UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Investigating the neurobiology regulating cognitive effort allocation using a rodent model of cost/benefit… Silveira, Mason Manuel 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_november_silveira_mason.pdf [ 15.56MB ]
Metadata
JSON: 24-1.0372368.json
JSON-LD: 24-1.0372368-ld.json
RDF/XML (Pretty): 24-1.0372368-rdf.xml
RDF/JSON: 24-1.0372368-rdf.json
Turtle: 24-1.0372368-turtle.txt
N-Triples: 24-1.0372368-rdf-ntriples.txt
Original Record: 24-1.0372368-source.json
Full Text
24-1.0372368-fulltext.txt
Citation
24-1.0372368.ris

Full Text

 Investigating the Neurobiology Regulating Cognitive Effort Allocation Using a Rodent Model of Cost/Benefit Decision Making   by     Mason Manuel Silveira    M.A., The University of British Columbia, 2014    A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF  THE REQUIREMENTS FOR THE DEGREE OF     Doctor of Philosophy   in   THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES  (Psychology)     THE UNIVERSITY OF BRITISH COLUMBIA   (Vancouver)     OCTOBER 2018     © Mason Manuel Silveira, 2018        ii The following individuals certify that they have read, and recommend to the Faculty of Graduate and Postdoctoral Studies for acceptance, the dissertation entitled: Investigating the Neurobiology Regulating Cognitive Effort Allocation Using a Rodent Model of Cost/Benefit Decision Making   submitted by     Mason Manuel Silveira In partial fulfillment of the requirements for the degree of    Doctor of Philosophy   in The Faculty of Graduate and Postdoctoral Studies  Examining Committee: Dr. Catharine Winstanley Supervisor  Dr. Stanley Floresco Supervisory Committee Member  Dr. Rebecca Todd Supervisory Committee Member Dr. Catharine Rankin University Examiner Dr. Lynn Raymond University Examiner  Additional Supervisory Committee Members:  Supervisory Committee Member  Supervisory Committee Member      iii Abstract   Choosing adaptively among candidate actions requires a cost-benefit analysis, in which potential rewards are considered against the costs required to obtain them. One cost frequently encountered by humans is the cost of cognitive effort, in which executive processes spanning attention, working memory, reasoning, and the like are taxed. This form of effort is in contrast to the physical effort costs that have generally been the focus of the cost/benefit decision-making literature. This thesis reviews the currently available literature investigating the brain regions and neurotransmitter systems guiding these respective forms of decision making, and then carries out a set of experiments to further characterize the neurobiology guiding cognitive effort allocation. These experiments utilize an animal model of decision making known as the rodent Cognitive Effort Task (rCET), in which subjects decide whether to exert more attention in pursuit of larger rewards, or to obtain smaller reward for comparatively less attentional demand. In experiment 1, I use chemogenetics to downregulate cholinergic neurons of the basal forebrain as rats perform the rCET, to determine whether this neuronal population is responsible for the previously ascribed role of acetylcholine in regulating decision making with cognitive effort costs. Experiments 2 and 3 use standard inactivation techniques to investigate striatal and orbitofrontal cortex contributions to this form of decision making, and Experiment 4 uses a disconnection procedure to assess whether BLA- ACC signaling regulates     iv willingness to apply cognitive effort. Collectively, the current findings complement existing work in the domain of physical effort allocation, and support the notion that these two forms of effort-based decision making are mediated by distinct, albeit overlapping circuitries. While this work fundamentally contributes to an understanding of how organisms navigate their environment, it also has practical utility. Indeed, work with the rCET and related cognitive effort paradigms may help identify behavioural or pharmacological therapies that can boost cognitive willingness. Alternatively, information gained may shed light on the aberrant processes underlying a blunted desire to achieve lucrative outcomes, as observed in disorders like depression, schizophrenia, and Parkinson’s Disease.              v Lay Summary   As we interact with the world, we often face situations where we can decide to invest cognitive effort for rewarding outcomes. Take, for example, the student who strives to do well in school in hopes that it will lead to lucrative job prospects. In its purest form, cognitive effort allocation is a cost-benefit analysis, and it is valuable to know the neurotransmitters and brains regions guiding this process. This thesis uses a rodent model of cognitive effort allocation to experimentally assess how different neurotransmitters and brain areas regulate decision making with cognitive effort costs. The findings from this thesis identify a number of brain structures and neurotransmitters that interact to bias organisms to put in mental work for outcomes they want. Importantly, this information may shed light on how these processes are disrupted in disorders characterized by a blunted desire to achieve lucrative outcomes, such as depression, schizophrenia, and Parkinson’s Disease.            vi Preface   The candidate, Dr. Adams, and Dr. Winstanley designed Experiment 1 (Chapter 3). Sukhbir Kaur bred the rats and I assisted with genotyping. I conducted behavioural testing. Dr. Adams and I performed the viral transfection surgeries. I prepared and delivered the pharmacological challeneges. Emily Allin assisted with pharmacological challenges. I independently analyzed the data, interpreted the results, and wrote the chapter. Dr. Winstanley reviewed and revised the chapter prior to inclusion in the thesis. Experiment 2 (Chapter 4) has been previously published in manuscript form: Silveira, M.M., Tremblay, M., & Winstanley, C.A. 2018. Dissociable contributions of dorsal and ventral striatal regions on a rodent cost/benefit decision-making task requiring cognitive effort. Neuropharmacology, 137, 322-331. The candidate and Dr. Winstanley were responsible for study concept and design. I collected the behavioural data, performed cannulation surgeries, and conducted the inactivations. Dr. Tremblay assisted with surgeries and inactivations. I performed data analyses and drafted the manuscript. Dr. Winstanley reviewed and revised the manuscript prior to manuscript submission.  The candidate and Dr. Winstanley designed Experiment 3 (Chapter 5). Sukhbir Kaur bred the rats and I assisted with genotyping. Sebastian Wittenkindt conducted behavioural testing and assisted with inactivations. I performed the cannulations and the inactivations. Sophie Ebsary sectioned brains for     vii histological analysis. I independently analyzed the data, interpreted the results, and wrote the chapter. Dr. Winstanley reviewed and revised the chapter prior to inclusion in the thesis.   The candidate and Dr. Winstanley designed Experiment 4 (Chapter 6). Leili Mortazavi conducted behavioural testing and assisted with inactivations. I performed the cannulations and the inactivations. Leili Mortazavi sectioned brains for histological analysis. I independently analyzed the data, interpreted the results, and wrote the chapter. Dr. Winstanley reviewed and revised the chapter prior to inclusion in the thesis. The candidate and Dr. Winstanley designed Experiment 5 (Appendix 1). Sukhbir Kaur bred the rats and I assisted with genotyping. Sebastian Wittenkindt and Leili Mortazavi conducted behavioural testing. I prepared and delivered the drugs. Leili Mortazavi assisted with pharmacological challenges. I consulted Dr. Wendy Adams for advice on preparing the drugs. I independently analyzed the data, interpreted the results, and wrote the chapter. Dr. Winstanley reviewed and revised the chapter prior to inclusion in the thesis. Figures are reproduced and modified from Paxinos & Watson, 1998, The Rat Brain in Stereotaxic Coordinates: Academic Press. They have been approved by Elsevier for reproduction in this thesis. All animal testing was performed in accordance with the Canadian Council on Animal Care (CCAC) and received ethical approval by the University of British Columbia Animal Care Committee, certificate numbers A13-0011 (behavioural protocol) and A15-0011 (breeding protocol). Adeno-associated virus delivery was     viii performed in accordance with approved standard operating procedure ACC WINSTANLEY CAW116. All procedures are standard for working with rAAVs classified as Risk Group 1, and have been approved by the UBC Biosafety Committee (protocol B15-0027).            ix Table of Contents  Abstract ............................................................................................................... iii Lay Summary ....................................................................................................... v Preface ................................................................................................................. vi Table of Contents ................................................................................................ ix List of Tables ....................................................................................................... xi List of Figures ................................................................................................... xiii List of Abbreviations ........................................................................................ xvi Glossary .......................................................................................................... xviii Acknowledgements .......................................................................................... xxi  Chapter 1: General Introduction ......................................................................... 1 1.1 Overview of the Behavioural Paradigms Used to Investigate Effort-Based Decision Making in Humans ................................................................ 3 1.2 Effort-Based Decision Making in Humans: Summary of Behavioural Findings, Relationship to Traits and Disease States ...................................... 9 1.3 Neurotransmitter Systems Implicated in Effort Allocation from Human Studies .......................................................................................................... 12 1.4 Human Imaging Studies on Effort Allocation ........................................ 14 1.5 Limitations of the Current Human Research on Effort Allocation ......... 17 1.6 Animal Models of Effort-Based Cost/Benefit Decision Making ............. 18 1.7 Comparing Effort Allocation Tasks ....................................................... 28 1.8 Neurotransmitter Systems Implicated in Effort-Based Decision Making ..   .............................................................................................................. 30 1.9 Individual Differences and Physiological Factors Mediating Effort Allocation ...................................................................................................... 52 1.10 Functional Neural Systems Mediating Effort-Based Decision Making 56 1.11 Comparison of the Neurotransmitter and Neural Systems Guiding Physical and Cognitive Effort Allocation and Rationale for the Current Experiments .................................................................................................. 68 Chapter 2: General Methods ............................................................................. 74 2.1 Subjects ................................................................................................ 74 2.2 Behavioural Apparatus ......................................................................... 74 2.3 The Rat Cognitive Effort Task (rCET) Training and Testing ................. 75 2.4 Behavioural Measures .......................................................................... 80 2.5 Data Analysis ........................................................................................ 81 2.6 Table ..................................................................................................... 83 2.7 Figures .................................................................................................. 85     x Chapter 3: Chemogenetic Silencing of Basal Forebrain Neurons Does Not Affect Cognitive Effort Allocation in a Rodent Model of Decision Making with Cognitive Effort Costs ............................................................................... 88 3.1 Introduction ........................................................................................... 88 3.2 Additional methods ............................................................................... 91 3.3 Results .................................................................................................. 94 3.4 Discussion .......................................................................................... 101 3.5 Tables ................................................................................................. 111 3.6 Figures ................................................................................................ 116 Chapter 4: Dissociable Contributions of Dorsal and Ventral Striatal Regions on a Rodent Cost/Benefit Decision-Making Task Requiring Cognitive Effort ........................................................................................................................... 122 4.1 Introduction ......................................................................................... 122 4.2 Additional Methods ............................................................................. 127 4.3 Results ................................................................................................ 130 4.4 Discussion .......................................................................................... 135 4.5 Tables ................................................................................................. 146 4.6 Figures ................................................................................................ 148 Chapter 5: Investigating Orbitofrontal Cortex Contributions to Decision Making Involving Cognitive Effort Costs ....................................................... 151 5.1 Introduction ......................................................................................... 151 5.2 Additional Methods ............................................................................. 155 5.3 Results ................................................................................................ 159 5.4 Discussion .......................................................................................... 163 5.5 Table ................................................................................................... 173 5.6 Figures ................................................................................................ 174 Chapter 6: Investigating Amygdala-Prefrontal Cortical Circuit Involvement in the Allocation of Cognitive Effort ............................................................... 177 6.1 Introduction ......................................................................................... 177 6.2 Additional Methods ............................................................................. 181 6.3 Results ................................................................................................ 187 6.4 Discussion .......................................................................................... 196 6.5 Tables ................................................................................................. 205 6.6 Figures ................................................................................................ 208 Chapter 7: General Discussion....................................................................... 212 References ....................................................................................................... 231 Appendix 1: Investigating Serotonergic Contributions to Cognitive Effort Allocation Using a Rodent Model of Cost/Benefit Decision Making ........... 256         xi List of Tables   Table 2.1. rCET training protocol ......................................................................... 84 Table 3.1. Other behavioural measures following CNO administered ten weeks post viral transfection surgery ..................................................................... 111 Table 3.2. Other behavioural measures following CNO administered fifteen weeks post viral transfection surgery ......................................................... 112 Table 3.3. Other behavioural measures following 10 mg/kg CNO ..................... 113 Table 3.4. Other behavioural measures following systemic scopolamine administration ............................................................................................. 114 Table 3.5. Other behavioural measures following systemic donepezil administration ............................................................................................. 115 Table 4.1. Other rCET behavioural measures following DMS inactivation ........ 146 Table 4.2. Other rCET behavioural measures following nucleus accumbens core inactivation .................................................................................................. 146 Table 4.3. Summary of behavioural effects observed following dorsomedial striatum and nucleus accumbens core inactivation .................................... 147 Table 5.1. Other rCET behavioural measures following OFC inactivation ........ 173 Table 6.1. Other rCET behavioural measures following contralateral ACC-BLA inactivation .................................................................................................. 205 Table 6.2. Other rCET behavioural measures following ipsilateral ACC-BLA inactivation .................................................................................................. 206     xii Table 6.3. Other rCET behavioural measures following unilateral BLA inactivation .................................................................................................................... 207 Table 7.1. Other behavioural measures following systemic Ro-60-0175 administration ............................................................................................. 268 Table 7.2. Other behavioural measures following systemic M100 907 administration ............................................................................................. 269 Table 7.3. Other behavioural measures following systemic SB 242, 084 administration ............................................................................................. 270      xiii List of Figures   Figure 2-1. The standard 5-hole operant chamber used for the rCET ................. 85 Figure 2-2. The standard 5-hole operant chamber used for the rCET ................. 86 Figure 2-3. Schematic diagram showing the trial structure of the rCET .............. 87 Figure 3-1. Verification of HM4Di expression in cholinergic neurons of the basal forebrain...................................................................................................... 116 Figure 3-2. Effects of CNO administration on choice of HR trials (%), accuracy (%), and premature responses (%) ten weeks post viral transfection surgeries ..................................................................................................... 117 Figure 3-3. Effects of CNO administration on choice of HR trials (%), accuracy (%), and premature responses (%) 15 weeks post viral transfection surgeries .................................................................................................................... 118  Figure 3-4. Effects of high dose CNO administration on choice of HR trials (%), accuracy (%), and premature responses (%) ............................................. 119 Figure 3-5. Effects of scopolamine administration on (a) choice of HR trials (%), (b) accuracy (%), (c) premature responses (%), and (d) response omissions (%) .............................................................................................................. 120 Figure 3-6. Effects of donepezil administration on (a) choice of HR trials (%), (b) accuracy (%), and (c) premature responses (%) ........................................ 121 Figure 4-1. Histological analysis of cannulae implantation ................................ 148 Figure 4-2. Effects of DMS Inactivations on (a) choice of HR trials (%), (b)     xiv accuracy, (c) premature responses (%), and response omissions (%) ...... 149 Figure 4-3. Effect of nucleus accumbens core (NAcC) inactivations on (a) trials initiated, (b) response omissions (%), and (c) choice of HR trials (%). ...... 150 Figure 5-1. Histological analysis of cannulae implantation ................................ 174 Figure 5-2. Effects of OFC inactivations on (a) choice of HR trials (%), (b) accuracy (%), (c) premature responses (%), (d) choice omissions, and (e) response omissions (%) ............................................................................. 175 Figure 6-1. Histological analysis of cannulae implantation ................................ 208 Figure 6-2. Effects of contralateral ACC-BLA inactivation on (a) choice of HR trials, (b) accuracy (%), (c) premature responses, and (d) choice latencies .................................................................................................................... 209 Figure 6-3. Effects of ipsilateral ACC-BLA inactivation on (a) choice of HR trials (%), (b) accuracy (%), (c) premature responses (%), and (d) choice latencies .................................................................................................................... 210 Figure 6-4. Effects of (a) unilateral BLA, (b) unilateral IPSI-ACC, and (c) unilateral contra-ACC inactivation on choice of HR trials (%) .................... 211 Figure 7-1. Effects of the 5-HT2C receptor agonist Ro-60-0175 on (a) choice of HR trials (%), (b) accuracy, (c) premature responses (%), and (d) response omissions (%) ............................................................................................. 271 Figure 7-2. Effects of the 5-HT2A receptor antagonist M100 907 on (a) choice of HR trials (%), (b) accuracy (%), (c) premature responses (%), and (d) response omissions (%) ............................................................................. 272 Figure 7-3. Effects of the 5-HT2C receptor antagonist SB 242, 084 on (a) choice     xv of HR trials (%), (b) accuracy, (c) premature responses (%), and (d) response omissions (%) ............................................................................. 273      xvi List of Abbreviations   ACC – anterior cingulate cortex Ach – acetylcholine ANOVA- analysis of variance BLA- basolateral amygdala BOLD- blood-oxygen-level-dependent  CRF – corticotrophin releasing factor  DA – dopamine DLS – dorsolateral striatum  DMS – dorsomedial striatum  EEfRT- Effort Expenditure for Rewards Task FR- fixed ratio  HR – high cost/high reward  IL- infralimbic LR- low cost/low reward kg – kilogram  ITI- inter-trial interval mg  - milligram  mPFC- medial prefrontal cortex NA – noradrenaline NAcc- nucleus accumbens     xvii OFC- orbitofrontal cortex PFC- prefrontal cortex PL – prelimbic PR- progressive ratio rCET- rodent Cognitive Effort Task  SEM- standard error of the mean SV- subjective value  VMAT - vesicular monoamine transporter-2 VS- ventral striatum VTA- ventral tegmental area     xviii Glossary   Dopaminergic Drugs 7-OH-DPAT – D3 agonist A77 636 - D1 agonist Buproprion (Wellbutrin) - dopamine transporter (DAT) reuptake inhibitor Cis-flupenthixol - nonspecific DA antagonist Eticlopride - D2 antagonist GBR12909 - dopamine transporter (DAT) reuptake inhibitor Haloperidol – D2 antagonist  PRX-14040 - dopamine transporter (DAT) reuptake inhibitor U-99194 – D3 antagonist SCH23390 - D1 antagonist SCH 39166 (ecicopam) - D1 antagonist SKF 38393 – D1 agonist  SKF 81297 – D1 agonist Tetrabenazine - vesicular monoamine transporter-2 Tolcapone - catchecol-O-methyltransferase (COMT) inhibitor  Adrenergic Drugs Atomoxetine - norepinephrine transporter (NET) reuptake inhibitor Desipramine - NET reuptake inhibitor     xix Yohimbine - a2-adrenergic antagonist   Cholinergic Drugs Mecamylamine – nonspecific nicotine antagonist  Nicotine – nonspecific nicotine receptor agonist  Oxotremorine – muscarinic agonist  Pilocarpine - muscarinic agonist Scopolamine - muscarinic antagonist  Cannabinoid Drugs AM 251 - CB1 receptor inverse agonist/antagonist  AM 4113 - neutral CB1 antagonist AM 630 - CB2 antagonist arachidonyl-2-chloroethylamide (ACEA) - synthetic CB1 receptor agonist Rimonabant - CB1 receptor inverse agonist/antagonist THC - CB1 receptor agonist URB 597 – fatty acid amide hydrolase inhibitor WIN 55, 212-2 - CB1 receptor agonist  Adenosine Drugs Caffeine – nonspecific adenosine antagonist  DPCPX - A1 adenosine antagonist      xx GS21680 - A2 agonist  KW6002 - A2 adenosine antagonist MSX-3 - A2 adenosine antagonist  Serotonergic Drugs Fluoxetine - serotonin transporter (SERT) reuptake inhibitor M100 907 – 5-HT2A antagonist  Ro-60-0175 - 5-HT2C agonist  SB 242, 084 – 5-HT2C antagonist  Other Drugs alpha-helical CRF - corticotrophin-releasing factor antagonist DRN - ERβ agonist Estradiol – nonspecific ER agonist  Ketamine – NMDA antagonist  PPT - estrogen receptor (ER)-α agonist      xxi Acknowledgements   First and foremost I would like to thank Dr. Catharine Winstanley for her valuable guidance over the course of my PhD. She continues to be an inspiring mentor as I continue on in my research career.  I am indebted to the members of the Winstanley lab whose assistance was integral to the completion of this project, and who were valued friends outside the lab.  I would like to thank Dr. Stan Floresco and Dr. Rebecca Todd for the guidance they have provided as my committee members.  I would also like to thank the UBC Psychology Department and the Natural Science and Engineering Research Council of Canada for the financial support they have provided over the course of my PhD.  Lastly, I want to thank all the animal care technicians who helped take care of my animals, and I want to acknowledge all the rodents that were involved in the experiments making up this thesis.                1 Chapter 1: General Introduction     Choosing adaptively among candidate actions requires a cost-benefit analysis, in which potential rewards are considered against the costs required to obtain them (Kahneman and Tversky, 1979; Winstanley and Floresco, 2016). Many decisions faced by humans involve costs resulting from the effects of action on the environment, including delays to reward receipt (Kable and Glimcher, 2007), risk of loss (Bechara et al, 1994; Kuhnen and Knutson, 2005), and social exclusion (Ruff and Fehr, 2014). However, one cost- specifically the cost of effort- is uniquely intrinsic to action itself (Kool et al, 2010a). It is well established that all things being equal, an organism will gradually learn to prefer options low in physical effort producing reinforcement (Hull, 1943). This Law of Least Effort has received extensive support, and for many years a number of researchers have studied the mechanisms guiding physical effort allocation (Salamone et al, 2009a; Walton et al, 2006). Only until recently, however, has a similar Law of Least Mental Effort been experimentally verified (Kool et al, 2010a). Work in this new area of research has demonstrated that humans generally avoid cognitively demanding tasks in domains spanning working memory, cognitive flexibility, response inhibition, and attention (Kool et al, 2010; Westbrook et al, 2013, Dixon and Christoff, 2012; Reddy et al, 2015, Chong et al, 2017). Indeed, the subjective value of rewarding outcomes is discounted by difficult cognitive challenge, and cognitive allocation towards harder tasks can be     2 biased by incentives, suggesting that like physical effort, organisms may decide between options in ways which simultaneously maximize gains while minimizing effort costs (Chong et al, 2017; Dixon and Christoff, 2012; Massar et al, 2015).   Thus, while there is evidence that both forms of effort appear to factor into value-based decision making, it is generally unknown whether overlapping or distinct brain processes mediate cognitive and physical effort allocation. To this end, a number of human paradigms and preclinical models have been developed to assess the ways in which organisms weigh potential rewards against the inherent effort costs required to obtain them. While the human literature has identified key neural substrates guiding effort allocation, preclinical models have also proven invaluable by allowing behavioural neuroscientists to address how different neurotransmitters and brains region causally regulate valuations of effort. In the sections that follow, I review the effort-based decision-making tasks developed for use in humans, and briefly summarize the main neurotransmitter and brain systems implicated in effort allocation on the basis of imaging and systemic pharmacology studies. This discussion then segues into an exhaustive review of: 1) animal models of effort allocation, 2) the results of systemic pharmacology completed using these tasks, as well as a 3) survey of the regions and circuits implicated in effort-based decision making on the basis of lesions or temporary inactivations. There are many findings for which data is available in the physical domain, but the equivalent work has not been modeled with cognitive effort, and so a thorough comparison of these processes is limited. I then propose a set of experiments in which I use a rodent model of cognitive     3 effort allocation to answer these outstanding questions. My goal is that the knowledge gained from these experiments will allow for a point-to-point comparison of the neural systems guiding effort in the physical and cognitive realm. The outline of these experiments and their objectives can be found in section 1.11.     1.1 Overview of the Behavioural Paradigms Used to Investigate Effort-Based Decision Making in Humans  1.1.1 Physical Effort Tasks  Physical effort-based paradigms generally fall into two categories: those that manipulate effort costs via repeated button pressing, and those that operationalize effort through forceful grip. A popular button pressing paradigm is the Effort Expenditure for Rewards Task (EEfRT) developed by Treadway and Colleagues (Treadway et al, 2009). In this task, subjects make repeated decisions between two levels of task difficulty. Successful completion of “hard-task” trials requires the subject to make 100 button presses using the nondominant little finger within 21 seconds, while successful completion of “easy-task” trials requires the subject to make 30 button presses, using the dominant index finger within 7 seconds (Treadway et al, 2009). Subjects can earn $1 for completing easy trials, but can earn higher amounts upon successful completion of hard trials, within a range of $1.24 – $4.30. Additionally, trials differ     4 in their probability of reward receipt, ranging from low (12.5%), medium (50%), and high (88%), allowing researchers to assess how effort-based decision making is modulated by reward magnitude, probability of reward receipt and expected value (Treadway et al, 2009). Healthy subjects normally increase effortful choice as probability of reward receipt increases (Treadway et al, 2009). A variant of the task has also been developed which titrates effort demands for each individual, which has proven useful when assessing patients suffering from motor impairments (Fervaha et al, 2013; Reddy et al, 2015). Similarly, in the Balloon Effort Task, subjects decide between two response alternatives in attempts to inflate a balloon on a computer screen until it bursts. For easy trials, subjects must alternately press game controller buttons 10 times to receive $1 hypothetical reward, whereas hard trials require 100 alternating buttons presses and are associated with 5 different levels of reward varying from $3 to $7 (Gold et al, 2013). In this paradigm, choice of the more effortful option increases as the associated reward levels increase. Other button pressing paradigms have been developed, but are generally similar to the tasks described above.   Forceful grip has similarly been used to vary effort costs in decision-making tasks. The individual parameters vary from study to study, but generally require subjects to decide between squeezing a hand-grip device with varying levels of intensity (Chong et al, 2017; Kurniawan et al, 2010, 2013; Prevost et al, 2010; Reddy et al, 2015; Schmidt et al, 2012). Easy trials require simply holding a hand-grip, or squeezing with less than 50% of maximal force for small monetary reward. In contrast, the high effort response alternative requires 80-    5 90% maximal force exertion, but is associated with larger payout. Force parameters are titrated based on an assessment of maximal force prior to testing, and feedback on the levels of force achieved in a single trial is provided by an image on-screen, such as a thermometer with rising mercury levels (Kurniawan et al, 2010, 2013; Prevost et al, 2010) or a ladder with rungs (Schmidt et al, 2012). The tasks have generally used fictive currency as reward, but at least one study has used erotic stimuli (Prevost et al, 2010), which is interesting given evidence that primary and secondary reward representations recruit distinct neural substrates (Sescousse et al, 2010).  Recently, a novel effortful grip task has been developed by Husain and colleagues, in which the decision is not between concurrent easy or hard grip alternatives, but whether subjects are willing to put in the effort for a specific stake (Chong et al, 2015). During the experiment, participants are shown cartoons of apple trees, and are told to collect as many apples as possible based on the combined stake and effort presented. Potential rewards are indicated by the number of apples on the tree (1, 3, 6, 9, 12, 15), while the associated effort is specified by the height of a yellow bar on the tree trunk, ranging over six levels as a function of participants’ maximal force (60%, 70%, 80%, 90%, 100%, 110%) (Chong et al, 2015). An analysis of indifference points suggests that willingness to accept offers increases as stake level increases (Chong et al, 2015).   1.1.2 Cognitive Effort Tasks  Like physical effort, cognitive effort can take on many forms. It is important     6 to keep this in mind when discussing the cognitive effort literature, as the term may be too broad to reflect only the processes taxed in a single paradigm. As will be described below, a number of cognitive effort allocation paradigms have been developed, which probe processes ranging from working memory, attention, cognitive flexibility, and response inhibition. Kool and colleagues developed the first test of cognitive effort allocation, known as the Demand Selection Task (Kool et al, 2010b, 2013). In this paradigm, subjects choose between two options (card decks or cues in computerized versions) that differ in difficulty. Numbers ranging from 1 to 9 are presented in one of two colours that specify the mental activity required to successfully complete the trial. Depending on the color of the numeral (blue or yellow, for example), the participant must make a parity (odd/even) or magnitude (less/greater than five) judgment. The easy option requires little switching between colours and their corresponding rules (either no switching or 90% congruence between trials), while the hard option requires frequent rule switching (Kool et al, 2010a). Subjects readily avoid the more challenging option, suggesting that these trials are indeed more effortful. While the original task did not include reward incentives for choosing between options, subsequent studies have baited the hard option with larger fictive monetary rewards, and this works to bias choice towards the more demanding option (Reddy et al, 2015, 2017).  Additionally, while the original version of the task did not label choices for difficulty, subsequent iterations have been developed for patient populations that explicitly label options as “easy” and “hard” (Gold et al, 2014; Kool et al, 2013; Reddy et al, 2015). Most recently, Wardle and Colleagues have adapted the     7 physical EEfRT task to model cognitive effort costs using the Demand Selection Task described above. Like the EEfRT task, participants choose to exert cognitive effort across trials varying in reward magnitude and probability of reward receipt. However, rather than manipulating effort via button pressing, the easy option requires no switching between colours and their corresponding rules (4 judgments in 7 s), while the hard option requires frequent rule switching (19 judgments in 21s) (Lopez-Gamundi and Wardle, 2018).  Another task that assesses effort allocation in the domain of working memory is the Cognitive Effort Discounting Paradigm (COG-ED) (Westbrook et al, 2013). In this paradigm participants choose whether to perform a low-effort N-back task for a small monetary reward, or a high-effort N-back task for a larger reward. Multiple choices are made, and the amount offered for the low-effort task is titrated until the offers are subjectively equivalent (Westbrook et al, 2013). The objective load of the high-effort task is varied parametrically, akin to the estimation of discounting across a range of delays in delay discounting paradigms (Kable and Glimcher, 2007). A similar task developed by Massar and Colleagues requires subjects to choose between low effort, low reward, and high effort, high reward alternatives, where the cognitive effort is operationalized by the number of words subjects are willing to type out backwards (Massar et al, 2015). Likewise, another group has developed behavioural paradigms in which participants decide between easy or difficult trials of the Stroop, Go/No-Go, and Wisconsin Card Sorting Tasks for small or large reward, respectively (Dixon and Christoff, 2012). In line with the other effort allocation paradigms described     8 above, offering incentives biases subjects away from their preference for automatic, pre-potent responding in these tasks (Dixon and Christoff, 2012).   In contrast to the tasks described above which probe the domains of working memory and cognitive flexibility, recent task development has focused on attentional effort. The Perceptual Effort Task is based on the rodent Cognitive Effort Task created by Winstanley and colleagues (Cocker et al, 2012b; Reddy et al, 2015) (discussed in detail in section 1.6.2 ). In this task, the objective is to correctly identify a faint stimulus in one of seven possible locations, and task difficulty is manipulated by the amount of gray-scale contrast between the stimulus and background. Individual perceptual thresholds are determined for each participant, so that easy stimuli are 98% of the participant’s threshold for detection, while the hard stimuli have a gray-scale that is 101% of their perceptual threshold. Easy trials offer a consistent small reward of $.10, while the size of the large reward is either equal or higher ($.10, $.20, $.30, $.40). Another attentional effort task utilizes the rapid serial visual presentation paradigm, in which participants monitor one of two target streams to the left and right of a fixation point for a target number, and cognitive effort is parametrically varied over six levels by increasing the number of times attention has to be switched between streams from one to six (Apps et al, 2015; Chong et al, 2017). On each trial, participants choose between a fixed, low-effort/low-reward “baseline” option, and a variable high-effort/high-reward “offer”. The increasing effort demands imposed by greater spatial attention are perceived as subjectively effortful, and act to devalue possible rewards (Apps et al, 2015)     9    1.2 Effort-Based Decision Making in Humans: Summary of Behavioural Findings, Relationship to Traits and Disease States    Most of the effort paradigms described above have either shown that subjects tend to avoid options as cognitive demand increases (Chong et al, 2017; Massar et al, 2015, 2016; Prevost et al, 2010; Reddy et al, 2015; Westbrook et al, 2013), or have demonstrated that increasing incentives can bias subjects towards otherwise costly cognitive expenditure (Chong et al, 2015; Dixon and Christoff, 2012; Massar et al, 2016; Reddy et al, 2015). Thus effort, whether in the physical or cognitive domain, is inherently costly, but these costs can be offset by potential lucrative outcomes. Subsequent work has assessed the different factors that may modulate effort allocation. For example, older adults show greater cognitive effort discounting relative to young adults on the COG-ED task, and willingness to exert cognitive effort is correlated with self-reported need for cognition (a trait measure of daily engagement with and enjoyment of cognitively demanding activities), but not intelligence (Lopez-Gamundi and Wardle, 2018; Westbrook et al, 2013). In the physical domain, men make more high effort choices on the EEfRT task relative to females, and social influence can bias effort allocation towards hard options, especially on low-reward probability trials (Gilman et al, 2015; Treadway et al, 2009). Trait anticipatory pleasure predicts effort exertion in the EEfRT task when reward receipt is not     10 guaranteed, while healthy individuals with elevated self-reports of both trait and state anhedonia exhibit a reduced willingness to make choices requiring greater effort (Geaney et al, 2015; Treadway et al, 2009).  This last finding, that trait anhedonia is negatively correlated with high effort choice, is particularly notable given that many psychiatric disorders characterized by anhedonia and avolition are also associated with an impaired willingness to exert effort, both in the physical and cognitive domain. Indeed, there a rich literature assessing effort-based decision making in schizophrenia, as these effort tasks may provide an objective measure of the negative symptoms characteristic of the disorder (Gold et al, 2015; Green et al, 2015). Negative symptoms can be divided into motivational negative symptoms (i.e avolition, anhedonia, and asocialty), and diminished expression (i.e. blunted affect and alogia), and are strong predictors of poor functional outcome (Green et al, 2009). Across tasks, individuals with schizophrenia regularly show impairments in both physical (Barch et al, 2014; Docx et al, 2015; Fervaha et al, 2013; Gold et al, 2013; McCarthy et al, 2016; Treadway et al, 2015)  as well as cognitive effort allocation (Culbreth et al, 2016, but see Gold et al, 2014). Many of these studies reporting impaired effort allocation also find that these deficits are associated with negative symptoms (Barch et al, 2014; Culbreth et al, 2016b; Gold et al, 2013; Hartmann et al, 2015; Treadway et al, 2015; Wolf et al, 2014), while some studies have found no relationship (Docx et al, 2015; Fervaha et al, 2013; Gold et al, 2014), as well as positive relationships between negative symptoms and more effortful choices (McCarthy et al, 2016). Discrepancies in     11 the literature might be explained by recent findings suggesting that defeatist beliefs moderate the relationship between negative symptoms and effort-based decision making. Specifically, high negative symptoms are associated with impaired effort allocation, but only in individuals with high levels of defeatist beliefs (Reddy et al, 2017). In any case, effort-based decision-making tasks appear to hold great promise in providing objective measures of the negative symptoms inherent to schizophrenia.       Deficits in effort allocation are not just limited to schizophrenia, but to other disorders characterized by symptoms of anhedonia or apathy. For example, depressed individuals are less willing to exert effort than controls, and duration of current depressive episode is negatively associated with effort investment (Treadway et al, 2012a). Likewise, Parkinson’s patients are less willing to accept bids to exert effort for potential rewards (Chong et al, 2015), Specifically, patients with concurrent apathy are less likely to exert effort for small rewards, while those off dopaminergic medication are less willing to exert high effort for high rewards (Le Heron et al, 2018). It must be said that deficits in effort allocation are not exclusive to disorders characterized by negative symptoms. For example, obesity has also been linked to reduced willingness to exert physical effort for larger reward outcomes, which is predictive of poor adherence to a weight loss program (Mata et al, 2017). When considered with the research linking negative symptoms to functional outcomes in schizophrenia described above, it appears that boosting effort allocation may be a promising therapeutic target across a number of conditions.       12   1.3 Neurotransmitter Systems Implicated in Effort Allocation from Human Studies   A handful of human experiments have uncovered key substrates guiding effort-based decision making. Most of this work has focused on dopamine (DA) and its links to physical effort allocation, given the wealth of preclinical research attributing a crucial role for DA signaling in valuations of physical effort (see section 1.8.1). In support of this link, acute administration of d-amphetamine (10, 20 mg) increases high effort choice in the EEfRT task, effects which are most pronounced in low and medium probability blocks where effort does not guarantee reward (Wardle et al, 2011). Using the same EEfRT task, DA sensitivity assessed via PET imaging is positively associated with high-effort choice on low probability trials in the left caudate nucleus, the left inferior temporal gyrus and ventrolateral prefrontal cortex, and ventromedial prefrontal cortex bilaterally (Treadway et al, 2012b). In contrast, DA sensitivity is negatively correlated with high effort choice across all trials in the anterior insula bilaterally, which is interesting given this area has been previously shown to process response costs (Kuhnen and Knutson, 2005; Prevost et al, 2010). Parkinson’s patients are less likely to invest effort for larger rewards when off medication, but are motivated to invest more effort when on their dopaminergic medication (Chong et al, 2015; Le Heron et al, 2018). Together this work suggests that     13 dopamine acts to bias choice towards effortful, more lucrative outcomes. Indeed, the preclinical literature provides strong support for this conclusion. While theoretical accounts suggest DA likely plays a similar role in cognitive effort allocation (Westbrook and Braver, 2016), this has not been experimentally verified.           The cannabinoid system has also received due consideration in regards to effort allocation. Acute treatment with vaporized cannabis (8 mg) reduces the likelihood of high-effort choices in the EEfRT task (Lawn et al, 2016). Similarly, smoking a joint containing 3.58% Δ9-THC, the main psychoactive component in cannabis and a CB1/CB2 receptor agonist, reduces button pressing in a choice task where participants earn money by choosing between a lucrative progressive ratio (PR) option or a fixed-time schedule (Cherek et al, 2002). In line with this, evidence suggests that heavy smoking adolescent cannabis users are less likely to select the effortful, high reward option in an effort discounting task, but that adult cannabis dependent users are identical to controls in their choice of effortful, high reward options in the EEfRT task (Lane et al, 2005a; Lawn et al, 2016). Notably, the study showing a similar choice profile between controls and adult cannabis dependent users did find impaired reward learning processes in this group, which is in agreement with work in the risky decision-making literature suggesting that THC reduces sensitivity to win and loss outcomes (Lane et al, 2005b). Two other neurotransmitters that have received comparatively less attention are adenosine and serotonin. Acute administration of caffeine (200 mg)     14 does not affect effort expenditure in the EEfRT task, except in a subgroup of individuals with a high cardiovascular response to caffeine, where the stimulant decreases their high-effort choice (Wardle et al, 2012). While serotonin has not been assessed explicitly in a cost/benefit decision-making framework, chronic escitalopram treatment increases physical grip for reward relative to placebo control (Meyniel et al, 2016).  Notably, all the studies described above focused on physical effort allocation, while no human study has attempted to delineate pharmacological regulation of effort allocation in the cognitive domain.   1.4 Human Imaging Studies on Effort Allocation     When reviewing the imaging literature, it is important to distinguish between tasks that present participants with concurrent options they must choose from, versus tasks presenting subjects with a cue denoting how much effort is to be invested in a subsequent trial. In the former, activity reflects the brain processes as subjects decide between options varying in their effort-cost/reward contingencies, while in the latter activity reflects how a lucrative outcome is discounted by the imposition of effort costs.  In response to cues denoting how much button pressing is required to obtain reward, blood-oxygen-level-dependent (BOLD) activity in the ventral striatum, ACC, and midbrain reflects the net value of the course of action, signaling the expected amount of reward discounted by the amount of effort to be     15 invested (Croxson et al, 2009). When presented with options differing in the amount of hand grip effort that must be invested for presentation of primary reward (erotic stimuli), and where subjects must choose how much effort to invest, anterior cingulate cortex and anterior insula activity represent the decreasing subjective value of the effortful option, in line with these regions coding the expected expense of energy (Prevost et al, 2010). Activity in the caudal dACC also represents the effort costs and benefits of a chosen relative to an alternative option, and integrates this into a subjective value signal that correlates with participants’ choices between levels of forceful grip (Klein-Flugge et al., 2016). A similar study using fictive monetary rewards found that choosing to grip over choosing to passively hold a hand-grip device increases BOLD activity in the superior frontal gyrus, while activity in the caudate-putamen is higher in anticipation of low-effort relative to high-effort trials, akin to a net value signal (Kurniawan et al, 2010). Similarly, the ACC and dorsal striatum (dorsal putamen) signal the anticipation of physical effort independently of the prospect of winning or losing reward in a task modeling monetary gain, but also monetary loss (Kurniawan et al, 2013).   In regard to effort allocation in the cognitive domain, anticipation of an upcoming arithmetic task activates the bilateral striatum, right ACC, and left brainstem (Vassena et al, 2014). Notably, this finding is not in line with a value-related encoding, according to which anticipation of low effort should elicit a greater response. However, a number of studies have found evidence for valuation signals in anticipation of a cognitive challenge. Botvinick (2009) found     16 that ventral striatal BOLD signals track the net value of outcomes in the Demand Selection Task (reviewed in 1.1.2), where cognitive demand varies as a function of the number of rule switches necessary for correct performance. Notably, this study found that activity in the dACC predicts the subsequent VS bold signal, supporting the notion that the dACC may regulate the allocation of cognitive effort via interactions with striatal subregions (Botvinick et al, 2009). In another study, in which participants make decisions between exerting different levels of cognitive effort (i.e., spelling words backwords), subjective value of effortful options is encoded in a network comprising the anterior and posterior cingulate cortex, caudate nuclei, cerebellum, bilateral temporal and parietal cortex, and lateral OFC (Massar et al, 2015). Interestingly, this study found ACC activation is more strongly correlated to effort discounting relative to delay discounting (Massar et al, 2015).   To date, only two studies have investigated physical and cognitive effort allocation in the same subjects and scanning session, providing a powerful means to compare and contrast the circuitries guiding these two forms of decision making. In one study, participants alternated between exerting different levels of force and rule switching in order to obtain fictive monetary reward (Schmidt et al, 2012). Ventral striatal activity predicts expected reward in either form of effort, and psychophysical interaction (PPI) analysis revealed the VS is significantly connected with cognitive regions (mostly the caudate) when cognitive demand is high, and with motor regions (mostly the putamen) when physical effort is required. The authors take this as evidence that the VS may be     17 a common motivational node, prioritising cognitive circuits during mental effort and motor circuits during physical effort exertion (Schmidt et al, 2012). In a study directly assessing effort-based choice, reward devaluation by both forms of effort is observed in a network of structures including the dorsomedial and dorsolateral prefrontal cortex, the intraparietal sulcus, and the anterior insula (Chong et al, 2017). Notably, this study also found that the right amygdala processes the subjective value of rewards discounted by cognitive, but not effort, costs (Chong et al, 2017).    1.5 Limitations of the Current Human Research on Effort Allocation  While valuable, the interpretation of the imaging results described above must be considered in light of the limitations of imaging data, such as difficulties in ascribing causality or directionality to any specific set of findings. Likewise, our understanding of the neurotransmitter systems involved in effort-based decision making (described in section 1.3) is limited by ethical considerations, and is inferred from certain disorders characterized by aberrant neural processes, or by assessing patients on or off certain medications. Often, however, these phenomena are confounded. A powerful alternative to circumvent these issues is to develop animal models of cost/benefit decision making, which allow researchers to causally investigate the neural systems guiding effort allocation in ways not feasible in human studies. As will be discussed in the below, animal     18 models have offered behavioural neuroscientists unprecedented insights into the neural mechanisms guiding effort-based decision making.    1.6 Animal Models of Effort-Based Cost/Benefit Decision Making   Animal models have proven invaluable towards investigating the neurotransmitter systems and neural circuits guiding cost/benefit decision making (Winstanley and Floresco, 2016). Many models have been developed, but all are based on the premise that subjects will evaluate whether the cost investment associated with different options is “worth” the potential reward resulting from their choice. Most of these paradigms present rodents with a low-cost, low-reward response option that is available with minimal to no expense, while selection of the other response alternative results in more lucrative reward, but is discounted by greater price. These costs can vary, and researchers have been truly creative in the way they have modeled different forms of decision making under this basic framework. Indeed, models exist which probe cost/benefit decision making involving delays (Cardinal et al, 2001), uncertainty (Cardinal and Howes, 2005; Cocker et al, 2012a; St Onge and Floresco, 2009), avoidance (Friedman et al, 2015), risk of punishment (Simon et al, 2009), and of particular interest, effort. Notably, animal effort-based decision-making models inspired many of the human tasks described above, such as the EEfRT task, which is based on the early work of Salamone and colleagues (Cousins and Salamone,     19 1994), and recently the rodent Cognitive Effort task has been back-translated for use in humans by Young and colleagues (Reddy et al, 2015).The sections that follow describe the various effort tasks used by behavioural neuroscientists, and systematically review the neurotransmitter systems and neural regions causally implicated in effort-based cost/benefit decision making. Most of this work has focused on the physical effort costs that discount lucrative outcomes, but I also detail work our lab has carried out investigating the substrates guiding cognitive effort allocation in rodents.    1.6.1 Physical Effort Decision-Making Tasks       1.6.1.1 Concurrent Choice Paradigms A number of sophisticated animal models have been developed to probe the neural mechanisms guiding physical effort-based choice, the earliest of which was developed by Salamone and colleagues (1991). In this concurrent fixed ratio 5 (FR5)/chow feeding procedure, a rat is placed in an operant box and can decide whether to eat freely available chow placed in the corner of the box (typically 15-20g), or to lever press on a FR-5 schedule for a single sugar pellet (Salamone et al, 1991). Given their preference for the high carbohydrate sugar pellets, rats typically spend most of the task lever pressing and consume very little of the freely available chow. Manipulations that reduce effortful choice (such as D2 receptor antagonism with haloperidol or 6-OHDA depletions; see section     20 1.8.1) decrease lever pressing and increase the amount of chow consumed, whereas appetite suppressants (such as cannabinoid receptor antagonist AM 251, see section 1.8.4) decrease both lever pressing and chow consumed (Cousins et al, 1994; Sink et al, 2008b). Notably, a variant of this task has rats choose between different concentrations of sucrose solution, where the choice is made between 7 lever presses for 5% sucrose or freely available 0.3% sucrose (Pardo et al, 2015). A subsequent iteration of the FR task, known as the progressive (PROG)/chow concurrent choice task, uses a progressive ratio schedule where the response requirements for a sugar pellet are increased by one each time 15 reinforcements are obtained (FR1×15, FR2×15, FR3×15,…) (Randall et al, 2012). In this task, animals repeatedly make within-session choices between lever pressing and chow intake under conditions in which the ratio requirement is gradually increasing. This procedure generates much more variability in behaviour between individuals compared to the concurrent FR-5 choice procedure, which is useful when studying the neural mechanisms responsible for individual differences in behaviour (Randall et al, 2012). Most recently a concurrent choice paradigm has been used by Izquierdo and colleagues, in which the required number of lever presses increases after five successive schedule completions, according to the formula ni = 5e(i/5)_5, where ni is equal to the number of presses required on the ith ratio, rounded to the nearest whole number (Hart and Izquierdo, 2017).  The advantage of these concurrent choice tasks is that they probe the cost/benefit evaluations of qualitatively different rewards (more/less preferred),     21 instead of differences in magnitude (larger/smaller) of the same reward (a feature of all the physical effort choice tasks discussed below). The disadvantage inherent to these tasks is the confound of delay costs associated with increasing response requirements. As described above, the imposition of delays can discount future rewards, and so can be difficult to interpret whether the switch to chow consumption reflects an aversion to effort costs, or an unwillingness to wait for the preferred sugar reward, unless appropriate control experiments are also conducted.   1.6.1.2 T-maze Choice Task  In the concurrent choice tasks described above, the responses required to obtain more-preferred sugar or less-preferred chow are quite distinct: lever pressing is an instrumental (or “preparatory”) response whereas feeding on chow is a consummatory (“to complete”) behaviour (Salamone et al, 1994; Salamone and Correa, 2012). While chow consumption also relies on simple responses such as locomotion to food, it is critical to evaluate cost/benefit decisions where explicit instrumental processes are compared. To address this, Salamone and colleagues developed a T-maze task in which rats decide between two arms varying in their reward/response cost ratios (Salamone et al, 1994). Subjects begin in a start arm and must choose between entering one of two opposing arms, where one arm end is associated with a small amount of freely available reward, whereas the other contains a higher magnitude of reward blocked by a scalable barrier. Typically, the low reward (LR) and high reward (HR) is set at two     22 and four pellets, respectively, while the scalable barrier has ranged in height from 25 cm (Schweimer, 2005), to the standard 30 cm used in most studies (Floresco and Ghods-Sharifi, 2007; Walton et al, 2002), up to 44 cm used in the original task (Salamone et al, 1994). At the beginning of each session rats typically receive two forced choice trials to expose them to the contingencies associated with each arm, and then receive 10-30 free choice trials where the number of high-effort/high reward arm entries or % HR choice is reported (Floresco and Ghods-Sharifi, 2007; Rudebeck et al, 2006; Salamone et al, 1994). Rats willingly scale the barrier for larger reward, and choice of the hard lever typically falls between 70-90% across studies. Alterations in choice behaviour due to motor, spatial memory, or discrimination deficits can be ruled out by equalizing response costs via insertion of an identical barrier in the LR arm, and sensitivity to reward magnitudes or effort costs can be assessed by altering the reward ratio differential (such as 5-1 versus 4-2) or changing the height of the barrier (e.g. from 30cm to 20 cm) (Walton et al, 2003a). A notable variation of the standard T-maze choice task employs a discounting procedure, where each time a rat chooses the HR arm, the amount of food in the arm is reduced by one pellet on the next trial (Bardgett et al, 2009). Such a procedure allows one to determine an indifference point for each rat, which provides a sensitive measure of reward value for each subject (Richards et al, 1997). In another variant of the T-maze task, Euston and colleagues have changed the cost parameter by replacing the scalable barrier with a weighted lever that rats must depress for larger sucrose reward (Holec et al, 2014). This is an interesting take on the traditional T-maze     23 task, allowing researchers to assess how manipulations affect allocation of distinctly effortful actions.   The T-maze choice task is the most commonly employed test of physical-effort based decision making, which has also been adapted for use with mice (Iodice et al, 2017; Pardo et al, 2012). However, training and testing is more involved than the operant box-based tasks described below, and again there is the confound of delay costs invariably imposed by scaling a barrier for reward. While some investigators have recorded latencies at different phases towards HR receipt (e.g. start to choice point, choice point to top of barrier, top of barrier to food) (Denk et al, 2005), few have factored latency data into their statistical analyses when discussing shifts in choice from HR to LR arms ( but see Yohn et al, 2015a).      1.6.1.3 Lever-Based Effort Discounting Tasks Two groups have developed lever-based operant tasks of physical effort allocation. The first, developed by Walton et al., (2006), requires rats to decide between two response alternatives, one of which involves investing effort by lever pressing on a high-fixed ratio to gain high reward (four food pellets, HR), while the other produces lower reward through completion of a less effortful FR requirement (two food pellets, LR) (Walton et al, 2006). Each session consists of repeating blocks of two-forced choice (one to each lever) and four free-choice trials, a protocol that allows sufficient assessment of decision-making behaviour while ensuring that rats sample each option at least once out of every six trials.     24 The FR requirement on the LR lever remains constant at FR4, while the FR requirements on the HR lever also remain constant within a session (typically set at FR12). The second operant task developed by Floresco and colleagues is similar, but within each session rats go through four blocks of trials where the fixed ratio requirement for the high reward (HR) systemically increases (Floresco et al, 2008b). Each block of trials begins with two forced choice trials (one for each lever), and is followed by ten free choice trials where rats can press a lever once for an immediate reward of two sugar pellets, or can complete a more effortful fixed ratio requirement on the other lever, resulting in four sugar pellets. The fixed ratio responses required to obtain the HR varies systematically across each block (starting at 2, then 5, 10 and 20 presses), resulting in a discounting curve showing how the value of HR reward is depreciated by increasing physical effort costs. An advantage of operant based paradigms is the ability to assess choice latencies, rates of responding (an index of response vigor), and the ability to equate the delay to obtain either reward. Indeed, a notable variant of the effort discounting task by Floresco et al, known as the “Effort Discounting Tasks with Equivalent Delays”, eliminates the confound of delay costs associated with increased responding on a lever, and is the only physical effort task to adequately control for time to reward receipt (Floresco et al, 2008b). Specifically, after responding on the LR lever two pellets are delivered after a delay equivalent to the time required for rats to complete the FR ratio on the HR lever. This delay to LR reward receipt increases across trial blocks, and is calculated based on the average time required for rats to press the lever 2,5,10, and 20 times. When     25 delays to receiving LR or HR are equalized, choice of HR significantly increases, suggesting that delays to reinforcement incurred when animals must respond on a lever multiple times influence the preference for these options above and beyond their inherent response costs (Floresco et al, 2008b).   1.6.2 Modelling Cognitive Effort Costs in Rodents: The Cognitive Effort Task    While a number of paradigms have been developed to model decision making with physical effort costs, only one animal model of cognitive effort allocation currently exists. Developed by Winstanley and colleagues, the rodent Cognitive Effort Task (rCET) is an adaptation of the well-established 5-choice serial reaction time task (5-CSRTT) (Cocker et al, 2012b). The 5-CSRTT is an animal model of sustained and divided attention, based on the human continuous performance task originally developed by Mirsky and Rosvold, which has recently been back-translated for use in humans (Voon et al, 2014). In the original 5-CSRTT, rats scan a visual array of five stimulus holes and must correctly identify the presentation of a brief visual stimulus (typically 0.5s in duration), which is rewarded with a single sucrose pellet. In the rCET, rats are presented with two levers at trial onset, allowing them to choose whether to engage in an easy or difficult attentional challenge. Choice of the easy, low reward option (LR) requires accurate detection of a 1s stimulus, and results in the delivery of one sucrose reward if correctly identified. Alternatively, rats can opt to initiate hard, high     26 reward trials (HR) where the stimulus is much briefer (0.2s), but can result in double the reward. HR trials are more difficult, as evidenced by lower accuracy rates relative to LR trials, but are ultimately the advantageous option if rats can successfully complete them. Indeed, while all rats can detect the HR stimulus successfully (~65-70% HR accuracy rates, on average), there is great individual variation in preference for the hard option. On average, choice of the HR lever is ~70%, and subjects are split into “workers” and “slackers” depending on whether they tend to choose the HR lever more or less than 70% during a session, respectively. Notably, “workers” and “slackers” do not differ in their accuracy to detect the difficult HR stimulus, nor do they differ in their rates of premature responding (ability to wait for the stimuli to appear), and so differences in cognitive effort allocation are not mediated by disparities in ability to perform the task.  Notably, choice patterns on the rCET are influenced by the cognitive effort costs associated with HR trials, and cannot simply be explained by sensitivity to differential rates of reinforcement. In a yoked-control rCET task, rats were trained on a control task where the reinforcement probabilities on both the LR and HR trials depended on a master animal’s performance on the rCET. The average accuracy on LR and HR trials was calculated for each rat performing the rCET at baseline, and these accuracies were used to determine the reward schedule for a yoked control rat. The structure of the control task was identical to that of the rCET, with the exception that the stimulus light remained illuminated until the rat made a response. Baseline choice behavior of cognitive effort and     27 control animals was directly compared, and these analyses revealed significant differences in choice behaviour between tasks. While workers and yoked-workers selected the HR option at similar rates, yoked-slackers chose HR trials more than slackers, suggesting that the imposition of a cognitive cost (increased attention) discounts selection of the HR reward in the standard task.   Like the original 5-CSRTT, the rCET provides a number of cursory measures to help interpret the effects produced by experimental manipulation (Robbins, 2002). LR and HR accuracies provide measures of attention for stimuli of differing durations, and the percent prematures for each lever indexes rats’ willingness to withhold responding during the inter-trial-interval prior to stimulus presentation. This is identical to the premature responding measure in the 5-CSRTT, which is considered a sensitive assay of impulsive action (Robbins, 2002). Other measures recorded include choice omissions at the decision time point, response omissions following presentation of the stimulus, as well as latencies to make a decision, to correctly detect the visual stimulus, and to collect reward from the magazine tray. By taking into account the overall constellation of effects observed on these variables within a session, it is possible to disambiguate possible interpretations of the data. For example, increased choice, correct, and collect latencies along with increased omissions likely reflect sedative or motor effects, whereas a pronounced slowing in collect latency likely reflects a motivational impairment. Likewise, increases in response omissions that are accompanied by accuracy impairments, but not changes in collect latencies likely reflect attentional deficits.     28     1.7 Comparing Effort Allocation Tasks   As described above, a number of different tasks can be used to study effort allocation in rodents, and deciding which one to employ for a particular investigation will depend on a host of factors. For example, if individual differences are of primary interest, then the (PROG)/chow concurrent choice task and the rCET are advantageous because they generate greater variability in effort allocation, such that “high” versus “low” effort preference groups can be generated and included in subsequent analyses (Cocker et al, 2012b; Randall et al, 2012). If the primary interest is in modelling cognitive effort costs, then the rCET is the clear choice as it is currently the only task available. Importantly, the rCET manipulates cognitive effort by increasing attentional demand, while cognitive effort discounting has been observed in the domains of working memory, response inhibition, and cognitive flexibility as well (Dixon and Christoff, 2012; Kool et al, 2010a; Reddy et al, 2015; Westbrook et al, 2013). It is likely that different neural systems mediate the allocation of effort across these executive functions, and so discretion must be taken when extrapolating the results of the rCET to cognitive effort allocation as a general phenomenon. Likewise, physical effort is not a unitary construct, and the processes involved in scaling a barrier for larger reward may not be the same as those regulating lever pressing for lucrative outcomes (Holec et al, 2014). This must be kept in mind when     29 considering tasks that vastly differ in the instrumental responses defining the effort cost.  Some considerations are purely practical, such as equipment availability (operant box, T-maze) and the amount of resources that can be invested for behavioural training (e.g., rats must be trained for 6-8 months on the rCET before any experimental manipulations can begin). Other considerations, however, are more nuanced, and reflect subtleties in task design that may interact with a given manipulation. For example, consider the effects of nucleus accumbens (NAc) 6-OHDA depletions on the operant effort-discounting task developed by Walton et al., (2009), and the effect of NAc inactivations on the similar operant discounting task developed by Floresco et al., (2008). 6-OHDA depletions have no effect on choice in the former task, while NAc inactivations reduce choice of the effortful option in latter (Ghods-Sharifi and Floresco, 2010; Walton et al, 2009). Although the difference in NAc manipulation may explain these divergent results, it has been suggested that these disparate results are due to subtle differences in task design, interacting with the role of NAc dopamine in reward-seeking behaviour (Nicola, 2007, 2010). According to the Flexible Approach Hypothesis, NAc dopamine is recruited only when actions required to obtain reward are variable. As such, when an animal’s starting location varies across trials, different actions are required to reach a fixed goal, and NAc dopamine is required for those actions to occur. In contrast, when the start and end locations are fixed across trials, identical actions can bring the animal to the goal, and NAc dopamine is not required. Notably, in the task reporting no effects on choice, rats start at the     30 exact same position on each trial, because a nosepoke in the magazine tray is required to extend the left and right lever. In the study reporting reduced choice of the HR lever following striatal inactivation, trial initiation is not contingent on a nosepoke in the magazine tray, and so approach towards a given lever in this task is necessarily flexible (Nicola, 2010). This is but one example of the factors a behavioural neuroscientist must consider when deciding between effort-based decision-making tasks. In the sections that follow, I review the neurotransmitter and neural systems that have been implicated in effort allocation using these behavioural paradigms.    1.8 Neurotransmitter Systems Implicated in Effort-Based Decision Making   A number of neurotransmitter systems have been implicated in decision making with effort costs. The research in this section is primarily based on the results of systemic pharmacology studies, in which a drug is administered intraperitoneally (I.P) or subcuteanously (S.C) before task onset. Site-specific targeting of a neurotransmitter system has also been achieved by intracerebral drug infusion, and more recent studies have used transgenic or knockout animals to similar effect. Most of this work has been conducted in rodents, but where applicable the results of nonhuman primate and human studies will be briefly described. A review of the literature suggests that while many neurotransmitter systems play similar roles across effort domains, others have more nuanced     31 roles in decision making with either physical or cognitive effort costs.    1.8.1 Dopaminergic Systems   1.8.1.1 Physical Effort: Systemic Administration  Across a variety of tasks with physical effort costs, systemic administration of dopaminergic compounds overwhelmingly implicate the dopamine (DA) system in physical effort allocation.  In the concurrent (FR5)/chow feeding task, systemic administration of the nonspecific DA antagonist cis-flupenthixol (0.15,0.30 mg/kg)1, and the D2 receptor antagonists haloperidol (0.1-0.15 mg/kg), eticlopride (.05, 0.1 mg/kg) and sulpiride (100, 150 mg/kg) decrease lever pressing for sucrose and increase intake of freely available chow (Cousins et al, 1994; Salamone et al, 1991; Sink et al, 2008b). Initial studies revealed that the D1 receptor antagonist SCH23390 (0.05, 0.1, 0.15 mg/kg) reduces lever pressing and increases chow intake, although this drug displays high binding affinity for 5-HT2A and 5-HT2C receptors (Alburges et al, 1992). Subsequent studies have employed the D1 antagonist ecicopam (SCH 39166), which has high affinity for D1 receptors and low affinity for 5-HT receptors (Alburges et al, 1992). Systemic administration of this compound (.05, 0.1, and 0.2 mg/kg) also decreases lever pressing and increases chow intake, effects which are partially reversed by the D1 receptor agonists SKF 38393 (.05 mg/kg), SKF 81297 (.05 mg/kg), and A77 636 ( 0.5                                             1 Doses listed in this section are effective doses that had effects on decision-making, not all doses used in a specific investigation     32 mg/kg)  (Sink et al, 2008b; Yohn et al, 2015a). Another method of downregulating DA transmission has involved systemic administration of tetrabenazine, a vesicular monoamine transporter-2 (VMAT-2) inhibitor, which blocks the storage of monoamine and depletes their levels (Pettibone et al, 1984). Salamone and Colleagues have used this compound, in conjunction with effort-based decision-making tasks, to model the motivational symptoms of depression in animals (Nunes et al, 2013b). Administration of tetrabenazine (0.75, 1 mg/kg) decreases lever pressing and increases chow intake, effects which are attenuated by the dopamine transporter (DAT) reuptake inhibitors GBR12909 (2.5, 5 mg/kg), buproprion (10, 15 mg/kg) and PRX-14040 (2.5, 5,10 mg.kg), but not by the norepinephrine transporter (NET) reuptake inhibitor desipramine, or by the serotonin transporter (SERT) reuptake inhibitor fluoxetine (Randall et al, 2014; Yohn et al, 2016b, 2016d). Notably, the effects of these dopaminergic agents are not restricted to concurrent choice of qualitatively different rewards: in a variant of the FR(5)/chow concurrent choice task, ecicopram (0.1, 0.2 mg/kg), haloperidol (0.05, 0.1 mg/kg), and tetrabenazine (0.75, 1.0 mg/kg) decrease lever pressing on a FR7 schedule for 5% sucrose solution, and increase selection of  freely available 0.3% sucrose (Pardo et al, 2015). When administered prior to the (PROG)/chow concurrent choice task, the D1 receptor antagonist ecopipam (0.5, 1.0, 2.0 mg/kg) decreases lever pressing while concomitantly increasing chow intake (Randall et al, 2014). In contrast, eticlopride (.04, .08 mg/kg), haloperidol (0.1 mg/kg), and tetrabenazine (0.5, 0.75, 1.0 mg/lg) decrease lever pressing at increasing FR ratios for sucrose, but this is     33 not accompanied by increases in chow intake (Randall et al, 2012, 2014).This is unlikely to reflect a reduced motivation for food, as rats will still seek food in a session. Instead this likely reflects a ceiling effect, as normal chow consumption in the (FR5)/chow feeding task is relatively low (1-2 grams), whereas in the (PROG)/chow feeding task it is higher (7-8) grams, and approximates the maximum amount of food rats will consume within a 30 minute session without water (Randall et al, 2010). Tetrabenazine-induced declines in progressive ratio lever pressing for lucrative reward are attenuated by the DAT inhibitor and common antidepressant bupropion (Wellbutrin) (15 mg/kg), by the monoamine oxidase B inhibitor (MAOBi) deprenyl (5.0, 1.0 mg/kg), but not by the catchecol-O-methyltransferase (COMT) inhibitor tolcapone (Randall et al, 2014),   In the T-maze task, numerous investigators have demonstrated that doses of haloperidol ranging from 0.1 (Bardgett et al, 2009; Salamone et al, 1994) to 0.2 mg/kg (Denk et al, 2005; Walton et al, 2005) decrease willingness to scale the barrier for larger reward. A motor slowing, as indexed by increased trial latencies, routinely accompany these doses of haloperidol. Regardless, rats are still capable of scaling the barrier, because they resume selection of the high reward arm when costs in both arms are equated. Similar effects of haloperidol are observed in a mouse variant of the T-maze task (0.05, 0.1 mg/kg) (Pardo et al, 2012). The D1 receptor antagonist ecicopram (0.1, 0.2 mg/kg) similarly decreases choice of the HR arm in rats, an effect which is reversed by the D1 receptor agonist SKF 3893 (0.25, 0.5, 0.75 mg/kg) (Yohn et al, 2015a). In one of the few investigations targeting the D3 receptor subtype, Bardgett et al., (2009)     34 administered the D3 receptor agonist 7-OH-DPAT (0.1, 0.3 mg/kg) and D3 receptor antagonist U-99194 (6.25 mg/kg) on the T-maze, and found no effects of either on the number of trials it took rats to switch preference to the low reward arm. In keeping with its effects on the concurrent choice tasks, tetrabenazine (0.75 mg/kg) decreases rats’ willingness to scale the barrier for larger reward (Yohn et al, 2015b). These impairments in HR choice are attenuated by SKF 81297 (0.05 mg/kg), and by bupropion (10, 15 mg/kg) (Yohn et al, 2015b).  In reference to the lever-based operant effort discounting tasks, haloperidol (0.1 mg.kg) decreases choice of the high reward/ high FR (4 pellets/FR8) lever in lieu of increased selection of the low reward/ low FR (2 pellets/FR4) alternative (Walton et al, 2009). Cis-flupenthixol (0.25, 0.5 mg/kg), also decreases choice of the effortful option across trials blocks with increasing response costs, an effect which is observed even when the delays to receive LR and HR reward are matched (Floresco et al, 2008b). High dose eticlopride (0.06 mg/kg) and SCH 23390 (0.01 mg/kg) similarly increase effort discounting, but SCH 23390 only decreases choice of the HR lever when response requirements are high (20 lever presses) (Hosking et al, 2015b).  Most systemic work has focused on DA receptor antagonism, and few studies have targeted the DA receptor subtypes alone with selective DA agonists. In most cases, a hyperdopaminergic state has been achieved by administering DAT inhibitors. GBR 12909 (5, 10 mg/kg), bupropion (20, 40 mg/kg) and PRX-14040 (20, 40 mg/kg) increase lever pressing and decrease chow consumption in the (PROG)/chow concurrent choice task, suggesting that     35 excess DA may invigorate behaviour towards lucrative, yet costlier response alternatives (Randall et al, 2015; Yohn et al, 2016c, 2016d). Amphetamine has also served as a proxy for dopamine agonism, given substantial evidence that the drug acutely enhances extracellular levels of DA by blockade of DA reuptake (Hutson et al, 2014). In humans, amphetamine (10, 20 mg) increases willingness to expend effort for monetary reward, in a task where subjects can decide between easy (i.e. 30 button presses within 7s) or hard  (i.e.100 presses with nondominant pinky finger in 21s) response costs (Wardle et al, 2011). Notably, this effect is most pronounced for low reward probability blocks, where fulfilling the effort requirements produces reward on 12.5 or 50% of trials (but not 88%). When administered to rodents, amphetamine (2-3 mg/kg) decreases lever pressing and chow consumption on the concurrent (FR5)/chow feeding task, although these doses are substantially higher than those used to modulate other forms of decision making in this species (Silveira et al, 2014). On the T-maze, a lower dose of amphetamine (.75 mg/kg) increases choice of the HR arm, an effect that is antagonized by both SCH23990 (0.0125 mg/kg) and haloperidol (0.1 mg/kg), indicating a DA receptor-mediated mechanism of action (Bardgett et al, 2009). Likewise, a low dose of amphetamine (.25 mg/kg) increases choice of the HR option in the operant effort-discounting task, but a higher dose (0.5 mg/kg) has the opposite effect, systematically decreasing preference for the HR lever as the response requirement increases across trials blocks (Floresco et al, 2008b). An interesting follow-up revealed that when the delays to receive LR and HR were equated, low dose amphetamine no longer biases choice towards the     36 effortful option, but the higher dose continues to amplify effort discounting (Floresco et al, 2008b). This work suggests that low doses of amphetamine may promote effort investment towards larger reward by reducing the impact of delay costs inevitably associated with most high-effort requirements.  When considered across physical effort tasks, the results of systemic administration studies overwhelmingly indicate that dopamine, acting primarily on D1 and D2 receptors, biases organisms to exert more physical effort for lucrative reward. As will be discussed below, these effects are likely mediated by dopaminergic transmission in the nucleus accumbens.       1.8.1.2 Physical Effort: Intracerebral Infusions and 6-OHDA Lesions  Site-specific targeting of DA transmission has been investigated in two regions: the anterior cingulate cortex (ACC) and nucleus accumbens (NAcc) (see section 1.10.1 and 1.10.2 respectively for a general discussion of these areas). 6-hydroxydopamine (6-OHDA) lesions to the anterior cingulate decrease choice of the HR arm in the T-maze task, an effect that is mimicked by infusions of a D1 (SCH 23390), but not a D2 receptor (eticlopride) antagonist into this region (Schweimer et al, 2005; Schweimer and Hauber, 2006). This suggests DA in the ACC may be acting on D1 receptors to bias behaviour towards high effort, high reward options. However, a similar study also found no effects of ACC 6-ODHA lesions on effort allocation in the T-maze task (Walton et al, 2005). A number of methodological differences might account for this discrepancy, including the doses of 6-OHDA used (the study reporting effects on choice infused 3 µg at 12     37 sites, whereas the one reporting null effects infused 2 µg at 8 sites), the time from surgery to retesting, as well as rodent strain used (Spragues versus Lister Hooded, respectively) (Schweimer et al, 2005).  6-OHDA lesions of the nucleus accumbens similarly disrupt lever pressing for sucrose when chow is concurrently available, but rats exhibit normal levels of pressing on alternate days when chow is not available (Cousins and Salamone, 1994). Although a significant correlation between DA levels and lever pressing was observed in this study, the shifts in choice cannot be attributed to motor impairments induced by depletions of NAcc DA, as rats readily lever press for sucrose when the chow is not available (Cousins and Salamone, 1994). This can be contrasted with DA depletions of the ventrolateral striatum, which produce severe deficits in motor ability that interfere with lever pressing and chow consumption (Cousins et al, 1993). The effects of accumbal DA depletions are likely driven by reduced DA transmission in the accumbens core, as selective DA lesions to this region, but not the shell, impair optimal response allocation in the concurrent choice task (Sokolowski and Salamone, 1998). Accordingly, infusion of the D1 antagonist SCH 23990, D2 antagonist raclopride, or the VMAT inhibitor tetrabenazine into the NAc mimic the effects of DA depletions to this area on the (FR5)/chow concurrent choice task (Nowend et al, 2001; Nunes et al, 2013b). NAc 6-OHDA lesions similarly decrease choice of the HR arm in the effort T-maze task, suggesting that mesoaccumbens DA invigorates responding to overcome work-related costs for preferred reward (Salamone et al, 1994).    While NAc DA signaling is clearly implicated in decision making with effort     38 costs, it is unclear exactly how mesolimbic DA contributes to the cost/benefit computation. DA release in the NAc appears to operate under two modes- a “phasic” rapid signaling driven by midbrain DA neuron burst firing, and an extrasynaptic “tonic” mode which changes on a slower timescale and reflects the overall number of active DA neurons (Floresco, 2015). It is well-established that phasic signaling codes a temporal reward prediction error reflecting the difference in value between a received versus predicted reward at each moment in time (Schultz, 2016; Schultz et al, 1997). However, this phasic signaling is also thought to convey the subjective value or utility attached to reward-predictive stimuli and actions (Schultz et al, 1997), such that “cached” action values may be used to guide future choice behaviour (Lee et al, 2012). Phasic DA has been shown to signal many of the economic attributes that might guide choice behaviour, including objective expected value (Fiorillo et al, 2003; Roesch et al, 2007; Tobler et al, 2005), as well as subjective preference for risky (Sugam et al, 2012)  or delayed rewards (Day et al, 2010). However, a number of studies have demonstrated that the cost of effort is not faithfully encoded in the DA signal in either rodents (Gan et al, 2010; Hollon et al, 2014) or primates (Pasquereau and Turner, 2013). Rather than conveying a cached action value, it has been proposed that DA may modulate incoming inputs from other brain regions which themselves encode effort costs, such as the anterior cingulate cortex (Rudebeck et al, 2006), basolateral amygdala, and insula (Prevost et al, 2010). NAc DA may set the threshold for what is considered appropriate effort for a given reward, whereby action policies provided by cortical and limbic inputs are then modified     39 according to this DA tone (Hauber and Sommer, 2009; Phillips et al, 2007; Sesack and Grace, 2010). Specifically, high DA in the NAc may minimize response costs and increase willingness to exert effort for larger reward, while low DA tone (as observed following 6-OHDA lesions) reduces this willingness. This is then compared to a general utility threshold, where effortful options high in utility are selected, but those falling below this threshold are not (Phillips et al, 2007). This proposal does not differentiate between phasic and tonic signaling, but tonic DA may impose a steady state cost-benefit bias, whereas phasic signaling in response to unexpected rewards may provide an opportunistic window where higher effort might result in more reward (Phillips et al, 2007). Similar theories suggest that tonic DA may act as a running meter of available reward, which can increase response vigor in situations where average reward rate is high (Niv et al, 2007). In support of these theories, DAT knockout mice which exhibit increased tonic firing and elevated DA levels demonstrate increased lever pressing in a modified version of the FR/chow concurrent choice task (Cagniard et al, 2006).    1.8.1.3 Cognitive Effort  In contrast, DA signaling does not appear to play a critical role in the processes guiding cognitive effort allocation, at least in rodents. When administered prior to performance of the rCET, the D1 receptor antagonist SCH 233190 and the D2 receptor antagonist eticlopride do not affect choice of the hard lever option, even though both drugs decrease the number of trials     40 completed and increase omission rates (Hosking et al, 2015b). In contrast, these dopaminergic agents reduce hard lever choice in an operant physical effort choice task (Hosking et al, 2015b). It is worth noting that the design of the rCET necessarily makes it a physically effortful task; in a typical session, rats complete approximately 130 trials in which they hit a lever, orient towards the stimulus array to detect the stimulus, then turn around to either collect reward or commence the next trial. However, the cognitively effortful options do not differ in their physical requirements, and so it is notable that dopamine antagonism does not affect choice, but does affect trials completed and omission rates. This might suggest DA is not involved in cognitive effort allocation per se, but may regulate the amount of physical effort recruited in a typical rCET session. While computational modeling studies have suggested a role for DA in allocation of cognitive effort, no human work has directly tested this hypothesis (Westbrook and Braver, 2016). It will be interesting to test a broader range of dopaminergic agents on the rCET, to assess whether DA involvement reflects a fundamental distinction between decision making with physical or cognitive effort costs.        1.8.2 Adrenergic Systems  The existing evidence suggests that noradrenaline (NA) is not involved in encoding of effort costs, or in the decision to engage in an effortful challenge, regardless of whether it is in the physical or cognitive domain. Enhancing noradrenergic tone with the NET inhibitors desipramine (2.5, 5.0, 10.0 mg/kg) or     41 atomoxetine (0.125, 0.25, 0.5, 1.0 mg/kg) fails to increase response output in the (PROG)/chow concurrent choice task, and actually decreases lever pressing at higher doses or when administered repeatedly (Yohn et al, 2016c). On the operant effort discounting task, atomoxetine also has no effect on rats’ choice profile (0.1, 0.3, 1.0 mg/kg) (Hosking et al, 2014b). The a2-adrenergic receptor antagonist yohimbine (5 mg/kg) appears to have some minor effects on choice behavior, decreasing choice of the HR lever during the first two blocks when the response costs are two and ten lever presses, but this effect is not robust (Hosking et al, 2015b). Rather, NA is likely mobilized during the actual experience of physical effort, as the body recruits the necessary muscles and autonomic arousal processes necessary to fulfill an effortful challenge (Burnstock, 2009; Carter et al, 2010; Robbins, 1984). In line with this, Varazzani et al., (2015) showed that in response to cues signaling the anticipated effort level of an upcoming physical challenge, firing of dopamine neurons in the substantia nigra pars compacta (SNc) reflects both the size of expected reward and amount of effort required to obtain it. In contrast, modulation of LC activity reflects the difficulty of the task at hand, increasing with effort level during the action (Varazzani et al, 2015). This work suggests DA and NA may work cooperatively to regulate effort-based choice: DA may encode the expected value of an upcoming effortful challenge, whereas NA may react to information about a current challenge, and will mobilize the resources necessary to overcome it.     When administered on the rCET, atomoxetine (0.1, 0.3, 1.0 mg/kg) has no     42 effort on choice behavior, but increases choice latencies and choice omissions, and slightly impairs accuracy for easy trials in worker rats (Hosking et al, 2015b). Similarly, a high dose of yohimbine (5 mg/kg) does not affect choice, but impairs accuracy across all rats (Hosking et al, 2015b). This profile mirrors the contributions of NA to physical effort, where it does not affect the decision to engage, but does affect actual performance of the effortful challenge. Indeed, attentional effort is associated with autonomic arousal (Egeth and Kahneman, 1975; Howells et al, 2010), and NA appears to play a selective role in the 5-CRTT in novel or arousing circumstances (i.e variable ITI, random bursts of white noise), by potentially optimizing attentional performance in these situations (Carli et al, 1983; Cole and Robbins, 1987; Robbins, 2002).    1.8.3 Cholinergic Systems  A number of cholinergic drugs have been assessed on the rCET. These experiments suggest that acetylcholine has dissociable roles in regulating attention versus the decision to engage in an attentional challenge. Systemic administration of nicotine (1.0 mg/kg) decreases choice of the HR option in slacker rats, while concomitantly increasing their attentional accuracy on hard trials (Hosking et al, 2014c). Importantly, the improved accuracy in slackers following nicotine does not reflect amelioration of an attention deficit, as numerous studies have shown that workers and slackers do not differ in attentional ability (Cocker et al, 2012b). Rather, increasing neuronal activity via     43 nicotine receptor activation simultaneously improves ability while diminishing optimal choice, suggesting that cholinergic regulation of these processes is dissociable. Indeed, cholinergic regulation of attention versus decision making may map onto dissociable cholinergic neurons in the brain, such as the cortically projecting neurons of the basal forebrain, versus the locally acting striatal interneurons (Bubser et al, 2012). In contrast, systemic administration of the muscarinic receptor antagonist scopolamine (0.3 mg/kg) decreases choice of the HR option across all rats without affecting attentional accuracy, while the muscarinic receptor agonist oxotremorine and the nicotine antagonist mecamylamine are without effect (Hosking et al, 2014c). Collectively, these results suggest that perturbing cholinergic signaling may impair or improve decision making with cognitive effort costs, the results of which depend on individual propensity to work or slack at baseline. Additionally, these effects on choice behaviour are separable from the established role of acetylcholine in regulating attentional processes (Himmelheber et al, 2000; Mcgaughy et al, 2002; Passetti et al, 2000).   In comparison, relatively little work has focused on cholinergic contributions to physical effort-based decision making. The nonspecific muscarinic agonist pilocarpine (1.0, 2.0, 3.0 mg.kg) decreases lever pressing on the (FR5)/chow concurrent choice task, but does not mimic the increased chow consumption commonly observed following systemic administration of dopamine antagonists (Cousins et al, 1994). In contrast, direct infusions of pilocarpine into the NAc causes rats to reallocate their responding away from the lever and     44 towards the freely available chow, an effect which is antagonized by co-infusion of the antagonist scopolamine (Nunes et al, 2013a). Thus it appears cholinergic signaling does contribute to physical effort allocation, likely via interactions with DA signaling in the NAc, but this has not been directly investigated (Threlfell et al, 2012). In order to further characterize the contribution of cortically projecting cholinergic neurons to decision making with cognitive effort costs, I carried out an experiment (Chapter 3) in which I used chemogenetics to selectively downregulate basal forebrain cholinergic neurons as rats performed the rCET.   1.8.4 Cannabinoid Systems  Studies in both humans and rodents suggest that pertubrations of the cannabinoid system can influence effort-based decision making (See section 1.3 for a discussion of the available human evidence). In rodents, CB1 receptor inverse agonism/antagonism with AM 251 (2.0, 4.0, 8.0 mg/kg) decreases lever pressing in the FR5/chow concurrent choice task, but does not lead to a corresponding rise in chow consumption (Sink et al, 2008b). A similar effect of AM 251 (16.0) mg/kg is also observed in the (PROG)/chow concurrent choice task, and administration of the neutral CB1 anatagonist AM 4113 (4.0, 8.0, 16.0 mg/kg) likewise reduces lever pressing and decreases chow consumption (Randall et al, 2014). These effects are line with the known role of CB1 receptor antagonists as appetite suppressants, and their ability to reduce food-reinforced behavior (Salamone et al, 2007; Sink et al, 2008a). In contrast, direct infusions of     45 arachidonyl-2-chloroethylamide (ACEA), a potent synthetic CB1receptor agonist, into the ACC causes a robust shift in choice towards the small, unobstructed reward option, suggesting that rats are less willing to exert physical effort for preferred rewards (Khani et al., 2015). This shift to LR does not occur when the costs in reward arms are equalized, or when the antagonist AM 251 is microinjected into the ACC (Khani et al, 2015). Notably, intra-ACC ACEA infusions affect physical effort but not delay-based decision making, whereas infusions of ACEA into the OFC have the opposite effect (Khani et al, 2015). These results mimic the effects of lesions to the ACC and OFC on effort and delay-based decision making, respectively, and suggest that CB1 receptors in distinct prefrontal regions are necessary when choosing between options varying in delay or effort (Rudebeck et al, 2006). Most recently, it has been demonstrated that microinjection of the CB1 receptor agonist WIN 55, 212-2 (2, 10 and 50 µM) into the nucleus accumbens decreases HR choice in the T-maze task, an effect which is antagonized by co-infusion of AM 251 (45 µM) (Fatahi and Haghparast, 2018). Overall, it appears then that cannabionoid receptor activation, in both prefrontal and striatal areas, may bias choice away from rewards that are discounted by physically effortful endeavors.  In accordance with the available evidence from the physical effort literature, we have also shown that cannabinoid signaling regulates the willingness to exert cognitive effort. When administered on the rCET, THC (1.0, 2.0, 3.0 mg/kg) dose-dependently decreases choice of the difficult, high-reward option requiring accurate detection of a brief (0.2 s) light stimulus;     46 correspondingly, rats shift choice to the easier, low-reward option where the light stimulus is presented for a longer duration (1 s) (Silveira et al, 2016). Importantly, the lack of effect of THC on attentional accuracy suggests that the choice shift is not due to an inability to complete high-reward trials. Moreover, THC-induced choice impairments are correlated with CB1 receptor density in the medial prefrontal cortex (mPFC), indicating that prefrontal CB1 receptors may contribute to THC-induced alterations in effortful decision making. These findings complement those of Khani et al. (2015), where CB1receptor agonism in the ACC decreases physical effort-based decision making. Interestingly, and unlike the effects of THC, CB1 agonism by administration of the synthetic cannabinoid WIN 55, 212-2 does not shift choice on the rCET (Silveira et al, 2016). These discrepancies are likely related to the distinct pharmacodynamic profiles of these drugs, and/or differences in the intracellular signalling pathways they recruit, with THC a potent recruiter of the arrestin-2 pathway, and WIN 55, 212-2 a recruiter of the classical G-protein Gαi/o and Gβγ pathways (Laprairie et al, 2014). These data emphasize that not all forms of cannabinoid receptor activation produce converging effects, and animal studies aiming to model the psychoactive effects of cannabis would do well to administer THC in lieu of its synthetic counterparts. In contrast, modulating endogenous cannabinoid tone, via CB1 receptor inverse agonism with rimonabant, CB2 antagonism with AM 630, or by inhibi	  tion of endogenous anandamide hydrolysis by FAAH with URB 597, does not affect choice on the rCET (Silveira et al, 2016). Together, this work suggests that endocannabinoid signaling does not tonically regulate effort-based decision     47 making with cognitive effort costs, but that endogenous activation of CB1 receptors may temper selection of otherwise lurcrative, albeit cognitively demanding challenges (Silveira et al, 2017).    1.8.5 Adenosine Systems  An impressive body of work by Salamone and Colleagues has demonstrated an interesting role for the purine nucleoside adenosine in the allocation of physical effort. Systemic administration of the A2 adenosine receptor antagonists MSX-3 (2 mg/kg) and KW6002 (0.5 mg/kg), the A1 adenosine receptor antagonist DPCPX (0.75 mg/kg), and the nonspecific adenosine receptor antagonist caffeine (20 mg/kg) have no effect on response allocation in the (FR5)/chow concurrent choice task (Farrar et al, 2007; Salamone et al, 2009b). Similarly MSX-3 and DPCPX do not alter selection of the HR option in the T-maze effort task (Mott et al, 2009), but MSX-3 (2mg.kg) does increase lever pressing and decreases chow consumption in the (PROG)/chow concurrent choice task (Randall et al, 2012). A2 receptor knockout mice also display a similar preference for the HR arm relative to wild types in the effort T-maze task (Pardo et al, 2012). Taken at face value, these studies suggest that under baseline conditions, adenosine signaling does not play a major role in the valuations guiding choice with physical effort costs.  However, adenosine does play a critical role in mediating the effects of dopamine signaling on physical effort allocation. The decline in lever pressing     48 and increased chow consumption observed following haloperidol (0.1 mg/kg) administration in the (FR5)/chow concurrent choice task is attenuated by the A2 antagonists MSX-3 and KW6002 (0.125, 0.25, 0.5 mg/kg), and by the general adenosine antagonist caffeine (10, 20 mg/kg) (Salamone et al, 2009b). MSX-3 also attenuates the haloperidol-induced shifts to the LR arm in both rats (Mott et al, 2009) and mice (Pardo et al, 2012) on the T-maze. Similarly, choice deficits induced by the VMAT inhibitor tetrabenazine (0.75 mg/kg) in the effort T-maze and both the (FR5) and (PROG) concurrent choice tasks are reversed by MSX-3 (0.5, 1, 2 mg/kg) (Nunes et al, 2013b; Randall et al, 2014; Yohn et al, 2015b). These effects appear to be mediated by A2 and D2 receptor interactions, as MSX-3 (0.5, 1, 2 mg/kg) and KW6002 (0.25, 0.5 mg/kg) can fully attenuate the shift from lever pressing to chow consumption induced by the D2 receptor antagonist eticlopride (0.8 mg/kg), but MSX-3 only partially attenuates the choice impairments produced by the D1 antagonist SCH 39166 (0.2 mg/kg) (Nunes et al, 2010; Worden et al, 2009). In contrast, the A1 receptor antagonist DPCPX does not reverse haloperidol’s effects in the concurrent choice or T-maze tasks (Mott et al, 2009; Salamone et al, 2009b), and choice impairments produced by D1 receptor antagonist ecicopam (.2 mg/kg) are not reversed by the A1 receptor antagonists DPCPX or CPT (Nunes et al, 2010). As further evidence of A2-D2 interactions guiding effort allocation, A2 receptor knockout mice are insensitive to the effects of haloperidol (0.05, 0.1 mg/kg) on effort T-maze performance (Pardo et al, 2012). Notably, A2 adeosine receptors are abundant in striatal zones, including     49 the nucleus accumbens, and are colocalized with D2 receptors on medium spiny neurons (Jarvis and Williams, 1989). Stimulation of either receptor has opposing effects on cAMP signal transduction pathways, thus providing a potential mechanism by which A2 receptors may mediate the effects of mesoaccumbens DA signaling on effort-based choice. In line with this, intra-NAcc infusions of the A2 agonist GS21680 (24 ng/ .5 ul per side) mimic the effects of intra-NAcc DA antagonism on the (FR5)/chow concurrent choice task (Font et al, 2008), and systemic (1, 2 mg/kg) or intra-NAcc (1.25, 2.5, 5 ug/side) administration of MSX-3 blocks the shift from lever pressing to chow produced by intra-accumbal eticlopride infusion (2,4, ug/side) (Farrar et al, 2010). Collectively, these studies overwhelmingly suggest that A2 receptors in the nucleus accumbens may mediate the effects of dopamine signaling on effort-based choice, via interactions with colocalized D2 receptors in this region.  Like its effects on the physical effort tasks, adenosine receptor antagonism on the rCET does not have a pronounced role in cognitive effort allocation. Administration of caffeine does not affect HR choice across subjects, but has mild effects in worker rats, resulting in a trending decrease in HR choice at higher doses (10, 20 mg/kg) (Cocker et al, 2012b). Given that DA signaling does not play as prominent a role in rCET decision making (Hosking et al, 2015b), A2-D2 interactions in the nucleus accumbens are less likely to regulate effort in the cognitive domain, although this is an exciting avenue for future study.    1.8.6 Serotonergic Systems     50  In contrast to the neurotransmitter systems described earlier, serotonin does not appear to play a role in the mechanisms guiding physical effort allocation. In both the FR5 and PROG concurrent choice tasks, the SERT inhibitor fluoxetine decreases lever pressing (2.5, 5.0, 10.0 mg/kg) and either has no effect, or decreases consumption of freely available chow (Yohn et al, 2016b, 2016c). Denk et al., (2005) directly compared serotonergic contributions to delay and physical effort discounting, by administering systemic pCPA (300 mg/kg; 2 injections, administered 48h and 24h before testing) prior to testing on a T-maze in which the costs to HR were either physical (a 30 cm scalable barrier) or delay-based (15s). This dosing regimen reduces serotonin levels and its metabolite, 5-HT1AA, by more than 85% in the frontal cortex and hippocampus, for up to seven days (Hajós et al, 1998; Jakala et al, 1992). Serotonergic depletion increases delay discounting, causing more animals to opt for the small, immediate reward, but has no effect on physical effort allocation (Denk et al, 2005). These results are in accordance with the prominent role of the serotonergic system in mediating decisions with delay costs (Winstanley et al, 2003a, 2004a, 2004b) . Currently no work has investigated serotonergic contributions to cognitive effort-based decision making. The neurotransmitter does not figure into current theories of cognitive effort allocation, and no drugs targeting serotonin have been investigated on the rCET (although see Appendix 1 for a summary of my recent attempts to study the role of serotonin in decision making with cognitive effort costs).       51  1.8.7 Other Neurotransmitter Systems: Glutamate and GABA  Few investigations have directly assessed glutamatergic and GABAergic signaling in the domain of effort-based decision making, but they will be briefly mentioned here for the sake of completeness. In an operant-based lever pressing task, the NMDA antagonist ketamine (5, 10 mg/kg) increases effort discounting across blocks with increasing effort requirements (Floresco et al, 2008b). However, when tested in an equivalent task with delays to reward receipt equalized across the low-effort and high-effort options, ketamine no longer results in effort aversion. Indeed, it appears that ketamine may specifically increase delay discounting, which explains why an effect is observed in the original task where the increased response requirements necessarily increase the time it take for rats to receive the larger reward (Floresco et al, 2008b). Interestingly, a recent study using a knockout mouse line demonstrated that glutamatergic transmission onto midbrain DA neurons is necessary for mice to lever press on a FR5 schedule for preferred sugary reward, instead of opting for freely available chow (Hutchison et al, 2017). This work highlights the complex neurotransmitter interactions that guide effort allocation, which to date have not received much attention.  In the concurrent choice task, the barbiturate sodium pentobarbital (10 mg.kg) decreases lever pressing and increases consumption of freely available chow, while a higher dose (15 mg.kg) decreases both lever pressing and chow consumption, likely reflecting sedative effects (Cousins et al, 1994). In contrast,     52 ethanol does not affect rats’ willingness to exert cognitive effort for larger rewards on the rCET (Cocker et al, 2012b). Thus, drugs primarily acting to increase GABA in the brain do not necessarily decrease willingness to work for reward, although more work is needed in this area.    1.9 Individual Differences and Physiological Factors Mediating Effort Allocation  1.9.1 Individual and Sex Differences  The study of individual differences has been incorporated into a number of models. This work has demonstrated that basal willingness to exert effort is modulated by a number of neurotransmitters systems and brain areas. Randall and colleagues were the first to study individual differences as they relate to physical effort allocation, by using a median split to classify rats as high or low responders based on lever presses in the (PROG)/chow concurrent choice task (Randall et al, 2012). Notably, the progressive ratio variant of the concurrent choice task generates high response variability, with some rats pressing the lever less than 100 times for sucrose and instead preferring the freely available chow, while others press the lever more than 1000 times and consume small amounts of chow (Randall et al, 2012). High responders are more sensitive to the suppressive effects of haloperidol on lever pressing for preferred sucrose, and show increased pDARPP-32(Thr34) expression in the core region of the nucleus accumbens (Randall et al, 2012).      53 Likewise, in the cognitive effort domain, the rCET generates high variability in baseline willingness to exert attentional effort for lucrative rewards. In a typical rCET study, baseline choice of the high reward (HR) option falls within 65-70% across rats. Rats are categorized as “workers” if they choose HR on more than 70% of trials, while rats whose choice of HR is less than 70% are termed “slackers”. This subdivision is based on the mean split from the original rCET paper (Cocker et al, 2012), and has been used in subsequent studies to aid comparison of individual differences across studies using the task. Individual propensities to “work” or “slack” at baseline have been shown to mediate the effects of a number of pharmacological challenges (e.g. amphetamine (section 1.8.1), nicotine (section 1.8.3), as well as the effects of regional inactivations (e.g. the basolateral amygdala (section 1.10.3)).   Overwhelmingly, animal models of cost/benefit decision making have used male rodents as subjects. Only recently have attempts been made to assess females on tasks in which effort costs are manipulated. In the operant physical effort discounting task, baseline choice of the high effort, high reward (HR) lever does not fluctuate with the female estrous cycle, but ovariectomy reduces effort discounting (Uban et al, 2012). In contrast, estradiol (10ug) decreases choice of the HR lever 24 hours post-injection, suggesting that ovarian hormones appear to bias choice towards smaller, albeit more accessible rewards. Notably, co-administration of the selective estrogen receptor (ER)-α agonist PPT (10 ug) and the ERβ agonist DRN (10 ug) appear to mimic the effects of estradiol on effort discounting, but paradoxically each ligand administered alone increases effortful     54 choice (Uban et al, 2012). Estrous cycling has not been investigated on the rCET, but baseline choice approximates those of male subjects (personal observation). However, in line with previous 5-CSRTT literature, female rats generally display lower accuracy rates (Bayless et al, 2012), providing further evidence that willingness to engage in cognitive effort and cognitive performance itself are dissociable.    1.9.2 Stress, Inflammation, and Fatigue   A number of investigators have used animal effort models to probe the physiological processes, such as stress and inflammation, which may mediate valuations of effort. These have received due consideration because disorders characterized by anergia, such as depression and schizophrenia, are also associated with stress and biomarkers of inflammation (Felger and Treadway, 2017; Monroe and Harkness, 2005). In the operant physical effort-discounting task developed by Floresco and Colleagues (2008), one hour of acute restraint stress, but not systemic administration of corticosterone (1, 3 mg.kg), dramatically decreases choice of the HR option (Shafiei et al, 2012). This effort discounting is observed across all trial blocks with varying response costs (2, 5, 10, 20 lever presses), and is attenuated by the corticotrophin-releasing factor antagonist (CRF) alpha-helical CRF (Bryce and Floresco, 2016). Indeed, a similar effort discounting is observed following intra-ventricular or intra-ventral tegmental area (VTA) infusions of CRF, suggesting that the effects of acute stress are mediated by CRF-induced perturbations in DA transmission (Bryce     55 and Floresco, 2016). Similarly, systemic administration of the pro-inflammatory cytokines interleukin (IL) -1β (2, 4 mg/kg) and IL-6 (4,6,8 mg/kg) decrease lever pressing and increase chow consumption in the FR(5)/chow concurrent choice task, suggesting that like acute stress, inflammation reduces willingness to exert physical effort for lucrative rewards (Nunes et al, 2014; Yohn et al, 2016a). This is in agreement with recent human work showing that systemic administration of the bacterial endotoxin E. coli lipopolysaccharide (LPS) transiently increases IL-6 and decreases acceptance of high effort options in an accept/reject choice paradigm (Draper et al, 2017). Paradoxically, another human study found LPS-induced sleepiness increases high effort choice (Lasselin et al, 2017), in line with the notion that sickness might not generally decrease motivation, but may lead to a reorganization of priorities where individuals become more discerning in their effort allocation; at least one study in mice supports this (Vichaya et al, 2014).  Another physiological state which has received due consideration is fatigue. Until recently, it was unknown whether metabolic states produced by fatigue are factored in the cost-benefit analysis. Mice fatigued by running on a treadmill at 80% of their peak workload show a sharp choice reversal from HR to LR arms (Iodice et al, 2017). These effects are temporary and specific, as subjects resume selection of the HR option when tested at 60% peak workload or when the efforts costs in either arm are equated (Iodice et al, 2017). This work suggests that levels of fatigue are dynamically factored into the cost/benefit analysis, presumably by magnifying the cost of effort in relation to the benefit of expected reward.      56  1.10 Functional Neural Systems Mediating Effort-Based Decision Making  Animal models haven proven invaluable towards identifying the neural circuits guiding effort-based cost/benefit decision making. This research has generally implicated a cortical-limbic-striatal circuit in guiding choice processes. Below I review three regions that have been assessed for their contribution to physical and cognitive effort allocation: the prefrontal cortex, the striatum, and the basolateral amygdala. Most of these experiments have assessed performance on-task following irreversible lesion of a brain area, typically with quinolinic or ibotinic acid.  Other studies have relied on temporary inactivation via intracerebral drug infusion, most commonly achieved via gaba agonism, but lidocaine has also been used. Recently, researchers have also begun using chemogenetic and optogenetic techniques to address similar questions, albeit with greater specificity in select cases.   1.10.1 Prefrontal Cortices The utility of animal models to investigate the processes underlying effort-based decision making rests on the circuit homology between rats and primates. This becomes especially pertinent when comparing cortical regions across species, which show great variation. Indeed, the rat cortex is composed exclusively of agranular cortex. Subsequent studies have established that orbital and medial prefrontal areas are similar in relative position and connectivity     57 across species, suggesting these areas are comparable (Heidbreder and Groenewegen, 2003; Öngür and Price, 2000). The infralimbic cortex (IL) of rodents is most analogous to area 25 of the primate brain, based on their similar connectivity to the striatal “emotional processing network”, consisting of three well-conserved and easily identifiable striatal regions: the nucleus accumbens shell, hippocampal-striatal projections, and amygdala-striatal projections (Heilbronner et al, 2016) . Based on this shared corticostriatal projection, ventrolateral areas of the rodent PFC (medial orbital and ventrolateral orbital) correspond to primate OFC (medial orbital and centrolateral orbital), whereas the dorsomedial wall of the rodent PFC (prelimbic, ventral Cg1) is most akin to areas of the primate ACC (32, 24). Notably, corticostriatal projections across species share a dual pattern of focal and diffuse projections, and follow a similar topographic organization across limbic, associative, and sensorimotor striatal zones (Haber, 2016; Mailly et al, 2013). While the absence of a granular zone in rodent cortex makes it difficult to establish which areas in the rat brain are homologous to the primate dorsolateral prefrontal cortices, this region has been compared to rodent PL cortex based on functional similarities (Dalley et al, 2004; Granon and Poucet, 2000; Vertes, 2004). Thus circuit and functional homology across species suggest that questions of effort allocation are amenable to analysis with rodent models. Highlighted here are three areas that have been investigated in regard to effort allocation: the orbitofrontal cortex, the medial prefrontal cortex (encompassing prelimbic and infralimbic regions), and the anterior cingulate cortex.      58  1.10.1.1 Orbitofrontal Cortex The available evidence suggests that the orbitofrontal cortices do not play a critical role in the processes guiding physical effort allocation. Lesions or inactivations of this area do not affect rats’ propensity to scale a barrier for large reward, but do affect their ability to withstand delays for larger reward (Mobini et al, 2002; Rudebeck et al, 2006; Winstanley, 2004; Zeeb et al, 2010). This is in stark contrast to lesions of the anterior cingulate, which bias choice away from the effortful T-maze option, but which also leave delay-based decision making intact (Rudebeck et al, 2006). In line with this, intra-OFC infusions of the potent CB1 receptor agonist ACEA do not affect choice in the T maze effort task, but drastically reduce high effort, high reward choice when infused into the ACC (Khani et al, 2015). Such work is in agreement with the general assertion that dorsal and ventral areas of the prefrontal cortex appear to regulate different forms of decision making, based on the cost inherent to the choice at hand. While lesions or inactivations of the anterior cingulate cortex affect physical or cognitive effort-based valuations (Croxson et al, 2009; Hosking et al, 2014a; Rudebeck et al, 2006; Walton et al, 2003b), similar manipulations to the orbital frontal regions selectively affect decision making involving delay, risk, and probability costs (Barrus et al, 2017; Orsini et al, 2015a; Stopper et al, 2014; Winstanley, 2004; Winstanley and Floresco, 2016). Notably, dissociable roles for dorsal and ventral prefrontal areas in goal-directed choice are also observed in the human literature (Prevost et al, 2010; Rangel and Hare, 2010; Rushworth et     59 al, 2011a). Viewed another way, the OFC can be said to mediate decisions in which costs are inherent to the outcome (i.e. delay, risk), while costs incurred on the individual (i.e. physical, cognitive effort) appear to be mediated by dorsal prefrontal regions. In support of this distinction, lesions to either the OFC or ACC impair stimulus or action value learning, respectively (Camille et al, 2011; Rudebeck et al, 2008), and neurons in either area appear to preferentially encode information about stimulus or action value  (Hayden and Platt, 2010; Padoa-Schioppa and Assad, 2006). Notably, one study has found increased metabolic activity in the rodent left orbitofrontal cortex following performance of a physical effort-discounting task, in addition to increased activity in the left ACC and PL cortex, and decreased metabolic activity in the IL region (Endepols et al, 2010). Thus, it appears that decisions requiring an assessment of physical effort investment for reward evoke metabolic changes in a variety of prefrontal regions, but some areas, such as the OFC, play a less critical role.  To date, OFC contributions to decision making with cognitive effort costs has not been empirically investigated. While the physical effort literature suggests the OFC plays a nominal role in effort discounting, work with the rCET has demonstrated that the two forms of effort recruit overlapping, yet distinct, areas of the PFC (such as the medial prefrontal cortex, discussed below in section 1.10.1.1.2). Chapter 4 addresses this gap in the cognitive effort literature by investigating the effects of OFC inactivation on rCET performance.   1.10.1.2 Medial Prefrontal Cortex     60  Lesions to the medial prefrontal cortex (mPFC), encompassing prelimbic (PL) and infralimbic (IL) regions, do not affect choice behaviour on the T-maze effort task, suggesting a minimal role for this prefrontal area in the processes guiding physical effort allocation (Walton et al, 2003b). Notably, one study did find that unilateral lesion of the mPFC decreases HR choice in the physical T-maze task, effects which are greatest when rats have to choose to put in more work for reward when it is on the contralateral side of space to the lesion (Croxson et al, 2014). These lesions encompassed the ACC, and so the authors suggest the ACC may have a lateralised role in evaluating the costs and benefits of actions based on their spatial location. However, another study did find increased and decreased metabolic activity in the PL and IL cortex, respectively, in rats following performance of an operant lever pressing choice task (Endepols et al, 2010), and using a similar task lower D2 mRNA expression in the IL cortex is associated with greater willingness to exert effort for reward (Simon et al, 2013). It is possible these discrepancies are related to the parameters of effort (maze running versus repeated lever pressing) under investigation, but this has not been systematically studied.   In contrast, temporary inactivations of the mPFC affect rats’ willingness to exert cognitive effort, such that selectively targeting the PL or IL reduces choice of the HR option in the rCET (Hosking et al, 2015a). Notably, these reductions in effortful choice are accompanied by distinct performance deficits: inactivations of the PL cortex impair attentional accuracy and response omissions (an indirect index of attention), while inactivations of IL cortex spare accuracy but robustly     61 increase impulsive action as measured by rates of premature responding (Hosking et al, 2015a). Although it is difficult to interpret the directionality of these deficits, it is possible that with attentional and response control processes impaired, rats shift choice towards a low cost (i.e LR) strategy. Another possibility is that different networks in the mPFC simultaneously mediate decision making, attention, and impulsivity, and so inactivations of the mPFC impair these independent processes. Indeed, the involvement of the mPFC in the rCET reflects the demands of the task at hand, given the medial wall of the frontal cortex plays a critical role in regulating attention and impulsivity in the standard 5-CSRTT (Chudasama et al, 2003b; Muir et al, 1996; Passetti, 2002). These processes are likely mediated via serial flow of information to the striatum, and it is possible different corticostriatal “loops” regulate distinct facets of task performance (i.e., choice, attention, response inhibition), the resulting output of which can then bias motor output via palladal-thalamic-cortical pathways (Chudasama and Robbins, 2006; Haber and Behrens, 2014).   1.10.1.3 Anterior Cingulate Cortex   As reviewed in section 1.4, a number of imaging studies have implicated the anterior cingulate cortex (ACC) in valuations of effort. Individuals with localized lesions of the ACC demonstrate impaired performance on tests of real world decision making and are unable to report conscious feelings of effort (Naccache et al, 2005). In contrast, direct stimulation of the ACC can elicit a feeling of “gearing up” for an effortful challenge (Parvizi et al, 2013). Individual     62 neurons in the ACC are sensitive to changes in reward magnitude (Akkal et al, 2002; Shima, 1998), and relative to other regions of the PFC, the ACC has a high  proportion of neurons encoding the decision parameters (i.e. payoff, probability, physical effort cost) necessary for valuations of effort expenditure for lucrative outcomes (Cowen et al, 2012; Hillman and Bilkey, 2010; Kennerley et al, 2011). Individual ACC neurons have also been shown to encode progression through a series of steps toward reward, effectively monitoring the “effort” exerted towards a goal (Shidara & Richmond, 2002). Accordingly, lesions of the ACC decrease rats’ willingness to scale a barrier for larger reward, but have no effect when costs for the low and hard reward alternatives are equated (Rudebeck et al, 2006; Walton et al, 2002, 2003b). Lesions to the ACC also decrease lever pressing on a variant of the (PROG)/chow concurrent choice task, and decrease lever pressing for the HR option in an operant decision-making task (Hart et al, 2017; Walton et al, 2009). These effects on effortful choice are likely mediated via interactions with the nucleus accumbens and basolateral amygdala, as functional disconnections of the ACC and either region decreases rats’ willingness to scale the effortful barrier in the T-maze effort task (Floresco and Ghods-Sharifi, 2007; Hauber and Sommer, 2009) Collectively, this work suggests that the ACC is part of a larger cortico-limbic-striatal circuitry regulating allocation of physical effort.   However, a closer inspection of the literature suggests that the ACC does not play a ubiquitous role in decision making, especially when different manifestations of physical effort are considered. While a recent study found an     63 effect of NMDA ACC lesions on lever pressing in a progressive ratio/concurrent chow choice task (Hart et al, 2017), an earlier study found no effects in a similar task following quinolinic acid lesions to the ACC (Schweimer, 2005). Notably, the latter study reporting no effects trained rats on the concurrent choice task after being trained on the effort T-maze. Thus it is possible that this previous experience improved performance in the subsequent decision-making task, or enough time elapsed for other brain regions to compensate for the extent of the lesion (Hart et al, 2017). Notably, the decision-making deficit observed following ACC lesions on the effort T-maze task is also ameliorated if rats receive previous training with response costs equated on each arm (Rudebeck et al, 2006), perhaps suggesting that other areas (i.e. NAc, BLA) may also regulate effort allocation when cost-benefit contingencies are updated. In contrast, lesions to the ACC do not affect decision making when effort is operationalized as weighted lever pressing (Holec et al, 2014). Clearly, more work is needed to determine under what conditions the ACC is recruited in valuations of physical effort, as the available evidence indicates that it plays a more nuanced role than current consensus would suggest (Winstanley and Floresco, 2016).   The ACC has also been implicated in decision making with cognitive effort costs. Indeed, objective measures of task difficulty correlate with activity in the ACC, while lesions to this region can abolish conscious feelings of mental effort (McGuire and Botvinick, 2010a; Naccache et al, 2005). While the ACC has been ascribed a number of cognitive functions including error detection (MacDonald, 2000), conflict monitoring (Botvinick et al, 1999), and action selection (Posner et     64 al, 1988), a recent proposal suggests that these may all reflect a general role of the ACC in cognitive effort allocation, and specifically in computing the expected value of cognitive control (Shenhav et al, 2013). Under this proposal, the decision to engage in a cognitive episode and the intensity of this engagement is based on the expected value of the goal it corresponds to, which is discounted by the costs required to maintain on task (Shenhav et al, 2013; Westbrook A1, 2016). If expected value is high, ACC signals for increased cognitive control, likely mediated by adjacent dorsolateral prefrontal cortices and their interconnections with the striatum via corticostriatal loops (Frank et al, 2001; Haber and Behrens, 2014; O’Reilly et al, 2014). In line with this proposal, temporary inactivation of the ACC decreases choice of the high-effort, high-reward option on the rCET (Hosking et al, 2014a).     1.10.2 Striatum  The ventral striatum (VS) has figured prominently in most, if not all, models of cost/benefit decision making, which is perhaps unsurprising given the region is the nexus at which cortical, limbic, and midbrain dopaminergic afferents – regions critical in reward and cognitive processes- can influence behaviour (Mogenson et al, 1980; Sesack and Grace, 2010). A major focus of the physical effort literature has been on mesolimbic dopamine in the nucleus accumbens (reviewed in section 1.8.1.2), but lesions, as well as temporary inactivation, have been also been carried out. Lesions to the nucleus accumbens (NAcc) reduce rats’ willingness to choose the HR arm in the T-maze effort task, an effect which     65 is reproduced following contralateral disconnection of the ACC and NAcc (Hauber and Sommer, 2009). This work suggests that ACC signals relayed to the NAcc may be a means to bias motor output towards more lucrative, albeit costly, options; a human imaging study using a physical effort task supports this interpretation (Croxson et al, 2009). The core region of the NAcc appears to be particularly involved in effort discounting, as temporary inactivation of this region, but not the lateral shell, reduces choice of the high-effort lever in the operant effort discounting task (Ghods-Sharifi and Floresco, 2010). In contrast, the dorsal striatum does not appear to play a role in decision making with physical effort costs. While no study has temporarily inactivated or lesioned the area, ventral striatal manipulations that affect valuations of physical effort allocation (e.g. intra-accumbens adenosine, acetylcholine, or dopamine neurotransmission manipulation) are ineffective when targeted toward a dorsal striatal site, suggesting that the contributions of this striatal zone to decision making with physical effort costs is minimal (Cousins et al, 1993; Farrar et al, 2010; Font et al, 2008; Nunes et al, 2013a).  In contrast, the role of the striatum has not yet been probed on the rCET. Given its role as an integrator of cortical and limibic inputs, it is likely this area is involved in rCET performance, as cortical and amygaldar regions guide choice on the rCET (Hosking et al, 2014a, 2015a). Indeed, a human imaging study has shown that NAcc activity is inversely related to cognitive demand, and this is preceded by coordinated activity in the ACC (Botvinick et al, 2009). Unlike the physical effort literature, however, there is reason to believe that cognitive effort     66 allocation may recruit more dorsal regions of the striatum.  Indeed, human imaging data suggests that while the ventral striatum may be a common motivational node, cognitive and physical effort may preferentially recruit the dorsomedial and dorsolateral striatum, respectively (Schmidt et al, 2012). The dorsal striatum may play a particularly important role in decisions based on cognitive effort costs, given widespread areas of the cortex project to this “associative” area, and aberrant signaling in this region can induce deficits in a number of cognitive processes, such as attention, working memory, and behavioural flexibility across species (Bradfield et al, 2013; Cools et al, 2006; Ragozzino, 2007; Rogers et al, 2001). Chapter 5 assesses dorsal and ventral striatal contributions to cognitive effort allocation as probed by the rCET.    1.10.3 Basolateral Amygdala The amygdala is traditionally viewed as a critical substrate for appetitive and aversive learning, but contemporary view of amygdalar function emphasizes its role in acquiring, monitoring, and updating value (Seymour and Dolan, 2008; Wassum and Izquierdo, 2015). Neurons in the primate amygdala track the positive and negative value of both conditioned and unconditioned stimuli, and single neurons in the human amygdala respond linearly to the monetary value assigned to individual items (Belova et al, 2007, 2008; Jenison et al, 2011; Paton et al, 2006). Additionally, neurons in the rodent amygdala anticipate reward encounter, respond during reward consumption, and differentiate between low and high reward magnitude (Pratt and Mizumori, 1998). Interestingly the rodent,     67 primate, and human studies described above recorded from neurons in the basolateral nucleus of the amygdala (BLA), an area that appears to make unique contributions to decision making with effort costs. Inactivation of the BLA decreases choice of the effortful HR option in the operant effort discounting task, an effect which persists even when the delays on the easy lever are matched to those necessarily incurred when fulfilling the response requirements on the hard lever (Ghods-Sharifi et al, 2009a). Similarly BLA inactivation, during initial learning or following stable baseline, decreases effortful choice in a variant of the (PROG)/chow concurrent choice task, but has no effect on lever pressing for sucrose when chow is no longer concurrently available (Hart and Izquierdo, 2017). This suggests that BLA inactivations leave primary motivational processes intact, but impair the valuations which bias choice towards the lucrative and effortful option (Hart et al, 2017). Notably, the BLA appears to guide physical effort-based choice via interactions with the ACC, as contralateral disconnection of these two areas decrease choice of the HR option in the T-maze effort task (Floresco and Ghods-Sharifi, 2007)  Similarly, the BLA appears to plays a critical, albeit more nuanced role, in the mechanisms guiding cognitive effort allocation. Whereas inactivation of the BLA in physical effort paradigms systematically decreases choice of the HR option, the effects of BLA inactivation on the rCET are dependent on rats’ propensity to “work” or “slack” at baseline (Hosking et al, 2014a). Indeed, rats initially preferring the hard option at baseline decrease selection of this option, whereas rats opting for the easier of the two options at baseline now increase     68 hard lever choice following BLA inactivation (Hosking et al, 2014a). These nuances in BLA contributions to effortful choice are particularly interesting given recent imaging work in humans suggesting that activity in the right amygdala appears to process subjective value uniquely for rewards associated with cognitive, but not physical effort costs (Chong et al, 2017). In contrast to the amygdalostriatal circuitry that appears to guide physical effort-based decision making (Floresco and Ghods-Sharifi, 2007), this pathway has not been probed on the rCET. The goal of Chapter 6 is to assess whether communication between the ACC and BLA regulates cognitive effort allocation on the rCET.    1.11 Comparison of the Neurotransmitter and Neural Systems Guiding Physical and Cognitive Effort Allocation and Rationale for the Current Experiments  The available literature suggests that decision making characterized by physical or cognitive effort costs is subserved by overlapping, yet distinct, neural mechanisms. For example, cannabinoid signaling appears to play a similar role across domains of effort, while the physical effort literature has established a critical role for DA and adenosine signaling in such valuations. In contrast, DA and adeonsine do not appear to affect decision making with cognitive effort costs, but instead cholinergic signaling has pronounced effects on effort allocation in this realm. In reference to brain areas mediating effort-based     69 decision making, the ACC appears to bias choice towards HR options in either the physical or cognitive space, while the prelimbic and infralimbic areas are preferentially involved in cognitive, but not physical, effort. And while the BLA has been implicated in both forms of decision making, in the cognitive effort literature these effects interact with individual willingness to “work” or “slack” at baseline.   However, a complete understanding of how physical and cognitive effort allocation compare is limited because only recently have efforts been made to model cognitive effort costs in animals. To address this, I will carry out a series of studies in which I use the rCET to probe how distinct brain regions and neurotransmitter systems regulate cognitive effort allocation. The neurotransmitters and regions up for investigation are those that have been previously assessed in the physical effort domain, but for which no data exists in regard to cognitive effort. The results of these experiments, paired with previous work on the rCET, will yield a point-to-point comparison of the major neurotransmitter and neural systems guiding effort-based choice. A summary of these experiments is described below.    Experiment 1 (Chapter 3) will investigate cholinergic contributions to decision making on the rCET, using chemogenetics and a transgenic CHAT cre recombinase-driven rat line. We have previously shown that systemic administration of cholinergic drugs affects rats’ willingness to exert cognitive effort, and that these effects are dissociable from the role of acetylcholine in regulating attention (Hosking et al, 2014c). However, it is currently unknown     70 which cholinergic neuron population is responsible for regulating decision making with cognitive effort costs. To address this, I will selectively transfect cholinergic basal forebrain neurons with inhibitory Designer Receptors Exclusively Activated by Designer Drugs (DREADDs). Prior to task onset, I will administer clozapine-n-oxide, which will selectively bind to these DREADDs and silence transfected neurons (Roth, 2016). I focus on cholinergic BF neurons in this experiment, given this is the main source of cortical cholinergic tone, and we have previously implicated medial prefrontal cortical areas in regulating the willingness to exert cognitive effort (Hosking et al, 2014a, 2015a).   In experiment 2 (Chapter 4) I bilaterally inactivate dorsal and ventral parts of the striatum to assess their contributions to decision making with cognitive effort costs. The ventral striatum plays a critical role in decision making with physical effort costs, but it is unknown how this area contributes to effort allocation in the cognitive domain (Ghods-Sharifi and Floresco, 2010; Hauber and Sommer, 2009). Imaging work suggests that NAc activity in anticipation of reward is degraded by a working memory challenge, but the causal role of this region to cognitive effort has not been identified (Botvinick et al, 2009). In contrast, regions dorsal to the ventral striatum have not received as much attention in the cost/benefit decision-making literature, although it generally appears that this region is not involved in physical effort allocation (Cousins et al, 1993; Farrar et al, 2008; Nunes et al, 2013a). Indeed, while ventral striatal activity has been shown to track expected reward, as well as predict behavioral     71 performance in tasks requiring both cognitive and physical effort, the ventral striatum is only co-activated with the dorsomedial striatum when cognitive, but not physical, task demands are high (Schmidt et al, 2012) Connectivity also suggests that the dorsal striatum may play a particularly important role in cognitive-effort based decision making, given widespread areas of the cortex project to this area, and aberrant signaling in this region can induce deficits in attention, working memory, and behavioural flexibility across species (Balleine et al, 2007; Bradfield et al, 2013; Cools et al, 2006; Ragozzino, 2007; Rogers et al, 2001). Neurons in dorsal striatal regions display task-related activity that is graded by reward magnitude, and neurons in this region appear to encode value signals necessary to guide choice during decision-making tasks (Cromwell, 2003; Hollerman et al, 1998; Kawagoe et al, 1998; Kim et al, 2013; Kimchi and Laubach, 2009; Lau and Glimcher, 2007, 2008). Thus I predict that temporarily silencing both striatal zones will decrease willingness to exert cognitive effort for preferred outcomes.   The purpose of experiment 3 (Chapter 5) is to investigate orbitofrontal cortex contributions to decision making on the rCET via bilateral inactivation. To date, dorsal and medial regions of the prefrontal cortex have been investigated on the rCET, while the role of orbitofrontal regions remain unknown. As discussed in section 1.10.1.1, the OFC generally mediates decisions in which costs are inherent to the outcome, such as those observed in delay or risk-based decision making (Rushworth et al, 2011b). This can be     72 dissociated from dorsal prefrontal regions, which are generally recruited when decision costs are incurred on the individual (physical, cognitive effort) (Rangel and Hare, 2010; Rudebeck et al, 2006). However, recent rodent work suggests that the LOFC may play a more nuanced role in decision making, and is particularly involved when task parameters make the objectively best option unclear, such that subjective evaluation of different outcomes is more pronounced (Winstanley and Floresco, 2016). This will be a particularly interesting area to probe, given lesions of the OFC do not affect attentional processes, and only mildly affect response control on the 5-CSRTT, unlike dorsal PFC regions (Chudasama et al, 2003b; Muir et al, 1996; Passetti, 2002). The OFC also sends projections to dorsal striatal areas implicated on the rCET (see Chapter 4), where signals from different frontal areas could be integrated via diffuse and focal projections (Mailly et al, 2013; Silveira MM, Tremblay M, 2016). We will focus on the ventral lateral orbital frontal cortex (VLO), given most of the decision-making literature has focused on this area, because neurons in this region encode expected outcomes, and because relative to the MO, the VLO sends more projections to the ACC – a cortical area known to be involved in rCET performance specifically and effortful decision making generally (Hoover and Vertes, 2011; Schoenbaum et al, 2003).    Lastly, Experiment 4 (Chapter 6) uses a contralateral ACC-BLA disconnection procedure to determine how bidirectional amygdalocortical communication regulates decisions of cognitive effort allocation. The BLA     73 appears to play a unique role on the rCET, whereby inactivations of this region decrease choice of subjects’ generally preferred option (Hosking et al, 2014a). This finding corroborates recent imaging work in humans where the BLA encodes the subjective value of choices associated with cognitive, but not physical effort costs (Chong et al, 2017). Critically, the effect of BLA inactivations in models of cost/benefit decision making appears to depend on the nature of the costs and their severity specifically (Winstanley and Floresco, 2016). If costs are very aversive, such as signaled punishing timeouts or electric shock, the BLA appears to attenuate baseline choice of the more aversive option (Orsini et al, 2015a; Zeeb and Winstanley, 2011). In contrast, when rewards are especially salient relative to costs, such as omission of reward receipt, delays, or exerting physical effort, then the BLA drives choice towards more rewarding options (Ghods-Sharifi et al, 2009a). This may explain the divergent effects of BLA inactivation on choice in the rCET, in that workers decrease choice of the high effort, high reward (HR) option HR, whereas slackers increase HR choice (Hosking et al, 2014a). Indeed, this may suggest that the rewards associated with successful trial completion are more salient in workers, but that cognitive costs are especially aversive in slacker rats. BLA-ACC circuitry promotes effortful choice in a decision task involving physical effort costs, and we hypothesize the BLA similarly provides the ACC with value-related information needed to guide cognitive effort allocation on the rCET (Floresco and Ghods-Sharifi, 2007).     74 Chapter 2: General Methods    2.1 Subjects Unless otherwise specified (see Chapter 5 and Appendix 1), subjects were male Long–Evans rats (Charles River Laboratories, St. Constant, Quebec, Canada) weighing 275–300g at the start of testing. Two weeks following arrival, rats were food-restricted to 14 g of rat chow per day and maintained at 85% of their free-feeding weight. Water was available ad libitum. All subjects were pair or trio-housed in a climate-controlled colony room under 12 h reverse light–dark cycle (21◦C; lights off at 8 am). Behavioral testing took place 4-6 days per week. Housing and testing conditions were in accordance with the Canadian Council of Animal Care, and experimental protocols were approved by the UBC Animal Care Committee.    2.2 Behavioural Apparatus  Testing took place in 32 standard five-hole operant chambers (Figure 2.1), each of which was enclosed in a ventilated, sound-attenuating chamber (Med Associates Inc, Vermont). The dimensions of this chamber are 25 x 25 x 25cm (Figure 2.2). The front and back walls, and the roof of the chamber are made of clear polycarbonate. The front panel has a hinged door to allow the introduction of the animal. Both side walls are made of aluminum; one is curved and contains 5 square holes (2.5cm sides, 4cm deep) positioned 2cm above the metallic grid     75 floor. The floor is made of stainless steel rods. A stimulus light (28V, 100mA, 1W, 2.5cm diameter) is positioned at the back of each hole. The other side wall is not curved and contains a food well (5 x 5cm), which is 25cm away from each of the five holes. This side wall also contains a house light (28V, 100mA, 3W) located 1cm from the top, which is the source of illumination during the task, and a retractable lever on either side of the food well. Once the animal pokes its nose inside the food well, a sucrose pellet (45mg; Bioserv, New Jersey) is delivered from an automatic dispenser positioned outside the box. A horizontal infrared beam crosses the entrance of the holes and food well in order to record the animals’ response. The chamber is enclosed in a ventilated sound-attenuating cubicle (inner dimensions 55.9 x 40.6 x 38.1cm). A micro-camera is mounted on the ceiling of this cubicle to monitor the rats’ behaviour. The operant chambers were operated by software written in Med-PC by CAW, running on an IBM-compatible computer.  2.3 The Rat Cognitive Effort Task (rCET) Training and Testing Habituation to the operant chambers took place over two daily sessions, during which the chambers were turned on and 5-10 sucrose pellets were placed in the response holes and food magazine.  Rats were then exposed to a variant of 5-choice serial reaction training, in which they had to detect a stimulus light appearing in one of the five nosepoke holes. Rats initiated a trial by poking their nose in the illuminated food well.     76 Following a 2 s inter-trial interval (ITI), a 30 s light stimulus illuminated one of the five holes. The rat was required to scan the five holes and respond by poking its nose in the lit hole within 30 s, at which point a food pellet was dispensed to reward this correct response. If the rat responded before the light stimulus turned on (“premature response”), responded by poking its nose in an un-lit hole (“incorrect response”), or failed to respond (“omission”) within the limited hold (LH) period, this resulted in a 5 s time-out (TO) period where the house light was extinguished and no reward pellets were offered. Once the TO was complete, the food well illuminated to signal that the next trial could be initiated. These sessions lasted 30 minutes or 100 completed trials, whichever came first. Parameters were adjusted until rats could successfully complete at least 50 correct trials with a 5 s ITI, a stimulus duration of 10 s, and a 10 s limited hold period (equivalent to Stage 3 of standard 5-CSRTT training). This was achieved within 5-10 sessions.  In subsequent sessions, subjects were trained to respond on two retractable levers at a fixed ratio 1 schedule for reward.  At the beginning of the session, the food tray light illuminated and one of the response levers extended from its slot (half of the animals were presented with the left lever, the other half with the right). When the animal depressed the lever, one sugar pellet was dispensed into the food tray. The session expired after 30 minutes. In the next session, the opposite response lever was presented and the animal was similarly rewarded for pressing the lever. Training progressed when all animals made 50 or more presses on each lever (approximately 2-3 sessions).      77 Rats then progressed to the forced choice version of the rCET. Prior to testing, the response levers were permanently designated as either “easy” or “difficult”. For half of the animals, the left response lever triggered a “low-effort/low reward” (LR) trial and the right response lever triggered a “high effort/high reward” (HR) trial; this arrangement was reversed in the other half of the animals. At the beginning of the session, the food tray light illuminated to signal that the first trial was available. The animal nose-poked the food tray to initiate the trial, and the food tray light extinguished. One of the two levers extended, chosen at pseudorandom; this indicated whether the trial was “easy” or “difficult”. The animal pressed the response lever and a 5 s inter-trial interval (ITI) followed.  If the animal did not press the response lever within 10 s (choice omission), or if the animal nose-poked one of the five response holes during the ITI (premature response), the house light illuminated and the trial ended; after a 5 s “time out” punishment, the house light then extinguished, the food tray light illuminated, and the next trial could be initiated. Following the ITI, one of the five response hole lights illuminated, chosen at pseudorandom. The animal then had an opportunity to respond: Correct response: the animal nose-poked the response hole that was illuminated Incorrect response: the animal nose-poked in one of the four response holes that was not illuminated Response Omission: the animal took no action      78 A correct response resulted in pellet delivery at the food tray, and the food tray light illuminated to signal that the next trial was available. Incorrect responses or omissions resulted in the illumination of the house light and a 5 s “time out” punishment, wherein the next trial was not available; following the punishment, the house light extinguished and the food tray light illuminated, signalling that the next trial could be initiated. During the initial forced choice sessions, the stimulus duration (SD) for the LR and HR levers were set at 10 s, and correct detection of these stimuli resulted in 1 sugar pellet reward. The levers were matched for SD and reward magnitude until rats were able to successfully detect a 1 s SD on the HR lever with 70% accuracy (along with <20% choice omissions, <20% hard omissions, and more than 25 correct responses made on the HR lever- see Table 2.1). Following this, the levers began to diverge in difficulty and reward magnitude. The SD and reward for correct responses remained the same for the LR lever (1 s and 1 sugar pellet, respectively), but correct responses on the HR lever now resulted in 2 sugar pellets, and SD for the HR lever incrementally decreased by .1 s until reaching a final duration of .2 s. If the animal surpassed the criteria for their respective stage of training (e.g. greater than 25 correct HR trials, greater than 70% accuracy and < 20% omissions on HR trials), they progressed to the next stage with correspondingly shorter stimulus durations. Training continued until all animals reached the stage corresponding to the free-choice stimulus durations (i.e. an LR stimulus duration of 1 s and a HR stimulus duration time of 0.2 s). Rats typically required 60-80 sessions of forced choice training before moving onto the free choice version of the task.        79  The free choice version of the rCET is nearly identical to the forced choice training, except that both of the levers extended after the animal initiated a trial. Thus, in this version rats can choose between easy or hard trials by selecting the corresponding lever (Figure 2.3). To recap, subjects began each trial by nose-poking in the illuminated food tray, thereby extending the levers. Pressing a lever would set the trial as LR or HR, at which point the levers would retract and a 5 s inter-trial interval (ITI) would be initiated. Following this ITI, one of the five stimulus lights would be briefly illuminated, with stimulus durations of 1.0 s for LR trials and 0.2 s for HR trials. Animals then had 5 s to nosepoke within the previously illuminated response hole (correct response) for reward. Subjects were rewarded with one sugar pellet for a correct LR trial and two sugar pellets for a correct HR trial, at which point the tray light would re-illuminate to signal the opportunity to start the next trial.  Animals were tested 4-6 days per week in 30-minute sessions of no fixed trial limit.  Trials went unrewarded for a number of reasons: if animals failed to make a lever response within 10 s (a choice omission); if animals nosepoked during the ITI (a premature response, a long-used behavioural measure of motor impulsivity; Robbins, 2002); if animals nosepoked in any aperture other than the illuminated one (an incorrect response); and if animals failed to nosepoke any aperture within 5 s of stimulus light illumination (a response omission). All of these unrewarded trials were associated with a 5s time-out punishment period during which the houselight was illuminated and new trials could not be initiated     80 and thus reward could not be earned. Following the time-out, the tray light illuminated to signal that the rat could begin the next trial.   2.4 Behavioural Measures  Percent choice (rather than the absolute number of choices) was used to determine preference for lever/trial type, in order to minimize the influence of variation in the number of trials completed. Percent choice was calculated as follows: (number of choices of a particular lever / total number of choices) * 100. When baseline performance on the rCET was deemed statistically stable (no effect of session for choice, accuracy, and premature responding over the last three sessions when analyzed with a repeated-measures ANOVA; see “Data Analysis” below), animals were grouped as “workers” if they choice HR for >70% of trials and as “slackers” if they chose HR for < 70% of trials. This subdivision was based on the mean split from the original rCET paper (Cocker et al, 2012b), where workers and slackers were categorized based on their preference for greater than or less than the average of 70% HR trials. To maintain consistency when discussing individual differences across studies, we held the worker/slacker distinction at 70% HR trials for all experiments.     The following variables were analyzed separately for LR and HR trials: percent accuracy ((number of correct responses/ number of correct and incorrect responses made) * 100); percent response omissions ((number of trials omitted / number of correct, incorrect, and omitted trials) * 100); percent premature responses ((number of premature responses / total number of trials initiated) *     81 100); latency to choose between LR and HR levers (lever choice latency); latency to correctly nosepoke in the illuminated aperture (correct latency); latency to collect reward (collection latency). Failures to choose a lever at the beginning of the trial (choice omissions), number of trials initiated, and total number of trials completed (correct + incorrect + response omissions) were also analyzed.  2.5 Data Analysis Any analyses exclusive to specific experiments are described in subsequent chapters. All data were analyzed in SPSS (version 24.0; SPSS/IBM, Chicago, IL, USA). Variables expressed as a percentage were arcsine transformed to minimize the effects of an artificially imposed ceiling. Data were analyzed with a two-way repeated measures ANOVA with session (three levels: baseline sessions 1-3), inactivation (two levels: vehicle or inactivation), or dose (varying levels: vehicle plus different drug doses) as within-subjects factors, and group (two levels: worker or slacker) as a between-subjects factor for all analyses. A criterion was set wherein a subject had to select a given trial type (LR or HR) at least five times in order for behavioural measures associated with that lever to be analyzed (i.e. collect latencies, correct latencies, response omissions, choice latencies, premature responding, accuracy). Given that some rats selected the HR or LR option almost exclusively, behavioral measures for the non-sampled trial type were often unavailable.  Thus including choice as a within-subjects factor would remove these rats from all analyses, so instead     82 behavioral measures for LR or HR trials were analyzed separately. Violations of sphericity were assessed using Mauchly’s test, and when violated degrees of freedom were adjusted to more conservative values using the Greenhouse-Geisser correction. Corrected degrees of freedom are shown to the nearest integer. Any main effects or interactions of significance (p < .05) were further analyzed via post hoc one-way ANOVA or paired samples t-tests with a bonferonni correction for the number of comparisons made. Any p-values > .05 but < .10 were reported as a statistical trend.       83 2.6 Table Stage Limited Hold Size of Hard reward LR Stimulus duration (s) HR Stimulus Duration (s)  Criteria to Advance 3 10 1 10 10 < 20 choice omissions > 25 LR correct > 25 HR correct  4 5 1 5 5 < 20 choice omissions > 25 LR correct > 25 HR correct  5 5 1 2.5 2.5 < 20 choice omissions > 25 LR correct > 25 HR correct  6 5 1 1.25 1.25 < 20 choice omissions > 25 LR correct > 25 HR correct  7 5 1 1 1 Criteria for remaining Stages < 20% choice omissions > 25 HR correct > 70% HR accuracy < 20% HR omissions  8 5 2 1 .9   9 5 2 1 .8   10 5 2 1 .7   11 5 2 1 .6   12 5 2 1 .5   13 5 2 1 .4   14 5 2 1 .3   15 5 2 1 .2       84 Table 2.1. rCET training protocol  The animals are placed on a forced choice variant of the rCET, at the equivalent of 5-CSRTT stage 3 and are trained to the task’s final stimulus durations. Following this, animals are placed on the free-choice variant of the rCET until baseline behaviour is stable. LR: low-effort/low-reward; HR: high-effort/high-reward.                                     85 2.7 Figures        Figure 2-1. The standard 5-hole operant chamber used for the rCET  a) Front-facing view of operant chamber, with 5-hole stimulus array on the left and food tray on the right. b) The 5-hole stimulus array.          86  Figure 2-2. The standard 5-hole operant chamber used for the rCET  a) Aerial schematic of the 5-hole operant chamber, showing dimensions of the chamber, the 5-hole array, and the retractable levers on either side of the food well.     87   Figure 2-3. Schematic diagram showing the trial structure of the rCET  The task trial began with illumination of the tray light. A nose-poke response in the food tray extinguished the tray light, commencing a new trial and extending the levers. These levers were permanently designated to initiate either low-effort/LR or high-effort/HR trials. If one of the two levers was pressed, the levers retracted and a 5 s ITI would begin. Following the ITI, one of five stimulus lights would be briefly illuminated: 1.0 s for a LR trial and 0.2 s for a HR trial. A nose-poke response in the illuminated hole (i.e., correct response) led to a sugar reward-one pellet for a LR trial and two pellets for a HR trial-and the tray light would illuminate, indicating the opportunity to start the subsequent trial. A number of behaviors led to a 5 s time-out, signaled by house-light illumination: failure to make a lever response (choice omission); failure to withhold responding during the ITI (premature response); nose poke in an unlit hole following the stimulus (incorrect response); failure to make a nose-poke response following the stimulus (response omission.     88 Chapter 3: Chemogenetic Silencing of Basal Forebrain Neurons Does Not Affect Cognitive Effort Allocation in a Rodent Model of Decision Making with Cognitive Effort Costs   3.1 Introduction  The cholinergic system is a neuromodulator that is implicated in a number of processes including arousal (Jones, 2004), learning and memory (Hasselmo, 2006), and attention (Klinkenberg et al, 2011). Disorders associated with aberrant cholinergic signaling, such as schizophrenia (Higley and Picciotto, 2014; Raedler et al, 2007) and Alzheimer’s Disease (Court et al, 2001), are characterized by deficits in these cognitive domains, and for this reason many view the cholinergic system as a potential therapeutic target to ameliorate the cognitive impairments inherent to these disorders (Jones et al, 2011; Müller, 2007).  Emerging evidence suggests that the cholinergic system also contributes to the processes by which organisms evaluate choices in light of the costs and benefits associated with different candidate options. Indeed, impairments in cost/benefit decision making have been observed across Alzheimer’s Disease and schizophrenia (Sevy et al, 2007; Sinz et al, 2008), but it is unclear whether cholinergic impairments contribute to these decision-making deficits. Animal models of cost/benefit decision making suggest that perturbations in acetylcholine transmission affect decision making involving delay (Mendez et al,     89 2012) and uncertainty (Mendez et al, 2012, 2013; Silveira et al, 2014) costs, and one investigation has demonstrated that muscarinic receptor agonism in the nucleus accumbens decreases rats’ willingness to exert physical effort for lucrative rewards (Nunes et al, 2013a).  Recently, our lab has demonstrated that acetylcholine has dissociable roles in regulating attention versus the decision to engage in an attentional challenge, using the rodent Cognitive Effort Task (rCET) (Cocker et al, 2012b). In the rCET, rats decide whether to engage in easy or hard attentional trials, where correct detection of a 1s stimulus results in 1 sugar pellet reward, while correction detection of a difficult, 0.2s stimulus delivers 2 sugar pellets (Cocker et al, 2012b). The rCET is an adapted version of the 5-choice serial reaction time task (5-CSRTT), so in addition to measures of choice and accuracy, rates of premature responding are also assessed. Systemic administration of the muscarinic receptor antagonist scopolamine decreases choice of the high-effort/high-reward (HR) option across all rats without affecting attentional accuracy, while the muscarinic receptor agonist oxotremorine and the nicotine antagonist mecamylamine are without effect (Hosking et al, 2014c). In contrast, systemic administration of nicotine decreases choice of the HR option in a subset of rats who generally avoid the more demanding, albeit more lucrative attentional challenge. While nicotine decreases effortful choice in these “slacker” rats, it concomitantly increases their attentional accuracy on hard trials (Hosking et al, 2014c). Importantly, the improved accuracy in slackers following nicotine does not reflect amelioration of an attention deficit, as numerous studies have shown     90 that workers and slackers do not differ in attentional ability (Cocker et al, 2012b). Collectively, these results suggest that perturbing cholinergic signaling may impair or improve decision making with cognitive effort costs, the results of which depend on individual propensity to work or slack at baseline. Additionally, these effects on choice behaviour are separable from the established role of acetylcholine in regulating attentional processes (Himmelheber et al, 2000; Mcgaughy et al, 2002; Passetti et al, 2000).  Indeed, cholinergic regulation of attention versus decision making may map onto dissociable cholinergic neuron populations in the brain, such as the cortically projecting neurons of the basal forebrain (BF), versus the locally acting striatal interneurons (Bubser et al, 2012). Eight distinct cholinergic populations have been identified, but how any of these might regulate decision making is currently unknown (Bubser et al, 2012; Mesulam et al, 1983).  To begin to investigate this question, we took advantage of a transgenic rat line selectively expressing cre recombinase in ChAT neurons, and paired this with infusions of cre-driven Designer Receptors Exclusively Activated by Designer Drugs (DREADDs) directed towards the BF (Witten et al, 2011). This procedure allows for cell-type specific insertion of DREADDs receptors into ChAT neurons of the BF, which are activated by the otherwise inert ligand clozapine N-oxide (Roth, 2016; Smith et al, 2016). The DREADD receptor utilized in this study is a mutated variant of the muscarinic M4 receptor, which upon ligand binding leads to neuron downregulation given its linkage to inhibitory GPCR signaling pathways (Urban and Roth, 2015). Rats were injected with CNO prior to rCET     91 performance, to assess how downregulation of BF cholinergic neurons affects performance on-task. We also included a transgene negative group to control for potential nonspecific effects of CNO (Gomez et al, 2017). We also delivered a number of pharmacological challenges to assess whether transgenic rats expressing cre recombinase in ChAT neurons behave similarly to transgene negative controls.   3.2 Additional methods    3.2.1 Subjects and Baseline Choice Subjects were 39 male Long-Evans rats bred in house. These rats were bred from breeding pairs obtained from Charles River Laboratories and the Rat Resource and Research Centre (RRRC, Columbia, MO) as part of a breeding program for transgenic rats that express cre recombinase (Cre) in neurons that contain choline acetyltranferase (ChAT; Long Evans –Tg (ChAT-Cre) 5.1 Deis, RRRC # 00658). Nineteen of these rats were positive for the transgene (TG+), while the remainder did not express the transgene (TG-). The transgenic status was included as a between-subjects variable for all analyses. One rat suffered health complications early on in training, leaving 38 males.  For this particular experiment, the mean choice of the HR option across rats was 76%. Animals were grouped as “workers” if they chose the HR option for >70% of trials (n= 29) and as “slackers” if they chose HR for ≤70% of trials (n= 9), as per previous work (Cocker et al., 2012), thereby enabling consistency     92 when discussing individual differences across studies.   3.2.2 Surgery   Animals were anesthetized with 2% isoflourane in O2 and administered .1 ml of 5 mg.ml anafen and 10 ml lactated ringer subcutaneously. Rats were then secured in a stereotaxic frame (David Kopf) with the incisor bar set at -3.3. When a surgical plane of anesthesia was reached, rats were infused with AAV5-hsyn-DIO-HM4DGi-mcherry (UNC Vector Core, 112 Durham, USA; titer 5.5 x 1012 gc/ml) bilaterally at two sites: Site 1: AP: -.8mm from bregma, ML: +/- 2.5 mm, DV: - 8 mm from dura; Site 2: AP: -1.5 mm, ML: + 2.5 mm, DV: -7 mm from dura. Using 32 G stainless steel injectors (Plastics One, Roanoke, VA), PE tubing (Instech) and 10 ul syringes (Hamilton), 1 ul of AAV was infused at each site a rate of 0.1 ul/minute, and injectors left in place for 10 minutes to ensure solution fully diffused from the injector tip. Animals were sutured and recovered for at least 5 days before resuming behavioural testing.    3.2.3 Drugs and Pharmacological Challenges  CNO was purchased from Toronto Research Chemical (North York, Ontario, Canada) and was dissolved in 5% DMSO in 0.9% saline. CNO was administered interperitoneally (I.P) 30 minutes before the start of the task. The muscarinic antagonist scopolamine hydrobromide was purchased from Tocris (Minneapolis, MN, USA), dissolved in 0.9% saline, and administered I.P right before task onset. The acetylcholinesterase inhibitor donepezil hydrochloride     93 monohydrate was purchased from Sigma (Oakville, Ontario, Canada), dissolved in 0.9% saline, and administered 30 minutes prior to task. All drugs were calculated as the salt and administered in a volume of 1ml/kg body weight.   3.2.4 Experimental Timeline  The transfection surgeries occurred while rats were being trained on the rCET. The first CNO (0, 0.3, 1.0, 3.0 mg/kg) Latin square occurred approximately 10 weeks after the final surgery, at which point their behaviour was stable. This was followed by a scopolamine Latin square (0, .03, 0.1, 0.3 mg/kg), and then by another CNO Latin square (0, 0.3, 1.0, 3.0 mg/kg) 15 weeks following the last transfection surgery. Rats then received crossover AB injections of vehicle or a high dose of CNO (10 mg/kg, approximately 20 weeks post-operation). The acetylcholinesterase inhibitor donepezil was then administered (0, 0.3, 1.0, 1.5 mg/kg). All latin squares were bookended by a one week washout period to prevent carryover effects, and stable baseline behaviour was established according to the procedures laid out in the general Data Analysis section in Chapter 2.    3.2.5 Immunohistochemistry  Rats were euthanized by transcardial perfusion of ice cold .9% saline in 1x PBS followed by 4% paraformaldehyde in 1x PBS. Rats were injected with 120 mg/kg ketamine and 15 mg/kg xylazine intraperitoneally, perfused, and brains were extracted and stored in 4% in 1x PBS paraformaldehyde for 48 hours     94 before being transferred to 30% sucrose solution in 1x PBS for at least 72 hours. The brains were then frozen and sliced at 35 µm on a cryostat. Sections encompassing the mPFC and BF were taken and either mounted onto gel-coated slides or stored as free-floating slices in 1x PBS.  Sections were processed for ChAT immunoreactivity. Sections were incubated in anti-ChAT primary antibody (Rabbit polyclonal, Millipore, cat # AB143, concentration used 1:200), and then secondary antibody (Goat anti Rabbit, Alexa Fluor 488conjugate, Invitrogen, cat # A11034, concentration used 1:500). Mounted slides were cover-slipped with VectaShield anti-fade Mounting Medium (Vector Laboratories, Burlingame, CA, USA). Expression of mCherry autoflourescence and ChAT was confirmed within the BF using an AxioZoom V16 microscope (Zeiss, Germany).   3.2.6 Data Analysis   For all analyses, transgene status was included as a between-subject factor. Any main effects of transgene status or interactions with dose and/or worker/slacker status were followed up post-hoc with a simple main effects analysis. Drug doses were compared to vehicle, and a bonferonni correction was applied to all p values to account for multiple comparisons. P values greater than .05 but less than .1 were reported as a statistical trend.   3.3 Results       95 3.3.1 Histology and Final Group Numbers   Verification of DREADDs transfection in ChAT neurons was confirmed by looking for co-expression of ChAT and mCherry flourescence. See Figure 3.1 for a representative image. hM4D(Gi) expression was primarily observed in the substantia innominata and globus pallidus, with less expression noted in the magnocellular preoptic nucleus. We did not observe expression in the diagonal band of Broca or the medial septum. We did not see expression of mCherry in 3 TG + samples, and so these rats were excluded. This left 16 TG + and 19 TG – rats for analysis. Including worker/slacker status, the final group numbers were as follows: TG  + worker, n = 11; TG + slacker: n = 5; TG – worker: n = 16; TG – slackers: n = 3).   3.3.2 CNO Administration 10 Weeks Post Viral Transfection Surgeries Choice, Accuracy, Premature Responding  Baseline behaviour has been described in detail in other chapters of this thesis, so will be briefly discussed here. As a group, rats chose the high-effort, high-reward (HR) option more than the low-effort, low-reward (LR) alternative (vehicle only- Choice: F(1, 31) = 28.399, p < .001). In keeping with their group categorization, workers chose the HR option more than slackers (vehicle only- Group: F(1, 31)  = 39.539, p < .001). Rats who were transgene positive or negative did not differ in choice of HR (vehicle only- Transgene group: F(1, 33)  = 0.271, NS). Transgene status also did not interact with lever choice or worker/slacker status (Choice x Transgene, Choice x Transgene x Group,     96 Transgene x Group: all Fs < .419, NS), suggesting the respective worker/slacker groups were equivalent across transgene status. Administration of CNO had no selective effects on the HR choice of transgene positive rats (Dose x Transgene: F(2, 56) = 1.432, NS; Dose x Transgene x Group: F(2, 93) = 0.51, NS, Figure 3.2), or on HR choice in general (Dose, Dose x Group: all Fs < 0.684, NS).  Animals were more accurate on LR versus HR trials (vehicle only- Choice: F(1, 31) = 133.765, p < .001). Workers and slackers did not differ in their performance across trials types (Choice x group: F(1, 31) = 0.00, NS; Group: F(1, 31) = 0.002, NS). Transgene positive and negative rats also did not differ in their attentional performance at baseline (Choice x Transgene, Choice x Transgene x Group, Transgene: all Fs < 0.372, NS). CNO had no effect on accuracy across trial types (LR and HR trials- Dose, Dose x Transgene, Dose x Group, Dose x Transgene x Group: all Fs < 2.315, NS, Figure 3.2).   There was a trending choice effect on premature responding, with rats making slightly more (3% versus 4%) premature responses on HR versus LR trials (vehicle only- Choice: F(1,31) = 4.023, p = .054). However, rates of premature responding did not differ between worker and slacker rats (Choice x Group, Group: all Fs < .048, NS). Rats with different transgene statuses also did not differ in rates of premature responding (vehicle only- Transgene x Choice, Transgene x Choice x Group, Transgene: all Fs < 2.135, NS). CNO had no effect on premature responding across trial types (LR and HR trials- Dose, Dose x Transgene, Dose x Group, Dose x Transgene x Group: all Fs < 1.455, NS, Figure 3.2).     97  Other Behavioural Measures   Rats initiated approximately 130 trials in an rCET session, and choice omissions following lever presentation were generally low (~4). This did not differ between worker/slacker groups (vehicle only- Group: all Fs < 1.764, NS), or transgene status (vehicle only- Transgene: all Fs <  0.050, NS). Latencies to make a choice and to collect reward did not differ across trial types, worker/slacker status, or transgene status (Choice, Group, Transgene: all Fs < 0.712, NS). Rats were faster to make a correct response on HR versus LR trials (vehicle only- Choice: F(1, 31) = 4.973, p = .033), and this did not differ between workers/slackers or transgene status (vehicle only- Transgene, Group: all Fs < 0.460, NS). Omissions following stimulus presentation were low and did not differ between trial types (vehicle only- Choice: F(1, 31) = 2.969, NS). Response omissions did not differ between transgene positive or negative rats or between workers and slackers (Choice x Transgene, Choice x Group, Choice x Group x Transgene, Transgene, Group: all Fs < 1.922, NS).   Administration of CNO increased trials initiated across all rats at the 1 mg.kg dose (Dose: F(3, 93) = 2.816, p = .043; sal versus 0.3 mg/kg: F(1, 34) = 0.487, NS; sal versus 1.0 mg/kg: F(1, 34) = 6.914, p = .013; sal versus 3.0 mg/kg: F(1, 34) = 0.117, p = .734), but this did not interact with worker/slacker or transgene status (Dose x Transgene, Dose x Group, Dose x Transgene x Group: all Fs < 1.839, NS). CNO had no effect on any other measure (all Fs < 1.358, NS) (Table 3.1).      98   3.3.3 CNO Administration 15 Weeks Post Viral Transfection Surgeries CNO administration had no effect on choice, accuracy or premature responding across rats (Dose, Dose x Transgene, Dose x Group, Dose x Group x Transgene: all Fs < 1.878, NS, Figure 3.3). CNO also did not affect any other rCET measure (Dose, Dose x Transgene, Dose x Group, Dose x Group x Transgene: all Fs < 2.377, NS, Table 3.2).   3.3.4 High Dose (10 mg/kg) CNO Administration   We subsequently increased the dose of CNO to 10 mg/kg, as some researchers have reported only seeing behavioural effects at this higher dose (Smith et al, 2016). Again, CNO administration had no effect on choice, accuracy or premature responding across rats (Dose, Dose x Transgene, Dose x Group, Dose x Group x Transgene: all Fs < 2.607, NS, Figure 3.4), although there was a trending increase in premature responding following CNO across all rats (Dose: F(1, 31) = 4.171, p = .050). The 10 mg/kg dose of CNO did not generally affect the other rCET measures (Dose, Dose x Transgene, Dose x Group, Dose x Group x Transgene: all Fs < 2.706, NS, Table 3.3). However, there was a significant Dose x Transgene X Group (F(1, 31) = 15.553, p < .001) effect of CNO on HR collect latencies, but this was not significant following post hoc follow-up testing.   3.3.5 Reanalysis of CNO Data Excluding Worker/Slacker Status      99 There were a limited number of slacker rats in both the transgene positive and transgene negative conditions (n = 5 and n =3, respectively), thus limiting our ability to probe potential worker/slacker differences. Inclusion of worker/slacker status in the analyses reduces the degrees of freedom and thus the statistical power to detect any potential effects of CNO, and so we reanalyzed all the data, but did not include worker/slacker status as a between-subjects factor in the mixed model ANOVA. When reanalyzed in this way, CNO dose still did not interact with transgene status for any measure (Transgene x Dose: all Fs < 2.830, NS).   3.3.6 Scopolamine Administration Choice, Accuracy, Premature Responding  One rat did not start the task following injection of the middle scopolamine dose, and so was excluded from all analyses. In keeping with previous reports, the highest dose of scopolamine decreased choice of HR trials across all rats (Dose: F(3, 90) = 4.464, p = .011; sal versus 0.03 mg/kg: F(1, 33) = 0.606, NS; sal versus 0.1 mg/kg: F(1, 33) = 4.908, NS ; sal versus 1.0 mg/kg: F(1, 33) = 16.594, p < .001, Figure 3.5).  This effect did not interact with worker/slacker designation or transgene status (Dose x Group, Dose x Transgene, Dose x Transgene x Group: all Fs < 2.128, NS). Scopolamine administration had no effect on accuracy or premature responding (Dose, Dose x Group, Dose x Transgene, Dose x Transgene x Group: all Fs < 1.99, NS, Figure 3.5).       100 Other Behavioural Measures   Scopolamine decreased the number of trials initiated across rats at the 0.1 and 0.3 mg/kg doses (Dose: F(2, 66) = 19.725, p < .001; sal versus 0.03 mg.kg: F(1, 33) = 3.186, NS; sal versus 0.1 mg/kg: F(1, 33) = 18.888, p < .001; sal versus 0.3 mg/kg: F(1, 33) = 72.233, p < .001; Dose x group, Dose x Transgene, Dose x Transgene x Group: all Fs < 1.545, NS) and increased choice omissions at all doses (Dose: F(3, 78) = 13.381, p < .001; sal versus 0.03 mg.kg: F(1, 33) = 9.425, p = .012; sal versus 0.1 mg/kg: F(1, 33) = 36.748, p < .001; sal versus 0.3 mg/kg: F(1, 33) = 38.069, p < .001; Dose x group, Dose x Transgene, Dose x Transgene x Group: all Fs < 0.920, NS, Table 3.4). Scopolamine sped up all rats’ latencies to make a HR choice (HR trials- Dose: F( 3, 90) = 3.5017, p = .018; sal versus .3 mg/kg: F(1, 33) = 15.50, p < .001; Dose x group, Dose x Transgene, Dose x Transgene x Group: all Fs < 1.981, NS, Table 3.4). Latencies to make a LR choice were unaffected (all Fs < 1.96, NS). Scopolamine increased latencies to make a correct response on LR, but not HR trials, but none of the individual doses were significantly different from saline when bonferonni corrections were applied (LR trials – Dose: F(2, 48) = 4.945, p = .012;  Dose x group, Dose x Transgene, Dose x Transgene x Group: all Fs < 1.403, NS; HR trials – all Fs < 0.62, NS). Latencies to collect reward were unaffected by scopolamine (all Fs < 1.451, NS). At the highest dose, scopolamine increased response omissions on LR and HR trials (LR trials – Dose: F(2, 74) = 3.983, p = .016; sal versus 0.3 mg/kg: F(1, 33) = 7.862, p = .024; Dose x group, Dose x Transgene, Dose x Transgene x Group: all Fs < 1.18, NS; HR trials- Dose: F(2, 55) = 14.561, p <     101 .001; sal versus 0.3 mg/kg: F(1, 33) = 39.904, p < .001; Dose x group, Dose x Transgene, Dose x Transgene x Group: all Fs < 1.70, NS, Table 3.4).   3.3.7 Donepezil  Administration  The acetylcholinesterase inhibitor donepezil did not affect choice or accuracy (Dose, Dose x Transgene, Dose x Group, Dose x Group x Transgene: all Fs < 1.878, NS, Figure 3.6). There was a significant dose x group interaction for premature responding on HR trials, but this was not significant following post hoc testing (LR trials –all Fs < 0.532, NS; HR trials – Dose x Group : F(3, 93) = 2.838, p = .042, Workers only and slackers only – all Fs < 1.925, NS, Figure 3.6).  Donepezil increased choice omissions at the middle dose (Dose: F(3, 93) = 4.276, p = .007; sal versus 0.3 mg.kg: F(1, 33) = 0.678, NS; sal versus 1.0 mg/kg: F(1, 33) = 7.986, p = .024; sal versus 1.5 mg/kg: F(1, 33) = 5.64, NS; Dose x group, Dose x Transgene, Dose x Transgene x Group: all Fs < 1.106, NS, Table 3.5). Latencies to make a choice, to make a correct response, and to collect reward were unaffected (Dose, Dose x group, Dose x Transgene, Dose x Transgene x Group: all Fs < 2.561, NS, Table 3.5). Donepezil decreased trials initiated significantly in transgene negative worker rats at the highest dose (Dose x Transgene x Group: F(2, 70) = 3.163, p = .043; transgene negative only : Dose x Group: F(2, 33) = 3.163, p = .056; transgene negative workers only: F(3, 45) = 3.513, p = .023; sal versus 1.5 mg/kg: F(1,15) = 10.388, p = .018).   3.4 Discussion     102   The aim of the current investigation was to assess whether the basal forebrain contributes to the willingness to expend cognitive effort, above-and-beyond its role in regulating attentional and other cognitive processes. To achieve this, we used a recently developed line of transgenic rats expressing cre recombinase in ChAT neurons (Witten et al, 2011), and combined this with a cre-driven viral vector expressing a mutated muscarinic inhibitory GPCR (DREADD). This allowed us to selectively target ChAT neurons in the basal forebrain (Figure 3.1), and to downregulate these neurons via administration of clozapine-n-oxide (CNO) – an otherwise inert ligand that selectively binds to DREADDs. CNO administered at two timepoints (T= 10 weeks and T = 15 weeks) or at a very high dose (10 mg.kg) did not affect decision making or attentional ability in rats expressing the DREADDs receptor (transgene positive rats). As a group, the rats responded normally to a scopolamine challenge previously shown to selectively decrease effort invested towards larger, albeit more cognitively costly rewards (Hosking et al, 2014c), suggesting they were sensitive to manipulations of the cholinergic system. Although these results must be considered in light of the limitations of using DREADDs to selectively control neuronal populations, the current findings suggest that the basal forebrain does not contribute to the regulation of cognitive effort allocation.   Before discussing the implications of the null effects observed in the current study, we must consider the possibility that the technique we carried out was insufficient to target the cholinergic neurons of the basal forebrain.     103 According to stereological estimates, ChAT neurons account for ~5% of all BF neurons, a third synthesize GABA, and the vast majority synthesize glutamate (Gritti et al, 2006). We based our coordinates for viral transfection on studies investigating BF contributions to different aspects of visual attention (Botly and De Rosa, 2012; Chudasama et al, 2004), and which used the immunotoxin 192 IgG-saporin that targets cholinergic cells (Wiley et al, 1991).  We delivered 2 ul of virus across two sites in each hemisphere, to ensure that we were able to transfect as many cholinergic BF neurons as possible. Indeed, we saw robust co-labelling of  ChAT and mCherry, suggesting the inhibitory DREADD was localized to ChAT-expressing neurons. To our knowledge, no one has evaluated the transfection quality of HM4Di targeted towards cholinergic neurons of the NBM, but a previous study using a Cre-inducible channelrhodopsin (ChR2) viral construct found that 90% of neurons expressing ChAT also expressed EYFP tagged to ChR2 (Witten et al, 2011). Other studies using DREADDs generally show that approximately 2/3 of neurons express DREADDs using common AAV8-synapsin vectors in several structures (Gremel and Costa, 2013; Vazey and Aston-Jones, 2014). However, it is difficult to compare transfection rates across studies, as this will differ based on a number of factors including specific DREADD used, neuronal population targeted, serotype, and infusion protocol. Another notable limitation in the current study is that we did not carry out the in vivo or in vitro studies to demonstrate that CNO binding to the HM4Di receptor resulted in silencing of ChAT neurons. However, other studies have shown that systemic administration of CNO suppresses neuronal firing within 10     104 minutes, with response offset occurring at 70 minutes (Chang et al, 2015), and so it is likely the activity of transfected BF neurons would have been downregulated for the duration of an rCET session.   We have previously implicated the cholinergic system in the processes by which organisms allocate their cognitive resources for lucrative outcomes, by showing that systemic administration of cholinergic agents was capable of affecting decision making with cognitive effort costs (Hosking et al, 2014c). This prompted us to ask whether a specific cholinergic neuronal population in the brain mediated this process, and we considered the basal forebrain cholinergic neurons prime candidates. Acetylcholine synthesizing neurons are scattered throughout the BF nuclei of the medial septum-diagonal band of Broca (MS-DBB) and the nucleus basalis magnocellularis/substantia innominata (nBM/SI) (Mesulam et al, 1983). The cholinergic neurons of the MS-DBB and nBM/SI project predominantly to the hippocampus and neocortex, respectively, and we targeted the nBM given its ascribed role in regulating visual attention, and because we have previously implicated a number of medial prefrontal cortical regions in regulating decision making on the rCET (Hosking et al, 2014a, 2015a). Temporary silencing of these cholinergic neurons with CNO did not affect rats’ choice of HR trials, nor did it affect attentional performance.  This latter finding is perhaps surprising considering a number of previous studies on the 5-CSRTT and related paradigms have implicated the nBM/SI in visual attention using the selective immunotoxin 192 IgG-saporin (Botly and De Rosa, 2012; Harati et al, 2008; Lehmann et al, 2003; Mcgaughy et al, 2002). In     105 one particular study, attentional impairments were most prominent in the high dose saporin group, which lesioned ChAT neurons in the nBM and the MS-DBB, while a lower dose selectively ablating ChAT neurons in the nBM produced subtler impairments in accuracy (Mcgaughy et al, 2002). A similar study also showed impairments in attentional accuracy at saporin doses that spared cholinergic neurons in the MS-DBB, suggesting that perturbation of cholinergic neurons in the nBM is sufficient to affect attentional processes (Harati et al, 2008). However, It is difficult to ascribe a purely attentional account to the deficits observed in some of these studies, as a number of cursory measures were also affected (including increased omissions, latencies, and premature responding), suggesting fundamental deficits in response control and motivation during the task (Lehmann et al, 2003; Mcgaughy et al, 2002). In contrast, other studies have failed to find attentional impairments in the 5-CSRTT or related tasks following nBM saporin lesions (Chudasama et al, 2004; Risbrough et al, 2002). Specifically, Risbrough et al., 2002 demonstrated a weak initial attentional deficit following NBM saporin lesions that quickly returned to pre-surgery levels. Similarly, Chudasama et al., (2004) found spared attentional processes in a variant of the 5-choice task, known as the Combined Attention and Memory Task, where rats have to detect a light stimulus and then recall its location after a delay. These particular studies reported increased omission rates following stimulus presentation -often considered a cursory measure of impaired attention- but primary indices of attentional processing were unaffected.   Indeed, a closer inspection of the available literature suggests that the     106 cholinergic neurons of the basal forebrain are not recruited in all situations taxing visual attention. Rather, these neurons appear to be play a prominent role in attention-recruiting situations that are especially difficult or novel, perhaps falling more in line with the role of the acetylcholine in the processes of cortical arousal (Teles-Grilo Ruivo et al, 2017). For example, in the study by Mcgaughy et al., (2002) reporting effects of NBM saporin lesions on attention, the original 5-CSRTT was modified so that incorrect trials and omissions automatically reinstated the ITI, meaning rats did not need to nosepoke the magazine tray to commence a new trial (Mcgaughy et al, 2002). This was done to increase the attentional demands of the task (Dalley et al, 2001), and is in contrast to the rCET where incorrect trials and omissions are followed by a 5s time-out and where a nosepoke is required to start the next trial. Attentional deficits following BF cholinergic neuron ablation also become pronounced in novel challenge sessions, where the parameters of the task, such ITI length or stimulus duration, change in a single session (Lehmann et al, 2003; Mcgaughy et al, 2002). In contrast, the two studies reporting null effects of saporin lesions on attention either used the standard five-choice, or used a variant of the task which reduced the number of stimulus presentations in a single session (Chudasama et al, 2004; Risbrough et al, 2002). If the BF cholinergic nuclei are particularly recruited in highly arousing situations, this may preclude their involvement in performance on the rCET. Indeed, while the rCET is an attentionally demanding task, rats receive extensive training, and the design of the task means that at trial onset rats have the opportunity to pace themselves before choosing the difficulty of the     107 upcoming trial. Had we instituted a rCET “challenge” session, or modified the pace of the rCET, it is possible we would have seen an effect of BF cholinergic downregulation on decision making and attention.   What also becomes clear after reviewing the literature is that a substantial number of cholinergic neurons must be depleted in order for effects on executive function to be observed. In the studies reporting no-to-minimal effects of saporin lesions, there was a 45-60 % reduction in ChAT BF neurons (Mcgaughy et al, 2002; Risbrough et al, 2002), while those reporting more substantial 5-CSRTT deficits saw almost complete ChAT BF loss, and in some cases these lesions also encroached on the MS-DBB cholinergic nuclei (Harati et al, 2008; Lehmann et al, 2003; Mcgaughy et al, 2002), Indeed, the study by McGaughy et al., (2002) reported a positive correlation between ChAT neurons in the nBM and accuracy on the 5-CSRTT. Collectively this work suggests that a majority of BF cholinergic nuclei must be silenced in order for behavioural effects to be observed. This may be another reason why the DREADDs technique we employed did not result in any noticeable behavioural changes. DREADDs are GPCRs that dampen, rather than eliminate, ongoing neural activity. Indeed, labs have reported that CNO suppresses neuronal activity to 60% of pre-CNO rates (Chang et al, 2015; Mahler et al, 2014). Thus it appears that DREADDs may not be the most appropriate method to investigate BF cholinergic neurons to attention and effort allocation. An alternative approach could be to make use of optogenetics, which utilizes light-gated ion channels, and which has been recently used to investigate the involvement of BF cholinergic nuclei to cue detection and vigilance in mice     108 (Gritton et al, 2016). However, a potential issue with this technique is the sparseness of BF cholinergic nuclei, making it unlikely that light would be able to penetrate brain tissue to silence enough cholinergic neurons in larger rodents.   While we chose to focus on BF cholinergic nuclei, there are a number of other cholinergic nuclei that might be responsible for regulating decision making on the rCET. Cholinergic nuclei are also found in the brainstem pedunculo-pontine and lateral dorsal tegmental nuclei, a subset of thalamic nuclei, and in the striatum, where they act as local interneurons (CINs) (Ballinger et al, 2016; Mesulam et al, 1983). Although CINs represent only a small fraction of the striatum total cell population (1–3% in rodents) they have dense and extensive axonal arborisation (Kawaguchi et al, 1995). These CINs would be another interesting target to pursue, given we have shown that the dorsomedial striatum regulates cognitive effort allocation (see Chapter 4). Cholinergic inputs to the dorsomedial striatum have also been detected originating from brainstem pedunculo-pontine nuclei (Dautan et al, 2014), thus providing another means by which cholinergic input may affect the striatal processes regulating decision making.   We subsequently carried out a number of pharmacological challenges to assess whether rats were sensitive to the effects of cholinergic drugs on the rCET. In keeping with a previous study, we demonstrated that scopolamine decreases HR choice, but does not affect ability to complete this more difficult challenge (Hosking et al, 2014c). We also administered the acetycholineserase inhibitor donepezil, but this did not affect decision making, attention, or     109 impulsivity. Notably, none of these drug effects interacted with transgene status, aside from donepezil significantly decreasing trials initiated in transgene negative workers. Overall, transgene positive rats demonstrated an identical behavioural phenotype to transgene negative rats, which is valuable to know given emerging evidence suggesting that method of transgenic development can lead to phenotypic variation. Indeed, a recent study has shown that ChAT– cre mouse lines created with the bacterial artificial chromosome (method) demonstrate significant deficits in intravenous nicotine self-administration, which are paralleled by an increase in vesicular acetylcholine transporter and choline acetyltransferase (ChAT) hippocampal expression (Chen et al, 2018). This was the same method used to generate the ChAT recombinase-driver rat lines (Witten et al, 2011), suggesting transgene status may pose as a confounding variable in some studies using these rat strains.   In conclusion, we were unable to identify a role for BF cholinergic nuclei in cognitive effort allocation and attention. We suggest that this reflects the selective recruitment of the BF cholinergic neurons in regulating arousal associated with demanding or novel conditions, which might explain the inconsistent literature implicating this neuronal population in visual attention as assessed by the 5-CSRTT and related paradigms. This work has important implications, and suggests that researchers must exercise caution when considering chemogenetic techniques to control cholinergic basal forebrain nuclei in relation to their role in mediating executive function. We also propose that cholinergic signalling in the striatum, mediated via CINs, may be another     110 potential candidate for regulating decision making with cognitive effort costs.      111 3.5 Tables   Table 3.1. Other behavioural measures following CNO administered ten weeks post viral transfection surgery Data are collapsed across transgene status and worker/slacker designation. Means are presented (+ SEM). # p < .1; * p < .05; ** p < .001    Vehicle 0.3 mg/kg 1.0 mg/kg 3.0 mg/kg Choice omissions 4.40 + 0.86 4.71 + 0.76 4.11 + 0.75 5.14 + 1.01 LR choice latency 3.70 + 0.25 3.80 + 0.28 3.39 + 0.21 3.63 + 0.30 HR choice latency 3.67 + 0.14 3.76 + 0.14 3.64 + 0.13 3.73 + 0.15 LR correct latency 0.56 + 0.03 0.58 + 0.04 0.51 + 0.02 0.51 + 0.02 HR correct latency 0.47 + 0.02 0.47 + 0.02 0.47 + 0.02 0.49 + 0.02 LR collect latency 1.99 + 0.11 1.77 + 0.06 1.80 + 0.10 1.91 + 0.08 HR collect latency 1.84 + 0.23 1.64 + 0.06  1.57 + 0.06 1.69 + 0.10 LR response omissions (%) 7.03 + 1.76 5.74 + 1.64 5.00 + 1.62 5.94 + 1.34 HR response omissions (%) 5.98 + 0.74 5.79 + 0.66 4.37 + 0.66 5.03 + 0.74 Trials initiated 132.77 + 3.15 134.17 + 2.82 139.34 + 2.58* 133.40 + 2.91     112  Table 3.2. Other behavioural measures following CNO administered fifteen weeks post viral transfection surgery Data are collapsed across transgene status and worker/slacker designation. Means are presented (+ SEM). # p < .1; * p < .05; ** p < .001  Vehicle 0.3 mg/kg 1.0 mg/kg 3.0 mg/kg Choice omissions 3.62 + 0.67 3.31 + 0.90 3.77 + 0.71 4.43 + 1.02 LR choice latency 3.50 + 0.24 3.38 + 0.26 3.26 + 0.23 3.62 + 0.26 HR choice latency 3.60 + 0.15 3.41 + 0.12 3.59 + 0.14 3.60 + 0.13 LR correct latency 0.51 + 0.02 0.50 + 0.02 0.49 + 0.02 0.50 + 0.02 HR correct latency 0.46 + 0.02 0.46 + 0.02 0.46 + 0.02 0.48 + 0.02 LR collect latency 1.81 + 0.08 1.91 + 0.12 1.93 + 0.18 2.00 + 0.24 HR collect latency 1.55 + 0.05 1.70 + 0.15  1.55 + 0.05 1.55 + 0.05 LR response omissions (%) 5.44 + 1.45 4.20 + 1.40 4.89 + 1.33 6.65 + 3.03 HR response omissions (%) 4.83 + 0.77 4.96 + 0.74 4.75 + 0.82 5.53 + 0.80 Trials initiated 139.46 + 3.37 141.57 + 3.38 139.69 + 3.32 138.39 + 3.31     113  Table 3.3. Other behavioural measures following 10 mg/kg CNO  Data are collapsed across transgene status and worker/slacker designation. Means are presented (+ SEM). # p < .1; * p < .05; ** p < .001         Vehicle 10.0 mg/kg Choice omissions 5.23 + 1.07 6.97 + 1.22 LR choice latency 3.39 + 0.26 3.52 + 0.25 HR choice latency 3.55 + 0.12 3.63 + 0.14 LR correct latency 0.54 + 0.03 0.55 + 0.02 HR correct latency 0.47 + 0.02 0.49 + 0.02 LR collect latency 1.95 + 0.12 2.00 + 0.14 HR collect latency 1.64 + 0.05 1.60 + 0.05 LR response omissions (%) 13.42 + 4.35 7.39 + 1.37 HR response omissions (%) 6.16 + 0.65 7.27 + 1.14 Trials initiated 130.66 + 3.02 127.63 + 3.48     114      Table 3.4. Other behavioural measures following systemic scopolamine administration Data are collapsed across transgene status and worker/slacker designation. Means are presented (+ SEM). # p < .1; * p < .05; ** p < .001  Vehicle 0.03 mg/kg 0.1 mg/kg 0.3 mg/kg Choice omissions 3.68 + 0.75 7.94 + 1.69* 13.65 + 1.94** 14.59 + 1.82** LR choice latency 2.88 + 0.25 3.30 + 0.26 2.80 + 0.19 2.57 + 0.17 HR choice latency 3.61 + 0.16 3.60 + 0.15 3.26 + 0.12 2.97 + 0.13** LR correct latency 0.47 + 0.02 0.48 + 0.02 0.55 + 0.04 0.63 + 0.04 HR correct latency 0.47 + 0.02 0.46 + 0.02 0.46 + 0.01 0.50 + 0.03 LR collect latency 1.79 + 0.07 1.81 + 0.08 1.96 + 0.13  2.54 + 0.36 HR collect latency 1.54 + 0.05 1.59 + 0.06  1.63 + 0.05 1.70 + 0.09 Trials initiated 138.79 + 3.47 132.56 + 4.56 117.12 + 5.10** 89.88 + 6.30**     115   Table 3.5. Other behavioural measures following systemic donepezil administration Data are collapsed across transgene status and worker/slacker designation. Means are presented (+ SEM). # p < .1; * p < .05; ** p < .001  Vehicle 0.3 mg/kg 1.0 mg/kg 1.5 mg/kg Choice omissions 5.27 + 0.89 5.80 + 0.80 7.40 + 1.16* 7.30 + 1.14 LR choice latency 3.51 + 0.24 3.45 + 0.33 3.84 + 0.31 3.87 + 0.34 HR choice latency 3.51 + 0.13 3.47+ 0.12 3.71 + 0.14 3.72 + 0.12 LR correct latency 0.50 + 0.03 0.54 + 0.03 0.50 + 0.03 0.53 + 0.02 HR correct latency 0.45 + 0.02 0.47 + 0.02 0.47 + 0.02 0.47 + 0.02 LR collect latency 3.14 + 1.07 2.37 + 0.30 1.88 + 0.12 2.06 + 0.19 HR collect latency 1.57 + 0.06 1.67 + 0.08  1.67 + 0.07 1.76 + 0.15 Trials initiated 132.76 + 3.76 131.00 + 3.78 121.89 + 5.00 119.94 + 5.00     116 3.6 Figures  Figure 3-1. Verification of HM4Di expression in cholinergic neurons of the basal forebrain A) Red circle corresponds to location of mCherry expression in image below brain atlas plate. hM4D(Gi) expression was primarily observed in the substantia innominata and globus palladus, with less expression noted in the magnocellular preoptic nucleus. We did not observe expression in the diagonal band of Broca or the medial septum. B) Cell bodies in the BF of a Tg + rat expressing ChAT (top), mCherry (middle), or co-expressing both mCherry and ChAT (bottom).      117  Figure 3-2. Effects of CNO administration on choice of HR trials (%), accuracy (%), and premature responses (%) ten weeks post viral transfection surgeries Data shown are the mean for each variable (±SEM). LR: Low effort/low reward trial; HR: High effort/high reward trial. Data is shown separated by transgene status (transgene positive on left, transgene negative on right). CNO did not affect choice, accuracy, or premature responding.     118  Figure 3-3. Effects of CNO administration on choice of HR trials (%), accuracy (%), and premature responses (%) 15 weeks post viral transfection surgeries Data shown are the mean for each variable (±SEM). LR: Low effort/low reward trial; HR: High effort/high reward trial. Data is shown separated by transgene status (transgene postivie on left, transgene negative on right). CNO did not affect choice, accuracy, or premature responding.     119               Figure 3-4. Effects of high dose CNO administration on choice of HR trials (%), accuracy (%), and premature responses (%) Data shown are the mean for each variable (±SEM). LR: Low effort/low reward trial; HR: High effort/high reward trial. Data is shown separated by transgene status (transgene positive on left, transgene negative on right). CNO did not affect choice, accuracy, or premature responding.      120  Figure 3-5. Effects of scopolamine administration on (a) choice of HR trials (%), (b) accuracy (%), (c) premature responses (%), and (d) response omissions (%) Data shown are the mean for each variable (±SEM). LR: Low effort/low reward trial; HR: High effort/high reward trial. Transgene status did not interact with dose for any measure, so data is collapsed across transgene status. The highest dose of scopolamine decreased HR choice and increased response omissions across rats but did not affect accuracy or premature responding.       121 Figure 3-6. Effects of donepezil administration on (a) choice of HR trials (%), (b) accuracy (%), and (c) premature responses (%) Data shown are the mean for each variable (±SEM). LR: Low effort/low reward trial; HR: High effort/high reward trial. Transgene status did not interact with dose for any measure, so data is collapsed across transgene status. Donepezil did not affect choice, accuracy, or premature responding.    122 Chapter 4: Dissociable Contributions of Dorsal and Ventral Striatal Regions on a Rodent Cost/Benefit Decision-Making Task Requiring Cognitive Effort   4.1 Introduction  In order to successfully navigate their environment, an organism must choose between actions based on the subjective evaluation of the costs and benefits associated with the likely outcomes. The magnitude of reward, the motivational state of the animal, as well as the amount of effort required to obtain preferred goals are all variables used to guide optimal decision making. And while such processes may seem trivial when considering the routine decisions of day-to-day life, optimal cost/benefit decision making becomes crucial when choices carry with them the possibility of substantial gain or loss. Indeed, perturbations in decision making are observed in almost every mental illness (Goschke, 2014).    A substantial body of experimental data using both human and animal subjects has demonstrated that decisions involving physical effort costs are mediated by a cortico-limbic-striatal circuit generally including, but not limited to, the anterior cingulate cortex (ACC), basolateral amygdala (BLA), and ventral striatum (Winstanley and Floresco, 2016). Lesions or inactivations of these     123 regions in rodents decrease their willingness to traverse a barrier or press a lever repeatedly for larger reward (Floresco and Ghods-Sharifi, 2007; Ghods-Sharifi et al, 2009a; Ghods-Sharifi and Floresco, 2010; Hauber and Sommer, 2009; Rudebeck et al, 2006; Walton et al, 2003b), and imaging studies have found increased activity in these regions as humans or rodents engage in the evaluation of physically effortful options (Croxson et al, 2009; Endepols et al, 2010).  However, these manipulations fail to assess effort that is cognitive in nature, which is arguably more representative of the decisions humans face on a daily basis. Such cognitive effort is intrinsically costly; humans avoid cognitively demanding tasks, the subjective value of rewarding outcomes is discounted by difficult cognitive challenge, and cognitive allocation towards harder tasks can be biased by incentives (Kool et al, 2010a; Westbrook et al, 2013). While a single operational definition of cognitive effort remains elusive, recent proposals suggest cognitive effort, and its allocation, should be considered in an economic framework (Botvinick and Braver, 2015; Westbrook and Braver, 2015). Executive processes constitute a limited resource, and as such are valuable because they promote behaviour in pursuit of desired goals. Viewed in this way, effort investment towards any particular goal precludes allocation to other options, thus incurring opportunity costs. Under such a framework, studies in humans have employed standard choice experiments and demonstrated that as cognitive demands and incentives increase, the subjective value of a given option decreases and increases, respectively (Chong et al, 2017; Dixon and Christoff,     124 2012; Kool et al, 2010a; Westbrook et al, 2013). Notably, this has been observed in tasks probing working memory, cognitive flexibility, response inhibition, and attention (Kool et al, 2010; Westbrook et al, 2013, Dixon and Christoff, 2012; Reddy et al, 2015, Chong et al, 2017). Cognitive effort is thus an intrinsically valuable resource, appropriate deployment of which can maximize gains, but which is also subjectively and objectively costly.  Whether cognitive effort, and its representation and regulation, should be conceptualized as unitary, pan-domain entities, or domain-specific processes, remains to be determined. Indeed, the mechanisms underlying the recruitment and allocation of cognitive effort in any aspect of cognition have yet to be firmly established.  To begin to address this discrepancy, we developed a rodent cognitive effort task (rCET), based on the 5-choice serial reaction time task, in which greater visuospatial attention is required to successfully obtain larger amounts of sucrose reward (Cocker et al, 2012b). As in human choice experiments manipulating cognitive effort costs, rats can decide at trial outset whether to detect a lengthy (1s) stimulus across five possible locations for one sugar pellet reward, or to earn two sugar pellets by detecting a brief (.2s) stimulus, which is much more difficult to identify (as reflected by lower task accuracy (ie, performance). Similar to animal physical-effort discounting tasks, decision making on the rCET is sensitive to inactivations of the ACC and BLA, but in the case of the BLA, is dependent on subjects’ subjective preference for the harder option at baseline (“workers” vs “slackers”; (Hosking et al, 2014a)). Additionally, and unlike the physical effort literature, inactivations of the prelimbic and     125 infralimbic regions of medial prefrontal cortex (mPFC) decrease willingness to expend cognitive effort, suggesting the two forms of decision making share interrelated-yet-distinct neurobiological mechanisms (Hosking et al, 2015a; Walton et al, 2002, 2003b).  These discrepancies may likewise imply that the role played by striatal subregions may differ with respect to decisions involving the evaluation of cognitive versus physical effort.  The ventral striatum, or nucleus accumbens (NAc), is histochemically and connectively segregated into a “core” and “shell” region, with previous work establishing a role for the core, but not shell, in decision making based on physical effort costs (Ghods-Sharifi and Floresco, 2010; Zahm, 2000). The NAc receives topographic excitatory inputs from the mPFC and BLA, and individual accumbal neurons receive converging inputs from various corticolimbic sites (Britt et al, 2012; Brog et al, 1993; Floresco et al, 2001; Mcdonald, 1991; Reynolds and Zahm, 2005). This organization allows the NAc to integrate signals from multiple regions, such as the effort and net value signals observed in the ACC, as well as valence signals observed from single unit recordings of the BLA (Belova et al, 2007; Cowen et al, 2012; Croxson et al, 2009; Jenison et al, 2011; Kennerley et al, 2009, 2011; Shidara and Richmond, 2002). For this reason the NAc has been deemed a “limbic-motor” interface, whereby these regions can influence motor output via accumbal projections to motor effector sites in the basal ganglia (Mogenson et al, 1980).  Although imaging studies suggest ventral striatal activity is inversely related to cognitive demand (Botvinick et al, 2009), the exact contribution of the nucleus accumbens     126 to decisions based on cognitive effort costs remains unknown. Moreover, effort tasks have mainly focused on the ventral, but not dorsal striatum, even though this latter region has been implicated in other forms of decision making and goal-directed behaviour (Balleine et al, 2007; Ferguson et al, 2013; Friedman et al, 2015). The dorsomedial portion of the striatum (DMS) receives efferent topographic projections from the medial prefrontal cortex, ACC and BLA (Mailly et al, 2013; Mcdonald, 1991; Sesack et al, 1989), and is thus ideally situated to integrate decision-related signals that can then bias action output. Indeed, the dorsal-ventral striatal boundary is an arbitrary one, with some proposing a ventromedial-to-dorsolateral striatal gradient based on functional organization (Voorn et al, 2004). It is possible that different corticostriatal “loops” processing facets of task performance (i.e., choice, attention, response inhibition) are integrated in the striatum, the resulting output of which can then bias motor output via palladal-thalamic-cortical pathways (Chudasama and Robbins, 2006; Haber and Behrens, 2014; Zahm and Brog, 1992). However, the contributions of striatal regions to cognitive effort allocation-versus-execution have not been investigated. The goal of the current investigation was therefore to characterize the role of the DMS and NAc core on rCET performance, using pharmacological inactivation methods comparable to that used in investigations of the role played by these regions in evaluating physical effort costs (Ghods-Sharifi and Floresco, 2010).       127 4.2 Additional Methods   4.2.1 Subjects and Baseline Choice Profile  Subjects were 32 male Long Evans rats. When baseline performance on the rCET was deemed statistically stable, the mean choice of the HR option was 70%. Animals were grouped as “workers” if they chose the HR option for >70% of trials (n = 15) and as “slackers” if they chose HR for < 70% of trials (n = 15), as per previous work (Cocker et al, 2012b), thereby enabling consistency when discussing individual differences across studies. Two rats failed to achieve criterion during rCET training and were thus excluded and not categorized.  4.2.2 Surgery to Implant Guide Cannulae When baseline performance was deemed statistically stable, animals were implanted with 22-gauge stainless steel guide cannulae (Plastics One) bilaterally into the dorsomedial striatum (DMS) using standard stereotaxic techniques. Animals were anesthetized with 2% isoflurane in O2 and cannulae implanted at the following coordinates: AP = +1.5 mm, ML = ±1.9 mm from bregma, DV = −3.8 mm from dura (Paxinos and Watson, 1998). Guide cannulae were fixed to the skull via 3-4 stainless steel screws and dental acrylic, and obdurators with dustcaps were inserted which ended flush with the end of the cannulae. Animals were given at least 1 week of recovery in their home cages before subsequent testing. Three animals were excluded due to poor recovery.       128 4.2.3 Microinfusion Procedure Following recovery, animals performed 10 free-choice sessions, after which all individuals displayed stable behavior. Animals were then habituated to the microinfusion process with 2 mock infusions, during which 30-gauge injectors with tips extending 1 mm beyond the guide cannulae were inserted for 2 min but no infusion was performed, followed by a behavioural testing session. Infusions adhered to a 3-day cycle starting with a baseline session, followed by a drug or saline injection session, and then by a non-testing day. The DMS was inactivated by a mixture of the GABAB agonist baclofen and the GABAA agonist muscimol (Sigma–Aldrich), prepared separately at 0.5 µg/µL in saline, and mixed together in equal volumes to form a 0.25 µg/µL solution. 0.5 µL per hemisphere injections of saline or baclofen/muscimol (i.e. 0.125 µg of drug per hemisphere) were administered bilaterally at a rate of 0.4 µL/min, and injectors were left in place for an additional minute to allow diffusion. Pilot infusions indicated that this dosing regimen caused severe behavioural disruptions when infused into the nucleus accumbens core, and so the protocol was adjusted so that baclofen and muscimol were prepared separately at 0.25 µg/µL in saline, and mixed together in equal volumes to form a 0.125 µg/µL solution. 0.3 µL per hemisphere injections of saline or baclofen/muscimol (i.e. 0.0375 µg of drug per hemisphere) were administered into the nucleus accumbens core (NAcC) bilaterally at a rate of 0.4 µL/min, and injectors were left in place for an additional minute to allow diffusion. The NAcC was targeted using the same guide cannulae, but with injectors extending 3 mm beyond the cannuale tips.      129 Once the microinfusions were completed, injectors were removed, obdurators replaced, and animals were returned to their home cages for 10 min before being placed in the operant chambers and performing the rCET. Animals underwent 4 infusion sessions in total: On the first infusion day, half of the rats received saline infusions to the DMS (via injectors with +1 mm tips) whereas the other half received baclofen/ muscimol to the DMS; these administrations were reversed on the second infusion day, allowing for a within-subjects comparison. On the third infusion day, half of the rats received saline infusions to the NAcC (via injectors with +3 mm tips) whereas the other half received baclofen/muscimol; and again these administrations were reversed on the fourth infusion day. Animals were given a minimum of 1 week drug-free testing between DMS and NAcC inactivations, to minimize any carryover effects. DMS inactivation caused severe behavioral disruption in 1 animal and thus this rat was removed from analyses. Thus data from n = 15 workers and n = 11 slackers were included for behavioural analysis.   4.2.4 Histology Following completion of all behavioral testing, animals were anesthetized with isoflurane and sacrificed by carbon dioxide exposure. Brains were removed and fixed in 4% formaldehyde for at least 24 h, transferred to a 30% sucrose solution, and then frozen and cut via cryostat into 40-µm coronal sections. These sections were stained with cresyl violet for visualization, and the projected locations of the injector tips protruding from the guide cannulae were mapped     130 onto standard sections from Paxinos and Watson (1998).  4.2.5 Data Analysis   Data were analyzed with a two-way repeated measures ANOVA with session (three levels: baseline sessions 1-3) or inactivation (two levels: vehicle or inactivation) as within-subjects factors, and group (two levels: worker or slacker) as a between-subjects factor for all analyses. Groups proved stable across the experiment: at rCET baseline and all saline infusions, workers chose a significantly greater percentage of HR trials than slackers (Group: all Fs > 22.553, p < .001). Given the behavioural disruptions observed following NAcC inactivation, choice data was only included from rats that made more than 10 choices.   4.3 Results   4.3.1 Cannula Placements The locations of all acceptable placements are depicted in Figure 4.1. All DMS placements were acceptable, but one rat was excluded from the NAcC analysis due to inaccurate placement in one hemisphere, leaving a total of 26 (n = 15 workers; n = 11 slackers) and 25 (n = 15 workers; n = 10 slackers) rats for the DMS and NAcC analyses, respectively.   4.3.2 Dorsomedial Striatum Inactivation      131 Choice, Accuracy, and Premature Responses.  Baseline performance on the rCET has been discussed in detail elsewhere, and so will be only briefly discussed here. Animals chose the high-effort/high-reward (HR) option more than the low-effort/low-reward (LR) option when the dorsomedial striatum (DMS) was infused with saline (choice: F(1, 24) = 22.967, p < .001).  In keeping with their group categorization, workers chose HR trials significantly more than slackers (saline only- group: F(1, 24)  = 24.910, p < .001). Inactivation of the DMS decreased all animals’ choice of HR trials (inactivation: F(1, 24) = 9.538, p = .005; inactivation x group: F(1, 24) = 0.169, NS; Figure 4.2).  As shown previously, animals were more accurate on LR versus HR trials (saline only- choice: F(1, 21) = 135.561, p < .001) and attentional performance did not differ between groups (saline only- choice x group: F(1, 21) = 1.934, NS; group: F(1, 21) = .028, NS), suggesting that subjects’ choice of HR trials was not solely driven by their ability to complete these trials.  Inactivation of the DMS decreased LR accuracy to a greater degree in worker rats, as revealed by a significant inactivation x group interaction (LR trials- inactivation: F(1, 21) = 39.358, p < .001; inactivation x group: F(1, 21) = 4.549, p = .045). However, subsequent post-hoc tests revealed that slackers also experienced attentional deficits on LR trials (Workers-LR: F(1, 11) = 30.104, p < .001; Slackers-LR: F(1, 10) = 10.952, p = .008). Inactivation of the DMS decreased HR accuracy to a similar extent across all rats (HR trials- inactivation: F(1, 23) = 15.30, p = .001; inactivation x group: F(1, 23) = 2.402, p = .135; Figure 4.2).      132 Premature responding was higher on HR relative to LR trials when the DMS was infused with saline (saline only- choice: F(1, 21) = 10.063, p = .005). As per previous cohorts, rates of premature responding did not differ between groups, again indicating that individuals’ choice preferences were not driven by differences in motor impulsivity (saline only- choice x group/group: all Fs < .544, NS). Inactivating the DMS significantly increased levels of premature responding on LR trials across subjects (LR trials- inactivation: F(1, 22) = 65.915, p < .001; inactivation x group: F(1, 22) = .506, NS). Although both groups exhibited increased impulsive responding on HR trials, this effect was again more pronounced in worker rats (HR trials- inactivation: F(1, 23) = 52.265, p < .001; inactivation x group: F(1, 23) = 4.950, p = .036; Workers-HR: F(1, 14) = 46.918, p < .001; Slackers-HR: F(1, 9) = 14.837, p = .004; Figure 4.2).   Other behavioural measures.  Rats initiated about 140 trials and made few (<5) choice omissions when the DMS was infused with saline, regardless of worker/slacker distinction (saline only- all Fs < .147, NS; Table 4.1). Under normal conditions latencies to make a choice and collect reward were similar across levers and did not differ between groups (saline only- choice/ choice x group/ group: all Fs < 2.983, NS). While workers and slackers did not differ in their latencies to make a correct response, rats generally were quicker to make responses for HR trials, perhaps unsurprisingly given the shorter stimulus duration (saline only- choice: F(1, 20) = 13.282, p = .002; choice x group / group: all Fs < 2.204, NS). When infused with     133 saline, response omissions were equivalent for LR and HR trials across groups (choice/ choice x group / group: all Fs < 2.789, NS). In line with the impairments in accuracy, inactivation of the DMS increased response omissions for both trial types, suggesting all rats were impaired in allocating attention to the five-hole array (LR trials- inactivation: F(1, 22) = 9.99, p = .005; inactivation x group: F(1, 22) = 4.187, p = .053; HR trials- inactivation: F(1, 23) = 12.944, p = .002; inactivation x group: F(1, 23) = 0.686, NS; Figure 4.2). This increase in response omissions does not seem indicative of general motor impairments, given choice omissions were unchanged, and the trending decline in trials initiated was modest following DMS inactivation (choice omissions- inactivation/ inactivation x group: all Fs < 0.663, NS; trials initiated- inactivation: F(1, 24) = 4.065, p = .055; inactivation x group: F(1, 24) = 2.466, NS; Table 4.1). While trials completed did decline following DMS inactivation (inactivation: F(1, 24) = 54.211, p < .001), this is likely due to the pronounced increase in premature responding observed for LR and HR trials (see above). Additionally, choice latencies actually sped up following inactivation of the DMS (LR trials- inactivation: F(1, 23) = 5.135, p = .033; inactivation x group: F(1, 23) = 1.139 NS; HR trials- inactivation: F(1, 22) = 9.564, p = .005; inactivation x group: F(1, 23) = 1.331, NS). And while both groups took longer to make a correct response on LR and HR trials following DMS inactivation, latencies to collect reward for either trial type were unaffected, suggesting the inactivation did not affect motivation for sucrose reward (correct latency: LR trials- inactivation: F(1, 21) = 6.164, p = .022; inactivation x group: F(1, 21) = 0.32, NS; HR trials- inactivation: F(1, 23) = 5.821, p = .024; inactivation     134 x group: F(1, 23) = 1.338, NS; Collect latency: LR and HR trials – inactivation/ inactivation x group: all Fs < 2.157, NS; Table 4.1).   4.3.3 Nucleus Accumbens Core Inactivation In stark contrast to DMS inactivations, infusions of baclofen/muscimol into the nucleus accumbens rendered most rats incapable of performing the rCET. Both workers and slackers showed a drastic decline in trials initiated, only starting about 33% of trials relative to infusion of saline in the NAcC  (Figure 4.3- inactivation: F(1, 23) = 79.924, p < .001; inactivation x group: F(1, 23) = 0.827, NS). Indeed, the high rate of response omissions for both trials types across groups (LR trials- inactivation: F(1, 21) = 31.704, p < .001; inactivation x group: F(1, 21) = 1.049, NS; HR trials- inactivation: F(1, 19) = 19.043, p <.001; inactivation x group: F(1, 19) = 0.003, NS) limits a meaningful analysis of choice patterns, attention, and impulsivity (Figure 4.3). Among rats that made at least ten choices (n = 22), inactivation of the NAcC produced a trending decline in choice of the HR option (inactivation: F(1, 20) = 3.109, p =.093; inactivation x group: F(1, 20) = 0.088, NS) (Figure 4.3). Visual inspection of the data also suggests this was accompanied by impairments in attentional accuracy, although these effects failed to reach significance (LR trials- inactivation: F(1, 6) = 2.548, NS; inactivation x group: F(1, 6) = 0.001, NS; HR trials- inactivation: F(1, 12) = 0.949,  NS; inactivation x group: F(1, 12) = 0.158, NS), likely owing to the few number of cases analyzed for LR and HR trials (see table 4.2). Inactivation of the NAcC appeared to decrease premature responding, but only a trending decline     135 was observed for LR trials across groups (LR trials- inactivation: F(1, 14) = 3.881, p = .069; inactivation x group: F(1, 14) = 0.041, NS; HR trials- inactivation: F(1, 15) = 0.778, NS; inactivation x group: F(1, 15) = 0.479, NS) (Table 4.2). This decline in premature responding may reflect general motor impairments, but it is interesting to note that rats were quicker to initiate HR trials (LR trials- inactivation: F(1, 14) = 0.008, NS; inactivation x group: F(1, 14) = 0.117, NS; HR trials- inactivation: F(1, 15) = 6.623, p = .021; inactivation x group: F(1, 15) = 0.005, NS), and generally latencies to make a correct response and collect reward were unaffected following NAcC inactivation (all Fs < 3.284, NS) (Table 4.2). Lastly, inactivation of the NAcC did not affect choice omissions (inactivation: F(1, 23) = 2.320, NS; inactivation x group: F(1, 23) = 0.166, NS).    4.4 Discussion   Here we show that temporarily silencing dorsal and ventral regions of the striatum produce dissociable patterns of effects on a complex decision-making task requiring the evaluation and application of cognitive effort. In line with recent theories of cognitive effort, this study viewed its allocation as an economic choice, in which larger rewards were discounted by the cognitive effort costs associated with increased attentional demand (Westbrook and Braver, 2015).  Inactivation of the dorsomedial striatum (DMS) decreased all animals’ choice of the high effort/high reward (HR) option, while concomitantly impairing accuracy, increasing premature responding, and increasing omission rates. This     136 behavioural profile mimics lesions of the dorsomedial striatum on the standard 5-CSRTT (Rogers et al, 2001), and as suggested below, these deficits may arise not only from attentional impairments per se, but more generally from an inability to organize both attentional and response control processes necessary for optimal task performance. With these processes impaired, rats appeared to shift their choice towards a low cost strategy. In contrast, inactivations of the nucleus accumbens core (NAcC) region of the ventral striatum resulted in profound behavioural impairments that rendered rats incapable of successfully engaging in the task. The current investigation helps to delineate the cortico-striatal-limbic circuits guiding decision making based on cognitive effort costs, albeit restricted to the domain of attentional demands. As noted earlier, cognitive effort discounting has also been reported in the domains of working memory, response inhibition, and cognitive flexibility (Dixon and Christoff, 2012; Kool et al, 2010; Reddy et al, 2015; Westbrook et al, 2013). Future studies can now address whether similar circuits are involved in the allocation of cognitive effort across these executive functions. It is possible that DMS inactivation impaired subjects’ ability to discriminate LR from HR trials, such that the increased choice of LR trials reflects an indiscriminate sampling of the two levers, rather than a deliberate selection of the easier option. If so, then selection of the HR option should approximate 50% for both workers and slackers after DMS inactivation, yet this is clearly not the case. Additionally, temporary inhibition of direct pathway striatal neurons does not impair performance in a simple magnitude discrimination task, and likewise     137 inactivation of the ventral striatum does not impair selection of lever options associated with larger sucrose reward (Ferguson et al, 2013; Ghods-Sharifi and Floresco, 2010). Thus, there is little evidence to suggest that subjects were unaware of the specific outcomes associated with the different levers following DMS inactivation. Although this manipulation impaired accuracy on HR trials, choice of the LR option cannot be considered “advantageous”, as accuracy was also impaired on these easier trials. Alternatively, increased preference for the LR option, along with impaired accuracy, could reflect decreased motivation for sucrose reward. Indeed, prefeeding produces a similar behavioural profile as that observed following DMS inactivation, including a shift towards LR, impaired accuracy, and decreased completed trials (Cocker et al, 2012b). Again this explanation seems unlikely, as the decline in trials initiated was modest following inactivation, and latencies to collect reward - considered indirect measures of motivation for sucrose reward in the standard 5-CSRTT (Robbins, 2002) - were otherwise unaffected. Lesions to the DMS also do not affect progressive ratio performance for sucrose, and do not affect the amount of sucrose eaten during a consumption test (Eagle et al, 1999; Rogers et al, 2001).  Although the striatum plays an integral role in motor control, the task deficits observed here are unlikely to reflect general impairments in motor ability given the robust number of trials completed, faster response latencies, and increased premature responding seen after DMS inactivations.  4.4.1 Dorsomedial Striatum Contributions to Executive Control and     138 Decision Making Following inactivation of the DMS, rats displayed impaired accuracy and an increase in premature responding (Table 4.3), which together suggests that the striatum is necessary for the integration of cognitive and motor signals underlying the executive control required for optimal task performance. Subjects were unable to adequately time and organize their behaviour during a trial, such that they were quicker to make a choice, but were unable to wait for the light stimulus, or nosepoke at the appropriate (illuminated) aperture (Table 4.3). This account of the DMS deficit is in accordance with various theories of basal ganglia function, which suggest that the direct and indirect pathways are involved in context-appropriate action selection, sequence programming, as well as the initiation and termination of motor output (Macpherson et al, 2014; Schroll and Hamker, 2013). Our findings replicate previous work with the standard 5-CSRTT, in which lesions of the medial striatum produced long-lasting attentional impairments as well as impairments in response inhibition (Rogers et al, 2001), and accord with more recent pharmacological evidence suggesting that glutamatergic, dopaminergic and serotonergic signaling in this region regulates attention and response control (Agnoli et al, 2012; Agnoli and Carli, 2011, 2012). As noted above, this contrasts with lesions of the lateral striatum, which render rats incapable of engaging with the task (Rogers et al, 2001). Similar deficits are also observed following disconnection of the medial prefrontal cortex-dorsomedial striatal pathway, suggesting that a mPFC-DMS projection regulates aspects of     139 visual attention and response control (Christakou et al, 2001). Indeed deficits in attention and impulsivity induced by altered cortical signaling (via intra-mPFC NMDA antagonism) can be remedied by infusions of dopaminergic or serotonergic antagonists into the DMS (Agnoli et al, 2012; Agnoli and Carli, 2012), further emphasizing that, at baseline, cortical and striatal networks are fine tuned to optimize executive function and behaviour. Notably, bilateral dorsal mPFC lesions impair accuracy and response control, akin to lesions of the dorsal striatum (Chudasama et al, 2003b; Muir et al, 1996; Passetti, 2002; Rogers et al, 2001). Disrupting dorsal striatal activity may also affect its interactions with other cortical inputs converging on this region, such as the posterior anterior cingulate cortices, which regulate response inhibition but not attention (Muir et al, 1996). Thus lesions or inactivations of DMS may result in more widespread deficits relative to manipulations that target specific cortical regions. Indeed, inactivations of the anterior cingulate, prelimbic, and infralimbic cortices lead to distinct attentional or response control impairments on the rCET (Hosking et al, 2014a, 2015a), but the DMS inactivation here mirrored a global cortical dysfunction.  Unlike previous work with the 5-CSRTT, we show here that the attentional and response control deficits observed following DMS silencing are also accompanied by alterations in decision making (Table 4.3). DMS inactivation decreased choice of the hard option across both worker and slacker rats, raising an interesting question as to whether the shift in choice is driven by, or independent of, the impaired ability to complete the more cognitively difficult HR trials. Indeed, we have shown previously that decision-making impairments     140 following cortical inactivation are accompanied by performance deficits on the rCET, but that silencing of the BLA can affect decision making devoid of effects on attention and impulsivity (Hosking et al, 2014a, 2015a). We have also shown that these behavioural variables- choice vs attentional performance vs impulsivity- can be independently affected by pharmacological challenges (Cocker et al, 2012b; Hosking et al, 2014c; Silveira et al, 2016). As described above, the shift to LR trials does not reflect an advantageous decision in this instance, as rats were similarly incapable of completing LR trials. The current behavioural data may suggest then that rats are making a shift to LR choice based on their learned experience with the easier option, but are not updating this representation with their current ability to complete these trials.  While possible, another and perhaps more parsimonious explanation is that the dorsal striatum, like the ventral striatum, may play a role in enabling animals to overcome cognitive effort-related costs associated with larger rewards (Floresco et al, 2008a). To our knowledge, no previous studies have temporarily inactivated or lesioned the dorsal striatum to assess its contributions to effort-based choice. However, ventral striatal manipulations that affect valuations of physical effort allocation (e.g. intra-accumbens adenosine, acetylcholine, or dopamine neurotransmission manipulation) are ineffective when targeted toward a dorsal striatal site, suggesting that the contributions of this striatal zone to decision making with physical effort costs is minimal (Cousins et al, 1993; Farrar et al, 2010; Font et al, 2008; Nunes et al, 2013a). Indeed, human imaging data suggests that while the ventral striatum may be a common motivational node,     141 cognitive and physical effort may preferentially recruit the dorsomedial and dorsolateral striatum, respectively (Schmidt et al, 2012). The dorsal striatum may play a particularly important role in decisions based on cognitive effort costs, given widespread areas of the cortex project to this “associative” area, and aberrant signaling in this region can induce deficits in a number of cognitive processes, such as attention, working memory, and behavioural flexibility across species (Bradfield et al, 2013; Cools et al, 2006; Ragozzino, 2007; Rogers et al, 2001).  In line with this, a medial-prefrontal-dorsal striatal circuit has recently been shown necessary for cost/benefit valuations under approach-avoidance conditions. During a T-maze task, in which rats chose between rewards discounted by an aversive overhead light, optogenetic inhibition of a medial prefrontal-DMS pathway increased rats’ choice of the high-cost/high-reward arm (Friedman et al, 2015). Subsequent experiments revealed that activation of this pathway during the approach-avoidance task dampens striatal neuron activity via an inhibitory interneuron microcircuit, suggesting that under basal conditions DMS activation may bias choice towards more rewarding, yet costlier “avoidable” options. This is in line with more general evidence that activity in dorsal striatal neurons display task-related activity that is graded by reward magnitude (Cromwell, 2003; Hollerman et al, 1998; Kawagoe et al, 1998), and that phasic neuronal activity in this region covaries with action value and chosen value, thus providing a means by which striatal activity may guide choice and evaluate consequent outcomes (Kimchi and Laubach, 2009; Lau and Glimcher, 2007,     142 2008; Tai et al, 2012). Evidence that signals for previous goal choice also persist in the DMS provide further evidence that this region may be actively involved in the cost-benefit computations necessary for rCET decision making (Kim et al, 2013).  The effect of DMS inactivation on attentional, response control, and decision-making processes are also in accordance with evidence that corticostriatal “loops” may regulate distinct facets of task performance (i.e., choice, attention, response inhibition), the resulting output of which can then bias motor output via palladal-thalamic-cortical pathways (Chudasama and Robbins, 2006; Haber and Behrens, 2014). For example, while the mPFC-DMS circuit appears necessary for 5-CSRTT performance, the ACC-NAc pathway is not, even though lesions to each area increase premature responding on task (Christakou et al, 2001, 2004; Muir et al, 1996). In contrast, this ACC-NAc projection is necessary for cost/benefit valuations in which rats must decide to overcome physical barriers for larger reward (Hauber and Sommer, 2009). And while we focus our discussion here on corticostriatal pathways, we cannot preclude the contribution of other basal ganglia structures and striatal circuits to task performance. Indeed, the STN, as well as the projection between the mPFC and STN, regulate attention and impulsivity on the standard 5-CSRTT (Baunez, 1997; Chudasama et al, 2003a). This structure is also implicated in some forms of decision making (Adams et al, 2017b). Additionally, amygdala projections to the striatum are involved in reward processing, and it is possible such amygdalo-striatal inputs may guide decision making on the rCET (Namburi et al, 2015;     143 Stuber et al, 2011).    4.4.2 Nucleus Accumbens Inactivations Preclude Responding on Complex Cognitive Tasks In stark contrast to inactivations of the DMS, inactivation of the nucleus accumbens core region rendered rats incapable of performing the rCET (Table 4.2). For this reason meaningful interpretation of choice, accuracy, and premature responding data is limited. Visual inspection of the data suggests that rats decreased choice of the HR option following inactivation, but this was only a trend. Decreased choice of the HR option is in line with the general role of the ventral striatum in biasing animals towards high reward options, even when discounted by physical effort, delay, or uncertainty costs (Cardinal et al, 2001; Cardinal and Howes, 2005; Ghods-Sharifi and Floresco, 2010; Stopper and Floresco, 2011), and supports human neuroimaging data suggesting that activity in the ventral striatum varies as a function of the reward and costs associated with different options (Botvinick et al, 2009; Croxson et al, 2009; Levy et al, 2010).  While rats are still capable of performing the 5-CRSTT following lesions of the nucleus accumbens core region, emerging evidence suggests that temporarily inactivating the region prior to task onset may preclude task engagement (Feja et al, 2014). The discrepancy observed following lesions and inactivations of the nucleus accumbens highlights the compensatory mechanisms associated with lesion studies, which might explain why the two     144 techniques can ascribe different roles to the ventral striatum and the corticostriatal pathways targeting it (Christakou et al, 2004; Feja and Koch, 2015). It is unlikely that the cessation of behaviour reflects a decreased motivation for sucrose reward, given NAcC inactivations do not affect responding for sucrose on a FR1 schedule, and do not affect preference for larger sucrose reward in a reward magnitude discrimination test (Ghods-Sharifi and Floresco, 2010; Moscarello et al, 2010). Rather, the effects of NAcC inactivation on-task might reflect general impairments in motor function, such that subjects were not able to keep pace with the rapid nature of an rCET trial. Indeed, inactivations of the core, but not shell region, decrease locomotor activity and increase response latencies on a probability discounting task (Pothuizen et al, 2005; Stopper and Floresco, 2011).   Another possibility is that the cessation of rCET engagement following NAcC inactivation might reflect a profound inability for rewarding outcomes to invigorate approach or effortful behavior (Floresco, 2015). Lesions of the nucleus accumbens decrease lever pressing for increasing PR ratios, and, inactivations or lesions of the NAcC decrease choice of high effort/high reward options in T-maze or lever-based versions of physical effort-based discounting tasks (Ghods-Sharifi and Floresco, 2010; Hauber and Sommer, 2009; Moscarello et al, 2010). Importantly, rats are still capable of engaging in these tasks, presumably because of the low number of trials completed in the former, or the relatively minimal movement required in the latter.  In contrast, the rCET requires animals to execute a complex series of carefully timed motor movements involving both     145 response selection and inhibition, and this may be impossible with the NAcC completely off-line.   4.4.3 Conclusion When considered with previous studies completed using the rCET, the current investigation contributes to our understanding of the cortico-limbic-striatal circuitry mediating the allocation of cognitive effort, at least in the attentional domain. Indeed, the current findings may relate to the apathy observed in basal ganglia disorders such as Huntington’s disease, as well as the anergia and avolition observed in schizophrenia, whose etiology is based in part on deregulated corticostriatal signaling (Culbreth et al, 2016b; Duijn et al, 2014; Shepherd, 2013). Work with the rCET may identify the aberrant mechanisms giving rise to the motivational deficits characteristically observed in such patient populations, and may one day reveal ways in which these can be treated.     146 4.5 Tables   LR - Saline LR - Bac/Mus HR - Saline HR - Bac/Mus Choice latency 2.94 + 0.24 2.39 + 0.22 2.83 + 0.15 2.22 + 0.15 Correct latency 0.61 + 0.02 0.69 + 0.04 0.52 + 0.02 0.68 + 0.08 Collect latency 1.74 + 0.07 2.22 + 0.31 1.63 + 0.09 1.81 + 0.14 Choice omissions (For Session) 3.54 + 0.97 4.57 + 1.03 Trials Initiated (For Session) 139. 23 + 4.49 124.50 + 7.26 Trials Completed  (For Session 121.46 + 5.42 77.85 + 6.04  Table 4.1. Other rCET behavioural measures following DMS inactivation Means are presented (+ SEM). Values presented for choice omissions, trials initiated, and trials completed reflect data for saline and inactivation session.  # p < .1; * p < .05; ** p < .001.      LR - Saline LR - Bac/Mus HR - Saline HR - Bac/Mus Accuracy (%) 88.46 + 2.64 65.47 + 12.33 66.02 + 3.77 42.62 + 11.68 Premature (%) 15.61 + 4.09 6.37 + 2.80 15.53 + 3.49 13.98 + 4.29 Choice latency 2.46 + 0.23 2.48 + 0.35 3.20 + 0.31 2.25 + 0.25 Correct latency 0.61 + 0.05 1.02 + 0.26 0.55 + 0.03 0.61 + 0.06 Collect latency 1.82 + 0.18 2.64 + 0.37 1.86 + 0.24 2.67 + 0.91 Choice omissions (For Session) 5.16 + 1.37  2.92 + 0.92  Table 4.2. Other rCET behavioural measures following nucleus accumbens core inactivation Means are presented (+ SEM). Values presented for choice omissions reflect data for saline and inactivation session. # p < .1; * p < .05; ** p < .001.          147  Dorsomedial striatum Nucleus accumbens core  Choice of HR   ê  ê#	   Accuracy (LR/HR)   ê/	  ê  - / -  Premature responding (LR/HR)   é/é  ê#/	  -­‐	   Omissions at choice onset   -  -  Response omissions following stimulus presentation (LR/HR)    é/é   é/é  Latencies to make choice (LR/HR)   ê/	  ê  -­‐	  /	  ê  Latencies to make a correct response (LR/HR)   é/é  - / -  Latencies to collect reward (LR/HR)  - / -   - / -  Trials initiated  -   ê  Table 4.3. Summary of behavioural effects observed following dorsomedial striatum and nucleus accumbens core inactivation HR = high effort/high reward; LR = low effort/low reward. Note that the severe behavioural disruptions observed following NAcC inactivation limit a meaningful analysis of choice patterns, attention, and impulsivity. # denotes a trending effect.     148 4.6 Figures                 Figure 4-1. Histological analysis of cannulae implantation Location of all acceptable dorsal striatal (a) and nucleus accumbens core (b) infusions. Coordinates are relative to bregma. Plates modified from Paxinos and Watson (1998).     149  Figure 4-2. Effects of DMS Inactivations on (a) choice of HR trials (%), (b) accuracy, (c) premature responses (%), and response omissions (%) Data shown are the mean for each variable (±SEM). LR: Low effort/low reward trial; HR: High effort/high reward trial. DMS inactivation decreased choice of the HR option across both groups, and impaired accuracy for both trial types. DMS Inactivation also increased premature responding and response omissions for LR and HR trials. # p < .1; * p < .05; ** p < .001     150     Figure 4-3. Effect of nucleus accumbens core (NAcC) inactivations on (a) trials initiated, (b) response omissions (%), and (c) choice of HR trials (%). 	  Data shown are the mean for each variable (±SEM). LR: Low effort/low reward trial; HR: High effort/high reward trial. NAcC inactivation decreased trials initiated and increased response omissions across both groups. There was a trending decline in HR choice across rats (individual choice data inlaid). # p < .1; * p < .05; ** p < .001       151 Chapter 5: Investigating Orbitofrontal Cortex Contributions to Decision Making Involving Cognitive Effort Costs   5.1 Introduction  The student choosing to study for an exam, the worker spending late nights at the office, and the scientist who devotes years of work towards a single research question are all similar in that they have decided to allocate their cognitive resources towards a particular task, in the hopes that it will lead to lucrative future outcomes (e.g. degrees, promotions, scientific discovery). In all cases described above, executive processes constitute a limited resource, and so engagement in a particular cognitive episode precludes effort allocation towards alternative options that might yield beneficial, albeit less rewarding, outcomes. The decision to expend cognitive effort can thus be parlayed as an economic decision, whereby cognitive resources will only be allocated to a particular task if the expected value of the goal is high, thereby offsetting the opportunity costs incurred from persevering with a cognitively effortful option (Botvinick and Braver, 2015; Westbrook and Braver, 2015)   Accordingly, a number of regions have been implicated in the processes by which the brain mediates such valuations of mental effort allocation, but none have received as much attention as the dorsal anterior cingulate cortex (dACC)     152 (Shenhav et al, 2013). Previous research has shown that dACC blood-oxygen-level dependent (BOLD) signals reflect the subjective value of a chosen cognitively effortful reward, and that activity in this region is higher during cognitive effort versus delay discounting tasks (Chong et al, 2017; Massar et al, 2015). More generally, objective measures of task difficultly correlate with activity in the ACC, and stimulation of this region can elicit a feeling of “gearing up” for an effortful challenge, whereas lesions to this region can abolish conscious feelings of mental effort (McGuire and Botvinick, 2010b; Mulert et al, 2005; Naccache et al, 2005; Parvizi et al, 2013).  Recently, our group has also demonstrated a causal role for the ACC in valuations of cognitive effort, using a rodent model of cost-benefit decision making known as the rodent Cognitive Task (rCET). In this task, rats choose between two levers associated with easy versus hard attentional trials, whose successful completion is associated with one or two sugar pellet rewards, respectively (Cocker et al, 2012b). In agreement with the human cognitive effort literature, inactivations of the rodent dorsal mPFC (including prelimbic and pregenual Cg1 regions, corresponding to areas of the primate dACC (Heilbronner et al, 2016)) uniformly decrease choice of the more effortful option on the rCET, regardless of individuals’ preference for hard or easy trials at baseline (Hosking et al, 2014a, 2015a).      Thus, while dorsal prefrontal cortical areas have been implicated in cognitive effort allocation across species, few, if any, studies have implicated the orbital frontal cortices in such processes. Indeed, the animal cost/benefit     153 decision-making literature generally supports dissociable roles for the ventral and dorsal cortical areas in decision making, in which lesions or inactivations of the anterior cingulate cortex affect physical or cognitive effort-based valuations (Croxson et al, 2009; Rudebeck et al, 2006; Walton et al, 2003b), while similar manipulations to the orbital frontal regions selectively affect decision making involving delay, risk, and probability costs (Barrus et al, 2017; Orsini et al, 2015a; Stopper et al, 2014; Winstanley, 2004; Winstanley and Floresco, 2016). Similar dissociable roles for dorsal and ventral prefrontal areas in goal-directed choice are also observed in the human literature (Mobini et al, 2002; Prevost et al, 2010; Rangel and Hare, 2010; Rushworth et al, 2011a).   However, the role of the orbital frontal regions in cognitive effort allocation has never been causally investigated, and the tendency to consider effort as a unitary construct may mask distinct neural substrates regulating cognitive versus physical effort. Indeed, there is mounting evidence across species that these forms of decision making are mediated by unique mechanisms (Chong et al, 2017; Hosking et al, 2015b; Schmidt et al, 2012), and our own with work with the rCET suggests the orbital areas may play a selective role in cognitive effort-based decision making. For one, rats display a diverse propensity to “work” or “slack” at baseline, a behavioural distribution which is not generally observed in the physical effort literature (although see Randall et al, 2012), given most subjects in these paradigms willingly scale a barrier or lever press at high ratios for larger reward (Cocker et al, 2012b). A review of the animal model decision-making literature     154 has led to a recent proposal that the OFC is recruited in situations where task parameters make the objectively best option unclear, such that subjective evaluation of different outcomes is most pronounced (Winstanley and Floresco, 2016). Given baseline choice patterns on the rCET are not driven by differences in cognitive ability (i.e. attention and response control), the OFC may be a critical region for establishing subjective preference for cognitive episodes associated with lucrative outcomes.         Secondly, the OFC sends dense projections to the dorsomedial striatum, and inactivation of this region decreases choice of the hard lever on the rCET, suggesting that at baseline this striatal zone may bias choice towards high effort, high rewards options (Mailly et al, 2013; Silveira MM, Tremblay M, 2016).  Thus, signals from the OFC, along with input from other prefrontal cortical, thalamic, and limbic areas may interact to bias activity in the dorsomedial striatum to optimize decisions with cognitive effort costs.   In this study, we use standard pharmacological inactivation techniques to test whether the orbitofrontal cortices play a causal role in deciding how to allocate cognitive resources for beneficial outcomes. Although there is increasing appreciation of the functional heterogeneity within the rat PFC as it relates to reward learning and decision making (Izquierdo, 2017), we focus on the ventral lateral orbital frontal cortex (VLO). This area is the most often targeted in other rodent decision-making paradigms, thus facilitating direct comparisons of how this region contributes to cost/benefit decision making with distinct costs.       155 5.2 Additional Methods  5.2.1 Subjects and Baseline Choice  Subjects were 24 female Long-Evans rats bred in house. These rats were bred from breeding pairs obtained from Charles River Laboratories and the Rat Resource and Research Centre (RRRC, Columbia, MO) as part of a breeding program for transgenic rats that express cre recombinase (Cre) in neurons that contain choline acetyltranferase (ChAT; Long Evans –Tg (ChAT-Cre) 5.1 Deis, RRRC # 00658). Twelve of these rats were positive for the transgene (TG+), while the remainder did not express the transgene (TG-). The transgenic status of the rats was not utilized for the current study, but instead the subjects were considered naïve given they exhibit a similar behavioural profile to Long Evans rats purchased from a commercial supplier. To account for potential differences, transgene status (TG + or TG -) was included as a between-subjects factor for all analyses. These results will not be reported, unless a significant a main effect of TG status or interaction with TG status is observed for any of the rCET behavioural measures. Animals weighed at least 200 g at the start of the experiment and were food restricted to 85% of their free-feeding weight (maintained on 10 g rat chow daily). Water was available ad libitum. Rats were housed in groups of three or four in a climate controlled colony room maintained at 21° C on a reverse 12-hr light-dark schedule (lights off at 8 am).  For this particular experiment, the mean choice of the HR option across rats was 66%. Animals were grouped as “workers” if they chose the HR option     156 for >70% of trials (n=12) and as “slackers” if they chose HR for ≤70% of trials (n=12), as per previous work (Cocker et al., 2012), thereby enabling consistency when discussing individual differences across studies.  5.2.2 Surgery to Implant Guide Cannulae Following stable baseline performance, animals were implanted with 22-gauge stainless steel guide cannulae (Plastics One) bilaterally into the lateral orbitofrontal cortices using standard stereotaxic techniques. Animals were anesthetized with 2% isoflurane in O2 and cannulae implanted at the following coordinates: AP = +3.8 mm, ML = ±2.6 mm from bregma, DV = −2.6 mm from dura (Paxinos and Watson, 1998). The cannulae were fixed to the skull by 3-4 stainless steel screws and dental acrylic. Obdurators flush with the cannulae tip were inserted to prevent clogging prior to testing. Animals recovered for at least 5 days before resuming behavioural testing. Two rats died during surgery, and so 22 rats moved on to microinfusions.    5.2.3 Microinfusion Procedure Once recovered, rats performed 10 free-choice sessions, at which point their behaviour was deemed statistically stable (see data analysis). To habituate rats to the microinfusion process, two mock infusion sessions were carried out prior to the critical testing sessions. During these mock infusions, obdurators were removed and 28-gauge injectors extending 1 mm beyond the guide cannulae were inserted for two minutes, but no infusion occurred. Ten minutes     157 following this, rats were tested on the rCET. Infusions adhered to a 3-day cycle, starting with a baseline session, followed by a drug or saline injection, and ending with a non-testing washout day. The LOFC was inactivated by a mixture of the GABAB agonist baclofen and the GABAA agonist muscimol (Sigma–Aldrich), prepared separately at 0.5 µg/µL in saline, and mixed together in equal volumes to form a 0.25 µg/µL solution. 0.5 µL per hemisphere injections of saline or baclofen/muscimol (i.e. 0.125 µg of drug per hemisphere) were administered bilaterally at a rate of 0.4 µL/min, and injectors (injectors extending +1 mm from cannula tips) were left in place for an additional minute to allow diffusion. Once the microinfusions were completed, injectors were removed, obdurators replaced, and animals were returned to their home cages for 10 min before being placed in the operant chambers and performing the rCET. Animals underwent two infusions in total: on the first infusion day, half of the rats received saline and the other half received baclofen/muscimol These infusions were reversed on the second infusion day, thus permitting a within-subjects comparison of how silencing the LOFC affects rCET performance. Three rats did not have data from at least one microinfusion day due to health complications, and so they were excluded from the study. Thus complete data was obtained from 19 subjects.    5.2.4 Histology  When behavioural testing terminated, animals were anesthetized with 2% isoflurane in O2 and then euthanized via C02. Brains were extracted and submerged in 4% paraformaldehyde for 24h, before being transferred to a 30%     158 sucrose solution where they remained until sectioning. Brains were frozen and cut into 40-µm coronal sections using a cryostat. The sections were placed on gelatin-coated slides and stained with cresyl violet for visualization. The projected locations of the injector tips protruding from the guide cannulae were mapped onto the coronal section images from Paxinos and Watson (1998).    5.2.5 Data Analysis Data were analyzed with a repeated measures ANOVA with session (three levels: baseline sessions 1-3) or inactivation (two levels: vehicle or inactivation) as within-subjects factors, and group (two levels: worker or slacker) as a between-subjects factor for all analyses. TG status (two levels: TG+ or TG-) was also included as a between-subjects factor all analyses. These results will not be reported unless there is a significant main effect of TG status, or interaction with TG status for any of the variables analyzed.       159  5.3 Results   5.3.1 Cannula Placements The locations of all acceptable placements are depicted in Figure 5.1. Two rats were excluded because the infusions were ventral to the OFC, leaving a total of 17 rats (n = 7 workers and n = 10 slackers) for analyses.    5.3.2 Inactivation Analyses Choice, accuracy, and premature responses.    Baseline performance on the rCET has been discussed in detail previously (Cocker et al, 2012b), and so will be briefly discussed here. As a group, rats chose the high-effort/high-reward (HR) option more than the low-effort/low-reward (LR) option when the orbitofrontal cortex (OFC) was infused with saline (saline only- choice: F(1, 13) = 43.473, p < .001).  In line with their categorization, workers (M = 86.70%) chose the HR option significantly more than slackers (M = 51.53%) (saline only- group: F(1, 13)  = 39.252, p < .001). Temporary inactivation of the OFC did not affect choice of the HR option across both groups of rats (inactivation: F(1, 13) = .963, NS; inactivation x group: F(1, 13) = 0.151, NS; Figure 5.2a).  Accuracy on HR trials was lower than accuracy on LR trials, suggesting that HR trials were indeed more difficult for subjects to attend to (saline only- choice: F(1, 12) = 56.701, p < .001). In line with previous work on the task,     160 workers and slackers did not differ in attentional performance (saline only- choice x group: F(1, 12) = 3.402, NS; group: F(1, 12) = .411, NS). This suggests that slackers’ reduced preference for HR was not due to a profound inability to successfully attend to the .2s stimulus. Inactivation of the OFC did not affect accuracy on LR (LR trials- inactivation: F(1, 12) = 0.155, NS; inactivation x group: F(1, 12) = 1.75, NS) or HR trials (HR trials- inactivation: F(1, 11) = 2.293, NS; inactivation x group: F(1, 11) = 0.087, NS) for either group (Figure 5.2b).  Premature responding did not differ between LR or HR trial types (saline only- choice: F(1, 12) = 0.185, NS), but groups did differ at baseline on rates of premature responding (saline only- choice x group: F(1, 12) = 7.036, p = .021). While workers and slackers did not differ in rates of premature responding on HR trials (HR trials-saline only: F(1, 13) = 0.81, NS), there was a trend for workers to be more impulsive on easy, LR trials (LR trials-saline only: F(1, 12) = 3.494, p = .086), This trend was driven by one worker rat who infrequently made LR choices, but who acted impulsively on most of these trials. In any case, slackers’ reduced choice of HR cannot be explained by an inability to wait for the .2s stimulus on HR trials. Across all animals, inactivation of the OFC robustly increased premature responding for both trials types (LR trials- inactivation: F(1, 12) = 17.286, p < .001; inactivation x group: F(1, 12) = .372, NS; HR trials- inactivation: F(1, 11) = 5.187, p = .044; inactivation x group: F(1, 11) = .434, NS, Figure 5.2c).   Other behavioural measures.      161 When infused with saline, rats initiated 130 trials on average, and this did not differ between worker and slacker rats (saline only-group: F(1, 13)  = 0.946, NS). OFC inactivation decreased trials initiated by both groups, but rats still initiated over 100 trials during the experimental test session (inactivation: F(1, 13) = 13.128, p = .003; inactivation x group: F(1, 13) = 0.896, NS). Following trial initiation and extension of both HR and LR levers, rats typically made few choice omissions under baseline conditions and this did not differ between groups (saline only-group: F(1, 13)  = 0.73, NS). However, when the OFC was infused with baclofen/muscimol, choice omissions almost tripled (inactivation: F(1, 13) = 30.788, p < .001; inactivation x group: F(1, 13) = 0.187, NS, Figure 5.2d). Latencies to make a trial type choice, to correctly detect the visual stimulus, and to collect sucrose from the magazine tray following successful trial completion did not differ between LR or HR trial types, nor did they differ between workers and slackers when the OFC was infused with saline (saline only- choice/ choice x group/ group: all Fs < 4.571, NS). Following OFC inactivation, there was a trend for rats to be slower to initiate LR trials (choice latency: LR trials- inactivation: F(1, 12) = 4.430, p = .057; inactivation x group: F(1, 12) = 3.602, p = .082), but otherwise latencies were unaffected (HR choice latency, LR and HR correct latency, LR and HR collect latency- inactivation/ inactivation x group: all Fs < 2.608, NS). Previous work has shown that OFC inactivation increases general latencies to make a choice in a probability discounting task (St Onge and Floresco, 2010), and so we collapsed latencies across trial types and analyzed this general choice latency measure. However, OFC inactivation did not affect     162 overall choice latencies (inactivation, inactivation x group: all Fs < 1.902, NS). In line with the increased attentional demands of HR trials, response omissions following stimulus presentation were slightly higher for this trial type, but did not differ between worker and slacker rats (saline only- choice: F(1, 12) = 12.476, p = .004; choice x group/ group: all Fs < 2.441, NS). Unlike the pronounced increase in choice omissions, response omissions for either trial type were otherwise unaffected (LR and HR trials- inactivation / inactivation x group: all Fs < 3.378, NS). Infusions of the OFC significantly decreased the number of trials completed (inactivation: F(1, 13) = 47.957, p < .001; inactivation x group: F(1, 13) = 1.194, NS, Figure 5.2e), reflecting the rise in incomplete trials due to increased choice omissions and premature responding.       163 5.4 Discussion   The cost/benefit decision-making literature generally supports dissociable roles for dorsal and ventral cortical regions in effort allocation. Across species, the anterior cingulate cortices are implicated in the processes by which organisms decide to allocate effort towards lucrative outcomes, while no work has shown a causal role for the orbitofrontal cortices in these cost/benefit computations (Winstanley and Floresco, 2016). However, this literature has almost entirely focused on the physical effort costs that discount preferred rewards, and has neglected the cognitive effort costs that are arguably more representative of the decisions individuals face in modern societies. Like physical effort, cognitive effort is an inherently costly resource whose deployment can be biased by incentives, suggesting it is amenable to the cost/benefit framework that has been used to study physical effort allocation (Westbrook and Braver, 2015). Here we directly investigated the role of the OFC in mediating valuations of cognitive effort allocation while rats performed the rodent cognitive effort task (rCET). Inactivations of the OFC did not affect rats’ willingness to exert cognitive effort for larger rewards, nor did they affect the attentional ability required to complete difficult, high reward trials. However, inactivations selectively and dramatically increased choice omissions, and in keeping with previous literature, increased measures of impulsive action (Chudasama et al, 2003b). The effects are in striking contrast to previous work on the rCET, in which inactivation of medial and dorsal cortical areas simultaneously affect willingness to exert     164 cognitive effort as well as executive performance (Hosking et al, 2014a, 2015a). Indeed, the current work suggests that cognitive performance deficits induced by cortical perturbation are not necessarily accompanied by shifts in effort allocation. More generally, this work provides further support for a dorsal/ventral divide in effort-based decision making, with dorsal, but not ventral, prefrontal regions recruited when decision costs require effort expenditure in the physical or cognitive domain.   Before discussing the behavioural effects observed following OFC inactivation, it is worth speculating whether our infusion protocol was sufficient to produce changes in choice behaviour.  The pharmalogical inactivations we carried out encompass a relatively small region (~1mm, see Floresco, 2006) relative to the lesions previously used to investigate the role of the OFC in physical effort-based decision making (Ostrander et al, 2011; Rudebeck et al, 2006). However, the infusion protocol is identical to those previously used to investigate prelimbic, infralimbic, and ACC contributions to rCET performance (Hosking et al, 2014a, 2015a). As will be discussed below, inactivations of these regions alter effort allocation while concomitantly affecting other measures of performance. While OFC inactivation did not affect decision making, it had pronounced effects on premature responding and choice omission rates, suggesting the infusions were selective and behaviourally effective.  Our infusions targeted the lateral orbitofrontal cortices, but there is growing recognition that the orbitofrontal cortex is a functionally heterogenous structure, and that behavioural neuroscientists may oversample some areas     165 while undersampling from others (Izquierdo, 2017). We focused on the lOFC because most of the previous cost/benefit decision making literature has targeted this region, thus allowing us to easily compare and contrast our observed effects with those seen in tasks probing decision making with risk (Orsini et al, 2015a), uncertainty (Barrus et al, 2017; St Onge and Floresco, 2010; Zeeb and Winstanley, 2011), delay (Mar et al, 2011; Winstanley, 2004), and most importantly effort (Ostrander et al, 2011; Rudebeck et al, 2006) costs. Additionally, this area was previously lesioned on the 5-choice serial reaction time task, the predecessor of the rCET, thus allowing us to directly relate our behavioural effects on attention and impulsivity with those seen previously (Chudasama et al, 2003b). Some researchers have investigated medial orbital (MO) contributions to decision making under conditions of uncertainty (Stopper et al, 2014) and delay (Mar et al, 2011) and reported effects on choice. While lesions targeting the infralimbic/prelimbic areas and encompassing MO do not affect physical effort allocation (Walton et al, 2003b), this may be an interesting target for future studies on the rCET.  We opted to use female rats in the current study, given that they display similar patterns of behaviour on the rCET to the more commonly used males. Indeed, choice of the HR lever in this study was 66 %, while HR choice has ranged from 56% (Hosking et al, 2014a) to 73% (Hosking et al, 2014b) across male cohorts (HR choice is typically ~70%). Accuracy and premature responding rates are also similar to those typically observed in male cohorts. While we did not track females’ estrous cycle, previous research using a risk discounting task     166 and lever-based physical effort discounting task have shown that this does not affect baseline choice patterns (Orsini et al, 2015b; Uban et al, 2012). It could be argued that a different pattern of behavioural results might have been observed if males were used. This is possible, in light of research showing c-fos expression in the lateral OFC is correlated with choice in a rodent analog of the Iowa Gambling Task in females, but not males (van Hasselt et al, 2012), as well as some human imaging studies showing differential OFC activity between sexes in risk-based decision making tasks (Bolla et al, 2004; Lee et al, 2009). However, we are not aware of any preclinical or human literature suggesting that the OFC is differentially recruited between males and females in regard to effort allocation, and cannot find any example of a brain lesion or inactivation effect fundamentally affecting one sex selectively. Given we also replicated an impulse control deficit previously documented on the 5-CSRTT in males (Chudasama et al, 2003b), it is likely the current study can be directly compared to previous investigations on the rCET using male rats.   5.4.1 The OFC is Not Involved in the Cost/Benefit Computations Regulating Cognitive Effort Allocation   Following OFC inactivation, worker and slacker rats did not alter their baseline preference for the HR option. This is in agreement with work in the domain of physical effort allocation, in which OFC lesions similarly do not affect rats’ willingness to scale a barrier for larger rewards (Ostrander et al, 2011; Rudebeck et al, 2006). Rather, dorsal prefrontal regions, encompassing the     167 anterior cingulate cortex (ACC), play a central role in processes guiding effort-based decision making. In the physical effort domain, imaging studies have shown increased activity in this region in anticipation of an effortful physical challenge (Croxson et al, 2009) as well as correlates of subjective value as subjects choose between options differing in their effort/reward contingencies (Chong et al, 2017; Prevost et al, 2010).  Lesions of the ACC decrease rats’ willingness to scale a barrier (Rudebeck et al, 2006; Schweimer, 2005; Walton et al, 2002, 2003a), or to repeatedly press a lever for larger reward (Hart et al, 2017; Walton et al, 2009) -processes regulated by dopaminergic (Schweimer et al, 2005; Schweimer and Hauber, 2006) and cannabinoid signaling (Khani et al, 2015) in this region. In contrast, intra-OFC CB1 receptor signaling does not regulate valuations of physical effort, providing further evidence that this form of decision making is localized to the ACC (Khani et al, 2015). Likewise, cognitive effort allocation appears to be mediated primarily by dorsal prefrontal regions, as temporary inactivation of the region decreases effortful choice on the rCET (Hosking et al, 2014a). This is in accordance with imaging studies showing coordinated activity in the ACC in anticipation of a mental challenge (Botvinick et al, 2009), and aligns with recent work showing that the value of a cognitively effortful option is encoded in the ACC (Chong et al, 2017; Massar et al, 2015). Indeed, according to recent theories of effort allocation, the ACC plays an executive role in calculating the “expected value” of cognitive control, and uses this information to select among competing cognitive tasks in ways that are most worthwhile (Shenhav et al, 2013).       168  In contrast, the OFC appears to play a central role in other forms of decision making, including those involving delays (Kable and Glimcher, 2007; Mar et al, 2011; Prevost et al, 2010; Winstanley, 2004), uncertainty (Barrus et al, 2017; Mobini et al, 2002; St Onge and Floresco, 2010; Stopper et al, 2014), and risk of punishment (Orsini et al, 2015a). Inherent to all these costs is the fact that none of them have a perceptible effect on the decider at the time of choice, but manifest themselves only when the outcomes are revealed. Said another way, the OFC may preferentially mediate decisions in which costs are inherent to the outcome, while dorsal prefrontal regions are recruited when decision costs are incurred on the individual, akin to physical or cognitive effort (Rangel and Hare, 2010; Rushworth et al, 2011b). These dissociable roles in decision making may stem from the roles of the OFC and ACC in stimulus learning and action learning, respectively, as lesions to either area disrupt subjects’ ability to use feedback about visual stimuli or actions to learn about outcomes (Camille et al, 2011; Rudebeck et al, 2008).   While OFC inactivation did not affect choice profiles, it increased choice omissions three-fold. There was also a trending increase in latency to choose the LR option following inactivation, but unlike previous work we did not observe a general rise in latencies to make a choice following this manipulation (St Onge and Floresco, 2010). This striking rise in choice omissions is unlikely to reflect a motivational deficit, as rats initiated a high number of trials, response omissions following stimulus presentation were unaffected, and latencies throughout the trial did not change. There is a rich literature implicating the OFC in the     169 processes by which stimuli or cues can be used to predict expected (Baxter et al, 2000; Schoenbaum et al, 1998), as well as inferred future outcomes (Takahashi et al, 2013). Previous work in the realm of delay-based decision making has shown that cues can exert powerful effects on baseline choice profiles, effects which are mediated by the OFC (Mobini et al, 2002; Zeeb et al, 2010; Zeeb and Winstanley, 2011). Indeed, while OFC lesions do not affect rats’ willingness to scale a barrier for larger reward under normal conditions, one study has demonstrated that pre-training OFC lesions can affect physical effort allocation if cues predict reward availability in the high effort arm (Ostrander et al, 2011). In the rCET, the HR and LR lever are identical but vary in their spatial location, their cognitive demand, as well as their reward outcomes both within a trial (1 vs 2 sugar pellets) and across a session. To the extent that the OFC may be involved in associating outcomes to each lever, or in establishing a cognitive map of “task space” as has been recently proposed (Wilson et al, 2014), the increase in choice omissions following OFC inactivation might reflect a disruption in representing this information when both levers are extended. Aberrations in these processes might then affect the cost/benefit computations occurring in other prefrontal regions, such as the medial prefrontal and anterior cingulate cortices, which are specifically involved in allocating cognitive resources in worthwhile ways (discussed below). Indeed, some computational models have proposed a goal-processing “pyramid” in the brain, in which the OFC may relay stimulus identity information to the ACC, which then uses this information to guide motor control during goal pursuit (O’Reilly et al, 2014). Relative to medial     170 areas of the OFC, the LOFC we targeted sends more projections to the medial wall of the frontal cortex, and so connectivity supports this theoretical account of OFC-ACC communication during decision making (Hoover and Vertes, 2011).  In further support of this cortical hierarchy, OFC inactivations appear to preferentially play a role in decision making during initial acquisition when task contingencies (e.g stimulus-outcomes) are first being learned, but play less of a role once acquired (Zeeb and Winstanley, 2011), and a PET study in rats showed increased metabolic activity in the OFC while rats performed a physical effort allocation task, even though this region does not affect this type of decision making when lesioned (Endepols et al, 2010). Given the extensive training on-task, disrupting OFC signaling may affect the component processes required to make an effort-based decision, but ultimately other prefrontal areas can compensate when this information is otherwise offline.   5.4.2 Cortical Contributions to Decision Making on the rCET  Previous work on the rCET has demonstrated that the medial wall of the rodent prefrontal cortex plays an important role in allocating attentional resources for lucrative outcomes. Specifically, selective inactivations of the prelimibic, inflralimbic, and anterior cingulate cortex all decrease choice of the HR option in both worker and slacker rats (Hosking et al, 2014a, 2015a). However, these shifts in decision making are also associated with deficits in the executive processes necessary to successfully complete trials and gain reward: inactivations of the prelimbic cortex selectively affect the ability to attend to LR     171 and HR stimuli (Hosking et al, 2015a), while inactivations of the infralimbic and anterior cingulate cortices increase premature responding (Hosking et al, 2014a, 2015a), thus precluding stimulus presentation. While attentional difficulty is explicitly manipulated between easy and hard trials, rats ultimately must exhibit a certain degree of restraint in order for the stimulus to appear. Thus, deficits in both impulse control and/or attentional ability might affect the processes by which rats choose between easy and hard response alternatives during the rCET. Notably, any attentional or impulse control deficits that have been observed following inactivation of these cortical regions are observed on both easy and hard trials, suggesting that the shift in choice toward the cognitively easier option is not the adoption of an “advantageous” choice strategy. Instead, we speculate rats are defaulting to the easier trials when executive processes are impaired, based on their previous experience with these trial types.  Unlike the simultaneous effects on cognitive allocation and ability following inactivation of the infralimbic, prelimbic, and anterior cingulate cortices, the current OFC findings suggest that at the level of the prefrontal cortex, cognitive allocation and ability can in part be dissociated. OFC inactivation did not affect effort allocation, but robustly increased impulsive action, replicating previous lesion work on the 5-CSRTT (Chudasama et al, 2003b). Thus, rats were impaired in their ability to complete successfully complete trials, but this did not affect their decision-making profiles. This is in contrast to the role of the basolateral amygdala on the rCET, which does not affect measures of performance, but appears to regulate rats’ baseline willingness to “work” or “slack” at baseline     172 (Hosking et al, 2014a). The observed effects are also distinct from inactivations of the dorsomedial striatum, which simultaneously decrease HR choice, impair accuracy, and increase impulsivity (Chapter 4), in line with this being a downstream cortical target that may integrate the various cortical signals used to rCET performance. Considered together, this work suggests that while many cortical areas regulate optimal performance on the rCET, the medial wall plays a selective role in allocating cognitive effort in pursuit of desired outcomes, above and beyond its role in attention and impulse control.   5.4.3 Conclusion  To summarize, the present findings suggest that valuations of cognitive effort allocation are not dependent on OFC signaling. Rather, the current work supports evidence for a dorsal-ventral divide in decision-making processes, whereby cognitive effort costs are processed by medial prefrontal structures encompassing the prelimbic, infralimbic, and anterior cingulate cortices in rodents. This is in agreement with current conceptualizations of cognitive effort allocation in humans, which attribute a prominent role to the anterior cingulate cortex in adjudicating cognitive resources for maximal gain (Shenhav et al, 2013). How cognitive effort allocation is specifically mediated by interactions with the ACC and other cortical and subcortical regions is still unknown, but future work on the rCET holds great promise towards delineating the cortical-striatial-limbic interactions guiding effort allocation in the cognitive domain.         173  5.5 Table 	   LR	  -­‐	  Saline	   LR	  -­‐	  Bac/Mus	   HR	  -­‐	  Saline	   HR	  -­‐	  Bac/Mus	  Choice	  latency	   3.28	  +	  0.26	   3.69	  +	  0.22#	   3.28	  +	  0.22	   3.40	  +	  0.22	  Correct	  latency	   0.53	  +	  0.02	   0.77	  +	  0.11	   0.52	  +	  0.03	   0.68	  +	  0.09	  Collect	  latency	   5.04	  +	  3.12	   1.82	  +	  0.09	   1.45	  +	  0.07	   1.52	  +	  0.12	  	  Trials	  Initiated	   (For	  Session)	   130.06	  +	  5.47	   108.06	  +	  7.93	  Trials	  Completed	  	   (For	  Session)	   103.65	  +	  6.46	   52.65	  +	  6.73	    Table 5.1. Other rCET behavioural measures following OFC inactivation  Means are presented (+ SEM). Values presented for trials initiated and trials completed reflect data for saline and inactivation session. # p < .1; * p < .05; ** p < .001.       174 5.6 Figures   Figure 5-1. Histological analysis of cannulae implantation Location of all acceptable orbitofrontal cortex infusions. Coordinates are relative to bregma. Plates modified from Paxinos and Watson (1998).        175  Figure 5-2. Effects of OFC inactivations on (a) choice of HR trials (%), (b) accuracy (%), (c) premature responses (%), (d) choice omissions, and (e) response omissions (%) Data shown are the mean for each variable (±SEM). LR: Low effort/low reward trial; HR: High effort/high reward trial. OFC inactivation increased premature     176 responding and choice omissions but did not affect choice, accuracy, or response omissions.     177 Chapter 6: Investigating Amygdala-Prefrontal Cortical Circuit Involvement in the Allocation of Cognitive Effort     6.1 Introduction   Humans must frequently evaluate the amount of effort that must be invested for potential rewards. Research in this area of cost/benefit decision making has focused on the physical effort costs that discount lucrative outcomes, but recent work has emphasized the cognitive effort costs that are arguably more representative of the decisions individuals face in modern societies. Current theories conceptualize cognitive effort allocation as a neuroeconomic choice, whereby mental processes will only be allocated to a particular task if the expected value of the goal is high, thus offsetting the opportunity costs incurred from persevering with an effortful option in lieu of simpler, less rewarded alternatives (Shenhav et al, 2013; Westbrook and Braver, 2015). In this framework, cognitive effort costs are those that are non-physical in nature and tax limited executive resources, as observed through psychological constructs such as working memory, attention and response inhibition (Schmidt et al, 2012). While such allocation processes may seem trivial when considering the routine decisions of day-to-day life, the consequences of not applying cognitive effort can be quite severe, as observed by the angeria, avolition, and impairments in effort     178 allocation inherent to conditions such as depression and schizophrenia (Culbreth et al, 2016a; Duijn et al, 2014; Gold et al, 2013; Treadway et al, 2009).      Two areas known to be critical for cost/benefit decision making are the anterior cingulate cortex (ACC) and basolateral amygdala (BLA). Patients with lesions to either area display deficits in tests of real-world decision making, in which the magnitude of rewards and costs are varied across options (Bechara et al, 1999; Brand et al, 2007; Manes et al, 2002; Naccache et al, 2005). Single unit recordings similarly suggest that BLA and ACC neurons may register reward, cost, and value information which could be used to guide choice behavior (Belova et al, 2008; Jenison et al, 2011; Shidara and Richmond, 2002). Regarding effort-based decision making, stimulation of the ACC can elicit a feeling of “gearing up” for an effortful challenge, whereas lesions to this region can abolish conscious feelings of mental effort (Mulert et al, 2005; Naccache et al, 2005; Parvizi et al, 2013). Functional imaging studies also demonstrate coordinated activation in the ACC in anticipation of effort exertion (Botvinick et al, 2009; Kurniawan et al, 2013; Vassena et al, 2014) and signals encoding the subjective value of a physically or cognitively effortful option have been observed in both the ACC and BLA (Chong et al, 2017; Massar et al, 2015; Prevost et al, 2010).   Recently, we have demonstrated a causal role for the ACC and BLA in the processes by which rodents must decide how to allocate attentional effort for lucrative outcomes, by using a rodent model of cost/benefit decision making     179 known as the rat Cognitive Effort Task (rCET) (Hosking et al, 2014a). The task is a modified version of the standard 5-choice serial reaction time task (Bari et al, 2008), in which rats can decide at trial outset whether to respond for a visuospatial stimulus that is either easy (1s) or hard (0.2s) to detect across five possible locations, and where successful completion of easy or hard trials results in one or two sugar pellets, respectively (Cocker et al, 2012b). Bilateral inactivations of the ACC uniformly decrease choice of the high-effort/high reward option (.2 s stimulus/ 2 pellet reward; HR), thus mimicking the effect ACC perturbations have on rodent tasks of physical effort which require rats to scale a barrier (Rudebeck et al, 2006; Schweimer, 2005; Walton et al, 2003b) or repeatedly press a lever (Hart et al, 2017; Walton et al, 2009) to access larger reward. In contrast, BLA inactivations move subjects away from their propensity to “work” or “slack” at baseline on the rCET: rats initially preferring the hard option at baseline decrease selection of this option, whereas rats opting for the easier of the two options at baseline now increase hard lever choice following BLA inactivation (Hosking et al, 2014a). These findings are in contrast to the observation that BLA inactivation decreases choice of physically effortful options across rats (Floresco and Ghods-Sharifi, 2007; Ghods-Sharifi et al, 2009b; Hart and Izquierdo, 2017). These nuances in BLA contributions to choice involving the evaluation of different forms of effort are particularly interesting given recent imaging work in humans suggesting that activity in the right amygdala appears to process subjective value uniquely for rewards associated with cognitive, but not physical effort costs (Chong et al, 2017).     180 However, interpreting the contributions of these regions to effort-based decision making in isolation lacks external validity (Floresco et al, 2008a). Rather, it is important to assess how different brain areas, such as the ACC and BLA, communicate to bias choice towards options differing in cognitive effort. Imaging work has provided some insight into the pathways regulating effort allocation, but these findings are correlational and have exclusively focused on corticostriatal pathways (Botvinick et al, 2009; Croxson et al, 2009; Schmidt et al, 2012). A previous study has demonstrated that contralateral disconnection of the BLA-ACC pathway decreases rats’ willingness to scale a barrier for larger reward, suggesting that under normal conditions this circuit biases subjects towards lucrative, physically effortful options (Floresco and Ghods-Sharifi, 2007). However, given differences in the role of the BLA in physical versus cognitive effort, and growing appreciation that these manifestations of effort are regulated by dissociable neural mechanisms, it is important to address the role of the BLA-ACC signaling to decisions with cognitive effort costs (Hosking et al, 2014b).   The current study utilizes a disconnection procedure to assess how ACC-BLA communication regulates aspects of cognitive effort allocation. Specifically, unilateral inactivation of the ACC is paired with unilateral inactivation of the contralateral BLA, thus disrupting serial communication between these structures in each hemisphere. Variations of this technique have been employed to successfully disrupt serial transfer of information between brain regions in both rodents (Chudasama et al, 2003a; St. Onge et al, 2012; Zeeb and Winstanley, 2013) and nonhuman primates (Baxter et al, 2000; Eldridge et al, 2015). We     181 hypothesized that BLA-ACC disconnection would mimic the effects of bilateral BLA inactivation on choice, suggesting that the BLA may send information about rats’ subjective preference for easy or difficult mental challenges to the ACC, which then uses this information to compute whether a lucrative outcome is worth the effort (Shenhav et al, 2013). To preview, we failed to confirm this hypothesis. In the course of carrying out the inactivations, we discovered that a single unilateral inactivation of the BLA was sufficient to affect cognitive effort allocation in a subset of rats. Although this finding precludes a discussion of how BLA-ACC circuitry regulates cognitive effort allocation using the current design, it does suggest that the BLA plays a critical role in this type of decision making.   6.2 Additional Methods  6.2.1 Baseline Choice  Thirty-seven Long-Evans rats were used in the currently study. For this particular experiment, the mean choice of the HR option across rats was 69% following post-operative baseline. Animals were grouped as “workers” if they chose the HR option for >70% of trials (n= 20) and as “slackers” if they chose HR for ≤70% of trials (n= 17), as per previous work (Cocker et al., 2012), thereby enabling consistency when discussing individual differences across studies.  6.2.2 Guide Cannulae Implantation       182  Thirty-seven rats were run on the free choice version of the rCET daily, at which point their behaviour was stable for at least three sessions (session main effect and interactions p > .05 when a repeated measures measures ANOVA was carried out on the choice, accuracy, and premature responding variables). Animals were anesthetized with 2% isoflourane in O2 and administered .1 ml of 5 mg.ml anafen and 10 ml lactated ringer subcutaneously. Rats were then secured in a stereotaxic frame (David Kopf) with the incisor bar set at -3.3. When a surgical plane of anesthesia was reached, rats were implanted with guide cannulae targeting the ACC or BLA. Twenty-two gauge bilateral cannulae were used to target the ACC in both hemispheres, whereas a single 22-gauge cannula was aimed at either the left or right BLA. The hemisphere chosen for BLA cannulation was counterbalanced across rats, accounting for worker/slacker and hard lever (left or right) designations to ensure a similar number of left or right BLA were targeted across these between-subjects factors. The coordinates (Paxinos and Watson, 1998) for ACC and BLA cannulation were as follows: ACC: anterioposterior (AP) = + 2.0 mm, mediolateral (ML) = + .75 mm from bregma, dorsoventral (DV) = - .8 mm from dura; BLA: AP = -3.1 mm, ML = + 5.2 from bregma, DV: - 6.5 mm from dura. These coordinates are similar to the ones used previously to independently target these structures on the rCET (Hosking et al, 2014a). Cannulae were fixed to the skull by 3-4 stainless steel screws and dental acrylic. Obdurators flush with the cannulae tip were inserted to prevent clogging prior to testing. Animals recovered for at least 5 days before resuming behavioural testing.       183 6.2.3 Disconnection Design          Disconnections are a sophisticated method to assess the functional interaction between two brain regions. They rely on the assumption that information flows serially from one area to an efferent region in both hemispheres in parallel. In this procedure, a structure in one hemisphere is lesioned or inactivated, and the corresponding brain region in the functional circuit is similarly lesioned or inactivated in the alternate hemisphere. For example, if decision making on the rCET is dependent on the serial flow of information from the BLA to the ACC, then inactivating the BLA in one hemisphere will prevent any signals from being generated and sent to the ipsilateral ACC. In the alternate hemisphere, the BLA is able to generate and send information to the ipsilateral ACC, but the ACC is offline and thus unable to act on those incoming signals. Thus, while there is an intact BLA and ACC in each hemisphere to carry out the functions dependent on those structures, the functional interaction between them is compromised.  Also implicit to a disconnection design is the assumption that ipsilateral inactivation of a structure alone will not have an effect on behaviour. To address this possibility, a condition is included where both ipsilateral structures are inactivated, because the intact ipsilateral circuit should be able to compensate for the unilateral disruption of function. Finally, in order for a disconnection to be successful, contralateral connections between structures should be minimal. In regards to the pathway of interest, the ACC sends both ipsi-and contralateral     184 projections to the rostral portions of the BLA, whereas efferents from the BLA to the ACC are primarily ipsilateral. Thus, any disruption observed following this disconnection is likely mediated by the serial transfer of information from the BLA to the ACC, given the procedure will spare projections from one ACC to the contralateral BLA. The positioning of the BLA inactivation towards its caudal pole would also block signal flow from the BLA, while leaving ACC inputs at the rostral pole intact. However, while the available evidence suggests that any behavioural effects observed following the contralateral disconnection are likely due to impaired information flow from the BLA-ACC, we cannot preclude the involvement of top-down ACC –> BLA signaling with the current design.    In accordance with the disconnection procedure described above, each rat will receive three sets of infusions: 1) unilateral saline infusion into the ACC and into the contralateral BLA (control); 2) unilateral inactivation of the ACC and BLA ipsilaterally (ipsilateral inactivation control; and 3) unilateral inactivation of the ACC and the contralateral BLA (disconnection).  6.2.4 Microinfusion   Following recovery from surgical implantation of cannula, rats performed an additional 10 baseline rCET sessions, at which point their behaviour was deemed stable (see Data Analysis). Rats were habituated to the microinfusion process via two mock infusions. During these mocks, dust caps and obdurators were removed, 28-gauge injectors extending 1 mm below the cannula tip were inserted, but no infusion was performed. The injectors were left in place for two     185 minutes before being removed and replaced with obdurators. Ten minutes following the mock infusion, rats began an rCET session. One rat got sick following their second mock infusion, leaving n= 36 (n = 19 workers, n = 17 slackers) rats who moved on to the infusion stage. Inactivations adhered to a three-day cycle, consisting of a baseline session, a test session where rats received intracerebral infusion or drug, followed by a washout day in which rats were not run. Inactivations were achieved using a mixture of the GABAB agonist baclofen and the GABAA agonist muscimol (Sigma–Aldrich), prepared separately at 0.5 µg/µL in saline, and mixed together in equal volumes to form a 0.25 µg/µL solution. 0.5 µL per hemisphere injections of saline or baclofen/muscimol (i.e. 0.125 µg of drug per hemisphere) were administered at a rate of 0.4 µL/min, and injectors (extending +1mm from cannula tips) were left in place for an additional minute to allow diffusion. Once the microinfusions were completed, injectors were removed, obdurators replaced, and animals were returned to their home cages for 10 min before being placed in the operant chambers and performing the rCET. In keeping with the disconnection protocol described above, each rat received three sets of infusions: 1) unilateral saline infusion into the ACC and into the contralateral BLA (control); 2) unilateral inactivation of the ACC and BLA ipsilaterally (ipsilateral inactivation control; and 3) unilateral inactivation of the ACC and the contralateral BLA (disconnection). The order of these infusions was counterbalanced across rats. Immediately after these infusions, rats received an additional three sets of infusions, the order of which was counterbalanced across     186 rats. These included a 1) unilateral BLA inactivation, a 2) unilateral left ACC inactivation, and a 3) unilateral right ACC inactivation. The three-day cycle (baseline, infusion, washout) was similarly carried out with these infusions.    6.2.5 Histology  When behavioural testing terminated, animals were anesthetized with 2% isoflurane in O2 and then euthanized via C02. Brains were extracted and submerged in 4% paraformaldehyde for 24h, before being transferred to a 30% sucrose solution where they remained until sectioning. Brains were frozen and cut into 40-µm coronal sections encompassing the ACC and BLA were obtained using a cryostat. The sections were placed on gelatin-coated slides and stained with cresyl violet for visualization. The projected locations of the injectors tips protruding from the guide cannulae were mapped onto the coronal section images from Paxinos and Watson (1998)  6.2.6 Data Analysis  Data analysis was carried out as described in Chapter 2. However, ipsilateral and contralateral ACC-BLA inactivation caused some rats to complete less than ten trials. Specifically, two rats in the ipsilateral condition and 9 rats in the contralateral condition completed less than ten trials, and so were excluded from the data analysis. Conducting a one-way repeated measures ANOVA across both conditions would exclude many rats from the entire analysis, so the contralateral and ipsilateral conditions were individually compared to the saline     187 condition.    6.3 Results  6.3.1 Cannula Placements     The locations of all acceptable ACC and BLA placements are shown in Figure 6.1. Given we used bilateral mounted cannulae to target the ACC, six placements were too lateral in one hemisphere and too medial in another. Four BLA placements were located dorsal to the BLA. When considered together, this meant 8 rats were excluded from the ipsilateral and contralateral analyses, leaving n = 14 workers and n = 14 slackers. In subsequent experiments where we analyzed the effects of unilateral ACC or BLA inactivation, rats were only excluded if they had missed placements in the area of interest.    6.3.2 Baseline rCET Behavior   6.3.2.1  Choice, Accuracy, and Premature Responses   When infused with saline, rats as an entire group showed a preference for the high-effort/high-reward (HR) option compared to the low-effort/low-reward option (saline only- choice: F(1, 27) = 12.074, p = .002). In keeping with their baseline categorization, workers (M = 85.74%) chose the HR option significantly more than slackers (M = 44.96%) (saline only- group: F(1, 26)  = 83.730, p < .001).      188  Accuracy for LR trials was significantly higher than accuracy for HR trials, suggesting that the HR trials were indeed more demanding (saline only- choice: F(1, 26) = 131.055, p < .001). While a repeated measures ANOVA did not suggest a difference in HR trial accuracy between workers and slackers (saline only- choice x group: F(1, 26) = 0.223, NS; group: F(1, 26) = 2.822, NS), there was a 10% difference in accuracy between groups (workers: M = 63.62, slackers: M = 53.79) in the saline condition. A follow-up one-way ANOVA comparing HR accuracy between groups suggests that this difference is significant (saline only- group: F(1, 26)  = 7.852, p = .009). This finding is in contrast to all other investigations in the rCET, in which workers and slackers do not differ in their attentional ability. Thus, in the current study we cannot rule out the possibility that slackers’ preference for LR trials is due to their impaired ability to successfully complete HR trials.   Premature responding did not differ between LR or HR trial types (saline only- choice: F(1, 26) = 2.350, NS) and groups did not differ at baseline on rates of premature responding (saline only- choice x group: F(1, 26) = 1.362, NS; group: F(1, 26) = 0.533, NS). This suggests that slackers’ reduced choice of HR cannot be explained by an inability to wait for the .2s stimulus on HR trials.   6.3.2.2  Other Behavioural Measures   When infused with saline, rats initiated about 130 trials on average, and this did not differ between worker and slacker rats (saline only-group: F(1, 26)  = 4.011, NS). Following trial initiation and extension of both LR and HR levers, rats     189 typically made few choice omissions (~6 on average), with slackers omitting more at this phase compared to workers (saline only- group: F(1, 26) = 4.791, p = .038). Latencies to make a choice and to collect reward did not differ between trial types or groups (saline only- choice/ choice x group/ group: all Fs < 2.151, NS). Rats were generally quicker to make responses for HR trials, perhaps unsurprisingly given their shorter stimulus duration (saline only- choice: F(1, 26) = 7.717, p = .010). Response omissions following stimulus presentation were low, and this did not differ between groups (saline only- choice/ choice x group/ group: all Fs < 0.446, NS).   6.3.3 Contralateral and Ipsilateral Inactivation Analyses   6.3.3.1 Trials Completed  We noticed that upon inactivating the ACC and BLA either ipsilaterally or contralaterally, some rats completed less than 10 total trials, and this appeared to be disproportionally higher following contralateral inactivation. Specifically, two slackers (/14) failed to complete at least ten trials in the ipsilateral condition, whereas in the contralateral inactivation condition this included four workers (/14) and five slackers (/14). Conducting a one-way repeated measures ANOVA across both conditions would thus exclude many rats from the entire analysis, so the contralateral and ipsilateral conditions were individually compared to the saline condition.        190 6.3.3.2 Contralateral Inactivation Analyses: Choice, Accuracy, and Premature Responding  The following analyses are based on the 10 workers and 9 slackers who completed at least ten trials following contralateral ACC-BLA inactivation. When considered as a homogenous group, inactivation of the ACC and contralateral BLA did not affect choice of the HR option (inactivation: F(1, 17) = .497, NS). Rather, the effects of this inactivation condition depended on rats’ worker/slacker status (inactivation x group: F(1, 17) = 6.051, p = .025). A simple main effects analysis revealed that contralateral ACC-BLA inactivation specifically decreased HR choice in workers, but not slacker rats (workers only- inactivation: F(1, 9) = 4.840, p = .055) (Figure 6.2). While a visual inspection of the data suggests that slackers’ choice of HR increased following inactivation, this effect was not significant (slackers only- F(1, 8) = 1.632, p = .237). Contralateral ACC-BLA inactivation did not affect accuracy on LR or HR trials (inactivation/ inactivation x group: all Fs < 3.060, NS), nor did it affect measures of premature responding for either option (inactivation/ inactivation x group: all Fs < 1.779, NS) (Figure 6.2).   6.3.3.3 Contralateral Inactivation Analyses: Other Behavioural Measures  Contralateral inactivation of the ACC-BLA significantly increased latencies to make a LR and HR choice across both groups (LR choice- inactivation: F(1, 17) = 10.554, p = .005, inactivation x group: F(1, 17) = 0.009, NS; HR choice- inactivation: F(1, 17) = 4.658, p = .046, inactivation x group: F(1, 17) = 1.22, NS), and increased choice omissions when both levers extended at trial onset     191 (inactivation: F(1, 17) = 23.451, p < .001; inactivation x group: F(1,17) = 0.046, NS) (Table 6.1). Latencies to make a correct HR response likewise increased across groups, but LR correct response latencies increased to a greater extent in worker rats (HR correct latency- inactivation: F(1, 17) = 4.382, p = .052, inactivation x group: F(1,17) = 0.942, NS; LR correct latency- inactivation: F(1,17) = 18.405, p < .001, inactivation x group: F(1,17) = 4.369, p = .052; workers only- inactivation: F(1, 9) = 22.842, p < .001; slackers only- inactivation: F(1, 8) = 2.155, NS). In contrast, latencies to collect reward on either trial type were unaffected (inactivation/inactivation x group: all Fs < 1.905, NS). Following stimulus presentation, response omissions on LR trials were unaffected by ACC-BLA inactivation, but this manipulation did increase response omissions on HR trials (LR response omissions- inactivation/inactivation x group: all Fs < 3.361, NS; HR response omissions- inactivation: F(1, 17) = 22.409, p < 001; inactivation x group: F(1,17) = 0.391, NS). Trials initiated and trials completed significantly decreased following contralateral ACC-BLA inactivation (trials initiated- inactivation: F(1, 17) = 27.363, p < .001, inactivation x group: F(1, 17) = 0.000, NS; trials completed- inactivation: F(1, 17) = 44.412, p < .001, inactivation x group: F(1, 17) = 0.591, NS) (Table 6.1).   6.3.3.4 Ipsilateral Inactivation Analyses: Choice, Accuracy, and Premature Responding   The following analyses are based on the 14 workers and 12 slackers who completed at least ten trials following ipsilateral ACC-BLA inactivation. Similar to     192 the effects on HR choice in the contralateral condition, ipsilateral inactivation did not have any discernible effect on the entire group (inactivation: F(1, 24) = 0.130, NS). However, there was a trending inactivation x group interaction (inactivation x group: F(1, 24) = 3.544, p = .072), but subsequent simple main effects analyses were not significant, even though a visual inspection of the data suggests this condition had a similar effect on HR choice compared to the contralateral condition (workers only- inactivation: F(1, 13) = 2.37,  p = .147; slackers only- inactivation: F(1, 11) = 1.303, p = .278) (Figure 6.3). In contrast to contralateral ACC-BLA inactivation, ipsilateral ACC-BLA inactivation decreased accuracy on both LR and HR trials across groups (LR trials- inactivation: F(1, 19) = 7.759, p = .012, inactivation x group: F(1,19) = 0.016, NS; HR trials- inactivation: F(1, 22) = 6.649, p = .017, inactivation x group: F(1, 22) = 0.416, NS), while rates of premature responding were unaffected (inactivation/ inactivation x group: all Fs < 1.090, NS) (Figure 6.3).   6.3.3.5 Ipsilateral Inactivation Analyses: Other Behavioural Measures  Latencies to make a choice increased for HR, but not LR, trials following ipsilateral ACC-BLA inactivation (HR trials- inactivation: F(1, 23) = 7.845, p = .010, inactivation x group: F(1, 23) = 1.588, NS; LR trials- inactivation/ inactivation x group: all Fs < 1.343, NS) and so did the number of choice omissions following lever presentation (inactivation: F(1, 24) = 25.998, p < .001; inactivation x group: F(1, 24) = 1.469, NS) (Table 6.2). Latencies to make a correct response and to collect reward were not affected for LR trials     193 (inactivation/ inactivation x group: all Fs < 3.091, NS), but these latencies increased on HR trials following ipsilateral ACC-BLA inactivation (correct latency- inactivation: F(1, 21) = 5.415, p = .03, inactivation by group: F(1, 21) = 0.112, NS; collect latency- inactivation: F(1, 21) = 19.046, p < .001; inactivation x group: F(1, 21) = 1.228, NS).  Response omissions also increased for both trial types following stimulus presentation (LR trials- inactivation: F(1, 20) = 8.102, p  = .01, inactivation x group: F(1, 20) = 0.071, NS; HR trials- inactivation: F(1, 21) = 9.334, p = .006, inactivation x group: F(1, 21) = 0.721, NS). Lastly, both trials initiated and trials completed significantly decreased following ipsilateral ACC-BLA inactivation (trials initiated- inactivation: F(1, 24) = 33.065, p < .001; inactivation x group: F(1, 24) = 1.318, NS; trials completed- inactivation: F(1, 24) = 54.966, p < .001; inactivation x group: F(1, 24) = 0.464, NS) (Table 6.2).   6.3.4 Follow-up Unilateral Analyses  A core assumption of a disconnection design is that ipsilateral inactivation should have minimal effects on behavior, given communication between two areas is still intact. Additionally, unilateral inactivation of a brain region should not affect behaviour, because the intact structure in the opposite hemisphere is still on-line to mediate the processes normally carried out by the structures bilaterally. Our observation that ipsilateral inactivation caused a trending choice impairment that mirrored the effects of the contralateral condition prompted us to ask whether these effects are primarily due to a unilateral inactivation of the ACC or BLA. We subsequently infused the BLA, left ACC, and right ACC unilaterally, and     194 tested rats on the rCET. We recoded the left and right ACC data to correspond to whether it was the ACC targeted in the original ipsilateral or contralateral conditions. Two rats suffered health complications from prior infusions, and so could not contribute data to any of these subsequent conditions. Additionally one rat completed less than 10 total trials in each of the unilateral BLA, ipsilateral ACC and contralateral ACC conditions, while an additional rat in each of the ipsilateral and contralateral unilateral ACC conditions experienced cannula blockages and thus data could not be obtained. For this reason each dose was individually compared to the original saline condition, in which the BLA and contralateral ACC were infused with saline. The sample size for each analysis is as follows:  unilateral BLA inactivation (n = 14 workers, 15 slackers), unilateral ipsi-ACC inactivation (n = 15 workers, n = 12 slackers), and unilateral contra-ACC inactivation (n = 15 workers, n = 12 slackers).   6.3.4.1 Effects of Unilateral Inactivations on Choice   Like the choice profiles observed in the ipsilateral and contralateral inactivation conditions, unilateral BLA inactivation did not affect choice as a homogenous group, but differentially affected rats based on their worker-slacker status (inactivation: F(1, 27) = 2.07, NS; inactivation x group: F(1, 27) = 4.716, p = .039. Specifically, workers’ choice of the HR option decreased, while slackers’ choice of HR was unaffected (workers only- inactivation: F(1, 13) = 7.485, p = .017; slackers only- inactivation: F(1, 14) = .243, NS). This unilateral effect persisted when looking specifically at rats included in the contralateral ACC-BLA     195 inactivation condition (inactivation: F(1, 15) = 0.338, NS; inactivation x condition: F( 1, 15) = 5.039, p = .040; workers only- F(1, 8) = 4.567, p = .065, slackers only – F(1, 7) = 1.209, NS),  as well as rats included in the ipsilateral ACC-BLA inactivation condition (inactivation: F(1, 22 = 0.366, NS; inactivation x condition: F( 1, 22) = 8.197, p = .009; workers only- F(1, 12) = 6.630, p = .024, slackers only – F(1, 10) = 2.324, NS). In contrast, unilateral inactivation of the ACC did not affect choice, regardless of it whether it was the ACC targeted ipsilateral or contralateral to the BLA (inactivation/ inactivation x group: all Fs < .400 NS). Unilateral ACC inactivation similarly did not affect choice when we re-ran the analyses according to whether the left or right ACC was targeted (inactivation/ inactivation x group: all Fs < .751, NS). A comparison of the three choice profiles  - unilateral BLA, unilateral ipsi-ACC, and unilateral contra-ACC -is provided in Figure 6.4.   6.3.4.2 Effects of Unilateral BLA Inactivation on Other Measures    We focus on the general effects of unilateral BLA inactivation given this is the only unilateral inactivation that affected HR choice. BLA inactivation impaired accuracy on LR, but not HR trials (LR trials inactivation: F(1, 26) = 11.511, p = .002, inactivation x group: F(1, 26) = 0.64, NS; HR trials- inactivation, inactivation x group: all Fs < 1.00, NS), and did not affect measures of premature responding (inactivation/ inactivation x group: all Fs < 2.232, NS). Latencies to make a HR choice increased, but LR choice latencies were unaffected (HR trials – inactivation: F(1, 26) = 5.749, p = .024; inactivation x group: F(1, 26) = .803, NS;     196 LR trials- inactivation/ inactivation x group: all Fs < 1.906, NS), and otherwise latencies to make a correct response and collect reward were not changed following unilateral BLA inactivation (all Fs < 3.00, NS). Response omissions did not change, but choice omissions increased significantly (choice omissions: inactivation: F(1, 26) = 9.935, p = .004, inactivation x group: F(1, 26) = 2.207, NS; LR and HR trial response omissions: all F< 1,273, NS). Lastly, unilateral BLA inactivation decreased the number of trials initiated, but did not affect trials completed (trials initiated- inactivation: F(1, 27) = 4.413, p = .045, inactivation x group: F(1,27) = 0.267, NS; trials completed- inactivation/ inactivation x group: all Fs < 2.56, NS). A summary of these behavioural results is provided in Table 6.3.  6.4 Discussion   The purpose of this study was to investigate whether BLA-ACC signaling contributes to decision making where cognitive effort must be exerted in pursuit of lucrative rewards. This was based in part on previous work with the rCET, in which bilateral inactivations of the basolateral amygdala (BLA) and anterior cingulate cortex (ACC) affect this form of decision making independently (Hosking et al, 2014a), and also by related work showing that BLA-ACC signaling regulates effort allocation in the physical domain (Floresco and Ghods-Sharifi, 2007). Following unilateral inactivation of the ACC and contralateral BLA, intended to produce a functional disconnection between these structures, we observed a decline in choice of the high-effort/high-reward option selectively in     197 “worker” rats inherently displaying a preference for this option at baseline. However, when we carried out ipsilateral inactivations of the ACC and BLA, a condition intended to act as a control, we observed an identical shift in HR choice in worker rats. Subsequent unilateral inactivations of the ACC and BLA revealed this effect on choice could be entirely attributed to a unilateral inactivation of the BLA. Although this finding precludes a discussion of how BLA-ACC circuitry regulates decision making with cognitive effort costs, it reinforces the critical role of BLA signaling in valuations of cognitive effort. Indeed, as will be discussed below, this adds to a growing literature suggesting that the BLA may play a unique role in decision making with cognitive effort costs, over and above its role in decision making in the physical effort domain.   Before discussing the observed effects on choice as an aberration in decision-making processes, it is important to consider alternative explanations. For one, the inactivations we carried out in this study, and in particular the unilateral BLA inactivation, could have produced indifference between options, as if rats were unable to recall the contingencies associated with each lever. This seems unlikely, given workers’ choice of the HR option was still well above 50% following inactivation. It is also possible that the inactivations reduced motivation for sucrose reward. Pre-feeding before the task decreases trials completed, decreases HR choice, increases collect latencies, and impairs accuracy (Cocker et al, 2012b). However, we did not see a similar profile following any of our inactivations, and collect latencies – considered indirect measures of motivation (Robbins, 2002) - were generally unaffected. This also seems unlikely     198 considering other studies showing that inactivation or lesion of the amygdala and ACC do not affect preference for larger sucrose rewards in similar tasks (Ghods-Sharifi et al, 2009b; Rudebeck et al, 2006).   One notable limitation in the current study is that slackers’ accuracy on HR trials was significantly lower than their worker counterparts. Indeed, this is the first rCET study where accuracy on the HR lever fell below 60%. Thus, it is possible rats’ reduced choice is not just based on their aversion to cognitive effort costs as previously established (Cocker et al, 2012b), but also by their fundamental inability to successfully perform the HR trials. This is important to consider when discussing the role of the BLA in cognitive effort allocation, because we have previously shown that this region may regulate subjective preference to “work” or “slack” at baseline (Hosking et al, 2014a). Indeed, a visual inspection of the contralateral and ipsilateral inactivation data, as well as the unilateral BLA inactivation data, appears to replicate the effects of bilateral BLA inactivation in slackers, but this was not significant. Given the LR response option might now be the objectively better option for some slackers, this might have precluded BLA involvement in this form of decision making. In the ensuing discussion, it is important to consider whether a unilateral BLA inactivation would have had an effect on slacker rats had their accuracy been similar to previous cohorts.   Inherent to a disconnection design is the assumption that unilateral inactivation of a structure will have minimal to no effect on behaviour, given there is still an intact structure in the alternate hemisphere capable of carrying out its     199 normal functions. Along these lines, ipsilateral inactivation of two connected brain regions in one hemisphere should have minimal impact, given serial transfer information is unobstructed in the other hemisphere. Following unilateral inactivation of the ACC and contralateral BLA, intended to produce a functional disconnection of BLA-ACC signaling, choice of the HR option decreased in worker rats, while slackers’ choice of HR appeared to increase, though this was not significant. However, a similar pattern of effects was observed following ipsilateral inactivation of the BLA and ACC. While there is some evidence that interhemispheric inactivations of the BLA and striatal or cortical regions can have effects on behaviour (Fuchs et al, 2007; Lasseter et al, 2014; St. Onge et al, 2012), this is generally rare. Rather, the similar effect on HR choice across ipsi- and contralateral inactivation groups could also reflect a mass action effect (Lashley, 1931), where the magnitude of the behavioral impairment is related simply to the volume of tissue damage, regardless of its location. However, this account is unlikely, as a single unilateral BLA inactivation was sufficient to replicate the effects observed in the contralateral and ipsilateral inactivation conditions.  Notably, the magnitude of the unilateral BLA effect (~12.5% decline in choice HR) was almost identical to that observed following contra- or ipsilateral ACC-BLA inactivation (~13%). This was specific to unilateral BLA inactivation, as unilateral inactivations of the ACC were without effect. This is a rare example of unilateral BLA disruption of function, and to our knowledge is the first demonstration of a decision-making deficit induced by unilateral BLA     200 perturbation. Previous work has established a number of functions that are impaired when a single BLA is inactivated or lesioned, but most of this work has been in the domain of Pavlovian conditioning or simple instrumental responding. For example, unilateral pre-training BLA lesions partially attenuate fear conditioning to cues (LaBar and LeDoux, 1996), while post-training lesions abolish contextual conditioned fear in rats (Flavell and Lee, 2012). Deficits in PPI are also observed following unilateral BLA inactivation (Forcelli et al, 2012), and this manipulation attenuates context-induced renewal of Pavlovian-conditioned alcohol seeking (Chaudhri et al, 2013). Unilateral BLA perturbation also disrupts numerous aspects of reward processing, including reinforcer devaluation in monkeys (Izquierdo, 2004), as well instrumental learning and reward memory in rodents (Baldwin et al, 2000; Coleman-Mesches et al, 1996).   The effects here are in stark contrast to the effects of unilateral BLA inactivation in the related domain of physical effort allocation, in which BLA-ACC signaling biases rats towards larger reward obstructed by a scalable barrier (Floresco and Ghods-Sharifi, 2007).  In this study unilateral inactivation of the BLA with lidocaine caused a slight, nonsignificant decrease in choice of the HR option, but the magnitude of this effect was much smaller than the choice shift observed following functional BLA-ACC disconnection. This is interesting considering the differential roles the BLA appears to play in either form of decision making. While bilateral BLA inactivations uniformly decrease willingness to exert physical effort across rats, regardless of whether the effort is defined by barrier climbing or lever pressing (Floresco and Ghods-Sharifi, 2007; Ghods-    201 Sharifi et al, 2009a), the effects of BLA inactivation on cognitive effort allocation appear to depend on rats’ baseline preference to “work” or “slack”. Specifically, workers decrease their choice of the high-effort/high (HR) reward option, while slackers increase HR choice following inactivation (Hosking et al, 2014a). A visual inspection of the current data suggests unilateral inactivation of the BLA reproduced the effects of bilateral BLA inactivation on rCET choice, but the choice shift was not significant in the slacker group (possible reasons for this were described above). It is possible the unilateral BLA inactivation was effective in producing a similar choice profile via directly or indirectly downregulating the contralateral, intact BLA. While diffuse interamygdaloid connectivity has been observed in rats via the anterior commissure, a previous study showed that unilateral lesion of the BLA does not affect measures of fos activity in the contralateral BLA (Flavell and Lee, 2012). Alternatively, the effects might reflect a functional breakdown in the intra- and interhemispheric projections originating from the BLA.  Regardless of the mechanism, BLA signaling appears to play a more prominent role in valuations of effort in the cognitive domain, as both nuclei are required for optimal effort allocation to occur.   This regulation of cognitive effort is also interesting considering recent imaging data showing that while a network of structures appear to process the subjective value of effortful options (including the dACC, dmPFC, dlPFC, intraparietal sulcus, and anterior insula), the amygdala appears to uniquely encode the subjective value of rewards discounted by cognitive, but not physical effort (Chong et al, 2017). The same study also found distinct computational     202 functions encode the subjective value of rewards discounted by physical or cognitive effort costs (Chong et al, 2017). Together, the current and previous work on the rCET support the notion that the BLA might specifically encode the subjective value of lucrative outcomes discounted by mental effort. In contrast, rather than supporting subjective value computations in the domain of physical effort, BLA involvement in rodent physical effort tasks may preferentially encode expected reward, thus biasing rats towards the high effort, high reward response option. This appears to be true given the effects of BLA inactivations on rodent decision-making tasks appear to depend on the nature and severity of the costs (Winstanley and Floresco, 2016). If costs are very aversive, such as signaled punishing timeouts or electric shock, the BLA appears to attenuate baseline choice of the more aversive option (Orsini et al, 2015a; Zeeb and Winstanley, 2011). In contrast, when rewards are especially salient relative to costs, such as omission of reward receipt, delays, or exerting physical effort, then the BLA appears to drive choice towards more rewarding options (Ghods-Sharifi et al, 2009a; Winstanley, 2004). Comparatively, the physical effort costs employed in rodent tasks may not be considered severe, perhaps because rats are always guaranteed reward following effort expenditure. In contrast, this may explain the divergent effects of BLA inactivation on choice in the rCET, whereby the rewards associated with successful trial completion are more salient in workers, but cognitive costs are especially aversive in slacker rats. This hypothetical role for the BLA in cost/benefit decision making may be mediated by distinct neuron populations shown to encode positive and negative value, as well as by distinct     203 circuits processing positive (BLA-NAc) or negative (BLA-CEA) associations (Belova et al, 2007, 2008; Namburi et al, 2015; Paton et al, 2006).     While we are the first group to show a unilateral BLA contribution to value-based decision making, we are not the first to show a lateralized cognitive deficit resulting from unilateral silencing of a structure in the domain of effort allocation. Croxson and colleagues demonstrated that a unilateral mPFC lesion, encompassing the prelimbic (PL), infralimbic (IL), and anterior cingulate cortices (ACC), was sufficient to decrease HR arm choice in the effort T-maze paradigm (Croxson et al, 2014). This effect is likely driven by an interaction of the PL/IL and ACC lesions, as selective lesions of the PL/IL do not affect decision making in this paradigm (Walton et al, 2003a), but other investigators have unilaterally lesioned or inactivated the ACC without effect (Floresco and Ghods-Sharifi, 2007; Hauber and Sommer, 2009). Notably, unilateral mPFC lesions were particularly effective at decreasing HR choice when the HR arm was contralateral to the lesion, suggesting that the mPFC may play an important role in evaluating the costs and benefits of actions directed to specific spatial locations (Croxson et al, 2014). We carried out a similar analysis, coding whether the BLA inactivation was ipsi- or contralateral to the HR option, but found no evidence that the inactivation interacted with the relative location of the HR option (data not shown).  While we cannot speculate on the role of ACC-BLA signaling in cognitive effort allocation given the nature of the unilateral BLA effect, it is worth noting that many rats in the contralateral inactivation condition did not engage with the task.     204 Specifically, nine rats completed less than 10 total trials following contralateral ACC-BLA inactivation, compared to two rats in the ipsilateral condition. It is difficult to say what is driving this profound inability to engage, but it is unlikely a motor or motivational impairment, considering effects on general latencies and omissions rates were more pronounced in the ipsilateral inactivation condition (see tables 6.1 and 6.2). It is also not due to an impaired ability to perform the attentional challenge, as accuracy and rates of premature responding were unaffected. Rather, this might reflect a fundamental deficit in willingness to apply effort generally, as has been previously described following functional disconnection of the ACC and BLA (Floresco and Ghods-Sharifi, 2007).  It is worth noting that the design of the rCET necessarily makes it both a cognitively and physically demanding task; in a typical session, rats initiate approximately 130 trials in which they hit a lever, orient towards the stimulus array to detect the stimulus, then turn around to either collect reward or commence the next trial. If the BLA and ACC work in tandem to energize organisms towards lucrative outcomes discounted by an effortful challenge, then downregulating this circuit might preclude task rCET engagement.  In conclusion, we were unable to implicate ACC-BLA circuitry in the processes guiding decision making with cognitive effort costs.  However, we discovered the BLA plays a powerful, albeit nuanced, role in guiding cognitive effort allocation.    205  6.5 Tables  Table 6.1. Other rCET behavioural measures following contralateral ACC-BLA inactivation Means are presented (+ SEM). Data in brackets shows means and SEM for all rats, including those excluded from analyses because they completed less than 10 trials. Values presented for choice omissions, trials initiated, and trials completed reflect data for saline and inactivation session.  # p < .1; * p < .05; ** p < .001      LR - saline LR - inactivation HR - saline HR - inactivation Choice omissions (For Session) 5.00 + 1.32 (6.50 + 1.40) 17.00 + 2.27* (13.46 + 1.87)  Correct latency (s) 0.58 + 0.03 (0.57 + 0.03)  0.75 + 0.04 (0.73 + 0.04) 0.47 + 0.02 (0.47 + 0.02) 0.75 + 0.13* (0.87 + 0.17) Collect latency (s) 3.21 + 1.23 (3.12 + 1.17) 2.20 + 0.19 (2.19 + 0.10)   1.83 + 0.29 (1.79 + 0.27) 2.60 + 0.50 (2.65 + 0.48) Response Omissions (%) 7.01 + 2.65 (6.67 + 2.54) 11.67 + 2.64 (11.09 + 2.57) 3.68 + 1.06 (3.82 + 1.04) 18.90 + 3.47* (24.25 + 5.13) Trials Initiated (For Session) 129.42 + 5.57 (127.71 + 4.44) 91.47 + 7.63* (65.21 + 8.99)  Trials Completed  (For Session) 102.18 + 7.32 (98.41 + 5.62) 58.21 + 6.85* (43.73 + 6.89)     206   Table 6.2. Other rCET behavioural measures following ipsilateral ACC-BLA inactivation Means are presented (+ SEM). Data in brackets shows means and SEM for all rats, including those excluded from analyses because they completed less than 10 trials. Values presented for choice omissions, trials initiated, and trials completed reflect data for saline and inactivation session. # p < .1; * p < .05; ** p < .001  LR - saline LR - inactivation HR - saline HR - inactivation Choice omissions (For Session) 5.65 + 1.31 14.58 + 1.88* Correct latency (s) 0.57 + 0.02 (0.55 + 0.02) 0.84 + 0.15 (0.84 + 0.14) 0.47 + 0.02 (0.46 + 0.02) 0.75 + 0.11* (0.75 + 0.11)  Collect latency (s) 3.11 + 1.12 (3.09 + 0.98) 2.57 + 0.35 (2.51 + 0.33) 1.51 + 0.07 (1.49 + 0.07) 1.83 + 0.06* (1.84 + .08)  Response Omissions (%) 7.20 + 2.42 (6.60 + 2.25) 20.72 + 3.87* (22.57 + 4.57) 5.41 + 1.33 (5.41 + 1.33) 29.34 + 4.09* (29.34 + 4.09)  Trials Initiated (For Session) 129.42 + 4.55 (127.71 + 4.44) 83.12 + 6.64* (78.29 + 7.02)  Trials Completed (For Session) 101.52 + 5.55 (98.41 + 5.62) 56.42 + 6.23* (52.75 + 6.31)     207   Table 6.3. Other rCET behavioural measures following unilateral BLA inactivation Means are presented (+ SEM). Data in brackets shows means and SEM for all rats, including the one rat that completed less than ten trials following unilateral BLA inactivation. This rat did not contribute individual lever data because they initiated less than five trials for each lever type. Values presented for choice omissions, trials initiated, and trials completed reflect data for saline and inactivation session. # p < .1; * p < .05; ** p < .001    LR - saline LR - inactivation HR - saline HR - inactivation Choice latency (s) 3.44 + 0.26  3.70 + 0.26  3.13 + 0.17 3.55 + 0.17 Choice omissions (For Session) 6.48 + 1.36 (7.03 + 1.42) 12.82 + 1.61* (12.93 + 1.56) Premature responses (%) 16.93 + 3.48  11.18 + 2.14  21.41 + 3.29 15.86 + 1.97 Accuracy (%) 88.49 + 1.92  80.48 + 2.34  59.79 + 2.11  53.41 + 2.82  Correct latency (s) 0.55 + 0.02 0.59 + 0.03  0.48 + 0.02 0.53 + 0.02 Collect latency (s) 2.87+ 0.85  2.57 + 0.65  1.66 + 0.21 1.66 + 0.11 Response Omissions (%) 7.00 + 2.03  5.60 + 1.39  5.43 + 1.23 7.04 + 1.99 Trials Initiated (For Session) 127.74 + 4.24 (126.93 + 4.17)  113.97 + 5.37* (110.87 + 6.04) Trials Completed (For Session) 98.95 + 5.43 (98.45 + 5.27) 87.69 + 5.27 (84.90 + 5.81)     208 6.6 Figures    Figure 6-1. Histological analysis of cannulae implantation Location of all acceptable a) anterior cingulate cortex (ACC) and b) basolateral amygdala infusions.  On the ACC plates, white circles reflect ipsilateral infusions, while black circles are contralateral infusions. Black triangles are placements that were not acceptable for the ipsilateral/contralateral analyses, but were included in the unilateral analyses. Coordinates are relative to bregma. Plates modified from Paxinos and Watson (1998)    209   Figure 6-2. Effects of contralateral ACC-BLA inactivation on (a) choice of HR trials, (b) accuracy (%), (c) premature responses, and (d) choice latencies Data shown are the mean for each variable (±SEM). LR: Low effort/low reward trial; HR: High effort/high reward trial. Contralateral ACC-BLA inactivation decreased HR choice in workers and increased choice latencies across groups.      210   Figure 6-3. Effects of ipsilateral ACC-BLA inactivation on (a) choice of HR trials (%), (b) accuracy (%), (c) premature responses (%), and (d) choice latencies Data shown are the mean for each variable (±SEM). LR: Low effort/low reward trial; HR: High effort/high reward trial. Ipsilateral ACC-BLA inactivation appeared to decrease HR choice in workers and increase HR choice in slackers, but this was not significant when the inactivation x group interaction was followed up with a simple main effects analysis. This condition also impaired accuracy for both trial types and increased HR choice latencies across groups.    211              Figure 6-4. Effects of (a) unilateral BLA, (b) unilateral IPSI-ACC, and (c) unilateral contra-ACC inactivation on choice of HR trials (%) Data shown are the mean for each variable (±SEM). LR: Low effort/low reward trial; HR: High effort/high reward trial.     212 Chapter 7: General Discussion    The purpose of the experiments described in the preceding chapters was to improve our understanding of the neural processes guiding the allocation of cognitive effort. This was primarily focused on specific brains regions, circuits, and neurotransmitter systems that have been studied in the related field of physical effort allocation, but which had not been studied in the cognitive domain. This limited a full comparison of the processes underlying both forms of effort allocation, and has often led to generalizations regarding the substrates guiding cognitive effort.  To this end, I identified brain regions (i.e., OFC, NAc) that appear to play similar roles across effort types, others that play a more nuanced role in valuations of cognitive effort (i.e., BLA), and some which exclusively mediate decision making with cognitive effort costs (i.e., DMS).  Together, these studies provide some of the first causal evidence of the neural processes regulating decision making with cognitive effort costs. My hope is this work will inform the corresponding human literature, and that it might motivate behavioural neuroscientists to develop preclinical models investigating effort allocation in other executive domains.  In experiment 1, I used chemogenetic techniques to temporarily donwregulate basal forebrain cholinergic neurons as rats performed the rCET. This manipulation did not affect performance on the rCET, even though multiple doses and timepoints were assessed. However, I was able to replicate a     213 previously observed effect of scopolamine of rCET choice, in which muscarinic receptor antagonism decreased high-effort/high-reward choice while leaving attentional processes intact (Hosking et al, 2014c) . Thus, cholinergic signaling appears to play an important role in rCET decision making, but the current experiments suggest that other cholinergic neuron populations are mediating these effects. I suggest that striatal cholinergic interneurons are a worthy target for future work on the rCET, given the role of the striatum in cognitive effort allocation, as identified in experiment 2.   In experiment 2, I showed that the striatum plays an integral role in performance of the rCET. Temporary inactivation of the dorsomedial striatum decreased choice of the HR option, whilst impairing accuracy and increasing impulsivity. In contrast, nucleus accumbens core inactivation precluded task engagement, which I suggest might reflect a profound unwillingness to engage with the rCET. Indeed, while this work complements human imaging work associating striatal areas with cognitive effort discounting (Botvinick et al, 2009; Schmidt et al, 2012), this is the first experiment to causally implicate the striatal regions in cognitive effort allocation.  The purpose of experiment 3 was to assess whether orbital frontal cortices contribute to the processes guiding rCET decision making. Temporary bilateral inactivation of this region did not affect decision making or attention on the rCET, but led to a pronounced increase in motor impulsivity and latency to make a decision. These results have important implications. For one, they establish specificity in the prefrontal regions regulating cognitive effort allocation; while     214 orbital areas do not appear play a role, the medial wall of the prefrontal cortex has been previously implicated in decision making on the rCET (Hosking et al, 2014a, 2015a). Additionally, this work shows that deficits in ability to perform the task – in this case a lack response control- are not always accompanied by shifts in decision making. This is in stark contrast to the effects observed following inactivation other prefrontal regions, which affect decision making while simultaneously impairing attention (i.e prelimbic cortex) or response control (i.e infralimbic, anterior cingulate cortex). Indeed, the results of experiment 3 confirm that, at the level of the prefrontal cortex, the processes guiding cognitive willingness and cognitive ability are dissociable.  Lastly, in experiment 4 I attempted to examine how cortical and limbic regions interact to guide decision making with cognitive effort costs, by functionally disconnecting the BLA and ACC while rats performed the rCET. However, proper investigation of BLA-ACC signaling was precluded by the observation that unilateral inactivation of the BLA was capable of affecting decision making in isolation. This is a surprising result, given few studies have reported disruption of behavioural function following unilateral lesion or inactivation of a brain structure. However, this is not the first study to report effects on decision making following unilateral inactivation (Croxson et al, 2014), and the results are in keeping with recent imaging work showing that the right BLA appears to encode the subjective value of options characterized by cognitive, but not physical, effort costs (Chong et al, 2017). Indeed, the results of the experiment 4 suggest that encoding in the BLA bilaterally strongly regulates     215 innate willingness to exert cognitive effort for lucrative outcomes.   Comparing Effort Allocation Across Cognitive and Physical Domains   The goal of the experiments in this thesis was to further characterize the neurobiology regulating decision making with cognitive effort costs, so that this type of effort allocation could be compared and contrasted with existing work in the physical effort literature. Indeed, the current experiments add support to the notion that cognitive and physical effort allocation are subserved by overlapping, albeit distinct, neural mechanisms. At the level of the prefrontal cortex, both forms of effort allocation appear to depend on neural activity in dorsal regions encompassing the anterior cingulate cortex (Hosking et al, 2014a; Rudebeck et al, 2006), while orbital regions appear to play a nominal role in either form of decision making (Rudebeck et al, 2006, Chapter 4). This finding lends further support for “dorsal-ventral” streams in decision making, by showing that costs intrinsic to the individual (i.e. physical or cognitive effort) are preferentially mediated by dorsal prefrontal cortical regions (Klein-Flugge et al, 2016; Rangel and Hare, 2010). In contrast, the prelimbic and infralimbic regions of the rodent prefrontal cortex appear to play a selective role in cognitive effort allocation (Hosking et al, 2015a), which is likely owing to the nature of the cognitive challenge, given these areas regulate the processes needed to successfully complete rCET trials (Chudasama et al, 2003b; Passetti, 2002). To the extent that the PL/IL regions are functionally analogous to the primate dorsolateral prefrontal cortex (reviewed in section 1.10.1), these findings fit with current     216 theories of cognitive effort allocation, which suggest that signalling between the ACC and dorsolateral prefrontal areas of the human PFC direct and intensify cognitive resources to maximize lucrative outcomes (Shenhav et al, 2013).  The striatum is another area this thesis has implicated in cognitive effort allocation, and is one area that appears to differentially regulate effort allocation depending on the nature of the effortful challenge. Decision making with physical effort costs appears to be exclusively mediated by the ventral striatum (and specifically the nucleus accumbens core region) (Ghods-Sharifi and Floresco, 2010). In contrast, I identified a possible role for the nucleus accumbens in deciding between options that differ in cognitive load, but also a prominent role for the dorsal striatum in mediating this process (Chapter 4). The preferential involvement of this region in cognitive effort allocation is likely related to the number of cortical inputs to this region from areas previously implicated in rCET decision making (i.e. prelimbic, infralimbic, and anterior cingulate cortices) (Hosking et al, 2014a, 2015a; Mailly et al, 2013).   Previously our lab has shown that the BLA plays a nuanced role in decision making with cognitive effort costs, such that inactivation of the region interacts with subjects’ baseline preference to work or slack at baseline (Hosking et al, 2014a). This is in contrast to effort allocation in the physical domain, where inactivation of BLA decreases effortful choice across subjects (Ghods-Sharifi et al, 2009b). BLA-ACC signaling regulates rats’ willingness to scale a larger barrier for larger reward, and I wanted to assess whether this circuit plays a similar role in biasing rats towards hard cognitive challenges. However, a proper assessment     217 of this pathway was precluded by the fact that a unilateral inactivation of the BLA was capable of affecting choice in a subset of rats preferring the high-effort/high-reward option at baseline. This partially replicates the effects of bilateral BLA inactivation, where choice of HR decreases in worker rats but increases in slackers, as if rats were moving away from their generally preferred option. The findings of experiment 4(Chapter 6) are in contrast to work with physical effort costs, where unilateral inactivation of the BLA Is without effect. Given the major difference between the two types of decision making is the type of cost, it appears then that intact bilateral BLA signaling is necessary for valuations of cognitive effort, which may reflect a specialized role for the BLA in subjective valuations of cognitive effort. Indeed, recent human imagining works suggests that activity in the right BLA encodes the subjective value of a choice with cognitive, but not physical, effort costs (Chong et al, 2017).  In Chapter 1 I carried out an exhaustive review of the neurotransmitter systems involved in either form of effort-based decision making. Some neurotransmitters, like the cannabinoid system, play similar roles regardless of whether the challenge is physical or cognitive (section 1.8.4), while others, such as noradrenaline and serotonin, do not appear to contribute to the cost/benefit computations regulating either form of effort allocation (Section 1.8.2, 1.8.6, and appendix 1).  In contrast, some neurotransmitters appear to be preferentially involved in physical effort allocation, such as dopamine (section 1.8.1) and adenosine (1.8.5), whereas the cholinergic system appears to play a more prominent role in valuations of cognitive effort (sections 1.8.3). We attempted to     218 further characterize this role of acetylcholine in this form of decision making, by selectively downregulating cholinergic neurons of the basal forebrain using chemogenetics. This manipulation did not affect choice on the rCET, which I take as evidence that this neuron population is not critical for mediating decision making with cognitive effort costs (although in Chapter 3 I discuss potential limitations of using chemogenetics to target basal forebrain cholinergic neurons). However, I replicated the effect of scopolamine on rCET decision making, suggesting that these rats were sensitive to challenges of the cholinergic system. While I was unable to identify which cholinergic neuron population might regulate cognitive effort allocation, experiment 1 does help narrow down the targets for future studies (such as striatal cholinergic interneurons, as discussed in Chapter 3).   Implications, Outstanding Questions, and Future Directions Aside from permitting a thorough comparison of the processes guiding cognitive and physical effort allocation, the current thesis experiments add to a relatively new literature studying valuations of cognitive effort (Westbrook and Braver, 2015). However, interest in decision making with cognitive effort costs is growing, as evidenced by the number of human behavioural paradigms that have recently been developed (section 1.1.2). Indeed, the number of cognitive allocation tasks now trumps the number of available physical effort discounting paradigms (reviewed in Section 1.1.1). However, only one animal model of decision making with cognitive effort costs currently exists, thus limiting causal     219 investigations of the neural mechanisms guiding this decision process. In the sections that follow I propose an expansion of the tasks developed to probe valuations of cognitive effort in rodents, and suggest avenues for future research that I believe will make important contributions to this area of decision making literature.   Are the Processes Guiding Effort allocation Domain-Specific or Generalized?   A core feature of the rCET is that it manipulates cognitive effort via titrating the attention required to detect a visuospatial stimulus. The task is a variant of the 5-choice serial reaction time task, and so there is a rich literature we can directly draw from when assessing how different manipulations affect decision making, attention, and response control (Chudasama and Robbins, 2006; Robbins, 2002). Another clear advantage of the rCET design is that there are no delay confounds between trial types, as latencies to reward delivery are similar. However, the rCET models only one type of effort, and it is interesting to consider whether different neurotransmitters and circuits would be recruited if another executive function were being taxed. Indeed, the human literature has models probing visuospatial attention, in addition to tasks manipulating working memory, cognitive flexibility, and response inhibition (Dixon and Christoff, 2012; Kool et al, 2010b; Massar et al, 2015; Westbrook et al, 2013). Generally speaking, the imaging literature suggests that valuations of cognitive effort are subserved by a similar network, including the ACC and dorsolateral prefrontal cortex (Chong et     220 al, 2017; Massar et al, 2015). However, some regions appear to be domain specific, such as the role of the BLA in encoding the subjective value of a difficult attentional challenge (Chong et al, 2017). This might be related to the role of the BLA in motivationally guided attention, via its connections to the basal forebrain as previously described (Peck and Salzman, 2014). Indeed, the BLA appears to plays a critical role in the evaluative processes guiding choice on the rCET, but this might be driven by the attentional nature of the cognitive challenge (Hosking et al, 2014a). Would the BLA still be implicated in cognitive effort allocation if a working memory challenge were used instead? This is an open question, but a review of the animal decision-making literature suggests it would play some role (Winstanley and Floresco, 2016).  To this end, I think it is imperative that other models be developed which assess cognitive effort allocation in other domains. For example, an operant task could be developed in which rats choose between easy or hard options that differ in the working memory load they require. In this task rats would nosepoke one of the five lit stimulus holes, orient towards the magazine and make a sustained nosepoke response, wait a specified delay period, and then try to make a nosepoke in the original lit aperture. On low-effort/ low reward trials, the lit hole could remain constant across trials, and so working memory requirements on a given trial would be minimal, while on high-effort/ high-reward trials rats would need to hold online the location of the correct stimulus hole as it varied between trials. This design is loosely based on the combined attention-memory (CAM) task developed by Chudasama and colleagues (Chudasama et al, 2004). This     221 procedure would also allow a researcher to increase the delay period across blocks of trials, thus producing a discounting curve showing how reward is devalued by increasing working memory load. Regardless of the specific nature of the task, other animal models would help us determine whether decision making with cognitive costs is mediated by a common system, or by distinct processes. This is an exciting avenue for future work.   The Role of Other Brain Regions and Site-Specific Neurotransmitter Regulation in Cognitive Effort Allocation   As described in Chapter 1, a number of prefrontal cortical, striatal, and limbic regions have been implicated in the processes guiding effort allocation. Indeed, in the realm of cost/benefit decision making, the literature has focused almost exclusively on the OFC, ACC, and medial prefrontal areas, the nucleus accumbens, and the basolateral amygdala (Winstanley and Floresco, 2016). In this thesis I identified a role for the dorsomedial striatum in cognitive effort allocation, which is notable considering this striatal zone has not been studied in many other domains. However, it is unlikely these regions are the sole regulators of decision making, and there are a number of regions I think would be useful to investigate on the rCET. I am primarily interested in the role of the insula in valuations of cognitive effort, as imagining evidence suggests that this area is recruited during valuations of physical effort (Arulpragasam et al, 2018; Prevost et al, 2010), and activity in this region correlates with the subjective value of choices that are either physically or cognitively effortful (Chong et al, 2017).     222 Another notable area to assess would be the lateral habenula, as work from Floresco and colleagues has shown that the region strongly regulates subjective preference in decision making with probability and delay costs (Stopper and Floresco, 2014)    Now that we have identified the general regions and neurotransmitter systems guiding decision making with cognitive effort costs, future work should specifically investigate where certain neurotransmitters are mediating their effects on decision making. For example, cholinergic and cannabinoid signaling regulate rCET decision making, but it is currently unknown where this regulation is taking place. We attempted to address this by studying the contribution of basal forebrain cholinergic neurons to the rCET, but we suggest that the chemogenetic technique might have been inadequate to fully test this hypothesis. Instead, lesions of BF cholinergic or striatal interneurons could be carried out using the immunotoxin 192- IgG-saproin as previously described (Chudasama et al, 2004; Mcgaughy et al, 2002), and likewise intra-PFC or intra-striatal infusions of scopolamine or similar cholinergic antagonists could also be used. In relation to cannabinoid signaling, CB1 receptor activation in the ACC and NAcc biases rats towards low effort, low reward options, and work with the rCET suggests a similar role for the CB1 receptor in the mPFC (Fatahi and Haghparast, 2018; Khani et al, 2015; Silveira et al, 2016).   Trial-by-Trial Analyses  The baseline propensity to “work” or “slack” at baseline interacts with a     223 number of pharmacological challenges and appears to be uniquely mediated by some brain regions, suggesting that different neural substrates may regulate baseline willingness to exert cognitive effort. Our understanding of exactly why workers and slackers differ in their baseline choice is limited, but it appears that slackers view cognitive costs as especially aversive, rather than their decreased HR choice being reflective of an insensitivity to reward (Cocker et al, 2012b). It would be interesting to conduct a trial-by-trial analysis of rCET performance, to see if workers and slackers fundamentally differ in the pattern by which they allocate their cognitive resources in a session. For example, there is evidence that individuals with Parkinson’s Disease are less willing to exert effort because they need to recuperate after a difficult challenge (Müller and Apps, 2018, personal communication). It would be interesting to assess whether something similar is happening in slackers, whereby following an HR trial they select the LR lever for a few trials before resuming selection of HR. This and other decision-making patterns can be evaluated, thus identifying key ways that workers and slackers differ in valuations of a cognitive challenge.   Physiological Factors Affecting Cognitive Effort Allocation and Modeling Disease States  The utility of the rCET is that it is designed to model the decision making humans face on a day-to-day basis. Indeed, compared to the costs modeled in other decision-making paradigms, humans consistently find themselves in settings where cognitive effort expenditure is necessary to achieve desired or     224 lucrative outcomes. Notably, physiological states such stress, fatigue, and sickness are associated with impaired academic and workplace performance (Kronholm et al, 2015; LeBlanc, 2009; Odeen et al, 2013), which are prime situations where individuals must decide how to allocate cognitive resources.  Indeed, these impairments can result in significant personal and economic consequences, and subdomains of economics and occupational psychology study how these adverse consequences can be ameliorated by environmental or behavioural modifications. It is possible relationships between adverse physiological states and impaired functional outcomes are mediated by a reduced willingness to exert cognitive effort, and the rCET provides an ideal model to test these hypotheses directly. For example the effects of acute, chronic, or unpredictable stress could be manipulated as rats perform the rCET (Bryce and Floresco, 2016; Shafiei et al, 2012), or fatigue could be induced by having rats treadmill run for an extended period of time or sleep depriving them prior to task performance (Iodice et al, 2017). Similarly, sickness could be induced by the administration of inflammatory cytokines or lipopolysaccharide as has been recently investigated in the domain of physical effort allocation (Nunes et al, 2014). If stress, fatigue, and sickness affect willingness to engage in a cognitive challenge, then the rCET provides a useful model to identify pharmacological or behavioural therapies that can reverse these impairments.      During the writing of this thesis, I became familiar with the work of Salamone and colleagues, and specifically their attempt to model the motivational symptoms inherent to depression (Nunes et al, 2013b). This work     225 was borne out of the fact that typical antidepressant SSRIs are ineffective in treating the motivational deficits of apathy and anergia in depression (Rothschild et al, 2014). Administration of the VMAT inhibitor tetrabenazine decreases willingness to exert physical effort for larger rewards across physical effort allocation paradigms, and these effects are reversed by a number dopaminergic and adrenergic drugs, but not by typical SSRIs (Yohn et al, 2016b). It would be interesting to test whether the effects of DA depletion via tetrabenazine similarly affect decision making in the cognitive domain. Although previous work on the rCET has shown that DA plays no role in guiding decision making with cognitive effort costs, selective D1 and D2 antagonists were administered that do not mirror the mechanism of action of tetrabenazine (Hosking et al, 2014b). Additionally, motivational deficits are not just limited to depression, but are also observed in Parkinson’s Disease and schizophrenia (Chong et al, 2015; Reddy et al, 2015), and at least one study has shown that individuals with schizophrenia display deficits in cognitive effort allocation (Culbreth et al, 2016b; Gold et al, 2014). To this end, it would be interesting to test animal models of Parkinson’s Disease (i.e. 6-OHDA lesions) and schizophrenia (I.e. isolation rearing, neonatal ventral hippocampal lesion, gestational methylazoxymethanol (Jones et al, 2011)) to assess whether they emulate the decision-making deficits inherent to the disorders they model.   Limitations and Critical Considerations    The rCET is has proven itself a valuable tool in probing the brain     226 mechanisms guiding effort allocation in the cognitive domain. As with any study, however, the findings must be considered in light of certain limitations.    Firstly, two of our experiments (Chapter 5 and Appendix 1) were notable in that they used female subjects. Indeed, the animal cost/benefit decision-making literature has almost exclusively focused on male subjects, with few exceptions (Orsini et al, 2015b; Uban et al, 2012). Female performance on the rCET is similar to males, but we did not assess how this behaviour may fluctuate with estrous cycle. Previous work in the physical effort allocation literature suggests that this phenomenon does not map onto changes in estrous cycling (Uban et al, 2012), but this is something that might be worth investigating in future rCET studies.   The rCET requires extensive training, and this might preclude involvement of some brain regions and neurotransmitters that might otherwise be involved in effort allocation in novel situations. In the human literature, participants experience a task once or twice, but in the rCET rats are trained for months on a task where contingencies do not change. Areas such as the OFC appear to be preferentially recruited early on during task acquisition, and neurotransmitters like dopamine play an important role in learning stimulus-outcome associations- processes that are probably not occurring following extended training (Schulz et al, 2013; Zeeb and Winstanley, 2011). Indeed, this may be one reason why DA does not appear to play a role in decision making on the rCET, but the physical effort literature and computational models of cognitive effort implicate DA in effort allocation (Salamone and Correa, 2012; Westbrook and Braver, 2016). One     227 possible way to get around this might be to administer rCET “challenge” sessions in which parameters such as stimulus duration or inter-trial-interval change. These challenges were used to great effect in the original 5-CSRTT literature, and were used to probe the extent of effects of different lesions and pharmacological manipulations. The extensive training on the rCET also precludes any investigation of decision making at earlier developmental time points, such as adolescence. This latter point is more a comment on the animal decision-making literature in general, as the training required extends beyond the 15-20 day window considered adolescence in rodents.   Another limitation requires a theoretical consideration of why a given cognitive challenge is difficult in the first place. A cognitive task may be aversive because it is inherently costly, but might also be aversive because it is less like to be successfully completed (Westbrook and Braver, 2016). In the rCET, difficulty between challenges is evidenced by the lower accuracy on hard trials, suggesting that one of the reasons the HR option is aversive is because it is less likely to result in reward. Previous work has shown that the effort discounting observed on the rCET is not simply related to the probability of receiving reward for LR and HR trials types (Cocker et al, 2012b). Rather, the imposition of cognitive load biases rats’ away from the HR option, above and beyond its probability of reinforcement. In any case, many human behavioural paradigms are designed so that performance on easy and hard trials is equivalent or similar, so that any effort discounting is directly due to the inherent cost of cognitive control. It might be interesting for future studies to play with the parameters of the     228 rCET, to see if there is a set of stimulus durations that result in similar accuracy rates, but which are valued differently on the basis of preference experiments where reward magnitudes are matched.   A final limitation that is not specific to the current thesis, but is pervasive across the animal cost/benefit decision-making literature is that choices are made between concurrently available options. Indeed, most paradigms present animals with a low cost, low reward response option, or a high-cost, high-reward alternative, while other models present subjects with multiple, albeit concurrently available choices (Cousins and Salamone, 1994; de Visser et al, 2011; Winstanley and Floresco, 2016). These paradigms were designed to model the decisions humans face on a day-to-day basis, but it is not always the case that humans choose between concurrently available options. Rather, the choice is deciding whether to stick with a current course of action, or to move on in hopes that a more fruitful course of action will be available. This “explore” versus “exploit” form of decision making, also known as foraging choice, recruits distinct neural circuits relative to concurrent choice paradigms, and may be related to the accept/reject format used by some newer human effort-based decision-making models (Chong et al, 2015; Daw et al, 2006; Kolling et al, 2012). Currently no rodent model has been developed to assess this type of decision making, but it is interesting to consider whether similar brain regions regulate cognitive effort when the decision is whether to continue engaging in a cognitive episode, or to disengage and instead seek out other cognitively challenging opportunities that might be more lucrative.       229  Conclusion  The experiments carried out in this thesis contribute to a growing literature investigating decision making with cognitive costs. I identified striatal contributions to this process, and confirmed that the BLA plays a critical, albeit nuanced role in the processes guiding cognitive effort allocation. I also demonstrate that akin to decision making with physical effort costs, the orbitofrontal cortices play a minimal role in valuations of cognitive effort. Additionally, I further characterized the previously ascribed role of the cholinergic system in rCET performance. The current findings complement existing work in the domain of physical effort, and support the notion that these two forms of effort-based decision making are mediated by distinct, albeit overlapping circuitries. Now that the main brain regions and neurotransmitters systems have been investigated on the rCET, future work can address specific circuits and neurotransmitter interactions guiding this allocation process. And while this work fundamentally contributes to our understanding of how organisms navigate their environment, it may also have practical utility. Indeed, the rCET may help to identify behavioural or pharmacological therapies that can boost cognitive willingness, and conversely has the potential to better understand the decision-making deficits observed in disorders such as Parkinson’s Disease, schizophrenia, and depression. The motivational dysfunction observed in some of these disorders is predictive of poor functional outcomes (Green et al, 2015), and an inherent unwillingness to exert cognitive effort may preclude attempts to     230 seek out or adhere to often challenging treatment regimens. In this regard, identifying the mechanisms giving rise to these cognitive deficiencies may be a therapeutically lucrative target warranting further investigation.         231 References  Adams WK, Barkus C, Ferland JMN, Sharp T, Winstanley CA (2017a). Pharmacological evidence that 5-HT2C receptor blockade selectively improves decision making when rewards are paired with audiovisual cues in a rat gambling task. Psychopharmacology (Berl) 234: 3091–3104. Adams WK, Haar CV, Tremblay M, Cocker PJ, Silveira MM, Kaur S, et al (2017b). Deep-Brain Stimulation of the Subthalamic Nucleus Selectively Decreases Risky Choice in Risk-Preferring Rats STN-DBS decreases risky decision making Deep-brain stimulation of the subthalamic nucleus selectively decreases risky choice in risk-preferring ra. doi:10.1523/ENEURO.0094-17.2017. Agnoli L, Carli M (2011). SYNERGISTIC INTERACTION OF DOPAMINE D 1 AND GLUTAMATE N-METHYL- D -ASPARTATE RECEPTORS IN THE RAT DORSAL STRIATUM CONTROLS ATTENTION. NSC 185: 39–49. Agnoli L, Carli M (2012). Dorsal – striatal 5-HT 2A and 5-HT 2C receptors control impulsivity and perseverative responding in the 5-choice serial reaction Time Task. 633–645doi:10.1007/s00213-011-2581-0. Agnoli L, Mainolfi P, Invernizzi RW, Carli M (2012). Dopamine D 1 -Like and D 2 -Like Receptors in the Dorsal Striatum Control Different Aspects of Attentional Performance in the Five-Choice Serial Reaction Time Task Under a Condition of Increased Activity of Corticostriatal Inputs. Neuropsychopharmacology 38: 701–714. Akkal D, Bioulac B, Audin J, Burbaud P (2002). Comparison of neuronal activity in the rostral supplementary and cingulate motor areas during a task with cognitive and motor demands. Eur J Neurosci 15: 887–904. Alburges ME, Hunt MAE, McQuade RD, Wamsley JK (1992). D1-receptor antagonists: Comparison of [3H]SCH39166 to [3H]SCH23390. J Chem Neuroanat 5: 357–366. Apps MAJ, Grima LL, Manohar S, Husain M (2015). The role of cognitive effort in subjective reward devaluation and risky decision-making. Sci Rep 5: . Arulpragasam AR, Cooper JA, Nuutinen MR, Treadway MT (2018). Corticoinsular circuits encode subjective value expectation and violation for effortful goal-directed behavior. Proc Natl Acad Sci 201800444doi:10.1073/PNAS.1800444115. Baldwin AE, Holahan MR, Sadeghian K, Kelley AE (2000). N-methyl-D-aspartate receptor-dependent plasticity within a distributed corticostriatal network mediates appetitive instrumental learning. Behav Neurosci 114: 84–98. Balleine BW, Delgado MR, Hikosaka O (2007). The role of the dorsal striatum in reward and decision-making. J Neurosci 27: 8161–8165. Ballinger EC, Ananth M, Talmage DA, Role LW (2016). Basal Forebrain Cholinergic Circuits and Signaling in Cognition and Cognitive Decline. Neuron 91: 1199–1218. Barch DM, Treadway MT, Schoen N (2014). Effort, anhedonia, and function in     232 schizophrenia: Reduced effort allocation predicts amotivation and functional impairment. J Abnorm Psychol 123: 387–397. Bardgett ME, Depenbrock M, Downs N, Points M, Green L (2009). Dopamine modulates effort-based decision making in rats. Behav Neurosci 123: 242–251. 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: 759–767. Barnes NM, Sharp T (1999). A review of central 5-HT receptors and their function. Neuropharmacology 38: 1083–1152. Barrus MM, Hosking JG, Cocker PJ, Winstanley CA (2017). Inactivation of the orbitofrontal cortex reduces irrational choice on a rodent Betting Task. Neuroscience 345: 38–48. Baunez C (1997). Bilateral lesions of the subthalamic nucleus induce multiple deficits in an attentional task in rats. Eur J Neurosci 9: 2086–2099. Baxter MG, Parker A, Lindner CCC, Izquierdo AD, Murray E a (2000). Control of response selection by reinforcer value requires interaction of amygdala and orbital prefrontal cortex. J Neurosci 20: 4311–9. Bayless DW, Darling JS, Stout WJ, Daniel JM (2012). Sex differences in attentional processes in adult rats as measured by performance on the 5-choice serial reaction time task. Behav Brain Res 235: 48–54. Bechara A, Damasio A, Damasio H, Anderson S (1994). Insensitivity to future consequences following damange to human prefrontal cortex. Cognition 50: 7–15. Bechara A, Damasio H, Damasio AR, Lee GP (1999). Different Contributions of the Human Amygdala and Ventromedial Prefrontal Cortex to Decision-Making. 19: 5473–5481. Belova MA, Paton JJ, Morrison SE, Salzman CD (2007). Expectation Modulates Neural Responses to Pleasant and Aversive Stimuli in Primate Amygdala. Neuron 55: 970–984. Belova MA, Paton JJ, Salzman CD (2008). Moment-to-moment tracking of state value in the amygdala. J Neurosci 28: 10023–30. Bolla KI, Eldreth DA, Matochik JA, Cadet JL (2004). Sex-related differences in a gambling task and its neurological correlates. Cereb Cortex 14: 1226–1232. Botly LCP, Rosa E De (2012). Impaired visual search in rats reveals cholinergic contributions to feature binding in visuospatial attention. Cereb Cortex 22: 2441–2453. Botvinick M, Nystrom LE, Fissell K, Carter CS, Cohen JD (1999). Conflict monitoring versus selection-for-action in anterior cingulate cortex. Nature 402: 179–181. Botvinick MM, Braver TS (2015). Motivation and Cognitive Control: From Behavior to Neural Mechanism. Annu Rev Psychol 66: 83–113. Botvinick MM, Huffstetler S, McGuire JT (2009). Effort discounting in human nucleus accumbens. Cogn Affect Behav Neurosci 9: 16–27. Bradfield L, Bertran-Gonzalez J, Chieng B, Balleine BW (2013). The Thalamostriatal Pathway and Cholinergic Control of Goal-Directed Action:     233 Interlacing New with Existing Learning in the Striatum. Neuron 79: 153–166. Brand M, Recknor EC, Grabenhorst F, Bechara A (2007). Decisions under ambiguity and decisions under risk: correlations with executive functions and comparisons of two different gambling tasks with implicit and explicit rules. J Clin Exp Neuropsychol 29: 86–99. Britt JP, Benaliouad F, McDevitt RA, Stuber GD, Wise RA, Bonci A (2012). Synaptic and Behavioral Profile of Multiple Glutamatergic Inputs to the Nucleus Accumbens. Neuron 76: 790–803. Brog JS, Salyapongse A, Deutch AY, Zahm DS (1993). The patterns of afferent innervation of the core and shell in the “accumbens” part of the rat ventral striatum: Immunohistochemical detection of retrogradely transported fluoro-gold. J Comp Neurol 338: 255–278. Bryce CA, Floresco SB (2016). Perturbations in Effort-Related Decision-Making Driven by Acute Stress and Corticotropin-Releasing Factor. Neuropsychopharmacology 41: 2147–2159. Bubser M, Byun N, Wood MR, Jones CK (2012). Muscarinic receptor pharmacology and circuitry for the modulation of cognition. Handb Exp Pharmacol 208: . Burnstock G (2009). Autonomic Neurotransmission: 60 Years Since Sir Henry Dale. Annu Rev Pharmacol Toxicol 49: 1–30. Cagniard B, Beeler JA, Britt JP, McGehee DS, Marinelli M, Zhuang X (2006). Dopamine Scales Performance in the Absence of New Learning. Neuron 51: 541–547. Camille N, Tsuchida A, Fellows LK (2011). Double Dissociation of Stimulus-Value and Action-Value Learning in Humans with Orbitofrontal or Anterior Cingulate Cortex Damage. J Neurosci 31: 15048–15052. Cardinal RN, Howes NJ (2005). Effects of lesions of the nucleus accumbens core on choice between small certain rewards and large uncertain rewards in rats. BMC Neurosci 6: 37. Cardinal RN, Pennicott DR, Sugathapala CL, Robbins TW, Everitt BJ (2001). Impulsive choice induced in rats by lesions of the nucleus accumbens core. Sci (New York, NY) 292: 2499–2501. Carli M, Robbins TW, Evenden JL, Everitt BJ (1983). Effects of lesions to ascending noradrenergic neurones on performance of a 5-choice serial reaction task in rats; implications for theories of dorsal noradrenergic bundle function based on selective attention and arousal. Behav Brain Res 9: 361–380. Carter ME, Yizhar O, Chikahisa S, Nguyen H, Adamantidis A, Nishino S, et al (2010). Tuning arousal with optogenetic modulation of locus coeruleus neurons. Nat Neurosci 13: 1526–1535. Chang SE, Todd TP, Bucci DJ, Smith KS (2015). Chemogenetic manipulation of ventral pallidal neurons impairs acquisition of sign-tracking in rats. Eur J Neurosci 42: 3105–3116. Chaudhri N, Woods CA, Sahuque LL, Gill TM, Janak PH (2013). Unilateral inactivation of the basolateral amygdala attenuates context-induced renewal of Pavlovian-conditioned alcohol-seeking. Eur J Neurosci 38: 2751–2761.     234 Chen E, Lallai V, Sherafat Y, Grimes NP, Pushkin AN, Fowler J, et al (2018). Altered Baseline and Nicotine-Mediated Behavioral and Cholinergic Profiles in ChAT-Cre Mouse Lines. J Neurosci 38: 1433–17. Cherek DR, Lane SD, Dougherty DM (2002). Possible amotivational effects following marijuana smoking under laboratory conditions. Exp Clin Psychopharmacol 10: 26–38. Chong TT-J, Apps M, Giehl K, Sillence A, Grima LL, Husain M (2017). Neurocomputational mechanisms underlying subjective valuation of effort costs. PLOS Biol 15: e1002598. Chong TTJ, Bonnelle V, Manohar S, Veromann KR, Muhammed K, Tofaris GK, et al (2015). Dopamine enhances willingness to exert effort for reward in Parkinson’s disease. Cortex 69: 40–46. Christakou A, Robbins TW, Everitt BJ (2001). Functional disconnection of a prefrontal cortical-dorsal striatal system disrupts choice reaction time performance: Implications for attentional function. Behav Neurosci 115: 812–825. Christakou A, Robbins TW, Everitt BJ (2004). Prefrontal Cortical – Ventral Striatal Interactions Involved in Affective Modulation of Attentional Performance  : Implications for Corticostriatal Circuit Function. 24: 773–780. Chudasama Y, Baunez C, Robbins TW (2003a). Functional disconnection of the medial prefrontal cortex and subthalamic nucleus in attentional performance: evidence for corticosubthalamic interaction. J Neurosci 23: 5477–5485. Chudasama Y, Dalley JW, Nathwani FF, Bouger P, Robbins TW, Nathwani FF (2004). Cholinergic modulation of visual attention and working memory: dissociable effects of basal forebrain 192-IgG-saporin lesions and intraprefrontal infusions of scopolamine. Learn Mem 11: 78–86. Chudasama Y, Passetti F, Rhodes SE V, Lopian D, Desai A, Robbins TW (2003b). 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: 105–119. Chudasama Y, Robbins TW (2006). Functions of frontostriatal systems in cognition: Comparative neuropsychopharmacological studies in rats, monkeys and humans. Biol Psychol 73: 19–38. Cocker PJ, Dinelle K, Kornelson R, Sossi V, Winstanley CA (2012a). Irrational Choice under Uncertainty Correlates with Lower Striatal D2/3 Receptor Binding in Rats. J Neurosci 32: 15450–15457. Cocker PJ, Hosking JG, Benoit J, Winstanley C a (2012b). Sensitivity to cognitive effort mediates psychostimulant effects on a novel rodent cost/benefit decision-making task. Neuropsychopharmacology 37: 1825–37. Cole BJ, Robbins TW (1987). Amphetamine impairs the discrimination performance of rats with dorsal bundle lesions on a 5-choice serial reaction time task: new evidence for cental dopaminergic-noradrenergic interactions. Psychopharmacology (Berl) 91: 458–466. Coleman-Mesches K, Salinas JA, McGaugh JL (1996). Unilateral amygdala inactivation after training attenuates memory for reduced reward. Behav     235 Brain Res 77: 175–180. Cools R, Ivry RB, D’Esposito M (2006). The Human Striatum is Necessary for Responding to Changes in Stimulus Relevance. J Cogn Neurosci 18: 1973–1983. Court J, Martin-Ruiz C, Piggott M, Spurden D, Griffiths M, Perry E (2001). Nicotinic receptor abnormalities in Alzheimer’s disease. Biol Psychiatry 49: 175–184. Cousins MS, Salamone JD (1994). Nucleus accumbens dopamine depletions in rats affect relative response allocation in a novel cost/benefit procedure. Pharmacol Biochem Behav 49: 85–91. Cousins MS, Sokolowski JD, Salamone JD (1993). Different effects of nucleus accumbens and ventrolateral striatal dopamine depletions on instrumental response selection in the rat. Pharmacol Biochem Behav 46: 943–951. Cousins MS, Wei W, Salamone JD (1994). Pharmacological characterization of performance on a concurrent lever pressing/feeding choice procedure: effects of dopamine antagonist, cholinomimetic, sedative and stimulant drugs. Psychopharmacology (Berl) 116: 529–537. Cowen SL, Davis G a., Nitz D a. (2012). Anterior cingulate neurons in the rat map anticipated effort and reward to their associated action sequences. J Neurophysiol 107: 2393–2407. Cromwell HC (2003). Effects of Expectations for Different Reward Magnitudes on Neuronal Activity in Primate Striatum. J Neurophysiol 89: 2823–2838. Croxson PL, Walton ME, Boorman ED, Rushworth MFS, Bannerman DM (2014). Unilateral medial frontal cortex lesions cause a cognitive decision-making deficit in rats. Eur J Neurosci 40: 3757–3765. Croxson PL, Walton ME, O’Reilly JX, Behrens TEJ, Rushworth MFS (2009). Effort-based cost-benefit valuation and the human brain. J Neurosci 29: 4531–41. Culbreth A, Westbrook A, Barch D (2016a). Negative Symptoms Are Associated With an Increased Subjective Cost of Cognitive Effort. J Abnorm Psychol 125: 528–536. Culbreth A, Westbrook A, Barch D (2016b). Negative symptoms are associated with an increased subjective cost of cognitive effort. J Abnorm Psychol 125: 528–536. Dalley JW, Cardinal RN, Robbins TW (2004). Prefrontal executive and cognitive functions in rodents: neural and neurochemical substrates. Neurosci Biobehav Rev 28: 771–84. Dalley JW, McGaughy J, O’Connell MT, Cardinal RN, Levita L, Robbins TW (2001). Distinct changes in cortical acetylcholine and noradrenaline efflux during contingent and noncontingent performance of a visual attentional task. J Neurosci 21: 4908–4914. Dautan D, Huerta-Ocampo I, Witten IB, Deisseroth K, Bolam JP, Gerdjikov T, et al (2014). A Major External Source of Cholinergic Innervation of the Striatum and Nucleus Accumbens Originates in the Brainstem. J Neurosci 34: 4509–4518. Daw ND, O’Doherty JP, Dayan P, Seymour B, Dolan RJ (2006). Cortical     236 substrates for exploratory decisions in humans. Nature 441: 876–9. Day JJ, Jones JL, Wightman RM, Carelli RM (2010). Phasic nucleus accumbens dopamine release encodes effort- and delay-related costs. Biol Psychiatry 68: 306–309. Denk F, Walton ME, Jennings KA, Sharp T, Rushworth MFS, Bannerman DM (2005). Differential involvement of serotonin and dopamine systems in cost-benefit decisions about delay or effort. Psychopharmacology (Berl) 179: 587–596. Dixon ML, Christoff K (2012). The Decision to Engage Cognitive Control Is Driven by Expected Reward-Value: Neural and Behavioral Evidence. PLoS One 7: . Docx L, La Asuncion J De, Sabbe B, Hoste L, Baeten R, Warnaerts N, et al (2015). Effort discounting and its association with negative symptoms in schizophrenia. Cogn Neuropsychiatry 20: 172–185. Draper A, Koch RM, Meer JW van der, Apps M, Pickkers P, Husain M, et al (2017). Effort but not Reward Sensitivity is Altered by Acute Sickness Induced by Experimental Endotoxemia in Humans. Neuropsychopharmacology 43: 1107–1118. Duijn E Van, Craufurd D, Hubers AAM, Giltay EJ, Bonelli R, Rickards H, et al (2014). Neuropsychiatric symptoms in a European Huntington ’ s disease cohort ( REGISTRY ). 1411–1418doi:10.1136/jnnp-2013-307343. Eagle DM, Humby T, Dunnett SB, Robbins TW (1999). Effects of regional striatal lesions on motor, motivational, and executive aspects of progressive-ratio performance in rats. Behav Neurosci 113: 718–731. Egeth H, Kahneman D (1975). Attention and Effort. Am J Psychol 88: 339. Eldridge MAG, Lerchner W, Saunders RC, Kaneko H, Krausz KW, Gonzalez FJ, et al (2015). Chemogenetic disconnection of monkey orbitofrontal and rhinal cortex reversibly disrupts reward value. Nat Neurosci 19: 37–39. Endepols H, Sommer S, Backes H, Wiedermann D, Graf R, Hauber W (2010). Effort-based decision making in the rat: an [18F]fluorodeoxyglucose micro positron emission tomography study. J Neurosci 30: 9708–14. Farrar  a M, Segovia KN, Randall P a, Nunes EJ, Collins LE, Stopper CM, et al (2010). Nucleus accumbens and effort-related functions: behavioral and neural markers of the interactions between adenosine A2A and dopamine D2 receptors. Neuroscience 166: 1056–67. Farrar AM, Font L, Pereira M, Mingote S, Bunce JG, Chrobak JJ, et al (2008). Forebrain circuitry involved in effort-related choice: Injections of the GABAA agonist muscimol into ventral pallidum alter response allocation in food-seeking behavior. Neuroscience 152: 321–330. Farrar AM, Pereira M, Velasco F, Hockemeyer J, Müller CE, Salamone JD (2007). Adenosine A2Areceptor antagonism reverses the effects of dopamine receptor antagonism on instrumental output and effort-related choice in the rat: Implications for studies of psychomotor slowing. Psychopharmacology (Berl) 191: 579–586. Fatahi Z, Haghparast A (2018). Activation of the cannabinoid system in the nucleus accumbens affects effort-based decision making. Pharmacol     237 Biochem Behav 165: 29–35. Feja M, Hayn L, Koch M (2014). Nucleus accumbens core and shell inactivation differentially affects impulsive behaviours in rats. Prog Neuro-Psychopharmacology Biol Psychiatry 54: 31–42. 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: 1291–1302. Felger JC, Treadway MT (2017). Inflammation Effects on Motivation and Motor Activity: Role of Dopamine. Neuropsychopharmacology 42: 216–241. Ferguson SM, Phillips PEM, Roth BL, Wess J, Neumaier JF (2013). Direct-pathway striatal neurons regulate the retention of decision-making strategies. J Neurosci 33: 11668–76. 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: 1590–1596. Fiorillo CD, Tobler PN, Schultz W (2003). Discrete Coding of Reward Probability and Uncertainty by Dopamine Neurons Published by  : American Association for the Advancement of Science Stable URL  : http://www.jstor.org/stable/3833875 Accessed  : 14-04-2016 13  : 57 UTC. Science (80- ) 299: 1898–1902. Flavell CR, Lee JLC (2012). Post-training unilateral amygdala lesions selectively impair contextual fear memories. Learn Mem 19: 256–263. Fletcher PJ, Rizos Z, Noble K, Higgins G a (2011). Impulsive action induced by amphetamine, cocaine and MK801 is reduced by 5-HT(2C) receptor stimulation and 5-HT(2A) receptor blockade. Neuropharmacology 61: 468–77. Fletcher PJ, Tampakeras M, Sinyard J, Higgins GA (2007). Opposing effects of 5-HT(2A) and 5-HT (2C) receptor antagonists in the rat and mouse on premature responding in the five-choice serial reaction time test. Psychopharmacol 195: 223–234. Floresco SB (2006). Dissociable Roles for the Nucleus Accumbens Core and Shell in Regulating Set Shifting. J Neurosci 26: 2449–2457. Floresco SB (2015). The Nucleus Accumbens: An Interface Between Cognition, Emotion, and Action. Annu Rev Psychol 66: 25–52. Floresco SB, Blaha CD, Yang CR, Phillips AG (2001). Modulation of hippocampal and amygdalar-evoked activity of nucleus accumbens neurons by dopamine: cellular mechanisms of input selection. J Neurosci 21: 2851–2860. Floresco SB, Ghods-Sharifi S (2007). Amygdala-prefrontal cortical circuitry regulates effort-based decision making. Cereb Cortex 17: 251–60. Floresco SB, St Onge JR, Ghods-Sharifi S, Winstanley C a (2008a). Cortico-limbic-striatal circuits subserving different forms of cost-benefit decision making. Cogn Affect Behav Neurosci 8: 375–89. Floresco SB, Tse MTL, Ghods-Sharifi S (2008b). Dopaminergic and     238 glutamatergic regulation of effort- and delay-based decision making. Neuropsychopharmacology 33: 1966–1979. Font L, Mingote S, Farrar AM, Pereira M, Worden L, Stopper C, et al (2008). Intra-accumbens injections of the adenosine A2A agonist CGS 21680 affect effort-related choice behavior in rats. Psychopharmacology (Berl) 199: 515–526. Forcelli PA, West EA, Murnen AT, Malkova L (2012). Ventral pallidum mediates amygdala-evoked deficits in prepulse inhibition. Behav Neurosci 126: 290–300. Frank MJ, Loughry B, O’Reilly RC (2001). Interactions between frontal cortex and basal ganglia in working memory: A computational model. Cognitive 1: 137–160. Friedman A, Homma D, Gibb LG, Amemori KI, Rubin SJ, Hood AS, et al (2015). A corticostriatal path targeting striosomes controls decision-making under conflict. Cell 161: 1320–1333. Fuchs RA, Eaddy JL, Su ZI, Bell GH (2007). Interactions of the basolateral amygdala with the dorsal hippocampus and dorsomedial prefrontal cortex regulate drug context-induced reinstatement of cocaine-seeking in rats. Eur J Neurosci 26: 487–498. Gan JO, Walton ME, Phillips PEM (2010). Dissociable cost and benefit encoding of future rewards by mesolimbic dopamine. Nat Neurosci 13: 25–27. Geaney JT, Treadway MT, Smillie LD (2015). Trait Anticipatory Pleasure Predicts Effort Expenditure for Reward. 1–17doi:10.5061/dryad.nm13s. Ghods-Sharifi S, Floresco SB (2010). Differential effects on effort discounting induced by inactivations of the nucleus accumbens core or shell. Behav Neurosci 124: 179–91. Ghods-Sharifi S, St Onge JR, Floresco SB (2009a). Fundamental contribution by the basolateral amygdala to different forms of decision making. J Neurosci 29: 5251–9. Ghods-Sharifi S, St Onge JR, Floresco SB (2009b). Fundamental contribution by the basolateral amygdala to different forms of decision making. J Neurosci 29: 5251–9. Gilman JM, Treadway MT, Curran MT, Calderon V, Evins AE (2015). Effect of social influence on effort-allocation for monetary rewards. PLoS One 10: . Gold JM, Kool W, Botvinick MM, Hubzin L, August S, Waltz JA (2014). Cognitive effort avoidance and detection in people with schizophrenia. Cogn Affect Behav Neurosci 15: 145–154. Gold JM, Strauss GP, Waltz JA, Robinson BM, Brown JK, Frank MJ (2013). Negative symptoms of schizophrenia are associated with abnormal effort-cost computations. Biol Psychiatry 74: 130–136. Gold JM, Waltz JA, Frank MJ (2015). Effort Cost Computation in Schizophrenia: A Commentary on the Recent Literature. Biol Psychiatry 78: 747–753. Gomez JL, Bonaventura J, Lesniak W, Mathews WB, Sysa-Shah P, Rodriguez LA, et al (2017). Chemogenetics revealed: DREADD occupancy and activation via converted clozapine. Science (80- ) 357: 503–507. Goschke T (2014). Dysfunctions of decision-making and cognitive control as     239 transdiagnostic mechanisms of mental disorders: advances, gaps, and needs in current research. Int J Methods Psychiatr Res 23 Suppl 1: 41–57. Granon S, Poucet B (2000). Involvement of the rat prefrontal cortex in cognitive functions  : A central role for the prelimbic area. Psychobiology 28: 229–237. Green MF, Butler PD, Chen Y, Geyer M a, Silverstein S, Wynn JK, et al (2009). Perception measurement in clinical trials of schizophrenia: promising paradigms from CNTRICS. Schizophr Bull 35: 163–81. Green MF, Horan WP, Barch DM, Gold JM (2015). Effort-based decision making: A novel approach for assessing motivation in schizophrenia. Schizophr Bull 41: 1035–1044. Gremel CM, Costa RM (2013). Orbitofrontal and striatal circuits dynamically encode the shift between goal-directed and habitual actions. Nat Commun 4: . Gritti I, Henny P, Galloni F, Mainville L, Mariotti M, Jones BE (2006). Stereological estimates of the basal forebrain cell population in the rat, including neurons containing choline acetyltransferase, glutamic acid decarboxylase or phosphate-activated glutaminase and colocalizing vesicular glutamate transporters. Neuroscience 143: 1051–1064. Gritton HJ, Howe WM, Mallory CS, Hetrick VL, Berke JD, Sarter M (2016). Cortical cholinergic signaling controls the detection of cues. Proc Natl Acad Sci 113: E1089–E1097. Haber SN (2016). Corticostriatal circuitry. Dialogues Clin Neurosci 18: 7–21. Haber SN, Behrens TEJ (2014). The Neural Network Underlying Incentive-Based Learning: Implications for Interpreting Circuit Disruptions in Psychiatric Disorders. Neuron 83: 1019–1039. Hajós M, Richards CD, Székely AD, Sharp T (1998). An electrophysiological and neuroanatomical study of the medial prefrontal cortical projection to the midbrain raphe nuclei in the rat. Neuroscience 87: 95–108. Harati H, Barbelivien A, Cosquer B, Majchrzak M, Cassel JC (2008). Selective cholinergic lesions in the rat nucleus basalis magnocellularis with limited damage in the medial septum specifically alter attention performance in the five-choice serial reaction time task. Neuroscience 153: 72–83. Hart EE, Gerson JO, Zoken Y, Garcia M, Izquierdo A (2017). Anterior cingulate cortex supports effort allocation towards a qualitatively preferred option. Eur J Neurosci 46: 1682–1688. Hart EE, Izquierdo A (2017). Basolateral amygdala supports the maintenance of value and effortful choice of a preferred option. Eur J Neurosci 45: 388–397. Hartmann MN, Hager OM, Reimann A V., Chumbley JR, Kirschner M, Seifritz E, et al (2015). Apathy but Not Diminished Expression in Schizophrenia Is Associated with Discounting of Monetary Rewards by Physical Effort. Schizophr Bull 41: 503–512. Hasselmo ME (2006). The role of acetylcholine in learning and memory. Curr Opin Neurobiol 16: 710–715. Hasselt FN van, Visser L de, Tieskens JM, Cornelisse S, Baars AM, Lavrijsen M, et al (2012). Individual variations in maternal care early in life correlate with later life decision-making and c-Fos expression in prefrontal subregions of     240 rats. PLoS One 7: . Hauber W, Sommer S (2009). Prefrontostriatal circuitry regulates effort-related decision making. Cereb Cortex 19: 2240–2247. Hayden BY, Platt ML (2010). Neurons in Anterior Cingulate Cortex Multiplex Information about Reward and Action. J Neurosci 30: 3339–3346. Heidbreder CA, Groenewegen HJ (2003). The medial prefrontal cortex in the rat: Evidence for a dorso-ventral distinction based upon functional and anatomical characteristics. Neurosci Biobehav Rev 27: 555–579. Heilbronner SR, Rodriguez-Romaguera J, Quirk GJ, Groenewegen HJ, Haber SN (2016). Circuit-Based Corticostriatal Homologies Between Rat and Primate. Biol Psychiatry 80: 509–521. Heron C Le, Plant O, Manohar S, Ang Y-S, Jackson M, Lennox G, et al (2018). Distinct effects of apathy and dopamine on effort-based decision making in Parkinson’s disease. Brain 1455–1469doi:10.1093/brain/awy110. Higley MJ, Picciotto MR (2014). Neuromodulation by acetylcholine: Examples from schizophrenia and depression. Curr Opin Neurobiol 29: 88–95. Hillman KL, Bilkey DK (2010). Neurons in the Rat Anterior Cingulate Cortex Dynamically Encode Cost-Benefit in a Spatial Decision-Making Task. J Neurosci 30: 7705–7713. Himmelheber AM, Sarter M, Bruno JP (2000). Increases in cortical acetylcholine release during sustained attention performance in rats. Cogn Brain Res 9: 313–325. Holec V, Pirot HL, Euston DR (2014). Not all effort is equal: the role of the anterior cingulate cortex in different forms of effort-reward decisions. Front Behav Neurosci 8: 12. Hollerman JR, Tremblay L, Schultz W (1998). Influence of reward expectation on behavior-related neuronal activity in primate striatum. J Neurophysiol 80: 947–963. Hollon NG, Arnold MM, Gan JO, Walton ME, Phillips PEM (2014). Dopamine-associated cached values are not sufficient as the basis for action selection. Proc Natl Acad Sci 111: 18357–18362. Hoover WB, Vertes RP (2011). Projections of the medial orbital and ventral orbital cortex in the rat. J Comp Neurol 519: 3766–3801. Hosking JG, Cocker PJ, Winstanley CA (2014a). Dissociable contributions of anterior cingulate cortex and basolateral amygdala on a rodent cost/benefit decision-making task of cognitive effort. Neuropsychopharmacology 39: 1558–67. Hosking JG, Cocker PJ, Winstanley C a. (2015a). Prefrontal Cortical Inactivations Decrease Willingness to Expend Cognitive Effort on a Rodent Cost/Benefit Decision-Making Task. Cereb Cortex 1–10doi:10.1093/cercor/bhu321. Hosking JG, Floresco SB, Winstanley CA (2014b). Dopamine Antagonism Decreases Willingness to Expend Physical, But Not Cognitive, Effort: A Comparison of Two Rodent Cost/Benefit Decision-Making Tasks. Neuropsychopharmacology 1–11doi:10.1038/npp.2014.285. Hosking JG, Floresco SB, Winstanley CA (2015b). Dopamine antagonism     241 decreases willingness to expend physical, but not cognitive, effort: a comparison of two rodent cost/benefit decision-making tasks. Neuropsychopharmacology 40: 1005–15. Hosking JG, Lam FCW, Winstanley CA (2014c). Nicotine increases impulsivity and decreases willingness to exert cognitive effort despite improving attention in “slacker” rats: insights into cholinergic regulation of cost/benefit decision making. PLoS One 9: e111580. Howells FM, Stein DJ, Russell VA (2010). Perceived mental effort correlates with changes in tonic arousal during attentional tasks. Behav Brain Funct 6: . Hull CL (1943). Principles of Behavior: An Introduction to Behavior Theory. J Abnorm Soc Psychol 39: 377–380. Hutchison MA, Gu X, Adrover MF, Lee MR, Hnasko TS, Alvarez VA, et al (2017). Genetic inhibition of neurotransmission reveals role of glutamatergic input to dopamine neurons in high-effort behavior. Mol Psychiatry 1213–1225doi:10.1038/mp.2017.7. Hutson PH, Tarazi FI, Madhoo M, Slawecki C, Patkar A a (2014). Preclinical pharmacology of amphetamine: implications for the treatment of neuropsychiatric disorders. Pharmacol Ther 143: 253–64. Iodice P, Ferrante C, Brunetti L, Cabib S, Protasi F, Walton ME, et al (2017). Fatigue modulates dopamine availability and promotes flexible choice reversals during decision making. Sci Rep 7: 1–11. Izquierdo A (2004). Combined Unilateral Lesions of the Amygdala and Orbital Prefrontal Cortex Impair Affective Processing in Rhesus Monkeys. J Neurophysiol 91: 2023–2039. Izquierdo A (2017). Functional Heterogeneity within Rat Orbitofrontal Cortex in Reward Learning and Decision Making. J Neurosci 37: 10529–10540. Jakala P, Sirvio J, Jolkkonen J, Riekkinen  Jr. P, Acsady L, Riekkinen P (1992). The effects of p-chlorophenylalanine-induced serotonin synthesis inhibition and muscarinic blockade on the performance of rats in a 5- choice serial reaction time task. Behav Brain Res 51: 29–40. Jarvis MF, Williams M (1989). Direct autoradiographic localization of adenosine A2 receptors in the rat brain using the A2-selective agonist, [3H]CGS 21680. Eur J Pharmacol 168: 243–246. Jenison RL, Rangel A, Oya H, Kawasaki H, Howard M a (2011). Value encoding in single neurons in the human amygdala during decision making. J Neurosci 31: 331–8. Jones BE (2004). Activity, modulation and role of basal forebrain cholinergic neurons innervating the cerebral cortex. Acetylcholine Cereb Cortex Volume 145: 157–169. Jones C a, Watson DJG, Fone KCF (2011). Animal models of schizophrenia. Br J Pharmacol 164: 1162–94. Kable JW, Glimcher PW (2007). The neural correlates of subjective value during intertemporal choice. Nat Neurosci 10: 1625–1633. Kahneman D, Tversky A (1979). Prospect Theory: An Analysis of Decision under Risk. Econometrica 47: 263–292. Kawagoe R, Takikawa Y, Hikosaka O (1998). Expectation of reward modulates     242 cognitive signals in the basal ganglia. Nat Neurosci 1: 411–416. Kawaguchi Y, Wilson CJ, Augood SJ, Emson PC (1995). Striatal interneurones: chemical, physiological and morphological characterization. Trends Neurosci 18: 527–535. Kennerley SW, Dahmubed AF, Lara AH, Wallis JD (2009). Neurons in the frontal lobe encode the value of multiple decision variables. J Cogn Neurosci 21: 1162–78. Kennerley SWS, Behrens TTEJ, Wallis JJD (2011). Double dissociation of value computations in orbitofrontal and anterior cingulate neurons. Nat Neurosci 14: 1581–1589. Khani A, Kermani M, Hesam S, Haghparast A, Argandoña EG, Rainer G (2015). Activation of cannabinoid system in anterior cingulate cortex and orbitofrontal cortex modulates cost-benefit decision making. Psychopharmacology (Berl) 232: 2097–112. Kim H, Lee D, Jung MW (2013). Signals for previous goal choice persist in the dorsomedial, but not dorsolateral striatum of rats. J Neurosci 33: 52–63. Kimchi EY, Laubach M (2009). Dynamic encoding of action selection by the medial striatum. J Neurosci 29: 3148–3159. Klein-Flugge MC, Kennerley SW, Friston K, Bestmann S (2016). Neural Signatures of Value Comparison in Human Cingulate Cortex during Decisions Requiring an Effort-Reward Trade-off. J Neurosci 36: 10002–10015. Klinkenberg I, Sambeth A, Blokland A (2011). Acetylcholine and attention. Behav Brain Res 221: 430–442. Kolling N, Behrens TEJ, Mars RB, Rushworth MFS (2012). Neural mechanisms of foraging. Science (80- ) 335: 95–98. Kool W, McGuire JT, Rosen ZB, Botvinick MM (2010a). Decision making and the avoidance of cognitive demand. J Exp Psychol Gen 139: 665–682. Kool W, McGuire JT, Rosen ZB, Botvinick MM (2010b). Decision making and the avoidance of cognitive demand. J Exp Psychol Gen 139: 665–682. Kool W, McGuire JT, Wang GJ, Botvinick MM (2013). Neural and Behavioral Evidence for an Intrinsic Cost of Self-Control. PLoS One 8: . Kronholm E, Puusniekka R, Jokela J, Villberg J, Urrila AS, Paunio T, et al (2015). Trends in self-reported sleep problems, tiredness and related school performance among Finnish adolescents from 1984 to 2011. J Sleep Res 24: 3–10. Kuhnen CM, Knutson B (2005). The neural basis of financial risk taking. Neuron 47: 763–70. Kurniawan IT, Guitart-masip M, Dayan P, Dolan RJ (2013). Effort and Valuation in the Brain  : The Effects of Anticipation and Execution. 33: 6160–6169. Kurniawan IT, Seymour B, Talmi D, Yoshida W, Chater N, Dolan RJ, et al (2010). Choosing to make an effort: the role of striatum in signaling physical effort of a chosen action. J Neurophysiol 104: 313–21. LaBar KS, LeDoux JE (1996). Partial disruption of fear conditioning in rats with unilateral amygdala damage: Correspondence with unilateral temporal lobectomy in humans. Behav Neurosci 110: 991–997.     243 Lane SD, Cherek DR, Pietras CJ, Steinberg JL (2005a). Performance of heavy marijuana-smoking adolescents on a laboratory measure of motivation. Addict Behav 30: 815–28. Lane SD, Cherek DR, Tcheremissine O V, Lieving LM, Pietras CJ (2005b). Acute marijuana effects on human risk taking. Neuropsychopharmacology 30: 800–9. Laprairie RB, Bagher AM, Kelly MEM, Dupré DJ, Denovan-Wright EM (2014). Type 1 cannabinoid receptor ligands display functional selectivity in a cell culture model of striatal medium spiny projection neurons. J Biol Chem 289: 24845–62. Lashley KS (1931). Mass Action in Cerebral Function. Science (80- ) 73: 245–54. Lasselin J, Treadway MT, Lacourt TE, Soop A, Olsson MJ, Karshikoff B, et al (2017). Lipopolysaccharide Alters Motivated Behavior in a Monetary Reward Task: A Randomized Trial. Neuropsychopharmacology 42: 801–810. Lasseter HC, Xie X, Arguello AA, Wells AM, Hodges MA, Fuchs RA (2014). Contribution of a mesocorticolimbic subcircuit to drug context-induced reinstatement of cocaine-seeking behavior in rats. Neuropsychopharmacology 39: 660–669. Lau B, Glimcher PW (2007). Action and Outcome Encoding in the Primate Caudate Nucleus. J Neurosci 27: 14502–14514. Lau B, Glimcher PW (2008). Value Representations in the Primate Striatum during??Matching Behavior. Neuron 58: 451–463. Lawn W, Freeman TP, Pope RA, Joye A, Harvey L, Hindocha C, et al (2016). Acute and chronic effects of cannabinoids on effort-related decision-making and reward learning: an evaluation of the cannabis “amotivational” hypotheses. Psychopharmacology (Berl) 233: 3537–3552. LeBlanc VR (2009). The effects of acute stress on performance: implications for health professions education. Acad Med 84: S25-33. Lee D, Seo H, Jung MW (2012). Neural Basis of Reinforcement Learning and Decision Making. Annu Rev Neurosci 35: 287–308. Lee TMC, Chan CCH, Leung AWS, Fox PT, Gao J-H (2009). Sex-related differences in neural activity during risk taking: an fMRI study. Cereb Cortex 19: 1303–12. Lehmann O, Grottick AJ, Cassel J, Higgins GA (2003). A double dissociation between serial reaction time and radial maze performance in rats subjected to 192 IgG-saporin lesions of the nucleus basalis and / or the septal region. 18: 651–666. Levy I, Snell J, Nelson AJ, Rustichini A, Glimcher PW (2010). Neural representation of subjective value under risk and ambiguity. J Neurophysiol 103: 1036–1047. Lopez-Gamundi P, Wardle MC (2018). Psychological Assessment The Cognitive Effort Expenditure for Rewards Task (C- EEfRT): A Novel Measure of Willingness to Expend Cognitive Effort The Cognitive Effort Expenditure for Rewards Task (C-EEfRT): A Novel Measure of Willingness to Expend Cognitiv. doi:10.1037/pas0000563. MacDonald AW (2000). Dissociating the Role of the Dorsolateral Prefrontal and     244 Anterior Cingulate Cortex in Cognitive Control. Science (80- ) 288: 1835–1838. Macpherson T, Morita M, Hikida T (2014). Striatal direct and indirect pathways control decision-making behavior. 5: 1–7. Mahler S V, Vazey EM, Beckley JT, Keistler CR, McGlinchey EM, Kaufling J, et al (2014). Designer receptors show role for ventral pallidum input to ventral tegmental area in cocaine seeking. Nat Neurosci 17: 577–585. Mailly P, Aliane V, Groenewegen HJ, Haber SN, Deniau J-M (2013). The Rat Prefrontostriatal System Analyzed in 3D: Evidence for Multiple Interacting Functional Units. J Neurosci 33: 5718–5727. Manes F, Sahakian B, Clark L, Rogers R, Antoun N, Aitken M, et al (2002). Decision-making processes following damage to the prefrontal cortex. Brain 125: 624–639. Mar AC, Walker ALJ, Theobald DE, Eagle DM, Robbins TW (2011). Dissociable Effects of Lesions to Orbitofrontal Cortex Subregions on Impulsive Choice in the Rat. J Neurosci 31: 6398–6404. Massar SAA, Libedinsky C, Weiyan C, Huettel SA, Chee MWL (2015). Separate and overlapping brain areas encode subjective value during delay and effort discounting. Neuroimage 120: 104–113. Massar SAA, Lim J, Sasmita K, Chee MWL (2016). Rewards boost sustained attention through higher effort: A value-based decision making approach. Biol Psychol 120: 21–27. Mata F, Treadway M, Kwok A, Truby H, Yücel M, Stout JC, et al (2017). Reduced Willingness to Expend Effort for Reward in Obesity: Link to Adherence to a 3-Month Weight Loss Intervention. Obesity 25: 1676–1681. McCarthy JM, Treadway MT, Bennett ME, Blanchard JJ (2016). Inefficient effort allocation and negative symptoms in individuals with schizophrenia. Schizophr Res 170: 278–284. Mcdonald AJ (1991). Topographical organization of amygdaloid projections to the caudatoputamen, nucleus accumbens, and related striatal-like areas of the rat brain. Neuroscience 44: 15–33. Mcgaughy J, Dalley JW, Morrison CH, Everitt BJ, Robbins TW (2002). Selective Behavioral and Neurochemical Effects of Cholinergic Lesions Produced by Intrabasalis Infusions of 192 IgG-Saporin on Attentional Performance in a Five-Choice Serial Reaction Time Task. J Neurosci 22: 1905–1913. McGuire JT, Botvinick MM (2010a). Prefrontal cortex, cognitive control, and the registration of decision costs. Proc Natl Acad Sci U S A 107: 7922–7926. McGuire JT, Botvinick MM (2010b). Prefrontal cortex, cognitive control, and the registration of decision costs. Proc Natl Acad Sci 107: 7922–7926. Mendez I a, Damborsky JC, Winzer-Serhan UH, Bizon JL, Setlow B (2013). Α4Β2 and Α7 Nicotinic Acetylcholine Receptor Binding Predicts Choice Preference in Two Cost Benefit Decision-Making Tasks. Neuroscience 230: 121–31. Mendez I a, Gilbert RJ, Bizon JL, Setlow B (2012). Effects of acute administration of nicotinic and muscarinic cholinergic agonists and antagonists on performance in different cost-benefit decision making tasks in rats.     245 Psychopharmacology (Berl) 224: 489–99. Mesulam MM, Mufson EJ, Wainer BH, Levey AI (1983). Central cholinergic pathways in the rat: An overview based on an alternative nomenclature (Ch1-Ch6). Neuroscience 10: 1185–1201. Meyniel F, Goodwin GM, William Deakin JF, Klinge C, Macfadyen C, Milligan H, et al (2016). A specific role for serotonin in overcoming effort cost. Elife 5: . Mobini S, Body S, Ho MY, Bradshaw C, Szabadi E, Deakin J, et al (2002). Effects of lesions of the orbitofrontal cortex on sensitivity to delayed and probabilistic reinforcement. Psychopharmacology (Berl) 160: 290–298. Mogenson GJ, Jones DL, Yim CY (1980). From motivation to action: Functional interface between the limbic system and the motor system. Prog Neurobiol 14: 69–97. Monroe SM, Harkness KL (2005). Life stress, the “kindling” hypothesis, and the recurrence of depression: considerations from a life stress perspective. Psychol Rev 112: 417–45. Moscarello JM, Ben-Shahar O, Ettenberg A (2010). External incentives and internal states guide goal-directed behavior via the differential recruitment of the nucleus accumbens and the medial prefrontal cortex. Neuroscience 170: 468–477. Mott AM, Nunes EJ, Collins LE, Port RG, Sink KS, Hockemeyer J, et al (2009). The adenosine A2A antagonist MSX-3 reverses the effects of the dopamine antagonist haloperidol on effort-related decision making in a T-maze cost/benefit procedure. Psychopharmacology (Berl) 204: 103–112. Muir JL, Everitt BJ, Robbins TW (1996). The Cerebral Cortex of the Rat and Visual Attentional Function: Dissociable Effects of Mediofrontal, Cingulate, Anterior Dorsolateral, and Parietal Cortex Lesions on a Five-Choice Serial Reaction Time Task. Cereb Cortex 6: 470–481. Mulert C, Menzinger E, Leicht G, Pogarell O, Hegerl U (2005). Evidence for a close relationship between conscious effort and anterior cingulate cortex activity. Int J Psychophysiol 56: 65–80. Müller T (2007). Rivastigmine in the treatment of patients with Alzheimer’s disease. Neuropsychiatr Dis Treat 3: 211–218. Müller T, Apps MAJ (2018). Motivational fatigue: A neurocognitive framework for the impact of effortful exertion on subsequent motivation. Neuropsychologia doi:10.1016/j.neuropsychologia.2018.04.030. Naccache L, Dehaene S, Cohen L, Habert MO, Guichart-Gomez E, Galanaud D, et al (2005). Effortless control: Executive attention and conscious feeling of mental effort are dissociable. Neuropsychologia 43: 1318–1328. Namburi P, Beyeler A, Yorozu S, Calhoon GG, Halbert SA, Wichmann R, et al (2015). A circuit mechanism for differentiating positive and negative associations. Nature 520: 675–678. Nicola SM (2007). The nucleus accumbens as part of a basal ganglia action selection circuit. Psychopharmacology (Berl) 191: 521–50. Nicola SM (2010). The Flexible Approach Hypothesis: Unification of Effort and Cue-Responding Hypotheses for the Role of Nucleus Accumbens Dopamine in the Activation of Reward-Seeking Behavior. J Neurosci 30: 16585–16600.     246 Niv Y, Daw ND, Joel D, Dayan P (2007). Tonic dopamine: Opportunity costs and the control of response vigor. Psychopharmacology (Berl) 191: 507–520. Nowend KL, Arizzi M, Carlson BB, Salamone JD (2001). D1 or D2 antagonism in nucleus accumbens core or dorsomedial shell suppresses lever pressing for food but leads to compensatory increases in chow consumption. Pharmacol Biochem Behav 69: 373–382. Nunes EJ, Randall P a, Podurgiel S, Correa M, Salamone JD (2013a). Nucleus accumbens neurotransmission and effort-related choice behavior in food motivation: effects of drugs acting on dopamine, adenosine, and muscarinic acetylcholine receptors. Neurosci Biobehav Rev 37: 2015–25. Nunes EJ, Randall PA, Estrada A, Epling B, Hart EE, Lee CA, et al (2014). Effort-related motivational effects of the pro-inflammatory cytokine interleukin 1-beta: Studies with the concurrent fixed ratio 5/ chow feeding choice task. Psychopharmacology (Berl) 231: 727–736. Nunes EJ, Randall PA, Hart EE, Freeland C, Yohn SE, Baqi Y, et al (2013b). Effort-Related Motivational Effects of the VMAT-2 Inhibitor Tetrabenazine: Implications for Animal Models of the Motivational Symptoms of Depression. J Neurosci 33: 19120–19130. Nunes EJ, Randall PA, Santerre JL, Given AB, Sager TN, Correa M, et al (2010). Differential effects of selective adenosine antagonists on the effort-related impairments induced by dopamine D1 and D2 antagonism. Neuroscience 170: 268–280. O’Reilly RC, Hazy TE, Mollick J, Mackie P, Herd S (2014). Goal-Driven Cognition in the Brain: A Computational Framework. at <http://arxiv.org/abs/1404.7591>. Odeen M, Magnussen LH, Maeland S, Larun L, Eriksen HR, Tveito TH (2013). Systematic review of active workplace interventions to reduce sickness absence. Occup Med 63: 7–16. Onge JR St., Stopper CM, Zahm DS, Floresco SB (2012). Separate Prefrontal-Subcortical Circuits Mediate Different Components of Risk-Based Decision Making. J Neurosci 32: 2886–2899. Öngür D, Price JL (2000). The Organization of Networks within the Orbital and Medial Prefrontal Cortex of Rats , Monkeys and Humans. 206–219. Orsini C a, Trotta RT, Bizon JL, Setlow B (2015a). Dissociable Roles for the Basolateral Amygdala and Orbitofrontal Cortex in Decision-Making under Risk of Punishment. J Neurosci 35: 1368–79. Orsini CA, Willis ML, Gilbert RJ, Bizon JL, Setlow B (2015b). Sex Differences in a Rat Model of Risky Decision Making. Behav Neurosci 130: 50–61. Ostrander S, Cazares VA, Kim C, Cheung S, Gonzalez I, Izquierdo A (2011). Orbitofrontal cortex and basolateral amygdala lesions result in suboptimal and dissociable reward choices on cue-guided effort in rats. Behav Neurosci 125: 350–359. Padoa-Schioppa C, Assad JA (2006). Neurons in the orbitofrontal cortex encode economic value. Nature 441: 223–226. Pardo M, López-Cruz L, Miguel NS, Salamone JD, Correa M (2015). Selection of sucrose concentration depends on the effort required to obtain it: Studies     247 using tetrabenazine, D1, D2, and D3 receptor antagonists. Psychopharmacology (Berl) 232: 2377–2391. Pardo M, Lopez-Cruz L, Valverde O, Ledent C, Baqi Y, Müller CE, et al (2012). Adenosine A2Areceptor antagonism and genetic deletion attenuate the effects of dopamine D2antagonism on effort-based decision making in mice. Neuropharmacology 62: 2068–2077. Parvizi J, Rangarajan V, Shirer WR, Desai N, Greicius MD (2013). The will to persevere induced by electrical stimulation of the human cingulate gyrus. Neuron 80: 1359–67. Pasquereau B, Turner RS (2013). Limited Encoding of Effort by Dopamine Neurons in a Cost-Benefit Trade-off Task. J Neurosci 33: 8288–8300. Passetti F (2002). The Frontal Cortex of the Rat and Visual Attentional Performance: Dissociable Functions of Distinct Medial Prefrontal Subregions. Cereb Cortex 12: 1254–1268. Passetti F, Dalley JW, O’Connell MT, Everitt BJ, Robbins TW (2000). Increased acetylcholine release in the rat medial prefrontal cortex during performance of a visual attentional task. Eur J Neurosci 12: 3051–3058. Paton JJ, Belova MA, Morrison SE, Salzman CD (2006). The primate amygdala represents the positive and negative value of visual stimuli during learning. Nature 439: 865–870. Paxinos G, Watson C (1998). The Rat Brain in Stereotaxic Coordinates Fourth Edition. Acad Press doi:0125476191. Peck CJ, Salzman CD (2014). The Amygdala and Basal Forebrain as a Pathway for Motivationally Guided Attention. J Neurosci 34: 13757–13767. Pettibone DJ, Totaro JA, Pflueger AB (1984). Tetrabenazine-induced depletion of brain monoamines: Characterization and interaction with selected antidepressants. Eur J Pharmacol 102: 425–430. Phillips PEM, Walton ME, Jhou TC (2007). Calculating utility: Preclinical evidence for cost-benefit analysis by mesolimbic dopamine. Psychopharmacology (Berl) 191: 483–495. Posner M, Petersen S, Fox P, Raichle M (1988). Localization of cognitive operations in the human brain. Science (80- ) 240: 1627–1631. Pothuizen HHJ, Jongen-Relo AL, Feldon J (2005). The effects of temporary inactivation of the core and the shell subregions of the nucleus accumbens on prepulse inhibition of the acoustic startle reflex and activity in rats. Neuropsychopharmacology 30: 683–696. Pratt WE, Mizumori SJY (1998). Characteristics of basolateral amygdala neuronal firing on a spatial memory task involving differential reward. Behav Neurosci 112: 554–570. Prevost C, Pessiglione M, Metereau E, Clery-Melin M-L, Dreher J-C (2010). Separate Valuation Subsystems for Delay and Effort Decision Costs. J Neurosci 30: 14080–14090. Raedler TJ, Bymaster FP, Tandon R, Copolov D, Dean B (2007). Towards a muscarinic hypothesis of schizophrenia. Mol Psychiatry 12: 232–46. Ragozzino ME (2007). The contribution of the medial prefrontal cortex, orbitofrontal cortex, and dorsomedial striatum to behavioral flexibility. Ann N     248 Y Acad Sci 1121: 355–375. Randall PA, Lee CA, Nunes EJ, Yohn SE, Nowak V, Khan B, et al (2014). The VMAT-2 inhibitor tetrabenazine affects effort-related decision making in a progressive ratio/chow feeding choice task: Reversal with antidepressant drugs. PLoS One 9: 23–26. Randall PA, Lee CA, Podurgiel SJ, Hart E, Yohn SE, Jones M, et al (2015). Bupropion increases selection of high effort activity in rats tested on a progressive ratio/chow feeding choice procedure: Implications for treatment of effort-related motivational symptoms. Int J Neuropsychopharmacol 18: 1–11. Randall PA, Pardo M, Nunes EJ, López Cruz L, Vemuri VK, Makriyannis A, et al (2012). Dopaminergic Modulation of Effort-Related Choice Behavior as Assessed by a Progressive Ratio Chow Feeding Choice Task: Pharmacological Studies and the Role of Individual Differences. PLoS One 7: 18–22. Randall PA, Vemuri VK, Segovia KN, Torres EF, Hosmer S, Nunes EJ, et al (2010). The novel cannabinoid CB1 antagonist AM6545 suppresses food intake and food-reinforced behavior. Pharmacol Biochem Behav 97: 179–184. Rangel A, Hare T (2010). Neural computations associated with goal-directed choice. Curr Opin Neurobiol 20: 262–270. Reddy LF, Horan WP, Barch DM, Buchanan RW, Dunayevich E, Gold JM, et al (2015). Effort-based decision-making paradigms for clinical trials in schizophrenia: Part 1 - Psychometric characteristics of 5 paradigms. Schizophr Bull 41: 1045–1054. Reddy LF, Horan WP, Barch DM, Buchanan RW, Gold JM, Marder SR, et al (2017). Understanding the Association Between Negative Symptoms and Performance on Effort-Based Decision-Making Tasks: The Importance of Defeatist Performance Beliefs. Schizophr Bull doi:10.1093/schbul/sbx156. Reynolds SM, Zahm DS (2005). Specificity in the projections of prefrontal and insular cortex to ventral striatopallidum and the extended amygdala. J Neurosci 25: 11757–67. Richards JB, Mitchell SH, Wit H de, Seiden LS (1997). Determination of discount functions in rats with an adjusting-amount procedure. J Exp Anal Behav 67: 353–366. Risbrough V, Bontempi B, Menzaghi F (2002). Selective immunolesioning of the basal forebrain cholinergic neurons in rats: Effect on attention using the 5-choice serial reaction time task. Psychopharmacology (Berl) 164: 71–81. Robbins TW (1984). Cortical noradrenaline, attention and arousal. Psychol Med 14: 13–21. Robbins TW (2002). The 5-choice serial reaction time task: behavioural pharmacology and functional neurochemistry. Psychopharmacology (Berl) 163: 362–80. Roesch MR, Calu DJ, Schoenbaum G (2007). Dopamine neurons encode the better option in rats deciding between differently delayed or sized rewards. Nat Neurosci 10: 1615–1624.     249 Rogers RD, Baunez C, Everitt BJ, Robbins TW (2001). Lesions of the medial and lateral striatum in the rat produce differential deficits in attentional performance. Behav Neurosci 115: 799–811. Roth BL (2016). DREADDs for Neuroscientists. Neuron 89: 683–694. Rothschild AJ, Raskin J, Wang CN, Marangell LB, Fava M (2014). The relationship between change in apathy and changes in cognition and functional outcomes in currently non-depressed SSRI-treated patients with major depressive disorder. Compr Psychiatry 55: 1–10. Rudebeck PH, Behrens TE, Kennerley SW, Baxter MG, Buckley MJ, Walton ME, et al (2008). Frontal Cortex Subregions Play Distinct Roles in Choices between Actions and Stimuli. J Neurosci 28: 13775–13785. Rudebeck PH, Walton ME, Smyth AN, Bannerman DM, Rushworth MFS (2006). Separate neural pathways process different decision costs. Nat Neurosci 9: 1161–8. Ruff CC, Fehr E (2014). The neurobiology of rewards and values in social decision making. Nat Rev Neurosci 15: 549–562. Rushworth MFS, Noonan MAP, Boorman ED, Walton ME, Behrens TE (2011a). Frontal Cortex and Reward-Guided Learning and Decision-Making. Neuron 70: 1054–1069. Rushworth MFS, Noonan MP, Boorman ED, Walton ME, Behrens TE (2011b). Frontal Cortex and Reward-Guided Learning and Decision-Making. Neuron 70: 1054–1069. Salamone JD, Correa M (2012). The Mysterious Motivational Functions of Mesolimbic Dopamine. Neuron 76: 470–485. Salamone JD, Correa M, Farrar AM, Nunes EJ, Pardo M (2009a). Dopamine, behavioral economics, and effort. Front Behav Neurosci 3: 13. Salamone JD, Cousins MS, Bucher S (1994). Anhedonia or anergia? Effects of haloperidol and nucleus accumbens dopamine depletion on instrumental response selection in a T-maze cost/benefit procedure. Behav Brain Res 65: 221–229. Salamone JD, Farrar AM, Font L, Patel V, Schlar DE, Nunes EJ, et al (2009b). Differential actions of adenosine A1and A2Aantagonists on the effort-related effects of dopamine D2antagonism. Behav Brain Res 201: 216–222. Salamone JD, McLaughlin PJ, Sink K, Makriyannis A, Parker LA (2007). Cannabinoid CB1receptor inverse agonists and neutral antagonists: Effects on food intake, food-reinforced behavior and food aversions. Physiol Behav 91: 383–388. Salamone JD, Steinpreis R, McCullough L, Smith P, Grebel D, Mahan K (1991). Haloperidol and nucleus accumbens dopamine depletion suppresses lever pressing for food but increase free food consumption in a novel food-choice procedure. Psychopharmacology (Berl) 104: 515–521. Schmidt L, Lebreton M, Cléry-Melin M-L, Daunizeau J, Pessiglione M (2012). Neural mechanisms underlying motivation of mental versus physical effort. PLoS Biol 10: e1001266. Schoenbaum G, Chiba  a a, Gallagher M (1998). Orbitofrontal cortex and basolateral amygdala encode expected outcomes during learning. Nat     250 Neurosci 1: 155–159. Schoenbaum G, Setlow B, Saddoris MP, Gallagher M (2003). Encoding predicted outcome and acquired value in orbitofrontal cortex during cue sampling depends upon input from basolateral amygdala. Neuron 39: 855–867. Schroll H, Hamker FH (2013). Computational models of basal-ganglia pathway functions  : focus on functional neuroanatomy. 7: 1–18. Schultz W (2016). Dopamine reward prediction-error signalling: a two-component response. Nat Rev Neurosci 17: 183–95. Schultz W, Dayan P, Montague PR (1997). A neural substrate of prediction and reward. Science (80- ) 275: 1593–1599. Schulz S, Becker T, Nagel U, Ameln-Mayerhofer A von, Koch M (2013). Chronic co-administration of the cannabinoid receptor agonist WIN55,212-2 during puberty or adulthood reverses 3,4 methylenedioxymetamphetamine (MDMA)-induced deficits in recognition memory but not in effort-based decision making. Pharmacol Biochem Behav 106: 91–100. Schweimer J (2005). Involvement of the rat anterior cingulate cortex in control of instrumental responses guided by reward expectancy. Learn Mem 12: 334–342. Schweimer J, Hauber W (2006). Dopamine D1 receptors in the anterior cingulate cortex regulate effort-based decision making. 777–782doi:10.1101/lm.409306.role. Schweimer J, Saft S, Hauber W (2005). Involvement of catecholamine neurotransmission in the rat anterior cingulate in effort-related decision making. Behav Neurosci 119: 1687–1692. Sesack SR, Deutch AY, Roth RH, Bunney BS (1989). Topographical organization of the efferent projections of the medial prefrontal cortex in the rat: An anterograde tract-tracing study with Phaseolus vulgaris leucoagglutinin. J Comp Neurol 290: 213–242. Sesack SR, Grace A a (2010). Cortico-Basal Ganglia reward network: microcircuitry. Neuropsychopharmacology 35: 27–47. Sescousse G, Redoute J, Dreher J-C (2010). The Architecture of Reward Value Coding in the Human Orbitofrontal Cortex. J Neurosci 30: 13095–13104. Sevy S, Burdick KE, Visweswaraiah H, Abdelmessih S, Lukin M, Yechiam E, et al (2007). Iowa gambling task in schizophrenia: a review and new data in patients with schizophrenia and co-occurring cannabis use disorders. Schizophr Res 92: 74–84. Seymour B, Dolan R (2008). Emotion, decision making, and the amygdala. Neuron 58: 662–71. Shafiei N, Gray M, Viau V, Floresco SB (2012). Acute stress induces selective alterations in cost/benefit decision-making. Neuropsychopharmacology 37: 2194–2209. Shenhav A, Botvinick MM, Cohen JD (2013). The expected value of control: an integrative theory of anterior cingulate cortex function. Neuron 79: 217–40. Shepherd GMG (2013). Corticostriatal connectivity and its role in disease. Nat Rev Neurosci 14: 278–91.     251 Shidara M, Richmond BJ (2002). Anterior Cingulate  : Single Neuronal Signals Related to Degree of Reward Expectancy. 296: 2000–2003. Shima K (1998). Role for Cingulate Motor Area Cells in Voluntary Movement Selection Based on Reward. Science (80- ) 282: 1335–1338. Silveira MM, Adams WK, Morena M, Hill MN, Winstanley CA (2016). Δ 9 -Tetrahydrocannabinol decreases willingness to exert cognitive effort in male rats. J Psychiatry Neurosci 42: 1–8. Silveira MM, Arnold JC, Laviolette SR, Hillard CJ, Celorrio M, Aymerich MS, et al (2017). Seeing through the smoke: Human and animal studies of cannabis use and endocannabinoid signalling in corticolimbic networks. Neurosci Biobehav Rev 76: 380–395. Silveira MM, Malcolm E, Shoaib M, Winstanley C a (2014). Scopolamine and amphetamine produce similar decision-making deficits on a rat gambling task via independent pathways. Behav Brain Res 281C: 86–95. Silveira MM, Tremblay M & WC (2016). Dissociable contributions of dorsal and ventral regions of the striatum on a rodent cost/benefit decision-making task requiring cognitive effort. San Diego Soc Neurosci . Simon NW, Beas BS, Montgomery KS, Haberman RP, Bizon JL, Setlow B (2013). Prefrontal cortical-striatal dopamine receptor mRNA expression predicts distinct forms of impulsivity. Eur J Neurosci 37: 1779–1788. Simon NW, Gilbert RJ, Mayse JD, Bizon JL, Setlow B (2009). Balancing risk and reward: A rat model of risky decision making. Neuropsychopharmacology 34: 2208–2217. Sink KS, McLaughlin PJ, Wood JAT, Brown C, Fan P, Vemuri VK, et al (2008a). The novel cannabinoid CB1receptor neutral antagonist AM4113 suppresses food intake and food-reinforced behavior but does not induce signs of nausea in rats. Neuropsychopharmacology 33: 946–955. Sink KS, Vemuri VK, Olszewska T, Makriyannis  a, Salamone JD (2008b). Cannabinoid CB1 antagonists and dopamine antagonists produce different effects on a task involving response allocation and effort-related choice in food-seeking behavior. Psychopharmacology (Berl) 196: 565–74. Sinz H, Zamarian L, Benke T, Wenning GK, Delazer M (2008). Impact of ambiguity and risk on decision making in mild Alzheimer’s disease. Neuropsychologia 46: 2043–2055. Smith KS, Bucci DJ, Luikart BW, Mahler S V (2016). DREADDs: Use and Application in Behavioral Neuroscience. Behav Neurosci 130: 137–155. Sokolowski JD, Salamone JD (1998). The role of accumbens dopamine in lever pressing and response allocation: Effects of 6-OHDA injected into core and dorsomedial shell. Pharmacol Biochem Behav 59: 557–566. St Onge JR, Floresco SB (2009). Dopaminergic modulation of risk-based decision making. Neuropsychopharmacology 34: 681–97. St Onge JR, Floresco SB (2010). Prefrontal cortical contribution to risk-based decision making. Cereb Cortex 20: 1816–28. Stopper CM, Floresco SB (2011). Contributions of the nucleus accumbens and its subregions to different aspects of risk-based decision making. Cogn Affect Behav Neurosci 11: 97–112.     252 Stopper CM, Floresco SB (2014). What’s better for me? Fundamental role for lateral habenula in promoting subjective decision biases. Nat Neurosci 17: 33–35. Stopper CM, Green EB, Floresco SB (2014). Selective involvement by the medial orbitofrontal cortex in biasing risky, but not impulsive, choice. Cereb Cortex 24: 154–162. Stuber GD, Sparta DR, Stamatakis AM, Leeuwen WA Van, Hardjoprajitno JE, Cho S, et al (2011). nucleus accumbens facilitates reward seeking. Nature 475: 377–380. Sugam JA, Day JJ, Wightman RM, Carelli RM (2012). Phasic nucleus accumbens dopamine encodes risk-based decision-making behavior. Biol Psychiatry 71: 199–205. Tai L-H, Lee AM, Benavidez N, Bonci A, Wilbrecht L (2012). Transient stimulation of distinct subpopulations of striatal neurons mimics changes in action value. Nat Neurosci 15: 1281–9. Takahashi YK, Chang CY, Lucantonio F, Haney RZ, Berg B a, Yau H-J, et al (2013). Neural estimates of imagined outcomes in the orbitofrontal cortex drive behavior and learning. Neuron 80: 507–18. Teles-Grilo Ruivo LM, Baker KL, Conway MW, Kinsley PJ, Gilmour G, Phillips KG, et al (2017). Coordinated Acetylcholine Release in Prefrontal Cortex and Hippocampus Is Associated with Arousal and Reward on Distinct Timescales. Cell Rep 18: 905–917. Threlfell S, Lalic T, Platt NJ, Jennings K a, Deisseroth K, Cragg SJ (2012). Striatal dopamine release is triggered by synchronized activity in cholinergic interneurons. Neuron 75: 58–64. Tobler PN, Fiorillo CD, Schultz W (2005). Adaptive coding of reward value by dopamine neurons. Science 307: 1642–5. Treadway MT, Bossaller NA, Shelton RC, Zald DH (2012a). Effort-based decision-making in major depressive disorder: A translational model of motivational anhedonia. J Abnorm Psychol 121: 553–558. Treadway MT, Buckholtz JW, Cowan RL, Woodward ND, Li R, Ansari MS, et al (2012b). Dopaminergic Mechanisms of Individual Differences in Human Effort-Based Decision-Making. J Neurosci 32: 6170–6176. Treadway MT, Buckholtz JW, Schwartzman AN, Lambert WE, Zald DH (2009). Worth the “EEfRT”? The effort expenditure for rewards task as an objective measure of motivation and anhedonia. PLoS One 4: . Treadway MT, Peterman JS, Zald DH, Park S (2015). Impaired effort allocation in patients with schizophrenia. Schizophr Res 161: 382–385. Uban KA, Rummel J, Floresco SB, Galea LAM (2012). Estradiol Modulates Effort-Based Decision Making in Female Rats. Neuropsychopharmacology 37: 390–401. Urban DJ, Roth BL (2015). DREADDs (Designer Receptors Exclusively Activated by Designer Drugs): Chemogenetic Tools with Therapeutic Utility. Annu Rev Pharmacol Toxicol 55: 399–417. Varazzani C, San-Galli  a., Gilardeau S, Bouret S (2015). Noradrenaline and Dopamine Neurons in the Reward/Effort Trade-Off: A Direct     253 Electrophysiological Comparison in Behaving Monkeys. J Neurosci 35: 7866–7877. Vassena E, Silvetti M, Boehler CN, Achten E, Fias W, Verguts T (2014). Overlapping neural systems represent cognitive effort and reward anticipation. PLoS One 9: . Vazey EM, Aston-Jones G (2014). Designer receptor manipulations reveal a role of the locus coeruleus noradrenergic system in isoflurane general anesthesia. Proc Natl Acad Sci 111: 3859–3864. Vertes RP (2004). Differential Projections of the Infralimbic and Prelimbic Cortex in the Rat. Synapse 51: 32–58. Vichaya EG, Hunt SC, Dantzer R (2014). Lipopolysaccharide reduces incentive motivation while boosting preference for high reward in Mice. Neuropsychopharmacology 39: 2884–2890. Visser L de, Homberg JR, Mitsogiannis M, Zeeb FD, Rivalan M, Fitoussi A, et al (2011). Rodent versions of the iowa gambling task: opportunities and challenges for the understanding of decision-making. Front Neurosci 5: 109. Voon V, Irvine M a, Derbyshire K, Worbe Y, Lange I, Abbott S, et al (2014). Measuring “waiting” impulsivity in substance addictions and binge eating disorder in a novel analogue of rodent serial reaction time task. Biol Psychiatry 75: 148–55. Voorn P, Vanderschuren LJMJ, Groenewegen HJ, Robbins TW, Pennartz CMA (2004). Putting a spin on the dorsal-ventral divide of the striatum. Trends Neurosci 27: 468–474. Walton ME, Bannerman DM, Alterescu K, Rushworth MFS (2003a). Functional Specialization within Medial Frontal Cortex of the Anterior Cingulate for Evaluating Effort-Related Decisions. 23: 6475–6479. Walton ME, Bannerman DM, Alterescu K, Rushworth MFS (2003b). Functional specialization within medial frontal cortex of the anterior cingulate for evaluating effort-related decisions. J Neurosci 23: 6475–6479. Walton ME, Bannerman DM, Rushworth MFS (2002). The role of rat medial frontal cortex in effort-based decision making. J Neurosci 22: 10996–11003. Walton ME, Croxson PL, Rushworth MFS, Bannerman DM (2005). The mesocortical dopamine projection to anterior cingulate cortex plays no role in guiding effort-related decisions. Behav Neurosci 119: 323–8. Walton ME, Groves J, Jennings KA, Croxson PL, Sharp T, Rushworth MFS, et al (2009). Comparing the role of the anterior cingulate cortex and 6-hydroxydopamine nucleus accumbens lesions on operant effort-based decision making. Eur J Neurosci 29: 1678–1691. Walton ME, Kennerley SW, Bannerman DM, Phillips PEM, Rushworth MFS (2006). Weighing up the benefits of work: Behavioral and neural analyses of effort-related decision making. Neural Networks 19: 1302–1314. Wardle MC, Treadway MT, Mayo LM, Zald DH, Wit H de (2011). Amping Up Effort: Effects of d-Amphetamine on Human Effort-Based Decision-Making. J Neurosci 31: 16597–16602. Wardle MC, Treadway MT, Wit H De (2012). Caffeine increases psychomotor performance on the effort expenditure for rewards task. Pharmacol Biochem     254 Behav 102: 526–531. Wassum KM, Izquierdo A (2015). The basolateral amygdala in reward learning and addiction. Neurosci Biobehav Rev 57: 271–283. Westbrook A, Braver T (2015). Cognitive effort: A neuroeconomic approach. Cogn Affect Behav Neurosci 15: 395–415. Westbrook A, Braver T (2016). Dopamine Does Double Duty in Motivating Cognitive Effort. Neuron 89: 695–710. Westbrook A, Kester D, Braver TS (2013). What Is the Subjective Cost of Cognitive Effort? Load, Trait, and Aging Effects Revealed by Economic Preference. PLoS One 8: . Westbrook A1 BT (2016). Dopamine Does Double Duty in Motivating Cognitive Effort. Neuron 89: 695–710. Wiley RG, Oeltmann TN, Lappi DA (1991). Immunolesioning: selective destruction of neurons using immunotoxin to rat NGF receptor. Brain Res 562: 149–153. Wilson RC, Takahashi YK, Schoenbaum G, Niv Y (2014). Orbitofrontal cortex as a cognitive map of task space. Neuron 81: 267–278. Winstanley CA (2004). Contrasting Roles of Basolateral Amygdala and Orbitofrontal Cortex in Impulsive Choice. J Neurosci 24: 4718–4722. Winstanley C a, Dalley JW, Theobald DEH, Robbins TW (2003a). Global 5-HT depletion attenuates the ability of amphetamine to decrease impulsive choice on a delay-discounting task in rats. Psychopharmacology (Berl) 170: 320–31. Winstanley CA, Chudasama Y, Dalley JW, Theobald DEH, Glennon JC, Robbins TW (2003b). Intra-prefrontal 8-OH-DPAT and M100907 improve visuospatial attention and decrease impulsivity on the five-choice serial reaction time task in rats. Psychopharmacology (Berl) 167: 304–314. Winstanley CA, Dalley JW, Theobald DEH, Robbins TW (2004a). Fractioning impulsivity: Contrasting effects of central 5-HT depletion on different measures of impulsive behaviour. Neuropsychopharmacology 29: 1331–1343. Winstanley CA, Floresco SB (2016). Deciphering Decision Making: Variation in Animal Models of Effort- and Uncertainty-Based Choice Reveals Distinct Neural Circuitries Underlying Core Cognitive Processes. J Neurosci 36: 12069–12079. Winstanley CA, Theobald DEH, Dalley JW, Glennon JC, Robbins TW (2004b). 5-HT2A and 5-HT2C receptor antagonists have opposing effects on a measure of impulsivity: Interactions with global 5-HT depletion. Psychopharmacology (Berl) 176: 376–385. Witten IB, Steinberg EE, Lee SY, Davidson TJ, Zalocusky KA, Brodsky M, et al (2011). Recombinase-driver rat lines: Tools, techniques, and optogenetic application to dopamine-mediated reinforcement. Neuron 72: 721–733. Wolf DH, Satterthwaite TD, Kantrowitz JJ, Katchmar N, Vandekar L, Elliott MA, et al (2014). Amotivation in schizophrenia: Integrated assessment with behavioral, clinical, and imaging measures. Schizophr Bull 40: 1328–1337. Worden LT, Shahriari M, Farrar AM, Sink KS, Hockemeyer J, Müller CE, et al     255 (2009). The adenosine A2A antagonist MSX-3 reverses the effort-related effects of dopamine blockade: Differential interaction with D1 and D2 family antagonists. Psychopharmacology (Berl) 203: 489–499. Yohn SE, Arif Y, Haley A, Tripodi G, Baqi Y, Müller CE, et al (2016a). Effort-related motivational effects of the pro-inflammatory cytokine interleukin-6: pharmacological and neurochemical characterization. Psychopharmacology (Berl) 233: 3575–3586. Yohn SE, Collins SL, Contreras-Mora HM, Errante EL, Rowland MA, Correa M, et al (2016b). Not All Antidepressants Are Created Equal: Differential Effects of Monoamine Uptake Inhibitors on Effort-Related Choice Behavior. Neuropsychopharmacology 41: 686–694. Yohn SE, Errante EE, Rosenbloom-Snow A, Somerville M, Rowland M, Tokarski K, et al (2016c). Blockade of uptake for dopamine, but not norepinephrine or 5-HT, increases selection of high effort instrumental activity: Implications for treatment of effort-related motivational symptoms in psychopathology. Neuropharmacology 109: 270–280. Yohn SE, Gogoj A, Haque A, Lopez-Cruz L, Haley A, Huxley P, et al (2016d). Evaluation of the effort-related motivational effects of the novel dopamine uptake inhibitor PRX-14040. Pharmacol Biochem Behav 148: 84–91. Yohn SE, Santerre JL, Nunes EJ, Kozak R, Podurgiel SJ, Correa M, et al (2015a). The role of dopamine D<inf>1</inf> receptor transmission in effort-related choice behavior: Effects of D<inf>1</inf> agonists. Pharmacol Biochem Behav 135: 217–226. Yohn SE, Thompson C, Randall PA, Lee CA, Müller CE, Baqi Y, et al (2015b). The VMAT-2 inhibitor tetrabenazine alters effort-related decision making as measured by the T-maze barrier choice task: Reversal with the adenosine A2Aantagonist MSX-3 and the catecholamine uptake blocker bupropion. Psychopharmacology (Berl) 232: 1313–1323. Zahm DS (2000). An integrative neuroanatomical perspective on some subcortical substrates of adaptive responding with emphasis on the nucleus accumbens. Neurosci Biobehav Rev 24: 85–105. Zahm DS, Brog JS (1992). On the significance of subterritories in the “accumbens” part of the rat ventral striatum. Neuroscience 50: 751–767. Zeeb FD, Floresco SB, Winstanley CA (2010). Contributions of the orbitofrontal cortex to impulsive choice: Interactions with basal levels of impulsivity, dopamine signalling, and reward-related cues. Psychopharmacology (Berl) 211: 87–98. Zeeb FD, Winstanley C a (2011). Lesions of the basolateral amygdala and orbitofrontal cortex differentially affect acquisition and performance of a rodent gambling task. J Neurosci 31: 2197–204. Zeeb FD, Winstanley C a (2013). Functional disconnection of the orbitofrontal cortex and basolateral amygdala impairs acquisition of a rat gambling task and disrupts animals’ ability to alter decision-making behavior after reinforcer devaluation. J Neurosci 33: 6434–43.     256 Appendix 1: Investigating Serotonergic Contributions to Cognitive Effort Allocation Using a Rodent Model of Cost/Benefit Decision Making    Introduction Serotonin plays a limited role in physical effort allocation (Denk et al, 2005), but it is still unknown what, if any, contribution this neurotransmitter makes to decision making with cognitive effort costs. To address this gap in the literature, I will investigate how serotonergic agents working at different receptors affect decision making on the rCET. While the serotonin system is notoriously complex, comprised of over 14 distinct receptors subtypes (Barnes and Sharp, 1999), I focus here on the 5-HT2A and 5-HT2c receptor subtypes. Specifically, I will administer the 5-HT2c agonist Ro-60-0175, the 5-HT2c antagonist SB 242,084, as well as the 5-HT2A antagonist M100907. I chose to study these subtypes because 5-HT2A and 5-HT2c receptors have received particular attention for their roles in regulating motor impulsivity, as assessed by the original 5-CSRTT. For example, the 5-HT2c agonist Ro-60-0175 and 5-HT2A antagonist M100907 both decrease motor impulsivity (Fletcher et al, 2011; Winstanley et al, 2003b, 2004b), while the 5-HT2c antagonist SB 242,084 has the opposite effect (Fletcher et al, 2007; Winstanley et al, 2004b). Thus, I will be able to assess whether doses normally affecting impulse control likewise affect allocation of cognitive resources.      257  Additional Methods   Subjects and Baseline Choice   Subjects were 24 female Long-Evans rats bred in house. These rats were bred from breeding pairs obtained from Charles River Laboratories and the Rat Resource and Research Centre (RRRC, Columbia, MO) as part of a breeding program for transgenic rats that express cre recombinase (Cre) in neurons that contain choline acetyltranferase (ChAT; Long Evans –Tg (ChAT-Cre) 5.1 Deis, RRRC # 00658). Twelve of these rats were positive for the transgene (TG+), while the remainder did not express the transgene (TG-). The transgenic status of the rats was not utilized for the current study, but instead the subjects were considered naïve given they exhibit a similar behavioural profile to Long Evans rats purchased from a commercial supplier. To account for potential differences, transgene status (TG + or TG -) was included as a between-subjects factor for all analyses. These results will not be reported, unless a significant a main effect of TG status or TG x dose interaction is observed for any of the rCET behavioural measures. Animals weighed at least 200 g at the start of the experiment and were food restricted to 85% of their free-feeding weight (maintained on 10 g rat chow daily). Water was available ad libitum. Rats were housed in groups of three or four in a climate controlled colony room maintained at 21° C on a reverse 12-hr light-dark schedule (lights off at 8 am).  For this particular experiment, the mean choice of the HR option across     258 rats was 66%. Animals were grouped as “workers” if they chose the HR option for >70% of trials (n=12) and as “slackers” if they chose HR for ≤70% of trials (n=12), as per previous work (Cocker et al., 2012), thereby enabling consistency when discussing individual differences across studies. We experienced a box issue during the first Latin square (Ro-60-0175), and so one worker was excluded from this analysis, but was included in subsequent Latin squares.  Effects of Systemic R0-60-0175, M100907, and SB 242,084 on rCET Performance  Animals received Ro-60-0175 (0, .1, .3, .6 mg/kg), then M100907 (0, .01, .03, 0.1 mg/kg) and lastly SB 242,084 (0, 0.1, 0.25, 0.5 mg/kg) in a series of counterbalanced Latin squares (four doses: ABCD, BDAC, CDAB, DCBA). Ro-60-0175 was administered subcutaneously, whereas M100907 and SB 242,084 were injected intraperitoneally. All drugs were administered 15 minutes prior to starting the behavioural task. Drug injections were given on a three-day cycle, starting with a baseline session. On the second day, rats received a drug or vehicle injection prior to testing, and on the third day they were not tested. To prevent carry-over effects, animals were tested drug-free for a minimum of one week before starting the next Latin square.   Drugs  All drugs were purchased from Tocris. A stock solution of the highest concentration was prepared, and four aliquots frozen at -20°C. One aliquot was     259 thawed on each drug day and diluted as necessary to produce the required doses. Ro-60-0175 was dissolved in 0.9% saline. M100907 was dissolved in 0.9% saline and pH adjusted to 6.25 using 0.1 M NaOH and 0.1 M HCl. SB 242,084 was dissolved in 25 mM citric acid in 8% cyclodextrine in 0.9% saline. Drug doses were calculated as the salt. Systemic injections of drugs were given in a volume of 1 ml/kg body weight. Doses were selected on the basis of a previously published study from out lab (Adams et al, 2017a).  Data Analyses    Data were analyzed with a repeated measures ANOVA with session (three levels: baseline sessions 1-3) or dose (four levels: vehicle plus three doses) as within-subjects factors, and group (two levels: worker or slacker) and TG (+ or -) status as between-subjects factors for all analyses. If analyses produced significant main effects of dose or dose x group interaction, doses were compared post hoc to vehicle with paired samples t-tests. A bonferonni correction was applied to these subsequent analyses. Significance was set at p < .05 and p values < .1 reported as statistical trend.    Results  Ro-60-0175 Administration  Choice, Accuracy, and Premature Responding     260  Baseline behaviour on the rCET has been discussed in detail previously, and so will be briefly discussed here. As an group, rats selected HR trials more often than LR trials (saline only- choice: F(1, 19) = 26.838, p < .001). In keeping with their group categorization, workers (M = 80.83%) chose HR significantly more than slackers (M = 52.76%)  (saline only- group: F(1, 21)  = 22.739, p < .001). Systemic administration of Ro-60-0175 did not affect choice of HR trials (Dose: F(3, 57) = .257, NS; dose x group: F(3, 57) = 2.192, NS; Figure 7.1a).   Accuracy on HR trials was lower than accuracy on LR trials, in keeping with this option being more attentionally demanding (saline only- choice: F(1, 19) = 102.914, p < .001). In line with previous studies, workers and slackers did not differ in attentional performance (saline only- choice x group: F(1, 19) = 0.00, NS; group: F(1, 19) = .067, NS), suggesting slackers’ reduced preference for the HR option is not driven by a deficit in ability. Across rats Ro-60-0175 affected LR and HR accuracy (LR trials- Dose: F(3, 51) = 3.045, p = .037; Dose x group: F(3, 51) = 1.705, NS; HR trials- Dose: F(2, 37) = 4.535, p = .007; Dose x group: F(2, 37) = 0.582, NS). Follow-up tests revealed that LR accuracy was not significantly affected by Ro-60-0175, but HR accuracy significantly increased at the 0.1, 0.3, and 0.6 mg/kg doses (LR trials- sal vs. 0.1 mg/kg: F(1, 20) = 3.522, p = .225, sal vs. 0.3 mg/kg: F(1, 20) = 1.125, p = .903, sal vs. 0.6 mg/kg: F(1, 20) = 0.600, p = 1.00; HR trials- sal vs. 0.1 mg/kg: F(1, 19) = 12.924, p = .006, sal vs. 0.3 mg/kg = F(1, 19) = 6.776, p = .051, sal vs. 0.6 mg/kg = 13.935, p = .003, Figure 7.1b).    Rates of premature responding did not differ between LR and HR trials (saline only- choice: F(1, 19) = .046, NS), and did not differ between groups     261 (saline only- choice x group: F(1, 19) = 2.913, NS; group: F(1, 19) = .355, NS). Ro-60-0175 produced a trending decrease in premature responding, but none of the individual doses were significant following post hoc tests (Dose: F(3, 51) = 2.284, p = .09; Dose x group: F(1 , 51) = 1.901, NS; all post hoc tests NS). Ro-60-0175 did not affect HR premature responding (Dose: F(3, 54) = 1.164, NS, Dose x group: F(3, 54) = 0.263, NS, Figure 7.1c).    Other Behavioural Measures  When injected with saline, rats initiated ~130 trials on average and made few choice omissions when the levers were extended. This did not differ between groups (all Fs < 0.436, NS). Latencies to make a lever choice, to correctly detect the light stimulus, and to collect reward did not differ between trial types (Saline only- Choice: all Fs < 4.00, NS). Workers and slackers generally did not differ  In latencies to make a choice and to make a correct response (Saline only- Choice x group, Group: all Fs < .718, NS), but slackers were generally slower to collect reward across trials (Saline only- Group: F(1, 19): 7.350, p = .014). Response omissions following stimulus presentation were higher for HR trials, in line with this being the more attentionally demanding trial type (Saline only- Choice: F(1, 19) = 11.630, p = .003). Omissions rates did not differ between groups (Saline only- Choice x group, Group: all Fs < 1.059, NS).   Following Ro-60-0175 administration, trials initiated and trials completed significantly decreased at all doses across groups (Trials initiated- Dose: F(3, 57) = 30.133, p < .001, Dose x group: F(3, 57) = 0.05, NS; sal vs. 0.1 mg/kg: F(1, 22)     262 = 8.90, p = .007, sal vs. 0.3 mg/kg: F(1, 22) = 33.33, p < .001, sal vs. 0.6 mg/kg: F(1, 22) = 75.78, p < .001; Trials completed- Dose: F(3, 57) = 33.595, p < .001, Dose x group: F(3, 57) = 0.028, NS; sal vs. 0.1 mg/kg: F(1, 22) = 10.634, p = .012, sal vs. 0.3 mg/kg: F(1, 22) = 30.758, p < .001, sal vs. 0.6 mg/kg: F(1, 22) = 77.86, p < .001). There was a trending effect on choice omissions, but none of the post hoc comparisons were significant (Dose: F(2, 45) = 2.895, p = .063, all posthoc tests Fs < 5.457, NS). None of the latency measures were affected by Ro-60-0175 (Choice latency, Correct Latency, Collect latency – Dose, Dose x group: all Fs < 1.726, NS). However, the drug had biphasic effects on response omissions, decreasing response omissions on HR trials at the lowest dose, and increasing response omissions for both trial types at the middle dose (LR response omissions- Dose: F(3, 51) = 2.852, p = .048, Dose x group: F(3, 51) = 0.045, NS; sal vs. 0.1 mg/kg: F(1, 20) = 0.822, p = 1.00, sal vs. 0.3 mg/kg: F(1, 20) = 12.447, p = .006, sal vs. 0.6 mg/kg: F(1, 20) = 2.637, p = .360; HR response omissions- Dose: F(3, 54) = 3.214, p = .030, Dose x group: F(3, 54) = 1.477, NS; sal vs. 0.1 mg/kg: F(1, 21) = 8.976, p = .007, sal vs. 0.3 mg/kg: F(1, 21) = 9.078, p = .007, sal vs. 0.6 mg/kg: F(1, 21) = 3.276, p = .085, Figure 7.1d).  M100 907 Administration Choice, Accuracy, and Premature Responding  M100 907 did not affect choice (Dose: F(3, 60) = 1.95, NS; Dose x group: F(3, 60) = 1.31, NS, Figure 7.2a), and did not affect accuracy for LR or HR trials (LR trials- Dose: F(3, 57) = 1.051, NS, Dose x group: F(3, 57) = 2.577, p = .063;     263 HR trials – Dose: F(3, 57) = 0.321, NS, Dose x group: F(3, 57) = 1.716, NS, Figure 7.2b). Systemic administration of M100 907 did not affect premature responding for LR trials (LR trials- Dose: F(3, 57) = 1.367, NS, Dose x group: F(3, 57) = 0.322, NS) but decreased motor impulsivity at all doses for HR trials (HR trials- Dose: F(3, 57) = 6.115, p < .001, Dose x group: F(3, 57) = 1.095, NS; veh versus 0.01 mg/kg: F(1, 22) = 16.17, p = .003; veh versus 0.03 mg/kg: F(1, 22) = 7.96, p = .03; veh versus 0.1 mg.kg: F(1, 22) = 11.73, p = .006, Figure 7.2c).   Other Behavioural Measures   M100 907 increased choice omissions at the middle and high dose when both levers were extended at trial onset (Dose: F(3, 60) = 7.309, p < .001, Dose x group: F(3, 60) = 0.057, NS; veh versus 0.01 mg/kg: F(1, 23) = 0.03, NS; veh versus 0.03 mg/kg: F(1, 23) = 8.11, p = .03; veh versus 0.1 mg.kg: F(1, 23) = 11.12, p = .009), and increased the latencies to make an LR or HR choice (LR trials- Dose: F(2, 45) = 4.910, p = .009, Dose x group: F(2, 45) = 2.508, p = .084; veh versus 0.01 mg/kg: F(1, 22) = 2.17, NS; veh versus 0.03 mg/kg: F(1, 22) = 6.02, NS; veh versus 0.1 mg.kg: F(1, 22) = 13.14, p = .003; HR trials- Dose: F(2, 40) = 8.509, p = .001, Dose x group: F(2, 40) = 1.380, NS; veh versus 0.01 mg/kg: F(1, 22) = 4.101, NS; veh versus 0.03 mg/kg: F(1, 22) = 18.122, p < .001; veh versus 0.1 mg.kg: F(1, 22) = 10.24, p = .012). Latencies to make a correct response were affected for HR, but not LR trials at all doses (HR trials- Dose: F(3, 57) = 8.138, p < .001, Dose x group: F(3, 57) = 0.587, NS; veh versus 0.01     264 mg/kg: F(1, 22) =13.513, p = .003; veh versus 0.03 mg/kg: F(1, 22) = 19.246, p < .001; veh versus 0.1 mg.kg: F(1, 22) = 17.825, p < .001; LR trials- Dose: F(2, 38) = 2.756, p = .076, Dose x group: F(2, 38) = 0.671, NS). However, latencies to collect reward were unaffected (LR and HR trials- Dose, Dose x group: all Fs < 2.716, NS). Response omissions significantly increased for HR but not LR trials  (HR trials- Dose: F(2, 41) = 5.789, p = .005, Dose x group: F(2, 41) = 1.478, NS; veh versus 0.01 mg/kg: F(1, 22) = 12.738, p = .006; veh versus 0.03 mg/kg: F(1, 22) = 12.917, p = .006; veh versus 0.1 mg.kg: F(1, 22) = 11.216, p = .009; LR trials- Dose: F(3, 57) = 0.856, NS, Dose x group: F(3, 57) = 0.479, NS, Figure 7.2d). Lastly, M100 907 decreased the number of trials completed at all doses (Dose- F(3, 60) = 10.668, p < .001, Dose x group: F(3, 60) = 0.305, NS; veh versus 0.01 mg/kg: F(1, 23) = 8.346, p = .024; veh versus 0.03 mg/kg: F(1, 23) = 15.713, p = .003; veh versus 0.1 mg.kg: F(1, 23) = 21.856, p < .001).   SB 242 Administration  Choice, Accuracy, and Premature Responding  SB 242, 084 did not affect choice (Dose: F(3, 60) = 2.139, NS; Dose x group: F(3, 60) = 0.446, NS, Figure 7.3a), and did not affect accuracy for LR or HR trials (LR trials- Dose: F(2, 33) = 1.106, NS, Dose x group: F(2, 33) = 1.830, NS; HR trials – Dose: F(3, 60) = 1.875, NS, Dose x group: F(3, 60) = 0.097, NS, Figure 7.3b). However, SB 242, 084 increased premature responding across trial types at all doses tested (LR trials- Dose: F(3, 57) = 8.503, p < .001, Dose x group: F(3, 60) = 0.643, NS; veh versus 0.1 mg/kg: F(1, 22) = 8.83, p = .021; veh     265 versus 0.25 mg/kg: F(1, 22) = 14.74, p = .003; veh versus 0.5 mg.kg: F(1, 22) = 45.62, p < .001; HR trials- Dose: F(3, 60) = 7.221, p < .001, Dose x group: F(3, 60) = 2.426, NS; veh versus 0.1 mg/kg: F(1, 23) = 12.84, p = .006; veh versus 0.25 mg/kg: F(1, 23) = 14.34, p = .003; veh versus 0.5 mg.kg: F(1, 23) = 11.34, p = .009 Figure 7.3c).   Other Behavioural Measures  SB 242, 084 appeared to invigorate responding on the rCET, increasing trials completed at all doses tested (Dose- F(3, 60) = 14.332, p < .001, Dose x group: F(3, 60) = 2.727, NS; veh versus 0.1 mg/kg: F(1, 23) = 22.629, p < .001; veh versus 0.25 mg/kg: F(1, 23) = 19.892, p < .001; veh versus 0.5 mg.kg: F(1, 23) = 19.874, p < .001). This was accompanied by faster latencies to make a choice or a correct response on either trial type (choice latency: LR trials- Dose: F(3, 57) = 14.438, p < .001, Dose x group: F(3, 57) = 0.326, NS; veh versus 0.1 mg/kg: F(1, 22) = 20.693, p < .001; veh versus 0.25 mg/kg: F(1, 22) = 22.137, p < .001; veh versus 0.5 mg.kg: F(1, 22) = 23.814, p < .001; HR trials- Dose: F(3, 60) = 36.418, p < .001, Dose x group: F(3, 60) = 0.766, NS; veh versus 0.1 mg/kg: F(1, 23) = 40.679, p < .001; veh versus 0.25 mg/kg: F(1, 23) = 71.605, p < .001; veh versus 0.5 mg.kg: F(1, 23) = 61.12, p < .001; Correct latency: LR trials- Dose: F(3, 57) = 9.147, p < .001, Dose x group: F(3, 57) = 0.581, NS; veh versus 0.1 mg/kg: F(1, 22) = 13.928, p = .003; veh versus 0.25 mg/kg: F(1, 22) = 20.512, p < .001; veh versus 0.5 mg.kg: F(1, 22) = 5.712, p = .078; HR trials- Dose: F(3, 60) = 3.943, p = .012, Dose x group: F(3, 60) = 1,152, NS; veh versus 0.1 mg/kg:     266 F(1, 23) = 11.243, p = .009; veh versus 0.25 mg/kg: F(1, 23) = 2.424, NS; veh versus 0.5 mg.kg: F(1, 23) = 10.831, p = .009). In contrast, SB 242, 084 did not affect latencies to collect larger reward following successful stimulus detection, but decreased latencies to collect reward selectively in slacker rats on LR trials at the highest dose (HR trials-Dose: F(2, 32) = 1.753, NS; Dose x group: F(2, 32) = 1.357, NS; LR trials-Dose: F(2, 33) =  5.722, p = .009; Dose x group: F(2, 33) = 4.900, p = .017, Workers only – all Fs < 1.172, NS, Slackers only – Dose: F(2, 15) = 4.226, p = .048, veh versus 0.1 mg/kg: F(1, 11) = 3.042, NS; veh versus 0.25 mg/kg: F(1, 11) = 4.834, NS; veh versus 0.5 mg.kg: F(1, 11) = 11.492, p = .018). Lastly, SB 242, 084 administration did not affect choice or response omissions (Dose, Dose x group: all Fs < 2.379, NS).          267 Discussion and Conclusion   The aim of this experiment was to assess serotonergic contributions to decision making on the rCET. I show here that serotonergic signaling at the 5-HT2A and 5-HT2C receptors has dissociable roles in regulating attention, impulsivity, and decision making as probed by the rCET. None of the drugs affected cognitive effort allocation, but all had subtle effects on either attention or impulsivity. The 5-HT2C agonist Ro-60-175 improved HR trial accuracy, while the 5-HT2C antagonist SB 242, 084 and the 5-HT2A antagonist M100 907 increased and decreased impulsivity, respectively. These effects are in keeping with previous work on the standard 5-CSRTT (Fletcher et al, 2007, 2011, Winstanley et al, 2003b, 2004b), as well as behavioural paradigms such as the rodent Gambling Task which provide cursory measures of impulsive action (Adams et al, 2017a). While the focus of these experiments was limited to the 5-HT2 receptor subtype, the results clearly demonstrate that doses affecting different dimensions of cognitive function were unable to shift rats’ baseline willingness to allocate cognitive resources for reward. When considered with the negligible role of serotonin signaling in decision making with physical effort costs (Denk et al, 2005), the current works suggests that this neurotransmitter system play a minor role in decision making with effort costs.             268  Tables  Table 7.1. Other behavioural measures following systemic Ro-60-0175 administration Means are presented (+ SEM). # p < .1; * p < .05; ** p < .001 Vehicle 0.1 mg/kg 0.3 mg/kg 0.6 mg/kg Choice omissions 7.04 + 1.33 8.70 + 1.59 8.00 + 1.21 11.43 + 1.55 LR choice latency 3.51 + 0.22 3.74 + 0.28 3.68 + 0.25 3.82 + 0.26 HR choice latency 3.07 + 0.22 3.21 + 0.19 3.47 + 0.24 3.38 + 0.21 LR correct latency 0.55 + 0.02 0.57 + 0.02 0.58 + 0.02 0.59 + 0.02 HR correct latency 0.51 + 0.03 0.52 + 0.03 0.50 + 0.02 0.52 + 0.03 LR collect latency 2.04 + 0.22 2.03 + 0.13 2.22 + 0.25 3.61 + 1.50 HR collect latency 1.54 + 0.18 1.61 + 0.23 2.02 + 0.55 1.50 + 0.07 Trials completed 114.30 + 5.52 101.17+ 6.63* 90.09 + 6.60** 69.35 + 5.81**     269   Table 7.2. Other behavioural measures following systemic M100 907 administration Means are presented (+ SEM). # p < .1; * p < .05; ** p < .001  Vehicle 0.01 mg/kg 0.03 mg/kg 0.1 mg/kg Choice omissions 5.63 + 1.45 5.50 + 1.16 8.04+ 1.44* 9.42 + 1.67* LR choice latency 3.22 + 0.22 3.38 + 0.22 3.70 + 0.24 3.82 + 0.24* HR choice latency 3.13 + 0.21 3.34 + 0.21 3.69 + 0.20** 3.69 + 0.21* LR correct latency 0.53 + 0.02 0.58 + 0.03 0.59 + 0.02 0.60 + 0.03 HR correct latency 0.46 + 0.02 0.54 + 0.03* 0.56 + 0.03** 0.58 + 0.03** LR collect latency 2.10 + 0.50 2.75 + 0.64 2.36 + 0.22 2.14 + 0.15 HR collect latency 1.34 + 0.05 1.47 + 0.06 1.60 + 0.12 1.50 + 0.05 Trials completed 125.25 + 4.99 117.50+ 6.08* 108.50 + 5.51* 107.58 + 5.76**     270  Table 7.3. Other behavioural measures following systemic SB 242, 084 administration Means are presented (+ SEM). # p < .1; * p < .05; ** p < .001  Vehicle 0.1 mg/kg 0.25 mg/kg 0.5 mg/kg Choice omissions 4.75 + 1.23 2.58 + 1.12 3.58 + 1.20 3.46 + 0.99 LR choice latency 3.25 + 0.22 2.47 + 0.18** 2.60 + 0.18** 2.33 + 0.15** HR choice latency 3.31 + 0.17 2.50 + 0.16** 2.49 + 0.13** 2.30 + 0.12** LR correct latency 0.55 + 0.02 0.49 + 0.02* 0.46 + 0.01** 0.50 + 0.02 HR correct latency 0.49 + 0.02 0.43 + 0.02* 0.45 + 0.02 0.43 + 0.02* LR collect latency 1.97 + 0.21 1.62 + 0.11 1.51 + 0.12 1.48 + 0.08 HR collect latency 1.71 + 0.29 1.26 + 0.05 1.33 + 0.08 1.46 + 0.17 Trials completed 121.21 + 5.04 136.71+ 4.81** 134.92 + 5.66** 137.08 + 4.27**     271  Figures   Figure 7-1. Effects of the 5-HT2C receptor agonist Ro-60-0175 on (a) choice of HR trials (%), (b) accuracy, (c) premature responses (%), and (d) response omissions (%) Data shown are the mean for each variable (±SEM). LR: Low effort/low reward trial; HR: High effort/high reward trial. Ro-60-0175 did not affect choice at any dose tested, and did not affect rates of premature responding. However, all doses tested increased HR trial accuracy, and the low (0.1 mg/kg) and middle (0.3 mg/kg) doses increased HR response omissions.       272   Figure 7-2. Effects of the 5-HT2A receptor antagonist M100 907 on (a) choice of HR trials (%), (b) accuracy (%), (c) premature responses (%), and (d) response omissions (%) Data shown are the mean for each variable (±SEM). LR: Low effort/low reward trial; HR: High effort/high reward trial. M100 907 did not affect choice at any dose tested, and did not affect accuracy. However, all doses tested decreased premature responding and increased response omissions on HR trials.     273   Figure 7-3. Effects of the 5-HT2C receptor antagonist SB 242, 084 on (a) choice of HR trials (%), (b) accuracy, (c) premature responses (%), and (d) response omissions (%) Data shown are the mean for each variable (±SEM). LR: Low effort/low reward trial; HR: High effort/high reward trial. SB 242, 084 did not affect choice at any dose tested, and did not affect accuracy or response omissions. However, all doses tested increased premature responding across trial types.   

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            data-media="{[{embed.selectedMedia}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.24.1-0372368/manifest

Comment

Related Items