7931-46. Philpot BD, Lim JH, Brunjes PC (1997) Activity-dependent regulation of calcium-binding proteins in the developing rat olfactory bulb. J Comp Neurol 387: 12-26. Pi HJ, Hangya B, Kvitsiani D, Sanders JI, Huang ZJ, Kepecs A (2013) Cortical interneurons that specialize in disinhibitory control. Nature 503: 521-4. Piantadosi PT, Floresco SB (2014) Prefrontal cortical GABA transmission modulates discrimination and latent inhibition of conditioned fear: relevance for schizophrenia. Neuropsychopharmacology 39: 2473-84. Piantadosi PT, Khayambashi S, Schluter MG, Kutarna A, Floresco SB (2016) Perturbations in reward-related decision-making induced by reduced prefrontal cortical GABA transmission: Relevance for psychiatric disorders. Neuropharmacology 101: 279-90. Pirker S, Schwarzer C, Wieselthaler A, Sieghart W, Sperk G (2000) GABA(A) receptors: immunocytochemical distribution of 13 subunits in the adult rat brain. Neuroscience 101: 815-50. Pizzagalli DA (2014) Depression, stress, and anhedonia: toward a synthesis and integrated model. Annu Rev Clin Psychol 10: 393-423. Pothuizen HH, Zhang WN, Jongen-Rêlo AL, Feldon J, Yee BK (2004) Dissociation of function between the dorsal and the ventral hippocampus in spatial learning abilities of the rat: a within-subject, within-task comparison of reference and working spatial memory. Eur J Neurosci 19: 705-12. Pouzet B, Welzl H, Gubler MK, Broersen L, Veenman CL, Feldon J, Rawlins JN, Yee BK (1999) The effects of NMDA-induced retrohippocampal lesions on performance of four spatial memory tasks known to be sensitive to hippocampal damage in the rat. Eur J Neurosci 11: 123-40. Povysheva NV, Zaitsev AV, Gonzalez-Burgos G, Lewis DA (2013) Electrophysiological heterogeneity of fast-spiking interneurons: chandelier versus basket cells. PLoS One 8: e70553. Prasad JA, Chudasama Y (2013) Viral tracing identifies parallel disynaptic pathways to the hippocampus. J Neurosci 33: 8494-503. 233  Price RB, Shungu DC, Mao X, Nestadt P, Kelly C, Collins KA, Murrough JW, Charney DS, Mathew SJ (2009) Amino acid neurotransmitters assessed by proton magnetic resonance spectroscopy: relationship to treatment resistance in major depressive disorder. Biol Psychiatry 65: 792-800. Puig O, Penadés R, Gastó C, Catalán R, Torres A, Salamero M (2008) Verbal memory, negative symptomatology and prediction of psychosocial functioning in schizophrenia. Psychiatry Res 158: 11-7. Quarta D, Naylor CG, Morri