UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Light-based mapping of mouse sensorimotor cortex Harrison, Thomas Clarke 2013

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

Item Metadata

Download

Media
24-ubc_2013_spring_harrison_thomas.pdf [ 9.65MB ]
Metadata
JSON: 24-1.0073652.json
JSON-LD: 24-1.0073652-ld.json
RDF/XML (Pretty): 24-1.0073652-rdf.xml
RDF/JSON: 24-1.0073652-rdf.json
Turtle: 24-1.0073652-turtle.txt
N-Triples: 24-1.0073652-rdf-ntriples.txt
Original Record: 24-1.0073652-source.json
Full Text
24-1.0073652-fulltext.txt
Citation
24-1.0073652.ris

Full Text

LIGHT-BASED MAPPING OF MOUSE SENSORIMOTOR CORTEX  by  Thomas Clarke Harrison B.Sc. (Hons.) University of Victoria 2008     A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  DOCTOR OF PHILOSOPHY  in  The Faculty of Graduate Studies (Neuroscience)     THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  April 2013    © Thomas Clarke Harrison, 2013  ii  Abstract   The motor cortex controls voluntary, skilled movements. It possesses a basic somatotopic organization, but its fine structure remains controversial. Although the reorganization of motor cortex after brain injury is believed to be a critical part of behavioral recovery, our understanding of this process has been constrained by technical limitations. Conventional methods for mapping the motor cortex rely on the insertion of intracortical stimulating electrodes into the brain. We developed a new method for light-based motor mapping that is faster, less invasive, and better suited to repeated mapping in both acute and longitudinal studies. Using this technique, we identified a functional subdivision of the mouse forelimb motor representation according to the direction of movement. Pharmacological and anatomical experiments revealed that the expression of complex movements requires the intact function of the intracortical circuitry, whereas the basic topography of movement in motor cortex may arise primarily from the arrangement of output projections. We also refined methods for intrinsic optical signal sensory imaging, which can be combined with light-based motor mapping to obtain a more complete map of sensorimotor cortex. Performing light-based mapping of sensorimotor cortex for weeks before and after targeted photothrombotic stroke allowed us to monitor cortical reorganization on an unprecedented timescale. We found that if forelimb somatosensory cortex was destroyed by targeted stroke, the forelimb motor cortex was able to incorporate a reorganized sensory map. As a consequence of this increased integration of sensory and motor function after stroke, however, the structure of the motor map was altered, evidenced by a decrease in spatial autocorrelation. Strokes in motor cortex caused increased motor output in peri-infarct cortex, but had no effect on the location or sensitivity of the forelimb somatosensory representation. Light-based motor mapping has provided new insights into both the functional organization of motor cortex and its capacity for spontaneous reorganization after stroke.     iii Preface  My supervisor, Tim Murphy, made important contributions to all published and unpublished material in this dissertation. His contribution is implied where not stated explicitly.  Chapter 1, section 1.1.2 is based on a review article submitted to Frontiers in Neural circuits. I wrote the manuscript with editing assistance from Tim Murphy.  Chapter 2 is based on a publication in Nature Methods: Ayling, O. G. S*., Harrison, T. C.*, Boyd, J. D.*, Goroshkov, A. & Murphy, T. H. (*Equal contributors ) Automated light-based mapping of motor cortex by photoactivation of channelrhodopsin-2 transgenic mice. Nat. Methods 6, 219–224 (2009). Together with Oliver Ayling, I was responsible for collecting and analyzing the data presented in this publication. Jamie Boyd developed the software required for these experiments and Alexander Goroshkov designed the hardware and optics. Oliver Ayling, Tim Murphy and I wrote the manuscript together.  Chapter 3 is based on a publication in the Journal of Neuroscience Methods: Harrison, T. C., Sigler, A. & Murphy, T. H.  Simple and cost-effective hardware and software for functional brain mapping using intrinsic optical signal imaging. J. Neurosci. Methods 182, 211–218 (2009). I collected and analyzed all of the data presented in this publication, and wrote the majority of the manuscript. Albrecht Sigler developed the software described in the publication.  Chapter 4 is based on a publication in Neuron: Harrison, T. C., Ayling, O. G. S. & Murphy, T. H. Distinct cortical circuit mechanisms for complex forelimb movement and motor map topography. Neuron 74, 397–409 (2012). I collected and analyzed all of the data presented in this publication, and wrote the majority of the manuscript.   iv Chapter 5 is based on a manuscript submitted to the Journal of Neuroscience. I collected ~95% of the data, with the remainder collected by Greg Silasi. Jamie Boyd assisted with the development of software for data analysis. I analyzed the data and wrote the manuscript with Tim Murphy. All animal experiments were approved by the Animal Care Committee of the University of British Columbia (Protocols A09-0665 and A10-0140).  v  Table of contents Abstract ...............................................................................................................................ii Preface ............................................................................................................................... iii Table of contents .................................................................................................................v List of tables .....................................................................................................................viii List of figures......................................................................................................................ix List of symbols and abbreviations .....................................................................................xi Acknowledgements...........................................................................................................xiii Dedication.........................................................................................................................xiv Chapter 1: General introduction........................................................................................1 1.1 THE CORTICAL MOTOR SYSTEM ..........................................................................................................................1 1.1.1 General organization of the cortical motor system..................................................................................1 1.1.2 Cortical Movement representations and their underlying microcircuitry..............................................4 1.1.2.1 Movement tuning in motor cortical neurons .......................................................................................................5 1.1.2.2 Macroscopic organization of motor cortex..........................................................................................................6 1.1.2.3 Origins of movement tuning.................................................................................................................................7 1.1.2.4 Relating movement tuning to microcircuit properties ........................................................................................8 1.1.2.5 Projection identity and movement tuning............................................................................................................9 1.1.2.6 Future directions for research in motor microcircuits.......................................................................................11 1.2 THE CORTICAL SOMATOSENSORY SYSTEM.......................................................................................................12 1.2.1 General organization of the somatosensory system...............................................................................12 1.2.2 Integration of sensory and motor systems ..............................................................................................13 1.3 REORGANIZATION OF SENSORIMOTOR CORTEX AFTER STROKE......................................................................14 1.3.1 Stroke and its consequences ....................................................................................................................14 1.3.2 Cortical plasticity .....................................................................................................................................16 1.3.2.1 Cortical plasticity during learning......................................................................................................................16 1.3.2.2 Circuit mechanisms of cortical plasticity ..........................................................................................................17 1.3.2.3 Sensorimotor plasticity after stroke ...................................................................................................................19 1.3.3 Experimental models of ischemic stroke.................................................................................................19 1.3.3.1 Rose bengal photothrombosis.............................................................................................................................19 1.3.3.2 Surgical occlusion of arteries .............................................................................................................................20 1.3.3.3 Endothelin-1 injection.........................................................................................................................................20 1.4 OPTOGENETIC BRAIN STIMULATION .................................................................................................................21 1.4.1 Channelrhodopsin-2: a light activated cation channel..........................................................................21 1.4.2 Engineered variants of Channelrodopsin-2............................................................................................23 1.4.3 Other optogenetic tools ............................................................................................................................25 1.4.4 Techniques for transgene delivery ..........................................................................................................25 1.4.5 Strategies for light delivery .....................................................................................................................26 1.5 RESEARCH HYPOTHESES AND OBJECTIVES.......................................................................................................27 Chapter 2: Automated light-based mapping of motor cortex by photoactivation of channelrhodopsin-2 transgenic mice ................................................................................28 2.1 INTRODUCTION..................................................................................................................................................28 2.2 METHODS ..........................................................................................................................................................29 2.2.1 ANIMALS AND SURGERY ................................................................................................................................29 2.2.2 Optical stimulation parameters ...............................................................................................................30  vi 2.2.3 Optical imaging ........................................................................................................................................31 2.2.4 Testing for the effect of photodamage on motor maps...........................................................................32 2.2.5 Intracortical microstimulation ................................................................................................................33 2.2.6 Motor output recordings ..........................................................................................................................33 2.2.7 Electroencephalography ..........................................................................................................................34 2.2.8 Effects of glutamate receptor antagonists on EMG maps .....................................................................34 2.2.9 Characterization of photoactivation area...............................................................................................35 2.2.10 Software ..................................................................................................................................................36 2.2.11 EMG and pixel based motor map analysis ...........................................................................................37 2.2.12 Histology.................................................................................................................................................38 2.3 RESULTS ............................................................................................................................................................38 2.3.1 Automated mapping of motor cortex using laser light...........................................................................38 2.3.2 Photostimulation elicited homogeneous cortical excitation..................................................................39 2.3.3 Mapping light-evoked muscle potentials in ChR2 mice.........................................................................43 2.3.4 Fine motor map structure ........................................................................................................................52 2.4 DISCUSSION .......................................................................................................................................................56 Chapter 3: Simple and cost-effective hardware and software for functional brain mapping using intrinsic optical signal imaging…………………………………………...59 3.1 INTRODUCTION..................................................................................................................................................59 3.2 METHODS ..........................................................................................................................................................60 3.2.1 Objective-mounted LED ring lights ........................................................................................................60 3.2.2 Powering the LED ring light ...................................................................................................................63 3.2.3 Testing illumination stability ...................................................................................................................63 3.2.4 Cameras and data acquisition .................................................................................................................65 3.2.5 Surgical preparation ................................................................................................................................68 3.2.6 Experimental design.................................................................................................................................69 3.2.7 Image analysis ..........................................................................................................................................70 3.3 RESULTS ............................................................................................................................................................72 3.4 DISCUSSION .......................................................................................................................................................74 3.4.1 Advantages of LED ring lights ................................................................................................................74 3.4.2 Combining IOS with two-photon microscopy.........................................................................................74 3.4.3 Variable illumination wavelength and spectroscopy .............................................................................75 Chapter 4: Distinct cortical circuit mechanisms for complex forelimb movement and motor map topography .....................................................................................................77 4.1 INTRODUCTION..................................................................................................................................................77 4.2 METHODS ..........................................................................................................................................................78 4.2.1 Animals and surgery ................................................................................................................................78 4.2.2 Light-based motor mapping.....................................................................................................................79 4.2.3 Map analysis .............................................................................................................................................79 4.2.4 Video capture of evoked movements .......................................................................................................80 4.2.5 Intracortical microstimulation ................................................................................................................80 4.2.6 Virus injections and anatomical tracing.................................................................................................81 4.2.7 Pharmacology...........................................................................................................................................81 4.2.8 Local field potential recordings ..............................................................................................................82 4.3 RESULTS ............................................................................................................................................................84 4.3.1 Movement-based mapping of motor cortex ............................................................................................84 4.3.2 Forelimb motor cortex is subdivided into functional subregions..........................................................86 4.3.3 Prolonged stimulation of abduction and adduction representations drives movements to distinct positions in space...............................................................................................................................................89 4.3.4 Electrical and optogenetic stimulation evoke similar movements ........................................................95 4.3.5 Specificity of complex movements evoked from different cortical areas requires intracortical synaptic transmission ........................................................................................................................................95 4.3.6 Movement topography is preserved during blockade of intracortical synaptic transmission ............96  vii 4.3.7 Topical application of glutamate receptor antagonists disrupts cortical input without preventing direct activation of ChR2-expressing output neurons ...................................................................................101 4.3.8 Divergent projections from Mab and Mad ...........................................................................................105 4.4 DISCUSSION .....................................................................................................................................................105 4.4.1 Mechanistic basis of multiple motor representations ..........................................................................106 4.4.2 Movement representations in rodents and primates ............................................................................107 4.4.3 Comparison of optogenetic and electrical motor mapping .................................................................108 4.4.4 A rodent model of motor circuitry for complex movements ................................................................110 Chapter 5: Longitudinal light-based mapping of vicarious function in sensorimotor cortex after targeted stroke in mice................................................................................111 5.1 INTRODUCTION................................................................................................................................................111 5.2 METHODS ........................................................................................................................................................112 5.2.1 Animals and surgery ..............................................................................................................................112 5.2.2 Intrinsic optical signal sensory mapping..............................................................................................113 5.2.3 Light-based motor mapping...................................................................................................................113 5.2.4 Photothrombotic stroke..........................................................................................................................114 5.2.5 Histology .................................................................................................................................................115 5.3 RESULTS ..........................................................................................................................................................116 5.3.1 Longitudinal light-based mapping of sensory and motor forelimb representations..........................116 5.3.2 Spatial properties of sensorimotor reorganization ..............................................................................119 5.3.3 Changes in sensorimotor excitability after stroke................................................................................120 5.3.4 Effects of stroke on the integrity of motor representations..................................................................124 5.4 DISCUSSION .....................................................................................................................................................127 5.4.1 Multifunctional peri-infarct cortex........................................................................................................128 5.4.2 Motor map structure after stroke ..........................................................................................................129 5.4.3 Longitudinal mapping of motor reorganization ...................................................................................130 5.4.4 Light-based mapping as an assay for therapeutic and rehabilitative strategies ...............................131 Chapter 6: General discussion........................................................................................132 6.1 STRENGTHS AND LIMITATIONS OF LIGHT-BASED SENSORIMOTOR MAPPING ................................................132 6.1.1 Intrinsic signal imaging .........................................................................................................................132 6.1.2 Light-based motor mapping...................................................................................................................132 6.1.3 Comparison to human non-invasive stimulation techniques...............................................................134 6.1.4 Physiological consequences of light stimulation..................................................................................136 6.1.5 Future improvements for light-based mapping ....................................................................................137 6.2 FUNCTIONAL ORGANIZATION OF MOTOR CORTEX .........................................................................................138 6.2.1 Comparison to other cortical regions ...................................................................................................138 6.2.2 Somatotopy as an organizing principle and its limitations .................................................................140 6.2.3 What does the motor cortex encode? ....................................................................................................140 6.2.4 Relationship between macroscopic maps and neuronal circuits ........................................................141 6.3 SENSORIMOTOR REORGANIZATION AFTER STROKE .......................................................................................144 6.3.1 The brain’s innate capacity for repair after brain injury....................................................................144 6.3.2 Therapeutic strategies for enhancing recovery from stroke................................................................145 6.3.3 Considerations for translation of animal research to clinical practice .............................................145 References........................................................................................................................148   viii List of tables  TABLE 2.1 MOTOR MAP COORDINATES .......................................................................................................................55 TABLE 2.2 MOTOR AND SENSORY MAP OVERLAP .......................................................................................................56   ix List of figures FIGURE 1.1 ORGANIZATION OF PRIMATE AND RODENT SENSORIMOTOR CORTEX .......................................................2 FIGURE 1.1 THE LIGHT SENSITIVE CATION CHANNEL CHR2 ......................................................................................21 FIGURE 1.2 BIOPHYSICAL PROPERTIES OF CHR2 ........................................................................................................22 FIGURE 1.3 COMPARISON OF ENGINEERED VARIANTS OF CHR2................................................................................24 FIGURE 2.1 AUTOMATED LIGHT-BASED MAPPING (LBM) OF THE MOUSE MOTOR CORTEX .....................................39 FIGURE 2.2 CHR2-MEDIATED EEG RESPONSES CAN BE ELICITED FROM ALL REGIONS OF THE EXPOSED CORTEX..40 FIGURE 2.3 CHR2-NEGATIVE ANIMALS SHOW NO RESPONSE TO PHOTOSTIMULATION .............................................41 FIGURE 2.4 CHR2 IS EXPRESSED THROUGHOUT THE SENSORY-MOTOR CORTEX IN LAYER 5 CELLS ........................42 FIGURE 2.5 HIGH-RESOLUTION OPTICALLY STIMULATED MOTOR MAPS....................................................................44 FIGURE 2.6 RESPONSE LATENCY IS INVERSELY RELATED TO EMG AMPLITUDE.......................................................45 FIGURE 2.7 ICMS AND LBM MOTOR MAPS OBTAINED FROM THE SAME MOUSE......................................................46 FIGURE 2.8 CORTICAL APPLICATION OF GLUTAMATE RECEPTOR ANTAGONISTS HAVE LITTLE INITIAL EFFECT ON LIGHT-EVOKED EMG AND EEG ACTIVITY .......................................................................................................47 FIGURE 2.9 ESTIMATES OF CHR2 AND ELECTRODE BASED CORTICAL ACTIVATION SPREAD USING IOS IMAGING .49 FIGURE 2.10 FOCAL AND REPEATED PHOTOSTIMULATION OF MOTOR CORTEX DOES NOT CAUSE DEGRADATION OF MOTOR MAP ........................................................................................................................................................50 FIGURE 2.11 MOTOR MAPS CAN BE EVOKED WEEKS APART WITHIN THE SAME ANIMALS ........................................51 FIGURE 2.12 STIMULATION-EVOKED MOVEMENTS DETECTED BY EMG AND LASER MOTION SENSOR....................52 FIGURE 2.13 MOTOR MAPS ARE STABLE AND REPEATABLE .......................................................................................54 FIGURE 2.14 MOTOR AND SENSORY CORTICAL LIMB REPRESENTATIONS..................................................................55 FIGURE 3.1 LED RING LIGHTS .....................................................................................................................................62 FIGURE 3.2 COMPUTER SYSTEMS AND CIRCUITRY FOR AUTOMATED DATA COLLECTION.........................................65 FIGURE 3.3 STEREOTACTIC APPARATUS FOR MECHANICALLY STABLE IOS IMAGING ..............................................69 FIGURE 3.4 MATHEMATICAL OPERATIONS PERFORMED DURING IOS ANALYSIS ......................................................71 FIGURE 3.5 IOS MAPS OF MOUSE SOMATOSENSORY REPRESENTATIONS...................................................................73 FIGURE 4.1 SPATIAL HETEROGENEITY OF EVOKED MOVEMENTS REVEALED BY LIGHT-BASED MAPPING ................83 FIGURE 4.2 MOTOR MAPS ARE STABLE FOR MONTHS IN ANIMALS IMPLANTED WITH CRANIAL WINDOWS ..............85 FIGURE 4.3 MOVEMENTS EVOKED BY STIMULATION OF VIRALLY-EXPRESSED CHR2 ..............................................86 FIGURE 4.4 RELATIVE POSITIONS OF MOTOR AND SOMATOSENSORY REPRESENTATIONS ........................................88 FIGURE 4.5 COMPLEX MOVEMENTS EVOKED BY PROLONGED STIMULATION OF THE ABDUCTION OR ADDUCTION REPRESENTATIONS..............................................................................................................................................89 FIGURE 4.6 MOVEMENT TRAJECTORIES ARE CONSISTENT FOR A GIVEN SITE OF CORTICAL STIMULATION..............91 FIGURE 4.7 MOVEMENT TOPOGRAPHY IS EVIDENT WITH SHORT PULSES, BUT COMPLEX MOVEMENTS REQUIRE PROLONGED STIMULATION .................................................................................................................................92 FIGURE 4.8 MOTOR MAP AREA IS DEPENDENT ON STIMULUS INTENSITY, BUT COMPLEX MOVEMENTS ARE NOT ....93 FIGURE 4.9 OPTOGENETIC AND ELECTRICAL STIMULATION EVOKE SIMILAR COMPLEX MOVEMENTS .....................94 FIGURE 4.10 GLUTAMATE RECEPTOR ANTAGONISTS DEGRADE THE DIFFERENCES BETWEEN COMPLEX MOVEMENTS EVOKED BY PROLONGED STIMULATION OF MAB AND MAD .............................................................................98 FIGURE 4.11 GLUTAMATE RECEPTOR ANTAGONISTS CAUSE MAP EXPANSION WITHOUT ABOLISHING MOVEMENT REPRESENTATIONS..............................................................................................................................................99 FIGURE 4.12 EFFECT OF REDUCED GABAERGIC INTRACORTICAL TRANSMISSION ON MOVEMENTS EVOKED BY PROLONGED STIMULATION...............................................................................................................................100 FIGURE 4.13 GABA RECEPTOR ANTAGONISTS CAUSE SELECTIVE EXPANSION OF MAD WITHOUT ABOLISHING MOVEMENT REPRESENTATIONS........................................................................................................................101 FIGURE 4.14 GLUTAMATE RECEPTOR ANTAGONISTS BLOCK CORTICAL SYNAPTIC TRANSMISSION BUT NOT DIRECT ACTIVATION OF CHR2-POSITIVE NEURONS .....................................................................................................103 FIGURE 4.15 MAB AND MAD HAVE ADJACENT, NON-OVERLAPPING CORTICOFUGAL PROJECTION PATHWAYS ....104 FIGURE 5.1 OVERVIEW OF EXPERIMENTAL DESIGN ..................................................................................................115 FIGURE 5.2 REPRESENTATIVE EXAMPLES OF SENSORIMOTOR REORGANIZATION AFTER STROKE..........................117 FIGURE 5.3 SPATIAL DISPLACEMENT OF MOTOR SOMATOSENSORY MAPS AFTER STROKE .....................................118 FIGURE 5.4 SENSORY RESPONSES AFTER STROKE.....................................................................................................120 FIGURE 5.5 REGIONAL CHANGES IN MOTOR EXCITABILITY AFTER STROKE ............................................................121  x FIGURE 5.6 OVERALL MOTOR EXCITABILITY IS CONSERVED AFTER STROKE ..........................................................123 FIGURE 5.7 MOTOR MAPS DEVELOP A DIFFUSE STRUCTURE AFTER STROKE ...........................................................126 FIGURE 5.8 MODEL OF CORTICAL PLASTICITY UNDERLYING SENSORIMOTOR MAP REORGANIZATION ..................127 FIGURE 6.1 STRATEGIES FOR TARGETING NEURONS IN MOTOR CORTEX BASED ON THEIR PROJECTION IDENTITY 143          xi List of symbols and abbreviations AAV = adeno-associated virus AMPA = a-amino-3- hydroxyl-5-methyl-4-isoxazole-propionate BNC = bayonet neill-concelman (connector) CCD = charge-coupled device CFA = caudal forelimb area ChR2 = channelrhodopsin-2 CNQX = 6-cyano-7-nitroquinoxaline-2,3-dione DC = direct current EEG = electroencephalogram EMG = electromyogram FL = forelimb GABA = gamma-Aminobutyric acid HEPES = (4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid HL = hindlimb ICMS = intracranial microstimulation IOS = intrinsic optical signal LASER = light amplification by stimulated emission of radiation LBM = light-based mapping LED = light-emitting diode LFP = local field potential LOT = laminar optical tomography M1 = primary motor cortex Mab = motor cortex abduction representation Mad = motor cortex adduction representation mFL = motor representation of contralateral forelimb mHL = motor representation of contralateral hindlimb MK801 = [5R,10S]-[+]-5-methyl-10,11- dihydro-5H-dibenzo[a,d]cyclohepten-5,10-imine MSRI = multispectral reflectance imaging NA = numerical aperture  xii NMDA = N-methyl-D-aspartatic acid PMv = ventral premotor cortex RFA = rostral forelimb area sFL  = sensory representation of contralateral forelimb sHL = sensory representation of contralateral hindlimb Thy1 = thymocyte differntiation antigen 1 (promoter) TIFF = tagged image format file TMS = transcranial magnetic stimulation TTL = transistor-transistor logic VSD = voltage sensitive dye YFP = yellow fluorescent protein     xiii Acknowledgements   I would like to recognize the mentorship of my supervisor, Tim Murphy. Tim has been my teacher, guide and advocate throughout my graduate degree. I also thank the members of my supervisory committee: Nick Swindale, Jeremy Seamans, and Kerry Delaney.   The talented staff of the Murphy lab have contributed their skills, ingenuity, and time to the research that I conducted during my time here. This work would not have been possible without Alexander Goroshkov, Jamie Boyd, Heidi Erb, Cindy Jiang, Jeff LeDue, and Pumin Wang. To the many past and present trainees in the Murphy lab who have been friends and collaborators – thanks for the camaraderie, commiseration, and cookies.   Lastly, I thank my wife, family and friends for their support. Their interest in the functional organization of mouse sensorimotor cortex may have been feigned, but their affection never was.      xiv Dedication  To the scientists who came before me and provided the foundation and inspiration for this work.              1 Chapter 1: General introduction  1.1 The cortical motor system  1.1.1 General organization of the cortical motor system  The motor system of a vertebrate animal can be considered to comprise all neuronal elements that influence the activity of the skeletal musculature to effect movement. This influence is exerted through the final common pathway of the alpha motor neurons of the spinal cord, which form the presynaptic side of neuromuscular junctions and directly elicit muscular contractions. Multiple pathways descend from the cortex and brainstem to terminate on spinal motor neurons, but the corticospinal tract is of primary importance for the dexterous control of the distal musculature for the performance of skilled movements (Lemon, 2008). Primates possess a specialized population of corticospinal neurons whose axons descend from the cortex to directly target a single motoneuron in the spinal cord (Cheney and Fetz, 1984). These corticomotoneuronal (CM) cells are believed to be an evolutionary adaptation that enhances fine control of the distal musculature and their somata are clustered within a frontal region of motor cortex (Rathelot and Strick, 2009). The corticospinal tract originates from many cortical regions and is involved in motor control, gating of nociceptive inputs, modulation of spinal reflexes and control of the autonomic nervous system (Lemon, 2008). The variety of these functions is reflected in the diversity of cortical regions contributing to the corticospinal tract (Dum and Strick, 2002, 2005; Maier et al., 2002), but here we focus on those regions of frontal and parietal cortex most directly related to motor control.  In primates, commonly accepted divisions of the cortical motor system include primary motor cortex (M1), the supplementary motor area (SMA), and premotor cortex, which is further subdivided into dorsal, ventral regions and rostral and caudal subregions (Squire, 2003; Graziano and Aflalo, 2007b). These regions are mostly within the agranular cortical areas 4 and 6 according to Brodmann’s cytoarchitectonic parcellation of cortex. An  2 alternative naming scheme involves areas F1-F7 in frontal cortex and PE, PF, PFG, and PG in parietal cortex (Matelli et al., 1991; Rizzolatti and Luppino, 2001). The situation in rodents has traditionally been considerably simpler, with the only deviation from simple somatotopy being duplicated rostral and caudal forelimb motor areas (RFA and CFA) (Neafsey et al., 1986). Recently, additional detail has been detected in the whisker and forelimb motor representations in mice, which been subdivided according to their output properties and the categories of movement evoked upon stimulation (Matyas et al., 2010; Harrison et al., 2012).  Figure 1.1 Organization of primate and rodent sensorimotor cortex At top is a schematic diagram of motor and somatosensory areas in the cortex of the rhesus macaque, an exemplar of the primate order. CG = cingulate gyrus; FEF = frontal eye field; SMA = supplementary motor area; PM = premotor cortex, subdivided into dorsal, ventral, rostral and caudal sectors; M1 = primary motor cortex; S1 = primary somatosensory cortex; P = parietal motor areas. Modified from Rizzolatti and Luppino 2001. At bottom is a schematic diagram of motor and somatosensory areas associated with the forelimb and hindlimb in the mouse cortex. RFA = rostral forelimb area; CFA = caudal forelimb area; mHL = motor hindlimb cortex; sFL = somatosensory forelimb cortex; sHL = somatosensory hindlimb cortex. The subregions of forelimb motor cortex defined by intracortical microstimulation are overlaid with subregions identified by light-based motor mapping (Mab = motor cortex abduction representation; Mad = motor cortex adduction representation).  3  The cortical motor system receives inputs from other cortical regions, primarily prefrontal, somatosensory and parietal cortex (Rouiller et al., 1993). The multiple subregions of motor cortex can be categorized according to the primary source of their intracortical inputs. Pre-SMA receives input from prefrontal cortex, while SMA and M1 receive largely somatosensory input and premotor cortex receives much of its input from parietal cortex (Luppino and Rizzolatti, 2000; Rizzolatti and Luppino, 2001). M1 also receives input from premotor cortex, but not pre-SMA (Luppino and Rizzolatti, 2000). Posterior motor regions that receive input from parietal cortex project directly to the spinal cord, whereas the pre- SMA projects to the brainstem and not the spinal cord (He et al., 1993, 1995; Luppino and Rizzolatti, 2000).  The motor cortex, like all cortical areas with the exception of primary auditory and visual cortices, sends topographically organized inputs to the basal ganglia. Sensorimotor axons terminate in the dorsolateral striatum, where a single corticostriatal axon may contact thousands of striatal medium spiny neurons. The motor cortex also makes extensive projections to the cerebellum, which are relayed through the pontine nuclei to the cerebrocerebellum. Feedback from the cerebellum to motor cortex originates in the deep cerebellar nuclei and is relayed via the thalamus (Purves, 2004).  The brainstem motor nuclei represent another major target of output projections from motor cortex. The motor cortex contains both corticorubral neurons, which project to the red nucleus, and corticospinal neurons which extend collaterals into the red nucleus before continuing on to the spinal cord. Like the corticospinal tract, the rubrospinal tract is involved in the control of the distal musculature. The rubrospinal tract plays a more prominent role in rodents than it does in primates, which have greater cortical control over movements of the hands. Other descending motor pathways originating in the brainstem include the reticulospinal, vestibulospinal, and tectospinal tracts, which are primarily involved in postural control of the proximal musculature (Squire, 2003). The contributions of the various descending motor pathways were famously dissected in a series of lesion experiments in monkeys. Animals subjected to transections of the corticospinal and rubrospinal tracts were  4 able to jump and climb immediately after surgery, but were impaired in skilled movements of the digits. In contrast, transection of the reticulospinal and vestibuspinal tracts degraded postural control and locomotion, but these animals were still capable of making fine movements of the hand if their bodies were supported (Lawrence and Kuypers, 1968a, 1968b).  Thalamic input to motor cortex includes information arriving from loops with the basal ganglia and cerebellum. In primates, M1 and PMv are interconnected with the thalamic nucleus VPL and VLp, whereas SMA is connected with VLo. This means that M1 and PMv primarily receive information from the basal ganglia, whereas SMA is more strongly associated with the cerebellum (Squire, 2003). Somatotopy within M1 is partially preserved in its connections with VLp. There is also specificity for projections from rostral and caudal M1, which primarily project to VLa and VLp, respectively (Stepniewska et al., 1994).  1.1.2 Cortical movement representations and their underlying microcircuitry  One of the great successes in neuroscience is our detailed understanding of the circuitry and function of the visual system. A well-defined anatomical framework and an established parameter space for visual stimulation have expedited research on the computations performed by the visual cortex. A particularly productive approach has been to develop circuit models of the visual cortex based upon its multiple input channels and to associate these microcircuits with the macroscopic functional areas of visual cortex (Sincich and Horton, 2005). Canonical circuits may be conserved across all cortical regions, but it is apparent that motor cortex, largely devoid of the granular input layer that has anchored the study of the visual system, must be considered in a unique manner (Poggio and Bizzi, 2004; Shipp, 2005; Shepherd, 2009).  Recent experiments have begun to unravel the microcircuitry of the motor cortex using its deep output layers as a reference point. The next step will be to examine how these local microcircuits vary with macroscopic maps of motor function.  Whereas the visual cortex is known to contain multiple overlaid maps of neurons tuned to retinotopic space, ocular dominance, orientation etc. (Swindale, 2000), the  5 topography of motor cortex is not yet fully defined. In addition to the widely accepted somatotopic organization of motor cortex (Penfield and Boldrey, 1937), evidence is accumulating for a mapping of movement directions (Graziano et al., 2002a; Ramanathan et al., 2006; Harrison et al., 2012). The firing of individual neurons in motor cortex can be related to many parameters of movement, but this tuning is less established and more controversial than in sensory cortex. Here, we review circuit properties of neurons in motor cortex that are likely to confer movement tuning in an effort to link local microcircuitry with macroscopic functional maps and motor behavior.  1.1.2.1 Movement tuning in motor cortical neurons  The movement tuning of a neuron in motor cortex refers to the relationship between its firing rate and behavioral variables such as the speed, direction, joint angle, or endpoint of a movement, typically of the contralateral forelimb. Directional tuning of neurons in motor cortex was first observed in recordings made from awake primates performing a two- dimensional centre-out reaching task (Georgopoulos et al., 1982, 1986). The firing rate of individual neurons is coarsely tuned to the direction of arm movement, but the activity of a population of neurons can be linearly transformed into a vector that predicts the speed and direction of arm movement (Georgopoulos et al., 1988; Moran and Schwartz, 1999). However, similar tuning can be demonstrated in monkeys that resist externally applied forces in various directions without moving their forelimbs (Kalaska et al., 1989). Movement tuning can also be affected by the changing the posture of the forelimb (Scott and Kalaska, 1995). For these reasons, there has been debate as to whether preferred movement directions are indeed a fundamental property of motor cortex, or whether they are an epiphenomenon emerging from activity more closely related to control of the peripheral musculature (Todorov, 2000; Scott, 2004). It has also been argued that dynamic neuronal activity encodes movement trajectories rather than instantaneous variables such as direction, speed, or force (Hatsopoulos et al., 2007; Reimer and Hatsopoulos, 2009; Churchland et al., 2012). Regardless of the theoretical framework, both kinetic and kinematic information from motor cortex has been productively exploited in the development of brain-machine interfaces that control the movement of computer cursors, artificial limbs, or paralyzed muscles (Wessberg  6 et al., 2000; Hochberg et al., 2006; Chestek et al., 2007; Velliste et al., 2008; Ethier et al., 2012).  1.1.2.2 Macroscopic organization of motor cortex  The motor cortices of primates (Leyton and Sherrington, 1917; Penfield and Boldrey, 1937; Rizzolatti and Luppino, 2001; Dum and Strick, 2002; Gharbawie et al., 2011a) and rodents (Neafsey and Sievert, 1982; Li and Waters, 1991; Tennant et al., 2010) have long been recognized to possess a topographic map of body parts. Beyond this broad somatotopic parcellation, finer structure has been proposed to exist in motor cortex. The minimal stimulus parameters adopted by practitioners of intracortical microstimulation (ICMS) mapping led to the interpretation of motor cortex as a mosaic of individual columns, each controlling a single muscle in the periphery (Asanuma, 1975). This hypothesis has since been refuted based on electrophysiological and anatomical evidence of multiple colonies of cortical neurons that are distributed broadly throughout cortex yet innervate a single muscle (Jankowska et al., 1975; Rathelot and Strick, 2006). Furthermore, individual corticospinal neurons target multiple pools of motoneurons and can facilitate or suppress several muscles simultaneously (Shinoda et al., 1981; Cheney et al., 1985). Finally, EMG-based mapping has revealed substantial overlap of muscle representations in motor cortex (Donoghue and Wise, 1982; Donoghue et al., 1992; Park et al., 2001; Ayling et al., 2009).  In contrast to the view of motor cortex obtained by mapping with brief, low-intensity stimuli, experiments with prolonged electrical or optogenetic stimulation have reported an organization of motor cortex output based on movement direction or category (Graziano et al., 2002a, 2005; Haiss and Schwarz, 2005; Stepniewska et al., 2005; Ramanathan et al., 2006; Harrison et al., 2012). These stimulation-based experiments broaden the definition of directional tuning in motor cortex, which has traditionally been based on recordings made from neurons during reaching behavior (Georgopoulos et al., 1986). Moreover, they corroborate the clustering of preferred movement directions in motor cortex reported from electrophysiological recordings in primates (Amirikian and Georgopoulos, 2003; Ben-Shaul et al., 2003; Georgopoulos et al., 2007). Directional tuning has also been detected in human  7 motor cortex using functional magnetic resonance imaging (Eisenberg et al., 2010); suggesting the existence of clustering given the relatively coarse spatial resolution of functional magnetic resonance imaging.  The complex topography of motor cortex may reflect the reduction of multiple dimensions of information onto the two-dimensional cortical surface (Aflalo and Graziano, 2006), reminiscent of the organization of multiple feature maps in visual cortex (Swindale et al., 2000). Motor maps may contain clusters of similarly tuned neurons, but the level of detail currently detected in these maps is less than that of tuning maps in sensory areas (Bonhoeffer and Grinvald, 1993; Schreiner and Winer, 2007). For example, calcium imaging has a revealed clustering of neuronal tuning properties in layer 2/3 neurons of rodent motor cortex (Dombeck et al., 2009), but this is less pronounced than that of orientation maps in cat visual cortex (Ohki et al., 2005). Whether this difference is attributable to differences between the species, the nature of the mapped parameter, or the cortical region is an open question.  1.1.2.3 Origins of movement tuning  Movement tuning is defined by firing rate, and must therefore arise from the pattern of input that drives a particular neuron to fire. As in other cortical areas, inhibitory neurons in motor cortex are hypothesized to act in concert with excitatory neurons to shape the tuning of downstream cells (Shapley et al., 2003; Georgopoulos and Stefanis, 2007; Merchant et al., 2008, 2012). Electrophysiological recordings have found that fast-spiking interneurons in motor cortex exhibit broadly tuned response profiles, suggesting that they may contribute to movement tuning by restricting all but the most excited neurons from firing (Isomura et al., 2009; Murray and Keller, 2011). In both rodents and primates, inhibitory neurons increase their firing rates during both movement preparation and execution, suggesting that they are likely to be involved in shaping movments rather than gating them through a sudden release of inhibition (Isomura et al., 2009; Kaufman et al., 2010). Finally, dendritic gating and amplification provide an additional means for the establishment of tuning beyond passive summation of inputs (London and Hausser, 2005; Harnett et al., 2012; Lee et al., 2012). Identifying and characterizing the many inputs that impart tuning to motor cortex output  8 neurons seems as daunting now as when Ramon y Cajal lamented the 'impenetrable thickets' of cortical connections (Ramón y Cajal, 1937). Given that excitatory connectivity within cortical microcircuits is specified by the identities of both the pre- and post-synaptic cells, however, knowledge of the projection identity of motor cortical neurons provides an indication of the source their inputs (Brown and Hestrin, 2009; Anderson et al., 2010). This makes the output layers of motor cortex a useful starting point for circuit analysis akin to the input layers of visual cortex (Shepherd, 2009).  1.1.2.4 Relating movement tuning to microcircuit properties  The fact that recordings from motor cortex can extract useful kinematic information for the control of neural prostheses demonstrates that the activity of neurons in the motor cortex encodes information relevant to the direction of intended movement (Chapin et al., 1999; Simeral et al., 2011). It remains to be determined how the circuit properties of cortical neurons confer their movement tuning. The majority of our knowledge about the firing properties of motor cortical neurons and movement tuning has come from primate studies, where relating neuronal activity to such microcircuit variables is often impractical (Sheets and Shepherd, 2011). The microcircuitry of motor cortex is now being studied intensively in rodents (Isomura et al., 2009; Anderson et al., 2010; Matyas et al., 2010; Mao et al., 2011), which possess many experimental advantages but also lack some features of the primate motor system such as direct corticomotoneurons (Lemon, 2008). An attempt to link motor microcircuits with movement tuning must draw upon both of these animal models while acknowledging the differences between them. In both primates and rodents, neurons can be classified by a set of inter-related attributes, including the region of cortex that they inhabit, laminar position, morphology, and their complement of transmitters and receptors.  The relationship between a neuron’s movement tuning and its location within cortex is not well defined. The existence of multiple movement maps and the broad distribution of motor-related neurons, particularly in primate motor cortex (Gharbawie et al., 2011b), makes it difficult to predict based on cortical position whether a neuron is likely to possess movement tuning. Recordings from M1 provide the most useful signal for neural prostheses  9 (Carmena et al., 2003; Vargas-Irwin et al., 2010), but a greater proportion of neurons in PMv possess “extrinsic-like” tuning to arm movement independent of posture (Kakei et al., 2001). Better established is the link between movement tuning and cortical depth. The concept of the cortical column has been applied to movement tuning in motor cortex, with consistent tuning reported across radial depths of ~500um in primates (Ben-Shaul et al., 2003; Georgopoulos et al., 2007).  Although signals useful for the control of brain machine interfaces can be extracted without penetrating the cortex (Wolpaw and McFarland, 2004), they are strongest in layers 5-6 (Parikh et al., 2009). Microstimulation studies have found movements to be most easily evoked from these deep cortical layers (Donoghue and Wise, 1982; Neafsey and Sievert, 1982; Young et al., 2011). Indeed, selective stimulation of ChR2- expressing L5b neurons yields a motor map subdivided by movement direction (Harrison et al., 2012). Taken together, these observations suggest that movement-tuned neurons in motor cortex are most likely to inhabit the deep cortical layers.  In sensorimotor cortex, as in all cortical areas, the layer occupied by a neuron’s soma is closely related to its projection identity (Hooks et al., 2011; Mao et al., 2011). Neurons in the superficial cortical laminae (2/3) form connections within their layer and send strong projections to layer 5 (Weiler et al., 2008). Connectivity in this pathway is determined by both the projection identity and radial position of the recipient neuron within layer 5 (Anderson et al., 2010). The projection identity of layer 5 neurons is also linked with their intralaminar connectivity (Brown and Hestrin, 2009; Kiritani et al., 2012), morphology (Gao and Zheng, 2004) and intrinsic electrophysiological profile (Hattox and Nelson, 2007; Sheets et al., 2011). Therefore, the projection identity of a motor cortical neuron can be related to both its inputs and its excitability, which together determine a neuron’s tuning properties (Lee et al., 2012).  1.1.2.5 Projection identity and movement tuning  Just as the response properties of granular neurons in visual cortex are largely derived from their inputs (Ferster and Miller, 2000; Huberman et al., 2008), it follows that the tuning of output neurons in motor cortex might be best predicted by their projections. Layer 5  10 pyramidal neurons are likely to be prime contributors to the movement tuning of a columnar microcircuit since they form the majority of the cortical output pathway. Layer 5 neurons are a heterogeneous population and project to many regions, including cortex, thalamus, brainstem, basal ganglia and spinal cord (Veinante and Deschênes, 2003; Kiritani et al., 2012). They can be further subdivided into layers 5A and 5B, each with characteristic gene expression, receptive field, and projections (Lund et al., 1988; Manns et al., 2004; Anderson et al., 2010; Mao et al., 2011). This diverse population of neurons includes corticospinal cells, which are required for skilled movements of the forelimb (Carmel et al., 2010). The corticospinal tract innervates the muscolotopically organized spinal cord (Levine et al., 2012), which orchestrates the synergistic activation of the musculature to achieve movements in stereotyped directions (Bizzi et al., 2002). In the primate corticospinal tract, there is also a group of corticomotoneuronal cells that contact spinal motoneurons directly and whose cell bodies are clustered within a region of motor cortex that is hypothesized to have evolved relatively recently to control fine movements of the distal musculature (Rathelot and Strick, 2009). Not only are these cells likely to possess movement tuning, it is possible that the nature of their tuning could be predicted based on the motoneuron they innervate. Similarly, sub-types of corticospinal neurons may have distinct movement tuning depending on their specific projection identity. Pyramidal (PT) type corticostriatal axons synapse in the ipsilateral striatum before continuing on toward the spinal cord. This differentiates them from intratelencephalic (IT) type corticostriatal neurons which may cross the midline to project to the striatum and contralateral cortex but do not leave the telencephalon. IT and PT type neurons preferentially innervate direct and indirect pathway neurons of the striatum, respectively (Reiner et al., 2010), and these connections are altered by the process of learning to control a neuroprosthetic device (Koralek et al., 2012).  Neurons forming the corticospinal tract are predisposed to possessing movement tuning, but there are many other descending pathways involved in motor control (Lemon, 2008). Nor is the corticospinal tract exclusively motor-related, with only 55 % of primate corticospinal neurons directly facilitating muscle activity (Lemon et al., 1986). Spinal motoneurons receive synaptic input from many sources, meaning that even direct corticomotoneurons have a variable influence on the muscles they innervate (Yanai et al.,  11 2007). In the rodent spinal cord, all descending inputs synapse onto spinal interneurons, making the contribution of this network to movement tuning a further consideration (Levine et al., 2012). The complexity of descending motor pathways provides an explanation for the observation that activity in motor cortex can become uncoupled from movement during sleep, motor imagery or control of brain machine interfaces (Schieber, 2011). In fact, the process of learning to control a brain-machine interface causes widespread changes in the preferred direction tuning of neurons throughout the motor cortex (Ganguly et al., 2011), with the greatest increase in performance coming from neurons in the supplementary motor area (Carmena et al., 2003).  1.1.2.6 Future directions for research in motor microcircuits  Movement tuning appears to be a property of many motor cortical neurons, particularly in the deep output layers. Whether directional tuning reflects a fundamental principle of cortical function or is an epiphenomenon that emerges from of other forms of coding remains controversial. However, these signals have been successfully exploited for direct cortical control of neural prostheses (Hochberg et al., 2006; Velliste et al., 2008). The deep layers of motor cortex contain the most useful signal for these brain machine interfaces and are also the site from which movements can most easily be evoked by stimulation. Neurons in layer 5 form outputs to a variety of structures involved in motor control, including the striatum, brain stem, and spinal cord. The advent of new experimental tools for combined anatomical and physiological circuit tracing based on retrograde transmission of cre-fused wheat germ agglutinin (Gradinaru et al., 2010) or modified rabies virus (Wickersham et al., 2007; Wall et al., 2010; Apicella et al., 2012; Kiritani et al., 2012) has made it possible to label neurons based on projection identity. Although some of these tools are currently used primarily in mice, efforts are underway to apply them in primate models (Diester et al., 2011). This will make it possible to both monitor and manipulate the activity of specific motor output pathways (e.g. corticospinal) to isolate their contribution to motor behavior. It will then become possible to identify the upstream circuit mechanisms that enable flexible control of these pathways during motor imagery or behavior. These  12 experiments will dramatically enhance our understanding of the cortical motor system and the nervous system as a whole.  1.2 The cortical somatosensory system  1.2.1 General organization of the somatosensory system   The exteroceptive somatosensory system serves to transmit information from the periphery to the brain, and it is best described by beginning in the periphery. Somatosensory information is transduced by a variety of receptors, each devoted to particular modalities (e.g. mechanoreceptors, thermoreceptors, nociceptors, and proprioceptors). These receptors are expressed by neurons in the dorsal root ganglia whose axons project centrally (Squire, 2003). Afferents can be classified according to the diameter of their axons. Small-diameter fibres are often unmyelinated, carry nociceptive and thermoceptive information, and decussate in the spinal cord to form the spinothalamic tract, terminating in the ventral posteromedial thalamus. Large-diameter fibres tend to be myelinated, faster conducting and carry proprioceptive and mechanoreceptive information. These fibres ascend to the gracile and cuneate nuclei of the medulla, where secondary neurons decussate and project to the ventral posterolateral thalamus. Somatosensory information from the head is transmitted by the transgeminal system, with an analogous division of fibres devoted to discriminative touch and pain or temperature (Squire, 2003).   From the ventrobasal thalamus, both of these pathways continue to the neocortex. In primates, the somatosensory cortex includes primary and secondary areas, which consist of Brodmann areas 1, 2, 3a and 3b and areas 40 and 43, respectively. The somatosensory cortex of primates is characterized by a somatotopic organization with enlarged representation of body parts with the greatest tactile acuity, such as the hands (Penfield and Boldrey, 1937; Kaas et al., 1979). Rodents also possess somatotopically organized primary and secondary somatosensory cortices (Carvell and Simons, 1986; Liao et al., 2010; Lim et al., 2012; Sweetnam et al., 2012). The somatosensory representation of the whiskers, known as the  13 barrel cortex, is among the best understood of all cortical structures and has become a model system for the study of cortical circuitry in general (Shepherd et al., 2003; Petersen, 2007; Huber et al., 2008; Kleinfeld and Deschênes, 2011; Perin et al., 2011)  1.2.2 Integration of sensory and motor systems   The performance of graceful, co-ordinated motor acts depends on the integration of sensory and motor information, and it should come as no surprise that the distinction between sensory and motor systems is blurred (Cisek and Kalaska, 2010). Proprioception is a faculty that cannot be arbitrarily assigned to either the motor or sensory modalities, and it highlights the functional interdependence of movement and sensation. For example, vibrating a tendon at high frequencies (~80 Hz) causes the illusory perception of movement in a stationary arm (Naito, 2004), and neurons in forelimb motor cortex respond to passive movements of the arm (Rosén and Asanuma, 1972; Strick and Preston, 1982). The response to passive movement by cortical motor neurons displays directional selectivity, with preferred directions matching those for active movement (Cheney and Fetz, 1984).   The cortical motor system operates under the constant influence of sensory information. One manifestation of this interdependence is short-latency afferent inhibition, a phenomenon described in the human transcranial magnetic stimulation literature whereby motor output from the forelimb can be inhibited by preceding stimulation of motor cortex with a brief forelimb sensory stimulus (Tokimura et al., 2000; Voller et al., 2006). The ongoing motor behavior of an animal can also modulate the sensitivity of the somatosensory system, particularly in haptic sensory systems such as the vibrissa of rodents (Petersen, 2007). For example, the latency and amplitude of sensory evoked responses in the rat whisker system depend upon the whisking behavior currently being performed (Ferezou et al., 2007). Local increases in motor cortex activity can facilitate these sensory-evoked responses in somatosensory neurons with a similar somatotopic receptive field  (Lee et al., 2008). Activity in motor cortex likely gates somatosensory afferents through disinhibition of thalamic nuclei, in particular the zona incerta in the case of the rodent vibrissal system (Urbain and Deschênes, 2007).  14   In primates, the posterior parietal cortex plays a central role in the sensorimotor integration that is necessary for coordinated movements of the eyes and hands. The pattern of inputs and outputs from this region make it well suited to multisensory integration and transformations from sensory to motor coordinate space (Kaas et al., 2012). In addition to its role in sensorimotor learning, the posterior parietal cortex is believed to encode movement intentions, making it an attractive target for neuroprosthetic devices (Andersen and Buneo, 2002). In rodents as in primates, parietal cortex is an association area involved in sensorimotor integration, and it is heavily interconnected with both motor (Buneo et al., 2002) and somatosensory regions (Lim et al. 2012).  1.3 Reorganization of sensorimotor cortex after stroke  1.3.1 Stroke and its consequences   The World Health Organization reports an incidence of 15 million strokes each year. Approximately 1/3 of these people die and an additional 1/3 are permanently disabled. In industrialized countries, stroke has an annual incidence of ~4 in 1000 and a mortality rate of 30%, making it the third leading cause of death and the primary cause of adult disability (Dirnagl et al., 1999). Many smaller strokes go unreported, but it is likely that their cumulative effects contribute to vascular dementia that closely resembles Alzheimer’s disease, particularly in elderly patients (Erkinjuntti et al., 2004; Shih et al., 2012).  The brain constitutes only 2% of the mass of an adult body, yet it consumes 20% of the body’s oxygen and 25% of its glucose (Bélanger et al., 2011). The brain’s high metabolic rate and reliance on glucose make it vulnerable to interruptions, even transient, in blood flow. In the majority (~80%) of cases, strokes are focal ischemic events caused by either atherosclerosis or embolism leading to localized decreases in blood flow. If perfusion is not restored within minutes, this decrease in blood flow triggers signaling cascades that lead to irreversible necrosis within the infarct core (Mergenthaler et al., 2004). This process begins with neuronal energy failure, which prevents the maintenance of the membrane potential and  15 reuptake of transmitters such as glutamate and culminates in excitotoxicity leading to cellular edema, increased intracellular [Ca2+] and free radical production (Hossmann, 2006).  The infarct core, where blood flow decreases to 20% of baseline, is rarely salvageable (Nedergaard et al., 1986). The core is surrounded by a penumbra of tissue where reductions in blood flow range from 20 - 70 % and therapeutic protection from delayed apoptotic mechanisms of cell death may be possible (Dirnagl et al., 1999). In the first minutes after stroke ischemic depolarization initiates the structural degradation of neurons within the penumbra, but these structural changes can be reversed by reperfusion as long as one hour after ischemic onset (Li and Murphy, 2008). As minutes turn to hours, waves of spreading depression propagate from the infarct and much of the penumbra succumbs and joins the growing lesion (Hossmann, 1994; Zhang et al., 1994; Fisher and Takano, 1995). In vivo imaging in mice has revealed that although neuronal dendrites are normally located no more than ~13 µm from a blood vessels, dendrites within 80 µm of flowing blood vessels are able to retain their neuronal structure. Functional impairment, however, extended an additional 400 µm into regions of cortex that showed no overt structural alterations (Zhang and Murphy, 2007). By the sixth hour after stroke onset, the infarct has typically expanded to occupy the entire penumbral region. Unfortunately this does not mark the end of cell death, as there is a final phase of injury lasting days to weeks. This delayed damage is related to inflammation, vasogenic edema and apoptotic processes (Hossmann, 2006). Attempts at rehabilitation during this subacute phase after stroke may actually exacerbate the damage and increase the magnitude of the lesion (Risedal et al., 1999).  The rim of tissue surrounding the mature lesion, known as peri-infarct cortex, contains neurons that survived the stroke but are nonetheless profoundly altered. The loss of synaptic partners from the infarct core results in diaschisis, a phenomenon which can also affect remote regions including the contralesional cortex (Witte, 1998). Another hallmark of peri-infarct cortex is hyperexcitability, which extends to regions that receive input from peri- infarct cortex (Domann et al., 1993; Buchkremer-Ratzmann et al., 1996; Buchkremer- Ratzmann and Witte, 1997). There is also an increase in tonic GABAergic inhibition in peri- infarct cortex (Clarkson et al., 2010), causing a disruption of the excitation-inhibition balance  16 that may further impair function in this region (Carmichael, 2012). In some stroke patients, disturbed inhibitory function gives rise to epileptic seizures that may partially underlie functional deficits after stroke (Beaumanoir et al., 1996; Ono et al., 1997). In addition to its heightened excitability, peri-infarct cortex undergoes changes in gene expression that foster increased angiogenesis (Ohab et al., 2006) and axonal sprouting (Li et al., 2010).  1.3.2 Cortical plasticity  1.3.2.1 Cortical plasticity during learning  All cortical circuits are capable of changing to manifest the effects of recent and long- term experience (Dan and Poo, 2006; Karmarkar and Dan, 2006). Within the somatotopic motor map, more dexterous body parts occupy larger cortical territories (Penfield and Boldrey, 1937). These representations expand during development (Chakrabarty and Martin, 2000) and can become even larger with repetitive use of a particular body part (Kleim et al., 1998), suggesting that motor maps may represent the physical embodiment of acquired motor skill (Monfils et al., 2005). This view is supported by the observation that disrupting a motor map impairs the expression of the skilled movement that the map encodes (Nudo et al., 1996; Harrison et al., 2012) and consistent evidence that rehabilitative training can restore both the map and the skilled behavior after brain damage (Biernaskie and Corbett, 2001; Johansen- Berg et al., 2002a; Plautz et al., 2003). Modulation of human motor cortex can enhance both the acquisition and consolidation of motor skill (Galea and Celnik, 2009; Reis et al., 2009), but the increase in motor area evoked by training occurs at a delay, suggesting that motor map reorganization reflects the consolidation of a skill and not its acquisition (Kleim et al., 2004).  The plasticity that follows motor skill learning or somatosensory deafferentation is dependent on the basal forebrain cholinergic system (Juliano et al., 1991; Conner et al., 2003, 2010; Ramanathan et al., 2009) and on continued protein synthesis within cortex (Kleim et al., 2003a). It is also specific to the acquisition of skilled motor acts, and is not activated by unskilled motor acts such as running on a wheel (Kleim et al., 2002), forelimb strength  17 training (Remple et al., 2001), or repetitive performance of an unskilled reaching task (Plautz et al., 2000). Map plasticity can also be observed after classical conditioning involving somatosensory (Siucinska and Kossut, 1996), auditory (Weinberger, 2007), and visual maps (Dan and Poo, 2006). In addition to map reorganization, motor training can increase motor output, as demonstrated by studies of the evoked muscle potentials evoked by transcranial magnetic stimulation in skilled athletes (Pearce et al., 2000).  1.3.2.2 Circuit mechanisms of cortical plasticity  Alterations in map structure that result from motor training or that coincide with the recovery of motor skill after a brain injury reflect alterations in the underlying cortical circuitry. Cortical circuits can be classified vertical (columnar) or horizontal (intralaminar). The columnar organization of cortical microcircuits has been posited as an explanation for the local clustering of neuronal properties (Asanuma, 1975; Mountcastle, 1997; Amirikian and Georgopoulos, 2003). Motor cortex is not merely a mosaic of independent columns, however. Instead, it functions as an interconnected network bound by long-distance horizontal projections that can extend as far as 7-8 mm in the cat (Keller, 1993; Capaday, 2004). Both vertical and horizontal connections in cortical circuits can be remodeled by experience and contribute to the reorganization of cortical maps.  Synaptic plasticity within columnar circuits is perhaps best understood at the layer 4 – layer 2/3 synapse in rodent barrel cortex, which displays pronounced long-term depression in columns linked to deprived or trimmed whiskers and long-term potentiation in the barrel columns of spared whiskers (Feldman, 2009). In the motor cortex, columnar organization has been inferred from vertical recording electrode penetrations made in behaving primates (Amirikian and Georgopoulos, 2003). The organization of vertical circuits continues to be a vigorous and productive area of research and the descending projection from layer 2/3 to layer 5 is now known to be the strongest pathway in rodent somatosensory and motor cortex (Weiler et al., 2008; Hooks et al., 2011). However, most of the evidence for synaptic plasticity mediating motor map reorganization has come from horizontal connections.   18 In brain slices prepared from rats trained in skilled forelimb motor behavior, horizontal connections in layer 2/3 of forelimb motor cortex are potentiated while hindlimb motor cortex is unaffected. This synaptic plasticity is presumed to be LTP-like, since skill acquisition saturates further potentiation of these synapses and facilitates the expression of LTD (Rioult-Pedotti et al., 1998, 2000; Monfils and Teskey, 2004). Although horizontal connections are widespread in motor cortex, their pattern of connectivity demonstrates specificity, with a preference for connections between functionally related representations (e.g. wrist and digit) and antagonistic muscles (Capaday et al., 1998). It is not yet clear whether the synaptic plasticity induced by motor training is similarly specific. Epileptogenic kindling or repetitive stimulation of rat motor cortex can cause an expansion of motor maps similar to that seen after training, providing an additional link between synaptic plasticity and map reorganization (Nudo et al., 1990; Teskey et al., 2002). Finally, pharmacological manipulations of cortical circuits can cause rapid alterations to motor map structure (Jacobs and Donoghue, 1991; Harrison et al., 2012).  In addition to synaptic plasticity within existing circuits, structural alterations such as dendritic remodeling (Brown et al., 2007; Yang et al., 2009), and axonal sprouting (Li et al., 2010) contribute to map reorganization. Dendritic spines are small structures that mark post- synaptic sites in the dendritic arbor (Alvarez and Sabatini, 2007). Spine turnover is considered to be a structural manifestation of synaptic plasticity, and spine density of layer 2/3 and layer 5 pyramidal neurons has been shown to increase along with motor map area during skill training (Monfils and Teskey, 2004). Dendritic spines can be imaged in vivo in longitudinal experiments using two-photon imaging to more directly link skill acquisition and retention with physical changes in brain structure (Trachtenberg et al., 2002). In mice that successfully learn a forelimb reaching task, spine turnover increases and newly formed spines are preferentially retained after several weeks (Xu et al., 2009; Yang et al., 2009). Axons are generally considered to be more stable than dendrites, but they have not yet been investigated during motor learning to the same extent as dendrites (De Paola et al., 2006; Yu and Zuo, 2011).   19 1.3.2.3 Sensorimotor plasticity after stroke  Although the physical damage caused by stroke is irreversible, the brain has an impressive capacity for spontaneous functional recovery (Dancause et al., 2005; Eisner- Janowicz et al., 2008) that parallels the cortical plasticity manifested during normal learning (Kleim et al., 1998; Conner et al., 2010; Hosp and Luft, 2011). Plasticity in cortical circuits is manifested in large-scale reorganization of cortical maps after stroke, which has been observed both in experimental animals and humans (Liepert et al., 2000a; Fridman et al., 2004; Murphy and Corbett, 2009). These mechanisms allow the brain to adapt after a stroke, with surviving neurons assuming roles previously performed by damaged brain regions (Winship and Murphy, 2008; Ghosh et al., 2010). This phenomenon, termed vicarious function, appears to be necessary for recovery of function after lesions that completely destroy a cortical area (Xerri et al., 1998; Dancause, 2006). Large strokes commonly induce bihemispheric reorganization, with the spared hemisphere subserving functions previously performed by homotopic counterparts destroyed by stroke (Bütefisch et al., 2008; Brown et al., 2009). In some cases, this may lead to an imbalance of interhemispheric inhibition, with the unaffected hemisphere becoming disinhibited and consequently increasing inhibition of the affected hemisphere (Liepert et al., 2000c; Perez and Cohen, 2009). Plasticity of sensorimotor cortex after stroke is enhanced by rehabilitative training (Nudo et al., 1996; Krakauer et al., 2012) and dependent on neuromodulatory inputs (Ramanathan et al., 2009; Conner et al., 2010).  1.3.3 Experimental models of ischemic stroke   1.3.3.1 Rose bengal photothrombosis   Photothrombotic stroke induction involves the injection of the light-sensitive dye rose bengal either into the peritoneum or directly into the vasculature, followed by illumination of a cortical region of interest (Watson et al., 1985). Irradiation with green light triggers a photochemical reaction that initiates a platelet aggregation cascade, leading to occlusion of  20 targeted vessels. Photothrombosis has the considerable advantage of generating a reproducibly-sized infarct that can be targeted to any surgically accessible region of the brain. The size of the infarct is dictated by the area of illumination, which can be reduced occlude single arterioles (Sigler et al., 2008). Its disadvantages include the possibility that reactive oxygen spaces produced by the photochemical reaction can damage neurons independent of ischemic effects (Gajkowska et al., 1997) and the spontaneous reperfusion that can occasionally occur after dissolution of photothrombotically generated clots (Zhang et al., 2005).  1.3.3.2 Surgical occlusion of arteries   The most common occlusion-based model involves inserting a suture into the internal carotid artery and advancing it until the middle cerebral artery is blocked (Longa et al., 1989). Because middle cerebral artery occlusion (MCAO) is a common cause of stroke in humans, the MCAO model is considered to be more clinically relevant than models such as focal photothrombosis. A second advantage is that the suture can be removed to model reperfusion. Cerebral blood flow can be measured (e.g. with laser Doppler) to improve the consistency of this model, but its surgical difficulty and the variability of the lesions that it produces are inevitable disadvantages.  1.3.3.3 Endothelin-1 injection   The vasoconstrictive peptide endothelin-1 binds ET-A receptors on smooth muscle cells of the vasculature, causing 50% decreases in blood flow for as long as 16 hours (Fuxe et al., 1997). This approximates levels of blood flow found in the penumbra, making endothelin-1 injections a useful means of testing therapeutic agents intended to protect neurons in that zone (Biernaskie et al., 2001; Shih et al., 2005). Endothelin-1 injection shares many of the advantages and weaknesses of rose bengal photothrombosis, including its ability to produce infarcts of consistent size in a targeted region but also its limited clinical validity.   21 1.4 Optogenetic brain stimulation  1.4.1 Channelrhodopsin-2: a light activated cation channel   Channelrhodopsin-2 (ChR2) was discovered in the green alga Chlamydomonas reinhardtii. In this photosynthetic organism ChR2 contributes to light-detection required for phototaxis, particularly in low light conditions (Nagel et al., 2003). Channelrhodopsin-2 is both a nonselective cation channel (Figure 1.1) and an outward proton pump  (Feldbauer et al., 2009) with a peak excitation wavelength of ~465 nm and a unitary conductance of ~50 fS (Nagel et al., 2003). Each ChR2 protein can pump one proton per photocycle, but during its open state it also acts as a channel permeable to cations (Feldbauer et al., 2009).   Figure 1.1 The light sensitive cation channel ChR2 Photoisomerization of retinal by blue light opens a channel in the seven transmembrane protein that is permeable to cations, especially sodium. From (Wong et al., 2012), reprinted with permission from Elsevier.  Upon light stimulation, ChR2-mediated currents have an initial peak that decreases to a smaller steady current. This peak current is diminished when a second light stimlus follows a brief dark period, leading to the proposal that ChR2 has a branched photocycle with light- adapted and dark-adapted states (Figure 1.2) (Berndt et al., 2010; Stehfest and Hegemann, 2010). In cultured cells and Xenopus oocytes, ChR2 has a 100-fold selectivity for  22 monovalent cations over divalent cations such as Ca2+ (Berndt et al., 2010). The monovalent cation conductance of ChR2 is relatively non-specific during the initial peak phase, but becomes increasingly selective for protons in the steady state (Berndt et al., 2010).  Figure 1.2 Biophysical properties of ChR2 A Typical ChR2 photocurrent elicited by a square pulse of light demonstrates an early peak (IP) that relaxes to a steady state (I∞). B ChR2 can exist in dark-adapted (left) and light-adapted states. The peak current (IO) is attributed to the open conformation of the dark-adapted state (O1) exclusively. From Berndt et al., 2010, reprinted with permission from Elsevier.  Remarkably, if ChR2 is expressed in mammalian neurons it can faithfully transduce light stimulation into depolarization at frequencies as high as 30Hz, even without supplementation of its requisite cofactor all-trans retinal (Boyden et al., 2005). This means that neurons expressing ChR2 gain sensitivity to blue light, permitting millisecond timescale control of genetically-targeted neurons. Although the ability to impart light sensitivity for  23 optical control of excitable cells was a long-standing goal in neuroscience and many alternative methods have been developed (Kramer et al., 2005), ChR2 is the first such tool to become widely adopted by the neuroscience community.  1.4.2 Engineered variants of Channelrodopsin-2   Wild-type ChR2 has attracted considerable attention both from neuroscientists who have employed it in their research and from synthetic biologists who have attempted to enhance it through directed mutagenesis. Engineering efforts have focused on improving the photocurrent, light sensitivity, spectral properties, photocycle length, and intracellular trafficking of ChR2 (Lin, 2012). Many of these variants represent a compromise between improvements in one of these domains and sacrifices in others. Mutations at amino acid position 123 (ChETA variant) accelerate the kinetics of activation  at the expense of photocurrent (Gunaydin et al., 2010; Berndt et al., 2011). Conversely, mutations at position 134 yield variants of ChR2 with increased photocurrent but slower kinetics (Nagel et al., 2005). Mutations at position 170 (ChIEF variant) also accelerate kinetics, but appear to cause intracellular aggregation of the opsin protein when expressed in neurons (Lin et al., 2009). In some cases, delayed kinetics are desirable. By mutating position 128, the open state life time of ChR2 can be extended to last for minutes, creating a step function opsin that evokes prolonged spiking with a minimum of light stimulation (Diester et al., 2011).    24  Figure 1.3 Comparison of engineered variants of ChR2 Utility of ChR2 variants in pyramidal neurons in vivo can be evaluated based on their photocurrent and speed, defined as the inverse of the time from light onset to peak current plus τoff. Opsins in the green range have relatively large photocurrents and fast kinetics, traits that are considered optimal for most applications. From (Berndt et al., 2011), reprinted with permission from Elsevier.   Beyond manipulating single amino acids to enhance the photocurrent kinetics of activation of ChR2, efforts have also been made to enhance its function through the fusion of multiple genes to create hybrid proteins. For example, adding a PDZ-binding domain to ChR2 can cause it to cluster at post-synaptic densities (Gradinaru et al., 2007). Fusions of ChR2 with a myosin-binding domain can target ChR2 expression to the somatodendritic compartment of neurons (Lewis et al., 2009), whereas ChR2 can be localized to the axon initial segment by creating a fusion of ChR2 and the ankyrinG-binding loop of voltage-gated sodium channels (Grubb and Burrone, 2010). Finally, fusions of multiple opsins (e.g. ChR2 and the light-activated chloride channel Halorhodopsin) can be exploited to ensure expression of multiple proteins in equal ratios (Kleinlogel et al., 2011).  25  1.4.3 Other optogenetic tools   After the blue light activated cation channel ChR2, the opsins most commonly used in neuroscience are variants of the yellow light activated neuronal silencers Halorhodopsin (NpHR), an inward chloride pump (Zhang et al., 2007b) and Archaerhodopsin (Arch), an outward proton pump (Chow et al., 2010). Since these opsins have activation spectra that are largely separate from that of ChR2, they can be used together with ChR2 for combined excitation and inhibition within the same group of neurons (Zhang et al., 2007b). The development of red-shifted excitatory opsins has also been pursued intensely, since this would allow independent activation of two discrete populations of neurons (e.g. pyramidal neurons and GABAergic interneurons). One such opsin is C1V1, a fusion of the naturally occurring opsins ChR1 and VChR1 (Yizhar et al., 2011). Chimaeric proteins combining opsins with GPCRs  (e.g. adrenergic receptors) have also been created for optical control of intracellular signaling (Airan et al., 2009). Potassium channels that can be activated and inactivated by distinct wavelengths are capable of prolonged inhibition (Janovjak et al., 2010).  In addition to neuronal actuators, optogenetic sensors such as genetically encoded calcium indicators (Tian et al., 2009) and voltage sensors (Perron et al., 2012) are now becoming widely used. The rapid development and introduction of these tools makes it imperative that they be carefully tested and validated by researchers to ensure suitability for their specific experimental model and application.  1.4.4 Techniques for transgene delivery   Multiple strategies exist for the expression of ChR2, each with its own advantages and drawbacks. The experiments described in this dissertation all involved transgenic mice that express ChR2 under the Thy-1 promoter. Within the cortex, expression is predominantly in layer 5b pyramidal neurons, but there is additional expression in other regions, including CA1 of the hippocampus (Arenkiel et al., 2007). Additional transgenic mouse lines have been created that express ChR2 under cell-type specific promoters such as VGAT (GABAergic cells) or ChAT (cholinergic cells) (Zhao et al., 2011).   26  Viral vectors including lentivirus and adeno-associated virus can also be used to express ChR2 in mice, rats, and primates (Blits et al., 2010; Diester et al., 2011). Viral transduction can be targeted to brain regions of interest in wild-type mice, making it a flexible tool. It is also possible to use cre-recombinase for combinatorial expression. In this system, a cre-dependent construct is injected into a mouse in which a subset of neurons express the bacterial enzyme cre recombinase under a cell-type specific promoter. Only those neurons which both express cre-recombinase and are transduced with the viral vector will express the construct (Cetin et al., 2006). Finally, it is also possible to inject plasmid DNA into the ventricles of a developing embryo, then apply a pulsed current to drive the DNA into newly born neurons lining the ventricle (Saito and Nakatsuji, 2001). This technique is technically challenging and labor intensive, but is an effective means of targeting cortical laminae based upon the stage of development at which neurons in that lamina are born (Adesnik et al., 2010).  1.4.5 Strategies for light delivery   Opsins must be activated with light, and several strategies have been devised to deliver this light to the appropriate place. In order to deliver light to freely-moving animals or locations deep within the brain, optical fibers can be introduced into the brain and cemented at the skull (Aravanis et al., 2007). Fibers can be targeted to the same position in the brain as an injection of virus using an implanted cannula (Aravanis et al., 2007), but a single fiber is not suitable for targeting multiple spatially separated brain regions. To overcome this limitation, multi-fiber arrays have been designed that can be customized to target light to multiple targets at varying depths (Zorzos et al., 2012).   If all of the targets are located in cortex or other easily accessible areas, a simpler solution involves the use of a collimated beam of laser or LED light that can readily be scanned to multiple locations in two dimensions. This approach has been used for scanning or mapping both in slice (Petreanu et al., 2009) and in vivo (Ayling et al., 2009). If spatial targeting is not required, light delivery can be as simple as attaching an LED light source directly to an animals head to transmit light through a cranial window or a region of thinned  27 skull (Wyatt et al., 2012). Finally, two-photon excitation of ChR2 (Rickgauer and Tank, 2009) can be used to photoactivate specific subcellular compartments (Packer et al., 2012). 1.5 Research hypotheses and objectives  Hypotheses: 1) Light-based mapping techniques are ideally suited to the investigation of the functional organization of mouse sensorimotor cortex and its plasiticity after stroke. 2) Mouse motor cortex contains a topography of movement direction. 3) Motor cortex is capable of adopting an expanded sensory role to compensate for the destruction of somatosensory cortex by stroke.  Aim 1: Develop methodology for non-invasive mapping sensory and motor representations of the forelimb in mouse neocortex.  Aim 2: Apply these techniques to investigate the functional organization of mouse sensorimotor cortex, determine whether mouse cortex contains distinct representations of different categories of movement, and identify the circuit mechanism that differentiates these representations.  Aim 3: Establish the extent to which surviving cortical areas can adopt the functions of neighbouring regions destroyed by stroke, characterize differences in the response of sensory and motor representations to targeted stroke, and assess the consequences of vicarious function on sensory and motor representations.   28 Chapter 2: Automated light-based mapping of motor cortex by photoactivation of channelrhodopsin-2 transgenic mice  2.1 Introduction  The motor cortex was the first region of the brain to be mapped and to have an overt function attributed to it (Penfield and Boldrey, 1937). Motor mapping technologies have been refined in the intervening years and now include intracortical microstimulation (ICMS) (Donoghue and Sanes, 1987) and surface stimulation with electrode arrays (Hosp et al., 2008). The advent of transcranial magnetic stimulation has made noninvasive motor mapping feasible in humans (Siebner and Rothwell, 2003). Each of these techniques has unique advantages and limitations. Transcranial magnetic stimulation is noninvasive but has poor spatial resolution. Electrode-based brain stimulation methods have common disadvantages: the inability to selectively target neuronal subtypes, indiscriminate activation of axons of passage and damage when impalements are made.  Recently it has become possible to stimulate neurons using light, either by uncaging neurotransmitters (Shepherd et al., 2003; Luo et al., 2008) or by directly activating light- sensitive channels (Zhang et al., 2007a; Huber et al., 2008). Channelrhodopsin-2 (ChR2) is a light- activated nonselective cation channel isolated from the green alga Chlamydomonas reinhardtii (Nagel et al., 2005), which when expressed in neurons can transduce light energy into neural activity (Boyden et al., 2005). Here we present hardware and software for light- based mapping (LBM) of motor cortex output from anesthetized mice expressing ChR2 (Arenkiel et al., 2007). High-resolution motor maps are generated quickly, reliably and accurately in mice using a stage scanning system and fixed laser. We offer investigators a tool with greatly improved speed and precision to interrogate the motor cortex and address questions about sensorimotor processing both in the normal brain and after training, injury or disease.  29 2.2 Methods  2.2.1 Animals and surgery  Animal protocols were approved by the University of British Columbia Animal Care Committee. Channelrhodopsin-2 transgenic mice were purchased from the Jackson Labs (line 18, stock 007612, strain B6.Cg-Tg(Thy1-COP4/EYFP)18Gfng/J). Adult mice aged 2-3 months and weighing 25-30 g were used for these experiments, and were maintained on a 12:12 hour light:dark schedule. Anesthesia was induced with isoflurane (1.5 % in air) and body temperature was maintained at 37 C using a feedback-regulated heating pad. A craniectomy was made over the right sensory-motor cortex while the anesthetized mouse was supported by ear and tooth bars. The skull was then fastened to a stainless steel plate with cyanoacrylate glue and dental cement, and the plate attached to 25.4 mm posts mounted on an aluminum plate that could be bolted to a stage. The exposed brain was covered with 1-1.5 % agarose (Type 3-A Sigma; A9793) dissolved in a HEPES buffered (pH 7.3) physiological salt solution (in mM): 135 NaCl, 5.4 KCl, 1 MgCl2, 1.8 CaCl2, and 5 HEPES, and sealed with a custom cut glass coverslip. Isoflurane anesthesia was maintained during IOS imaging of somatosensory representations, but was replaced by ketamine/xylazine prior to LBM. Consistent with previous work (Ferezou et al., 2007; Hosp et al., 2008), we found it easier to elicit an evoked response in ketamine/xylazine anesthetized animals. Ketamine/xylazine was administered in doses of 0.02 mL (20 mg/mL ketamine, 2 mg/mL xylazine) approximately every 30 minutes or as necessary to maintain a constant level of anesthesia.   After craniectomy, the mouse was fixed to the scanning stage, and the locations of its somatosensory forelimb and hindlimb representations were visualized using IOS imaging (Winship and Murphy, 2008). During craniectomy surgery and IOS imaging the mouse was anesthetized with isoflurane (1.5% in air). Ketamine-xylazine (100 mg/kg ketamine, 10 mg/kg xylazine) anesthetic was used during motor mapping.  30  2.2.2 Optical stimulation parameters  We generally collected several cortical EEG-based maps at the beginning of each experiment using low laser power (40 mW/mm) and short activation duration (1 ms). In some cases when responses were weak (usually when craniectomies were imperfect), we increased laser power (up to 200 mW/mm) and/ or duration (up to 5 ms). We then connected the EMG electrodes and laser motion sensor, and began collecting motor maps. These EMG experiments were typically conducted using increased laser power (40–600 mW/mm) and duration (up to 35 ms), with stimulus parameters adjusted to suprathreshold levels.   There are several possible methods for photoactivation of ChR2 transgenic mice. Fiber optic systems are the best option for targeting subcortical structures, but are not ideal for stimulating the cortical surface because of light divergence. Divergence becomes a considerable problem in mapping, where curvature of the brain results in a variable distance between the light source and cortical surface. By using lens based beam-conditioning optics we were able to generate a relatively collimated beam that could be varied in size from 100 to 220 µm in diameter by changing lens focal lengths and/or lens positions. Within the optical cage, a 25.4 mm plano convex 50 mm focal length lens (LA1131, Thor Labs, Newton NJ USA) could be moved relative to the first video lens (typically placed 145 mm from the first video lens) to alter the laser spot size on the brain surface. The beam XY position within the video image field could be adjusted by moving both the plano convex lens within its mount using a Thorlabs XY translator lens mount (HPT1) as well as a right angle silver mirror mounted on a XY adjustable holder (Linos 065087, Goettingen, Germany) within a 30 mm Linos cube (061081). A final level of adjustment was achieved using a Linos XY adjustable holder (065087).  The XY stage used to move the animal relative to the laser was driven by XY LS50 high-velocity motors and controlled by an MS2000 2-axis stage controller (Applied Scientific  31 Instrumentation, Eugene OR USA). All maps were created based on a random sequence of movements to a series of positions outlined in a grid of stimulus locations superimposed over a map of the brain (see Software section below for details). We chose a stage scanning system since it was capable of repeated optical stimulation at intervals of < 1 s and ensured accurate XY positioning. An advantage of a mechanical scanning system is that all movements are based on absolute distance with respect to the excitation laser, and therefore it is inconceivable that photoactivation power or position would be subject to errors due to lens aberration that may occur near the edges of an image field. Although it would be possible to reduce the time between stimulation points by using a galvanometer and mirror based beam steering system, shorter (< 1 s) interstimulus intervals may lead to unexpected interactions between stimulus pulses. For both EEG and EMG maps, stimulus parameters (especially duration) were increased gradually until deflections in the recorded traces became apparent upon visual inspection. Once this threshold was reached, we would increase the stimulus duration by an additional 50 % to ensure adequate stimulation.  2.2.3 Optical imaging   To perform IOS imaging and to create maps of the surface vasculature a Dalsa 1M60 camera was used (Waterloo Ont. Canada). The frame grabber for the camera was an E1DB from EPIX (Buffalo Grove IL USA) and was running EPIX XCAP version 2.2 software. The camera was mounted on a vertical milling machine (Sherline Tool #5430, Miami FL USA), and images were taken through a macroscope composed of front-to-front video lenses coupled with a 52 mm threaded adaptor ring (BH Photo, New York NY USA). The top lens (closer to the camera) was a 135 mm F2.8 Nikor and the lower lens was a 50 mm F1.4 Nikor lens. To direct the 473nm photoactivation laser (CrystaLaser BCL-473-050, Reno NV USA) a hole was cut in the side of the Dalsa 1M60 camera F-mount adaptor and a dichroic mirror was installed between the CCD camera and the first video lens. The dichroic mirror was an Olympus DM500 (500 nm cut-off). To direct the blue laser light an optical cage was  32 constructed using Thorlabs 5 mm rods and microbench parts similar to that previously reported by us for photoactivation of rose bengal (Sigler et al., 2008).   Prior to each motor mapping session, we conducted IOS imaging to define the locations of the somatosensory forelimb and hindlimb representations. Following a protocol described previously (Winship and Murphy, 2008), we used piezoceramic bending actuators (Piezo Systems Q220-AY-203YB, Cambridge MA USA) to deliver 1 s trains of 100 Hz vibrations to the forelimb and hindlimb alternately. 15 baseline images were compared to 15 images captured over a 1.5 s period following stimulation, and a custom-written ImageJ (NIH, Bethesda MD USA) plugin was used to calculate the percentage change in reflectance of 635 nm light. A 50 % threshold was then applied and the resulting maps color-coded.  2.2.4 Testing for the effect of photodamage on motor maps   To test for photodamage, we compared processed forelimb EMG responses from two animals evoked by stimulation within a region of interest (a square of 36 pixels, each pixel 300 µm2). We compared EMG responses from stimulus parameters for trials at the beginning and the end of an experiment. In one animal, there was no significant difference in evoked EMG amplitude after 143 intervening stimulus trials (p = 0.1220, paired t-test). The other animal showed an increase in EMG amplitude after 103 trials of stimulation (p = 0.0004, paired t-test), which is explainable by a decrease in anesthetic depth or stimulation-induced plasticity (Figure 2.10). The laser powers used (40-600 mW/mm2, 1-35 ms) were within the limits of the maximum permissible exposure to the human cornea (for 1 W/mm2, the maximum permissible exposure time specified by IEC 60825 standards is 1 ms for lasers with wavelengths 400-700 nm).     33 2.2.5 Intracortical microstimulation  ICMS was performed using a glass pipette (2-3 MOhm, made on a Narashige P83 vertical electrode puller) containing a 0.25 mm bare silver wire and filled with 3 M sodium chloride, with fast green (Sigma) added in order to facilitate visualization under the microscope. Five to ten 125 ms trains of stimulation, each with five 240 µs stimuli at maximum intensities of 200 µA, were delivered at a frequency of 40 Hz to a depth of 400-500 µm to target layer 5 motor neurons. Impalement sites were guided by somatosensory IOS maps, and spacing between sites was approximately 500 µm. EMG latencies for ICMS and LBM were calculated by measuring from stimulus onset to the point where a pre-defined threshold was exceeded (three times the standard deviation of the baseline noise).  2.2.6 Motor output recordings   Hindlimb EMGs were recorded from the biceps femoris and the vastus lateralis muscles using electrodes similar to those described by others (Pearson et al., 2005). These were constructed by twisting together a pair of 0.125 mm teflon coated silver wires and stripping the insulation from two non-overlapping contacts. The twisted bipolar electrodes were then inserted into the muscle using a 22.5-gauge needle. The forelimb EMG recordings were made from the triceps brachii and the extensor carpi radialis brevis muscles using single 0.125 mm teflon wires bared 2 mm from the end and inserted with a 26-gauge needle. The insulated tips of the wires were then bent over to secure them in place. A common ground for the two forelimb electrodes was inserted into a small incision in the footpad. The larger twisted bipolar electrodes were used exclusively in the hindlimb because of the small size of the forelimb muscles. Forelimb movements were also quantified using a laser motion sensor (LK-2000, Keyence, Osaka Japan).   34 2.2.7 Electroencephalography  To examine stimulation-evoked EEG responses at the cortical surface, we used a razor blade to bare ~1 mm from the tip of 125 µm diameter silver wires. The electrodes were inserted into the agarose near each of the four corners of the craniectomy. We then mapped the entire surface of the exposed cortex, performing 3-5 repetitions per map. The averaged maps recorded by each electrode were then normalized to each other and a mean map incorporating the information from all electrodes was created. The duration of individual EEG depolarizations was measured from stimulus onset to the time point where the trace returned to 85 % of the pre-stimulation baseline. EEGs and EMGs were sampled at 5 kHz. An unpaired t-test was used to compare stimulus duration vs. EEG depolarization duration.  2.2.8 Effects of glutamate receptor antagonists on EMG maps   Although light activation may be targeted to a selected region, adjacent areas of cortex could be activated through intra-cortical synaptic interactions. We have addressed this possibility by applying AMPA/NMDA glutamate receptor antagonists to the surface of the cortex at high concentrations that we have previously been shown to completely block sensory stimulation induced intrinsic optical signal maps (supplementary Fig. 4). These maneuvers would be expected to block intracortical synaptic transmission and potentially the spread of excitation. Despite using these antagonists, we found only modest change in the size area or amplitude of the light activated maps within the first 30-60 min. These experiments were performed as follows. After obtaining a set of baseline maps as described above, we applied CNQX (4.5 mM) and MK801 (300 µM, both in physiological saline solution) to the open craniectomy (with dura intact) and allowed the drugs to incubate for 30 minutes. Motor mapping resumed following this period and lasted for up to two hours after the incubation period. The drugs were reapplied (at the same concentrations) to the cortical surface at intervals of approximately 30 min. To compare EMG amplitudes before and after  35 application of the drugs, we calculated mean amplitudes for a region of 12 pixels (3 x 4) at the center of the motor map (as defined by a two-dimensional Gaussian fit). Paired t-tests were used to compare EMG and EEG amplitude at time points after drug application (30-60, 61- 90 min) to pre- drug EMG and EEG amplitudes. Glutamate antagonists failed to have a significant effect on motor map amplitude (P = 0.1393, n = 14 maps from four mice, paired t- test), or the cortical EEG (P = 0.0595, n = 10, paired t-test) elicited by light stimulation within 30-60 min of application in 5 of 6 animals examined (Figure 2.8). At later time points (90 min) map amplitude was depressed (P < 0.0001, n = 12 maps from four mice, paired t- test) (possible due to more distant drug action), but map boundaries were in part retained. A total of 5 animals were studied for comparison of EEG and EMG sensitivity to antagonists; one animal was not included since the EMG amplitude decreased by over 80 % within the first 30 min of antagonist application and the EEG was of poor quality compared to the results observed in the other 4 animals.   These observations suggest that light-based motor maps are not necessarily dependent on intracortical synaptic activity. At longer time points (90 min after MK801 and CNQX addition) we did observe an 80 % depression of map amplitude. We are currently exploring why the drugs had a delayed effect on amplitude, but presumably this reflects a more distant site of action, perhaps near the layer 5 somata or within the striatum or spinal cord. Nonetheless, despite pharmacological diminution of the maps by a factor of 3 in the amplitude, their general boundaries still apparent (Figure 2.8). This result is consistent with light-based mapping directly activating layer 5 output neurons leading to muscle potentials rather than a model where light-based pulses activate intracortical synaptic transmission, which does not necessarily reflect direct connections from motor cortex.  2.2.9 Characterization of photoactivation area   We used IOS imaging using 630 nm illumination as described above to assess the spread  36 of laser excitation following photostimulation. We compared 15 baseline images to 2 images collected 200-400 ms after photostimulation (100ms burst of 5 ms pulses delivered at 100 Hz with laser power between 156 and 469 mW/mm2. We also used an intracortical microelectrode (see ICMS section above) to deliver a 100 ms 100 Hz train of 200 µA of stimulation pulses using the parameters employed for ICMS. To obtain reliable IOS activation using the ICMS electrode we needed to increase the pulse duration to 5 ms. In analysis of 2 animals we found that varying the ICMS pulse duration from 0.5-5 ms did not strongly affect the width of the IOS activation (r2 = 0.325). The photoactivation profile was estimated from average images of changes in light reflectance and measuring the full width at half maximal amplitude of the response as well as a contour plot analysis of averaged group data (Figure 2.9)  2.2.10 Software  Custom software in Igor Pro (Wavemetrics) running on a standard PC controlled the scanning stage using serial commands, while a National Instruments board (PCI- 6036E) triggered the 473 nm laser with a TTL pulse and acquired analog outputs (5 kHz) from EEG, EMG, and the laser motion sensor. The software package includes a graphical user interface that allows the experimenter to modify all parameters of interest (e.g. stimulus duration, number of repetitions, inter-stimulus delay, channels recorded, sampling rate etc.). Within each repetition, stimuli were always delivered in a randomized fashion. Randomization was achieved by sorting the list of desired stimulation points by a list of random numbers generated by the Igor Pro random number function. At the beginning of each experiment, the number and location of stimulation points were defined with reference to an image of the exposed brain.    37 2.2.11 EMG and pixel based motor map analysis   EMG records were sampled at 5kHz, and band-pass filtered (0.5-500 Hz), full-wave rectified, the mean of the pre-stimulation baseline subtracted, and integrated to give the array of values displayed in pixel-based maps. In order to quantify the size of motor maps and locate their centers, we fitted a two-dimensional Gaussian curve to the pixel- based maps. Motor and sensory map areas were estimated from contour lines of Gaussian fits at 50 % of peak value. The mean areas of the Gaussian-fit cortical representations of the four muscles studied were then determined (n=9 animals), and a one-way ANOVA was performed (P = 0.0006) followed by the Tukey-Kramer multiple comparisons test. The baseline offset of each Gaussian fit (z0) was defined as the mean of the background noise. This value was obtained from trials in which stimulation was targeted over thick bone, where no response should be evoked. In approximately 5 % of cases, the maps were of relatively poor quality and could not be fit by the Gaussian function. Maps were excluded from further analysis if any of the following three empirically established criteria were not met: peak amplitude of map is more than five times greater than the standard deviation of the baseline noise; map width is at least 300 µm (typical size of one pixel) in the X and Y dimensions; and calculated map center must be within the area selected for photoactivation and imaging. Poor map quality could generally be attributed to imperfect craniectomies or anesthesia. The X and Y values of the map centers and widths were then averaged across several maps (3-6 per animal), and the means of these values were compared statistically. Before performing one- way ANOVA on the map area values of the different muscles, we tested the variation of the standard deviations and found that it was not significant (P = 0.131, Bartlett test). Because we did not record from all four muscles in some animals and because some maps were excluded, not all comparisons were made using the same number of animals. To generate maps based on laser motion sensor data, a two-sided Gaussian fit was applied to each trace and the peak displacement was plotted for each point of stimulation.   38 2.2.12 Histology   Brains were fixed for histology by transcardial perfusion with 4% paraformaldehyde and, coronal slices 100 µm thick were sectioned by vibratome and examined under epifluorescence as described by (Brown et al., 2007). Confocal microscopy image analysis was performed with 16 bit, 20 and 40x magnification at 1.6 and 3.2 µm per pixel.  2.3 Results  2.3.1 Automated mapping of motor cortex using laser light  For automated ChR2-based motor mapping we chose a relatively collimated 473 nm laser targeted through a simple microscope (Fig. 2.1a). To check the beam profile as it passes through brain tissue, we directed the beam into the cortical surface of a fixed mouse brain section (Fig. 2.1b). The beam width (measured using a monochrome camera) was 170 ± 3.7 µm at the cortical surface and 640 ± 220 µm at a depth of 250 µm (n = 7 measurements; all values are reported as mean ± s.d.; Fig. 2.1c). Examination of light intensity at different depths indicated that beam width decreased exponentially with a decay constant of ~450 µm.  For LBM, we moved the mouse relative to the laser using a fast scanning stage (13 mm/s) (Callaway and Katz, 1993). We moved the stage in random order to each of the predefined stimulation locations superimposed on the cortical map (Fig. 4.1a), and delivered a flash of laser light to each point while collecting an electromyogram (EMG) and a cortical electroencephalogram (EEG). We selected the intensity and duration of photostimulation based on their ability to elicit a suprathreshold EMG response.  39  Figure 2.1 Automated Light-based mapping (LBM) of the mouse motor cortex  (a) Experimental setup. Anesthetised mice were placed on a scanning stage and an array of cortical points (inset) was stimulated by a 473 nm collimated laser beam directed through a video microscope objective. Motor output was detected by EMG electrodes in forelimb and hindlimb muscles, and by a laser motion sensor fixed to the stage. (b) Photograph of a stimulation laser targeted at a coronal slice of fixed brain tissue embedded in carboxyfluorescein-containing agarose. (c) Intensity profile of the illuminated area as the beam passes through fluorescent agarose above the surface of the brain and 250 µm under the cortical surface (peaks were normalised for comparison). Images used for analysis were acquired using a high-resolution monochromatic camera. Scale bars, 1mm (a), 2mm (b), and 400 µm (c).   2.3.2 Photostimulation elicited homogeneous cortical excitation  After verifying that the stage scanning laser system was accurate in positioning, we tested its ability to evoke local excitation of the cortex by performing a craniectomy and then placing surface EEG electrodes made of silver wire in the four corners of the craniectomy. We mapped EEG responses over areas of up to 20 mm2 divided into activation sites of ~0.09 mm2 (300 µm spacing) and found that photostimulation excited all regions of the exposed  40 cortex (Fig. 2.2a–c). Homogeneity of cortical excitation ensured that differences in motor maps reflect local motor output circuitry and not the distribution of ChR2 responsiveness.   Figure 2.2 ChR2-mediated EEG responses can be elicited from all regions of the exposed cortex (a) Mean EEG responses evoked when the laser stimulated that cortical location from all four electrodes at the cortical surface. EEG amplitudes were normalised to the maximum value (within an electrode), and then the mean values from all four electrodes were averaged. Lighter colours signify a larger response. The linear scale was set to emphasise variations in cortical response. At points of stimulation where the cortical surface was obstructed by blood vessels or bone (coloured red and green respectively), responses were diminished or absent. Scale bar, 1mm. (b) Raw EEG traces from a single electrode. (c) traces (boxed in b) showing a representative EEG response evoked by stimulation over bone (top) and of exposed cortex (bottom). Optical stimulation began at the point marked by the asterisk. (d) The relative time courses of ChR2-evoked EEG and EMG responses are shown after a single 5 ms pulse of 160 mW/mm2 laser light (blue bar). Note the prolonged EEG depolarisation relative to stimulus duration.   In evaluating EEG recordings, we found that photostimulation as short as 1–5 ms evoked a response. These brief light flashes produced cortical depolarizations that were significantly longer than the stimulus duration (31.4 ± 5.4 ms, P < 0.0001, n = 15 trials in four mice, unpaired t-test; Fig. 2.2d). We also found that targeting the laser at the exposed EEG electrode caused a large photoelectric artifact that was different in kinetics from the results of cortical tissue excitation and was restricted to periods when the laser was activated. As expected, wild-type mice lacking ChR2 showed no response to photostimulation (n = 6) but did show the photoelectric artifact (Fig. 2.3).  41  Figure 2.3 ChR2-negative animals show no response to photostimulation (a) Stimulation was delivered to an array of points (red crosses), and cortical activity was recorded by an EEG electrode (at right). (b) Each pixel represents the response evoked when the laser stimulated that cortical location, with lighter colors signifying a larger response. Scale bars in a and b 1 mm. (c) Raw EEG traces. Scale bars 3 mV, 200 ms. (d) selected traces enlarged from c. Note the large stimulation artifact produced when the laser strikes the recording electrode (bottom trace), which has amplitude and time kinetics dissimilar to genuine EEG responses. Scale bars 1 mV, 200 ms.   To confirm the expression of ChR2-YFP protein reported by the developers of the mouse (Arenkiel et al., 2007) and the distributor (Jackson Labs), we performed a histological examination of ChR2-YFP fluorescence in a subset of mice (n = 3; Fig. 2.4). We corroborated the homogeneous distribution of ChR2-YFP fusion protein throughout the sensorimotor cortex and its restriction to tufted layer-5 neurons as originally reported, and this was consistent with other Thy1 promoter–driven mouse lines (Feng et al., 2000; Arenkiel et al., 2007). In two mice examined by confocal microscopy we saw no labeling of neuronal cell bodies in more superficial layers (Figure 2.4).  42  Figure 2.4 ChR2 is expressed throughout the sensory-motor cortex in layer 5 cells Expression of YFP (a,b) and ChR2-YFP fusion protein (c-f) in fixed coronal sections of mouse cortex. Expression was under control of the Thy1 promoter in both cases. (a) Low magnification wide-field fluorescence micrograph of a YFP-H mouse cortex. Medial is to the right and dorsal is to the top. Borders between primary motor cortex (m1), secondary motor cortex (m2), and primary somatosensory cortex (s1) are marked with arrows. Scale bar equals 500 µm, and also applies to c. (b) Higher magnification view of the border between m1 and m2. Apical dendrites of YFP-expressing neurons in layer 5 can be seen ascending to layer 1. Scale bar 250 µm. (c) Low power wide-field fluorescence micrograph from a mouse expressing the ChR2-YFP fusion protein. The areal and laminar expression pattern is similar to that shown in a for YFP expression in YFP-H line mice. (d,e) Maximum intensity projections over 20 µm from coronal slices of 2 mice expressing ChR2- YFP. Apical dendrites of layer 5 pyramidal neurons expressing ChR2-YFP extend into layer 1. No pyramidal neurons expressing ChR2-YFP are seen in layer 2/3. Arrow in (d) shows an axon entering the white matter (wm). Scale bar in e equals 200 µm and applies to f. (f) Higher magnification maximum intensity projection of 3 µm through layer 5 from the same coronal slice illustrated in d. Examples of individual neurons expressing ChR2-YFP are indicated with arrows. Scale bar 100 µm. (g-j) Coronal sections (100 µm thick), anterior to posterior, of mice expressing ChR2-YFP, scale bar 1mm.    43 2.3.3 Mapping light-evoked muscle potentials in ChR2 mice  By implanting silver EMG electrodes in the triceps brachii (extensor) and extensor carpi radialis brevis muscles of the forelimb, and the biceps femoris (flexor) and vastus lateralis (extensor) of the hindlimb, we established the parameters of LBM necessary to evoke contralateral EMG responses. We assessed the effect of light intensity (40–600 mW/mm2) and stimulus duration (1–35 ms), and found that these ranges of intensity and duration were sufficient to produce a motor response (Fig. 2.2d). Photoactivation of areas 170 µm in diameter reliably evoked a motor cortex EEG response and a delayed EMG response in contralateral forelimb and contralateral hindlimb muscles. We did not study smaller photo- activation areas because the arbors of layer-5 neurons are at least 300 µm across, and we would not expect any increase in detail with reduced photoactivation areas.  We assigned processed EMG responses a grayscale value on a linear scale from black (zero) to white (maximum response) to form a pixel-based map, typically created with grids of stimulation points using 300 µm spacing (Fig. 2.5a–d and Supplementary Methods online). Given some scattering of blue light by tissue14, this spatial frequency should efficiently excite the cortex between each of the points and is consistent with photoactivation areas used in previous brain-slice and in vivo work (Aravanis et al., 2007; Arenkiel et al., 2007).  44  Figure 2.5 High-resolution optically stimulated motor maps (a, b) Forelimb triceps brachii (a) and hindlimb biceps femoris (b) motor maps created with 320 µm spacing between laser stimulation points (single 15 ms pulses at 160 mW mm-2 ). Each map is the average of three repetitions. Absolute grayscale values are not equivalent for a and b. M, medial; L, lateral; R, rostral; and C, caudal. (c,d) One repetition of raw EMG traces for forelimb (c) and hindlimb (d), with individual traces arranged according to the cortical locations from which they were evoked by photostimulation. (e,f) Boxes in c and d identify expanded forelimb (e) and hindlimb (f) EMG traces with an asterisk indicated onset of the laser stimulation. Responses to optical stimulation of points outside the motor maps (top traces) and inside the motor maps (bottom traces) are shown. Scale bars, 1mm (a,b), 200 ms (c,d) and 20 ms (e,f).   Photostimulation in the center of motor maps yielded muscle excitation after a delay from the photostimulation onset of 10.8 ± 1.0 ms for contralateral forelimb and 19.4 ± 1.0 ms for contralateral hindlimb EMG (n = 4 mice). Analysis of the relationship between cortical EEG depolarization and evoked EMG signals (Fig. 2.2d) revealed the latency between cortical excitation and muscle excitation. As expected, optically evoked EMG responses exhibited latencies comparable to those of EMG responses produced by direct electrode- based stimulation of motor cortex in mice and other animals (Fig. 2.5e,f) (Rho et al., 1999). In ICMS experiments, the latency of ICMS-evoked EMG responses was 11.1 ± 1.1 ms for contralateral forelimb and 19.5 ± 0.9 ms for contralateral hindlimb (n = 4 mice), consistent  45 with values from photostimulation experiments. Cortical regions from which LBM evoked larger EMG responses tended to also produce responses with shorter latencies (Fig. 2.6).   Figure 2.6 Response latency is inversely related to EMG amplitude (a) High resolution forelimb motor map. White pixels are maximum EMG response, black is no response. Scale bar 1 mm. (b) Corresponding EMG latency map. Pixel values represent latency of EMG response from stimulus onset. Black pixels represent latencies greater than 40 ms or the absence of any response, white pixels represent latencies of less than 10 ms. Scale bar 1 mm.  In a mouse on which we performed both ICMS and LBM (Fig. 2.7), the positions and sizes of motor maps were generally in agreement. In this combined ICMS and LBM experiment we performed 26 penetrations to map the motor cortex, completing the ICMS map in approximately 1 h. In the same amount of time, we could map more than 3,000 points using LBM.  46  Figure 2.7 ICMS and LBM motor maps obtained from the same mouse (a) Points of electrode-based ICMS trains are displayed in blue (forelimb movement) and white (no forelimb movement). Purple contour lines represent the ChR2- derived LBM forelimb motor map created with single 20 ms, 160 mW mm-2 laser pulses (90% and 50% of peak response). IOS sensory maps are displayed in yellow for sensory forelimb sFL and red for sensory hindlimb sHL. Scale bar 1mm. (b,c) Raw EMG (top), full wave rectified response (bottom, solid line) and integrated response (bottom, dashed line) for the ICMS point of stimulation in a by the square (b) and the oval (c). Electrode symbol indicated stimulus onset.  Given that layer-5 neurons make corticospinal projections, it is likely that LBM does not require intracortical excitatory synaptic activity to stimulate muscles. Application of a- amino-3- hydroxyl-5-methyl-4-isoxazole-propionate (AMPA) and N-methyl-D-aspartatic acid (NMDA)-type glutamate receptor antagonists (Fig. 2.8 online) directly to the sensorimotor cortex at concentrations and durations previously shown to block sensory signals (Murphy et al., 2008) suggested that LBM activates corticofugal projections directly and not antagonist-sensitive circuitous intracortical routes of motor activation.  47  Figure 2.8 Cortical application of glutamate receptor antagonists have little initial effect on light-evoked EMG and EEG activity (a) Experimental timeline for antagonist experiments (MK-801 0.3mM and CNQX 4.5mM, applied directly to the intact cortical surface, dura intact). (b) Forelimb motor map before antagonist application. Scale bar, 1 mm. (c) Forelimb motor map 50 min after initial antagonist application. (d) Forelimb motor map 75 min after initial antagonist application. Motor map amplitude, indicated by the gray scale with scale bar expressed in mV.s on a linear scale. (e) EMG (black bars) and EEG (red bars) amplitudes normalised to pre-antagonist values (error bars SEM, n= 4 animals). Group data indicates that cortical EEG responses and light-evoked muscle potentials are relatively resistant to blockade of excitatory transmission in the cortex consistent with EMG maps reflecting direct activation of cortical spinal neurons and not indirect intracortical circuits. (f) Motor map in d with scale increased 2.5 times to highlight area of map rather than EMG amplitude.   To estimate the area of cortex activated by light pulses, we examined intrinsic optical signals (IOS) in response to 100 ms trains of light pulses and found them to spread over 1,012 ± 316 µm (n = 4 mice, measured at full width at half-maximal amplitude; Fig 2.9) consistent with the extent of light scattering observed at a depth of 250 µm (Fig. 2.1c). In comparison, we observed ICMS electrode activation widths of 690 ± 102 µm (n = 3 mice),  48 indicating that LBM activates an area only moderately larger than ICMS (~2.2-fold larger). The IOS response area within a contour plot drawn at 75% of the peak laser activation was considerably smaller (0.22 mm2 or about 0.5 µm in diameter; Fig. 2.9). These measurements suggest relative differences between ICMS and LBM activation areas; however, the use of IOS activation area to determine exactly what fraction of output neurons are activated with a single light pulse may be complicated by potential non-linearity associated with IOS measurements and by uncertainty of the relevant firing thresholds.  Regarding phototoxicity, we observed no consistent decrease in the amplitudes of evoked EEGs or EMGs during an experiment and no gross histological evidence of damage. In two mice involved in particularly long experiments, EMG amplitude showed no measurable reduction after 4100 stimulus repetitions made over the same areas (Figure 2.10). By making a sealed chronic cranial window and using a noninvasive laser–based measurement device (Ferezou et al., 2007) we found in two preliminary experiments that similar forelimb movement maps could be evoked in sessions 7–10 d apart, indicating that LBM does not lead to slowly developing toxicity (Fig. 2.11). The laser motion sensor was more sensitive to paw movements than visual assessment and provided data on an absolute scale that agreed with EMG-based maps (Fig. 2.12).  49  Figure 2.9 Estimates of ChR2 and electrode based cortical activation spread using IOS imaging (a) Image of brain surface with the location of blue laser light stimulation marked by a blue dot. A 100 ms train of 10, 5 ms laser pulses given at 100 Hz was used for optical stimulation. Intracortical microstimulation was performed in approximately the same area using a glass-stimulating electrode (see Supplementary Methods). (b) Image showing change in reflected light signal 200 ms after the onset of a train of blue light pulses. A small reduction in reflected light is observed consistent with local brain activation. The scale for panel B is between -0.03 to +0.02 %; data is the average of 140 trials. (c) Change in reflected light signal in response to ICMS train stimulation, the average of 60 trials is shown. (d) Plot of change in IOS reflectance measured using a horizontal rectangle 180 µm in height placed across the centre of activation for both channelrhodopsin activation and ICMS. The data plotted is from panels b and c. No light activated changes in brain reflectance were observed in 2 wild type animals examined, or in animals killed by anesthetic overdose. (e) Average laser light- induced IOS response from normalised data (each animal scaled from - 1.0 to 0) from n = 4 animals using the parameters described above. Contour lines indicate 50, 75, and 90 % of the peak response in this panel and f. (f) Average (ICMS) stimulating electrode induced IOS response from normalised data (each animal scaled from -1.0 to 0) from n = 3 animals using the parameters described above.   50    Figure 2.10 Focal and repeated photostimulation of motor cortex does not cause degradation of motor map (a) Image of cortex with region of focal and repeated stimulation displayed in red box. 103 repetitions of a 6 × 6 pattern of photostimulation (300 µm spacing between points) was delivered to the same region of cortex prior to motor mapping (up to 294 mW/mm2 and 50 ms). (b) Forelimb motor map created after focal and repeated photostimulation. There is a robust response in area corresponding to the position of focal stimulation after 100 repetitions (red box). Scale bars, 1 mm. (c) Plot of EMG amplitude over multiple repetitions of photostimulation. EMG amplitudes are taken as the average of multiple repetitions from the area corresponding to the red box in a and b (error bars are SEM). (d-f) Raw EMG traces corresponding to red box in a and b taken from the 1st, 50th, and 90th repetition of focal photostimulation. Scale bar 25 ms, 2 mV.  51  Figure 2.11 Motor maps can be evoked weeks apart within the same animals Preliminary maps of forelimb movements evoked by ChR2 activation within sensory-motor cortex from two separate animals implanted with chronic cranial windows are shown (measurements made with a laser-based motion sensor). The animal in (a) has been mapped two times (b,c), with at least one week between mapping sessions.  The animal shown in panel (d) has been mapped three times, the second (e) and third (f) maps are shown here. The map centers (defined by 2-DGaussian fit) are marked on each map and at the corresponding cortical location (in the top panel). The boundaries of the cranial windows are outlined in red. For all maps, black pixels represent cortical locations from which no movements were evoked, and white pixels represent the location of maximal response. Maps b and c are on the same scale, with black representing no evoked movement and white representing a limb displacement of 2.2 mm. Map e is similarly scaled from 0- 4.6 mm, and f is scaled from 0- 0.43 mm. Scale bar 1 mm. Note that limb displacement is strongly dependent on anesthetic state and is highly variable.  52  Figure 2.12 Stimulation-evoked movements detected by EMG and laser motion sensor (a) EMG-based motor map of the extensor carpi radialis muscle. Pixel values signify EMG amplitude, with white corresponding to peak response and black to the absence of any response. (b) Forelimb movement map from the same animal obtained using a laser motion sensor. Pixel values correspond to the displacement of the forelimb, with values ranging from 15 mm (white pixels) to no movement. Scale bars 1 mm. (c) EMG trace corresponding to pixel circled in a. Scale bars 100 mV, 100 ms. (d) Laser motion sensor trace corresponding to pixel circled in b. Scale bar 0.5 mm, 100 ms.  2.3.4 Fine motor map structure  Repeated LBM maps from the same mouse indicated that spatial heterogeneity in EMG amplitude was not due to noise or poor sampling but reflected the underlying properties of the motor representations (Fig. 2.12). To examine limb representations on a finer scale, we compared the size and center positions of two different muscles within a motor map of a single limb by one-way ANOVA (P = 0.0007) and Tukey post-hoc tests. Contralateral forelimb extensor muscle maps were similar in size when thresholded at 50%  53 of maximal amplitude: carpi radialis brevis and triceps brachii muscle maps were 1.65 ± 0.61 mm2 (n = 9 mice) and 1.60 ± 0.67 mm2 (n = 7), respectively (P > 0.05). The same was true for the hindlimb biceps femoris flexor and vastus lateralis extensor maps (0.71 ± 0.30 mm2 (n = 5) and 0.61 ± 0.28 mm2 (n = 7), respectively; P > 0.05). Both contralateral forelimb maps were significantly larger than either of the contralateral hindlimb maps (P < 0.05), which is consistent with epidural array-based mapping studies in the rat (Hosp et al., 2008). Similar to map area, differences in motor map position were significant only when comparisons were made between contralateral forelimb and contralateral hindlimb, and not between muscles within the same limb. Forelimb muscle representations had a mean center point that was separated from the center of the combined hindlimb map by an average distance of 0.46 ± 0.25 µm (P = 0.0005, n = 9 mice, one-sample two-tailed t-test). Our map analysis suggests that muscles working together to control a body part were represented in very similar regions of motor cortex, whereas muscles in different appendages overlapped less. In eight mice we defined map coordinates with reference to bregma (Table 2.1).  We examined the spatial relationships between sensory and motor representations of contralateral forelimb and contralateral hindlimb (n = 3 mice; Fig. 2.14). Approximately 50% of contralateral forelimb and contralateral hindlimb motor maps overlapped with sensory cortex (Table 2.2). Although there is some uncertainty about the motor map edge position (to within 500 um), the motor and sensory map center positions should be more precise. The distances between the centers of the forelimb motor and sensory maps were 1,217 ± 669 µm, and centers of hindlimb motor and sensory maps were 540 ± 454 µm apart.  54  Figure 2.13 Motor maps are stable and repeatable (a, b) Four consecutive replicates (numbered) of forelimb (a) and hindlimb (b) EMGs in response to laser stimulation using 10 ms, 160 mW mm-2 blue-light pulses. The resulting four motor maps were generated in ~100 s per repetition. In each array, individual EMG traces (200 ms long) are plotted according to the stimulation position from which they were evoked. These spatial relationships are preserved in the pixel-based maps of response amplitude. Stimulation was performed with 300 µm spacing between points, and each pixel represents a cortical area of 0.09 mm2. Scale bars, 500 ms (top) and 1 mm (bottom).   55  Figure 2.14 Motor and sensory cortical limb representations Sensory forelimb (sFL) and sensory hindlimb (sHL) representations were visualised using IOS imaging (thresholded at 50% of maximal response). Contour lines at 50% of peak response are shown for the extensor carpi radialis forelimb muscle (mFL) and the hindlimb biceps femoris (mHL) and vastus lateralis (dark blue) motor maps derived from single 5 ms, 330 mW mm-2 laser pulses. Scale bar, 1 mm.  Muscle Lateral from Bregma (mm) Posterior from Bregma (mm) Extensor carpi radialis brevis (FL extensor) 1.84± 0.37 0.42 ± 0.97 Triceps brachii (FL extensor) 1.96 ±0.29 0.15 ± 0.66 Biceps femoris (HL flexor) 1.47 ± 0.32 0.55 ± 1.04 Vastus lateralis (HL extensor) 1.60 ± 0.25 -0.33 ± 0.97  Table 2.1 Motor map coordinates Coordinates of the center point of cortical motor representations relative to bregma, as defined by two-dimensional Gaussian fitting. n= 8 animals, with 3-6 maps per animal. Note, both HL and FL muscles were not always assayed within the same animals and inter-animal variability can account for some variability in map centers.    56 Muscle %  Motor map in sensory territory. % Motor map in FL sensory territory % Motor map in HL sensory territory Extensor carpi radialis brevis (FL extensor) 51 ±15 24 ±8 27 ±8 Triceps brachii (FL extensor) 51 ±12 22 ±10 29 ±4 Biceps femoris (HL flexor) 55 ±14 0 ±0 55 ±14 Vastus lateralis (HL extensor) 50 ±12 3 ±5 47 ±11  Table 2.2 Motor and sensory map overlap Overlapping sensory and motor representations of forelimb and hindlimb (n=3 mice).  2.4 Discussion  Given that electrode impalements require several minutes each, we estimate that LBM is two orders of magnitude faster than electrode-based techniques. We anticipate that such an approach will be useful for determining changes in motor map structure before and after stroke or spinal cord injury (Raineteau and Schwab, 2001; Brown et al., 2007; Winship and Murphy, 2008). Although we performed mostly acute experiments, LBM is ideally suited to longitudinal experiments and can be performed multiple times on the same mouse through a chronic craniectomy (Trachtenberg et al., 2002) or possibly a thinned-skull preparation (Xu et al., 2007). Notably, we observed some activation through thinned bone at the edge of the craniectomy (Fig. 2.11). Repeated ICMS (on the same rat) has been conducted in the past (Teskey et al., 2002; Kleim et al., 2003a), but the likelihood of damaging the brain makes LBM a better choice for longitudinal studies of reorganization following experimental manipulations.  Other advantages of LBM over penetrating electrodes are related to sampling. With LBM, stimulation points can be arrayed in a perfect grid, ensuring a more uniform sampling of the cortex than is possible with ICMS. We found that the presence of large blood vessels did not completely block the photoactivation of ChR2, and that motor maps could be obtained even in areas occupied by large vessels, something that would not be possible with  57 ICMS (Fig. 2.2a). LBM appears to detect motor representations selectively as the resulting forelimb and hindlimb motor maps were located medial to the respective sensory maps (Fig. 6) in the approximate location expected for the mouse motor cortex (Ferezou et al., 2007) and in agreement with observations in rats (Hosp et al., 2008). Although the size and center of the forelimb and hindlimb motor representations were different, the two territories exhibited considerable spatial overlap. Possibly motor map overlap between limbs could reflect activation at off-target sites resulting from light scattering or spatial overlap between axonal or dendritic arbors of forelimb and hindlimb motor cortex. Alternatively, map overlap may be physiologically relevant and would suggest that specificity in motor output is achieved through additional regulation and not just the topographical layout of the motor cortex. Conceivably, LBM could be extended to single neurons to address whether excitation of individual neurons (Herfst and Brecht, 2008) within overlapping map areas can evoke both forelimb and hindlimb muscle excitation, or whether individual neurons are dedicated to specific limbs. Notably, we have shown previously that reorganization after a stroke can cause individual somatosensory neurons (normally preferentially activated by signals from a single limb) to process information from multiple limbs, suggesting that single neurons can assume multiple roles (Winship and Murphy, 2008). With regard to sensory maps, LBM shows that the centers of sensory and motor maps are generally ~0.5–1 mm apart (Fig. 2.14), supporting lower-resolution studies using ICMS in rats that had identified these areas as a mixed sensorimotor cortex (Neafsey et al., 1986).  The resolution of LBM depends on its ability to activate subsets of cortex despite the scattering of light and despite the presence of overlapping axons and dendrites from neurons with cells bodies outside of the activation area. Estimation of the cortical area LBM activates is a complex function of light scattering and depth-dependent changes in excitability. However, we can define a lower limit based on the size of the hindlimb motor map we observed (~0.65 mm2 or 0.9 mm in diameter). We estimated the area of cortex activated by LBM pulses using IOS imaging. The area showing > 50% of maximal activation was approximately 0.8 mm2 (measured at full width at half-maximal amplitude), about the size of the hindlimb motor map. Notably, IOS activation profiles of point-source ICMS electrodes were also relatively large, indicating that improvements in light delivery may not lead to  58 large gains in resolution. Given that excitation of motor neurons would not be linearly related to IOS changes and that the method does not directly read out activity within output neurons, it is possible that the spatial resolution of LBM is substantially greater than we estimated. Despite some uncertainty about map edge position, map center positions would be expected to be more precise and should accurately define the location of motor maps and potential changes after experimental manipulations. To improve the resolution of LBM, future work could use red-shifted variants of ChR2 (Zhang et al., 2008) using wavelengths of light that are less susceptible to scattering in tissue. Perhaps the largest gain in resolution would be from making a transgenic rat with ChR2 expression driven by the Thy1 promoter, where the motor cortex would be at least threefold larger (Kleim et al., 2003a).  Most previous motor-mapping studies have been conducted in rats or other larger species, but the variety of available transgenic mice makes them an increasingly attractive experimental model. As other strains become available, it will be interesting to conduct motor or intracortical mapping studies using mice that express ChR2 in other cortical layers or groups of neurons. LBM could also be extended to more complex movements using patterned stimulation or multisite activation (Graziano et al., 2002).  59 Chapter 3: Simple and cost-effective hardware and software for functional brain mapping using intrinsic optical signal imaging  3.1 Introduction  Intrinsic optical signal (IOS) imaging was developed in the late 1980s to visualize activity-dependent changes in the reflectance of brain tissue and delineate functional areas (Grinvald et al., 1986b). In the intervening years, IOS studies have contributed to our under- standing of the olfactory bulb (Grinvald et al., 1986) and the visual (Grinvald et al., 1986b; Bonhoeffer and Grinvald, 1993), auditory (Bakin et al., 1996), and somatosensory (Masino et al., 1993) cortices. Because it is a simple technique compatible with in vivo imaging preparations, IOS is commonly used to target cortical areas for interventions such as stroke or microinjections, or to select areas for two-photon imaging (Brown et al., 2007; Schummers et al., 2008; Winship and Murphy, 2008). Although one could attempt to locate functional areas based on brain atlas coordinates, these can vary even between littermate mice. Furthermore, functional and anatomical relationships will change after interventions such as deafferentation (Pons et al., 1991) or stroke (Winship and Murphy, 2008), making it necessary to combine functional macroscopic imaging with studies of microscopic structure. We describe a system that is simple and inexpensive to construct and use, and provide details of small animal surgery to ensure successful preparations. It is our hope that this information will encourage more researchers to incorporate IOS into their in vivo experiments.  60 3.2 Methods  3.2.1 Objective-mounted LED ring lights   IOS has traditionally been performed using a tungsten halogen lamp coupled to a focusing lens, filter set, light guide, and shutter (Toga and Mazziotta, 2002). Another possibility is to use light guides or through-lens LED illumination sources as employed for voltage-sensitive dye excitation or other forms of epifluorescence microscopy (Nishimura et al., 2006; Cescon et al., 2007; Albeanu et al., 2008). As a simpler alternative, we have devised an objective- mounted light-emitting diode (LED) light source. The low cost and small size of LEDs makes it possible to incorporate many lights into an unobtrusive ring light, which can be conveniently mounted to the objective lens (Fig. 3.1A and 3.1B). These LED ring lights produce spatially homogeneous light, and mounting to the objective auto- matically fixes their position relative to the focal plane. Because LEDs of appropriate wavelength for IOS are commercially available, no bandpass filters are required. A simple transistor circuit allows LEDs to be silently toggled on and off under electronic TTL control, eliminating the need for a mechanical shutter.  We construct ring lights from 1mm-thick aluminum strips drilled to accommodate 5 mm-diameter through-hole mounted LEDs (Fig. 3.1C and D). IOS imaging can be performed with red LEDs only, but it is advantageous to have both green LEDs (525nm, KSB1340-1P) for imaging brain vasculature at high contrast and red LEDs (625nm, Knight Lites KSB1385- 1P) to emphasize the rapid changes in blood oxygenation that accompany neuronal activity (Frostig et al., 1990). Wavelengths of light between 600 and 630 nm produce IOS maps with the greatest signal-to-noise ratio and spatial resolution (Toga and Mazziotta, 2002). In this range of wavelengths, the decrease in reflectance associated with increased deoxyhe- moglobin concentration lasts several seconds, and is followed by an increase in reflectance that can last up to 10 s as oxyhemoglobin concentration rise (Frostig et al., 1990). By restricting acquisition of response images to the first 1.5s after stimulation onset, we predominantly detect the initial decrease in reflectance (Fig. 3.5D).  61  We have designed two ring sizes (Fig. 3.1C and D) for mounting either to macroscopes (Ratzlaff and Grinvald, 1991) or standard microscope objectives. The smaller ring (Fig. 3.1D) is used primarily with a 4 ×, 0.13 NA UPlan Fl Olympus objective 26 mm in diameter. The same ring light can also be attached to a 2.5 ×, 0.075 NA Plan Neofluar Zeiss objective, which is useful for maintaining a large field of view when using cameras with smaller CCD chips (e.g. Sony XC-ST70). The use of higher magnification objective lenses for functional mapping is not recommended because they limit both the field of view and the number of photons collected. In addition to the microscopes described above, we conduct IOS experiments using a macroscope constructed from front-to-front video lenses coupled with a 52 mm threaded adaptor ring (Nikon #4598). The top lens (closer to the camera) is a 135 mm F2.8 Nikkor and the bottom lens is a 50 mm F1.4 Nikkor manual focus lens (Nikon). Macroscopes are advantageous for IOS experiments due to their wide field of view (4.3 mm with Dalsa 1M60) and efficient light gathering. We also conduct IOS imaging using a combination of 50 and 35 mm lenses (top and bottom), which form a macroscope with a field of view of 8.2 mm (when used with a Dalsa 1M60 camera). This arrangement of lenses is useful for bilateral imaging experiments with mice or unilateral experiments with rats.  62  Figure 3.1 LED ring lights (A) Large ring light mounted to the objective of a macroscope. (B) Smaller ring lights fitted to a 4×microscope objective. Both rings hold 12 LEDs, with alternating red and green LEDs evenly spaced around the ring. (C and D) Dimensions of small (C) and large (D) rings. Dashes mark lines along which the 1 mm aluminum plate should be bent. (E) Noise plots of three IOS systems expressed as a percentage change from baseline. Imaging data from three IOS systems were collected using an inert test specimen and analyzed to give images representing percentage change in reflectance of 625 nm light. The percentage change was then determined across a 2 mm × 200 µm segment of each image to generate the noise plots (see Section 3). (F) Standard deviations of the three noise plots.  63  3.2.2 Powering the LED ring light   A stable power supply is the foundation of a good illumination system. Fluctuations of DC power will degrade IOS signals and preclude mapping even if power varies by as little as ∼0.1% over tens of milliseconds. In many cases, standard laboratory constant-current power supplies are sufficiently stable for intrinsic optical signal imaging (e.g. Circuit-test PS- 3330) (Fig. 3.1E and F). A large battery regulated with a potentiometer is also an ideal power supply. It is possible to use feedback-based systems employing photodiodes to ensure constant light output (Beach and Duling, 1993). In our hands, however, such devices only function well over a small range of currents and in some cases may produce more noise than they are designed to remove. Ideally, power supplies should have a current limiting resistor in place to protect the LEDs from excessive current. If the LEDs are connected in parallel, the maximum allowable current is equal to the maximum current for a single LED multiplied by the total number of LEDs in the circuit. LEDs wired in series gain no such safety in numbers. If it is not possible to set a current limit on the power supply, a safety resistor should be added to the ring light circuit (Fig. 3.2A). The value of the resistor should be equal to the LED operating voltage divided by the maximum current, with an additional 10% added as a safety margin. For example, the red LEDs described above operate at 2.5 V with a maximum current of 0.1 A. If there are five LEDs wired in parallel, the maximum current is 0.5 A and so the safety resistor should have a value of 5 ohms plus an additional 10%. If the maximum voltage of the supply can exceed the operating voltage of the LEDs, the value of the resistor will need to be proportionally increased to protect the lights (e.g. a 50 ohm resistor should be used in the above circuit if it is powered by a 0–25 V supply).  3.2.3 Testing illumination stability  To assess the stability of illumination, we first determine the amount of LED light needed to fully illuminate the cortex. We operate our CCD cameras at approximately 80% of saturation (measured using EPIX XCAP image capture software), a value reached by  64 adjusting the camera’s exposure time and LED intensity. In general, signal-to-noise ratios improve with each additional photon collected. If light levels are insufficient to bring the CCD to 80% saturation (despite maximal exposure times and LED current), the ring light may require additional LEDs. We have found that with six of the red LEDs described and an exposure time of 100 ms, the camera can easily be brought to 80% saturation with sub- maximal LED current. Light levels should not exceed 90% of saturation anywhere in the image, as saturated pixels will degrade the IOS maps and render them useless. Once the amount of light (LED current) necessary to bring the CCD camera to ∼80% of capacity is established by adjusting brain illumination, one can begin experiments designed to check stability by collecting images in the absence of sensory stimulation or, better yet, by using an inert test specimen (Fig. 3.1E and F). Using the software provided in the supplementary materials, variation in baseline can be calculated by expressing signals as a percent change in light reflectance. If significant baseline variation is observed, e.g. a 0.05% variation in data integrated from 40 trials, one should optimize the illumination and/or detection systems. Using similar amounts of illumination light (measured with a power meter or photodiode, or estimated using the camera), different stabilized DC power sources can be evaluated. It may be necessary to employ a feedback system (Beach and Duling, 1993) to maintain a constant light level. It is important to optimize LED stability using a current level similar to that used to illuminate the brain because the variability of the power source and electronics is current-dependent. For a reflective non-biological test specimen we use a piece of polystyrene foam, altering reflectance (detected signal) to match that of the brain by covering the polystyrene with dark fabric or de-focusing. A flat and relatively homogeneous specimen such as a polystyrene foam block also allows one to adjust the position of LEDs for homogeneous illumination by bending individual tabs within the ring light, or by shifting the position of the entire unit on the objective lens. This should be done with the specimen in focus so that the lights are optimized for illumination of the focal plane. The microscope and sample should always be shielded from potentially varying ambient light by an opaque curtain during data acquisition.   65  Figure 3.2 Computer systems and circuitry for automated data collection (A) Two computers are used to control the lights, camera, and somatosensory stimulators. One computer runs the electrophysiology program Clampex (Axon Instruments) and controls the timing of the ring light via a transistor-transistor logic (TTL)-controlled switch. This computer controls (via TTL) a second machine running XCAP (EPIX), which acquires imaging data from the camera and triggers the stimulator. (B) LED switch circuit for TTL control of LEDs. The LED switch incorporates an opto- isolator (NTE3084) to toggle the LEDs under TTL control and shield the LEDs from any electrical noise transmitted via the TTL lines. Manual switches can over-ride TTL control to activate the lights. When adjusting DC power levels to the LEDs, it is important to turn on the LEDs via TTL control and not with the manual switch, as these two circuits have different resistances and will result in different light levels. The connections from the DC power supply and the LED outputs are banana plugs, and the TTL connections are BNC-type. A safety resistor (A) should be placed between the LED switch box and the ring light to protect the LEDs (see Section 2). The relay box (C) receives amplified input from the stimulator and TTL input from the data acquisition computer, and transfers power to each piezo actuator alternately using a simple 5 V relay (Song Chuan 842-1C-C). All connectors are BNC- type. Both the LED switch and the relay circuits are housed in plastic casings.  3.2.4 Cameras and data acquisition  Stimulus-evoked intrinsic optical signals are relatively small changes in light reflectance, and even under optimal conditions tend to be less than one part in 400 (Grinvald et al., 1988). Given that 8-bit cameras can only measure one part in 255, it is not immediately clear how such a camera could be used to image these slight changes in signal. To allow  66 detection of such small signals with relatively low-cost 8-bit cameras such as the Sony XC- ST70 (currently ∼$900), the following steps are taken. First, a camera with a relatively high photon well capacity (such as the above-mentioned) should be chosen. Cameras with high well capacities are able to absorb more photons before saturating, improving signal-to-noise ratios by lowering Poisson noise (estimated by taking the square root of the number of photons collected) (Pawley, 2006). To increase sensitivity and ensure linearity, the infrared blocking filter should be removed from the chip face and automatic gain control disabled. To extend the 8-bit dynamic range we collect images at relatively high frame rates (30 Hz) and collapse across time to effectively boost well capacity. To increase signals further, we also integrate light from multiple stimulation trials and in space when regions of interest are used for calculations, and always perform calculations with 32-bit precision.  Although it is possible to perform IOS imaging with an 8-bit camera, greater bit depth is preferable. The camera we have the most experience with is the 12-bit Dalsa 1M60, which has a well capacity of 600,000 photons (Dalsa specifications). Compared with the Sony XC- ST70, the increased bit depth allows us to use lower frame rates (10 Hz) and reduce the total amount of data handled while still improving the signal-to-noise ratio. With the Dalsa 1M60 camera, we have found that a 10 Hz frame rate with 2 × 2 pixel binning is a good compromise between the number of photons collected and the size of image files. Given that IOS signals are slow, it would be possible to use frame rates lower than 10 Hz to reduce the amount of data collected. For example, one could take a single frame per second, using one pre-stimulation image as baseline and a second post-stimulation image as response. Such a procedure would work well for cameras with a very large photon well capacity, but with the Dalsa 1M60 camera this is disadvantageous since relatively fewer total photons can be collected. It is also important to point out that with the Dalsa 1M60, binning or integration will not change the total number of photons that can be collected per pixel. To collect the maximum number of photons from a specimen, one needs to increase frame rates and reduce binning (resulting in increased file size) since the number of photons collected per pixel is fixed. After images are acquired, analysis software can be used to bin or average floating point data to increase precision by sampling relatively larger numbers of photons.  67 Some practical considerations will apply to any camera. For example, lowering the focal plane to ∼250 µm beneath the cortical surface helps to minimize biological noise from large surface blood vessels. If the curvature of the brain is very pronounced, it may be necessary to lower the focal plane further. A heat sink (e.g. Graftek AC-MS-0102) or active cooling device fitted to the Dalsa camera will also help to reduce noise by minimizing thermally generated pixels. However, these points are moot if excessive movement artifacts degrade the signal, so the animal should be held securely and the entire apparatus placed on a solid table. If a macroscope is used, mounting to an XYZ-translator (e.g. Sherline Tool #5430) facilitates both centering the sample (with respect to defined coordinates such as bregma) and focusing the image. The XYZ translator can be adapted for holding a camera by removing the tool mount and tapping the attachment point for ¼ - 20 mounting bolts.  It is common for researchers to use multiple computers to control the camera, deliver stimulation, and acquire data (Fig. 3.2A). For data acquisition, a frame grabber system must be capable of collecting image sequences at precise time intervals and reproducibly delivering an output signal. We currently use the EPIX frame grabber, but alternatives exist. Given variations in timing between different software packages, we find it best if the imaging software (EPIX XCAP) triggers the somatosensory stimulator directly (rather than another program). This will ensure that sensory stimulation is synchronized with image acquisition. Other parameters such as control of LED lights (Fig. 3.2B) or changing stimulation modalities (e.g. switching stimulation between fore- and hindlimbs, Fig. 3.2C) can be delegated to a second computer, as some degree of variation in timing is tolerable given that LED lights are switched on well before trials begin (>1 s). For a further check of synchrony between image acquisition and stimulation, we recommend mounting a red LED in place of the specimen and triggering it to coincide with stimulation. In this way an increase in LED signal will appear during stimulation and can be used to assess variation in synchronization between different trials. If one was concerned about synchronization during experiments, an LED pulse could be placed at the end of the trial to provide an internal standard for the entire time course of the experiment.   68 3.2.5 Surgical preparation  In our laboratory, we perform IOS imaging using three different types of animal preparations: the acute craniotomy, the chronic cranial window (Chen et al., 2008; Mostany and Portera-Cailliau, 2008), and the thinned skull preparation (Yoder and Kleinfeld, 2002; Xu et al., 2007). These procedures are approved by the University of British Columbia’s animal care committee, in accordance with the guidelines of the Canadian Council on Animal Care. IOS imaging will only be successful if the animal’s head is held firmly in position, so we have designed a baseplate (Fig. 3.3A) to accommodate a Kopf Instruments Mouse Head Holder (model 926) on an adjustable arm fixed to the baseplate by a pivoting base (Thor Labs mounting hardware). The plate allows the mouse to be moved from the surgical area to the microscope stage, and is especially practical when the same animal is used for multiple experiments performed on different apparatuses (e.g. IOS and photothrombotic stroke). This freedom of movement allows the head to be tilted until the cortical surface is level. For experiments conducted under isoflurane anesthesia, a gas anesthesia mask (Kopf model 907) can be fitted to the head holder. When acute craniotomies are performed, the head holder is used in combination with a headplate (Fig. 3.3B) that is cemented to the skull to frame the craniotomy (Yoder and Kleinfeld, 2002). A well around the window holds agarose made in a saline solution, and a coverslip is cut to fit this well. The headplate can be bolted to two upright posts on the baseplate (Fig. 3.3C) (Grutzendler and Gan, 2007). For chronic cranial window and thinned skull preparations, the headplate is omitted in favor of ear bars (Kopf model 921).  69  Figure 3.3 Stereotactic apparatus for mechanically stable IOS imaging (A) Baseplate designed to accommodate stereotaxic apparatus and mount to a microscope stage. The plate is made of 1 cm thick anodized aluminum. (B) Headplate used to stabilize the preparation during acute experiments. (C) Diagram of mouse surgical preparation. The anesthetized mouse is placed on a feedback-controlled heating pad (not shown) on the baseplate. The mouse is then fitted with a stereotaxic head holder and either a headplate or earbars. The head holder is attached to the baseplate by an adjustable arm. 3.2.6 Experimental design  Although any of the sensory cortices can be studied with IOS imaging, in our research we focus on the somatosensory cortex. Activation of the somatosensory cortex to elicit an IOS response can be achieved by stimulating various body parts, including the limbs, whiskers, tail, hips, shoulders, and trunk. For stimulation of the contralateral fore- and hindlimbs, we use an isolated pulse stimulator (AM Systems Model 2100) to drive two piezoelectric bending actuators (Piezo Systems Q220-AY-203YB) with a bi-phasic 4.5 V signal amplified to ± 90 V by a high-voltage booster (Piezo Systems EPA-007-012) (Fig.  70 3.2A). Although the stimulator can produce a ± 100 V signal, without the voltage booster it lacks sufficient current to properly drive the piezo. Each piezo is connected to one of the limbs by means of graphite pencil leads glued to the paw. A transistor-transistor logic (TTL)- controlled relay is used to alternate stimulation between the limbs (Fig. 3.2C). In these experiments, we delivered 1 s stimulus trains of 5 ms pulses at 100 Hz.  To reliably generate IOS maps, we perform up to 40 trials of stimulation per limb. It is usually possible to identify the activated areas with 10 trials, and after 20 trials the maps are generally well defined. Each trial begins by triggering the red lights and then collecting 15 baseline images at a frame rate of 10 Hz (Dalsa 1M60). One second of stimulation follows immediately after baseline acquisition, and 15 response images are collected during and after stimulation. Baseline and response images are compiled as multi-frame temporal stacks of TIFF files. Due to the slow time course of the hemodynamic response, we allow 20 s between stimulus trains to avoid any temporal overlap between trials.  3.2.7 Image analysis  Image analysis was conducted using a plugin for ImageJ (National Institutes of Health, Bethesda, MD; http://rsb.info.nih.gov/ij). The plugin, included as supplementary material and available in updated form online (neuroscience.ubc.ca/faculty/Murphy.html), creates IOS maps of cortical sensory representations based on IOS or voltage-sensitive dye (VSD) imaging data. The following analysis routine (Fig. 4) was performed using the ImageJ plugin described above: images were first converted to 32-bit depth to display results as decimals with high precision, and then the multi-frame image stacks were averaged across 40 trials. Next, we collapsed across time, averaging the 15 baseline and 15 response frames to a single frame each. After Gaussian filtering against a kernel (10 um) in the X and Y dimensions, we compared baseline and response either by calculating their ratio or by dividing their difference by the baseline. As an additional step, each frame in the temporal sequence was normalized to the mean baseline image (%!I/I0), yielding a video of percent changes in light reflectance over time.  71  Figure 3.4 Mathematical operations performed during IOS analysis (A) Raw data are collected as temporal sequences of 30 sequential frames. (B) The image sequences are converted to 32-bit depth and then averaged across all trials. (C) The frames from the first half of the sequence, taken before stimulation, are averaged together to give a single baseline image. The average of the last half of the sequence, taken during and after stimulation, represents the response. (D) The ratio of the response and baseline can be calculated to generate a functional map. (E) To create a normalized video of changes in reflectance, each frame in the sequence has the baseline average subtracted from it and is then divided by the same baseline average. Scale bar applies to (B)–(E).  72 3.3 Results  The ability to detect slight changes in reflectance is the key to successful IOS imaging, and requires consistent illumination. We characterized the temporal stability of our LED ring lights using a polystyrene test specimen as described in Section 2. This test was performed on three IOS systems used in our laboratory. System #1 incorporates a Sony XC- ST70 8-bit camera on an upright Olympus BX50WI microscope with a 2.5 × objective. System #2 uses the same ring light (Fig. 3.1B) as system #1, but is fitted to a 4 × objective on an Olympus BX51WI with a 12-bit Dalsa 1M60 camera. System #3 also uses a 12-bit Dalsa 1M60 camera, but images are captured through a macroscope and so a larger ring light is used (Fig. 3.1A). 40 trials were run on each system, and the data were analyzed using the same routine performed for IOS mapping. The percentage change between baseline and response was plotted for a region of interest of 2 mm in the x-dimension, averaged over 200 µm in the y-dimension. These plots (Fig. 3.1E) represent both the level of noise inherent to the system as well as spatial homogeneity across an image. To compare noise levels across the three systems, we calculated the standard deviations of the noise plots (Fig. 3.1F). For system #1 (Sony XC-ST70, 2.5× objective), this value was 0.0068 %. System #2 (Dalsa 1M60, 4 × objective) had a better noise profile (SD = 0.0029 %), and system #3 (Dalsa 1M60, macroscope) was similar (SD = 0.0028 %). After generating somatosensory maps (Fig. 3.5A–C), we performed a simple test to check for the presence of artifacts. Ideally, time plots of reflectance from regions of interest outside of sensory representations will show relatively constant reflectance over time, suggesting that changes within IOS maps are specific to somatosensory stimulation (Fig. 3.5D).  73  Figure 3.5 IOS maps of mouse somatosensory representations (A) The cortical surface illuminated with green light (525 nm). Anterior–posterior and medial–lateral directions are marked. Bregma is not marked, but the forelimb representation is located an average of 2.5 mm lateral and 0.5 mm anterior of bregma, with the hindlimb at 1.9 mm lateral and 0.6 mm posterior of bregma (n = 5 mice). (B) Forelimb and hindlimb (C) somatosensory maps of cortical area (A). Grey values are on a linear scale from a 0.02 % increase in reflectance (white pixels) to a 0.08% decrease in reflectance (black). Scale bar in (A) also applies to (B) and (C). (D) Changes in reflectance of 625 nm light over time for the regions of interest numbered in (B) and (C). Regions 1 and 4, within the areas of activation, show a decrease in reflectance following somatosensory stimulation (marked with bar). Cortical regions outside of the somatosensory representations (2 and 3) show little change in reflectance after stimulation. The plots are the mean of 10 trials from a single animal, with a region of interest 0.07 mm2.     74 3.4 Discussion  3.4.1 Advantages of LED ring lights  Ring lights have several advantages over traditional illumination sources. By virtue of the spatial homogeneity of illumination they provide, LEDs represent an ideal light source for IOS. Even when 8-bit cameras are used to collect imaging data, the consistent illumination provided by LED ring lights permits good signal-to-noise ratios. When detecting such faint changes in reflectance, noise introduced by fluctuations inherent to arc or incandescent lamps can rapidly degrade the signal (Albeanu et al., 2008). Because they are fixed to the objective lens, LED ring lights maintain their position with respect to the objective focal plane, something that would be difficult to achieve with light guides. One alternative to light guides or ring lights would be to direct illumination light through the microscope objective. However, this requires an additional mirror and/or splitter, complicating the design and dividing the reflected light. Focusing light through the objective may also cause uneven illumination because the curvature of the brain results in a variable working distance. Ring lights work without the filter sets, shutters, and optics used in lamp-based illumination systems, simplifying the IOS system and reducing cost.  3.4.2 Combining IOS with two-photon microscopy  Two-photon microscopy is an excellent tool for investigating the fine structure of neurons, and can be combined with Ca2+ imaging in vivo to study cell function (Stosiek et al., 2003). This information is valuable, but needs to be placed in the context of the larger cortical areas in which these cells exist. IOS is a simple, non-invasive tool that can be used to delineate functional areas and guide higher-resolution investigations of cortical function. IOS can be combined with any form of sensory stimulation, permitting researchers to apply calcium indicators or other probes accurately in the sensory representation of interest (Brown et al., 2007; Schummers et al., 2008; Winship and Murphy, 2008). The simplicity of our ring lights makes the IOS apparatus easy to incorporate into two-photon microscopes to permit  75 low power IOS mapping in parallel with two-photon imaging of fine structure and function. IOS experiments require a CCD camera and the ability to produce a reflected light image of the brain. IOS imaging using ring lights is conducted without the use of excitation or emission filters, so a blank slot in the microscope’s filter set is necessary. Most upright microscopes used for two-photon microscopy are already equipped with the hardware necessary to redirect the beam path from two-photon excitation/emission to a CCD camera, making it a simple process to accommodate both IOS and two-photon on the same microscope. For a description of the adaptations necessary to perform both CCD camera and two-photon imaging on a BX-51WI Olympus microscope see (Sigler et al., 2008).  3.4.3 Variable illumination wavelength and spectroscopy  The wavelength of illumination will determine the physiological signal detected by IOS. We have used 625 nm LEDs to monitor rapid changes in blood oxygenation that accompanies neuronal activity (Frostig et al., 1990) and to optimize the spatial resolution of the resulting maps (Toga and Mazziotta, 2002). Changes in total blood volume are best observed by illuminating the brain with 550 or 570 nm light, which is absorbed equally by oxy- and deoxyhemoglobin (Toga and Mazziotta, 2002). Finally, longer wavelengths (> 630 nm) can reveal activity-related light scattering believed to reflect cell-swelling and other blood-independent processes (Cohen, 1973; Andrew and MacVicar, 1994; Joshi and Andrew, 2001). One important observation is that all of these different response signals localize to the same cortical region, and therefore IOS maps will be very similar regardless of the imaging wavelength (Frostig et al., 1990).  Advances in optical design have made it possible to image the brain with multiple wavelengths of light simultaneously, allowing a three-dimensional representation of spectroscopic signals from living tissue to be constructed. This method, termed laminar optical tomography (LOT), detects functional changes in oxygenation levels in addition to providing structural information based on endogenous chromophores and fluorophores (Burgess et al., 2008). Multispectral reflectance imaging (MSRI), another extension of IOS  76 involving multiple wavelengths of illuminating light, allows total hemoglobin concentration and blood oxygenation to be quantified (Devor et al., 2003; Jones et al., 2008). Although LOT requires a complex optical system, ring lights could conceivably be adapted to incorporate LEDs of several different wavelengths in order to perform MSRI experiments. Spectral overlap between LEDs would be a limitation of ring lights however, and placing many different types of LEDs sparsely within the same ring light could also cause inhomogeneous illumination. These problems could likely be addressed with excitation filters, careful design of the ring lights, and possibly corrections performed in software.   77 Chapter 4: Distinct cortical circuit mechanisms for complex forelimb movement and motor map topography  4.1 Introduction   The motor cortex has long been known to play a central role in the generation of movement (Fritsch and Hitzig, 1870), but fundamental questions remain to be answered about the functional organization of its subregions and their neuronal circuits. Results from electrical brain stimulation have traditionally been interpreted with an emphasis on somatotopy (Penfield and Boldrey, 1937; Asanuma and Rosén, 1972), but the utility of this principle has diminished with the discovery of multiple representations of the body (Neafsey and Sievert, 1982; Luppino et al., 1991; Schieber, 2001). A more nuanced view has since developed, with recordings made during voluntary movements in monkeys demonstrating that neurons in motor cortex encode information related to the force (Evarts, 1968), direction (Georgopoulos et al., 1986), and speed of movements (Moran and Schwartz, 1999; Churchland et al., 2006). The activity of cortical neurons also reflects both preparation for movement (Sanes and Donoghue, 1993; Paz et al., 2003) and the interpretation of actions performed by others (Gallese et al., 1996; Hari et al., 1998). Recently, experimentation with prolonged trains of stimulation has suggested that the brain’s multiple motor representations may be organized according to classes of behavior (Graziano et al., 2002a; Stepniewska et al., 2005; Ramanathan et al., 2006).  Despite the detailed knowledge gleaned from these efforts, our understanding of the macroscopic organization of motor cortex remains incomplete. Much of our understanding about the motor cortex comes from experiments in which stimulation or recording is performed at a few cortical points. Technical limitations have traditionally made it difficult to probe the cortical circuitry underlying motor representations in a uniform, quantitative manner. Recently, we and others have developed a novel method for rapid automated motor mapping based on light activation of Channelrhodopsin-2 (ChR2) that has facilitated  78 experiments which were previously impossible (Ayling et al., 2009; Hira et al., 2009; Komiyama et al., 2010). This technique has the advantage of objectively and reproducibly sampling the movements evoked by stimulation at hundreds of cortical locations in mere minutes. Here, we apply light-based motor mapping to investigate the functional subdivisions of the motor cortex and their dependence on intracortical activity.  The ability to repeatedly map the motor cortex over timescales ranging from minutes to months has allowed us to appreciate the dynamic nature of movement representations and facilitated the comparison of motor maps generated before and after pharmacological perturbations of the intracortical circuitry. We have exploited the predominant expression of Channelrhodopsin-2 in layer 5B pyramidal neurons of Thy-1 transgenic mice (Arenkiel et al., 2007; Wang et al., 2007; Yu et al., 2008; Ayling et al., 2009) to target this class of corticofugal cells directly, exposing their contribution to motor cortex topography and identifying a functional subdivision of the mouse forelimb representation based on movement direction. Prolonged trains of light or electrical stimulation revealed that activation of these subregions drives movements to distinct positions in space. To identify mechanisms that could account for the different movement types evoked by stimulation of these cortical subregions, we performed pharmacological manipulations of the intracortical circuitry and targeted anatomical tracing experiments.  4.2 Methods 4.2.1 Animals and surgery  Animal protocols were approved by the University of British Columbia Animal Care Committee. Channelrhodopsin-2 transgenic mice (Arenkiel et al., 2007) from Jackson Labs (line 18, stock 007612, strain B6.Cg-Tg(Thy1-COP4/EYFP)18Gfng/J) established a breeding colony. Adult mice aged 2-6 months and weighing 20-30 g were used for these experiments. Isoflurane anesthesia was used during surgery and intrinsic optical signal imaging of somatosensory representations, but was replaced by ketamine/xylazine (100/10 mg/kg, supplemented at 1/10th initial dose as necessary) prior to motor mapping. Craniectomies were  79 performed on transgenic mice used in acute experiments, but virally-transduced mice (see section below for details on injections) were mapped through the intact skull due to concern that multiple cranial surgeries could damage the cortex. Chronic mapping was performed through a cranial window (Harrison et al., 2009).  4.2.2 Light-based motor mapping  Light-based mapping methodology has been described in detail (Ayling et al., 2009). Briefly, we used a scanning stage (ASI MS-2000) controlled by custom Igor Pro software (Wavemetrics) to direct a fixed 473 nm laser beam (Crystalaser, focused to 100 µm diameter, 10 ms pulses, 1-10 mW total or 127-1270 mW/mm2) to an array of cortical sites (typically 13x13, with 300 µm spacing between sites). This process was repeated 3-5 times to obtain a mean value for each pixel of the map. Stimulation was delivered in a semi-random order with identical stimulus intensity for all sites within a map. Movements were detected using laser range finders with mm sensitivity targeted to the forelimb and hindlimb (Keyence LK-081). In order to exclude artifacts (e.g. from breathing or electrical noise), responses were considered to be genuine only if their amplitude exceeded three times the standard deviation of the 500 ms pre-stimulus period within 100 ms after stimulus onset.  4.2.3 Map analysis  Motor maps were generated by plotting the peak amplitude of the mean movement profile corresponding to each cortical site of stimulation. Amplitude was quantified within a 300 ms time window after laser stimulation. If the amplitude of the movement evoked at that site was positive, the corresponding pixel was added to the adduction map. If the amplitude had a negative value with respect to the baseline, that site was added to the abduction map. In the case of bidirectional movement profiles where both the positive and negative components satisfied the amplitude criteria, the corresponding site was included in both the abduction and adduction maps and counted as overlap between maps. For each map, the center of gravity was calculated along with the mean amplitude and latency for the nine pixels closest to the  80 center point. Maps with mean amplitude of < 0.1 mm at the center were excluded from further analysis. Separation between Mab and Mad was defined as the distance between the centers of gravity for each map.  4.2.4 Video capture of evoked movements  After completing 2-5 motor maps, mice were raised into a sitting posture with their forelimbs hanging freely. Stimulus sites were placed as close to the centers of the abduction and adduction representations as possible without targeting major blood vessels, since these absorb light strongly (Ayling et al., 2009). 51 frames were captured at a rate of 100 Hz beginning 10 ms prior to laser stimulus onset, and paw trajectories were generated from the raw image sequences using the plugin “MTrack2” for ImageJ. 10-20 repetitions were then averaged for each trial, and speed and angle profiles were calculated based on this average trajectory.  For awake experiments, ChR2 transgenic mice were implanted with optical fibers (Thorlabs BFH48-200) extending to the cortical surface and terminating in a ferrule connector (Precision Fiber Products) fixed to the skull with dental acrylic and bone screws. Two fibers were implanted, targeted to the mean co-ordinates of the Mab and Mad map centers. These locations were stimulated alternately (5 mW 5 ms pulses at 100 Hz for 500 ms) using a 473 nm laser (IKECOOL IKE-473-100-OP) connected via an optical commutator (Doric). Stimulus evoked behavior was recorded by a CCD camera (Dalsa 1M60) and frame grabber (EPIX). Limb trajectories were analyzed in the same manner as the anesthetized data, except that paw position was tracked using the plugin “Manual Tracking” for ImageJ.  4.2.5 Intracortical microstimulation  Glass pipets (tip width 10-20 µm) containing a 0.25 mm bare silver wire were filled with 1% fast green in 3M sodium chloride. A micromanipulator (Sutter) was used to advance  81 the pipet to a depth of 700 µm. Stimulation sites were matched with those targeted by laser stimulation in the same animals. Trains of 200 µs 100 µA pulses at 200 Hz with 10-500 ms durations were generated by an AM systems stimulator and a WPI stimulus isolator.  4.2.6 Virus injections and anatomical tracing  For motor mapping experiments involving virally-transduced mice, 1-2 µL of adeno- associated virus (serotype 2/1 CAG-ChR2-GFP) was injected through a burr hole into the sensorimotor cortex of ChR2-negative mice 2 mm lateral of bregma at a depth of 500 µm using a 5 µL Hamilton syringe with a 33 gauge needle and a syringe pump (WPI). Mice recovered for 2-4 weeks before being used in experiments. For anatomical tracing experiments, Mab and Mad were identified by light-based mapping through the intact skull of ChR2 transgenic mice (Hira et al., 2009). Injections were made using a custom pressure injection system (Cetin et al., 2006). At each site, 250 nL of virus (turboRFP, mCerulean, or eGFP, with matched serotypes 2/1 or 2/9) was injected over 10 min at a depth of 500 µm. Fluorophore placement in Mab vs. Mad was alternated between animals. In 3 of 7 animals motor maps could not be produced by transcranial stimulation, and injections were targeted to the mean co-ordinates of Mab and Mad. Three weeks after injection, the mice were transcardially perfused and 100 µm coronal sections were sliced on a vibratome, with every third section mounted for epifluorescence imaging. Fluorescence plots from midline were smoothed and averaged, and the mean position of peak fluorescence was calculated for each animal.  4.2.7 Pharmacology  For experiments involving glutamate receptor antagonists, CNQX (4.5 mM) and MK801 (300 µM), gabazine (1 µM), or picrotoxin (100 µM) in physiological saline solution were applied to the craniectomy. The compounds were allowed to incubate for 30 min before mapping resumed, and were replenished (at the same concentration) every ~30 min  82 throughout the experiment. Control experiments were identical except that saline solution was applied in place of the drugs.  4.2.8 Local field potential recordings  A NeuroNexus multi-site electrode (A1-X16-3mm-50-413) was lowered 800 µm into sensorimotor cortex using a micromanipulator (Sutter), and a reference electrode was immersed in the saline bathing the cortical surface. In each experiment, at least 50 trials of 1ms, 0.1 Hz electrical (1 mA) and ChR2 (10 mW 473 nm) stimulation were recorded, and then CNQX and MK801 were applied to the cortical surface as above and incubated for 30 min before recordings were repeated. The mean peak-to-peak amplitude was measured in a time window of 300 ms after stimulus onset for each electrode contact. The mean amplitude of the baseline noise was subtracted, and adjacent electrode contacts were binned by averaging.   83  Figure 4.1 Spatial heterogeneity of evoked movements revealed by light-based mapping A Anesthetized, head-fixed mice were placed in the prone position with their contralateral forelimb suspended to allow free forward or backward movement (left). Forelimb movements evoked by optogenetic cortical stimulation were assayed as either abduction or adduction depending on the direction of movement recorded by a non-invasive motion sensor (right). Mapping was relatively non- invasive and could be performed repeatedly in the same animal (see figure s1). B By delivering three repetitions of stimulation to an array of cortical points in random order, a map of averaged evoked movements (C) was assembled. Note the heterogeneity of movements in this representative example. D,E The same movements classified by direction and scaled by amplitude to form separate maps of forelimb abduction and adduction. F Merged motor map, with sites from which abduction movements were evoked in green (Mab, center of gravity marked with an x), and adduction sites in red (Mad). Similar maps were generated in animals where expression of ChR2 was mediated by viral transduction (see figure s2). G Latencies from stimulus onset to movement onset for each of the cortical sites in F. All data in this figure are from the same representative animal.  84  4.3 Results  4.3.1 Movement-based mapping of motor cortex  We used optogenetic motor mapping to rapidly stimulate hundreds of cortical points in ChR2 transgenic mice (Arenkiel et al., 2007) and assemble maps based on evoked movements of the contralateral forelimb and hindlimb (Figure 4.1A-C). In these experiments, anesthetized mice were head-fixed in the prone position with their contralateral limbs suspended. In this posture, the limbs were able to move freely along the axis of measurement of a laser range finder. The resultant movement maps were centered at positions consistent with those obtained by EMG recording or visual observation (Forelimb: 2.2 ± 0.1 mm lateral, 0.05 ± 0.09 mm anterior of bregma; Hindlimb: 2.0 ± 0.11 mm lateral, 0.21 ± 0.1 mm posterior of bregma, n = 14 mice, all values ± SEM) (Pronichev and Lenkov, 1998; Ayling et al., 2009; Hira et al., 2009; Tennant et al., 2010). Composite maps based on the average of three repetitions were highly reproducible, with a shift in center position of 0.19 ± 0.02 mm (n = 12 mice) between mapping trials (~30 min per composite map). In a separate group of animals implanted with cranial windows, maps remained stable for months (Figure 4.2). Movement maps could also be generated in animals where ChR2 was expressed in pyramidal neurons of both superficial and deep cortical layers by transduction with adeno-associated virus (Figure 4.3).  85  Figure 4.2 Motor maps are stable for months in animals implanted with cranial windows  A Motor maps generated two weeks apart in the same representative animal. B In a group of mice mapped weekly for a total of 8 weeks, the centers of gravity of the adduction (red) and abduction (green) maps were highly stable, with a mean shift in center position comparable to the size of one map pixel (0.3 mm). There was no difference in weekly shift between Mab and Mad (p = 0.34, n = 7, paired t-test). C The latency from stimulus onset to movement onset was also stable over time for movements evoked from the centers of Mab and Mad (green and red lines, respectively; Fsite(1,8) = 2.65 p = 0.11, Ftime(1,8) = 1.79 p = 0.09, RM-ANOVA). D Same plot as C but for movement amplitudes, which were also consistent over time (Fsite(1,8) = 0.92 p = 0.34, Ftime(1,8) = 0.49 p = 0.86, RM-ANOVA). Error bars in all graphs are SEM.  86   Figure 4.3 Movements evoked by stimulation of virally-expressed ChR2 A Coronal section showing ChR2-eYFP expression three weeks after 2 mL of adeno-associated virus were injected into the sensorimotor cortex of a representative mouse. B Mean movements (average of five repetitions) evoked by stimulation (10 mW 10 ms) across an array of cortical sites in the same animal as A. C Motor map based on the same data as B. D Mean separation between Mab and Mad for virally-transduced (grey) and Thy-1 transgenic mice (black) (p = 0.7647, t-test). E Mean areas for Mab (green) and Mad (red) maps and their region of overlap (yellow) in virally-transduced mice. F Mean movement trajectories evoked by stimulation of Mab (green) and Mad (red) with a 500 ms train of laser stimulation (5 ms 10mW pulses at 100 Hz). Note the similarity to movements evoked by transgenic ChR2 or electrical stimulation (Figure 4). Error bars in all graphs are SEM.  4.3.2 Forelimb motor cortex is subdivided into functional subregions  Consistent with previous results, forelimb movements could be elicited by stimulation (10 ms pulses, 1-10 mW or 127-1270 mW/mm2) of a broad cortical area, up to 2 mm anterior and posterior of bregma (Ayling et al., 2009; Tennant et al., 2010). However, when forelimb movements were examined at stimulation sites across the motor cortex, a diversity of response types became apparent (Figure 4.1C-F). Evoked movements were divided into two classes depending on the direction of forelimb movement (abduction or adduction, Figure 4.1D-F). Stimulation sites that produced movements containing both abduction and  87 adduction components were considered as regions of overlap between abduction and adduction maps. This analysis revealed a functional subdivision of the motor cortex that was not apparent from EMG-based maps, even when antagonistic muscle pairs were compared (Ayling et al., 2009).   The motor cortex abduction representation (here termed Mab) was not different from the adduction representation in area (Mad) (4.7 ± 0.6 vs. 4.9 ± 0.7 mm2, n = 14 mice), but movements evoked from the center of Mab tended to be smaller than those evoked from the center of Mad (0.2 ± 0.02 vs. 0.5 ± 0.09 mm, p = 0.036 paired t-test, n = 14 mice).  Mab movements also began at a shorter latency from the onset of cortical stimulation (19.4 ± 0.9 vs. 24.6 ± 1.5 ms, p = 0.002 paired t-test, n = 14 mice) (Figure 4.1G). Mab was typically located anterior and lateral of Mad (Figure 4.4A,B). Mab and Mad were both centered within the boundaries of the caudal forelimb area defined by intracortical electrical microstimulation, but frequently extended into the reported territory of the rostral forelimb area (Tennant et al., 2010). The Mad portion of the forelimb map overlapped with hindlimb motor cortex to a greater extent than Mab (55.9 ± 8.7 vs. 43.9 ± 7.5 %, n = 14 mice, p < 0.01, paired t-test). Mad was also closer than Mab to the centers of the hindlimb somatosensory representation, whereas Mab was closer than Mad to the center of the forelimb somatosensory representation (Figure 4.4B).  Mab and Mad representations were not different in consistency, defined as the percentage of stimulus sites from which movements were evoked in all three repetitions of a composite map (8.3 ± 2.3 vs. 10.8 ± 3.0 %, n = 12 mice). The centers of gravity of Mab and Mad were separated from each other by an average of 0.6± 0.06 mm (p < 0.0001, single sample t-test vs. hypothetical mean 0, n = 14 mice). When a threshold was applied at 50% of each map’s peak amplitude, separation between Mab and Mad increased to 1.2 ± 0.07 mm (n = 14 mice), which is comparable to the distance between the centers of forelimb and hindlimb somatosensory maps (1.2 ± 0.2 mm, n = 7 mice). These observations demonstrate that the mouse forelimb motor cortex can be reproducibly subdivided according to a simple assay of evoked movement direction.  88  Figure 4.4 Relative positions of motor and somatosensory representations A Representative motor maps (Mab in green, Mad in red) thresholded at 0.1 mm of limb displacement and overlaid onto an image of the cortex. Somatosensory representations of the forelimb (sFL, purple) and hindlimb (sHL, cyan) were generated by intrinsic optical signal imaging and thresholded at 0.02 % change in reflectance of 635 nm light. White crosses mark the center of gravity for each representation. B Mean positions of the centers of gravity with respect to bregma for each of these representations, with the variability of the co-ordinates (standard deviation) represented by the lengths of the cross-bars (n = 6 mice for sensory maps and 14 mice for motor maps).   89 4.3.3 Prolonged stimulation of abduction and adduction representations drives movements to distinct positions in space  Figure 4.5 Complex movements evoked by prolonged stimulation of the abduction or adduction representations A Representative forelimb motor map generated with pulses (10 ms) of 473 nm light (left). After identifying the centers of gravity of the abduction (Mab) and adduction (Mad) representations in anesthetized animals, the center of each representation was stimulated with trains of light stimulation (left) while the resulting movements were captured by high-speed video to reconstruct movement trajectories (center). In separate experiments, Mab and Mad  were stimulated alternately via optical fibers in awake, freely-moving animals (right). B Mean trajectories of movements evoked in anesthetized (left) and awake (right) animals by stimulation of Mab (green) and Mad (red) marked with error bars (SEM). Movements evoked from a given site are highly reproducible within animals (see figure s4). Movement trajectories are strongly dependent on stimulus duration, but movement maps are not (see figure s5). C Mean forelimb displacement for the movements depicted in B. Dashed blue lines above the abscissae denote the period of stimulation. Movements evoked by stimulation of Mab are significantly larger than Mad in anesthetized animals (F(1,44)stim site = 12.36, p = 0.0025, F(1,44)interaction = 5.638, p < 0.0001, RM-ANOVA, n = 10) and in awake animals (F(1,49)stim site = 557.4, p < 0.0001, F(1,49)interaction = 1.661, p = 0.01 RM-ANOVA, n = 4). D Speed profiles for the movements shown in B. Note that despite differences in movement trajectory, speed profiles are almost identical for both anesthetized and awake animals. Error bars in all graphs are SEM.   90 It has been proposed that long stimulus trains may be more effective than shorter bursts at producing ethologically relevant movements and identifying cortical movement representations (Graziano et al., 2005). Despite the ability of light-based mapping to rapidly, quantitatively, and uniformly sample the motor output of a large cortical area, the restricted sampling of forelimb displacement in our method limits the information that can be gathered about the movements generated by stimulation of any particular cortical location. To better describe the properties of the Mab and Mad motor subregions, we used a high-speed CCD camera to record forelimb movements evoked by stimulation of sites near the center of each map. In these experiments, the centers of the Mab and Mad maps were defined with the mouse lying prone and the contralateral forelimb suspended parallel to the ground (Figure 4.5A, left). The anesthetized mice were then moved to a sitting posture, with their heads fixed and their forelimbs hanging free (Figure 4.5A, center).   With prolonged stimulus trains (500 ms), the forelimb tended to reach a final position within ~300 ms and remain there for the duration of the stimulus. Stimulation of Mab caused the contralateral forelimb to be raised and then brought toward the midline, whereas stimulation of Mad typically produced rhythmic movements lower in space, often coupled with movement of the hindlimb (Figure 4.5B). These movements were reproduced in anesthetized mice where ChR2 was locally expressed using adeno-associated virus (Figure 4.3) and in awake, freely moving ChR2 transgenic mice stimulated within Mab and Mad via optical fibers (Figure 4.5A,B right). In both anesthetized and awake mice, the displacement of the limb from its starting position was significantly greater when Mab was stimulated rather than Mad (Figure 4.5B,C). Although movement trajectories (Figure 4.5B) and displacements (Figure 4.5C) were clearly dependent on stimulus site for both awake and anesthetized mice, the speed profiles of Mab and Mad movements were nearly identical (Figure 4.5D). Movements evoked from each site were remarkably consistent from trial to trial, and the variability that they did exhibit had a temporal structure that depended on the site of stimulation (Figure 4.6). Increasing stimulus duration generally had little effect on movement map structure, despite apparent changes observed in movement trajectories (Figure 4.7). Consistent with previous results from electrical stimulation (Ramanathan et al., 2006), modulating optogenetic stimulus intensity did not affect movement trajectories  91 evoked by prolonged stimulation (Figure 4.8). These experiments complement the mapping study by exposing the distinct types of complex movement that can be evoked from Mab and Mad by prolonged stimulation in both anesthetized and awake mice.    Figure 4.6 Movement trajectories are consistent for a given site of cortical stimulation A Representative movement trajectories obtained by repeated stimulation (100 ms train of 5 ms 5 mW pulses at 100 Hz) of Mab (top row) or Mad (bottom row) in a single animal. Note the similarity and scale invariance of movement trajectories across repeated trials, and the consistent differences between trajectories evoked from Mab and Mad. Dashed blue lines along the mean trajectories at right denote the stimulation period. B Movement variability profile, expressed as the hypotenuse of XY limb position variability (SEM, n = 20 trials per animal) over time as a percentage of the mean peak displacement for that animal (n = 16 animals). Mab  (green) movements showed less variability in later movement stages than Mad  (red) movements (F(1,44)interaction = 2.395, p < 0.0001, RM-ANOVA, n = 16). Error bars are SEM.  92   Figure 4.7 Movement topography is evident with short pulses, but complex movements require prolonged stimulation A Similar motor maps generated from the same animal using 10 ms pulses of 5 mW light (left) or 100 ms trains (5 ms 5 mW pulses at 100 Hz). Scale below maps refers to movement amplitude in mm. Note that the green and red channels are scaled independently in this example. B Area of Mab and Mad maps was similar for maps evoked by pulses or trains of light (left, p > 0.05, n = 4, RM-ANOVA), as was separation between Mab and Mad representations (right, p = 0.43, n = 4, paired t-test). C Movement trajectories evoked by stimulation of Mab (green) or Mad (red) with pulses (left) or prolonged stimulus trains (center, right) of light. Error bars in all graphs are SEM.  93  Figure 4.8 Motor map area is dependent on stimulus intensity, but complex movements are not A Representative motor maps obtained using a range of laser stimulus intensities (1 mW = 127 mW/mm2 for laser spot diameter of 100 mm). B Forelimb map area decreased with decreasing laser power (r = -0.957, p = 0.011) while the distance between Mab and Mad map centers increased (C) (r=0.930, p = 0.022, n = 4 mice). D Mean trajectories of movements evoked by 100 ms trains were similar regardless of optogenetic stimulus intensity. E The peak displacement of the forelimb from its starting position was independent of light intensity (p > 0.05, n = 4, RM-ANOVA). Note that the data shown in D-E and A-C are from different animals. Error bars in all graphs are SEM.     94  Figure 4.9 Optogenetic and electrical stimulation evoke similar complex movements A Mean trajectories of movements evoked by 500 ms of electrical (left) or optogenetic stimulation (right) of the Mab (green) and Mad (red) representations in the same animals. B These movements did not differ in peak displacement, time to peak displacement, or angle from origin at peak displacement. Peak movement speed was greater for optogentically-evoked Mab and Mad  movements (paired t- tests). C Speed profiles for the movements depicted in (A). Solid black lines correspond to optogenetically evoked movements, dashed black lines to electrically evoked movements. Dashed blue lines above the abscissae denote the period of stimulation. Despite similar movement trajectories (A), speed profiles were strongly dependent on stimulus type for both Mab  (left, F(1,50)stim type = 28.41, p < 0.0001, F(1,50)interaction = 1.682, p = 0.0033, RM-ANOVA with Bonferonni post-test results indicated by asterisks on graph) and for Mad (right, F(1,50)stim type = 24.68, p < 0.0001, F(1,50)interaction = 2.798, p = 0.0033). Increasing stimulus intensity had no effect on complex movement trajectories, but did increase map area (see figure s6). Error bars in all graphs are SEM.   95 4.3.4 Electrical and optogenetic stimulation evoke similar movements  To determine whether these complex movements require selective stimulation of layer 5B neurons, we compared optogenetic stimulation (500 ms train of 5 ms, 5 mW pulses at 100 Hz) with trains of electrical intracortical microstimulation (ICMS) targeted to layer 5 of cortex (500 ms trains of 200 µs, 100 µA pulses at 200 Hz) (Ramanathan et al., 2006). Given the differences between ICMS and optogenetic stimulation, we were surprised to discover that ICMS was able to closely reproduce the complex movements characteristic of transgenic or viral optogenetic stimulation of Mab and Mad (Figure 4.9A, Figure 4.3). In addition to their overlapping trajectories, movements evoked by either method had comparable peak displacements, time to peak, and angle from origin at peak displacement (Figure 4.9B). Interestingly, although movements evoked by ICMS or optogenetic stimulation shared the same end point, ICMS-evoked movements were significantly slower (Figure 4.9C). These results suggest that the site of stimulation determines the trajectory of the resulting movement (Figure 4.5), whereas movement speed depends on the mechanism of stimulation (Figure 4.9).  4.3.5 Specificity of complex movements evoked from different cortical areas requires intracortical synaptic transmission  After characterizing the movement representations of the mouse motor cortex, we investigated their mechanistic basis. We hypothesized that the distinct movements produced by the Mab and Mad motor cortex subregions could be explained by differences either in their output projections (Rathelot and Strick, 2009; Matyas et al., 2010), or in the pattern of input they receive from recurrent intracortical circuits (Weiler et al., 2008; Anderson et al., 2010; Hooks et al., 2011) or subcortical loops (Flaherty and Graybiel, 1991; Hoover and Strick, 1993; Kelly and Strick, 2003). To test the extent to which cortical synaptic input contributes to the differences between Mab and Mad motor subregions, we compared movement trajectories generated before and after the application of glutamate receptor antagonists (CNQX 4.5 mM and MK-801 0.3 mM) or saline to the cortical surface (Figure 4.10A). In the  96 control condition Mab and Mad movements had non-overlapping trajectories that could be distinguished by plotting the angle of the forelimb from the starting position (Figure 4.10B, left). Disrupting glutamatergic transmission increased the extent to which Mab and Mad trajectories overlapped, biasing both toward medial rotation (Figure 4.10B, right). Glutamate receptor antagonists also had a site-specific effect on speed profiles, causing a delayed increase in movement speed for Mad, but not Mab (Figure 4.10C). These results suggest that differences between movements evoked by prolonged stimulation of Mab and Mad may reflect variation in the patterns of glutamatergic synaptic input that these areas receive.  4.3.6 Movement topography is preserved during blockade of intracortical synaptic transmission  We next examined the effects of pharmacological manipulations on the structure of motor maps evoked by brief (10 ms) pulses of light (Figure 4.11A,B). We had initially hypothesized that blocking cortical glutamatergic transmission would eliminate the contribution of facilitatory cortico-cortical projections from regions lacking direct motor output, causing a reduction in map area. Surprisingly, we found that Mab and Mad maps tended to increase in amplitude (Figure 4.11B) and expand in area (Figure 4.11C) after application of glutamate receptor antagonists, compared with no change after application of saline vehicle. This expansion in map area was also apparent in the hindlimb motor representation (134 ± 77%, p = 0.02, n = 9, paired t-test), but the expansion was most pronounced in Mad (Figure 4.11C). The region of overlap between abduction and adduction representations increased in the presence of glutamate receptor antagonists, but was unchanged by application of saline (Figure 4.11D). Because of its influence on map area (Figure 4.8A,B), stimulus intensity was held constant within animals for all pharmacology experiments.   Despite the fact that glutamate receptor antagonists caused map expansion and increased overlap between Mab and Mad, movement topography was not abolished. The Mab and Mad  maps could still be distinguished in the presence of glutamate receptor antagonists  97 (Figure 4.11B), with no reduction in the separation between their centers of gravity (Figure 4.11D). Application of glutamate receptor antagonists did not cause a significantly greater shift in map centers from their baseline positions than application of saline for Mab (0.5 ± 0.09 vs. 0.5 ± 0.1 mm respectively, p = 0.96, n = 9 vs. n = 5, t-test) or Mad (0.5 ± 0.09 vs. 0.2 ± 0.04 mm respectively, p = 0.06).   Although the increased movement durations (Figure 4.10C) and expansion of motor maps (Figure 4.11C) caused by disruption of excitatory synaptic transmission were unexpected, this may be explained by a loss of disynaptic inhibition (Kapfer et al., 2007; Silberberg and Markram, 2007; Helmstaedter et al., 2009; Murayama et al., 2009; Adesnik and Scanziani, 2010). To test this hypothesis, we repeated these experiments with GABAA receptor antagonists (gabazine 1 µM n = 4 or picrotoxin 100 µM n = 2, Figure 4.12). GABA receptor antagonists diminished differences between Mab and Mad movement trajectories, but had no effect on movement kinematics (Figure 4.12), and generally did not degrade functional subdivisions of the motor cortex. Disrupting GABAergic transmission did reproduce the increases in map amplitude (Figure 4.13C) and area (Figure 4.13D) seen during blockade of excitatory transmission. As with the delayed increase in movement speeds (Figure 4.10C), this effect was restricted to Mad. These effects are consistent with disinhibition causing the selective expansion of the Mad  subregion. The separation between Mab and Mad and the region of overlap between them was unchanged (Figure 4.13E). Like glutamate receptor antagonists, GABA receptor antagonists did not cause greater displacement of map centers than saline treatment for Mab (0.6 ± 0.1 vs. 0.5 ± 0.1 mm, p = 0.37, n = 6 vs. n = 5, t-test) or Mad (0.4 ± 0.1 vs. 0.2 ± 0.04 mm, p = 0.24).  98  Figure 4.10 Glutamate receptor antagonists degrade the differences between complex movements evoked by prolonged stimulation of Mab and Mad A Mean trajectories of movements evoked by stimulation (100 ms train of 5 ms pulses at 100 Hz) of Mab (green traces) and Mad (red traces) after application of CNQX and MK-801 (4.5 and 0.3 mM, respectively, right) or saline (left) to the surface of the sensorimotor cortex. B Plots of angle from the start point for the movement trajectories shown in A (see compass in A). Saline-treated control animals (left) displayed movement trajectories that were dependent on stimulus site (Finteraction (1,44) = 3.59, p < 0.001).  Glutamate receptor antagonists degraded the differences between Mab and Mad movements and biased both toward medial rotation (right) Finteraction (1,44) = 0.47 p=0.9984 (see also Figure s6). Dashed blue lines above the abscissae denote the period of stimulation. C Plots of change in speed for the post-treatment movements shown in A (pre-treatment speed profiles subtracted). There was no effect of saline application (left), but glutamate receptor antagonists caused a site-specific increase in delayed movement speed (right, Finteraction (1,44) = 2.079, P < 0.0001, RM-ANOVA, n = 7). GABA receptor antagonists similarly altered movement trajectories, but not kinematics (see figure s7). Error bars in all graphs are SEM.  99  Figure 4.11 Glutamate receptor antagonists cause map expansion without abolishing movement representations A Timeline for pharmacology experiments. Baseline maps were generated before applying either CNQX and MK801 or saline to the cranial window. Post-treatment mapping began after a 30 minute incubation period. B Representative movement maps from two different animals before (left) and after (right) pharmacological treatment. Compared with application of saline (top), incubating the cortex with CNQX and MK-801 (4.5 and 0.3 mM, respectively, middle) caused an enlargement both the Mab and Mad representations relative to baseline, but did not cause them to merge (see also Figures s6, s7). Scale bar at left applies to both maps. C Quantification of increases in map area after application of glutamate receptor antagonists or saline * p < 0.05, ** p < 0.01, paired t-test against baseline values). The number of animals per condition is marked for each group. D The region of overlap between Mab and Mad (yellow pixels in map) increased in the presence of glutamate, but separation between the centers of Mab and Mad was unchanged. Error bars in all graphs are SEM.  100  Figure 4.12 Effect of reduced GABAergic intracortical transmission on movements evoked by prolonged stimulation A Mean trajectories of movements evoked by stimulation (100 ms train of 5 ms pulses at 100 Hz) of Mab (green traces with square markers at 10 ms intervals) and Mad (red traces, round markers) before (left) and after application of gabazine (right). B,C Speed plots for the same movements. Error bars in all graphs are SEM.  101  Figure 4.13 GABA receptor antagonists cause selective expansion of Mad without abolishing movement representations A Representative movement maps from the same animal generated before (top) and after (bottom) pharmacological treatment with gabazine show expansion of Mad. B Timeline for pharmacology experiments. Baseline maps were generated before applying gabazine (1 µM, n = 4), picrotoxin (100 µM, n = 2) or saline to the cranial window. Post-treatment mapping began after a 30 min incubation period. C Quantification of increases in map amplitude and area (D) after application of GABA receptor antagonists or saline (* = p < 0.05, paired t-test against baseline values). The number of animals per condition is indicated for each group. E The mean regions of overlap and separation between the centers of Mab and Mad were unchanged.  4.3.7 Topical application of glutamate receptor antagonists disrupts cortical input without preventing direct activation of ChR2-expressing output neurons  The observation that disrupting intracortical synaptic transmission can impair the expression of diverse complex movements without abolishing the topography of movement maps was initially surprising, but may be explained by differences between the roles of intracortical and corticofugal circuits. It is possible that cortical application of receptor antagonists interferes with local circuit function and the generation of complex movements by prolonged stimulation, but does not alter the movement maps generated by the output of corticofugal cells directly activated by brief pulses of optogenetic excitation. To measure the effect of glutamate receptor antagonists on cortical activity evoked by ChR2 stimulation, we  102 recorded local field potentials (LFPs) in all cortical layers using a multi-channel probe (Figure 4.14). These recordings confirmed that glutamate receptor antagonists blocked synaptic input to the cortex driven by electrical stimulation of the contralateral forelimb. Glutamate receptor antagonists did not block direct activation of ChR2, but they did cause a decrease in delayed, presumably synaptic, components (Figure 4.14A). This effect was evident at all depths recorded (Figure 4.14B), but may have been primarily due to inactivation of the upper cortical layers, where drug concentrations are expected to be highest after topical application. Because optogenetic stimulation of ChR2-expressing neurons does not require synaptic activation, corticofugal neurons could still propagate their action potentials beyond the influence of the cortically applied glutamate receptor antagonists to evoke movements.  103   Figure 4.14 Glutamate receptor antagonists block cortical synaptic transmission but not direct activation of ChR2-positive neurons A Representative local field potentials (LFPs, mean of 50 trials) evoked by electrical stimulation of the contralateral paw (left, stimulus time marked with dashed line) or by cortical stimulation with blue light (right). After application of CNQX and MK-801 (bottom), activity evoked by paw stimulation, but not ChR2 stimulation, was blocked. B Mean LFP amplitudes recorded before (baseline, solid black lines) and after (CNQX+MK801, red lines) application of glutamate receptor antagonists. The peak-to-peak amplitude was measured in a time window 300 ms after paw stimulation (left) or ChR2 stimulation (right). After application of CNQX and MK801, LFP deflections evoked by paw stimulation were not greater than spontaneous fluctuations recorded in the absence of  stimulation (control, dashed black line; F(1,7) = 3.76, p = 0.06, RM-ANOVA, n = 7 mice). Conversely, ChR2-evoked LFP amplitudes were still present after application of CNQX and MK801 (F(1,7) = 25.78, p < 0.0001). Application of CNQX+MK801 caused a significant reduction in LFP amplitudes evoked by both paw (F(1,7) = 114.9, p < 0.0001) and ChR2 stimulation (F(1,7) = 18.29, p < 0.0001, RM-ANOVA, n = 7 mice). Error bars in all graphs are SEM.  104  Figure 4.15 Mab and Mad have adjacent, non-overlapping corticofugal projection pathways A Representative motor map generated through a thinned-skull preparation (left) to target the centers of gravity of Mab and Mad with injections of anterograde viral tracers (right). B Fibers from Mab (green) and Mad (red) in the dorsolateral striatum (left) and internal capsule (right) in another representative animal. C Magnified details of the inset sections above demonstrating that projections from Mab and Mad had little overlap. D Average distance from midline of peak fluorescence intensity for projections from Mab (green) and Mad (red) in the dorsolateral striatum (left, 1.97 + 0.11 mm vs. 2.49 + 0.07 mm, p = 0.03, t-test, n = 7 mice) and internal capsule (right, 1.92 + 0.07 mm vs. 2.26 + 0.02 mm, p = 0.03, t- test, n = 7 mice). Note that images were rotated until the midline was vertical before quantification. Error bars in all graphs are SEM.  105   4.3.8 Divergent projections from Mab and Mad  The fact that cortical application of glutamate receptor antagonists does not abolish movement topography (Figure 4.11) or prevent direct activation of corticofugal ChR2- expressing neurons (Figure 4.14) suggests that cortical output circuits may differentiate the Mab and Mad subregions.  To test this hypothesis, we injected the deep cortical layers of Mab and Mad with adeno-associated virus containing fluorescent marker constructs to label axonal projections throughout the brain (Figure 4.15A). In addition to reciprocal intracortical projections between these regions and trans-callosal projections to homotopic sensorimotor cortex, we observed adjacent, non-overlapping projections in the striatum and internal capsule (Figure 4.15B,C), with fibers originating in Mab occupying positions medial to those from Mad  (2.0 ± 0.1 vs. 2.5 ± 0.07 mm from midline in the dorsolateral striatum, p = 0.03, n = 7, paired t-test; Figure 4.15D). This observation further supports the hypothesis that movement map topography is a product of the pattern of corticofugal projections, whereas the generation of complex movements by prolonged stimulation requires input from recurrent intracortical circuits and/or loops with subcortical structures.  4.4 Discussion   We have applied light-based motor mapping to reveal that the mouse forelimb motor cortex is subdivided into distinct movement representations. Prolonged stimulation of these regions drives movements with similar speed profiles, but which terminate at different positions in space. Although complex movements evoked by prolonged stimulation were sensitive to perturbations of intracortical synaptic transmission, the topography of movement direction was not abolished by blockade of either excitatory or inhibitory synaptic transmission. The persistence of movement topography in spite of disrupted intracortical synaptic transmission may be due to the presence of segregated corticofugal pathways from the two movement representations.  106  4.4.1 Mechanistic basis of multiple motor representations  Functional differences between movement representations are likely the product of both their intracortical circuits (Jacobs and Donoghue, 1991; Rouiller et al., 1993) and their corticofugal pathways (Brown and Hestrin, 2009; Rathelot and Strick, 2009). The recurrent circuitry of the neocortex (Douglas and Martin, 2004; Hooks et al., 2011) provides computational power and allows flexible control of the more stereotyped connections between the spinal cord and the periphery. We have shown that the ability of prolonged cortical stimulation to generate complex movement patterns depends upon these intracortical circuits, and can be blocked by pharmacological manipulations. The contribution of recurrent cortical circuitry to movement representations is evidenced by their rapid modification in response to pharmacological manipulations (Jacobs and Donoghue, 1991) or inhibition of protein synthesis (Kleim et al., 2003a) and their rewiring after injury (Dancause et al., 2005). Expansion of representations after application of both glutamate and GABA receptor antagonists is presumably due to a loss of disynaptic inhibition, consistent with previous work (Jacobs and Donoghue, 1991; Aroniadou and Keller, 1993; Hess and Donoghue, 1994; Schneider et al., 2002; Foeller et al., 2005). The critical role of inhibitory circuits in cortical function and the profound change in brain state induced by application of GABA receptor antagonists complicates interpretation of our GABA experiments, but it is interesting to note that the effects of this manipulation were relatively specific to the Mad representation (Figure 4.13).  Our observation that distinct cortical movement representations persisted after the pharmacological disruption of intracortical synaptic transmission suggests that the corticofugal projections made by these regions play a key role in shaping movement representations, as has been reported for the whisker motor pathway of mice (Matyas et al., 2010) and monkey motor cortex (Rathelot and Strick, 2009). Light-based motor mapping using line 18 Thy-1 transgenic mice (Ayling et al., 2009; Hira et al., 2009; Komiyama et al., 2010) is particularly well-suited to defining the contribution of corticofugal projections to  107 motor topography since layer 5b pyramidal neurons are stimulated preferentially (Yu et al., 2008; Ayling et al., 2009).  The macroscopic parcellation of motor cortex into functionally distinct zones is particularly intriguing given that neuronal response types appear to be intermingled at the cellular level in rodents (Ohki et al., 2005; Dombeck et al., 2009; Komiyama et al., 2010; Wang et al., 2011). This apparent paradox may be resolved if movement representations are emergent phenomena that only materialize at the population level (Georgopoulos et al., 1986; Wessberg et al., 2000). Alternatively, this observation could reflect important differences between the layer 2/3 cortical neurons studied in many imaging experiments and the predominantly layer 5b neurons stimulated in light-based mapping.  4.4.2 Movement representations in rodents and primates  Multiple motor representations of the rodent forelimb have previously been described as the caudal and rostral forelimb areas (CFA and RFA) (Neafsey and Sievert, 1982). Although Mab and Mad occupy the same cortical territory as mouse CFA and RFA (Tennant et al., 2010), important differences exist between them. First, Mab and Mad are contiguous and equal in area, whereas CFA is larger than RFA and they are separated by a representation of the neck (Tennant et al., 2010). Second, RFA is not apparent in all experiments or animals (Tennant et al., 2010), whereas Mab and Mad almost always co-occur. It is interesting to note that in rats, mapping with short stimulus durations produces maps that include RFA and CFA, whereas long (500ms) durations reveal maps containing movement representations similar to Mab and Mad (Ramanathan et al., 2006).  Primate motor cortex is commonly described as a hierarchical arrangement of primary motor cortex, premotor areas, and supplementary motor cortex where premotor areas can facilitate motor output from primary motor cortex (Cerri et al., 2003).  It has been suggested based on their connectivity that rodent RFA and CFA are homologous to premotor and primary motor cortex, respectively (Rouiller et al., 1993). Our observation that Mad  108 expands after application of GABA receptor antagonists but Mab does not suggests that these regions may be differentially regulated by feedforward or lateral inhibition. Coupled with the relatively longer latencies for movements evoked from the more caudal Mad region, this could be viewed as evidence for a hierarchical arrangement of mouse motor cortex.  Although intracortical connections are obviously critical for motor function, it is also known that multiple motor cortical regions project in parallel to the spinal cord (Rouiller et al., 1993; Dum and Strick, 2002). This implies that multiple motor regions can contribute directly to movement, and may not be arranged hierarchically (Graziano and Aflalo, 2007b). This view is corroborated by the results of our experiments with glutamate and GABA receptor antagonists, which demonstrated that the Mab and Mad representations could function independently after a diminution of intracortical synaptic transmission. If multiple motor regions do not form a hierarchical chain, they may instead encode various behaviors or postures (Graziano et al., 2002a, 2005). This is consistent with our observation that stimulation of Mab and Mad drives limb movements to different end positions in space. This result could be produced with optogenetic or electrical stimulation, suggesting that it is not an artifact of passive electrical current spread from the stimulation site (Strick, 2002). Although we sometimes observed movements that resembled locomotion (combined rhythmic movements of contralateral forelimb and hindlimb upon stimulation of Mad), or manipulation (stimulation of Mab generally caused elevation and medial rotation to bring the contralateral forelimb to a central position in front of the body, see Harrison et al. 2012 for videos), we chose to focus our analysis on basic measures of motor behavior, such as movement direction.  4.4.3 Comparison of optogenetic and electrical motor mapping  Our light-based motor mapping technique has been optimized for speed and simplicity (Ayling et al., 2009); hence measurements of limb movement were made in a single dimension during mapping, or in two dimensions for video analysis. ICMS has been optimized to resolve select movements of single joints (Burish et al., 2008; Chakrabarty et  109 al., 2009; Young et al., 2011), something that is not observed with our technique in its present form. As a consequence, we are overlooking some of the complexity of evoked movements during mapping, and it is likely that the mouse motor cortex could be subdivided more finely based on a more advanced quantitative assay. These disadvantages of light-based mapping are offset by its unique ability to rapidly, objectively, and non-invasively quantify motor output of a defined cell type across the entire sensorimotor cortex.   The spatial resolution of light-based mapping is determined by physical scattering of light and by active spread of excitation. The influence of these factors is apparent from the observation that motor map area is strongly related to both stimulus intensity (Figure 4.8) and anesthetic depth (Tandon et al., 2008). A further limit on spatial resolution could be imposed by the width of ChR2-expressing pyramidal neurons’ overlapping dendritic arbors. Although the lateral resolution of light-based mapping may limit our ability to define exact boundaries of motor representations, we are able to resolve functional subregions of the forelimb motor cortex and generate maps of the hindlimb motor cortex that are often less than a mm in diameter (Ayling et al., 2009). Furthermore, blocking the synaptic spread of activation does not decrease the size of motor maps, suggesting that active spread of excitation does not substantially degrade map resolution (Figure 4.11). It is interesting to note that although motor map area decreases with reduced stimulus intensity, distinct Mab and Mad representations persist and separation between them actually increases (Figure 4.8). Furthermore, the cortical area activated by optogenetic stimulation is estimated to be only modestly larger than for electrical stimulation based on intrinsic signal imaging (Ayling et al., 2009). This difference may be offset by the selective expression of ChR2 in corticofugal output neurons, which could avoid stimulating axons of passage.  Light-based mapping also benefits from advantages in sampling, since stimulation sites can be distributed uniformly, spaced densely, and sampled repeatedly to accurately define the center of a motor map. Despite the biophysical differences between optogenetic and electrical stimulation, light- based maps generally resemble motor maps produced by electrical stimulation (Ramanathan et al., 2006; Tennant et al., 2010). Movement trajectories characteristic of Mab or Mad could be evoked using electrical or optogenetic stimulation, suggesting that similar neuronal populations are recruited by these methods. This finding supports the ability of ICMS to  110 selectively target restricted ensembles of cortical neurons.  4.4.4 A rodent model of motor circuitry for complex movements   The ability to reproducibly evoke distinct complex movements from multiple cortical sites presents an opportunity to perform further investigations of motor circuitry in a widely used model organism. More importantly, it will allow the advantages of genetic engineering in mice to be applied to the problem of motor cortex function and organization, either for optical circuit analysis (Zhang et al., 2007b; Tian et al., 2009; Chow et al., 2010) or in the search for future treatments for movement disorders, cortical injuries, and paralysis (Hodgson et al., 1999; Dancause, 2006; Murphy and Corbett, 2009; Dawson et al., 2010; Vargas-Irwin et al., 2010).  111 Chapter 5: Longitudinal light-based mapping of vicarious function in sensorimotor cortex after targeted stroke in mice  5.1 Introduction  Recovery from stroke depends upon the ability of surviving brain areas to reorganize and compensate for the loss of damaged regions (Liepert et al., 1998; Chen et al., 2002; Murphy and Corbett, 2009). Cortical regions that are in close proximity to the stroke or are functionally related to the damaged region are well positioned for this type of vicarious function (Xerri et al., 1998; Carmichael, 2006; Dancause, 2006). For example, destruction of the mouse forelimb sensory cortex by targeted stroke can cause a new sensory representation to emerge in the territory normally occupied by forelimb motor cortex (Winship and Murphy, 2008; Brown et al., 2009).   The ability of motor cortex to adopt the functions of somatosensory cortex is perhaps explained by the fact that these regions are naturally overlapping and interconnected in rodents (Ayling et al., 2009; Harrison et al., 2012; Lim et al., 2012). It remains unclear, however, whether the motor cortex can maintain its primary role in addition to shouldering the computational burden previously carried by the somatosensory cortex. The annexation of motor cortex by new sensory representations may require the underlying circuitry to abandon their original function, causing the motor map to be displaced. This type of maladaptive reorganization has been proposed as a mechanism for the secondary deficits that appear several weeks after stroke in some patients (Witte and Stoll, 1997; Dancause, 2006; Rijntjes, 2006).   Cortical reorganization continues for months after stroke (Ward et al., 2003) and is best studied with frequent mapping in longitudinal experiments. In the motor cortex such experiments have been constrained by the limitations of electrical stimulation, which is typically performed at one or two time points (Eisner-Janowicz et al., 2008; Gharbawie et al.,  112 2008). We made use of a recently developed method of light-based motor mapping using transgenic Channelrhodopsin-2 animals (Arenkiel et al., 2007) that express a light-sensitive cation channel in cortical neurons (Boyden et al., 2005). Light-based motor mapping has the advantages of being faster and less invasive than electrode based mapping and is easy to combine with intrinsic signal imaging (Grinvald et al., 1986) of somatosensory representations in thinned skull (Hira et al., 2009; Drew et al., 2010) or chronic cranial window preparations (Trachtenberg et al., 2002; Ayling et al., 2009). Here, we present the first longitudinal study of combined sensory and motor cortical reorganization after targeted stroke.  5.2 Methods  5.2.1 Animals and surgery  Animal protocols were approved by the University of British Columbia Animal Care Committee. Channelrhodopsin-2 transgenic mice (Arenkiel et al., 2007) from Jackson Labs (line 18, stock 007612, strain B6.Cg-Tg(Thy1-COP4/EYFP)18Gfng/J) established a breeding colony. Anesthesia was induced with isoflurane (1.5 % in air) and body temperature was maintained at 37˚C ± 0.2˚ C using a feedback-regulated heating pad. A craniectomy of approximately 5 mm by 5 mm was made over the right sensory-motor cortex using a dental drill. The craniectomy was sealed with a coverslip and dental acrylic (Trachtenberg et al., 2002; Harrison et al., 2009). These mice were allowed to recover for two months before being used in mapping experiments. The majority of cranial windows remained viable for the full extent of the experiment (17 of 24 mice), but in cases where the windows became unsuitable for intrinsic imaging, motor mapping was also terminated and no additional data were collected after that time point.  A total of 13 male and 11 female mice were used in the study.    113 5.2.2 Intrinsic optical signal sensory mapping  For details on intrinsic optical signal sensory mapping, refer to the methods paper published by our group (Harrison et al., 2009). In brief, mice were anesthetized with isoflurane (1% in air) and their heads immobilized in a stereotax while maintaining body temperature at 37 oC.  After capturing an image of the cortical surface under 525 nm green light, we switched to 635 nm red light and the focus was moved to ~200 µm beneath the cortical surface to obscure artefacts from surface vessels. Images were acquired through a video macroscope (4.3 mm field, 8.4 µm per pixel) using a Dalsa M-60 camera and XCAP software (EPIX). We completed 20-40 imaging trials per experiment, with each trial consisting of 15 frames collected over 1.5 s preceding a tactile stimulus delivered by a piezoelectric device (1 s of 5 ms square pulses at 100 Hz) and 15 frames collected during and after the stimulus. Trials of contralateral forelimb and hindlimb stimulation were alternated, with 10 s between trials. Images were analyzed using an ImageJ plugin described previously (Harrison et al., 2009) to create an image of percentage change in 635 nm light reflectance. We applied a threshold at 66 % of maximal response amplitude to calculate the area of sensory representations and produce color-coded maps of forelimb and hindlimb sensory cortex that could be overlaid onto the image of the cortical surface captured under green light. For analysis of sensory response profiles before and after stroke, 1mm diameter ROIs were placed in non-overlapping positions defined by the baseline positions of sFL and mFL to plot percent change of reflectance over time. IOS maps could not be obtained at all time points; this is reflected in the number of mice included for each measurement.  5.2.3 Light-based motor mapping  Light-based mapping methodology has been described in detail (Ayling et al., 2009). Briefly, we used a scanning stage (ASI MS-2000) controlled by custom Igor Pro software (Wavemetrics) to direct a fixed 473 nm laser beam (Crystalaser, focused to 100 µm diameter, 10 ms pulses, 1-5 mW total or 127-635 mW/mm2) to an array of cortical sites (typically 13 x 13, with 300 µm spacing between sites). This process was repeated 3 times to obtain a mean value for each pixel of the map. Stimulation was delivered in a semi-random order with  114 identical stimulus intensity for all sites within a map, with the requirement that sites within 750 µm of each other could not be stimulated consecutively. At each site, 500 ms of baseline movement data were collected before a 10 ms pulse was delivered and then an additional 990 ms of post-stimulus data were recorded. Movements were measured using laser range finders with µm sensitivity targeted to the forelimb and hindlimb (Keyence LK-081). Bracelets made of surgical tubing with a 12 mm diameter round coverslip glued to their sides were placed over the contralateral paws to provide a large, flat target for the laser range finders and to suspend the limbs slightly above the ground so that they could move freely along the axes of measurement of the range finders. In order to exclude artifacts (e.g. from breathing or electrical noise), responses were considered to be genuine only if their amplitude exceeded three times the standard deviation of the 500 ms pre-stimulus period within 100 ms after stimulus onset. A single injection of ketamine/diazepam (100/10 mg/kg) was administered at the beginning of the experiment. To minimize the possibility of anesthetic complications, ketamine/diazepam was not supplemented during mapping. If necessary, isoflurane (0.2- 0.5% in air) was administered to prolong anesthesia until completion of the map. In some cases, mapping was terminated before three map repetitions were completed. Motor maps with a mean amplitude of less than 0.1 mm from the 9 pixels closest to the centre of gravity were excluded from further analysis.  5.2.4 Photothrombotic stroke  To generate photothrombotic strokes, mice were injected with 1% Rose Bengal in phosphate-buffered saline (PBS, 100 mg/kg intraperitoneal) (Watson et al., 1985). A cortical area 1 mm in diameter was illuminated with the arc lamp of an epifluorescence microscope (10 mW green light, 10x objective, numerical aperture = 0.3) for 13 minutes. Strokes were targeted either to sFL (sensory stroke group) or a region of mFL separate from sFL (motor stroke group) as defined by baseline intrinsic signal sensory mapping and light-based motor mapping. Sham mice were injected with saline only and illumination was targeted to sFL.   115 5.2.5 Histology  After being deeply anesthetized with sodium pentobarbital, mice were transcardially perfused with 10 mL of 0.1 M PBS and then 10 mL of 4% paraformaldehyde in PBS. The brain was dissected, immersed in 4% paraformaldehyde for 24 hours, and then transferred to 30% sucrose in PBS for 24 hours. A vibratome (Leica) was used to cut 100 µm sections, with every third section mounted for epifluorescence imaging. Infarct volumes were calculated by identifying the section where the area of the infarct was largest, and multiplying this value by the anterior-posterior extent of the infarct.  Figure 5.1 Overview of experimental design A Mice aged at least two months were implanted with a cranial window and allowed two months recovery. Motor and somatosensory maps of the contralateral forelimb were generated before and after a photothrombotic stroke was targeted to the motor or somatosensory forelimb map. Eight weeks after stroke the mice were anesthetized and transcardially perfused and their brains recovered for histology. B Schematic diagram of motor and somatosensory forelimb maps overlaid onto an image of the cortical surface acquired through a cranial window. Note that because the skull over the midline was removed during cranial window implantation, bregma was not visible in any of the animals and its position here is estimated.  116  5.3 Results  5.3.1 Longitudinal light-based mapping of sensory and motor forelimb representations  24 Thy1-Channelrhodposin-2 transgenic mice (JAX line 18, Arenkiel et al., 2007) were implanted with cranial windows that covered sensorimotor cortex of the right hemisphere (5 mm x 5 mm, extending 1 mm across the midline and 2.5 mm anterior and posterior from bregma). Mice were allowed two months to recover from cranial surgery before mapping experiments began (Figure 5.1). Mapping was conducted weekly, with motor maps collected one day after sensory mapping. Three baseline motor and sensory mapping sessions were performed for each animal. Contralateral forelimb motor (mFL) maps were spatially stable, with a mean weekly shift in map centre position of 0.42 ± 0.22 mm (n = 24 mice, all values ± SEM unless otherwise stated, centre position defined by centre of gravity). Contralateral sensory forelimb (sFL) maps exhibited a similar weekly shift in centre position during the baseline period (0.42 ± 0.09 mm, n = 14 mice). The centres of sFL and mFL were separated by a mean of 0.88 ± 0.08 mm (n = 17 mice).  One day after the third baseline mapping session, a photothrombotic infarct was targeted to either sFL (sensory stroke group, Figure 5.2a), or the portion of mFL that did not overlap with sFL (motor stroke group, Figure 5.2b). Control mice were injected with saline instead of the photothrombotic agent rose bengal (Watson et al., 1985) and illumination was targeted to sFL (sham group, Figure 5.2c). Photothrombotic infarcts were of comparable volume for the sensory and motor-targeted groups (1.56 ± 0.37 vs. 1.32 ± 0.40 mm3 respectively, p = 0.69, t-test), but sensory-targeted infarcts were located more laterally (2.71 ± 0.27 vs. 1.87 ± 0.15 mm from midline, p = 0.0194 t-test). Each group contained 4 male and 4 female mice (except motor stroke group: 5 males, 3 females). Age at the onset of the study was consistent between groups (sensory 148.2 ± 10.5 days, motor 139.1 ± 16.9 days, sham 133.75 ± 11.8 days, p = 0.72 ANOVA).  117  Figure 5.2 Representative examples of sensorimotor reorganization after stroke A Paired motor (top) and somatosensory (bottom) maps of the contralateral forelimb from one of three baseline timepoints (first column), followed by maps generated after sensory-targeted stroke. Pixels in motor maps reflect the amplitude of movement evoked at that stimulation site. Sensory maps are thresholded at 67 % of maximal intrinsic signal response (see Methods for details). At right is a coronal section demonstrating the extent of an infarct in somatosensory cortex. B Maps generated before and after a motor-targeted stroke and sham (C) stroke.   118  Figure 5.3 Spatial displacement of motor somatosensory maps after stroke A Displacement of the centre of the forelimb sensory map from its mean baseline position before and after strokes targeted to sensory cortex (blue), motor cortex (orange), or sham strokes (black). The same color convention applies to all subsequent figures and to the asterisks signifying probability values (2-way ANOVA F(2) = 6.4, p = 0.002, asterisks correspond to results of Bonferonni’s post- test). B Similar plot for displacement of forelimb motor maps. 2-way ANOVA F(2) = 9.572, p = 0.0002. C Mean positions of sensory and motor forelimb maps for the average of all time points before and after sham strokes (top left), strokes targeted to mFL (top right), or strokes targeted to sFL (bottom). The mean baseline position of each map is marked with an arrow pointing toward the mean post- stroke position and the location of the stroke is marked with a dashed circle. RM-ANOVA tests were performed for medial-lateral and anterior-posterior displacement of each map, asterisk indicates significant result of Bonferonni’s post-test for sFL displacement in the medial-lateral dimension after sensory stroke.  D Correlation between shifts in sFL and mFL. Error bars in this and all subsequent figures are SEM.   119 5.3.2 Spatial properties of sensorimotor reorganization  We first sought to determine whether destruction of sFL does indeed cause a new sensory map to emerge in mFL, and if so whether this causes the mFL map to displace from its original position. To address this question, we calculated the centre of gravity of the forelimb sensory and motor representations at each experimental timepoint to track spatial reorganization of sensorimotor cortex (Figure 5.3). Strokes targeted to sFL caused a shift in that representation (Figure 5.3a), with the post-stroke sFL map taking up position near the baseline location of the mFL map (Figure 5.3c). The shift of the sFL map into the mFL region did not cause significant displacement of mFL, however. Only strokes targeted directly to mFL caused the centre of that map to be displaced from its baseline position (Figure 5.3a). Unlike in the case of sensory stroke, the direction of mFL shift was not consistent across animals (Figure 5.3c). Strokes in mFL did not cause a subsequent shift of the neighbouring sFL map (Figure 5.3b,c). Sham strokes caused no spatial reorganization of sensorimotor cortex (Figure 5.3a-c).  Interestingly, more pronounced displacement of sFL, which tends to be into mFL territory, was correlated with increased displacement of mFL from its pre-stroke position (Figure 5.3d). The extent of mFL displacement after sensory stroke was approximately the same as after sham stroke, however (Figure 5.3d), meaning that spatial reorganization was effectively confined to the stroke-damaged functional area. Displacement of sFL, but not mFL was correlated with infarct volume after both sensory stroke (r = 0.82 and -0.26, p = 0.48 and 0.61 respectively, n = 6 mice) and motor stroke (r = 0.92 and -0.65, p = 0.003 and 0.11 respectively, n = 7 mice). Post-stroke displacement of sFL and mFL was not correlated with the extent of overlap between these maps, defined by the pre-stroke separation between their centres of gravity (sensory stroke r = 0.35 and -0.71, p = 0.50 and 0.11, respectively, n = 6 mice; motor stroke r = -0.16 and 0.19, p = 0.73 and 0.69, respectively, n = 7 mice). Both sFL and mFL returned toward their original location over time and within one month post- stroke displacement from baseline position was no longer significant (Figure 5.3a,b). This corresponds well with the time course of behavioural recovery seen after photothrombotic strokes of similar magnitude (Brown et al., 2009; Sweetnam et al., 2012).  120   Figure 5.4 Sensory responses after stroke Each panel contains intrinsic signal responses to stimulation of the contralateral forelimb at time points before and after stroke, with the time point indicated by the colour of the line. The left column contains responses from a ROI defined by the mean pre-stroke position of sFL, and the right column contains responses from a non-overlapping ROI defined by the mean pre-stroke position of mFL. Asterisks indicate probability values from Bonferonni’s post-test from a 2-way ANOVA of each post- stroke timepoint against the baseline.  5.3.3 Changes in sensorimotor excitability after stroke  To assess the responsiveness of the sensorimotor cortex to incoming somatosensory stimuli, intrinsic optical signals (IOS) were quantified in non-overlapping regions of interest defined by the baseline positions of sFL and mFL. Applying a 1 s 100 Hz vibrotactile stimulus to the contralateral forepaw caused an IOS response in sFL and, to a lesser extent, mFL. These sensory responses were unaffected by sham or motor-targeted strokes (Figure 5.4). Strokes targeted to sFL, however, caused a persistent deficit in sensory responses within somatosensory cortex (Fig 5.4, upper left panel). Responses to sensory stimulation were initially disrupted in mFL, but returned after 6-8 weeks.  121  Figure 5.5 Regional changes in motor excitability after stroke To generate each panel, post-stroke maps were normalized to their baseline average, and then maps from multiple animals were aligned according to the location of the infarct and averaged. Finally, the aligned maps were averaged across two consecutive timepoints. The colorscale indicates the mean percentage change in movement amplitude relative to baseline, with red pixels marking regions of increased motor output after stroke (i.e. sites that produce larger movements upon stimulation).  To examine the effect of targeted stroke on motor representations, each animal’s post- stroke motor maps were normalized to their baseline mean and then aligned according to the position of the infarct and averaged. Sensory-targeted strokes caused a modest increase in motor output posterior to the infarct and a decrease anteriorly (Figure 5.5). Strokes in mFL caused a decrease in motor output from the infarct core (Figure 5.5), but this was balanced by  122 a substantial increase in peri-infarct motor output not seen after sensory-targeted or sham strokes (Figure 5.5). Plotting map area and overall motor output across the entire mapping window (Figure 5.6) revealed no significant differences between groups (2-way ANOVA p = 0.90 and p = 0.14, respectively). Normalized map area and motor output after stroke were not significantly correlated with infarct size (Sensory stroke r = 0.33 and 0.36, respectively, p = 0.59 and 0.55, n = 6 mice; motor stroke r = -0.55 and -0.42, p = 0.16 and 0.30, n = 7 mice).  By considering the changes in map position and motor output, differences between the reorganization of motor and somatosensory representations after stroke become apparent. When strokes afflict motor cortex, the loss of the infarcted region is countered by an increase in motor output from peri-infarct cortex (Figure 5.5), allowing the motor map to remain centred near its original location (Figure 5.3).  Motor-targeted strokes have little effect on neighbouring somatosensory maps in terms of their position and excitability (Figure 5.3, 5.4). Strokes targeted to sFL permanently destroy the ability of the infarcted area to respond to sensory stimulation (Figure 5.4), leading to the emergence of new sensory maps in motor cortex (Figure 5.2, 5.3, 5.4). These new sFL maps in motor cortex did not force the mFL map to abandon its original position, and sensory-targeted stroke had no significant effect on overall motor output.  123  Figure 5.6 Overall motor excitability is conserved after stroke A Motor map area, defined by the number of cortical sites capable of generating movement, is not significantly altered by stroke (2-way ANOVA p > 0.05, see Methods for details). B There were no significant group differences in normalized motor output (mean movement amplitude from all stimulus sites, 2-way ANOVA p > 0.05). Note that the two final time points (post-stroke weeks six and eight) are binned by averaging.   124 5.3.4 Effects of stroke on the integrity of motor representations   Although stroke targeted to motor cortex had no significant group effects on the size or overall excitability of motor maps, we observed that motor maps frequently became more diffuse after stroke (Figure 5.2). This effect was even more pronounced after stroke targeted to sensory cortex (Figure 5.7a). Because such changes in map structure may not be accurately reflected in a map’s centre of gravity (Figure 5.3) or overall motor output (Figure 5.6), we calculated a spatial correlation index (Figure 5.7b) for all motor maps. Motor maps were more diffuse after motor-targeted stroke and especially after sensory-targeted strokes, with a decrease in correlation between neighbouring pixels (Figure 5.7c,d). This decrease in local correlation was negatively correlated with infarct volume for sensory strokes (r = -0.96, p = 0.01, n = 6 mice) but not for motor strokes (r = -0.40, p = 0.33, n = 7 mice).  125   126 Figure 5.7 Motor maps develop a diffuse structure after stroke A Representative motor maps obtained before and six weeks after a stroke targeted to sFL. Note that after stroke, the map has a more scattered appearance, with more variability in amplitude between adjacent pixels. B Spatial cross-correlation of representative motor maps in A. The correlation value between any given pixel and its neighbours is plotted as a function of distance for the pre-stroke (black) and post-stroke (red) maps, revealing a decrease in positive correlation between nearby pixels after stroke. C Mean spatial correlations for each group. Bonferroni’s post-test values marked for 2-way RM-ANOVA. D Changes in correlation strength between neighbouring pixels after stroke. Both sensory and motor-targeted stroke caused lasting decreases in local correlation (F(2) = 23.9, p < 0.0001, asterisks indicate probability values from Bonferonni’s post test against the sham group). E Plots of cumulative motor output in a single mapping session. Each dashed line represents an individual animal’s mean from timepoints before (black lines) and after (red lines) sensory-targeted stroke. Thick, solid lines represent the group mean.  Motor maps are generated by stimulating various sites in a random spatial order, which raises the possibility that decreased spatial correlation after stroke is caused by alterations in the temporal properties of motor output within an experiment. This is not the case, however, because performing linear regression on plots of cumulative motor output over the course of a mapping session (Figure 5.7e) revealed linear rates of motor output before and after stroke (sensory stroke r2 = 0.98 ± 0.01 vs. 0.99 ± 0.01 respectively, p = 0.38; motor stroke 0.99 ± 0.002 vs. 0.99 ± 0.001, p = 0.07; sham 0.98 ± 0.001 vs. 0.99 ± 0.002, p = 0.24, paired t-tests). There were no significant changes in the rate of motor output (sensory stroke = 0.05 ± 0.01 mm per stimulus pre-stroke vs. 0.04 ± 0.01 post-stroke, p = 0.65; motor stroke 0.02 ± 0.004 vs. 0.02 ± 0.005, p = 0.68; sham 0.04 ± 0.01 vs. 0.03 ± 0.007, p = 0.31, paired t-tests). Although changes in the rate of motor output did not reach statistical significance, they were strongly correlated with infarct volume after sensory stroke (r = 0.96, p = 0.003, n = 6 mice). The diffuse structure of motor maps after stroke may be a manifestation of the ongoing reorganization of the underlying cortical circuitry, as the process of accommodating a new sFL map forces mFL to either devote its neurons to a hybrid sensory/motor role or to become a mosaic of intermingled motor and sensory neurons (Figure 5.8).  127  Figure 5.8 Model of cortical plasticity underlying sensorimotor map reorganization A Mouse sensorimotor cortex (box outlined in coronal section at left) is schematized as overlapping populations of motor (orange) and sensory (blue) neurons (right). B After sensory stroke, the sensory region is initially destroyed (left), but the overlap region (represented with mixed orange/blue neurons) survives and expands to form the new sFL representation in motor cortex (right). C Strokes in motor cortex do not cause the motor map to expand into sFL. Instead, an increase in the excitability of peri- infarct neurons (indicated with bold lines) compensates for the partial loss of mFL. 5.4 Discussion   We have exploited the development of a new method for light-based motor mapping (Ayling et al., 2009) using Channelrhodopsin-2 transgenic mice (Arenkiel et al., 2007) to perform the first longitudinal study of sensorimotor reorganization after targeted stroke. We first attempted to confirm previous observations of reorganized sFL maps emerging in motor cortex after stroke in somatosensory cortex. We found that displaced sFL maps do tend to  128 shift toward the center of the mFL map, causing secondary displacement of mFL that was strongly correlated with the extent of sFL shift. The displacement of mFL maps after sensory stroke was not statistically significant, however, and was less than the map displacement seen after strokes within mFL itself. Neither sensory nor motor-targeted strokes caused significant changes in motor map size or excitability, perhaps because hyper-excitable peri-infarct (Figure 5.5) cortex began to counteract decreases in excitability in the infarcted region as soon as one week after stroke (Fujioka et al., 2010). Motor maps maintained their size and position after stroke, but their structure was altered as evidenced by a decrease in local spatial correlation. This effect was especially pronounced after sensory-targeted stroke, suggesting that the ability of motor cortex to host new sensory representations may come at the cost of alterations in the motor cortical network.  5.4.1 Multifunctional peri-infarct cortex   Remapping of cortical function is closely related to behavioural recovery (Cramer et al., 1997; van Meer et al., 2012). In particular, recovery tends to be best in patients (Johansen-Berg et al., 2002a, 2002b; Fridman et al., 2004; Cramer et al., 2006) or animal models (Dijkhuizen et al., 2001, 2003; van Meer et al., 2012) where reorganization occurs primarily within the perilesional cortex of the stroke affected hemisphere. The peri-infarct cortex exhibits heightened neuronal sprouting and angiogenesis after stroke (Carmichael et al., 2005; Hossmann, 2006; Ohab et al., 2006; Brown et al., 2007) and is believed to play a central role in remapping and recovery (Witte, 1998; Johnston et al., 2012). We observed that both motor and somatosensory maps initially displaced from their original location returned toward their baseline position within two months, typically occupying the peri-infarct region. After strokes in motor cortex, the peri-infarct cortex became hyperexcitable, generating larger movements upon stimulation than were observed during the pre-stroke baseline period. This effect was specific to the peri-infarct cortex and was not observed in control animals, suggesting that it was not caused by repeated brain stimulation. Increased excitability and synchronous activity in peri-infarct cortex may enhance plasticity and cortical reorganization after stroke (Witte, 1998; Carmichael and Chesselet, 2002; Carmichael, 2012). In contrast to the increased peri-infarct excitability seen after strokes in motor cortex, sham-operated mice  129 exhibited a uniform decrease in motor excitability throughout the map area (Figure 5.5). This could be due either to changes in the opacity of the cranial window over time or to the effects of repeated anesthesia and stimulation of the periphery (during sensory mapping) or cortex (during motor mapping). The decreased excitability of sham mice makes the peri-infarct hyperexcitability seen after motor stroke even more striking. This is the first study of its kind involving longitudinal light-based motor mapping and as such it will need to be compared with future experiments using alternate surgical preparations or stimulation parameters to further address this question.  5.4.2 Motor map structure after stroke   The diffuse structure of motor maps after stroke, evidenced by their decrease in local spatial correlation, has not previously been reported. Because motor maps are generated by stimulating an array of cortical sites in a random spatial order, it is possible that the changes in the spatial properties of maps could be caused by a change in the temporal properties of motor output. For example, if motor output was initially robust but quickly declined due to fatigue or other factors, the first sites to be stimulated would be over-represented in the motor map. The map would have a temporal structure, but the spatial structure would be disrupted. This does not appear to be the case however, since motor output is relatively constant over the course of a mapping session both before and after stroke (Figure 5.7). An alternative explanation is that the spatial structure of maps is altered by the incorporation of new regions of cortical output that were masked by inhibition prior to stroke (Perez and Cohen, 2009; Margolis et al., 2012). Map area remains constant after stroke (Figure 5.6), but this could reflect the addition of new, more distant regions to the map offsetting the loss of motor output from the area of the infarct (Figure 5.5).   Curiously, the disruption of motor map structure is most pronounced after strokes targeted to sensory cortex. This could be due to an expanded region of cortex devoted to mixed sensory/motor function (Figure 5.8). After stroke, neurons within this area may be more likely to perform a dual function (Winship and Murphy, 2008), or the region may  130 contain an intermingled mixture of neurons devoted solely to motor or sensory function. Either of these scenarios could account for the observation of diffuse motor map structure after stroke, but this hypothesis cannot be tested with the macroscopic mapping methods employed in this study. Because the methods used here are entirely light-based and minimally invasive, future studies could combine light-based mapping with imaging of microscopic cellular structure and function after stroke to gain additional detail (Mostany et al., 2010; Risher et al., 2012).  5.4.3 Longitudinal mapping of motor reorganization   Light-based motor mapping has been performed repeatedly within animals in the past (Ayling et al., 2009; Hira et al., 2009), but never in experiments lasting months before and after experimental manipulations such as stroke. We made several slight modifications to the mapping procedure originally published (Ayling et al., 2009) to optimize the method for longitudinal experiments. This included increasing intervals between mapping stimuli to 3s and substituting ketamine/xylazine anesthesia for ketamine/diazepam to avoid epileptogenesis (Hira et al., 2009). We also found that regrowth of bone at the edges of cranial windows was most pronounced along the medial edge of the window. Extending the cranial window across the midline (i.e. an asymmetric bilateral window) greatly reduced the problem of regrowth. The development of new methodology for longitudinal motor mapping will greatly facilitate studies of cortical plasticity after motor training (Kleim et al., 1998) or experimental lesions of the motor system (Ghosh et al., 2010). Future experiments will benefit from a combination of mapping and behavioural assessment. We chose not to add behavioural testing to the present study because of the likelihood that behavioural training and testing itself could alter sensory and motor map plasticity (Kleim et al., 1998; Barbay et al., 2012; Tennant et al., 2012), preventing an accurate measurement of the spontaneous capacity for cortical reorganization. It is known from earlier work, however, that photothrombotic strokes of this size targeted to somatosensory forelimb cortex cause impairment on the cylinder task, with recovery to baseline levels of performance within 8 weeks (Brown et al., 2009).   131 5.4.4 Light-based mapping as an assay for therapeutic and rehabilitative strategies  We have demonstrated the feasibility of longitudinal sensorimotor mapping and characterized the spontaneous cortical reorganization that occurs in the absence of any intervention. It will now be possible to test the efficacy of preventative, protective or rehabilitative therapies in the context of motor recovery while monitoring the organization of sensorimotor cortex (Biernaskie and Corbett, 2001; Kleim et al., 2003b; Hummel and Cohen, 2006). Ultimately, these optimized rehabilitation strategies could be translated to humans to enhance recovery from stroke and other forms of brain injury (Krakauer et al., 2012; Martin, 2012).  132 Chapter 6: General discussion  6.1 Strengths and limitations of light-based sensorimotor mapping  6.1.1 Intrinsic signal imaging   IOS imaging is effective as a simple, non-invasive means of mapping functional areas in cortex; its strengths relative to other imaging techniques derive from that simplicity (Grinvald et al., 1986). IOS imaging does not require the addition of any sensors or dyes such as are required for voltage-sensitive dye imaging, and it is far less expensive to perform than magnetic resonance imaging. Its disadvantages stem from the fact that imaging deoxygenation of the cortical vasculature is an indirect measure of neuronal activity (Frostig et al., 1990). The same disadvantage applies to functional magnetic resonance imaging, however. Voltage-sensitive dye imaging has unique caveats, as it reports subthreshold voltage changes from a variety of cells, including astrocytes (Chemla and Chavane, 2010). Compared with electrophysiological recordings performed at single points, IOS imaging is able to provide a measure of changes in activity across large cortical regions. IOS imaging is limited to the cortical surface and the information that it provides is far less detailed than that which can be obtained by intracortical electrophysiology, but it is less invasive and easier to perform repeatedly without causing damage to the cortex. The fact that IOS imaging detects changes in blood flow rather than neuronal activity per se is actually an advantage for certain applications, particularly in studies of the relationship between blood flow and neuronal function or cognition (Moore and Cao, 2008).  6.1.2 Light-based motor mapping   The standard approach for electrode-based motor mapping has traditionally been intracortical microstimulation (ICMS) (Stoney et al., 1968). This technique was developed to optimize the spatial resolution of mapping by employing minimal stimulus parameters. In contrast with the surface stimulation used to make earlier motor maps (Penfield, 1937),  133 ICMS makes use of sharp microelectrodes lowered to layer 5 of cortex, where the electrical current required to evoked movements is at a minimum (Young et al., 2011). ICMS motor mapping typically employs brief pulses of minimally suprathreshold stimuli, with the goal of minimizing passive current spread and restricting the area of activation to improve the spatial resolution of mapping. Because cortical points are stimulated individually, it is possible to identify the current threshold for each point, providing an additional layer of detail to the motor map (Tennant et al., 2010). This also allows the experimenter to carefully inspect and describe the nature of the movement evoked from each cortical point. The slow pace of ICMS is also one of its major disadvantages, however, as are its invasive nature and the subjective, qualitative method used to describe motor output.  One variation of electrode-based mapping that addresses some of these issues involves epidural stimulation through a thin-film electrode array (Hosp et al., 2008). Although epidural stimulation suffers from the disadvantages related to passive current spread discussed above, the use of an implanted electrode array permits rapid, automated stimulation of many cortical sites (~70 in the case of Hosp et al., 2008). An additional benefit is that the implanted epidural electrodes can be used to record sensory-evoked potentials, meaning that the same electrode array can be used for both motor and sensory mapping. This approach is also suitable for longitudinal experiments, although the implantation of the electrode array requires invasive surgery (Hosp et al. 2008). A disadvantage of the electrode array is that its opacity precludes imaging of the underlying cortex.  Compared with conventional forms of motor mapping, light-based mapping (Ayling et al., 2009; Hira et al., 2009; Komiyama et al., 2010) benefits from its minimal invasiveness, cell-type specificity, speed, and its automated, objective quantification of motor output. The application of optogenetics to motor mapping conferred the ability to target genetically defined populations of neurons. The ability to stimulate these cells using light makes it possible to perform mapping through a sealed cranial window or even the intact skull, making light-based mapping relatively benign and ideal for longitudinal experiments (Ayling et al., 2009; Hira et al., 2009). Compared with ICMS, light-based mapping is three orders of magnitude faster and can objectively quantify motor output (Ayling et al., 2009). A potential  134 disadvantage is its spatial resolution; the area activated by a laser stimulus is estimated to be approximately twice that activated by ICMS (Ayling et al., 2009). However, this disadvantage in spatial resolution may be offset by the ability of light-based mapping to selectively stimulate output neurons of motor cortex, avoiding axons of passage from other types of neurons. The rapid, automated nature of light-based mapping is an advantage in many ways, but it also hinders the careful examination of each individual stimulation site that characterizes ICMS mapping.   Some limitations of motor mapping apply equally to each of the above methods. For example, the vast majority of motor mapping experiments rely on ketamine to preserve motor tone that is lost under other forms of anesthesia (Kalkman et al., 1994; Simon et al., 2010). Ketamine is relatively short-acting, meaning that it must be supplemented in prolonged experiments. As a consequence, motor mapping experiments are performed with a fluctuating level of anesthesia, which is known to influence map area and structure (Tandon et al., 2008). In addition, mapping experiments only provide knowledge of the outputs from motor cortex that produce movement, despite the fact that the vast majority of neurons in the motor cortical circuit do not project to the spinal cord or brainstem motor areas (Keller, 1993). The full menagerie of motor cortical neurons can be studied using electrophysiology or imaging, but these methods have their own limitations. Electrophysiological recording limits the number of neurons that can be sampled, and tends to overemphasize neurons whose activity co-vary with motor behavior. Similarly, functional imaging detects active neurons, potentially overlooking the significance of silence in a population code (Pouget et al., 2000; Sanger, 2003). Finally, the contribution of recorded or imaged neurons to movement is always correlative, not causative.  6.1.3 Comparison to human non-invasive stimulation techniques  Non-invasive stimulation of the human brain has been practiced for more than 200 years in its simplest form, transcranial direct current stimulation (tDCS) (Nitsche et al., 2008). The technique requires nothing more than applying two electrodes to the skull and allowing current to flow from the cathode to the anode. The region of brain underlying the  135 cathode is inhibited, while anodal stimulation is generally excitatory. The placement of the two electrodes largely dictates the effects of stimulation and is therefore of the utmost importance in tDCS experiments (Nitsche et al., 2008). tDCS has been used both in investigations of motor physiology (Nitsche et al., 2005) and as an adjuvant to enhance rehabilitation after stroke (Hummel, 2005). Although tDCS is simple, inexpensive, and has the experimental advantage of bidirectional control of excitability, its spatial resolution limits its utility for applications such as motor mapping.  Transcranial magnetic stimulation (TMS) is a more recently developed form of non- invasive brain stimulation whose spatial resolution is considerably better than tDCS (Hallett, 2007). TMS involves passing a brief electrical current through a coil placed near the skull. The resultant magnetic field is transmitted through the skull and causes current to flow within the brain, depolarizing neurons near the cortical surface. Modifying the basic coil to a figure- of-eight shape increases the strength of the field at the center of the coil and yields more focal stimulation (Hallett, 2007). Like tDCS, TMS can be both excitatory or inhibitory. The effects of stimulation are strongly frequency dependent; repeated stimulation at rates of 5 Hz or higher are excitatory and at 1 Hz or less stimulation is inhibitory (Hallett, 2007). Patterned “theta burst” stimulation (3 pulses at 50 Hz separated by 200 ms) is excitatory if delivered intermittently, but inhibitory in continuous application (Hallett, 2007). TMS can be used to measure corticomotor conduction time, cortical excitability, or to generate motor maps. Paired pulses of TMS or pairing of cortical and peripheral stimulation can be used to measure forms of cortical plasticity such as intracortical inhibition and facilitation (Ziemann et al., 2008). Therapeutic applications of TMS have also been investigated, including the use of TMS to enhance recovery from stroke (Mansur et al., 2005; Fregni et al., 2006).  The spatial resolution of TMS is ~1 cm, making it suitable for motor mapping in humans (Huerta and Volpe, 2009). Unfortunately, the much smaller brains of rodents preclude the use of TMS in a spatially selective manner. One possible alternative is the use of ultrasound radiation for brain stimulation, which has a reported lateral resolution of 2 mm (Tufail et al., 2010). Optogenetic stimulation, which can also be delivered through the skull in a minimally invasive fashion, has been used in mice to reproduce findings from the human  136 TMS literature (Ayling et al., 2011). For example, short-latency afferent inhibition (SAI), initially described in humans (Tokimura et al., 2000), can be observed in mice stimulated with ChR2 (Ayling et al., 2011). Optogenetic stimulation has also been used to investigate other questions arising from clinical practice, such as the effects of deep brain stimulation on symptoms of Parkinson disease (Gradinaru et al., 2009; Kravitz et al., 2010). Far-fetched as it may seem, optogenetics could eventually be used to selectively activate or inhibit targeted circuits in the human brain to ameliorate neurological and psychiatric disorders.  6.1.4 Physiological consequences of light stimulation   The spread of activity through the cortical network following a single pulse of light stimulation is a matter of considerable interest to many experimenters and is largely an open question. We know that stimulation triggers a depolarization of the cortical electroencephalogram at short (< 1 ms) latency, and that the amplitude of this depolarization is relatively uniform throughout sensorimotor cortex (Ayling et al., 2009). Based on IOS imaging the spread of activation stemming from a single mapping stimulus can be estimated at ~700 µm, but this is an indirect measure (Ayling et al., 2009). The spread of subthreshold activity through the cortical network has also been examined with VSD imaging and found to flow preferentially between regions known to be anatomically interconnected (Lim et al., 2012).  Given the area of activation following a single mapping stimulus measured using IOS imaging (Ayling et al., 2009), it is possible to make an estimate of the number of neurons activated by each stimulus. The estimated density of neurons in layer 5 of mouse motor cortex is 5x104 per cubic mm (Schüz and Palm, 1989). If we assume that all of these neurons express ChR2 in the Thy-1 transgenic mice used in these experiments, and that all ChR2- positive neurons within the area of IOS activation are recruited ( 700 µm diameter), then we arrive at ~ 19000 neurons  activated. However, this is likely to be an over-estimate, since it is unlikely that all neurons in layer 5 express ChR2 or that every neuron within the area of IOS activation is activated by every stimulus. Unfortunately, however, the data necessary to improve this estimate are presently unavailable.  137 Phototoxicity is an obvious concern associated with optogenetic stimulation, but there is currently no post-mortem evidence of gross pathology in mice subjected to light-based motor mapping, nor is there any decrease in motor output within the course of a mapping experiment, even after thousands of stimuli (Ayling et al., 2009). It is known that prolonged stimulation can cause heating of brain tissue and introduce artifacts in functional magnetic resonance images, but the intermittent and spatially distributed pulses of light used in these experiments would be expected to cause negligible changes in brain temperature (Christie et al., 2012).  6.1.5 Future improvements for light-based mapping   Light-based mapping is a new technique and as such there are many possible means of improving it. New opsins are continually being engineered, with modifications being made to their activation spectra, intracellular trafficking, kinetics of activation and conductance (Gradinaru et al., 2010). The application of red-shifted opsins to light-based mapping, for example, could improve spatial resolution by employing longer stimulation wavelengths that can penetrate to the deep cortical output layers with less scattering (Bernstein et al., 2008). It would also permit two-color experiments, where separate populations of neurons could be simultaneously activated or inhibited in the same mapping experiment (Yizhar et al., 2011; Prigge et al., 2012). Additional advances could be made by extending light-based mapping to rats (Smith et al., 2012) or even primates (Gerits et al., 2012). The larger cortices of these animals would effectively improve the spatial resolution of motor mapping, and their increased behavioral repertoires would facilitate more sophisticated measurement of motor behavior.   It is also possible to use galvanometers to direct light to various cortical stimulation sites (Lim et al., 2012) rather than the scanning stage employed in the studies described here. This would avoid movement of the animal, which could potentially engage the vestibular system. It also permits light to be scanned across the cortex very quickly, permitting multiple sites to be targeted simultaneously with trains of stimulation. Alternatively, deformable  138 mirrors could be used for spatially patterned stimulation of multiple cortical areas (Rueckel et al., 2006).   Improvements to light-based mapping could  also be made in movement detection. Light-based mapping was initially performed using electromyography (EMG) to quantify movements evoked by stimulation (Ayling et al., 2009). EMG provides a sensitive readout of activity in identified muscle groups, but it is highly invasive and implantation of electrodes requires some surgical skill. For longitudinal experiments, EMG electrodes were replaced with non-invasive motion sensors (Harrison et al., 2012). These sensors do not provide the same level of physiological detail as EMG data, but they are simpler to use and have the useful benefit of reporting movement direction (Harrison et al., 2012). In order to be used effectively, however, the limb being assessed must be restrained so that it remains within the beam of the laser motion sensor. This restraint constrains the movements that can be expressed. It is also possible to detect stimulus evoked movements using magnets (Hira et al., 2009), accelerometers (Hosp et al., 2008), or video cameras (Harrison et al., 2012). Any of these methods, or some combination of them, could be developed to simultaneously detect unconstrained movements of multiple body parts in all three dimensions.  6.2 Functional organization of motor cortex  6.2.1 Comparison to other cortical regions   What cortical maps represent and why they should exist at all is a fundamental question that must be discussed in the context of brain mapping. Features map in sensory or motor cortex can be defined as the spatial clustering of neurons with similar functional properties. In primary visual cortex, these properties may be related to the receptive field of the neuron in retinotopic space or its tuning to the orientation of lines in the visual environment (Bonhoeffer et al., 1993). Primary auditory cortex contains clusters of neurons preferentially excited by a common frequency of sound (Schreiner and Winer, 2007).  Even outside the cortex, olfactory bulb glomeruli with sensitivity to similar odorant molecules  139 cluster together (Mori and Sakano, 2011). In the motor cortex there are clusters of neurons with similar projection identities (Rathelot and Strick, 2009; Harrison et al., 2012) or movement tuning (Amrikian and Georgopoulos, 2003; Dombeck et al., 2009). On a larger scale, entire cortical lobes (e.g. the occipital lobe) are collections of multiple subregions each devoted to some aspect of a common function (i.e. visual processing).  One explanation for clustering is that the smooth mapping of neuronal tuning across the cortex serves to decrease the distance between preferentially interconnected neurons with similar functional properties, thus reducing the wiring length of intracortical axons, minimizing cortical volume and axonal conduction time, and increasing the efficacy of signaling via diffusion of messengers (Koulakov and Chklovskii, 2001). Wiring optimization has been hypothesized as an explanation for many features of the nervous system, from the shape of dendritic arbors to the folds of the neocortex in gyrencephalic animals (Chklovskii and Koulakov, 2004).  The presence of multiple overlapping maps in regions such as primary visual cortex imposes additional constraints on functional organization. Each map must be arranged so that all feature combinations are represented at least once within each portion of retinotopic space, or perceptual blind spots would be introduced (Swindale, 1996). This principle of coverage, along with the continuity of feature representation imposed by the evolutionary pressure to minimize wiring length, may be applicable to all cortical maps. In motor cortex, self-organizing maps of complex movement can emerge from these principles along with a requirement for somatotopy (Aflalo and Graziano, 2006). When the multiple features believed to be encoded by motor cortex are reduced into two-dimensional space, these models reproduce many features of cortical motor maps, including repeated representations of the hand (Aflalo and Graziano, 2006; Graziano and Aflalo, 2007a). Despite similarities between motor maps and feature maps in sensory cortex, there remains the fundamental conceptual difference that sensory maps are defined by their inputs, whereas motor cortex lacks the granular input layer and is perhaps better understood in the context of its outputs.    140 6.2.2 Somatotopy as an organizing principle and its limitations   Somatotopy was one of the first features of sensorimotor cortex to be described (Penfield and Boldrey, 1937) and has been consistently identified as a key feature of these cortical regions. Although somatotopy is undeniably a key feature of cortical organization, it is only one of the organizing principles in motor cortex, perhaps akin to the role of retinotopy in visual cortex. Its importance may have been overstated due to the appeal of its pleasing simplicity, but the desire for a simple explanation to complex phenomena has long been recognized as a pitfall: “Nature seems unaware of our intellectual need for convenience and unity, and very often takes delight in complication and diversity.” (Ramón y Cajal, 1937)   Many investigators anticipated that the somatotopic organization of sensorimotor cortex would extend beyond the level of limbs to include ordered representations of individual joints, digits and muscles (Asanuma and Rosen, 1972). This is unlikely, however, given several key limitations to the principle of somatotopy (Schieber, 2001). Within the cortical region representing a body part such as the forelimb, representations of smaller structures are scattered widely and overlap extensively with each other. Output from multiple regions converges at single muscles (Dum and Strick 2005), whereas individual corticospinal axons diverge to target multiple spinal motoneurons (Shinoda et al., 1981). Furthermore, horizontal intracortical connections can extend for long distance to link disparate cortical regions, emphasizing that the motor cortex operates in a distributed and interconnected nature (Capaday, 2004). Taken together, these observations place a firm limit on the extension of somatotopy to the level of cortical columns, and rule out the interpretation of motor cortex as a mosaic of discrete muscle representations or a sort of piano keyboard upon which movements can be played out (Schieber, 2001).  6.2.3 What does the motor cortex encode?   The debate over the importance of somatotopy in motor cortex is one facet of the larger discussion about the function of motor cortex and the parameters that are encoded in its neurons. This debate is commonly reduced to “muscles vs. movements”, despite the  141 recognition on both sides that this is an oversimplification (Graziano et al., 2002b). Essentially, the controversy stems from disagreements about the relative importance of higher level kinematic and lower level kinetic parameters in motor cortex. Evidence for both have been uncovered by experiments involving stimulation or recording of motor cortical neurons (Kakei et al., 1999; Ramanathan et al., 2006).   Recording from single neurons in the motor cortex of behaving primates has revealed that neuronal firing rates are tuned to both movement (Georgopoulos et al., 1982, 1986) and to the direction of force generated in the absence of movement (Kalaska et al., 1989). Stimulation of motor cortex has been interpreted to produce twitches in single muscles or rotation of specific joints, but other experimenters report complex, behaviorally relevant movements (Graziano et al., 2002). These discrepancies result in part from differences in experimental design, but they also suggest that the motor cortex operates as a distributed network which simultaneously encodes both kinetic and kinematic parameters as it optimally controls ongoing motor behavior (Scott, 2004). Rather than examining the activity of individual neurons while animals perform unnatural behaviors, it may prove more informative to study cortical activity at the population level during learning and execution of motor tasks (Komiyama et al., 2010).  6.2.4 Relationship between macroscopic maps and neuronal circuits   The nature of the microcircuit mechanisms that differentiate the multiple maps within a somatotopic region, such as abduction and adduction subregions of mouse forelimb motor cortex, remains an open question in the field. It is possible that these subregions arise from the pattern of their corticofugal output, from the structure of the inputs that they receive from other regions or from recurrent intracortical circuitry, or from some combination thereof (Harrison et al., 2012).   Recurrent circuits provide computational power and flexibility to the cortex. Their contribution to movement representation is apparent from the rapid alterations in motor output seen after pharmacological manipulations of the cortex (Jacobs and Donoghue, 1991;  142 Harrison et al., 2012). The spatial organization of various classes of output neurons (Rathelot and Strick, 2009) and the preservation of movement representations after blockade of intracortical glutamate receptors (Harrison et al., 2012) demonstrate that patterns of corticofugal output also play a key role in the functional organization of motor cortex. The introduction of new techniques for labeling neurons according to their projection identity will make it possible to identify the relationship between microcircuits and macroscopic maps in motor cortex (Figure 6.1).  143   Figure 6.1 Strategies for targeting neurons in motor cortex based on their projection identity A Injecting a retrograde tracer or viral vector into a region downstream of motor cortex (for example, the spinal cord) will label neurons in motor cortex that project to that structure. B By expressing optogenetic activators such as Channelrhodopsin-2, corticospinal neurons can be selectively activated while resulting movements are measured. Motor maps generated by stimulating specific projection classes can be tested for topographies of evoked movement direction and compared with maps from other projection classes. C Alternatively, the activity of labeled neurons can be measured while an experimental animal makes voluntary movements in different directions. This could be accomplished either with electrical recordings or by imaging genetically encoded calcium or voltage indicators. Again, preferred movement directions can be examined at different cortical locations for neurons in a given projection class, and different projection classes can be compared by performing retrograde injections in multiple structures targeted by cortical output neurons.  144 6.3 Sensorimotor reorganization after stroke  6.3.1 The brain’s innate capacity for repair after brain injury   The damage wreaked by stroke can only rarely be avoided in the acute phase, making neural repair the only recourse for recovery of function. Stroke is the leading cause of adult disability, and it can be expected to exact an increasing toll on society as the population ages (Carlo, 2009). Most patients do experience an improvement in function over the months after stroke, although it is often unclear whether this represents true recovery or behavioral compensation, defined as a reliance on alternative strategies to accomplish a motor task (Murphy and Corbett 2009). One example of behavioral compensation is increased use of an unaffected limb to perform tasks that were previously accomplished with the stroke-affected limb (Murphy and Corbett, 2009). In contrast, true recovery requires renewed use of the circuitry and behaviors employed prior to the injury. If the relevant circuitry was destroyed rather than being transiently impaired by inflammation or diaschisis (Carmichael et al., 2004)2, true recovery may be impossible without neurogenesis or the further development and successful implementation of fledgling technologies such as stem cell therapy (Chen et al., 2001; Lichtenwalner and Parent, 2006).   Luckily, the brain has a natural ability to reorganize itself that is enhanced in the peri- infarct region after stroke (Murphy and Corbett, 2009). The peri-infarct cortex is profoundly influenced by the trauma in neighbouring regions, and undergoes a host changes that include alterations in the expression of growth-related genes (Carmichael et al., 2005) and increased angiogenesis (Ohab et al., 2006). There are also profound effects on synaptic function, marked by changes in GABAergic tone (Clarkson et al., 2010) and increased turnover of dendritic spines (Brown et al., 2007). Together with extensive axonal sprouting, these changes likely underlie the large-scale modification of cortical maps seen after stroke (Dancause et al., 2005).      145 6.3.2 Therapeutic strategies for enhancing recovery from stroke   Many therapeutic strategies have been devised to augment the spontaneous recovery that occurs after stroke. One possibility is to replace the cells killed by stroke by injecting stem cells into the brain or increasing the rate of neurogenesis (Chen et al., 2001; Lichtenwalner and Parent, 2006). The major impediment to this approach is that newly-born neurons are unlikely to be effectively integrated into cortical circuits. It is possible, however, for stem cells to secrete messengers or growth factors that could improve the function of surviving neurons (Barkho and Zhao, 2011). Encouraging surviving neurons to elongate their axons and form new connections may be another way to foster the creation of new circuits capable of performing functions impaired by stroke (Zhang and Chopp, 2009; Li et al., 2010).   Any therapy is likely to function best if it is applied in combination with rehabilitation to increase task-related activity in the cortical region mediating the impaired behavioral function (Biernaskie et al., 2001; Krakauer et al., 2012). One particularly effective technique for rehabilitation is constraint-induced movement therapy, which involves restraining the spared limb to force increased use of the stroke-affected limb (Liepert et al., 1998). Activity in targeted brain regions can also be increased artificially by means of electrical or transcranial magnetic stimulation, but this too is most effective if applied in conjunction with rehabilitation (Fregni et al., 2006; Adkins et al., 2008; Ackerley et al., 2010). Light-based mapping (Ayling et al., 2009; Hira et al., 2009; Komiyama et al., 2010) will be a valuable tool for assessing the ability of these therapies to enhance cortical reorganization after stroke. The potential also exists to use optogenetic stimulation directly as a therapy. Because opsin expression can be restricted to specific cell types, it will be possible to stimulate these populations selectively to test their contribution to post-stroke recovery (Chen et al., 2012).  6.3.3 Considerations for translation of animal research to clinical practice   Despite many promising findings from animal research and some recent successes (Cook et al., 2012; Hill et al., 2012), translating basic research into clinical practice has  146 proven to be extremely difficult (O’Collins et al., 2006). Failures in translation can be attributed to the design of both basic research and clinical trials. In order to improve the productivity of translational research, it is important for scientists to conduct their research with clinical utility in mind, and for clinicians to think carefully about the methodology and results of animal studies before attempting to apply these results to patients.   In basic research, oversights in experimental design have hobbled many studies. One of the simplest is the use of relatively young animals in stroke studies, which does not reflect the reality of the human population of stroke patients. Second, stroke patients frequently present with comorbidities, risk factors and chronic conditions such as hypertension, smoking or diabetes (Galimanis et al., 2009). The influence of these comorbidities on stroke is only infrequently examined (Sweetnam et al., 2012). The choice of animal model is also a major consideration. Mice, for example, have many advantages in genetic tractability and are less expensive to house than larger animals, but their lissencephalic cortex and small brain size relative to humans must be recognized. Finally, in many experiments that test the efficacy of a therapeutic intervention, the timing of that intervention (i.e. before stroke onset) is relevant only to a limited subset of clinical situations (e.g. global ischemia during cardiac surgery). At the same time, many translational studies have failed because they did not reproduce key elements of the animal research upon which they were based (Plow et al., 2009). The patient population for these studies must be chosen carefully to  replicate as closely as possible the infarct size and placement of the original animal study. The intervention tested should be applied with comparable intensity or frequency, and given at a comparable time after stroke onset.  The work described here represents a compromise between aspirations of clinical relevance and concessions made to experimental constraints. Mice were aged approximately five months at the beginning of the longitudinal mapping study; older mice were not used in order to minimize mortality from stroke and repeated anesthesia. The rose bengal photothrombosis model of ischemic stroke was used to reproducibly target functional areas of cortex, an advantage that outweighed its decreased external validity as compared with other models of stroke. Future experiments that use this work as a starting point will attempt  147 to identify therapeutic strategies for enhancing recovery from stroke, and these experiments will need to be prudently designed and executed to maximize their clinical utility.    The rodent sensorimotor cortex, frequently studied in the context of middle cerebral artery stroke (Nedergaard and Hansen, 1993; Nawashiro et al., 1997; Li and Murphy, 2008), is an ideal model system for longitudinal studies of plasticity and vicariation after stroke. Cortical reorganization is a slow process, and is best studied in longitudinal experiments. In the motor cortex, such experiments have been constrained by the surgical limitations of repeated electrical stimulation (Kleim et al., 2003b; Gharbawie et al., 2005). Using LBM and intrinsic optical signal (IOS) imaging (Grinvald et al., 1986) to repeatedly map the motor and somatosensory cortices before and after targeted stroke, we were able to assess the cortical response to injury on an unprecedented time scale.   As a form of non-invasive brain stimulation, LBM is also a useful mouse model of human transcranial magnetic stimulation (TMS), a technique that has been used both for motor mapping (Liepert et al., 2000b; Sawaki et al., 2008) and therapeutic stimulation (Mansur et al., 2005; Fregni et al., 2006) after stroke. The selective nature of optogenetic stimulation can be exploited in the mouse model to identify the specific cortical targets that might benefit most from therapeutic stimulation, and to determine whether the natural process of recovery can be enhanced or accelerated by cortical stimulation.     148 References  Adesnik, H., and Scanziani, M. (2010). Lateral competition for cortical space by layer- specific horizontal circuits. Nature 464, 1155–1160. Aflalo, T.N., and Graziano, M.S.A. (2006). Possible origins of the complex topographic organization of motor cortex: reduction of a multidimensional space onto a two-dimensional array. J. Neurosci. 26, 6288–6297. Airan, R.D., Thompson, K.R., Fenno, L.E., Bernstein, H., and Deisseroth, K. (2009). Temporally precise in vivo control of intracellular signalling. Nature 458, 1025–1029. Albeanu, D.F., Soucy, E., Sato, T.F., Meister, M., and Murthy, V.N. (2008). LED arrays as cost effective and efficient light sources for widefield microscopy. PLoS ONE 3, e2146. Alvarez, V.A., and Sabatini, B.L. (2007). Anatomical and physiological plasticity of dendritic spines. Annu. Rev. Neurosci 30, 79–97. Amirikian, B., and Georgopoulos, A.P. (2003). Modular organization of directionally tuned cells in the motor cortex: is there a short-range order? Proc. Natl. Acad. Sci. U.S.A. 100, 12474–12479. Andersen, R.A., and Buneo, C.A. (2002). Intentional maps in posterior parietal cortex. Annu. Rev. Neurosci. 25, 189–220. Anderson, C.T., Sheets, P.L., Kiritani, T., and Shepherd, G.M.G. (2010). Sublayer-specific microcircuits of corticospinal and corticostriatal neurons in motor cortex. Nat Neurosci 13, 739–744. Andrew, R.D., and MacVicar, B.A. (1994). Imaging cell volume changes and neuronal excitation in the hippocampal slice. Neuroscience 62, 371–383. Apicella, A.J., Wickersham, I.R., Seung, H.S., and Shepherd, G.M.G. (2012). Laminarly Orthogonal Excitation of Fast-Spiking and Low-Threshold-Spiking Interneurons in Mouse Motor Cortex. J. Neurosci. 32, 7021–7033. Aravanis, A.M., Wang, L.-P., Zhang, F., Meltzer, L.A., Mogri, M.Z., Schneider, M.B., and Deisseroth, K. (2007). An optical neural interface: in vivo control of rodent motor cortex with integrated fiberoptic and optogenetic technology. J Neural Eng 4, S143–156. Arenkiel, B.R., Peca, J., Davison, I.G., Feliciano, C., Deisseroth, K., Augustine, G.J., Ehlers, M.D., and Feng, G. (2007). In vivo light-induced activation of neural circuitry in transgenic mice expressing channelrhodopsin-2. Neuron 54, 205–218.  149 Aroniadou, V., and Keller, A. (1993). The patterns and synaptic properties of horizontal intracortical connections in the rat motor cortex. J. Neurophysiol 70, 1553–1569. Asanuma, H. (1975). Recent developments in the study of the columnar arrangement of neurons within the motor cortex. Physiol Rev 55, 143–156. Asanuma, H., and Rosén, I. (1972). Topographical organization of cortical efferent zones projecting to distal forelimb muscles in the monkey. Exp Brain Res 14, 243–256. Ayling, O.G.S., Harrison, T.C., Boyd, J.D., Goroshkov, A., and Murphy, T.H. (2009). Automated light-based mapping of motor cortex by photoactivation of channelrhodopsin-2 transgenic mice. Nat. Methods 6, 219–224. Ayling, O.G.S., Harrison, T.C., and Murphy, T.H. (2011). Optogenetic investigation of inhibitory circuits between sensory and motor cortex in mouse. (SfN, Washington DC),. Bakin, J.S., Kwon, M.C., Masino, S.A., Weinberger, N.M., and Frostig, R.D. (1996). Suprathreshold auditory cortex activation visualized by intrinsic signal optical imaging. Cereb. Cortex 6, 120–130. Barbay, S., Guggenmos, D.J., Nishibe, M., and Nudo, R.J. (2012). Motor Representations in the Intact Hemisphere of the Rat Are Reduced After Repetitive Training of the Impaired Forelimb. Neurorehabil Neural Repair. Barkho, B.Z., and Zhao, X. (2011). Adult neural stem cells: response to stroke injury and potential for therapeutic applications. Curr Stem Cell Res Ther 6, 327–338. Beach, J.M., and Duling, B.R. (1993). A light-emitting diode light standard for photo- and videomicroscopy. J Microsc 172, 41–48. Beaumanoir, A., André-Obadia, N., Nahory, A., and Zerbi, D. (1996). Special types of periodic lateralized epileptiform discharges associated with confusional state in cerebral circulation insufficiency. Electroencephalogr Clin Neurophysiol 99, 287–292. Bélanger, M., Allaman, I., and Magistretti, P.J. (2011). Brain energy metabolism: focus on astrocyte-neuron metabolic cooperation. Cell Metab. 14, 724–738. Ben-Shaul, Y., Stark, E., Asher, I., Drori, R., Nadasdy, Z., and Abeles, M. (2003). Dynamical organization of directional tuning in the primate premotor and primary motor cortex. J. Neurophysiol. 89, 1136–1142. Berndt, A., Prigge, M., Gradmann, D., and Hegemann, P. (2010). Two open states with progressive proton selectivities in the branched channelrhodopsin-2 photocycle. Biophys. J. 98, 753–761. Berndt, A., Schoenenberger, P., Mattis, J., Tye, K.M., Deisseroth, K., Hegemann, P., and Oertner, T.G. (2011). High-efficiency channelrhodopsins for fast neuronal stimulation at low light levels. Proc. Natl. Acad. Sci. U.S.A. 108, 7595–7600.  150 Bernstein, J.G., Han, X., Henninger, M.A., Ko, E.Y., Qian, X., Franzesi, G.T., McConnell, J.P., Stern, P., Desimone, R., and Boyden, E.S. (2008). Prosthetic systems for therapeutic optical activation and silencing of genetically-targeted neurons. Proc Soc Photo Opt Instrum Eng 6854, 68540H. Biernaskie, J., and Corbett, D. (2001). Enriched rehabilitative training promotes improved forelimb motor function and enhanced dendritic growth after focal ischemic injury. J. Neurosci 21, 5272–5280. Biernaskie, J., Corbett, D., Peeling, J., Wells, J., and Lei, H. (2001). A serial MR study of cerebral blood flow changes and lesion development following endothelin-1-induced ischemia in rats. Magn Reson Med 46, 827–830. Bizzi, E., D’Avella, A., Saltiel, P., and Tresch, M. (2002). Modular organization of spinal motor systems. Neuroscientist 8, 437–442. Blits, B., Derks, S., Twisk, J., Ehlert, E., Prins, J., and Verhaagen, J. (2010). Adeno- associated viral vector (AAV)-mediated gene transfer in the red nucleus of the adult rat brain: Comparative analysis of the transduction properties of seven AAV serotypes and lentiviral vectors. JOURNAL OF NEUROSCIENCE METHODS 185, 257–263. Bonhoeffer, T., and Grinvald, A. (1993). The layout of iso-orientation domains in area 18 of cat visual cortex: optical imaging reveals a pinwheel-like organization. J. Neurosci. 13, 4157–4180. Boyden, E.S., Zhang, F., Bamberg, E., Nagel, G., and Deisseroth, K. (2005). Millisecond- timescale, genetically targeted optical control of neural activity. Nat. Neurosci. 8, 1263– 1268. Brown, C.E., Aminoltejari, K., Erb, H., Winship, I.R., and Murphy, T.H. (2009). In vivo voltage-sensitive dye imaging in adult mice reveals that somatosensory maps lost to stroke are replaced over weeks by new structural and functional circuits with prolonged modes of activation within both the peri-infarct zone and distant sites. J. Neurosci 29, 1719–1734. Brown, C.E., Li, P., Boyd, J.D., Delaney, K.R., and Murphy, T.H. (2007). Extensive turnover of dendritic spines and vascular remodeling in cortical tissues recovering from stroke. J. Neurosci 27, 4101–4109. Brown, S.P., and Hestrin, S. (2009). Intracortical circuits of pyramidal neurons reflect their long-range axonal targets. Nature 457, 1133–1136. Buchkremer-Ratzmann, I., August, M., Hagemann, G., and Witte, O.W. (1996). Electrophysiological transcortical diaschisis after cortical photothrombosis in rat brain. Stroke 27, 1105–1109; discussion 1109–1111. Buchkremer-Ratzmann, I., and Witte, O.W. (1997). Extended brain disinhibition following small photothrombotic lesions in rat frontal cortex. Neuroreport 8, 519–522.  151 Buneo, C.A., Jarvis, M.R., Batista, A.P., and Andersen, R.A. (2002). Direct visuomotor transformations for reaching. Nature 416, 632–636. Burgess, S.A., Bouchard, M.B., Yuan, B., and Hillman, E.M.C. (2008). Simultaneous multiwavelength laminar optical tomography. Opt Lett 33, 2710–2712. Burish, M.J., Stepniewska, I., and Kaas, J.H. (2008). Microstimulation and architectonics of frontoparietal cortex in common marmosets (Callithrix jacchus). J. Comp. Neurol. 507, 1151–1168. Bütefisch, C.M., Wessling, M., Netz, J., Seitz, R.J., and Hömberg, V. (2008). Relationship between interhemispheric inhibition and motor cortex excitability in subacute stroke patients. Neurorehabil Neural Repair 22, 4–21. Callaway, E.M., and Katz, L.C. (1993). Photostimulation using caged glutamate reveals functional circuitry in living brain slices. Proc. Natl. Acad. Sci. U.S.A. 90, 7661–7665. Capaday, C. (2004). The integrated nature of motor cortical function. Neuroscientist 10, 207– 220. Capaday, C., Devanne, H., Bertrand, L., and Lavoie, B.A. (1998). Intracortical connections between motor cortical zones controlling antagonistic muscles in the cat: a combined anatomical and physiological study. Exp Brain Res 120, 223–232. Carlo, A.D. (2009). Human and economic burden of stroke. Age Ageing 38, 4–5. Carmel, J.B., Kim, S., Brus-Ramer, M., and Martin, J.H. (2010). Feed-forward control of preshaping in the rat is mediated by the corticospinal tract. Eur. J. Neurosci. 32, 1678–1685. Carmena, J.M., Lebedev, M.A., Crist, R.E., O’Doherty, J.E., Santucci, D.M., Dimitrov, D.F., Patil, P.G., Henriquez, C.S., and Nicolelis, M.A.L. (2003). Learning to control a brain- machine interface for reaching and grasping by primates. PLoS Biol. 1, E42. Carmichael, S.T. (2006). Cellular and molecular mechanisms of neural repair after stroke: making waves. Ann. Neurol 59, 735–742. Carmichael, S.T. (2012). Brain Excitability in Stroke The Yin and Yang of Stroke Progression. Arch. Neurol. 69, 161–167. Carmichael, S.T., Archibeque, I., Luke, L., Nolan, T., Momiy, J., and Li, S.L. (2005). Growth-associated gene expression after stroke: Evidence for a   growth-promoting region in peri-infarct cortex. Exp. Neurol. 193, 291–311. Carmichael, S.T., and Chesselet, M.-F. (2002). Synchronous neuronal activity is a signal for axonal sprouting after cortical lesions in the adult. J. Neurosci. 22, 6062–6070. Carmichael, S.T., Tatsukawa, K., Katsman, D., Tsuyuguchi, N., and Kornblum, H.I. (2004). Evolution of diaschisis in a focal stroke model. Stroke 35, 758–763.  152 Carvell, G.E., and Simons, D.J. (1986). Somatotopic organization of the second somatosensory area (SII) in the cerebral cortex of the mouse. Somatosens Res 3, 213–237. Cescon, C., Madeleine, P., Graven-Nielsen, T., Merletti, R., and Farina, D. (2007). Two- dimensional spatial distribution of surface mechanomyographical response to single motor unit activity. J. Neurosci. Methods 159, 19–25. Cetin, A., Komai, S., Eliava, M., Seeburg, P.H., and Osten, P. (2006). Stereotaxic gene delivery in the rodent brain. Nat Protoc 1, 3166–3173. Chakrabarty, S., Friel, K.M., and Martin, J.H. (2009). Activity-Dependent Plasticity Improves M1 Motor Representation and Corticospinal Tract Connectivity. J. Neurophysiol 101, 1283 –1293. Chakrabarty, S., and Martin, J.H. (2000). Postnatal development of the motor representation in primary motor cortex. J. Neurophysiol. 84, 2582–2594. Chapin, J.K., Moxon, K.A., Markowitz, R.S., and Nicolelis, M.A. (1999). Real-time control of a robot arm using simultaneously recorded neurons in the motor cortex. Nat. Neurosci. 2, 664–670. Chemla, S., and Chavane, F. (2010). Voltage-sensitive dye imaging: Technique review and models. Journal of Physiology-Paris 104, 40–50. Chen, B.E., Trachtenberg, J.T., Holtmaat, A.J.G.D., and Svoboda, K. (2008). Long-term, high-resolution imaging in the neocortex in vivo. CSH Protoc 2008, pdb.prot4902. Chen, J., Li, Y., Wang, L., Zhang, Z., Lu, D., Lu, M., and Chopp, M. (2001). Therapeutic benefit of intravenous administration of bone marrow stromal cells after cerebral ischemia in rats. Stroke 32, 1005–1011. Chen, R., Cohen, L.G., and Hallett, M. (2002). Nervous system reorganization following injury. Neuroscience 111, 761–773. Chen, S., Mohajerani, M.H., Xie, Y., and Murphy, T.H. (2012). Optogenetic analysis of neuronal excitability during global ischemia reveals selective deficits in sensory processing following reperfusion in mouse cortex. J. Neurosci. 32, 13510–13519. Cheney, P.D., and Fetz, E.E. (1984). Corticomotoneuronal cells contribute to long-latency stretch reflexes in the rhesus monkey. J. Physiol. (Lond.) 349, 249–272. Cheney, P.D., Fetz, E.E., and Palmer, S.S. (1985). Patterns of facilitation and suppression of antagonist forelimb muscles from motor cortex sites in the awake monkey. J. Neurophysiol. 53, 805–820. Chestek, C.A., Batista, A.P., Santhanam, G., Yu, B.M., Afshar, A., Cunningham, J.P., Gilja, V., Ryu, S.I., Churchland, M.M., and Shenoy, K.V. (2007). Single-neuron stability during repeated reaching in macaque premotor cortex. J. Neurosci. 27, 10742–10750.  153 Chklovskii, D.B., and Koulakov, A.A. (2004). Maps in the brain: what can we learn from them? Annu. Rev. Neurosci. 27, 369–392. Chow, B.Y., Han, X., Dobry, A.S., Qian, X., Chuong, A.S., Li, M., Henninger, M.A., Belfort, G.M., Lin, Y., Monahan, P.E., et al. (2010). High-performance genetically targetable optical neural silencing by light-driven proton pumps. Nature 463, 98–102. Christie, I.N., Wells, J.A., Southern, P., Marina, N., Kasparov, S., Gourine, A.V., and Lythgoe, M.F. (2012). fMRI response to blue light delivery in the naïve brain: Implications for combined optogenetic fMRI studies. Neuroimage 66C, 634–641. Churchland, M.M., Cunningham, J.P., Kaufman, M.T., Foster, J.D., Nuyujukian, P., Ryu, S.I., and Shenoy, K.V. (2012). Neural population dynamics during reaching. Nature 487, 51– 56. Churchland, M.M., Santhanam, G., and Shenoy, K.V. (2006). Preparatory Activity in Premotor and Motor Cortex Reflects the Speed of the Upcoming Reach. J. Neurophysiol 96, 3130 –3146. Cisek, P., and Kalaska, J.F. (2010). Neural Mechanisms for Interacting with a World Full of Action Choices. Annual Review of Neuroscience 33, 269–298. Clarkson, A.N., Huang, B.S., Macisaac, S.E., Mody, I., and Carmichael, S.T. (2010). Reducing excessive GABA-mediated tonic inhibition promotes functional recovery after stroke. Nature 468, 305–309. Cohen, L.B. (1973). Changes in neuron structure during action potential propagation and synaptic transmission. Physiol. Rev. 53, 373–418. Conner, J.M., Culberson, A., Packowski, C., Chiba, A.A., and Tuszynski, M.H. (2003). Lesions of the Basal forebrain cholinergic system impair task acquisition and abolish cortical plasticity associated with motor skill learning. Neuron 38, 819–829. Conner, J.M., Kulczycki, M., and Tuszynski, M.H. (2010). Unique Contributions of Distinct Cholinergic Projections to Motor Cortical Plasticity and Learning. Cereb. Cortex 20, 2739– 2748. Cook, D.J., Teves, L., and Tymianski, M. (2012). Treatment of stroke with a PSD-95 inhibitor in the gyrencephalic primate brain. Nature 483, 213–217. Cramer, S.C., Nelles, G., Benson, R.R., Kaplan, J.D., Parker, R.A., Kwong, K.K., Kennedy, D.N., Finklestein, S.P., and Rosen, B.R. (1997). A functional MRI study of subjects recovered from hemiparetic stroke. Stroke 28, 2518–2527. Cramer, S.C., Shah, R., Juranek, J., Crafton, K.R., and Le, V. (2006). Activity in the peri- infarct rim in relation to recovery from stroke. Stroke 37, 111–115.  154 Dan, Y., and Poo, M.-M. (2006). Spike timing-dependent plasticity: from synapse to perception. Physiol. Rev. 86, 1033–1048. Dancause, N. (2006). Vicarious function of remote cortex following stroke: recent evidence from human and animal studies. Neuroscientist 12, 489–499. Dancause, N., Barbay, S., Frost, S.B., Plautz, E.J., Chen, D., Zoubina, E.V., Stowe, A.M., and Nudo, R.J. (2005). Extensive cortical rewiring after brain injury. J. Neurosci 25, 10167– 10179. Dawson, T.M., Ko, H.S., and Dawson, V.L. (2010). Genetic animal models of Parkinson’s disease. Neuron 66, 646–661. Devor, A., Dunn, A.K., Andermann, M.L., Ulbert, I., Boas, D.A., and Dale, A.M. (2003). Coupling of total hemoglobin concentration, oxygenation, and neural activity in rat somatosensory cortex. Neuron 39, 353–359. Diester, I., Kaufman, M.T., Mogri, M., Pashaie, R., Goo, W., Yizhar, O., Ramakrishnan, C., Deisseroth, K., and Shenoy, K.V. (2011). An optogenetic toolbox designed for primates. Nat. Neurosci. 14, 387–397. Dijkhuizen, R.M., Ren, J., Mandeville, J.B., Wu, O., Ozdag, F.M., Moskowitz, M.A., Rosen, B.R., and Finklestein, S.P. (2001). Functional magnetic resonance imaging of reorganization in rat brain after stroke. Proc. Natl. Acad. Sci. U.S.A. 98, 12766–12771. Dijkhuizen, R.M., Singhal, A.B., Mandeville, J.B., Wu, O., Halpern, E.F., Finklestein, S.P., Rosen, B.R., and Lo, E.H. (2003). Correlation between brain reorganization, ischemic damage, and neurologic status after transient focal cerebral ischemia in rats: a functional magnetic resonance imaging study. J. Neurosci. 23, 510–517. Dirnagl, U., Iadecola, C., and Moskowitz, M.A. (1999). Pathobiology of ischaemic stroke: an integrated view. Trends Neurosci. 22, 391–397. Domann, R., Hagemann, G., Kraemer, M., Freund, H.J., and Witte, O.W. (1993). Electrophysiological changes in the surrounding brain tissue of photochemically induced cortical infarcts in the rat. Neurosci. Lett. 155, 69–72. Dombeck, D.A., Graziano, M.S., and Tank, D.W. (2009). Functional clustering of neurons in motor cortex determined by cellular resolution imaging in awake behaving mice. J. Neurosci 29, 13751–13760. Donoghue, J.P., Leibovic, S., and Sanes, J.N. (1992). Organization of the forelimb area in squirrel monkey motor cortex: representation of digit, wrist, and elbow muscles. Exp Brain Res 89, 1–19. Donoghue, J.P., and Sanes, J.N. (1987). Peripheral nerve injury in developing rats reorganizes representation pattern in motor cortex. Proc. Natl. Acad. Sci. U.S.A. 84, 1123– 1126.  155 Donoghue, J.P., and Wise, S.P. (1982). The motor cortex of the rat: cytoarchitecture and microstimulation mapping. J. Comp. Neurol 212, 76–88. Douglas, R.J., and Martin, K.A.C. (2004). Neuronal circuits of the neocortex. Annu. Rev. Neurosci 27, 419–451. Drew, P.J., Shih, A.Y., Driscoll, J.D., Knutsen, P.M., Blinder, P., Davalos, D., Akassoglou, K., Tsai, P.S., and Kleinfeld, D. (2010). Chronic optical access through a polished and reinforced thinned skull. Nat. Methods 7, 981–984. Dum, R.P., and Strick, P.L. (2002). Motor areas in the frontal lobe of the primate. Physiol Behav 77, 677–682. Dum, R.P., and Strick, P.L. (2005). Frontal lobe inputs to the digit representations of the motor areas on the lateral surface of the hemisphere. J. Neurosci. 25, 1375–1386. Eisenberg, M., Shmuelof, L., Vaadia, E., and Zohary, E. (2010). Functional organization of human motor cortex: directional selectivity for movement. J. Neurosci. 30, 8897–8905. Eisner-Janowicz, I., Barbay, S., Hoover, E., Stowe, A.M., Frost, S.B., Plautz, E.J., and Nudo, R.J. (2008). Early and late changes in the distal forelimb representation of the supplementary motor area after injury to frontal motor areas in the squirrel monkey. J. Neurophysiol 100, 1498–1512. Erkinjuntti, T., Roman, G., Gauthier, S., Feldman, H., and Rockwood, K. (2004). Emerging therapies for vascular dementia and vascular cognitive   impairment. Stroke 35, 1010–1017. Ethier, C., Oby, E.R., Bauman, M.J., and Miller, L.E. (2012). Restoration of grasp following paralysis through brain-controlled stimulation of muscles. Nature 485, 368–371. Evarts, E.V. (1968). Relation of pyramidal tract activity to force exerted during voluntary movement. J. Neurophysiol 31, 14–27. Feldbauer, K., Zimmermann, D., Pintschovius, V., Spitz, J., Bamann, C., and Bamberg, E. (2009). Channelrhodopsin-2 is a leaky proton pump. Proceedings of the National Academy of Sciences 106, 12317–12322. Feldman, D.E. (2009). Synaptic Mechanisms for Plasticity in Neocortex. Annu. Rev. Neurosci. 32, 33–55. Feng, G., Mellor, R.H., Bernstein, M., Keller-Peck, C., Nguyen, Q.T., Wallace, M., Nerbonne, J.M., Lichtman, J.W., and Sanes, J.R. (2000). Imaging neuronal subsets in transgenic mice expressing multiple spectral variants of GFP. Neuron 28, 41–51. Ferezou, I., Haiss, F., Gentet, L.J., Aronoff, R., Weber, B., and Petersen, C.C.H. (2007). Spatiotemporal dynamics of cortical sensorimotor integration in behaving mice. Neuron 56, 907–923.  156 Ferster, D., and Miller, K.D. (2000). Neural mechanisms of orientation selectivity in the visual cortex. Annu. Rev. Neurosci. 23, 441–471. Fisher, M., and Takano, K. (1995). The penumbra, therapeutic time window and acute ischaemic stroke. Baillieres Clin Neurol 4, 279–295. Flaherty, A.W., and Graybiel, A.M. (1991). Corticostriatal transformations in the primate somatosensory system. Projections from physiologically mapped body-part representations. J. Neurophysiol. 66, 1249–1263. Foeller, E., Celikel, T., and Feldman, D. (2005). Inhibitory sharpening of receptive fields contributes to whisker map   plasticity in rat somatosensory cortex. J. Neurophysiol. 94, 4387–4400. Fregni, F., Boggio, P.S., Valle, A.C., Rocha, R.R., Duarte, J., Ferreira, M.J.L., Wagner, T., Fecteau, S., Rigonatti, S.P., Riberto, M., et al. (2006). A sham-controlled trial of a 5-day course of repetitive transcranial magnetic stimulation of the unaffected hemisphere in stroke patients. Stroke 37, 2115–2122. Fridman, E.A., Hanakawa, T., Chung, M., Hummel, F., Leiguarda, R.C., and Cohen, L.G. (2004). Reorganization of the human ipsilesional premotor cortex after stroke. Brain 127, 747–758. Fritsch, G., and Hitzig, E. (1870). Electric excitability of the cerebrum (Uber die elektrische Erregbarkeit des Grosshirns). Epilepsy Behav 15, 123–130. Frostig, R.D., Lieke, E.E., Ts’o, D.Y., and Grinvald, A. (1990). Cortical functional architecture and local coupling between neuronal activity and the microcirculation revealed by in vivo high-resolution optical imaging of intrinsic signals. Proc. Natl. Acad. Sci. U.S.A. 87, 6082–6086. Fujioka, H., Kaneko, H., Fujii, M., Suzuki, S.S., Imoto, H., Nomura, S., Fujisawa, H., Yamakawa, T., and Suzuki, M. (2010). Epileptiform discharges and neuronal plasticity in the acute   peri-infarct cortex of rats. Neurol. Res. 32, 666–669. Fuxe, K., Bjelke, B., Andbjer, B., Grahn, H., Rimondini, R., and Agnati, L.F. (1997). Endothelin-1 induced lesions of the frontoparietal cortex of the rat. A possible model of focal cortical ischemia. Neuroreport 8, 2623–2629. Gajkowska, B., Frontczak-Baniewicz, M., Gadamski, R., and Barskov, I. (1997). Photochemically-induced vascular damage in brain cortex. Transmission and scanning electron microscopy study. Acta Neurobiol Exp (Wars) 57, 203–208. Galea, J.M., and Celnik, P. (2009). Brain polarization enhances the formation and retention of motor memories. J. Neurophysiol. 102, 294–301. Galimanis, A., Mono, M.-L., Arnold, M., Nedeltchev, K., and Mattle, H.P. (2009). Lifestyle and stroke risk: a review. Curr. Opin. Neurol. 22, 60–68.  157 Gallese, V., Fadiga, L., Fogassi, L., and Rizzolatti, G. (1996). Action recognition in the premotor cortex. Brain 119 ( Pt 2), 593–609. Ganguly, K., Dimitrov, D.F., Wallis, J.D., and Carmena, J.M. (2011). Reversible large-scale modification of cortical networks during neuroprosthetic control. Nature Neuroscience 14, 662–667. Gao, W.-J., and Zheng, Z.-H. (2004). Target-specific differences in somatodendritic morphology of layer V pyramidal neurons in rat motor cortex. J. Comp. Neurol. 476, 174– 185. Georgopoulos, A., Kalaska, J., Caminiti, R., and Massey, J. (1982). On the Relations Between the Direction of Two-Dimensional Arm Movements and Cell Discharge in Primate Motor Cortex. J. Neurosci. 2, 1527–1537. Georgopoulos, A.P., Kettner, R.E., and Schwartz, A.B. (1988). Primate motor cortex and free arm movements to visual targets in three-dimensional space. II. Coding of the direction of movement by a neuronal population. J. Neurosci. 8, 2928–2937. Georgopoulos, A.P., Merchant, H., Naselaris, T., and Amirikian, B. (2007). Mapping of the preferred direction in the motor cortex. Proc. Natl. Acad. Sci. U.S.A. 104, 11068–11072. Georgopoulos, A.P., Schwartz, A.B., and Kettner, R.E. (1986). Neuronal population coding of movement direction. Science 233, 1416–1419. Georgopoulos, A.P., and Stefanis, C.N. (2007). Local shaping of function in the motor cortex: motor contrast, directional tuning. Brain Res Rev 55, 383–389. Gerits, A., Farivar, R., Rosen, B.R., Wald, L.L., Boyden, E.S., and Vanduffel, W. (2012). Optogenetically induced behavioral and functional network changes in primates. Curr. Biol. 22, 1722–1726. Gharbawie, O.A., Gonzalez, C.L.R., Williams, P.T., Kleim, J.A., and Whishaw, I.Q. (2005). Middle cerebral artery (MCA) stroke produces dysfunction in adjacent motor cortex as detected by intracortical microstimulation in rats. Neuroscience 130, 601–610. Gharbawie, O.A., Stepniewska, I., and Kaas, J.H. (2011a). Cortical connections of functional zones in posterior parietal cortex and frontal cortex motor regions in new world monkeys. Cereb. Cortex 21, 1981–2002. Gharbawie, O.A., Stepniewska, I., Qi, H., and Kaas, J.H. (2011b). Multiple parietal-frontal pathways mediate grasping in macaque monkeys. J. Neurosci. 31, 11660–11677. Gharbawie, O.A., Williams, P.T.J.A., Kolb, B., and Whishaw, I.Q. (2008). Transient middle cerebral artery occlusion disrupts the forelimb movement representations of rat motor cortex. Eur. J. Neurosci. 28, 951–963.  158 Ghosh, A., Haiss, F., Sydekum, E., Schneider, R., Gullo, M., Wyss, M.T., Mueggler, T., Baltes, C., Rudin, M., Weber, B., et al. (2010). Rewiring of hindlimb corticospinal neurons after spinal cord injury. Nat. Neurosci 13, 97–104. Gorelick, P.B., Sacco, R.L., Smith, D.B., Alberts, M., Mustone-Alexander, L., Rader, D., Ross, J.L., Raps, E., Ozer, M.N., Brass, L.M., et al. (1999). Prevention of a first stroke - A review of guidelines and a   multidisciplinary consensus statement from the National Stroke Association. JAMA-J. Am. Med. Assoc. 281, 1112–1120. Gradinaru, V., Mogri, M., Thompson, K.R., Henderson, J.M., and Deisseroth, K. (2009). Optical deconstruction of parkinsonian neural circuitry. Science 324, 354–359. Gradinaru, V., Thompson, K.R., Zhang, F., Mogri, M., Kay, K., Schneider, M.B., and Deisseroth, K. (2007). Targeting and Readout Strategies for Fast Optical Neural Control In Vitro and In Vivo. J. Neurosci. 27, 14231–14238. Gradinaru, V., Zhang, F., Ramakrishnan, C., Mattis, J., Prakash, R., Diester, I., Goshen, I., Thompson, K.R., and Deisseroth, K. (2010). Molecular and cellular approaches for diversifying and extending optogenetics. Cell 141, 154–165. Graziano, M.S.A., and Aflalo, T.N. (2007a). Mapping behavioral repertoire onto the cortex. Neuron 56, 239–251. Graziano, M.S.A., and Aflalo, T.N. (2007b). Rethinking Cortical Organization: Moving Away from Discrete Areas Arranged in Hierarchies. Neuroscientist 13, 138–147. Graziano, M.S.A., Aflalo, T.N.S., and Cooke, D.F. (2005). Arm movements evoked by electrical stimulation in the motor cortex of monkeys. J. Neurophysiol 94, 4209–4223. Graziano, M.S.A., Taylor, C.S.R., and Moore, T. (2002a). Complex movements evoked by microstimulation of precentral cortex. Neuron 34, 841–851. Graziano, M.S.A., Taylor, C.S.R., Moore, T., and Cooke, D.F. (2002b). The cortical control of movement revisited. Neuron 36, 349–362. Grinvald, A., Frostig, R.D., Lieke, E., and Hildesheim, R. (1988). Optical imaging of neuronal activity. Physiol. Rev. 68, 1285–1366. Grinvald, A., Lieke, E., Frostig, R.D., Gilbert, C.D., and Wiesel, T.N. (1986). Functional architecture of cortex revealed by optical imaging of intrinsic signals. Nature 324, 361–364. Grubb, M.S., and Burrone, J. (2010). Channelrhodopsin-2 localised to the axon initial segment. PLoS ONE 5, e13761. Grutzendler, J., and Gan, W.-B. (2007). Long-term two-photon transcranial imaging of synaptic structures in the living brain. CSH Protoc 2007, pdb.prot4766.  159 Gunaydin, L.A., Yizhar, O., Berndt, A., Sohal, V.S., Deisseroth, K., and Hegemann, P. (2010). Ultrafast optogenetic control. Nat. Neurosci. 13, 387–392. Haiss, F., and Schwarz, C. (2005). Spatial Segregation of Different Modes of Movement Control in the Whisker Representation of Rat Primary Motor Cortex. The Journal of Neuroscience 25, 1579 –1587. Hallett, M. (2007). Transcranial magnetic stimulation: A primer. Neuron 55, 187–199. Hari, R., Forss, N., Avikainen, S., Kirveskari, E., Salenius, S., and Rizzolatti, G. (1998). Activation of human primary motor cortex during action observation: A neuromagnetic study. Proceedings of the National Academy of Sciences 95, 15061 –15065. Harnett, M.T., Makara, J.K., Spruston, N., Kath, W.L., and Magee, J.C. (2012). Synaptic amplification by dendritic spines enhances input cooperativity. Nature 491, 599–602. Harrison, T.C., Ayling, O.G.S., and Murphy, T.H. (2012). Distinct cortical circuit mechanisms for complex forelimb movement and motor map topography. Neuron 74, 397– 409. Harrison, T.C., Sigler, A., and Murphy, T.H. (2009). Simple and cost-effective hardware and software for functional brain mapping using intrinsic optical signal imaging. J. Neurosci. Methods 182, 211–218. Hatsopoulos, N.G., Xu, Q., and Amit, Y. (2007). Encoding of movement fragments in the motor cortex. J. Neurosci. 27, 5105–5114. Hattox, A.M., and Nelson, S.B. (2007). Layer V Neurons in Mouse Cortex Projecting to Different Targets Have Distinct Physiological Properties. J Neurophysiol 98, 3330–3340. He, S.Q., Dum, R.P., and Strick, P.L. (1993). Topographic organization of corticospinal projections from the frontal lobe: motor areas on the lateral surface of the hemisphere. J. Neurosci. 13, 952–980. He, S.Q., Dum, R.P., and Strick, P.L. (1995). Topographic organization of corticospinal projections from the frontal lobe: motor areas on the medial surface of the hemisphere. J. Neurosci. 15, 3284–3306. Helmstaedter, M., Sakmann, B., and Feldmeyer, D. (2009). Neuronal Correlates of Local, Lateral, and Translaminar Inhibition with Reference to Cortical Columns. Cereb Cortex 19, 926–937. Herfst, L.J., and Brecht, M. (2008). Whisker movements evoked by stimulation of single motor neurons in the facial nucleus of the rat. J. Neurophysiol 99, 2821–2832. Hess, G., and Donoghue, J.P. (1994). Long-term potentiation of horizontal connections provides a mechanism to reorganize cortical motor maps. J. Neurophysiol. 71, 2543–2547.  160 Hill, M.D., Martin, R.H., Mikulis, D., Wong, J.H., Silver, F.L., Terbrugge, K.G., Milot, G., Clark, W.M., Macdonald, R.L., Kelly, M.E., et al. (2012). Safety and efficacy of NA-1 in patients with iatrogenic stroke after endovascular aneurysm repair (ENACT): a phase 2, randomised, double-blind, placebo-controlled trial. Lancet Neurol 11, 942–950. Hira, R., Honkura, N., Noguchi, J., Maruyama, Y., Augustine, G.J., Kasai, H., and Matsuzaki, M. (2009). Transcranial optogenetic stimulation for functional mapping of the motor cortex. J. Neurosci Meth 179, 258–263. Hochberg, L.R., Serruya, M.D., Friehs, G.M., Mukand, J.A., Saleh, M., Caplan, A.H., Branner, A., Chen, D., Penn, R.D., and Donoghue, J.P. (2006). Neuronal ensemble control of prosthetic devices by a human with   tetraplegia. Nature 442, 164–171. Hodgson, J.G., Agopyan, N., Gutekunst, C.A., Leavitt, B.R., LePiane, F., Singaraja, R., Smith, D.J., Bissada, N., McCutcheon, K., Nasir, J., et al. (1999). A YAC mouse model for Huntington’s disease with full-length mutant huntingtin, cytoplasmic toxicity, and selective striatal neurodegeneration. Neuron 23, 181–192. Hooks, B.M., Hires, S.A., Zhang, Y.-X., Huber, D., Petreanu, L., Svoboda, K., and Shepherd, G.M.G. (2011). Laminar analysis of excitatory local circuits in vibrissal motor and sensory cortical areas. PLoS Biol. 9, e1000572. Hoover, J., and Strick, P. (1993). Multiple output channels in the basal ganglia. Science 259, 819 –821. Hosp, J.A., and Luft, A.R. (2011). Cortical Plasticity during Motor Learning and Recovery after Ischemic Stroke. Neural Plasticity 2011, 1–9. Hosp, J.A., Molina-Luna, K., Hertler, B., Atiemo, C.O., Stett, A., and Luft, A.R. (2008). Thin-film epidural microelectrode arrays for somatosensory and motor cortex mapping in rat. J. Neurosci. Methods 172, 255–262. Hossmann, K.-A. (2006). Pathophysiology and therapy of experimental stroke. Cell. Mol. Neurobiol 26, 1057–1083. Hossmann, K.A. (1994). Viability thresholds and the penumbra of focal ischemia. Ann. Neurol. 36, 557–565. Huber, D., Petreanu, L., Ghitani, N., Ranade, S., Hromádka, T., Mainen, Z., and Svoboda, K. (2008). Sparse optical microstimulation in barrel cortex drives learned behaviour in freely moving mice. Nature 451, 61–64. Huberman, A.D., Feller, M.B., and Chapman, B. (2008). Mechanisms underlying development of visual maps and receptive fields. Annu. Rev. Neurosci. 31, 479–509. Huerta, P.T., and Volpe, B.T. (2009). Transcranial magnetic stimulation, synaptic plasticity and network oscillations. Journal of NeuroEngineering and Rehabilitation 6, 7.  161 Hummel, F. (2005). Effects of non-invasive cortical stimulation on skilled motor function in chronic stroke. Brain 128, 490–499. Hummel, F.C., and Cohen, L.G. (2006). Non-invasive brain stimulation: a new strategy to improve neurorehabilitation after stroke? Lancet Neurol 5, 708–712. Isomura, Y., Harukuni, R., Takekawa, T., Aizawa, H., and Fukai, T. (2009). Microcircuitry coordination of cortical motor information in self-initiation of voluntary movements. Nat. Neurosci. 12, 1586–1593. Jacobs, K.M., and Donoghue, J.P. (1991). Reshaping the cortical motor map by unmasking latent intracortical connections. Science 251, 944–947. Jankowska, E., Padel, Y., and Tanaka, R. (1975). Projections of pyramidal tract cells to alpha-motoneurones innervating hind-limb muscles in the monkey. J. Physiol. (Lond.) 249, 637–667. Janovjak, H., Szobota, S., Wyart, C., Trauner, D., and Isacoff, E.Y. (2010). A light-gated, potassium-selective glutamate receptor for the optical inhibition of neuronal firing. Nature Neuroscience 13, 1027–1032. Johansen-Berg, H., Dawes, H., Guy, C., Smith, S.M., Wade, D.T., and Matthews, P.M. (2002a). Correlation between motor improvements and altered fMRI activity after rehabilitative therapy. Brain 125, 2731–2742. Johansen-Berg, H., Rushworth, M.F.S., Bogdanovic, M.D., Kischka, U., Wimalaratna, S., and Matthews, P.M. (2002b). The role of ipsilateral premotor cortex in hand movement after stroke. Proc. Natl. Acad. Sci. U.S.A. 99, 14518–14523. Johnston, D.G., Denizet, M., Mostany, R., and Portera-Cailliau, C. (2012). Chronic In Vivo Imaging Shows No Evidence of Dendritic Plasticity or Functional Remapping in the Contralesional Cortex after Stroke. Cerebral Cortex (New York, N.Y.: 1991). Jones, P.B., Shin, H.K., Boas, D.A., Hyman, B.T., Moskowitz, M.A., Ayata, C., and Dunn, A.K. (2008). Simultaneous multispectral reflectance imaging and laser speckle flowmetry of cerebral blood flow and oxygen metabolism in focal cerebral ischemia. J Biomed Opt 13, 044007. Joshi, I., and Andrew, R.D. (2001). Imaging anoxic depolarization during ischemia-like conditions in the mouse hemi-brain slice. J. Neurophysiol. 85, 414–424. Juliano, S.L., Ma, W., and Eslin, D. (1991). Cholinergic depletion prevents expansion of topographic maps in somatosensory cortex. Proc. Natl. Acad. Sci. U.S.A. 88, 780–784. Kaas, J.H., Nelson, R.J., Sur, M., Lin, C.S., and Merzenich, M.M. (1979). Multiple representations of the body within the primary somatosensory cortex of primates. Science 204, 521–523.  162 Kakei, S., Hoffman, D.S., and Strick, P.L. (2001). Direction of action is represented in the ventral premotor cortex. Nat. Neurosci. 4, 1020–1025. Kalaska, J.F., Cohen, D.A., Hyde, M.L., and Prud’homme, M. (1989). A comparison of movement direction-related versus load direction-related activity in primate motor cortex, using a two-dimensional reaching task. J. Neurosci. 9, 2080–2102. Kalkman, C., Chesnut, R., and Drummond, J. (1994). Effects of Droperidol, Pentobarbital, and Ketamine on Myogen... : Neurosurgery. Kapfer, C., Glickfeld, L.L., Atallah, B.V., and Scanziani, M. (2007). Supralinear increase of recurrent inhibition during sparse activity in the somatosensory cortex. Nat. Neurosci 10, 743–753. Karmarkar, U.R., and Dan, Y. (2006). Experience-dependent plasticity in adult visual cortex. Neuron 52, 577–585. Kaufman, M.T., Churchland, M.M., Santhanam, G., Yu, B.M., Afshar, A., Ryu, S.I., and Shenoy, K.V. (2010). Roles of monkey premotor neuron classes in movement preparation and execution. J. Neurophysiol. 104, 799–810. Keller, A. (1993). Intrinsic synaptic organization of the motor cortex. Cereb. Cortex 3, 430– 441. Kelly, R.M., and Strick, P.L. (2003). Cerebellar Loops with Motor Cortex and Prefrontal Cortex of a Nonhuman Primate. J. Neurosci 23, 8432 –8444. Kiritani, T., Wickersham, I.R., Seung, H.S., and Shepherd, G.M.G. (2012). Hierarchical connectivity and connection-specific dynamics in the corticospinal-corticostriatal microcircuit in mouse motor cortex. J. Neurosci. 32, 4992–5001. Kleim, J.A., Barbay, S., and Nudo, R.J. (1998). Functional Reorganization of the Rat Motor Cortex Following Motor Skill Learning. J Neurophysiol 80, 3321–3325. Kleim, J.A., Bruneau, R., Calder, K., Pocock, D., VandenBerg, P.M., MacDonald, E., Monfils, M.H., Sutherland, R.J., and Nader, K. (2003a). Functional organization of adult motor cortex is dependent upon continued protein synthesis. Neuron 40, 167–176. Kleim, J.A., Bruneau, R., VandenBerg, P., MacDonald, E., Mulrooney, R., and Pocock, D. (2003b). Motor cortex stimulation enhances motor recovery and reduces peri-infarct dysfunction following ischemic insult. Neurol. Res 25, 789–793. Kleim, J.A., Cooper, N.R., and VandenBerg, P.M. (2002). Exercise induces angiogenesis but does not alter movement representations within rat motor cortex. Brain Res. 934, 1–6. Kleim, J.A., Hogg, T.M., VandenBerg, P.M., Cooper, N.R., Bruneau, R., and Remple, M. (2004). Cortical synaptogenesis and motor map reorganization occur during late, but not early, phase of motor skill learning. J. Neurosci. 24, 628–633.  163 Kleinfeld, D., and Deschênes, M. (2011). Neuronal basis for object location in the vibrissa scanning sensorimotor system. Neuron 72, 455–468. Kleinlogel, S., Terpitz, U., Legrum, B., Gökbuget, D., Boyden, E.S., Bamann, C., Wood, P.G., and Bamberg, E. (2011). A gene-fusion strategy for stoichiometric and co-localized expression of light-gated membrane proteins. Nat. Methods 8, 1083–1088. Komiyama, T., Sato, T.R., O’Connor, D.H., Zhang, Y.-X., Huber, D., Hooks, B.M., Gabitto, M., and Svoboda, K. (2010). Learning-related fine-scale specificity imaged in motor cortex circuits of behaving mice. Nature 464, 1182–1186. Koralek, A.C., Jin, X., Long, J.D., 2nd, Costa, R.M., and Carmena, J.M. (2012). Corticostriatal plasticity is necessary for learning intentional neuroprosthetic skills. Nature 483, 331–335. Koulakov, A.A., and Chklovskii, D.B. (2001). Orientation preference patterns in mammalian visual cortex: a wire length minimization approach. Neuron 29, 519–527. Krakauer, J.W., Carmichael, S.T., Corbett, D., and Wittenberg, G.F. (2012). Getting neurorehabilitation right: what can be learned from animal models? Neurorehabil Neural Repair 26, 923–931. Kramer, R.H., Chambers, J.J., and Trauner, D. (2005). Photochemical tools for remote control of ion channels in excitable cells. Nat. Chem. Biol. 1, 360–365. Kravitz, A.V., Freeze, B.S., Parker, P.R.L., Kay, K., Thwin, M.T., Deisseroth, K., and Kreitzer, A.C. (2010). Regulation of parkinsonian motor behaviours by optogenetic control of basal ganglia circuitry. Nature advance online publication,. Lawrence, D.G., and Kuypers, H.G. (1968a). The functional organization of the motor system in the monkey. I. The effects of bilateral pyramidal lesions. Brain 91, 1–14. Lawrence, D.G., and Kuypers, H.G. (1968b). The functional organization of the motor system in the monkey. II. The effects of lesions of the descending brain-stem pathways. Brain 91, 15–36. Lee, D., Lin, B.-J., and Lee, A.K. (2012). Hippocampal Place Fields Emerge upon Single- Cell Manipulation of Excitability During Behavior. Science 337, 849–853. Lee, S., Carvell, G.E., and Simons, D.J. (2008). Motor modulation of afferent somatosensory circuits. Nat. Neurosci. 11, 1430–1438. Lemon, R.N. (2008). Descending Pathways in Motor Control. Annu. Rev. Neurosci. 31, 195– 218. Lemon, R.N., Mantel, G.W., and Muir, R.B. (1986). Corticospinal facilitation of hand muscles during voluntary movement in the conscious monkey. J Physiol 381, 497–527.  164 Levine, A.J., Lewallen, K.A., and Pfaff, S.L. (2012). Spatial organization of cortical and spinal neurons controlling motor behavior. Current Opinion in Neurobiology. Lewis, T.L., Jr, Mao, T., Svoboda, K., and Arnold, D.B. (2009). Myosin-dependent targeting of transmembrane proteins to neuronal dendrites. Nat. Neurosci. 12, 568–576. Leyton, A.S.F., and Sherrington, C.S. (1917). Observations on the Excitable Cortex of the Chimpanzee, Orang-Utan, and Gorilla. Exp Physiol 11, 135–222. Li, C.X., and Waters, R.S. (1991). Organization of the mouse motor cortex studied by retrograde tracing and intracortical microstimulation (ICMS) mapping. Can J Neurol Sci 18, 28–38. Li, P., and Murphy, T.H. (2008). Two-Photon Imaging during Prolonged Middle Cerebral Artery Occlusion in Mice Reveals Recovery of Dendritic Structure after Reperfusion. Journal of Neuroscience 28, 11970–11979. Li, S., Overman, J.J., Katsman, D., Kozlov, S.V., Donnelly, C.J., Twiss, J.L., Giger, R.J., Coppola, G., Geschwind, D.H., and Carmichael, S.T. (2010). An age-related sprouting transcriptome provides molecular control of axonal sprouting after stroke. Nat. Neurosci. 13, 1496–1504. Liao, C.-C., Chen, R.-F., Lai, W.-S., Lin, R.C.S., and Yen, C.-T. (2010). Distribution of large terminal inputs from the primary and secondary somatosensory cortices to the dorsal thalamus in the rodent. J. Comp. Neurol. 518, 2592–2611. Lichtenwalner, R.J., and Parent, J.M. (2006). Adult neurogenesis and the ischemic forebrain. J. Cereb. Blood Flow Metab. 26, 1–20. Liepert, J., Bauder, H., Miltner, W.H.R., Taub, E., and Weiller, C. (2000a). Treatment- Induced Cortical Reorganization After Stroke in Humans. Stroke 31, 1210–1216. Liepert, J., Graef, S., Uhde, I., Leidner, O., and Weiller, C. (2000b). Training-induced changes of motor cortex representations in stroke patients. Acta Neurol. Scand 101, 321–326. Liepert, J., Hamzei, F., and Weiller, C. (2000c). Motor cortex disinhibition of the unaffected hemisphere after acute stroke. Muscle Nerve 23, 1761–1763. Liepert, J., Miltner, W.H.R., Bauder, H., Sommer, M., Dettmers, C., Taub, E., and Weiller, C. (1998). Motor cortex plasticity during constraint-induced movement therapy in stroke patients. Neuroscience Letters 250, 5–8. Lim, D., Mohajerani, M., LeDue, J., Boyd, J., Chen, S., and Murphy, T. (2012). In vivo large-scale cortical mapping using channelrhodopsin-2 stimulation in transgenic mice reveals asymmetric and reciprocal relationships between cortical areas. Frontiers in Neural Circuits 6,.  165 Lin, J.Y. (2012). Optogenetic excitation of neurons with channelrhodopsins: light instrumentation, expression systems, and channelrhodopsin variants. Prog. Brain Res. 196, 29–47. Lin, J.Y., Lin, M.Z., Steinbach, P., and Tsien, R.Y. (2009). Characterization of Engineered Channelrhodopsin Variants with Improved Properties and Kinetics. Biophysical Journal 96, 1803–1814. London, M., and Hausser, M. (2005). Dendritic computation. In Annual Review of Neuroscience, (Palo Alto: Annual Reviews), pp. 503–532. Longa, E.Z., Weinstein, P.R., Carlson, S., and Cummins, R. (1989). Reversible middle cerebral artery occlusion without craniectomy in rats. Stroke 20, 84–91. Lund, J.S., Hawken, M.J., and Parker, A.J. (1988). Local circuit neurons of macaque monkey striate cortex: II. Neurons of laminae 5B and 6. J. Comp. Neurol. 276, 1–29. Luo, L., Callaway, E.M., and Svoboda, K. (2008). Genetic dissection of neural circuits. Neuron 57, 634–660. Luppino, G., Matelli, M., Camarda, R.M., Gallese, V., and Rizzolatti, G. (1991). Multiple representations of body movements in mesial area 6 and the adjacent cingulate cortex: an intracortical microstimulation study in the macaque monkey. J. Comp. Neurol 311, 463–482. Luppino, G., and Rizzolatti, G. (2000). The Organization of the Frontal Motor Cortex. News Physiol. Sci. 15, 219–224. Maier, M.A., Armand, J., Kirkwood, P.A., Yang, H.-W., Davis, J.N., and Lemon, R.N. (2002). Differences in the Corticospinal Projection from Primary Motor Cortex and Supplementary Motor Area to Macaque Upper Limb Motoneurons: An Anatomical and Electrophysiological Study. Cerebral Cortex 12, 281 –296. Manns, I.D., Sakmann, B., and Brecht, M. (2004). Sub- and suprathreshold receptive field properties of pyramidal neurones in layers 5A and 5B of rat somatosensory barrel cortex. J. Physiol. (Lond.) 556, 601–622. Mansur, C.G., Fregni, F., Boggio, P.S., Riberto, M., Gallucci-Neto, J., Santos, C.M., Wagner, T., Rigonatti, S.P., Marcolin, M.A., and Pascual-Leone, A. (2005). A sham stimulation- controlled trial of rTMS of the unaffected hemisphere in stroke patients. Neurology 64, 1802–1804. Mao, T., Kusefoglu, D., Hooks, B.M., Huber, D., Petreanu, L., and Svoboda, K. (2011). Long-range neuronal circuits underlying the interaction between sensory and motor cortex. Neuron 72, 111–123. Margolis, D.J., Lütcke, H., Schulz, K., Haiss, F., Weber, B., Kügler, S., Hasan, M.T., and Helmchen, F. (2012). Reorganization of cortical population activity imaged throughout long- term sensory deprivation. Nature Neuroscience 15, 1539–1546.  166 Martin, J.H. (2012). Systems neurobiology of restorative neurology and future directions for repair of the damaged motor systems. Clin Neurol Neurosurg 114, 515–523. Masino, S.A., Kwon, M.C., Dory, Y., and Frostig, R.D. (1993). Characterization of functional organization within rat barrel cortex using intrinsic signal optical imaging through a thinned skull. Proc. Natl. Acad. Sci. U.S.A. 90, 9998–10002. Matelli, M., Luppino, G., and Rizzolatti, G. (1991). Architecture of superior and mesial area 6 and the adjacent cingulate cortex in the macaque monkey. J. Comp. Neurol 311, 445–462. Matyas, F., Sreenivasan, V., Marbach, F., Wacongne, C., Barsy, B., Mateo, C., Aronoff, R., and Petersen, C.C.H. (2010). Motor control by sensory cortex. Science 330, 1240–1243. Van Meer, M.P.A., Otte, W.M., Van der Marel, K., Nijboer, C.H., Kavelaars, A., Van der Sprenkel, J.W.B., Viergever, M.A., and Dijkhuizen, R.M. (2012). Extent of bilateral neuronal network reorganization and functional recovery in relation to stroke severity. J. Neurosci. 32, 4495–4507. Merchant, H., De Lafuente, V., Peña-Ortega, F., and Larriva-Sahd, J. (2012). Functional impact of interneuronal inhibition in the cerebral cortex of behaving animals. Prog. Neurobiol. 99, 163–178. Merchant, H., Naselaris, T., and Georgopoulos, A.P. (2008). Dynamic sculpting of directional tuning in the primate motor cortex during three-dimensional reaching. J. Neurosci. 28, 9164–9172. Monfils, M.-H., Plautz, E.J., and Kleim, J.A. (2005). In search of the motor engram: motor map plasticity as a mechanism for encoding motor experience. Neuroscientist 11, 471–483. Monfils, M.-H., and Teskey, G.C. (2004). Skilled-learning-induced potentiation in rat sensorimotor cortex: a transient form of behavioural long-term potentiation. Neuroscience 125, 329–336. Moore, C.I., and Cao, R. (2008). The hemo-neural hypothesis: On the role of blood flow in information   processing. J. Neurophysiol. 99, 2035–2047. Moran, D.W., and Schwartz, A.B. (1999). Motor Cortical Representation of Speed and Direction During Reaching. J. Neurophysiol 82, 2676 –2692. Mori, K., and Sakano, H. (2011). How is the olfactory map formed and interpreted in the mammalian brain? Annu. Rev. Neurosci. 34, 467–499. Mostany, R., Chowdhury, T.G., Johnston, D.G., Portonovo, S.A., Carmichael, S.T., and Portera-Cailliau, C. (2010). Local hemodynamics dictate long-term dendritic plasticity in peri-infarct cortex. J. Neurosci. 30, 14116–14126. Mostany, R., and Portera-Cailliau, C. (2008). A craniotomy surgery procedure for chronic brain imaging. J Vis Exp.  167 Mountcastle, V.B. (1997). The columnar organization of the neocortex. Brain 120 ( Pt 4), 701–722. Murayama, M., Pérez-Garci, E., Nevian, T., Bock, T., Senn, W., and Larkum, M.E. (2009). Dendritic encoding of sensory stimuli controlled by deep cortical interneurons. Nature 457, 1137–1141. Murphy, T.H., and Corbett, D. (2009). Plasticity during stroke recovery: from synapse to behaviour. Nat. Rev. Neurosci 10, 861–872. Murphy, T.H., Li, P., Betts, K., and Liu, R. (2008). Two-photon imaging of stroke onset in vivo reveals that NMDA-receptor independent ischemic depolarization is the major cause of rapid reversible damage to dendrites and spines. J. Neurosci. 28, 1756–1772. Murray, P.D., and Keller, A. (2011). Somatosensory response properties of excitatory and inhibitory neurons in rat motor cortex. J. Neurophysiol. 106, 1355–1362. Nagel, G., Brauner, M., Liewald, J.F., Adeishvili, N., Bamberg, E., and Gottschalk, A. (2005). Light activation of channelrhodopsin-2 in excitable cells of Caenorhabditis elegans triggers rapid behavioral responses. Curr. Biol. 15, 2279–2284. Naito, E. (2004). Sensing limb movements in the motor cortex: How humans sense limb movement. Neuroscientist 10, 73–82. Nawashiro, H., Tasaki, K., Ruetzler, C.A., and Hallenbeck, J.M. (1997). TNF-alpha pretreatment induces protective effects against focal cerebral ischemia in mice. J. Cereb. Blood Flow Metab 17, 483–490. Neafsey, E.J., Bold, E.L., Haas, G., Hurley-Gius, K.M., Quirk, G., Sievert, C.F., and Terreberry, R.R. (1986). The organization of the rat motor cortex: A microstimulation mapping study. Brain Research Reviews 11, 77–96. Neafsey, E.J., and Sievert, C. (1982). A second forelimb motor area exists in rat frontal cortex. Brain Res 232, 151–156. Nedergaard, M., Gjedde, A., and Diemer, N.H. (1986). Focal ischemia of the rat brain: autoradiographic determination of cerebral glucose utilization, glucose content, and blood flow. J. Cereb. Blood Flow Metab. 6, 414–424. Nedergaard, M., and Hansen, A.J. (1993). Characterization of cortical depolarizations evoked in focal cerebral ischemia. J. Cereb. Blood Flow Metab 13, 568–574. Nishimura, M., Shirasawa, H., and Song, W.-J. (2006). A light-emitting diode light source for imaging of neural activities with voltage-sensitive dyes. Neurosci. Res. 54, 230–234. Nitsche, M.A., Cohen, L.G., Wassermann, E.M., Priori, A., Lang, N., Antal, A., Paulus, W., Hummel, F., Boggio, P.S., Fregni, F., et al. (2008). Transcranial direct current stimulation: State of the art 2008. Brain Stimul. 1, 206–223.  168 Nitsche, M.A., Seeber, A., Frommann, K., Mein, C.C., Rochford, C., Nitsche, M.S., Fricke, K., Liebetanz, D., Lang, N., Antal, A., et al. (2005). Modulating parameters of excitability during and after transcranial   direct current stimulation of the human motor cortex. J. Physiol.-London 568, 291–303. Nudo, R.J., Jenkins, W.M., and Merzenich, M.M. (1990). Repetitive microstimulation alters the cortical representation of movements in adult rats. Somatosens Mot Res 7, 463–483. Nudo, R.J., Wise, B.M., SiFuentes, F., and Milliken, G.W. (1996). Neural substrates for the effects of rehabilitative training on motor recovery after ischemic infarct. Science 272, 1791– 1794. O’Collins, V.E., Macleod, M.R., Donnan, G.A., Horky, L.L., Van der Worp, B.H., and Howells, D.W. (2006). 1,026 experimental treatments in acute stroke. Ann. Neurol. 59, 467– 477. Ohab, J.J., Fleming, S., Blesch, A., and Carmichael, S.T. (2006). A neurovascular niche for neurogenesis after stroke. J. Neurosci. 26, 13007–13016. Ohki, K., Chung, S., Ch’ng, Y.H., Kara, P., and Reid, R.C. (2005). Functional imaging with cellular resolution reveals precise micro-architecture in visual cortex. Nature 433, 597–603. Ono, S., Chida, K., Fukaya, N., Yoshihashi, H., and Takasu, T. (1997). Dysphasia accompanied by periodic lateralized epileptiform discharges. Intern. Med. 36, 59–61. Packer, A.M., Peterka, D.S., Hirtz, J.J., Prakash, R., Deisseroth, K., and Yuste, R. (2012). Two-photon optogenetics of dendritic spines and neural circuits. Nat. Methods. De Paola, V., Holtmaat, A., Knott, G., Song, S., Wilbrecht, L., Caroni, P., and Svoboda, K. (2006). Cell type-specific structural plasticity of axonal branches and boutons in the adult neocortex. Neuron 49, 861–875. Parikh, H., Marzullo, T.C., and Kipke, D.R. (2009). Lower layers in the motor cortex are more effective targets for penetrating microelectrodes in cortical prostheses. J Neural Eng 6, 026004. Park, M.C., Belhaj-Saïf, A., Gordon, M., and Cheney, P.D. (2001). Consistent features in the forelimb representation of primary motor cortex in rhesus macaques. J. Neurosci 21, 2784– 2792. Paz, R., Boraud, T., Natan, C., Bergman, H., and Vaadia, E. (2003). Preparatory activity in motor cortex reflects learning of local visuomotor skills. Nat Neurosci 6, 882–890. Pearce, A.J., Thickbroom, G.W., Byrnes, M.L., and Mastaglia, F.L. (2000). Functional reorganisation of the corticomotor projection to the hand in skilled racquet players. Exp Brain Res 130, 238–243.  169 Pearson, K.G., Acharya, H., and Fouad, K. (2005). A new electrode configuration for recording electromyographic activity in behaving mice. J. Neurosci. Methods 148, 36–42. Penfield, W., and Boldrey, E. (1937). Somatic motor and sensory representation in the cerebral cortex of man as studied by electrical stimulation. Brain 60, 389. Perez, M.A., and Cohen, L.G. (2009). Interhemispheric inhibition between primary motor cortices: what have we learned? J Physiol 587, 725–726. Perin, R., Berger, T.K., and Markram, H. (2011). A synaptic organizing principle for cortical neuronal groups. Proc. Natl. Acad. Sci. U.S.A. 108, 5419–5424. Perron, A., Akemann, W., Mutoh, H., and Knöpfel, T. (2012). Genetically encoded probes for optical imaging of brain electrical activity. Prog. Brain Res. 196, 63–77. Petersen, C. (2007). The functional organization of the barrel cortex. NEURON 56, 339–355. Plautz, E.J., Barbay, S., Frost, S.B., Friel, K.M., Dancause, N., Zoubina, E.V., Stowe, A.M., Quaney, B.M., and Nudo, R.J. (2003). Post-infarct cortical plasticity and behavioral recovery using concurrent cortical stimulation and rehabilitative training: a feasibility study in primates. Neurol. Res 25, 801–810. Plautz, E.J., Milliken, G.W., and Nudo, R.J. (2000). Effects of repetitive motor training on movement representations in adult squirrel monkeys: role of use versus learning. Neurobiol Learn Mem 74, 27–55. Plow, E.B., Carey, J.R., Nudo, R.J., and Pascual-Leone, A. (2009). Invasive cortical stimulation to promote recovery of function after stroke: a critical appraisal. Stroke 40, 1926–1931. Poggio, T., and Bizzi, E. (2004). Generalization in vision and motor control. Nature 431, 768–774. Pons, T.P., Garraghty, P.E., Ommaya, A.K., Kaas, J.H., Taub, E., and Mishkin, M. (1991). Massive cortical reorganization after sensory deafferentation in adult macaques. Science 252, 1857–1860. Pouget, A., Dayan, P., and Zemel, R. (2000). Information processing with population codes. Nat Rev Neurosci 1, 125–132. Prigge, M., Schneider, F., Tsunoda, S.P., Shilyansky, C., Wietek, J., Deisseroth, K., and Hegemann, P. (2012). Color-tuned channelrhodopsins for multiwavelength optogenetics. J. Biol. Chem. 287, 31804–31812. Pronichev, I.V., and Lenkov, D.N. (1998). Functional mapping of the motor cortex of the white mouse by a microstimulation method. Neurosci. Behav. Physiol 28, 80–85.  170 Raineteau, O., and Schwab, M.E. (2001). Plasticity of motor systems after incomplete spinal cord injury. Nat. Rev. Neurosci. 2, 263–273. Ramanathan, D., Conner, J.M., and Tuszynski, M.H. (2006). A form of motor cortical plasticity that correlates with recovery of function after brain injury. Proc. Natl. Acad. Sci. U.S.A 103, 11370–11375. Ramanathan, D., Tuszynski, M.H., and Conner, J.M. (2009). The basal forebrain cholinergic system is required specifically for behaviorally mediated cortical map plasticity. J. Neurosci. 29, 5992–6000. Ramón y Cajal, S. (1937). Recollections of My Life (MIT Press). Rathelot, J.-A., and Strick, P.L. (2006). Muscle representation in the macaque motor cortex: an anatomical perspective. Proc. Natl. Acad. Sci. U.S.A 103, 8257–8262. Rathelot, J.-A., and Strick, P.L. (2009). Subdivisions of primary motor cortex based on cortico-motoneuronal cells. Proc. Natl. Acad. Sci. U.S.A 106, 918–923. Ratzlaff, E.H., and Grinvald, A. (1991). A tandem-lens epifluorescence macroscope: hundred-fold brightness advantage for wide-field imaging. J. Neurosci. Methods 36, 127– 137. Reimer, J., and Hatsopoulos, N.G. (2009). The problem of parametric neural coding in the motor system. Adv. Exp. Med. Biol. 629, 243–259. Reiner, A., Hart, N.M., Lei, W., and Deng, Y. (2010). Corticostriatal Projection Neurons – Dichotomous Types and Dichotomous Functions. Front Neuroanat 4,. Reis, J., Schambra, H.M., Cohen, L.G., Buch, E.R., Fritsch, B., Zarahn, E., Celnik, P.A., and Krakauer, J.W. (2009). Noninvasive cortical stimulation enhances motor skill acquisition over multiple days through an effect on consolidation. Proc. Natl. Acad. Sci. U.S.A. 106, 1590–1595. Remple, M.S., Bruneau, R.M., VandenBerg, P.M., Goertzen, C., and Kleim, J.A. (2001). Sensitivity of cortical movement representations to motor experience: evidence that skill learning but not strength training induces cortical reorganization. Behav. Brain Res. 123, 133–141. Rho, M.J., Lavoie, S., and Drew, T. (1999). Effects of red nucleus microstimulation on the locomotor pattern and timing in the intact cat: a comparison with the motor cortex. J. Neurophysiol. 81, 2297–2315. Rickgauer, J.P., and Tank, D.W. (2009). Two-photon excitation of channelrhodopsin-2 at saturation. Proc. Natl. Acad. Sci. U.S.A. 106, 15025–15030. Rijntjes, M. (2006). Mechanisms of recovery in stroke patients with hemiparesis or aphasia: new insights, old questions and the meaning of therapies. Curr. Opin. Neurol. 19, 76–83.  171 Rioult-Pedotti, M.S., Friedman, D., and Donoghue, J.P. (2000). Learning-induced LTP in neocortex. Science 290, 533–536. Rioult-Pedotti, M.S., Friedman, D., Hess, G., and Donoghue, J.P. (1998). Strengthening of horizontal cortical connections following skill learning. Nat. Neurosci. 1, 230–234. Risedal, A., Zeng, J., and Johansson, B.B. (1999). Early Training May Exacerbate Brain Damage After Focal Brain Ischemia in the Rat. Journal of Cerebral Blood Flow & Metabolism 19, 997–1003. Risher, W.C., Croom, D., and Kirov, S.A. (2012). Persistent astroglial swelling accompanies rapid reversible dendritic injury during stroke-induced spreading depolarizations. Glia 60, 1709–1720. Rizzolatti, G., and Luppino, G. (2001). The cortical motor system. Neuron 31, 889–901. Rosén, I., and Asanuma, H. (1972). Peripheral afferent inputs to the forelimb area of the monkey motor cortex: input-output relations. Exp Brain Res 14, 257–273. Rouiller, E.M., Moret, V., and Liang, F. (1993). Comparison of the connectional properties of the two forelimb areas of the rat sensorimotor cortex: support for the presence of a premotor or supplementary motor cortical area. Somatosens Mot Res 10, 269–289. Rubin, B.D., and Katz, L.C. (1999). Optical imaging of odorant representations in the mammalian olfactory bulb. Neuron 23, 499–511. Rueckel, M., Mack-Bucher, J.A., and Denk, W. (2006). Adaptive wavefront correction in two-photon microscopy using coherence-gated wavefront sensing. Proc. Natl. Acad. Sci. U.S.A. 103, 17137–17142. Saito, T., and Nakatsuji, N. (2001). Efficient gene transfer into the embryonic mouse brain using in vivo electroporation. Dev. Biol 240, 237–246. Sanes, J.N., and Donoghue, J.P. (1993). Oscillations in local field potentials of the primate motor cortex during voluntary movement. Proc Natl Acad Sci U S A 90, 4470–4474. Sanger, T.D. (2003). Neural population codes. Curr. Opin. Neurobiol. 13, 238–249. Sawaki, L., Butler, A.J., Xiaoyan Leng, Wassenaar, P.A., Mohammad, Y.M., Blanton, S., Sathian, K., Nichols-Larsen, D.S., Wolf, S.L., Good, D.C., et al. (2008). Constraint-Induced Movement Therapy Results in Increased Motor Map Area in Subjects 3 to 9 Months After Stroke. Neurorehabilitation and Neural Repair 22, 505–513. Schieber, M.H. (2001). Constraints on somatotopic organization in the primary motor cortex. J. Neurophysiol 86, 2125–2143. Schieber, M.H. (2011). Dissociating motor cortex from the motor. J. Physiol.-London 589, 5613–5624.  172 Schneider, C., Devanne, H., Lavoie, B.A., and Capaday, C. (2002). Neural mechanisms involved in the functional linking of motor cortical points. Exp Brain Res 146, 86–94. Schreiner, C.E., and Winer, J.A. (2007). Auditory Cortex Mapmaking: Principles, Projections, and Plasticity. Neuron 56, 356–365. Schummers, J., Yu, H., and Sur, M. (2008). Tuned responses of astrocytes and their influence on hemodynamic signals in the visual cortex. Science 320, 1638–1643. Schüz, A., and Palm, G. (1989). Density of neurons and synapses in the cerebral cortex of the mouse. J. Comp. Neurol. 286, 442–455. Scott, S. (2004). Optimal feedback control and the neural basis of volitional motor control. NATURE REVIEWS NEUROSCIENCE 5, 534–546. Scott, S.H., and Kalaska, J.F. (1995). Changes in motor cortex activity during reaching movements with similar hand paths but different arm postures. J. Neurophysiol. 73, 2563– 2567. Shapley, R., Hawken, M., and Ringach, D.L. (2003). Dynamics of Orientation Selectivity in the Primary Visual Cortex and the Importance of Cortical Inhibition. Neuron 38, 689–699. Sheets, P.L., and Shepherd, G.M. (2011). Cortical Circuits for Motor Control. Neuropsychopharmacology 36, 365–366. Sheets, P.L., Suter, B.A., Kiritani, T., Chan, C.S., Surmeier, D.J., and Shepherd, G.M.G. (2011). Corticospinal-specific HCN expression in mouse motor cortex: Ih-dependent synaptic integration as a candidate microcircuit mechanism involved in motor control. J. Neurophysiol. 106, 2216–2231. Shepherd, G.M.G. (2009). Intracortical cartography in an agranular area. Front Neurosci 3, 337–343. Shepherd, G.M.G., Pologruto, T.A., and Svoboda, K. (2003). Circuit analysis of experience- dependent plasticity in the developing rat barrel cortex. Neuron 38, 277–289. Shih, A.Y., Blinder, P., Tsai, P.S., Friedman, B., Stanley, G., Lyden, P.D., and Kleinfeld, D. (2012). The smallest stroke: occlusion of one penetrating vessel leads to infarction and a cognitive deficit. Nature Neuroscience. Shih, A.Y., Li, P., and Murphy, T.H. (2005). A small-molecule-inducible Nrf2-mediated antioxidant response provides effective prophylaxis against cerebral ischemia in vivo. J. Neurosci. 25, 10321–10335. Shinoda, Y., Yokota, J., and Futami, T. (1981). Divergent projection of individual corticospinal axons to motoneurons of multiple muscles in the monkey. Neurosci. Lett. 23, 7–12.  173 Shipp, S. (2005). The importance of being agranular: a comparative account of visual and motor cortex. Phil. Trans. R. Soc. B 360, 797–814. Siebner, H.R., and Rothwell, J. (2003). Transcranial magnetic stimulation: new insights into representational cortical plasticity. Exp Brain Res 148, 1–16. Sigler, A., Goroshkov, A., and Murphy, T.H. (2008). Hardware and methodology for targeting single brain arterioles for photothrombotic stroke on an upright microscope. J. Neurosci. Methods 170, 35–44. Silberberg, G., and Markram, H. (2007). Disynaptic inhibition between neocortical pyramidal cells mediated by Martinotti cells. Neuron 53, 735–746. Simeral, J.D., Kim, S.-P., Black, M.J., Donoghue, J.P., and Hochberg, L.R. (2011). Neural control of cursor trajectory and click by a human with tetraplegia 1000 days after implant of an intracortical microelectrode array. J Neural Eng 8, 025027. Simon, M.V., Michaelides, C., Wang, S., Chiappa, K.H., and Eskandar, E.N. (2010). The effects of EEG suppression and anesthetics on stimulus thresholds in   functional cortical motor mapping. Clin. Neurophysiol. 121, 784–792. Sincich, L.C., and Horton, J.C. (2005). THE CIRCUITRY OF V1 AND V2: Integration of Color, Form, and Motion. Annu. Rev. Neurosci. 28, 303–326. Siucinska, E., and Kossut, M. (1996). Short-lasting classical conditioning induces reversible changes of representational maps of vibrissae in mouse SI cortex--a 2DG study. Cereb. Cortex 6, 506–513. Smith, K.S., Virkud, A., Deisseroth, K., and Graybiel, A.M. (2012). Reversible online control of habitual behavior by optogenetic perturbation of medial prefrontal cortex. Proc. Natl. Acad. Sci. U.S.A. 109, 18932–18937. Squire, P.L.R. (2003). Fundamental Neuroscience (Academic Press). Stehfest, K., and Hegemann, P. (2010). Evolution of the channelrhodopsin photocycle model. Chemphyschem 11, 1120–1126. Stepniewska, I., Fang, P.-C., and Kaas, J.H. (2005). Microstimulation reveals specialized subregions for different complex movements in posterior parietal cortex of prosimian galagos. Proc. Natl. Acad. Sci. U.S.A. 102, 4878–4883. Stepniewska, I., Preuss, T.M., and Kaas, J.H. (1994). Thalamic connections of the primary motor cortex (M1) of owl monkeys. The Journal of Comparative Neurology 349, 558–582. Stoney, S., Thompson, W., and Asanuma, H. (1968). Excitation of Pyramidal Tract Cells by Intracortical Microstimulation - Effective Extent of Stimulating Current. J. Neurophysiol. 31, 659–&.  174 Stosiek, C., Garaschuk, O., Holthoff, K., and Konnerth, A. (2003). In vivo two-photon calcium imaging of neuronal networks. Proc. Natl. Acad. Sci. U.S.A. 100, 7319–7324. Strick, P.L. (2002). Stimulating research on motor cortex. Nat Neurosci 5, 714–715. Strick, P.L., and Preston, J.B. (1982). Two representations of the hand in area 4 of a primate. II. Somatosensory input organization. J. Neurophysiol. 48, 150–159. Sweetnam, D., Holmes, A., Tennant, K.A., Zamani, A., Walle, M., Jones, P., Wong, C., and Brown, C.E. (2012). Diabetes Impairs Cortical Plasticity and Functional Recovery Following Ischemic Stroke. J. Neurosci. 32, 5132–5143. Swindale, N.V. (1996). The development of topography in the visual cortex: a review of models. Network 7, 161–247. Swindale, N.V. (2000). How many maps are there in visual cortex? Cereb. Cortex 10, 633– 643. Swindale, N.V., Shoham, D., Grinvald, A., Bonhoeffer, T., and Hübener, M. (2000). Visual cortex maps are optimized for uniform coverage. Nat. Neurosci. 3, 822–826. Tandon, S., Kambi, N., and Jain, N. (2008). Overlapping representations of the neck and whiskers in the rat motor cortex revealed by mapping at different anaesthetic depths. Eur. J. Neurosci 27, 228–237. Tennant, K.A., Adkins, D.L., Donlan, N.A., Asay, A.L., Thomas, N., Kleim, J.A., and Jones, T.A. (2010). The Organization of the Forelimb Representation of the C57BL/6 Mouse Motor Cortex as Defined by Intracortical Microstimulation and Cytoarchitecture. Cereb Cortex. Tennant, K.A., Adkins, D.L., Scalco, M.D., Donlan, N.A., Asay, A.L., Thomas, N., Kleim, J.A., and Jones, T.A. (2012). Skill learning induced plasticity of motor cortical representations is time and age-dependent. Neurobiol Learn Mem 98, 291–302. Teskey, G.C., Monfils, M.-H., VandenBerg, P.M., and Kleim, J.A. (2002). Motor map expansion following repeated cortical and limbic seizures is related to synaptic potentiation. Cereb. Cortex 12, 98–105. Tian, L., Hires, S.A., Mao, T., Huber, D., Chiappe, M.E., Chalasani, S.H., Petreanu, L., Akerboom, J., McKinney, S.A., Schreiter, E.R., et al. (2009). Imaging neural activity in worms, flies and mice with improved GCaMP calcium indicators. Nat. Methods 6, 875–881. Todorov, E. (2000). Direct cortical control of muscle activation in voluntary arm movements: a model. Nature Neuroscience 3, 391–398. Toga, A.W., and Mazziotta, J.C. (2002). Brain Mapping: The Methods (Academic Press).  175 Tokimura, H., Di Lazzaro, V., Tokimura, Y., Oliviero, A., Profice, P., Insola, A., Mazzone, P., Tonali, P., and Rothwell, J.C. (2000). Short latency inhibition of human hand motor cortex by somatosensory input from the hand. J. Physiol. (Lond.) 523 Pt 2, 503–513. Trachtenberg, J.T., Chen, B.E., Knott, G.W., Feng, G., Sanes, J.R., Welker, E., and Svoboda, K. (2002). Long-term in vivo imaging of experience-dependent synaptic plasticity in adult cortex. Nature 420, 788–794. Tufail, Y., Matyushov, A., Baldwin, N., Tauchmann, M.L., Georges, J., Yoshihiro, A., Tillery, S.I.H., and Tyler, W.J. (2010). Transcranial pulsed ultrasound stimulates intact brain circuits. Neuron 66, 681–694. Urbain, N., and Deschênes, M. (2007). Motor cortex gates vibrissal responses in a thalamocortical projection pathway. Neuron 56, 714–725. Vargas-Irwin, C.E., Shakhnarovich, G., Yadollahpour, P., Mislow, J.M.K., Black, M.J., and Donoghue, J.P. (2010). Decoding Complete Reach and Grasp Actions from Local Primary Motor Cortex Populations. J. Neurosci. 30, 9659–9669. Veinante, P., and Deschênes, M. (2003). Single-cell study of motor cortex projections to the barrel field in rats. J. Comp. Neurol. 464, 98–103. Velliste, M., Perel, S., Spalding, M.C., Whitford, A.S., and Schwartz, A.B. (2008). Cortical control of a prosthetic arm for self-feeding. Nature 453, 1098–1101. Voller, B., St Clair Gibson, A., Dambrosia, J., Pirio Richardson, S., Lomarev, M., Dang, N., and Hallett, M. (2006). Short-latency afferent inhibition during selective finger movement. Exp Brain Res 169, 226–231. Wall, N.R., Wickersham, I.R., Cetin, A., Parra, M.D.L., and Callaway, E.M. (2010). Monosynaptic circuit tracing in vivo through Cre-dependent targeting and complementation of modified rabies virus. PNAS 107, 21848–21853. Wang, H., Peca, J., Matsuzaki, M., Matsuzaki, K., Noguchi, J., Qiu, L., Wang, D., Zhang, F., Boyden, E., Deisseroth, K., et al. (2007). High-speed mapping of synaptic connectivity using photostimulation in Channelrhodopsin-2 transgenic mice. Proceedings of the National Academy of Sciences 104, 8143 –8148. Wang, L., Conner, J.M., Rickert, J., and Tuszynski, M.H. (2011). Structural plasticity within highly specific neuronal populations identifies a unique parcellation of motor learning in the adult brain. Proc Natl Acad Sci U S A. Ward, N.S., Brown, M.M., Thompson, A.J., and Frackowiak, R.S.J. (2003). Neural correlates of motor recovery after stroke: a longitudinal fMRI   study. Brain 126, 2476–2496. Watson, B.D., Dietrich, W.D., Busto, R., Wachtel, M.S., and Ginsberg, M.D. (1985). Induction of reproducible brain infarction by photochemically initiated thrombosis. Ann. Neurol 17, 497–504.  176 Weiler, N., Wood, L., Yu, J., Solla, S.A., and Shepherd, G.M.G. (2008). Top-down laminar organization of the excitatory network in motor cortex. Nat Neurosci 11, 360–366. Weinberger, N.M. (2007). Associative representational plasticity in the auditory cortex: a synthesis of two disciplines. Learn. Mem. 14, 1–16. Wessberg, J., Stambaugh, C.R., Kralik, J.D., Beck, P.D., Laubach, M., Chapin, J.K., Kim, J., Biggs, S.J., Srinivasan, M.A., and Nicolelis, M.A.L. (2000). Real-time prediction of hand trajectory by ensembles of cortical neurons in primates. Nature 408, 361–365. Wickersham, I.R., Lyon, D.C., Barnard, R.J.O., Mori, T., Finke, S., Conzelmann, K.-K., Young, J.A.T., and Callaway, E.M. (2007). Monosynaptic Restriction of Transsynaptic Tracing from Single, Genetically Targeted Neurons. Neuron 53, 639–647. Winship, I.R., and Murphy, T.H. (2008). In vivo calcium imaging reveals functional rewiring of single somatosensory neurons after stroke. J. Neurosci 28, 6592–6606. Witte, O.W. (1998). Lesion-induced plasticity as a potential mechanism for recovery and rehabilitative training. Curr. Opin. Neurol 11, 655–662. Witte, O.W., and Stoll, G. (1997). Delayed and remote effects of focal cortical infarctions: secondary damage and reactive plasticity. Adv Neurol 73, 207–227. Wolpaw, J.R., and McFarland, D.J. (2004). Control of a two-dimensional movement signal by a noninvasive   brain-computer interface in humans. Proc. Natl. Acad. Sci. U. S. A. 101, 17849–17854. Wong, J., Abilez, O.J., and Kuhl, E. (2012). Computational optogenetics: A novel continuum framework for the photoelectrochemistry of living systems. Journal of the Mechanics and Physics of Solids 60, 1158–1178. Wyatt, R.M., Tring, E., and Trachtenberg, J.T. (2012). Pattern and not magnitude of neural activity determines dendritic spine stability in awake mice. Nat. Neurosci. 15, 949–951. Xerri, C., Merzenich, M.M., Peterson, B.E., and Jenkins, W. (1998). Plasticity of primary somatosensory cortex paralleling sensorimotor skill recovery from stroke in adult monkeys. J. Neurophysiol. 79, 2119–2148. Xu, H.-T., Pan, F., Yang, G., and Gan, W.-B. (2007). Choice of cranial window type for in vivo imaging affects dendritic spine turnover in the cortex. Nat. Neurosci. 10, 549–551. Xu, T., Yu, X., Perlik, A.J., Tobin, W.F., Zweig, J.A., Tennant, K., Jones, T., and Zuo, Y. (2009). Rapid formation and selective stabilization of synapses for enduring motor memories. Nature 462, 915–919. Yanai, Y., Adamit, N., Harel, R., Israel, Z., and Prut, Y. (2007). Connected corticospinal sites show enhanced tuning similarity at the onset of voluntary action. J. Neurosci. 27, 12349–12357.  177 Yang, G., Pan, F., and Gan, W.-B. (2009). Stably maintained dendritic spines are associated with lifelong memories. Nature 462, 920–924. Yizhar, O., Fenno, L.E., Prigge, M., Schneider, F., Davidson, T.J., O’Shea, D.J., Sohal, V.S., Goshen, I., Finkelstein, J., Paz, J.T., et al. (2011). Neocortical excitation/inhibition balance in information processing and social dysfunction. Nature 477, 171–178. Yoder, E.J., and Kleinfeld, D. (2002). Cortical imaging through the intact mouse skull using two-photon excitation laser scanning microscopy. Microsc. Res. Tech. 56, 304–305. Young, N.A., Vuong, J., Flynn, C., and Teskey, G.C. (2011). Optimal parameters for microstimulation derived forelimb movement thresholds and motor maps in rats and mice. J. Neurosci. Methods 196, 60–69. Yu, J., Anderson, C.T., Kiritani, T., Sheets, P.L., Wokosin, D.L., Wood, L., and Shepherd, G.M.G. (2008). Local-Circuit Phenotypes of Layer 5 Neurons in Motor-Frontal Cortex of YFP-H Mice. Front Neural Circuits 2, 6. Yu, X., and Zuo, Y. (2011). Spine plasticity in the motor cortex. Curr. Opin. Neurobiol. 21, 169–174. Zhang, F., Aravanis, A.M., Adamantidis, A., De Lecea, L., and Deisseroth, K. (2007a). Circuit-breakers: optical technologies for probing neural signals and systems. Nat. Rev. Neurosci 8, 577–581. Zhang, F., Prigge, M., Beyrière, F., Tsunoda, S.P., Mattis, J., Yizhar, O., Hegemann, P., and Deisseroth, K. (2008). Red-shifted optogenetic excitation: a tool for fast neural control derived from Volvox carteri. Nat. Neurosci. 11, 631–633. Zhang, F., Wang, L.-P., Brauner, M., Liewald, J.F., Kay, K., Watzke, N., Wood, P.G., Bamberg, E., Nagel, G., Gottschalk, A., et al. (2007b). Multimodal fast optical interrogation of neural circuitry. Nature 446, 633–639. Zhang, F., White, J.G., and Iadecola, C. (1994). Nitric oxide donors increase blood flow and reduce brain damage in focal ischemia: evidence that nitric oxide is beneficial in the early stages of cerebral ischemia. J. Cereb. Blood Flow Metab. 14, 217–226. Zhang, S., Boyd, J., Delaney, K., and Murphy, T.H. (2005). Rapid reversible changes in dendritic spine structure in vivo gated by the degree of ischemia. J. Neurosci. 25, 5333–5338. Zhang, S., and Murphy, T.H. (2007). Imaging the impact of cortical microcirculation on synaptic structure and sensory-evoked hemodynamic responses in vivo. PLoS Biol. 5, e119. Zhang, Z.G., and Chopp, M. (2009). Neurorestorative therapies for stroke: underlying mechanisms and translation to the clinic. Lancet Neurol 8, 491–500.  178 Zhao, S., Ting, J.T., Atallah, H.E., Qiu, L., Tan, J., Gloss, B., Augustine, G.J., Deisseroth, K., Luo, M., Graybiel, A.M., et al. (2011). Cell type–specific channelrhodopsin-2 transgenic mice for optogenetic dissection of neural circuitry function. Nat. Methods 8, 745–752. Ziemann, U., Paulus, W., Nitsche, M.A., Pascual-Leone, A., Byblow, W.D., Berardelli, A., Siebner, H.R., Classen, J., Cohen, L.G., and Rothwell, J.C. (2008). Consensus: Motor cortex plasticity protocols. Brain Stimul. 1, 164–182. Zorzos, A.N., Scholvin, J., Boyden, E.S., and Fonstad, C.G. (2012). Three-dimensional multiwaveguide probe array for light delivery to distributed brain circuits. Opt Lett 37, 4841– 4843. (2004). Neuroscience (Sinauer Associates).    

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}]}"
                            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.24.1-0073652/manifest

Comment

Related Items