UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The dendritic morphology of local patch and callosal neurons in the superficial layers of cat visual… Thejomayen, D. Moira 1994

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

Item Metadata


831-ubc_1994-894431.pdf [ 3.24MB ]
JSON: 831-1.0088165.json
JSON-LD: 831-1.0088165-ld.json
RDF/XML (Pretty): 831-1.0088165-rdf.xml
RDF/JSON: 831-1.0088165-rdf.json
Turtle: 831-1.0088165-turtle.txt
N-Triples: 831-1.0088165-rdf-ntriples.txt
Original Record: 831-1.0088165-source.json
Full Text

Full Text

THE DENDRITIC MORPHOLOGY OF LOCAL PATCH AND CALLOSAL NEURONSIN THE SUPERFICIAL LAYERS OF CAT VISUAL CORTEX (AREA 18)ByD. MOIRA THEJOMAYENB.V.Sc., The University of Peradeniya, Sri Lanka, 1985A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFDOCTOR OF PHILOSOPHYinTHE FACULTY OF GRADUATE STUDIESFACULTY OF MEDICINEDEPARTMENT OF ANATOMYWe accept this thesis as confirming to the required standardTHE UNIVERSITY OF BRITISH COLUMBIAApril 1994© D. Moira Thejomayen, 1994In presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.(Signature)_______________Department of AThe University of British ColunbiaVancouver, CanadaDate A’ni 2, IDE-6 (2188)ABSTRACTUsing the double-labeling technique with fluorescent tracers local patch and callosalneurons in area 18 of the cat were studied. The main goal was to determine whether singlecortical neurons participate in both the local patch networks and callosal projection. Callosaland local patch neurons were found in overlapping and interdigitating patch networks.Double-labeled cells were not found in the regions of overlap between the two populationsof labeled cells, or in any other locations in area 18, suggesting that the callosal and localpatch neurons form separate neuronal populations in the superficial layers of cat area 18.The morphology of local patch neurons was next examined and compared to callosalneurons in the superficial layers of area 18. The local patch neurons were mostly spinypyramidal neurons including modified, small to medium sized standard, and star pyramids.Smooth, multipolar neurons, presumably of the basket cell type were also present. In thetangential plane, the basal dendritic trees of about half of the local patch pyramidal neuronsin layer 2/3 had mediolaterally elongated fields and the rest had circular fields. The callosalneurons in layer 2/3 were standard, fusiform, and star pyramids. Callosal standardpyramids were, on average, larger than the local patch standard pyramids based on meanvalues of somatic areal measurements. The mean dendritic field width of the callosal cellswas also larger than that of the local cells. Callosal cells also possesed a more complexdendritic branching pattern, with numerous branch points. Both features appear to beindependent of cell size. Finally, almost all the callosal cells had closely circular dendriticfields when viewed in the tangential plane. Thus, the local patch and callosal neurons canbe distinguished from each other by morphological and morphometric parameters.11TABLE OF CONTENTSAbstract iiTable of Contents iiiList of Figures viList of Abbreviations xAcknowledgement xiDedication xiiCHAPTER 1: INTRODUCTION 11.1. Historical Overview 11.2. General organization of the cerebral cortex 41.3. Neuronal organization of the visual cortex 61.3.1. Columnar structure of visual cortex 61.3.2. Lamination of the visual cortex of cat 81.3.3. Types of neurons 111.3.4. Afferent and Efferent connections 191.3.5. Intrinsic connections in visual cortex 221.4. Structure and function of axons and dendrites in cortex 231.4.1. Relationship of dendrites and axons to cortical laminae 231.4.2. Normal development of dendrites in sensory cortex 251.4.3. Effects of deprivation on the development of dendritesand axons in sensory cortex 271.5. Patchy connections in visual cortex 291.5.1. Intrinsic local connections 291.5.2. Corticocortical connections 311.5.3. Callosal connections 321.5.4. Functional significance of patchy intrinsic connections111in visual cortex.331.5.5. Patchy networks and their relationship to each other 341.6. Rationale 36CHAPTER 2: EXPERIMENTAL PROCEDURES 392.1. Injection of retrograde tracers and preparation of brain slices 392.1.1. Local patch population 392.1.2. Callosal population 432.2. Intracellular injection with Lucifer Yellow (LY) -Local patch and Callosal populations 442.2.1. Preparation of micropipettes 442.2.2. Injection technique 442.3. Processing of tissue for subsequent examination, imagingand analysis - Local patch and callosal populations 452.4. Data analysis - Local patch cells and callosal cells 452.5. Double-labeling study 47CHAPTER 3: RESULTS 493.1. Morphological characteristics of local patch neurons 493.1.1. General findings 493.1.2. Soma size and location 493.1.3. Diversity of cell shapes 563.1.3.1. Pyramidal cells 563.1.3.2. Nonpyramidal cells 653.1.4. Organization of basal dendrites in coronal, sagittaland tangential planes 723.2. Morphological characteristics of callosal neurons 853.2.1. General Findings 85iv3.2.2. Soma size and location.853.2.3. Diversity of cell shapes 943.2.4. Organization of basal dendrites in coronal andtangential planes 1053.3. Comparison of the morphology of local and callosal neurons 1193.3.1. Cell types 1193.3.2. Organization of basal dendrites 1193.4. Double-labeling study 129CHAPTER 4: DISCUSSION 1344.1. Introductory remarks 1344.2. Summary of the findings 1364.3. Technical considerations 1384.3.1. Advantages of using fixed slices 1394.3.2. Advantages of confocal imaging 1414.4. Comparison of the intrinsic and extrinsic neurons in layer 2/3 1424.5. Dendritic organization of local and callosal cells:Possible functional implications 1444.6. Future Directions 1484.7. Conclusion 150REFERENCES 155VLIST OF FIGURES1. A, Schematic diagram, and B, Photograph of a Nissl section illustrating thelamination of the cat visual cortex, area 18 102. A, Photograph of a cat brain showing the visual cortex, area 17 and 18, andthe part of area 18 used in this study. B, Schematic diagram illustrating thedifferent planes of sectioning of area 18 used in this study 423. Lower-power project image of the fluorescein-dextran labeled neurons in alocal patch in area 18 514. Higher-power project image of layer 2/3 neurons in a local patch, retrogradelylabeled with fluorescein-dextrans 535. Histogram showing the somatic area of the local patch neurons injected withLY in this study 556. Somatic area of the LY-filled local patch neurons in layer 2/3 graphed as afunction of their distance from the pial surface 587. Project image of an LY-filled local patch modified pyramidal neuron oflayer 2 608. Project images of two LY-filled local patch small pyramidal neurons oflayer 2 629. Project images of two LY-filled local patch star pyramidal neurons oflayer 3 6410. Camera lucida drawings of two LY-filled local patch standard pyramidalneurons of layer 3 6711. A, Lower-power project image of a local patch at the level of layer 3. B,Higher-power project image of the layer 3 nonpyramidal neuron withsmooth dendrites seen in Figure 1 1A 6912. Project image of an LY-filled large layer 3 multipolar neuron, with beadedvidendrites .7113. A, Project image of a local patch taken at the level of lower layer 3 showingnonpyramidal neurons (arrows) at the border of the patch. B, Project imageof another local patch with a nonpyramidal neuron (arrow) within the patch 7414. Histogram showing the number of primary basal dendrites of intrinsic localpatch neurons injected with LY 7615. A, Graph showing the dendritic field width of LY-filled local patch pyramidalneurons in layer 2/3 with respect to their soma size. B, Histogram illustratingthe width of the basal dendritic trees of 80 local patch standard pyramidal cellsin layer 2/3 7916. Basal dendritic fields of LY-filled layer 2/3 local patch pyramidal cells in thetangential plane 8117. A, Graph showing the mediolateral (ML) and anteroposterior (AP) dendriticfield widths of 46 local patch pyramids of layer 2/3, in the tangential plane.B, Scattergram showing the distribution of ML:AP and AP:ML ratio valuesof the basal dendritic fields of local patch cells in the tangential plane 8318 Lower-power project image of rhodamine dextran-labeled callosal neuronsin area 18 8719. Higher-power project images of rhodamine dextran-labeled callosal neuronsin area 18 taken at the level of layer 3 8920. Histogram showing the somatic area of the callosal neurons injected withLY in this study 9121. Somatic area of the LY-fihled callosal neurons in layer 2/3 graphed as afunction of their distance from the pial surface 9322. Higher-power project image of the basal dendrites of a callosal standardpyramidal neuron which shows spines 9623. Project image of a LY-filled callosal standard pyramidal neuron of layer 3,viiin coronal plane .9824. Project image of a LY-filled callosal fusiform pyramidal neuron of layer 3,in coronal plane 10025. Project image of a callosal star pyramidal neuron of layer 3, in coronalplane 10226. Project image of a callosal pyramidal neuron in coronal plane with dendritesoriginating from the upper portions of the cell body 10427. Project image of a callosal pyramidal neuron in coronal plane withbifurcated apical dendrite 10728. Project image of a callosal pyramidal neuron in coronal plane with shortapical dendrite 10929. Histogram showing the number of primary dendrites of callosal neuronsinjected with LY 11130. Graph showing the dendritic field width of the callosal pyramidal neurons inlayer 2/3 with respect to their soma size 1133 1. Basal dendritic fields of LY-filled callosal pyramidal cells of layer 3 in thetangential plane 11632. A, Graph showing the mediolateral (ML) and anteroposterior (AP) widthsof the basal dendritic fields of callosal cells in the tangential plane.B, Scattergram showing the distribution of ML:AP and AP:ML ratio valuesof the basal dendritic fields of callosal cells in the tangential plane 11833. Graph showing the soma sizes of the local patch and callosal cells in layer2/3 with respect to their distance from the pial surface 12234. Graph showing the number of branch points of callosal and local patchpyramidal cells in layer 2/3 with respect to their soma size 12435. Graph showing the basal dendritic field widths of the callosal and localpatch pyramidal cells in layer 2/3 with respect to their soma size 126viii36. A, Graph showing the mediolateral (ML) and anteroposterior (AP) widthsof the basal dendritic fields of local patch and callosal cells in the tangentialplane. B, Scattergram showing the distribution of ML:AP and AP:ML ratiovalues of the basal dendritic fields of local patch and callosal cells in thetangential plane 12837. A chart which shows retrogradely labeled local patch and callosal cells in anoverlap region of area 18 in a double-labeling experiment 13138 . Confocal dual channel images of rhodamine- and fluorescein- dextranlabeled area 18 neurons in A, a double-labeling experiment and B, a controlexperiment 133ixLIST OF ABBREVIATIONSAP: AnteroposteriorCO: Cytochrome oxidase2-DG: 2-deoxyglucoseEM: Electron microscopeGABA: Gamma-amino butyric acidHRP: Horseradish peroxidaseLGN: Lateral geniculate nucleusLS: Lateral suprasylvian areaLY: Lucifer YellowML: MediolateralPB: Phosphate bufferPHA-L: Phaseolus vulgaris lectinPMLS: Posteromedial lateral suprasylvian areaWGA-HRP: Wheat germ agglutinin conjugated to HRPxACKNOWLEDGEMENTI wish to express my deepest gratitude to my supervisor, Dr. Joanne A. Matsubara,for her enthusiastic encouragement, knowledgeable guidance, patience, and supportthroughout the course of my research in her laboratory. The confidence that Dr. Matsubarashowed in my abilities allowed me to complete this study and made the course of research arewarding process.I am very greatful to the members of my thesis advisory committee, Dr. WayneVogl, Dr. Joanne Weinberg, and Dr. John OKusky for their valuable suggestions andconstructive comments which made this project interesting and relevant.I would like to express my sincere thanks to Dr. Rob Douglas, Dr. Gavin Thurstonand Bill Kiss for helpful advice on confocal imaging. I also wish to thank all the othermembers of the departments of Anatomy and Ophthalmology who have helped methroughout my graduate education.Special thanks are due to Jamie Boyd and Jin Zhang for technical assistance andhelpful suggestions throughout this study.My sincerest appreciation is extended to Dr. Ronald Leslie for the part played byhim in starting my post-graduate career.The research presented here was supported by grants awarded to Dr. Joanne A.Matsubara from the Medical Research Council of Canada and by Scholarship providedfrom the Association of Universities and Colleges of Canada, for which I would like toexpress my appreciation.xiThis thesis is dedicated to my husband Thejomayen Sammy for all the support, love andencouragement he has given me over the years and to my son Michael Thejomayen.xiiCHAPTER 1INTRODUCTION1.1 HISTORICAL OVERVIEWInvestigation of the cerebral cortex began in the eighteenth century, but the earlyefforts to study the anatomical organization of the cortex were hindered by the lack ofsuitable staining techniques. Golgi discovered the silver stain in 1873 and by 1883, he haddescribed pyramidal and stellate cells and made clear distinctions between axons anddendrites in cortex (Golgi 1883). The invention of the Golgi method provided investigatorswith a powerful tool to examine the morphology of individual neurons and the relations thatmay exist between them. In the early 1 900s, Cajal and his pupil, Lorente de No, used thisnew technique and provided detailed morphological descriptions about the anatomicalorganization of the cerebral cortex (Cajal 1911; Lorente de No 1933, 1949). Cajal usedmainly human brain tissue, supplemented by some tissue from rodents and cats. Lorente deNo confirmed many of the cell types described by Cajal and made pertinent observationsabout the relationships of dendrites of certain pyramidal cells and the terminals ofthalamocortical axons. Since then, systematic analyses of Golgi—stained cortical neuronaltypes continued over the years. There are a number of descriptive Golgi studies subsequentto Lorente de NO’s which made significant contributions towards the classification of thecortical neurons. These include the study of the anatomical organization of the striate cortexby O’Leary (1941), the first quantitative assesment of the shape of neurons and theirdendritic organization by Sholl (1953), a program which was later continued by Colonnier(1964), and the studies of Lund (1973) who defined two classes of nonpyramidal cells,spiny and nonspiny, and the studies of Jones (1975) who defined distinctive cell classes ofnonpyramidal cells based on their axonal arborizations.1In the middle of this century, a new phase of descriptive morphology using theGolgi technique emerged with the invention of the electron microscope. Gray (1959) madethe first significant contribution with the use of this new microscope. He identifiedasymmetrical and symmetrical synapses (which he called types I and II) on the basis ofpostsynaptic membrane differentiations and described dendritic spines as biological entitiesrather than, as some previous reports had suggested, artifacts of the Golgi technique. Grayalso found that the two types of synapses were located on different parts of the neurons.Subsequently, Hamlyn (1963) found similar distribution of Gray’s type I and type IIsynapses in the hippocampus, and Hamori and Szentágothai (1965) reported that the basketcell terminals form Gray’s type II synapses on the soma of Purkinje cells. At this timeaxosomatic contacts were shown to be inhibitory (Andersen et al. 1963) and axodendriticsynapses were shown to be excitatory (Andersen and Lomo 1966). These observations ledto the generalization that Gray’s type II synapses are inhibitory and Gray’s type I synapsesare excitatory. With the introduction of aldehyde mixtures as the primary fixative of choicein electron microscopy, two further observations of significance were made by Colonnier(1968), which subsequently formed the basis of many fine structural analyses of thecerebral cortex. Colonnier (1968) showed that the axon terminals making Gray’sasymmetrical synapses contained spherical shaped synaptic vesicles, while those makingsymmetrical synapses contained flattened or pleomorphic vesicles. He also pointed out thatpyramidal and nonpyramidal cells could be distinguished in the numbers and proportions ofdifferent types of synapses they receive.The introduction of microelectrode recording enabled electrophysiologists toinvestigate the properties of individual neurons and interpret the functional organization ofthe cerebral cortex. Over the past two decades, electrophysiological studies have providedsignificant contributions to our understanding of the function of the cortex at the single celllevel. Direct electrical stimulation of the brain, or evoked potential recordings from it, have2shown that the cerebral cortex is parcelled into functional areas whose relative positions aresimilar in all mammalian species (Woolsey 1964; Welker 1976). It is largely due the effortsinitiated by these earlier studies that we now accept the notion that the cerebral cortex ofdifferent species is built according to some common, basic plan and that the arrangement ofcell types and input and output pathways are strikingly similar across species.The more recent development of sensitive and reliable microassays for enzymes,and the generation of specific antisera against neurotransmitters and their receptors haveprovided an explosive knowledge of the neurochemical organization of the cerebral cortex.The knowledge of different aspects of cortical organization, namely the neurochemical,neuroanatomical and neurophysiological, provide a more integrative picture of the circuitryby which various neuronal types communicate.More recent technical innovations include the labeling of cortical cells by retrogradeand anterograde axoplasmic tracers and the classification of individual cell classes in termsof their receptive field properties by intracellular marking techniques. The tract tracingstudies done with the use of retrograde axonal transport method have shown that thepyramidal cells are the major output cells of the cortex and the nonpyramidal cells areusually intrinsic (e.g., Gilbert and Kelly 1975; Lund et al. 1975; Jones and Wise 1977).These studies have identified the cells of origin of most corticofugal pathways from avariety of areas and have shown that the different layers of the cortex have different outputconnections. Intracellular marking technique was commenced by Kelly and Van Essen(1974). They utilized intracellular recording to define the cellts receptive field propertiesand then used intracellular injection of the fluorescent dye, procion yellow, as the cellmarker to identify the cells morphology. This work has been continued by Gilbert andWiesel (1979), Martin et al. (1983), Somogyi et al. (1983), Kisvárday et al. (1983) and byMartin and Whitteridge (1984a) using horseradish peroxidase (HRP) as the intracellular3marker. Intracellular injection of HRP has the advantage that the synaptic connections ofHRP-filled neurons can also be examined with the electron microscope. With so many newtechnical innovations there is hope that in the near future we may understand why corticalcells adopt individual morphologies and connectivities, why they express differenttransmitters and how the cortical circuits act to produce functional properties of corticalcells. Understanding how information is processed in cerebral cortex greatly depends onthe ability to identify its different types of neurons and the sources and the types ofsynapses linking them.1.2. GENERAL ORGANIZATION OF THE CEREBRAL CORTEXThe cerebral cortex is characterized by a highly ordered anatomical arrangementwhich has provided a framework for anatomical and physiological investigations. About90% of the cerebral cortex is made up of neocortex, the remainder is paleocortex andarchicortex (e.g., Nolte 1988). Almost all the cortex seen from the outside of the brain isneocortex which has increased enormously in size and surface area by the appearance offissures and convolutions in higher mammals and man. The paleocortex is restricted to thatpart of the cerebral cortex which forms part of the olfactory system. The archicortexcomprises the hippocampal formation which is a component of the limbic system. Theneocortex is also referred to as the homogenetic cortex or isocortex since its different partsdevelop in the same manner through a six-layered structure. In contrast, the paleocortexand archicortex have fewer than six layers, mostly three layers, and are thus collectivelycalled the heterogenetic cortex or the allocortex (e.g., Nolte 1988).In spite of the common features of all cortical areas, clear-cut differences instructure were recognized among various areas of cortex. Different layers were seen toexhibit varying thicknesses, varying densities of cells and varying sizes and types of cells.4On the basis of these structural differences, several regions were identified within thecortex by a number of anatomists. One labeling system most extensively used is that ofBrodmann (1909), who divided the cortex of each hemisphere into 52 cytoarchitectonicareas.In line with the system used by Brodmann (1909), each cytoarchitectonic area ofthe cortex can be divided into six basic horizontal layers based on the size and packingdensity of the neuronal cell bodies it contains. Axonal processes that arise locally or thatenter the area from other regions of the brain also contribute to cortical lamination. Basedon local variations, the six basic laminae are often further divided into sublayers.Anatomical studies have shown that individual layers can be distinguished not only by cellsize and density, but also by the unique morphological cell types they contain (e.g., Cajal1911;LorentedeNó 1933; Lund 1973).The layers are named roughly according to the resident cell type. Layer 1 is themolecular (or plexiform) layer which is the most superficial and cell-poor layer. Layer 2 isthe external granular layer which contains densely packed small pyramids and some stellatecells. Layer 3 is the external pyramidal layer which predominantly consists of mediumsized and moderately large superficial pyramids. Layer 4 is the internal granular layerwhich predominantly consists of stellate cells and some pyramidal cells. Layer 5 is theinternal pyramidal layer which is composed mainly of the large deep pyramids. Layer 6 isthe multiform layer which contains predominantly polymorphic or fusiform cells (e.g.,Barr 1979, Nolte 1988).The basic six layered arrangement is clearest in areas of the neocortex referred to ashomotypic cortex and is somewhat modified in areas termed as heterotypic cortex(Brodmann 1909). Heterotypic cortex can be divided into a granular type characteristic of5sensory areas and an agranular type characteristic of motor areas. Granular cortex ischaracterized by its richness of granular cells and a well developed layer 4. In agranularcortex granular cells are nearly absent and pyramidal cells dominate layers 2 through 5 tothe extent that individual layers are no longer obvious (e.g, Barr 1979, Nolte 1988).Coexisting with the horizontal laminar arrangement of the cortex, is a vertical aspectof organization that was first revealed by electrophysiological methods (Mountcastle 1957;Abeles and Goldstein 1970; Hubel and Wiesel 1962, 1977). The finding that vertical arraysof neurons display similar response properties was first reported by Mountcastle (1957) insomatosensory cortex. Subsequently, vertical arrays of cells that preferentially respond tosame or closely similar frequencies in auditory cortex were made (Abeles and Goldstein1970). In 1962, Hubel and Wiesel found vertical arrays of cells that preferentially respondto stimuli presented to one eye. The vertical arrangements of cells which respondedpreferentially to visual stimuli viewed through one eye were called ocular dominance‘columns’.1.3. NEURONAL ORGANIZATION OF THE VISUAL CORTEX1.3.1. COLUMNAR STRUCTUREThe first demonstration of functional columns using anatomical methods wasprovided by Hubel and Wiesel (1969). They demonstrated ocular dominance columns inmonkey visual cortex by making lesions in single laminae of the lateral geniculate nucleus(LGN) and staining the resulting terminal degeneration in layer 4 of striate cortex with theNauta method. They found that the afferents from the LGN segregate into roughly 0.5-mm-wide, alternating left- and right- eye dominated columns in layer 4. Subsequently,ocular dominance columns were demonstrated by Shatz et al. (1977) and LeVay et al.(1978) using transneuronal transport of radioactive substances injected into an eye.6Shortly after discovering that most neurons in the mammalian visual cortexresponded optimally to oriented line segments, Hubel and Wiesel (1962) found that cellswith similar orientation preferences were grouped together in columns which were referredto as orientation ‘columns’. Orientation columns were demonstrated in the monkey (Hubelet al. 1978), and cat (Albus 1979; Löwel et al. 1987), by autoradiography of cortical tissuemetabolically activated by specifically oriented stimuli, following intravascularadministration of radioactively tagged 2-deoxyglucose (2-DG). 2-DG is a glucose analoguethat labels cells which are activated specifically by the visual stimulus presented. Studies onfunctional ‘columns’ have shown that in many areas of cortex they are arranged more likebands, especially when viewed from the surface or in the tangential plane of section.Another set of anatomically defined columnar structures, known as the cytochromeoxidase (CO) blobs, extending mainly through layers 2 and 3, has been observed in visualcortex stained for the mitochondrial enzyme cytochrome oxidase (Wong-Riley 1979), anenzyme associated with cortical zones of high metabolic activity. CO blobs have beendemonstrated in primary visual cortex of the monkey (Horton and Hubel 1981; Livingstoneand Hubel, 1984a) and cat (Murphy et al. 1990; Dyck and Cynader 1992; Boyd andMatsubara, submitted). The CO staining revealed distinct processing streams in monkeywhich deal with different types of information, for example, color and form. Physiologicalstudies have shown that cells situated within the CO blobs are usually color selective, andonly a minority are orientation selective, whereas cells between CO blobs (interbiobneurons) respond best to orientation selective stimuli, and only a minority are colorselective (Livingstone and Hubel, l984b). Anatomical studies have shown that the COblobs and interblobs have unique corticocortical and intracortical connectivity, i.e, the samecompartments interconnect. Blobs in area 17 are reciprocally connected selectively to a setof CO-rich, stripe-like regions in area 18, whereas the interblobs of area 17 are reciprocally7connected to CO-poor interstripes in area 18 (Livingstone and Hubel 1984a; Hendrickson1985). Intrinsic connections are made reciprocally among CO blobs or interblobs but notbetween them, in striate cortex of monkey (Livingstone and Hubel 1984b) and cat (Boydand Matsubara 1992).1.3.2. LAMINATION OF THE VISUAL CORTEX OF CATDetailed descriptions about the lamination of the visual cortex of the cat have beenprovided for the primary visual cortex, area 17, by O’Leary (1941) and by Otsuka andHassler (1962) and for the secondary visual area, area 18, by Otsuka and Hassler (1962).According to O’Leary’s scheme of area 17, layer 1 is a cell deficient layer which liesimmediately under the pial surface. Layers 2 and 3 are usually considered together becausethe boundary between them is difficult to determine with certainty in Nissl preparations.They contain an abundance of small to medium pyramids and few stellates and are limitedinferiorly by moderately large pyramidal cells. The first three layers form roughly one-thirdof the depth of the cat visual cortex. Layer 4 is wide and occupies the middle one-third ofthe cortical thickness. It is divided into sublayer 4a which contains large stellate cells andstar pyramids and sublayer 4b which contains small stellate cells. The lower one-third ofthe cat visual cortex is divided approximately equally between layers 5 and 6. Layer 5contains a lower density of cells than layer 6 and is divided into sublayer 5a which containssmaller pyramidal cells and sublayer Sb which contains large pyramidal cells. Layer 6contains an abundance of medium sized cells with pyramidal or fusiform cell bodies. Thebasic layering pattern of area 18 is similar to that of area 17, and therefore, O’Leary’sscheme is used in my study (Figure 1).8Figure 1. A, Schematic diagram illustrating the lamination and the main cell types present inindividual layers of cat area 18 according to the scheme of O’Leary (1941). The relativesizes of pyramidal and stellate cells are shown in the laminar distribution. B,Photomicrograph of a Nissi section through the cat visual cortex, area 18, cut in the coronalplane showing the six basic horizontal layers.9A12 A’A AAAAA3AAA4a*4b5 AAAAAAAAA6AAA>4h.. :.S.a—B101.3.3. TYPES OF NEURONSThe classical categorization of cortical neurons was based mainly on themorphology of the cell body and dendritic trees, and to a lesser extent on the morphologyof the axon. In recent years, several other features have been used to categorize neurons.Among these are the brain region to which the axon projects, the type of synapse made andthe transmitter released by the neuron. The neurons of the cerebral cortex can be classifiedinto spiny neurons which include both pyramidal neurons and spiny stellates, and smoothor sparsely spiny nonpyramidal neurons. In general, the spiny cells and the smooth cellshave characteristic synaptic patterns. Electron microscopic examination of Golgi-stainedneurons demonstrated that the axons of the spiny neurons generate only asymmetricsynapses with other neurons and thus the postsynaptic effects of the spiny cells areexcitatory. The distribution of synapses received by the spiny cells is also typical (Gray1959; Colonnier 1968; LeVay 1973; Somogyi 1978). The cell body and the dendritic shafts(particularly the initial portions) receive only symmetrical synapses, whereas the dendriticspines receive generally asymmetrical synapses. In contrast, the smooth or sparsely spinyneurons receive both symmetrical and asymmetrical synapses intermixed on soma anddendrites (Colonnier 1968; LeVay 1973; Somogyi 1977; Peters and Fairén 1978). Theaxons of the smooth and sparsely spiny nonpyramidal neurons make symmetrical synapsesonto other neurons and thus the postsynaptic effects of these cells are inhibitory.Pyramidal cellsPyramidal cells are present in cortical layers 2 through 6 and have the followingfeatures in common: A cell body possesing a triangular profile (hence its name), a singledominant apical dendrite with a larger diameter than the other dendrites, which originatesfrom the apex of the cell body and extends towards the pial surface, and several basal11dendrites, which radiate more or less horizontally or obliquely downwards from the base ofthe cell body. The dendrites of the pyramidal cells are thickly covered with smallappendages, the dendritic spines. The axon of the pyramidal cells originates typically fromthe base of the cell body, or less frequently from the proximal portion of a basal dendrite,and projects into the white matter, giving off collateral branches on the way.The apical dendrites of pyramidal cells give off few branches, usually two to five,in an obliquely outward direction, on their way towards the pial surface, near which mostapical dendrites branch out extensively to form an apical tuft. The vertically oriented apicaldendrite, which is a prominent, identifying feature of the pyramidal cells, typically crossesseveral cortical layers and thus is well placed to receive input from the multiplicity of axonalpathways that are known to ramify within each cortical layer. This arrangement, whichallows for the integration by single neurons of a large variety of lamina-specific inputs, canbe considered a crucial aspect in the design of pyramidal cells.The side branches of the apical dendrites exhibit laminar specificity (Lorente de No1949) and it has been suggested that this pattern may in turn be related to the laminarspecificity of the various axon terminals present within the cortex (Lorente de No 1949;Lund and Boothe 1975). An example in which pyramidal cell dendrites appear to bespecifically distributed such that they preferentially receive certain inputs and not others isshown by layer 6 pyramidal cells of monkey visual cortex. The apical dendrites of theselayer 6 cells give off many oblique side branches in layers 5 and 3 but display a lack of sidebranches on their ascent through two sublaminae of layer 4, 4CB and layer 4A, the samesublayers that receive input from the parvocellular sublayers of LGN (Lund and Boothe1975). A similar laminar specificity also seems to apply to the length of the apical dendriteof pyramidal cells. For example, Lorente de No (1949) described pyramidal cells of layer 5and 6 whose apical dendrites terminated within or below layer 4 and he called these cells12the ‘short’ pyramidal cells. Lund and Booth (1975) based on Golgi studies in monkeyvisual cortex also have reported short pyramidal cells in layer 5 and 6.The basal dendritic trunks (primary branches) of pyramidal cells shortly afterleaving the cell body, branch out profusely in all directions and form secondary and tertiarybranches. A significant literature has developed on the study of the organization of basaldendritic fields as a whole, and on the detailed, quantitative analysis of the length ofindividual dendritic segments, the total length of dendrites, the number of dendriticbranches and the number of branch points (e.gs., Sholl 1953; Lindsay and Scheibel 1974;Smit and Uylings 1975). Most quantitative studies on the organization of basal dendritesare based on the method devised by Sholl (1953). Dendrites are stained with Golgitechnique, and the analysis is performed on drawings. Concentric circles, 20 pm apart andcentered on the cell body are superimposed on the drawing. The area enclosed by adjacentcircles is referred to as a shell. Dendritic branches and branch points are counted withinsuccessive shells and the number of dendritic intersections cutting each circle across ismeasured and quantified.Pyramidal cells project to subcortical structures and to other cortical areas of thesame and opposite hemispheres (Gilbert and Kelly 1975; Lund et al. 1975; Albus et al.1981; Hornung and Garey 1981; Segraves and Rosenquist 1982; Symonds and Rosenquist1984; Segraves and Innocenti 1985). Study of pyramidal cell subcortical projections, usingmodern tracing techniques, has shown a clear pattern in the laminar specificity of thepyramidal cell populations giving rise to various subcortical projections in cat visual cortex(Gilbert and Kelly 1975) and monkey visual cortex (Lund et al. 1975). Pyramidal cellswith axons that do not enter the white matter have also been observed and the output ofthese pyramidal cells is presumed to be entirely local or to adjacent cortical areas. Examplesof such cells are described in cat visual cortex (Gilbert and Wiesel 1983; Katz 1987) and13monkey visual cortex (Lund and Boothe 1975). Gilbert and Wiesel (1983) described alayer 2 pyramidal cell that had extensive intrinsic axon collaterals over 4 mm but no efferentaxon entering white matter. Katz (1987) described layer 6 pyramidal cells that lackedefferent axons. These cells have axons which leave the soma and proceed towards thewhite matter, and then turn towards the pia by making an abrupt U-turn. Before the U-turnand at the bottom of the U, the axon sends off several thin axon collaterals that turntowards the pia or arborize within layer 6 and occasionally in layer 5. Lund and Boothe(1975) have described layer 5 pyramidal cells that lacked efferent axon in the monkey.These layer 5 cells have only recurrent axons that ramify in the supragranular layers.Included in the category of pyramidal neurons are modified pyramids (OLeary1941), star pyramids (Lorente de No 1949), and inverted pyramids (Van der Loos 1965).Modified pyramids have very short apical dendrites which break up close to the cell bodyand form extensive terminal tufts (O’Leary 1941). Their basal dendrites ramify extensivelyand some extend to layer 1. Star pyramids appear similar to the spiny stellate cells in thatboth have dendrites originating from the cell body radially in all directions, but unlike thespiny stellate cells, star pyramids posses a slender apical dendrite that extends to layer 1(Lorente de No 1949). Inverted pyramidal cells resemble classical (standard) pyramidalcells except, as their name suggests, appear upside down so that their apical dendrites aredirected toward the white matter (Van der Loos 1965).Spiny stellate cellsSpiny stellate cells occur exclusively within the layer 4 of the primary sensory areasof many species (Lund 1984), and are most abundant in the visual cortex. They differ fromthe pyramidal cells in that they do not posses an apical dendrite. The dendrites of spinystellate cells are richly covered with spines, a feature that distinguishes from the other14classes of stellate cells. As discussed earlier in this section, the synaptic patterns of spinystellates are similar to pyramidal cells, and the postsynaptic effects of both cells areexcitatory. The dendrites of spiny stellates are for the most part restricted to layer 4 but theyoccasionally enter other layers (Lund 1973). Thus, it is clear that most of the input to themmust derive from the axonal pathways that terminate within this layer, and the thalamicafferents form the principal source of input to these cells. In contrast, pyramidal cells havedendrites that enable them to receive input from afferents that terminate within severalcortical laminae. Although some spiny stellate cells in the cat primary visual cortex havebeen shown to project to adjacent cortical areas (Meyer and Albus 1981), by and large, theaxons ramify within layer 4 or within adjacent regions of layers 3 and 5 (Lorente de No1949; Lund 1973; Lund et al. 1979; Gilbert and Wiesel 1979). Thus, the axons of spinystellate cells generally have a local quality. Axons of pyramidal cells, on the other hand,project not only locally, but also to laminae and regions of the brain situated at considerabledistances from their parent cell bodies. According to Lund et al. (1979) there are twosubtypes of spiny stellate cells in cat primary visual cortex. Those present in layer 4a arelarger in size; their axons which project horizontally for long distances, send verticalbranches into the superficial and deep layers. Those present in layer 4b are smaller in size;their axons which arborize mainly within layer 4, only occasionally send branches to thedeep layers.Smooth nonpyramidal cellsNeurons that lack an apical dendrite and are spine-free are considered to be smoothnonpyramidal cells, but included within this broad category are types of neurons that differmarkedly in the morphology of their cell bodies and in the shapes and patterns of theirdendrites and axons. Various classification schemes have been used for classifyingnonpyramidal neurons, but because of its simplicity, the one advanced by Feldman and15Peters (1978) has proven especially useful. According to this classification scheme,nonpyramidal cells are classified as multipolar, bipolar or bitufted according to the shapeof their dendritic tree and as smooth or sparsely spiny according to the concentration oftheir dendritic spines.Multipolar cells have dendrites that emerge from multiple points on the cell bodyand radiate out in all directions. Bipolar and bitufted neurons emit, respectively, either twodendrites or two groups of dendrites from the opposite poles of the cell bodies that tend tobe ovoid or spindle shaped. A more comprehensive scheme can be achieved by taking intoaccount the form and distribution of the axonal ramifications of the nonpyramidal neurons.As a general rule, the axonal ramifications of nonpyramidal cells do not leave the area ofcortex in which their parent cell body is situated and are usually considered local circuitelements.Basket cells are large multipolar cells having either smooth or sparsely spinydendrites that often extend well above and below the layer containing the parent cell body(Cajal 1911; Peters and Regidor 1981; Defelipe and Fairén 1982; Martin et al. 1983). Thedendrites of some basket cells are beaded (Somogyi et al. 1983). Basket cells are located inlayers 2 through 5 but are found mainly in layers 3 and 5 (Peters and Regidor 1981;Somogyi et al. 1983). The dendrites of basket cells radiate in all directions, but the verticalones predominate such that some cells have vertically elongate dendritic fields and abitufted appearance. The axons of these cells initially ascend or descend upon leaving thecell body and then run horizontally for as much as a millimeter (Jones and Hendry 1984).Thus, basket cells may provide a basis for long range horizontal inhibitory interactionsacross the cortex. Recent observations of DeFelipe and Fairén (1982) and Martin et al.(1983) in cat visual cortex clearly show that the axons of basket cells form symmetricsynapses. Initial descriptions of the basket cells emphasized the association of their axonal16branches, which form pericellular nests’ or “baskets” with the cell bodies and proximaldendrites of pyramidal cells (Caj al 1911). However, recent evidence demonstrates that lessthan half of the synapses made by the basket cells are on the cell bodies and proximaldendrites of pyramidal cells. The remainder of the synapses are with cell bodies anddendrites of other nonpyramidal neurons and with spines and distal dendrites of pyramidalneurons (Somogyi et al. 1983; Freund et al. 1986). Evidence that the basket cells areGAB Aergic has been obtained by the immunohistochemical staining of synaptic terminalsof basket cells that have been intracellularly filled with HRP (Somogyi and Soltész 1986;Kisvárday et al. 1987).Chandelier cells have been observed in layers 2 through 5 in a variety of corticalareas and species (Jones 1975; Szentagothai 1975; Fairén and Valverde 1980; Lund et al.1981; Somogyi et al. 1982). They may occur as multipolar or bitufted varieties even withina single region of the cortex. Their dendrites, which may span one or several layers of thecortex, have few branches and bear only occasional spines. Typically, the axons ofchandelier cells form a profuse plexus in the vicinity of the cell body, coextensive with orsomewhat above or below the distribution of the cell’s dendritic tree. Chandelier cells havebeen so named because of the resemblance of their axonal ramifications and verticallyoriented arrays of axon terminals to the branches and candles of the chandelier. Somechandelier cells in the superficial layers of the cortex may also possess a second, lessextensive axonal ramification in the deeper layers of the cortex (Fairén and Valverde 1980).Chandelier cell axons form symmetrical synapses which are presynaptic only to the axoninitial segments of pyramidal cells (Somogyi 1977; Somogyi et al. 1979; Fairén andValverde 1980; Freund et al. 1983; DeFelipe et al. 1985), and for this reason these cellshave also been referred to as axoaxonic interneurons (Somogyi et al. 1982). GABAergicchandelier cells have been identified by combining Gogi impregnation with GABAimmunohistochemistry (Freund et al. 1983; Somogyi et al. 1985). The preferential17distribution of chandelier axon terminals on axon initial segments, that is, the trigger zonefor the initiation of action potentials, provides chandelier cells with the ability to exert apowerful inhibition on the postsynaptic neurons.The name “double bouquet” has been applied to neurons whose dendriticarborizations exhibit a variety of different patterns and spine densities. But based on thedistribution of their local axon collaterals and the type of synapses they form, the doublebouquet cells may be divided into two basic types Axons of the first type form ascendingand descending collaterals and ramify in laminae situated both above and below the parentcell body, their axon terminals form asymmetrical synapses (Fairén et al. 1984). Their cellbodies occur in layers 2 through 5 and their bitufted dendritic fields tend to be verticallyelongated and coextensive with their axonal ramifications. In contrast, axons of the secondtype of double bouquet cell form mainly descending axonal arborizations and ramify for themost part within laminae well beneath their dendritic arborizations, their axon terminalsform symmetrical synapses (Fairén et al. 1984). Their cell bodies occur only within layers2 and 3. Examples of this type of double bouquet cell have been shown to contain GABAin the visual cortex of the cat (Somogyi and Cowey 1981; Somogyi et al. 1981).Bipolar cells have been identified in the cortices of several species including thevisual cortex of the rat (Peters and Kimerer 1981), cat (Fairén et al. 1984) and monkey(Peters 1984) and are found predominantly in layer 4 of cat visual cortex. These cells sharemany morphological features with double bouquet cells. Single primary dendrites normallyemerge from the two opposite poles of the elongated cell body and branch only sparsely.The dendrites are either smooth or sparsely spiny and their extent is varied; some cells mayspan most of the cortical thickness but others may be contained within only two corticallayers. The axon tends to arise from one of the primary dendrites at some distance from thesoma. Bipolar cells form very narrow axonal fields, crossing many cortical layers, and it18has been suggested that these cells may serve to synchronize the output of cells residingwithin a cortical column. Like the double bouquet cells, some types of bipolar cells formasymmetrical synapses whereas others form symmetrical, presumably GABAergic,synapses (Peters and Harriman 1988).1.3.4. AFFERENT AND EFFERENT CONNECTIONSVarious cortical afferents branch out and form many synaptic contacts within thecortex, but the density of branches and contacts are not homogenous throughout the corticaldepth (Jones 1981). The thalamic afferents concentrate their terminals in the middle corticallayers, mostly in layer 4 of the sensory cortices. In different cortical areas and in differentanimals there are numerous variations on this general scheme. For instance, in the primaryvisual cortex of the cat, the specific thalamic afferents terminate not only in layer 4 but alsoin lower layer 3 and in the upper half of layers 6 and 1 (Levay and Gilbert 1976). It hasbeen shown that the afferents from individual laminae of the dorsal LGN project todifferent cortical layers (Levay and Gilbert 1976). The A geniculate laminae (A and Al)was shown to project in two main bands, one extending from the bottom of layer 4 tolower layer 3 and the other to layer 6. The C geniculate laminae was also shown to projectin two dense bands, one to the upper and lower borders of layer 4, bracketing the Alaminae projection, with some overlap, and the other to the upper half of layer 1. Inaddition, the study of LeVay and Gilbert (1976) shows that the LGN afferents from eachgeniculate laminae segregate into patches of about 500 urn width, which probably form theanatomical basis for ocular dominance columns. The LGN afferents to cat visual cortexarise from three major classes of neurons, X-, Y- and W cells, which receive inputs fromtheir retinal counterparts, X-,Y- and W ganglion cells (Sherman 1985, for review). The Alaminae of the LGN receive inputs from X- and Y ganglion cells and the C laminae receiveinputs from Y- and W ganglion cells (Leventhal 1982). The X cells of the cat LGN appears19to project exclusively to area 17, the Y- and W cells project additionally to other visualcortical areas, for example, areas 18, 19 and PMLS (Sherman 1985, for review). Recentstudies (Boyd and Matsubara, in preparation) from our lab shows that the A geniculatelaminae (X- and Y cells) form patchy projections to layers 4 (A and B) and 6, in area 17 ofcat visual cortex, in agreement with the findings of LeVay and Gilbert (1976). However, Cgeniculate laminae (Y- and W cells) were observed to project to layers 4A only, 3 and 5A,in area 17. The C laminae projection was also patchy and the patches aligned with COblobs. When the W cells were selectively labeled by injection in the lower part of Claminae, the patchy labeling was only present in layers 3, 5A and top half of 1. In area 18,the C laminae projection was found to terminate in patches in layers 3, 4 (A and B) and 5A,and in 1, and the patches again aligned with CO blobs (Boyd and Matsubara, inpreparation).Recent studies using Golgi-electron microscopic procedure have shown that inrodents, at least, thalamic afferents may terminate on the dendrites of any cell type withintheir termination zones (Peters et al. 1976). This observation has been confirmed in cat andprimate by many studies using the same or different techniques (e.g., Hornung and Garey1981; Freund et al. 1985a,b; Jones 1986; Kisvardy et al 1986; Somogyi and Soltész 1986).Thus, the assumption that thalamic afferents arrive at layer 4 and terminate only on localcircuit neurons which then serve as the sole relays to the various output pyramidal neuronslocated in other layers has become difficult to accept. As suggested by Jones (1988), itseems essential now to consider both direct pathways of thalamic afferent fibers on outputcells and indirect pathways between thalamic input and output which pass throughnumerous local circuits involving various chains of neurons.The other major inputs to the cortical areas are ipsilateral corticocortical afferentsand contralateral callosal afferents. These afferents tend to form dense connections in layer204 and all the superficial layers (Shatz 1977; Jones 1981; Segraves and Rosenquist 1982b;Henry et al. 1991). Thus, they overlap the thalamic fiber termination zones, but the extentto which these afferents terminate on the same cortical cells as those receiving thalamicafferents is not yet known (Jones 1981). The study of Henry et al. (1991) in cat visualcortex shows that the laminar distribution of the terminals of the corticocortical afferentsdiffer depending on the areas which are involved in the connection. The study shows thatthe corticocortical afferents from area 18 to area 17 terminate heavily in layers 1, 2, 3, 4Aand to a lesser extent in 5 but rarely in 6, whereas, the afferents from area 19 to area 17terminate mostly in layers 5, 6 and to a lesser extent in 1, 2 and 3, but almost completelyavoid layer 4. Thus, the corticocortical projection from area 19 to 17 has been observed tobe organized like a standard feed back projection that connects a higher order area to alower order area (Rockland and Pandya 1979; Maunsell and Van Essen 1983).The cortical efferents are also not homogenously distributed, and the outputs ofcortical regions tend to be anatomically segregated. For example, certain subcorticalprojections and a number of corticocortical projections are now known to arise fromdifferent cortical layers in the visual cortex of the cat (Gilbert and Kelly 1975) and monkey(Lund et al. 1975). There is a fairly clear segregation of cortical efferent projectionsbetween the supra- and infra- granular layers. Most of the ipsilateral corticocorticalefferents and the contralateral callosal efferents originate from the pyramidal cells of layers2 and 3. The subcortical efferents to the tectum originate from the pyramidal cells of layer5. The subcortical efferents that project back to the specific thalamic nuclei originate mostlyfrom pyramidal cells in layer 6 (e.g., Gilbert and Kelly 1975).211.3.5. INTRINSIC CONNECTIONS IN VISUAL CORTEXIt is now clear that numerous neuronal circuits can be identified in the cortex andthat some or all may function in determining the manner in which cortical cells respond toperipheral stimuli. A series of interlaminar and horizontal connections, has beendemonstrated, using a number of techniques, such as Golgi, degeneration, axoplasmictransport of tracers, and intracellular filling of axons of single neurons.Interlaminar connectionsThe vertical interlaminar projections start with the layer that receives thalamic input.A systematic flow of information from one cortical layer to another occurs through theseconnections. The thalamic input to the cortex arrives mainly to layer 4. Spiny stellates inlayer 4 project predominantly to layers 2 and 3. Pyramidal cells in layers 2 and 3 project toother cortical areas and give off substantial collaterals that project to layer 5. Pyramidal cellsin layer 5 project outside the cortex, but some have axon collaterals that extend to layer 6and travel laterally over long distances. Pyramidal cells in layer 6 project out of the cortex,but these have collaterals that ascend and terminate densely in layer 4 forming a closed ioop(Gilbert 1983). The interlaminar connectivity has been thought to underlie sequentialelaborative hierarchies in the visual cortex. For example, the widely extended collateralprojection from layer 5 to layer 6 has been implicated in the formation of the extremelylarge receptive fields of layer 6 visual cortical neurons (Gilbert 1983).Horizontal connectionsIn recent years it has become increasingly evident that there is also a significantamount of horizontal interaction, covering quite large distances. The original work22investigating horizontal connections in the cortex was done with degeneration techniques(Fisken et al. 1975, Creutzfeldt et al. 1975). Recently, the horizontal connections havebeen demonstrated using HRP as an extracellular marker. These horizontal connectionshave been shown to be patchy in nature (Rockland and Lund 1982) (see section 1.5.1).Intracellular injections of HRP, which allow the investigation of these connections at thesingle cell level also have also shown that the connections formed by individual cells canextend for considerable distances in the horizontal direction, and that the axons often formpatchy collaterals (Gilbert and Wiesel 1979; Martin and Whitteridge 1984). Gilbert andWiesel (1979) describe a layer 5 pyramidal cell in cat visual cortex that has a local axoncollateral that travels horizontally some 6-8 mm within layer 6. This layer 5 pyramidalneuron has an axon that enters the white matter, and a second axon collateral ascending tolayer 2/3. Long, intracortical collaterals were also subsequently observed to be emitted bypyramidal cells in the superficial layers of the cat visual cortex (Gilbert and Wiesel 1983;Martin and Whitteridge 1984; Kisvárday and Eysel 1992). One of these superficial layercells has long horizontal axon collaterals both in layer 2/3 and in layer 5, which are givenoff in clusters and the clusters of layer 2/3 were directly over the clusters of layer 5 (Gilbertand Wiesel 1983).1.4. STRUCTURE AND FUNCTION OF DENDRITES AND AXONS INCORTEX1.4.1. RELATIONSHIP OF DENDRITES AND AXONS TO CORTICALLAMINAESpatial relationships between the dendrites of pyramidal neurons and corticallaminae were investigated systematically by Lorente de No (1949), who observed that boththe length of the apical dendrite and the specific cortical laminae within which its sidebranches occur can be correlated with the laminar location of the parent cell body. Lund and23Boothe (1975) examined the lamina-neuronal relationships in detail and showed thatpyramidal neurons within a single layer can be divided into morphological subclassesaccording to the laminar distribution of their apical dendritic side branches. Theseanatomical studies using Golgi technique have distinguished morphological subclasses ofpyramidal cells based on their apical dendritic branching pattern. Subsequent studies ofKatz (1987) using the combination of retrograde tracing and intracellular injection in liveslices have shown that neurons within a morphological subclass may project to a singletarget area. The studies of Katz (1987) showed that in cat visual cortex, apical dendrites oflayer 6 pyramids that project to claustrum give off side branches mainly in layer 5 beforeterminating in layer 1, whereas apical dendrites of layer 6 pyramids that project to the LGNbranch extensively in layers 5 and 4 and terminate within or beneath lower layer 3.Specific laminar relationships have also been shown to characterize the spatialdistribution of basal dendrites belonging to pyramidal cells. For example, in all corticallayers, the basal dendrites of pyramidal cells tend to arborize preferentially within thelamina containing the parent cell body. The specificity is shown in an even more strikingmanner in monkey visual cortex by the pyramidal cells of sublayer 4A which have basaldendrites that turn sharply downward from the cell body to enter sublayer 4B where theyfan out horizontally, thus, avoiding any ramification within layer 4A (Lund 1973).The local axon collaterals of pyramidal cells exhibit a wide variety of branchingpatterns, and at first glance little order can be detected within them. However, careful studyof axonal ramifications provides considerable support to the belief that the distribution ofaxonal branches within the cortex is as systematic as that of the dendrites (Lorente de No1949) and follows certain rules. The arrangement of axon collaterals appears to be relatedin some way to the laminar location of the parent cell body. For instance, pyramidal cellsin the deep layers (5 and 6) often have ascending collaterals that ramify extensively within24the superficial layers (2 and 3) (Cajal 1911; Lorente de No 1938; Lund and Boothe 1975).Conversely pyramidal cells of superficial layers have been shown to typically senddescending collateral projections to layer 5 (Lund and Boothe 1975; Gilbert and Wiesel1979; Gilbert and Wiesel 1983). The pattern of axon collaterals has also been shown to berelated to the terminal site to which its principal axon projects. For example, Katz (1987)reports that pyramidal cells in layer 6 of cat visual cortex which project to claustrum havehorizontally projecting collaterals that arborize exclusively within layer 6 and lower layer 5.By contrast, most pyramids which project to LGN lack horizontal collateral branches inlayer 6, but they have vertical collaterals which arborize extensively in layer NORMAL DEVELOPMENT OF DENDRITES IN SENSORYCORTEXThe dendritic and axonal systems of cortical neurons appear to have preferentialrather than random orientations in their pattern of arborization. And, it has been establishedthat characteristic and highly organized patterns of dendrites in adult cortex can be achievedthrough selective elimination and/or preservation of dendrites. For example, according tothe studies of Greenough et al. (1988) in mouse somatosensory cortical barrels, thedevelopment of oriented dendritic fields within the barrel system appears to involve bothselective elimination of improperly oriented dendrites and selective growth of properlyoriented dendrites. Barrels are three-dimensional multicellular units in layer 4 of primarysomatosensory cortex of certain rodents (Woolsey and Van der Loos 1970). They receiveinput from vibrissa on the animal’s snout and consist of cell-dense cylindrical walls andcell-sparse hollows where thalamocortical afferents are concentrated (Simons and Woolsey1984). In adults, the orientation of dendrites in the barrel is such that, the dendriticbranches of the cells in the wall largely project toward the hollow. Greenough et al. (1988)analysed the dendrites of barrel neurons by superimposing a set of concentric rings on the25camera lucida drawings of barrel neurons and recorded the number of ring intersections ofdendrites towards the hollow and away from the hollow. They found that the number ofintersections increased towards the hollow and decreased away from the hollow as afunction of age. Thus, both the addition of dendrites on the hollow side and the eliminationof dendrites away from the hollow may be important in the development of orienteddendritic fields of barrel neurons. Selective dendritic elimination has also been shown toplay a role in shaping the distinct laminar arrangements of the apical dendrites of layer 5callosal neurons in rats (Koester and O’Leary 1992). Koester and O’Leary (1992) havedemonstrated that callosal and corticotectal neurons of layer 5 initially send apical dendritesto layer 1, but later in development the apical dendritic segments of layer 5 callosal neuronssuperficial to layer 4 are actively eliminated creating their characteristic short pyramidalmorphology, while the corticotectal neurons retain their apical dendrites to layer 1.Studies have shown that the spatial extent of dendrites in the normal adult animal isalso related to the segregation pattern of afferent fibers in the visual cortex. For example,Lund (1984) reported that the dendrites of several classes of neurons in layer 4 of the visualcortex of tree shrew and macaque monkey were restricted to single sublaminae of layer 4,possibly as a result of the segregation of parvo- and magnocellular thalamocortical afferentsinto these sublaminae. More recently, by combining fluorescent demonstration of eyedominance columns with the intracellular filling of neurons in vitro, Katz et al. (1989)reported that the dendrites of layer 4C8 cells near the borders between eye dominancecolumns appear to remain preferentially in the same eye dominance column as its cell body.This dendritic pattern of cells in layer 403, the recipient zone of parvocellular afferents,shows that the dendrites can also distinguish among parvocellular inputs from the two eyeswithin the same laminae. Thus, the segregated pattern of thalamocortical afferents appearsto have considerable influence on the spatial extent of dendrites of cells in layer 4CB.261.4.3. EFFECTS OF DEPRIVATION ON THE DEVELOPMENT OFDENDRITES AND AXONS IN SENSORY CORTEXDendrites and axons of cortical neurons have the capacity to undergo plastic changes inresponse to changes in their environment. The idea that neurons are able to modify theirstructure and pattern of connectivity was originally proposed by Cajal (1911) whoobserved plasticity of neurons, following spinal cord transection in young cats and dogs.Cortical neurons have been shown to reorganize their dendrites after the ablation of sensoryafferents. For example, in the visual cortex, the dendrites of layer 4 spiny stellates typicallyramify within layer 4 with branches in layers 3 and 5, but in mice enucleated at birth, thelayer 4 stellates tend to direct their dendrites away from layer 4 into adjacent layers 3 and 5,as if they were seeking for other afferents outside layer 4 (Valverde 1968). Anotherexample is in the mouse somatosensory cortex after ablation of individual vibrissae. Here,the dendritic territories of neurons in layer 4 stellate cells were shown to expand intodeafferented areas in response to the altered pattern of thalamic afferents (Harris andWoolsey 1979; Steffen and Vander Loos 1980). An interesting example of changes indendritic shape induced by deafferentation is demonstrated in auditory system of thechicken (Deitch and Rubel 1984, 1989). The dendrites of neurons in nucleus laminaris, abrainstem auditory nucleus, are segregated into dorsal and ventral dendritic tufts. Thedorsal and ventral tufts are known to receive spatially separated innervation from theipsilateral and contralateral nucleus magnocellularis respectively. Transection of afferents tothe ventral dendrites has been shown to result in rapid and specific atrophy of ventraldendrites with no change in the nondeafferented dorsal dendrites of the same cells (Deitchand Rubel 1984). Neurons have also been shown to reorganize their dendrites aftermanipulations of visual experience. For example, in cats reared under conditions in whichonly horizontal or vertical lines were presented for viewing, Tieman and Hirsh (1982)27found shifts in the dendritic field orientation of layer 3 pyramidal cells which werecorrelated with the rearing condition.A striking rearrangement of axonal arbors in cat and monkey primary visual cortex hasbeen demonstrated, due to inbalance in the visual exposure of the two eyes, during thecritical period of development in early postnatal life. In the newborn animals, oculardominance columns are not fully formed and the process of sorting out of the terminals ofleft- and right-eye geniculocortical afferents into alternating bands in layer 4 continuesthrough the critical period. In normal animals, geniculocortical afferents from the left andthe right eye sort out into columns of nearly equal width in layer 4. Depriving one eye ofvision by monocular eyelid suture (monocular deprivation) at birth causes thegeniculocortical afferents to sort out into columns of unequal width, those for the open eyebeing wider than for the closed eye (Wiesel and Hubel 1963; Hubel et al 1977; Shatz andStryker 1978). Physiological changes have also been observed after monoculardeprivation, most neurons can only be activated through the nondeprived eye and thedeprived eye responses are greatly reduced (Wiesel and Hubel 1963; Shatz and Stryker1978). In contrast, visually evoked responses of visual cortical neurons in binocularlydeprived animals appeared largely normal (Wiesel and Hubel 1965). It has been suggestedthat the effects of monocular deprivation depend on a type of competitive interactionbetween the afferents of the two eyes that is in operation during normal postnatal life.Recently, Antonnini and Stryker (1993), using the technique of anterograde filling withphaseolous lectin (PHA-L), have shown rapid remodeling of geniculocortical axonal arborsfollowing both a short-term (6-7 days) and long-term (4 weeks) monocular deprivation. Inthese animals, the total lengths and the number of branch points of the axonal arbors of thedeprived eye were significantly reduced compared to those of the nondeprived eye.281.5. PATCHY CONNECTIONS IN THE VISUAL CORTEX1.5.1. INTRINSIC LOCAL CONNECTIONSRecently, it was found that the intrinsic, local connections within a single corticalarea are organized into a periodic patchy system, made among groups of neurons. Whensmall amounts of a tracer such as HRP were injected into a region of visual cortex,extracellularly, patches of peroxidase labeling distributed at 1 mm intervals were observedsurrounding the injection site. These intrinsic patches were originally reported by Rocklandand Lund (1982) in primary visual cortex (area 17) of the tree shrew. The patches consistedof neuronal cell bodies, presumably labeled by retrograde axonal transport, and axonterminals, presumably labeled by anterograde transport. Since this original description thelocal intrinsic patches have been reported in area 17 of the tree shrew (Sesma et al., 1984),area 17 of the macaque and squirrel monkey (Rockland and Lund, 1983; Livingstone andHubel, 1984b), area 17 of the cat (Luhmann et al., 1986) and area 18 of the cat (Matsubaraet al., 1985, 1987; LeVay, 1988). Supportive evidence for the patchy nature of the localconnections is provided by the clustering of axonal terminals of individual pyramidalneurons intracellularly filled with HRP (Gilbert and Wiesel 1983; Martin and Whitteridge1984a). The patchy pattern of intrinsic connections in cat visual cortex has been shown toemerge early in development from an immature homogenous distribution (Price 1986;Luhmann et al. 1986; Callaway and Katz 1990), by a process of activity-dependent, axonalelimination (Callaway and Katz 1991; Katz and Callaway 1992 for review).In cat area 18, studies of local patchy connections with wheat germ agglutininconjugated to HRP (WGA-HRP), have demonstrated that the majority of the local patchesare closely circular in outline with a diameter of roughly 350 tim, in sections cut intangential plane (Matsubara et al. 1987; Matsubara 1988). The number of patches arisingfrom a single injection ranges from 2 to 10 (mean: 5.7, SD: 2, n=15) and most of the29patches are labeled within 1.4 mm from the injection site center, with an occasional patch asfar away as 3.4 mm (Matsubara et al. 1987; Matsubara 1988). LeVay (1988), reportedsimilar results using WGA-HRP in cat area 18. According to LeVay (1988), the individualpatches range in width from about 250 to 500 pPm, and the gaps between the patches arecomparable in width to the patches, so that the overall periodicity ranges between 500 and1000 urn. The patches close to the injection site are heavily labeled, and the patches becomeprogressively fainter with increasing distance from the injection site. The tangential extentof the patches, measured from the center of the injection site is usually between 1 and 4 mm(LeVay 1988).In sections cut in the coronal plane, the labeled cells in the patches are found invertical columns running through all 6 cortical layers, but certain layers, especially layers 2and 3, are more heavily labeled than others (Matsubara et al. 1987; Matsubara 1988; LeVay1988). The labeled cells in each patch include both pyramidal and smooth multipolar(GABAergic) types (Matsubara et al. 1987; Matsubara 1988). However, the number ofsmooth multipolar neurons is relatively low in comparison to the pyramidal neurons. Thismay merely reflect the finding that, on average, only 20% of the total number of neurons invisual cortex are GABAergic (Gabbott and Somogyi 1986). Thus, the apparently lownumbers of GABAergic neurons may not indicate that their role in local processing isminor, but rather that their inhibitory effects are more wideranging. However, on the otherhand, an EM study of patchy intrinsic projections of cat area 18 (LeVay 1988) reported thatnearly all (about 96%) of labeled axon terminals in the patches formed asymmetricsynapses. The synapses were found to be predominantly (86%) onto dendritic spines, withmost of the remainder being made onto dendritic shafts, suggesting that the patchyprojections predominantly arise from, and synapse on, neurons with spiny dendrites. Onecould argue that sampling bias associated with EM studies could account for the lack ofsymmetric, GABAergic synapses found in this study.301.5.2. CORTICOCORTICAL CONNECTIONSMany of the projections linking one visual area to the other are also patchy (Gilbertand Kelly 1975; Rockland and Pandya 1979; Tigges et al. 1981; Bullier et al. 1984;Symonds and Rosenquist 1984; Van Essen et al. 1986). Since the work of Gilbert andKelly (1975), it has been generally thought that, in the visual cortex of the cat, the neuronsprojecting to other cortical areas are located in lamina 2/3 whereas neurons projecting tothalamus are located in lamina 6 and the neurons projecting to superior colliculus arelocated in lamina 5. However, Bullier et al. (1984) found corticocortically projectingneurons in all the laminae although the neurons were relatively rare in layer 4. Theysuggested that the corticocortical neurons are found in all layers but their proportionchanges depending on the areas involved in the connection. Further their data showed thatthe laminar distribution of corticocortical neurons depends more on the area of origin thanon the target cortical area. For example, corticocortical neurons in the maingeniculorecipient areas, 17 and 18, belong mostly to supragranular layers (layers 2 and 3),whereas those belonging to area 20, which is far removed from area 17 and 18, belongalmost exclusively to the infragranular layers (layers 5 and 6). Area 19 andposteromediolateral suprasylvian area (PMLS), which are located between these twoextremes, contain similar proportion of neurons in the supra- and infragranular layers. Thisresult is interpreted as showing that the visual information flow from LGN to successivecortical areas is reflected in the laminar distribution of corticocortical neurons, and that thefeedforward (away from area 17) connections from low levels tend to arise fromsupragranular layers whereas the feedback (toward area 17) connections from higher tolower levels mostly arise from infragranular layers (Rockland and Pandya 1979; Bullier etal. 1984).311.5.3. CALLOSAL CONNECTIONSIt is known that the callosally projecting neurons are also organized into a patchysystem (Newsome and AIlman 1980; Van Essen et a!. 1982; Segraves and Rosenquist1982; Cusick et a!. 1984; Andersen et al. 1985) (However, see section 1.5.5). The cells oforigin and their terminal arborizations occupy irregular clusters separated by unlabeled orlightly labeled zones (e.g., Segraves and Rosenquist 1982; Voigt et al. 1988). Whenvisualized with retrograde tracers, for example, HRP, alone or bound to wheat-germagglutinin (WGA), or with fluorescent tracers, the largest fraction of retrogradely labeledcallosal neurons has been found in layer 3. Other layers (2,4,5 and 6) also contribute tocallosal projections, in varying proportions, depending on the species and area (Innocent!1986). For example, the second largest fraction of callosal neurons seems to be in layer 5in rodents, and in layer 6 in cats. In monkey, infragranular callosal neurons have beeninconsistently found in either layer 5 or 6 (Innocenti 1986). The cells of origin of thecallosal pathway in the cat are located in the lower part of layer 2/3, the upper part of layer4, layer 6 and occasionally, in layer 5 (Segraves and Rosenquist 1982; Voigt et al. 1988;Buhi and Singer 1989). Across areas and species, the callosally projecting neurons werefound to be pyramidal neurons (Innocenti 1986). In cat, layer 4 spiny stellate cells alsocontribute to the callosal projection (Innocenti 1980).There are also areal differences in the distribution of callosal neurons. For example,there is a stronger contribution of infragranular layers to callosal pathways in area 19 in thecat, as compared to areas 17 and 18 (Keller and Innocenti 1981; Segraves and Rosenquist1982). Cortical areas project callosal axons not only to homologous (homotypic) areas inthe contralateral hemisphere, but also to heterologous (heterotopic) areas. These two typesof projections originate from different layers. For example, the PMLS of the cat projects tothe contralateral homologue mainly from layer 3, and to contralateral areas 17 and 1832mainly from layer 6 (Keller and Innocenti 1981; Segraves and Rosenquist 1982; Segravesand Innocenti 1985).Various functional properties have been attributed to callosal connection. Theseinclude midline fusion of the two hemifields (Berlucchi 1980; Leporé and Guillemot 1986),binocular convergence, stereopsis (Mitchell and Blakemore 1970; Payne et al. 1980;Cynader et al. 1986) and interhemispheric transfer of learned visual discriminations (Sperry196 1; Myers 1962; Berlucchi 1972). The callosal neurons are, at birth, distributedcontinuously through the cortex (Innocenti et al. 1977; Innocenti and Caminiti 1980). In theearly postnatal period, the transitory callosal projections are removed by axonal eliminationrather than neuronal death, to create the tangentially discontinuous pattern in the adult(Innocenti 1981). The maturation of callosal connections is experience-dependent, sincemanipulation of sensory experience disrupts the normal development of callosalconnections (Innocenti and Frost 1979; Lund and Mitchell 1979; Berman and Payne 1983;Innocenti et aT. 1985). The studies of Innocenti and Clarke (1983) shows that neocorticalneurons decide at an early stage whether to send their axons to the ipsilateral or to thecontralateral hemisphere. Different neurons were found to project ipsilaterally andcontralaterally, and to different areas of the contralateral hemisphere with only very fewneurons projecting to both hemispheres by bifurcating axons (Innocenti and Clarke 1983).1.5.4. FUNCTIONAL SIGNIFICANCE OF PATCHY INTRINSICCONNECTIONS IN VISUAL CORTEXMany studies have investigated whether there is any systematic relationship between theintrinsic patchy system and the ocular dominance or orientation columnar systems ofpatches. Matsubara et al. (1987), in cat area 18, compared the distribution of the intrinsicpatches to the functional architecture of the same sample of tissue as mapped in multiple33electrode penetrations. They found a significant relationship between the orientationpreference of the injected and labeled regions. Their findings were that the patchy intrinsicprojections were between cell groups whose preferred orientations were different and onaverage orthogonal. Gilbert and Wiesel (1989), in cat area 17, combined the labeling ofintrinsic patches obtained by retrograde transport of rhodamine microspheres and thelabeling of orientation columns obtained by 2-DG autoradiography. They reported that thepatchy intracortical connections were between cell groups with similar preferredorientations. Blasdel et al. (1992), in monkey striate cortex, combining optical imagingwith small injections of biocytin have reported that the patchy intrinsic projections weremade between regions that preferred similar orientations in some cases and between regionsthat preferred different orientations in other cases. They suggest the possibility that thepatchy connectivity may be driven by properties other than orientation preference. Thesame group of researchers (Yoshioka et al. 1992), in monkey striate cortex, using the sametechnique, i.e., the combination of optical imaging and small injections of biocytin, havereported that the intrinsic patches from single injections of biocytin were predominantly(about 75%) in domains of same ocular dominance. They suggest that the intrinsicconnections primarily link cells of like property, but in addition possess a substrate forinteraction between cells of unlike properties.1.5.5. PATCHY NETWORKS AND THEIR RELATIONSHIP TO EACHOTHERThe patchy nature of connectivity is well known from anatomical studies of cerebralcortex. There is evidence that such connectivity patterns in turn reflect patchy elements inthe functional organization of cortex. Two distinct types of patchy connectivity have beenobserved following neuroanatomical tracing experiments. For ease of reference, theterminology used by Shipp and Grant (1991) is used here in the description of these two34types of patchy connectivity. In the first type, the connection from one area to the other isglobally continuous, but locally discontinuous, with each site in a given area connectingwith a number of patches in another area. When a large volume of tracer is injected into thearea concerned, the patchy pattern disappears and become continuous. This type of patchyconnectivity is referred to as the ‘soft’ pattern by Shipp and Grant (1991). In the secondtype, the distribution of the cells of origin or the terminals within one area which connectsto the other area is genuinely discontinuous or uneven and independent of injection size.This type is referred to as the ‘hard’ pattern by Shipp and Grant (1991).Examples for ‘soft’ pattern can be found in the intrinsic connectivity of primate area17 (Livingstone and Hubel 1984b) and cat area 18 (Matsubara et al. 1987). Livingstoneand Hubel (1984b), made small injections which were entirely inside a single CO blob orinterblob in area 17. They found that when the injections were within blobs, the labeledpatches were predominantly within the blobs and when the injections were in interblobzones, the labeled patches were found in the interblob areas. This result suggest that largerinjections covering both the blobs and interblobs would result in a continuous labelingincluding both the blob and interblob areas. Matsubara et al. (1987), made small injectionsof two different tracers into two neighbouring sites in area 18. Such injections resulted intwo sets of patches with the two different labels, which frequently fused together toproduce a continuous, less patchy distribution of labeled cells. This result suggests that thepatches observed after a single injection were only a part of the continuous ‘soft’ patchsystem of intrinsic connections in area 18.Examples of ‘hard’ patch patterns can be found in the corticocortical connections,from area 17 to area 18 (Ferrer et al. 1988), from area 17 to the lateral suprasylvian area(LS) (Shipp and Grant 1991) and from area 17 to area 19 and suprasylvian areas (Ferrer etal. 1992). In these studies large injections of retrograde tracers into area 18, 19 or LS,35resulted in projection patches in area 17, which were genuinely discontinuous, i.e., ‘hard’patterned. The callosal connections seem to have a combination of both ‘soft’ and ‘hard’patterns. Large injections of HRP in one hemisphere result in callosal labeling which iscontinuous along the 17/18 border, and discontinuous at more peripheral fieldrepresentations in cat visual cortex (Boyd and Matsubara, submitted). Similar studies inprimates (Cusick et al. 1984) show that callosal connections are continuously distributednear the 17/18 border. Away from the border, callosal labeling appears as patches thatoverlap with CO blobs, and in area 18, patches form bands that overlap with CO-richstripes, thus, indicating also the ‘hard’ pattern (Cusick et al. 1984).Since the intrinsic local, corticocortical, and callosal connections all independantlyexhibit patchy labeling, the relative distribution of these patches to each other in a givencortical area is of interest. These patchy systems, may fully or partially overlap orinterdigitate with one another or, they may have no recognizable relationship at all to oneanother. In macaque monkey prefrontal cortex, it has been shown that columns ofcorticocortical afferents interdigitate with columns of callosal terminations (Goidman-Rakicand Schwartz 1982), thus, forming separate and alternating patch networks. The study ofDeYoe and Van Essen (1985) in macaque monkey shows that the efferent cells in thesecondary visual area V2, projecting to two of its major target areas, MT and V4, arearranged in stripe-like patches which are largely segregated from each other and are closelyrelated to the pattern of alternate CO-rich and CO-sparse zones.1.6. RATIONALEAs discussed above in section 1.5., the intrinsic local patch networks and thecorticocortical and the callosal projections all tend to be segregated into periodic columnarpatches. For a better understanding of the functional organization of the cerebral cortex, it36seems necessary to examine the structural and functional relationship between these patchysystems. Pyramidal neurons in the superficial layers are known to participate in all threeprojection pathways, the local intrinsic projections (Matsubara et al. 1987; LeVay 1988) thecallosal (Jacobson and Trojanowski 1974; Shatz 1977; Innocenti 1980; Segraves andRosenquist 1982; Voigt et al. 1988; Buhl and Singer 1989) and the corticocorticalprojections (Gilbert and Kelly 1975; Bullier et al. 1984; Symonds and Rosenquist 1984;Rosenquist 1985; Ferrer et al. 1988). Thus, there appear to be at least three populations ofpyramidal cells in the superficial layers. What is not yet known is the spatial relationshipamong the patches of the three systems and whether there are morphological differencesamong the cells of these projection systems. Also, it is not known whether individualneurons in the intrinsic local patch networks also participate in cortical projection pathways.Evidence from intracellular studies of physiologically identified neurons in vivo (Gilbertand Wiesel 1983; Martin and Whitteridge 1984a) suggests that many pyramids in thesuperficial layers posses a descending axon trunk entering white matter, as well as localpatchy axon collaterals which spread within the same cortical area. These results suggestthat at least some neurons in the superficial layers may participate in the local intrinsicconnections and a distant projection pathway. The question then arises, whether the localpatch neurons represent a distinct group of intemeurons, separate from the main populationof projection neurons, or whether they are primarily projection neurons with local axoncollaterals ? While earlier studies focused on the organization of the patches and theirrelationship to functional properties, such as orientation selectivity and binocularity, little isknown about the detailed morphology of the neurons in these patches, or their potential rolein other periodic projection systems. With these considerations in mind, it was decided toundertake a qualitative and quantitative analysis of the dendritic morphology of local patchand callosal neurons in area 18 of the cat visual cortex and to perform doublelabelingexperiments to address whether the local patch neurons also participates in cortical outputpathways, for example, the callosal pathway.37Given the functional specialization of local cortical networks and callosal pathway in visualcortex, I hypothesize that each system will form a unique and separable population ofanatomical cell types. These unique anatomical cell types will reflect their functional roles invisual processing. This hypothesis will be tested by addressing the following goals.The goals of this project may be stated as follows:1. To determine which morphological types participate in local patchy projections. Inorder to study this issue, the dendritic morphology of local patch neurons in layer 2/3 of catarea 18 was examined using retrograde tracing and intracellular injection.2. To identify if the cells belonging to the callosal and local patch populations exhibitdifferences in dendritic morphology. In order to study this issue, the dendritic morphologyof callosal neurons in layer 2/3 of cat area 18 was examined, and compared to local patchneurons, in coronal and tangential planes.3. To determine whether the local patch neurons are primarily projection pyramidswith local axon collaterals, or whether they represent a distinct group, separate from theprojection neurons in layer 2/3. In order to study this issue, double-labeling experimentswere performed where both local patch and callosal neurons were labeled in area 18 and theregions of overlap of the two populations of labeled cells were examined for evidence ofdouble-labeling.4. To provide a basis for studies on the dendritic morphology of corticocorticalneurons (18 to 17 projecting neurons) in layer 2/3 of cat area 18.38CHAPTER 2EXPERIMENTAL PROCEDURES2.1. INJECTION OF RETROGRADE TRACERS AND PREPARATION OFBRAIN SLICES2.1.1. LOCAL PATCH POPULATIONNine adult cats of both sexes were used for the local patch study. The cats wereanaesthetized with intravenous sodium methohexital (Brietal) and were given atropinesulfate (0.05 mg/kg) and dexamethasone (0.15 mg/kg) to reduce salivation and brainedema, respectively. Under surgical anaesthesia, the animals were intubated and placed in astereotaxic frame at which time anaesthesia was supplemented with halothane gas. Acraniotomy was performed over the region of area 18 at stereotaxic coordinates +6 to -1 APand +4 to +1 ML. Microinjections of fluorescent dextrans (M.W. 10,000) (MolecularProbes) reconstituted as 10% solutions in distilled water were made into area 18 to identifythe local patches by retrograde transport. Dextrans have been shown to transport in bothanterograde and retrograde directions (Nance and Burns, 1990; Schmeud et al. 1990).When dextrans and microspheres were tested for use as tracers in this study, dextrans werefound to be superior to the latex microspheres (Lumafluor, mc) for retrograde labeling ofneurons associated with the local patches. To label the local patch neurons, two smallinjections (100 nl each) of the tracer, were made into area 18. These injections were spacedat least 3.5 mm apart, to prevent overlap of the label from adjacent injection sites. Thetracer solutions were drawn into glass micropipettes (inner tip diameter, 10-20 rim) andinjected by pressure, at a depth of 1 mm. One cat received injections of rhodamineconjugated microspheres into one hemisphere, three other animals received injections offluorescein-conjugated dextrans into one hemisphere, and the other five animals receivedinjections of either fluorescein-conjugated dextrans or lucifer yellow-conjugated dextrans39into each hemisphere. By utilizing two different fluorescent dextrans, local patch neuronsin both hemispheres were studied without complications arising from callosally labeledneurons.At the end of 7-14 days survival time the animals were deeply anesthetized andperfused intracardially with a rinse of 0.1 M phosphate buffer (PB), pH 7.2, with 0.5%sodium nitrite, followed by 1.0 liter of 4% paraformaldehyde in 0.1 M PB. The flow ratewas maintained at 70 ml/min with a perfusion pump and the total perfusion time lastedapproximately 30 minutes. The quality of tissue fixation was very crucial to the subsequentfilling of neurons with intracellular iontophoretic application of LY. If the tissue was fixedtoo strongly the dendritic filling was poor. Alternatively when the tissue was too lightlyfixed, LY quickly leaked out of the cell, presumably, because of the inadequate fixation ofthe membranes. After perfusion, the brains were blocked stereotaxically and blocks oftissue containing visual cortex were postfixed in 4% paraformaldehyde for 30 minutes-ihour at 4°C. Vibratome sections, 220 tm thick, in the coronal and sagittal planes weretaken and stored in cold PB. In the tangential plane, 150 urn thick sections were takenstarting from the pial surface and prelabeled cells in the center flat regions of the top 4sections were used for LY-filling. Of the 14 hemispheres used for this study, 6 were cut inthe coronal plane, 5 were cut in the sagittal plane and 3 were cut in the tangential plane.Thicker sections of tissue, up to 400 uim were utilized initially, however, because of thelimited working distance (w.d.) of the optimal lenses for confocal imaging (Nikon, 60Xoil, N.A. 1.4, w.d. 170 im and 40X oil, N.A. 1.3, w.d. 220 jim), it was moreappropriate to section the tissue no more than 220 jim in thickness. Neurons displayingobviously truncated processes were eliminated from the data sets. Figure 2A illustrates thepart of the area 18 of cat brain (shaded box) that was stereotaxically blocked andsubsequently used for obtaining slices. Figure 2B is a schematic diagram which shows thedifferent planes in which the area 18 was sectioned in this study.40Figure 2. A, Photograph of a cat brain showing the visual cortex area 17 and area 18. Thepart of the area 18 used for this study is indicated by the shaded box. B, Schematic diagramillustrating the different planes of sectioning used in this study. The area 18 was blockedstereotaxically along the hatched lines indicated in Figure 2A and subsequently sectionedwith a vibratome in coronal, sagittal and tangential planes.41422.1.2. CALLOSAL POPULATIONTen adult cats of both sexes were used for the callosal study. Eight animals were usedfor the study of the morphology of callosal neurons by retrograde labeling followed by LYinjection and two animals were used for double-labeling experiments to see whethercallosal and intrinsic patch neurons were the same population. The cats were anaesthetizedthe same way as for the local patch study. A large craniotomy was performed over the righthemisphere at stereotaxic coordinates +6 to -1 AP and +4 to + 1 ML and the dura wasincised. Six to eight large injections (250 ni each) of fluoroscein-conjugated dextransreconstituted as 10% solution in distilled water were placed at the right 17/18 border tolabel callosal neurons in left area 18 by retrograde transport. The tracer solution was drawninto glass micropipettes (inner tip diameter, 20-30 tm) and injected by pressure, at a depthof 1 mm.At the end of 7-14 days survival time, the animals were deeply anesthetized andperfused intracardially with a rinse of 0.1 M PB with 0.5% sodium nitrite, followed by 1.0liter of 4% paraformaldehyde and 0.01% glutaraldehyde in 0.1 M phosphate buffer.Glutaraldehyde in a minimum effective concentration (between 0.005 and 0.0 1%) has beenshown to reduce the damage of tissue during handling without affecting the quality ofneuronal filling (Tauchi and Masland 1985; Buhl et al. 1990). The perfusion procedure, thepostfixation and the subsequent slice preparation were the same as for the local patch study.Of the 8 hemispheres used for the morphological study, 5 were cut in the coronal plane and3 were cut in the tangential plane. Similar to the local patch study, 220 Im thick sections inthe coronal plane and 150 iim thick sections in the tangential plane were taken.432.2. INTRACELLULAR INJECTION WITH LV- LOCAL AND CALLOSALPOPULATIONS2.2.1. PREPARATION OF MICROPIPETTESAldehyde fixation causes the neuronal membranes to become rigid, and thereby reducestheir ability to form a tight seal around an impaling pipette, which may result in dye-leakageif pipettes with large tips are used. Thus, intracellular injection technique requires theproduction of micropipettes with ultrafine tips. Both vertical and horizontal pipette pullersprovide suitable pipettes for intracellular filling in fixed slices. By adjusting the heater andmagnetic pull control of the pullers it is possible to determine the optimal taper and tip sizeof microelectrodes. Omega dot glass tubing, 1.0 mm in diameter, was used for making thepipettes. This glass tubing contains a single fine glass fiber fused to the inner surface whichimproves the spread of LY into the pipette tips. The pipettes were prepared and filledimmediately before use. The most convenient measure of the size of pipette tips is theirresistance. The pipettes used in this study ranged in resistance between 100 and 200MOhm. The resistance of micropipettes suitable for filling purposes range between 90 and400 MOhm (Buhl and Lübke 1989).2.2.2. INJECTION TECHNIQUESlices were carefully mounted, with a brush, on a glass slide and were kept moist with0.1 M cold PB. Slices were scanned for prelabeled cells in area 18 using a Nikon Optiphotmicroscope equipped with epifluorescence and extralong working distance objectives (20 Xand 40 X, ELWD, Nikon). Micropipettes (usually about 100 MOhm impedance) werefilled with LY, Molecular Probes, 5-15% in distilled water or Sigma, 5% in distilled water.Both types of LY gave similar results. The LY-filled micropipettes were then heldapproximately at an angle of 40 degrees from the horizontal, by a stage mounted hydraulic44micromanipulator (Narashige, #204). Under visual guidance, prelabeled neurons in layer2/3 of area 18 (local patch neurons in the local patch study and callosal neurons in thecallosal study) were impaled with the micropipettes and iontophoretically injected (1-3 nA,electrode negative constant current, for 5-10 mi MSI iontophoretic unit, BH-2 system)with LY, until distal dendrites appeared brightly fluorescent. Successful impalement of aneuron resulted in rapid and intense filling of the neuron with LY. The fluorescent dextransclearly labeled the cell body, thus, targeting of neurons was quite successful, and inoptimally fixed tissue from a single animal, yields of 25-30 LY-filled, prelabeled neuronswere obtained.2.3. PROCESSING OF TISSUE FOR SUBSEQUENT EXAMINATION,IMAGING AND ANALYSIS - LOCAL AND CALLOSAL POPULATIONSSlices with LY-filled neurons were postfixed overnight in 4% paraformaldehyde andmounted on gelatin-subbed slides. These slices were then air-dried, cleared directly inmethyl salicylate for 1 minute and coverslipped with Fluoromount (Gurr). This method ofthick section preparation gave confocal images with the best possible resolution and theleast possible bleaching compared to dehydrating through alcohols and clearing with xyleneor mounting with an aqueous media (glycerol). Ultrathin covergiass (#00, 70 im thick)was used to enhance the working distance of the oil immersion objectives.2.4. DATA ANALYSIS - LOCAL AND CALLOSAL CELLSA total of 193 layer 2/3 local patch neurons and 177 layer 2/3 callosal neurons werefilled with LY in cat area 18. All the local patch cells filled in coronal and sagittal planes andall the callosal cells filled in coronal plane were used in the analysis of the mean somaticsize and size distribution of cells with respect to their distance from pial surface. Of the 19345local patch cells filled, 155 neurons (59 in the coronal plane, 50 in the sagittal plane and 46in the tangential plane) were used in the analysis of basal dendritic trees as the remainderhad truncated dendrites and were eliminated from the data set. Similarly, of the 177 callosalneurons filled, 152 callosal neurons (98 in the coronal plane and 54 in the tangential plane)were used in the analysis of basal dendritic trees as the rest had obvious truncation pointsalong their dendrites. The basal dendrites of the LY-filled neurons, unless truncated bysectioning, appeared to be filled completely to their distal ends. Apical dendrites weresometimes incompletely filled, especially their apical tufts. Axonal filling was variable, butin most cases the axon initial segment was filled. The neurons were imaged on a Bio-RadMRC 600 confocal microscope employing an argon ion laser (2 488 and 514 nm). Serialoptical sections at 1 pm intervals were collected and each optical section was digitallyprocessed by Kalman filtering (Woods and Radewan, 1977) to optimize the signal-to-noiseratio. The optical sections (Z series) were then stacked together to produce a ‘project’ imageof the entire neuron. Thus, a ‘project’ image is a photographic equivalent to a camera lucidadrawing, in that all processes are displayed in focus on one plane. For illustrationpurposes, reverse contrast on the project images was used, thus, the LY-filled neuronsappear black on white background. A few cells possesed basilar dendritic trees which weretoo large to be fully displayed on the computer monitor. In these instances, left and rightdendritic fields were imaged separately. Some cells were drawn with a camera lucida with100 X oil immersion objective.Quantitative analysis of width of the basal dendritic fields, number of primary dendriticbranches and total number of basal dendritic branch points was performed on the basaldendrites of the callosal and local patch pyramids, in order to explore the possibility thatquantitative differences may exist between the dendritic systems of the two populations ofcells. Measurements of the area of the soma and the total number of basal dendritic branchpoints of each neuron were made on the project images. The number of primary dendrites46was obtained by direct microscopic observation of the LY-fihled neurons. The width of thebasilar dendrites of each neuron and the distance from the pial surface was measured usinga standard epifluorescent compound microscope with a 10 X or 20 X objective and aneyepiece graticule which was calibrated against a stage micrometer scale. In the tangentialplane, the anteroposterior (AP) and mediolateral (ML) basal dendritic tree widths of the LYfilled cells were measured in the same way using a standard epifluorescence microscopewith a 10 X or 20 X objective and a calibrated eyepiece graticule. Laminar boundaries weredetermined according to the criteria of O’Leary (1941). The border between layer 3 and 4was identified by the transition from triangular cell bodies of pyramidal neurons which arepresent in the bottom of layer 3 to the circular cell bodies of stellate neurons present in layer4 (e.g., Figure 3). On average the border between layer 3 and 4 falls between 615 pm and640 pm from the pial surface.2.5. DOUBLE-LABELING STUDYThe animals used for the double-labeling experiments received 8-10 large injections(300 nl) of texas-red-conjugated dextrans or rhodamine-conjugated dextrans into the right17/18 border, to label as many callosal neurons as possible in the left area 18. In addition,these animals also received 2 small injections (100 nI) of fluorescein-conjugated dextrans inthe left area 18, to label the local patch neurons. After 7-14 days survival time, the animalswere perfused. The perfusion procedure was the same as for the local and callosal study.Serial sections were taken at 150 im thickness with the vibratome in coronal plane. Theprocessing of tissue was also similar to the local and callosal study. In a controlexperiment, a mixture of the two different fluorescent tracers that were used in the doublelabel study were injected into a single site in area 18, to determine whether the two tracerscan be taken up by the same neuron and transported retrogradely. The survival time,47perfusion, sectioning and processing of tissue in the control experiment were the same asfor the double-labeling study.In the double-labeling experiments, the regions with both callosal and local cells, wereexamined with a standard compound epifluorescence microscope and the number ofpatches labeled and the number of overlapping regions in area 18 were counted. Thedistribution of the labeled callosal and local cells was charted with camera lucida underepifluorescence. The overlap regions were then imaged with the confocal microscope. Inthe control experiments, the regions containing labeled cells were examined in the sameway as the double-labeling experiments. In both the double-labeling and the controlexperiments the regions containing labeled cells were carefully examined for containingboth fluorochromes.48CHAPTER 3RESULTS3.1. MORPHOLOGICAL CHARACTERISTICS OF LOCAL PATCHNEURONS3.1.1. GENERAL FINDINGSMicroinjections of fluorescent dextrans into area 18 produced discrete patches ofretrogradely labeled neurons surrounding the injection sites. Examples of retrogradelylabeled neurons in the superficial layers of the local patches are shown in Figures 3 and 4.The injections of fluorescent dextrans were all confined to the grey matter. Thus, theneurons in the local patches were most likely labeled by the uptake of tracer by axoncollaterals terminating in the injection sites. When viewed in the coronal plane, theprelabeled neurons extended through cortical layers 2 through 6 in a columnar fashion. Theretrograde labeling of local patch neurons seen with the fluorescent dextrans agrees withthat described using HRP (Matsubara et al. 1985). The majority of the neurons prelabeledin layer 2/3 were small to medium sized pyramidal neurons. A few nonpyramidal neuronswere also observed among the prelabeled population in layer 2/3. Of the 109 LY-filled localpatch cells studied in the coronal and sagittal planes, 94 were pyramidal neurons and 15were smooth multipolar neurons.3.1.2. SOMA SIZE AND LOCATIONThe somatic areas of LY-filled local patch neurons in layer 2/3 ranged from 90-550Jim2 (Figure 5). The mean size of local patch pyramidal neurons was 254 m2 (range: 90-550 m2, SD: 109), while the nonpyramidal neurons were typically much larger, with a49Figure 3. Lower-power project image of fluorescein-dextran labeled neurons in a localpatch in area 18. The transition from layer 3 to layer 4 can be identified by the change ofshape of the labeled neurons from pyramidal neurons in layer 3 to stellate neurons in layer4. Arrowheads point to pyramidal cells in layer 3 and curved arrows point to stellate cells inlayer 4. The layer 3/4 border is at a depth of 621 urn from the pial surface. Arrows point toblood vessels. Scale bar, 100 um.50cr1I.4* i4I%,et.14Figure 4. Higher-power project image of layer 2/3 neurons in a local patch, retrogradelylabeled with fluorescein-dextrans. The figure spans a cortical depth of 200 to 505 tm frompial surface in layer 2/3. Scale bar, 50 tim.52c)—4‘,eFigure 5. Histogram showing the somatic area of the local patch neurons injected with LYinthisstudy.5420 —Cl)——10—zz15 —pyramidal cellsD nonpyramidal cells5Somatic area (pm2)55mean of 347 jim2 (range: 147-550 jim2, SD:108). When the somatic areas of the localpatch neurons were related to depth from the pial surface, it was found that small neurons(less than 250 jim2) predominate up to about 250 jim from the pial surface. This depthappears to correspond to the transition from layer 2 to layer 3. Below 300 jim, in layer 3,the majority of the neurons were medium sized neurons (Figure 6).3.1.3. DIVERSITY OF CELL SHAPES3.1.3.1. Pyramidal cellsIn layer 2, based on morphological shape only, two types of pyramidal neuronswere found. The first type was the modified pyramid (O’Leary, 1941) present at the top oflayer 2 (Figure 7). The apical dendrite and one of the basal dendrites of this cell type extendinto layer 1 and fan-out beneath the pial surface. Frequently, branches of the other basaldendrites also curve upwards into layer 1. The dendrites of the modified pyramids bearnumerous spines. The second type of layer 2 neuron was the small standard pyramid with ashort apical dendrite, which branches almost immediately, or after a short distance whileextending to layer 1 (Figures 8 A, B). The spine density of the small standard pyramids isless than that of the modified pyramids.Two distinct forms of pyramidal neurons were also found in layer 3. One type wasthe star pyramid (Figures 9A, B). The basilar dendritic tree of the star pyramid moreclosely resembles a spiny stellate cell in that both have dendrites which radiate, like pointsof a star, from symmetrically located sites on the cell body. However, the star pyramidsunlike the spiny stellates also possess a slender apical dendrite which extends to layer 1.This mixed arrangement of dendrites on these neurons led to the introduction of the name,star pyramid (Lorente de No, 1949). The star pyramids often appear sparsely spiny. Theother type of layer 3 neuron was the medium sized standard pyramid with classical56Figure 6. Somatic area of the LY-filled local patch neurons in layer 2/3 graphed as afunction of their distance from the pial surface. The dotted line approximates the transitionfrom layer 2 to layer 3. Note that small neurons predominate in the top 250 1m in layer 2;medium sized neurons predominate below 300 Im in layer 3. Open symbols representsmooth, multipolar neurons and solid symbols represent pyramidal neurons.57Figure 7. Project image of an LY-filled local patch modified pyramidal neuron of layer 2.The apical as well as one of the basal dendrites of this cell extend into layer 1, and thismodification of the dendrites led to the introduction of the name modified pyramid(O’Leary, 1941). The section of tissue is cut in the coronal plane. Arrowheads indicatedextran prelabeled neurons in the patch. Scale bar, 50 jIm.59IcD1Figure 8. Project images of two LY-filled local patch small pyramidal neurons of layer 2.A, Layer 2 pyramid in coronal plane. Arrow points to axon initial segment. B, Layer 2pyramid in sagittal plane. Note the curved arrow in B, which indicates curvature point inone of the basilar dendrites. Scale bar, 50 rim.614w4w/IS-4’;1t4Spyramidal morphology. The apical dendrites of these neurons give off side branches inlayer 2/3 and extend to layer 1. Their basal dendrites branch profusely to form bushy trees(Figures bA, B). The dendrites of layer 3 pyramids have fewer spines scattered alongtheir length compared to the dendrites of layer 2 pyramids. Of the 94 LY- filled local patchpyramidal cells described in this study, 9 were modified pyramids, 5 were star pyramidsand 80 were standard pyramids. Nonpyramidal cellsIn addition to spiny pyramidal neurons, two types of smooth, multipolar cells werefound in the labeled local patches in layer 2/3. Multipolar neurons seen in this studycharacteristically had large somatic areas (Figure 5). One of the multipolar cell typesposseses a spherical soma and 4-6 primary dendrites that emerge radially from its surfacewithout any preferred sites of origin. Figure 1 lB illustrates an example of such a cell.None of the dendrites of these neurons appear beaded, nor did they possess spines.The other multipolar cell type posseses a large ovoid soma and 6-8 primarydendrites, the majority of which emerge from the upper and lower poles of the soma(Figure 12). The most prominent feature of this cell type was the presence of numerousdendritic beads. Some of the primary dendrites branch close to the soma to form secondarydendrites that commonly undergo no further branching, and usually follow a rather straightcourse as they extend distally (Figure 12). Based on the scheme used to classifynonpyramidal neurons on the basis of dendritic morphology alone (Jones 1975; Peters andRegidor 1981), the latter type of neurons discussed above may be classified as basket cells.The dendritic patterns used in the classification scheme by Jones (1975) and Peters andRegidor (1981) were also used by others (Somogyi et al. 1983; Martin et al. 1983) inaddition to axonal morphology to classify basket cells. However, without the axonal65Figure 10. Camera lucida drawings of two LY-filled local patch standard pyramidalneurons of layer 3. Note the similarity between the dendrites of these two standardpyramidal cells. Spines were not drawn. Scale bars, 50 urn.66IIFigure 11. A, Lower-power project image of a local patch at the level of layer 3. Straightarrow points to a nonpyramidal neuron at the edge of the patch filled with LY. Curvedarrow points to a blood vessel which serve as a landmark. Arrowheads point to dextranprelabeled neurons in the local patch. B, Higher-power project image of the layer 3nonpyramidal neuron with smooth dendrites seen in Figure 1 1A. Scale bars, 50 .tm.68‘-‘4B69Figure 12. Project image of an LY-filled large layer 3 multipolar neuron, with beadeddendrites. Scale bar, 50 JIm.70TI-;..a—aIi,—morphology it is difficult to confirm the types of the nonpyramidal neurons filled in thisstudy. Of the 15 multipolar neurons studied, 9 were located near the edges of the patches(Figures 1 1A, 13A) and 6 were located within the patches (Figure 13B). The study ofMatsubara and Boyd (1992) shows that the GABAergic smooth multipolar neurons foundin local patch networks can be divided into two groups which likely participate in differentinhibitory circuits. These are, the ‘short-range’ GABAergic cells which are found within aregion approximately 1 mm from the injection site center, and the ‘long-range’ GABAergiccells which are found as far away as 3 mm from the center of the injection site. The ‘long-range’ GABAergic cells are mostly large multipolar cells, presumably of the basket celltype and are usually located at the edges of the local patches.3.1.4. ORGANIZATION OF BASAL DENDRITES IN CORONAL,SAGITTAL AND TANGENTIAL PLANESThe local patch pyramidal neurons had 3-8 primary branches of basal dendrites (Figure14) which soon branched into second- and third-order dendrites. The mean of the totalnumber of basal dendritic branch points was 16 (range: 10-22, SD:3). Basilar dendritictrees of most of the pyramidal cells filled in tissue sections cut in the coronal plane appearedmore robust and covered a wider zone of cortex than those neurons filled in the sagittalplane (Figures 8A,B; 9A,B). In the sagittal plane, the dendritic trees of some of theseneurons often curved or drooped downward, toward white matter (Figures 8B). Thisdownward curvature resulted in a narrower zone of cortex covered by the dendritic trees.These qualitative distinctions were substantiated by measurements of the widths of thedendritic trees. The mean width of the basal dendritic fields of LY-filled layer 2/3 localpatch pyramids was 209 jim (range: 10 1-338 jim, SD: 40, n = 41) in the coronal plane and164 jim (range: 72-220 jim, SD: 34, n = 38) in the sagittal plane. A Student’s t-test on thisdata showed a highly significant difference in the widths of the basal dendritic fields72Figure 13. A, Higher-power project image of a local patch taken at the level of lower layer3 showing nonpyramidal neurons (arrows) at the border of the patch. This section, as allthe other sections, was not counterstained. Each cell present in the field represents adextran-prelabeled neuron. B, Project image of another local patch showing anonpyramidal neuron (arrow) located within the patch. Scale bars, 50 pm.73wIFigure 14. Histogram showing the number of primary basal dendrites of intrinsic localpatch neurons injected with LY in this study.7560—50—— -—40—-30—s—I I I•0 1 2 9 10Number of primary basal dendrites76between the coronal and sagittal groups (p < 0.0001). The mediolateral to anteroposterior(ML:AP) ratio calculated from the mean widths of dendritic fields of neurons from coronaland sagittal sections was 1.27. Figures 15A and B show the asymmetry observed in thewidths of the basilar dendritic trees of layer 2/3 local patch pyramidal neurons in coronaland sagittal planes.The data obtained from the layer 2/3 pyramidal neurons studied in the tangentialplane of area 18 also confirmed the above finding. Local patch pyramidal neurons (n=46)were studied from the central part of the tangential sections where the apical dendrites lieperpendicular to the plane of section. All of these neurons had spines scattered along theirdendrites which confirmed that they were pyramidal neurons. In tangential sections, thebasal dendritic trees of nearly half of the local patch pyramidal neurons were elongated(e.gs., Figures 16 B, C), the rest were circular in shape (e.g., Figure 16A). The elongateddendritic fields were oriented along the mediolateral (ML) axis in comparison to theanteroposterior (AP) axis. Figure 17A shows the ML and AP widths of the basal dendriticfields of local patch pyramidal cells in the tangential plane. In cells with a ML:AP ratiovalue of 1.25 we could, subjectively by eye, begin to detect a hint of an elongated field andhence, we chose 1.25 as the cut off between circular and elongated dendritic fields (Figure16). When the ML:AP ratio was 1.25 or more, the neuron was classified as a neuron withmediolaterally elongated dendritic field and when the ratio was between 1.25 and 1.0, theneuron was classified as a neuron with circular dendritic field (Figure 17B). An AP:MLratio was calculated for the neurons which had basal dendritic field width measurementslarger in the AP plane than the ML plane. Of the 46 layer 2/3 cells studied in the tangentialplane, only 4 cells (1 in layer 2, and 3 in layer 3) had their AP dendritic field width largerthan their ML width. But none of them had AP:ML ratio values more than 1.25 and thus,they were classified as neurons with circular basal dendritic fields (Figure 17B). In layer 2,77Figure 15. A, Graph showing the width of the basal dendritic fields of the LY-filled localpatch pyramidal cells in layer 2/3 with respect to their soma size. Solid symbols indicatecells in coronal plane with regression shown by solid line. Open symbols indicate cells insagittal plane with regression shown by broken line. Note that the dendrites of pyramidalcells extend greater distances in the coronal plane than in the sagittal plane as indicated bythe regression lines. Triangle, standard pyramid; square, modified pyramid; circle, starpyramid. B, Histogram illustrating the width of the basal dendritic trees of 80 LY-filledlocal patch standard pyramidal cells in layer 2/3. Dark bars represent cells measured incoronal plane while light bars represent cells in sagittal plane. Asterisks indicate meanvalues of dendritic width. Note that the basilar dendritic tree spans greater distances in thecoronal plane than in the sagittal plane.78350 —AcI——C.?z300 -C.?:? ‘‘—200 —150—100 -50 —AA.AAAA2AlA QA4A— A jAADA0I I I I100 200 300 4002Somatic area (urn)500B 15 *10 —5—0—5—10 —15—I-II I I I • I • I50 110 170 230 290 350Width of basal dendriticfields (urn)79Figure 16. Basal dendritic fields of LY-filled layer 2/3 local patch pyramidal cells in thetangential plane. A, Project image of a pyramidal cell with circular dendritic field possesingan ML:AP value of 1.06. Arrowheads point to prelabeled cells in patch. B, Project imageof a pyramidal cell with elongated dendritic field possesing an ML:AP value of 1.34.Arrow indicates blood vessel. A and B, Scale bar, 50 rim. C, Camera lucida drawing of thebasal dendritic field of a pyramidal cell possesing an elongated dendritic field, with anML:AP ratio of 1.68. The cell was drawn under epifluorescence using a 100 X oilimmersion lens. The axes indicate the anterior (a) and medial (m) directions on the corticalsurface of the brain. Scale bar, 50 pm.80*I.***;•I‘iVQ4F41Co—*j3•4*4c**.**Cma81aFigure 17. A, Graph showing the mediolateral (ML) and anteroposterior (AP) widths of thebasal dendritic fields of 46 local patch pyramids of layer 2/3, in the tangential plane.Regression lines for layer 2 cells (thin, solid line) and layer 3 (thin, dashed line) indicatevirtually identical slopes for the two cell groups. Note that almost all cells fall above theequivalence line (bold), indicating the bias of the dendritic fields towards mediolateraldirection. B, Scattergram showing the distribution of ML:AP ratio values of the basaldendritic fields of local patch cells in the tangential plane.82A /350-/01) //l 300- /oQ .. /.- .0 / .250-0/•••/cl•200-• 0.• /• 0.150 - 00 Layer 2 cells/ • I • Layer 3 cellsI Layer 2 regressionLayer 3 regression— equivalence line100 -_____________________________________________I I I I I100 150 200 250 300 350AP Width of basal dendritic fields (rim)B CIRCULAR CIRCULAR ML ELONGATEDILAYER 2I I I I Ill I 111111 CELLSLAYER 3I I II I I I I III I II I I I 1111 I CELLS— I I I I I I1.25 1.2 1.1 1.0 1.1 1.2 1.25 1.3 1.4 1.5 1.6 1.7AP:ML ratio ML:AP ratio838 cells had circular fields with ML:AP ratio values between 1 and 1.25 and 9 cells hadmediolaterally elongated fields with ML:AP ratio values more than 1.25 (Figures 17B). Inlayer 3, 10 cells had circular fields with ML:AP ratio values between 1 and 1.25 and 15cells had mediolaterally elongated fields with ML:AP ratio values more than 1.25(Figure 17B). The mean ML width of the basal dendritic fields of neurons measured in thetangential plane was 221 !Im (range: 132-362 llm, SD: 52), the mean AP width of the sameneurons in the tangential plane was 177 jim (range: 100-304 jim, SD: 48) and the ML:APratio calculated from the mean widths was 1.25. This ratio is virtually identical to the meanML:AP ratio obtained from the neurons studied in the sagittal and coronal planes asindicated earlier.843.2. MORPHOLOGICAL CHARACTERISTICS OF CALLOSAL NEURONS3.2.1. GENERAL FINDINGSThe large multiple injections of fluorescent dextrans into the 17/18 border produceda zone of retrogradely labeled callosal cells in the contralateral 17/18 border extendingroughly 2 cm into area 17 and 1.5 cm into area 18. The distribution of the labeled callosalcells was often patchy within the labeled zone (Figure 18). The labeling patterns seen withthe fluorescent dextrans agrees with those described using HRP (Segraves and Rosenquist1982; Voigt et al. 1988). Most callosal neurons labeled were in lower layer 3. Few labeledcells were also found in layers 2, 4 and 6. Pyramidal cells were the most commonmorphological type of callosal neurons labeled in these layers. In layer 4, neurons withround cell bodies, presumably corresponding to spiny stellate cells, were also labeled.These observations are in agreement with earlier studies which also investigated the callosalneurons with the use of retrograde tracers (Jacobson and Trojanowski 1974; Innocenti andFiore 1976; Shatz 1977; Innocenti 1980; Keller and Innocenti 1981; Segraves andRosenquist 1982). Figures 19A and B are two examples of rhodamine-dextran labeledcallosal neurons taken at higher power at the level of layer 2/3 in area SOMA SIZE AND LOCATIONThe somatic area of LY-filled callosal neurons in layer 2/3 ranged from 170-640jim2 (mean: 378 jim2, SD: 111) (Figure 20). Figure 21 shows the somatic areas of LYfilled callosal neurons in layer 2/3 with respect to their depth from the pial surface. Noteonly a few callosal cells were present in the top 250 jim in layer 2. The majority of callosalcells were labeled below 250 jim in layer 3. The callosal pyramids in layer 3 were primarilymedium to large pyramids (250-640 jim2) but also included a few small pyramids (170-250 jim2).85Figure 18. Lower-power project image of rhodamine dextran-labeled callosal neurons inarea 18. The cells are primarily medium to large pyramids but also include small pyramids.The callosally projecting cells are concentrated in lower layer 3 with few labeled cells inlayers 2, 4 and 6. The layer 3/4 border is at a depth of 635 pm from the pial surface. Scalebar, 100 lIm.86\*V$/VV11-4V4VVVV——VVVVVV—%4VVFigure 19. Higher-power project images of rhodamine-dextran labeled callosal neurons inarea 18 taken at the level of layer 3. A and B, The cells are medium to large pyramids, butalso include a few small pyramids. Note that the granular label of fluorescent dextransfrequently extended into the proximal portion of the apical (1 9B) and basal (1 9A) dendritesof callosal cells. The cell morphology to some extent is visible in the dextran prelabeledneurons, before LY-filling. Note in 19A, standard pyramids (arrowheads) and fusiformpyramid (arrow). In 19B, straight arrow points to a neuron with bifurcated apical dendriteand curved arrow points to a neuron with single apical dendrite, scale bar, 100 pm.88cxp*‘wr4.::•:.Figure 20. Histogram showing the somatic area of the callosal neurons injected with LY inthis study.9020 —10—z15 —5I I400 500Somatic area (Jim2)91Figure 21. Somatic area of the LY-fihled callosal neurons in layer 2/3 graphed as a functionof their distance from the pial surface.92£6(wil)rndwojuijsij09009occoococi’oovoc00EoczOOiOcTooioc0IIIIIIIIIIIIII—0IspweJidpiepuesvI—ocSp!WeJ/dies•IspweAdwJoJisn4•—001—ocTVIVVVVVq00l)V-oczVIVVVVvIVV—.VVV•V•VVVv-00EVVVVVVVVVVVVVvV10017VvVVVIVVVIVi-ocj7VVVIVVVVV7.VI-oocV•VVI.VVI-0ccVI—009VV-0c93.2.3. DIVERSITY OF CELL SHAPESAll the layer 2/3 callosal neurons filled with LY in this study were spinouspyramidal neurons (e.g., Figure 22). Smooth multipolar neurons were not present in thecallosal sample. The callosal pyramidal cells differed markedly in soma size, shape, originof primary basal dendrites and pattern of apical dendrites and thus can be divided intoseveral subclasses. Some aspects of the somatodendritic morphology of callosal cells wereapparent even before LY-filling with dextran retrograde labeling (Figure 19). The somashapes were either standard pyramidal (Figure 23), fusiform (Figure 24) or round inoutline (Figure 25) in the coronal plane. The round cell bodies labeled in layer 3 belong tothe star pyramidal neurons (Lorente de No, 1949). The star pyramids have basal dendriteswhich radiate from the cell body and a slender apical dendrite which extends to layer 1.Within the standard pyramidal category pyramids with different pattern of basal andapical dendrites were observed. The primary basal dendrites of most callosal standardpyramidal cells originate from the base of the cell body in the usual manner (Figure 23). Insome of the callosal pyramids, dendrites originate from regions high up along the sides ofthe cell body in addition to the base (Figure 26). These dendrites which originate from theupper portions of the cell body ascend obliquely and give off branches. The apicaldendrites of the majority of the callosal pyramidal cells in layer 2/3 extend to layer 1 aftergiving off 3-7 side branches at the base (e.g., Figure 23). In some callosal neurons theapical dendritic side branches, together with the basal dendrites, and if present, thedendrites from the upper portion of the soma, form a richly arborized spherical area aroundthe cell body, an arrangement that may allow these neurons to receive a dense input fromthe afferents in layer 3 (Figure 26). Some callosal pyramidal cells had apical dendriteswhich bifurcated close to the cell body before extending towards layer 1 (Figures 19B, 27).94Figure 22. Higher-power project image of a callosal standard pyramidal neuron showingspines on the basal dendrites.950Figure 23. Project image of a LY-filled callosal standard pyramidal neuron of layer 3, incoronal plane. Scale bar, 50 tim.97cc CoFigure 24. Project image of a LY-filled callosal fusiform pyramidal neuron of layer 3, incoronal plane. Scale bar, 50 pm.99100*Figure 25. Project image of a callosal star pyramidal neuron of layer 3, in coronal plane.Scale bar, 50 Lm.101*‘F’I,CI4>1Figure 26. Project image of a callosal pyramidal neuron in coronal plane with dendritesoriginating from the upper portions of the cell body. Arrowhead indicates truncation of theapical dendrite. Scale bar, 50 tim.103104A few callosal pyramidal cells in layer 3 had short apical dendrites that did not extend tolayer 1 but branched at lower levels in layer 2/3 (Figure 28).The majority of the LY-filled callosal cells were standard pyramids. Of the 98 LYfilled callosal cells described in this study, 86 were of this type. Of the rest, 6 werefusiform pyramids and 6 were star pyramids. Of the 86 standard pyramids, 10 werepyramids with dendrites originating from the upper portions of the cell body in addition tothe base, 7 were pyramids with bifurcated apical dendrites and 4 were pyramids with shortapical dendrites. The morphology of the callosal neurons described in this study agreeswith those previously described in cat visual cortex with HRP retrograde transport(Innocenti and Fiore 1976; Shatz 1977; Segraves and Rosenquist 1982) and intracellularLY injection (Voigt et al. 1988; Buhl and Singer 1989). Standard pyramids (Innocenti andFiore 1976; Shatz 1977; Segraves and Rosenquist 1982; Voigt et al. 1988; Buhi and Singer1989), star pyramids (Voigt et al. 1988) and fusiform pyramids (Buhi and Singer 1989) oflayer 3, were previously reported to contribute to the callosal projection in cat visual cortex.3.2.4. ORGANIZATION OF BASAL DENDRITES IN CORONAL ANDTANGENTIAL PLANESThe callosal pyramidal cells had 3-6 primary basal dendrites (Figure 29) whichbranched repeatedly up to the third branch point to form a bushy basal dendritic arbour offthe basal portion of the soma. The mean of the total number of branch points of basaldendritic trees of callosal neurons was 24 (range: 14-31, SD: 4). The mean width of thebasal dendritic fields of layer 2/3 callosal cells studied in the coronal plane was 304 urn(range: 182-393 urn, SD: 54, n 98). Figure 30 shows the width of the basal dendriticfields of callosal pyramidal neurons in layer 2/3 with respect to their somatic area.105Figure 27. Project image of a callosal pyramidal neuron in coronal plane with bifurcatedapical dendrite. Scale bar, 50 rim.106I107Figure 28. Project image of a callosal pyramidal neuron in coronal plane with shoit apicaldendrite. Scale bar, 50 jim.108109Figure 29. Histogram showing the number of primary dendrites of callosal neuronsinjected with LY.11040 —Cl)—-20-Ez 10-0_I I I I I i i0 1 2 3 4 5 6 7 8 9 10Number of primary basal dendrites111Figure 30. Graph showing the width of the basal dendritic fields of LY-filled callosalpyramidal neurons in layer 2/3 with respect to their soma size.112TT(unl)innptuos009ooc00V00IIIIspiuiiAdns•spiuiiiduuojsnj•spnuiJAdpmpu1lJsv001001VVVVVVVVVVVVVVVVV.ATVVVv7V#VV•‘‘YvVVVWvvwVVV•V•VVVVVVVVVVVV—001—ocT—001——.—001:—ocl—0017In the tangential plane, 54 LY-filled callosal neurons were studied. All of them hadspiny dendrites which confirmed that they were pyramidal neurons. In tangential sections,with the exception of one callosal cell which had basal dendritic field elongated along theML axis (Figure 3 1C), all others had closely circular basal dendritic fields (e.gs., Figure3 1A, B). The mean AP width of the basal dendritic fields of callosal cells was 271 pm(range: 152-398 rim, SD: 61, n=54), the mean ML width was 267 im (range: 144-375jim, SD: 57, n=54) and the AP:ML ratio calculated from the means was 1.01. Figures 32Aand B show the ML and AP widths of the basal dendritic fields, and the distribution of theML:AP and AP:ML ratio values respectively of 54 callosal neurons studied in the tangentialplane.114Figure 31. Basal dendritic fields of LY-filled callosal pyramidal cells of layer 3 in thetangential plane. A, Project image of a small callosal cell with circular dendritic fieldpossesing an AP:ML ratio value of 1.09. Note a partially filled cell can be seen above thefilled neuron. B, Project image of a large callosal cell with circular dendritic field possesingan AP:ML ratio value of 1.01. A and B, Arrowheads point to other prelabeled callosal cellsin the tangential plane. Scale bar, 50 tm. C, Camera lucida drawing of the basal dendriticfield of the callosal cell with elongated dendritic field possesing an ML:AP ratio of 1.31.The cell was drawn under epifluorescence using a 100 X oil immersion lens. The axesindicate the anterior (a) and medial (m) directions on the cortical surface of the brain. Scalebar, 50 jim.115* ICm1162.Figure 32. A, Graph showing the mediolateral (ML) and anteroposterior (AP) widths of thebasal dendritic fields of callosal cells in the tangential plane. Note only one cell had MLdendritic width larger than the AP width (for this cell ML width was equal to 1.3 1X APwidth). All other cells had closely circular dendritic fields which is shown by theirdistribution along the equivalence line. B, Scattergram showing the distribution of ML:APand AP:ML values of the basal dendritic fields of callosal neurons studied in the tangentialplane.117400 —_/. ..SA • .‘.300— • •4%. .._,—••250-S.4.200- •150—ioo— //50- /I I I I50 100 150 200 250 300 350 400A-P Width of basal dendritic fields (!lm)B CIRCULAR CIRCULAR ML ELONGATEDII U 11111111111 I III I III I I I I CP1_J_OSPLCELLS—I I I1.25 1.2 1.1 1.0 1.1 1.2 1.25 1.3 1.4 1.5 1.6AP:ML ratio ML:AP ratio1183.3. COMPARISON OF THE MORPHOLOGY OF LOCAL ANDCALLOSAL NEURONS3.3.1. CELL TYPESThe local patch neurons in layers 2/3 were small to medium sized pyramids andincluded modified pyramids of layer 2, small to medium sized standard pyramids of layers2 and 3, star pyramids of layer 3 and smooth multipolar neurons of layer 3 (Thejomayenand Matsubara, ‘93). The callosal pyramids were, on average, larger than the local patchpyramids (Figure 33), and included standard pyramids, star pyramids and fusiformpyramids. The modified pyramids and smooth multipolar neurons were not present in thecallosal population. The apical and basal dendrites of the callosal standard pyramidsexhibited a variety of branching patterns different from the local patch neurons. Forexample, pyramids with dendrites originating from the upper portion of the cell body,pyramids with bifurcated apical dendrites and pyramids with short apical dendrites were notpresent in the local patch population. In addition, in some callosal neurons the proximalside branches of the apical dendrites and the basal dendrites together formed a richlyarborized spherical area around the cell body.3.3.2. ORGANIZATION OF BASAL DENDRITESThe basal dendrites of callosal neurons possesed a more complex branching patternthan the local patch cells. The mean number of branch points of callosal cells (mean: 24,range: 14-31, SD= 4) was significantly larger than that of the local patch cells (mean: 16,range: 10-22, SD= 3) (Student’s t-test, p< 0.0001) (Figure 34). The mean width of thebasal dendritic fields of callosal cells in the coronal plane was also significantly larger(mean: 304 rim, range: 182-393 rim, SD:54) than that of the local patch cells (mean: 209jim, range: 101-338 jim, SD: 40) (Student’s t-test, p< 0.0001) (Fig.35). The differences in119the dendritic field width and the dendritic branch points of local and callosal cells areindependent of cell size, since a local cell could be distinguished from a same size callosalcell based on differences in these values. In the tangential plane, about half of the localpatch neurons had elongated basal dendritic fields and all of these elongated fields wereoriented parallel to the mediolateral cortical axis. The rest of the local patch neurons hadcircular dendritic fields. In contrast, in the tangential plane, with the exception of onecallosal neuron with a mediolaterally elongated dendritic field, all callosal cells had closelycircular basal dendritic fields (Figure 36A,B). A Student’s t-test showed a highlysignificant difference (p< 0.0001) in the ML and AP dendritic field widths between callosaland local patch cells, in the tangential plane.120Figure 33. Graph showing the soma size of the local patch and callosal cells in layer 2/3with respect to their distance from the pial surface.121650 —..... 00• ••.o o0•••. 0 0o .0 oo•ooo. •00 0 •••o Do0.00 0o OdD0Q.• •••oo. • 0o••.•..•# o. .•00600 —550 —500 —450 —400 —350 —300 —E 250 —200 —150 —100 —50 —0—0• 00o.•••.•o..o• • 0o•• 0 ••.•• 0 •• • •0•00• ..•• •0.00 00I I I I I I I I0 50 100 150 200 250 300 350 400Distance from pia (im)0 localcells. callosal cellsI I I I I450 500 550 600 650122Figure 34. Graph showing the total number of branch points of the basal dendrites ofcallosal and local patch pyramidal cells in layer 2/3 with respect to their soma sizes. Notethat, on average, the callosal pyramids have higher number of branch points compared tothe local patch cells.12340 —3O— .?? •* .•% I... . . . .... ....••:•%.•..: • ••.o.. • ... . . I .0)9)) 0 00 ) 0 • 00 oco0 0©10- 0 0z_________________I 0 localcells-coronalplaneI • callosalcells-coronalplane I0-I I I I I I I0 100 200 300 400 500 600 700Somatic area (Lm2)124Figure 35. Graph showing the widths of the basal dendritic fields of callosal and localpatch pyramidal cells in layer 2/3 with respect to their soma size. Note that on average thewidth of the basilar dendritic field of callosal pyramids is larger than that of the local patchpyramids.125500 —1400 -—100-0-.....• 1. :.“.• •..•.0 • • ••••• ‘..•o •: ••..•.o •o OPDf•&• 0000•o••o°.•c9c•0•o oO000 100o local cells - coronal plane• callosal cells - coronal plane200 300 400 500 600 700Somatic area (jim2)I I I I126Figure 36.A, Graph showing the mediolateral (ML) and anteroposterior (AP) widths of thebasal dendritic fields of callosal and local patch cells in the tangential plane. B, Scattergramshowing the distribution of the ML:AP and AP:ML values of the basal dendritic fields ofthe local patch and callosal cells in the tangential plane.127400ECrM‘S‘C0 ..0000 ...cP0035030025020015010050..Acallosal cellslocal cellscallosal regressionlocal regression50 100 150 200 250 300 350A-P Width of basal dendritic fields (lim)B400CIRCULAR CIRCULAR ML ELONGATEDII I I I III I 111111 I LAYER 2CELLSII I I I I Ill I II I I I 1111 I LYER 3CELLS111111 II III IIllIll liii iii I CELLS1.1 1.2 1.25 1.3 1.4 1.5 1.6ML:AP ratio1.25 1.2 1.1AP:ML ratio1 .0 1 .71283.4. DOUBLE-LABELING STUDYAfter injection of two different fluorescent dextrans to label both callosal and localpatch neurons in area 18, the rhodamine-dextran or texas-red-dextran labeled callosalneurons and the fluorescein-dextran labeled local patch neurons were labeled in patches thatoverlapped in certain regions. The two populations of labeled cells were found intermingledin the overlap regions. In the first cat, in which the callosal pathway was labeled withtexas-red-dextran and the local patch networks were labeled with fluorescein-dextran, 7local patches were labeled in area 18 and 3 of these overlapped with the callosal patches. Inthe second cat in which the callosal pathway was labeled with rhodamine-dextran and thelocal patch networks were labeled with fluoroscein-dextran, 9 local patches were labeled inarea 18 and 5 of these overlapped with the labeled callosal patches. Double-labeled cellswere not found in the regions of overlap. All the cells were labeled with either rhodamine-,texas-red- or fluoroscein-dextran indicating that none of the labeled cells participated inboth callosal projection and local patch networks. Figure 37 is a chart which shows thedistribution of retrogradely labeled local and callosal cells in an overlap region of area 18.Note that the patch contains both local and callosal cells, but no double-labeled cells. Figure38A is a confocal dual chanel image showing the retrogradely labeled local and callosalcells in another overlap region of area 18. Note that, even at higher magnification none ofthe cells contain both fluorochromes. In a control experiment where a mixture of the twofluorescent tracers was injected into a single site in area 18, double-labeled cells were found(Figure 38B) indicating that the double-labeling technique does work with the twofluorescent tracers used in cat area 18 and suggests that neurons associated with the localpatch networks and the callosal pathway comprise separate populations.129Figure 37. A charting of retrogradely labeled local and callosal cells in an overlap region ofarea 18 in a double-labeling experiment. The entire local patch and the callosal cells in thearea of local patch were drawn. The callosal labeling extended further than the charted area.Note that the two populations of cells were found intermingled in the overlap region. Nodouble-labeled cells were found.130pia100000o 00o 0.2/3 0 0• 00 •Q r0.o00 0 o’00 co0 00• 00 0 .0 .•• •0 0— _t —0 • 0 0• 0 •400— — — -50-- — —00 050a-0 •00• 0 06whitematter0.5mmI o local cells• callosal cells131Figure 38. A, Confocal dual channel image of labeled callosal and local patch neurons in anoverlap region of area 18 in a double-labeling experiment. The region where bothpopulation of cells overlapped is scanned with dual excitation wave length in the confocalmicroscope. Rhodamine-dextran labeled callosal neurons are shown in the left half and thefluorescein-dextran labeled local patch neurons are shown in the right half of the image.Arrowheads point to matching blood vessels. Note that none of the labeled cells are excitedwith both wave lengths which indicates that none of the cells contain both fluorochromesand that there are no double-labeled cells. Scale bar, 100 tm. B, Confocal dual channelimage of rhodamine- and fluorescein-dextran labeled neurons in the area 18 of the cat in acontrol experiment. The animal received a mixture of the two fluorescent tracers into asingle site in area 18. Arrowheads point to the same blood vessel. Arrows point to doublelabeled neurons. Star and asterisk indicate neurons single labeled with rhodamine-dextranand fluorescein-dextran, respectively. Scale bar, 100 lIm.132()CHAPTER 4DISCUSSION4.1. INTRODUCTORY REMARKSThe relationship between the neuronal morphology and its efferent projection target is asubject of extensive investigations. It is now known that the dendritic morphology of aneuron is correlated to that neuron’s projection target. This has been shown previously forneurons in the infragranular layers of visual cortex. The study by Katz (1987) showed thatthe corticoclaustral neurons in layer 6 of cat visual cortex have apical dendrites extending tolayer 1, with branches in layer 5 only, while corticothalamic neurons in layer 6 have apicaldendrites restricted to layer 3, with branches in layers 5 and 4. Later, Hübener et al.(1990), showed that this is also true for layer 5 pyramidal cells of cat visual cortex and thebranching pattern of both the apical and basal dendrites can be used in the identification oftheir projection target. The study of Hübener et al. (1990) showed that the layer 5corticotectal pyramids have long apical dendrites extending to layer 1 and densesymmetrical basilar dendrites, while layer 5 corticocortical pyramids have short apicaldendrites extending only to layer 2/3 and few basal dendrites. Hence, it has become evidentthat a neuron’s projection target can be predicted on the basis of its dendritic morphologyand laminar location.The laminar arrangement of dendritic branching patterns has also been well defined forlayer 5 pyramids of rodents (Hallman et al. 1988; Hübener and Bolz 1988). Hallman et al.(1988) and Hübener and Bolz (1988) reported that corticotectal neurons have a thick apicaldendrite that extends to layer 1, where it terminates in a highly branched tuft, whereascallosal neurons have a much shorter apical dendrite that usually ends in layer 4. This type134of dendritic structure was first described by Lorente de No (1949), who referred to thesepyramids as short pyramids, a population distinct from the tall pyramids.The above studies show correlations between mainly the apical dendrite and projectiontarget of pyramidal cells in the infragranular layers. In the superficial layers, it is knownthat the apical dendrites of almost all pyramids extend to layer 1, and thus my study hasfocused on the basal dendritic system of projection cells in the supragranular layers. Basedon the knowledge of local patch networks and the cortical projection pathways, there are atleast three populations of cells in the superficial layers. The local patch neurons (Matsubaraet al. 1987, Matsubara 1988; LeVay 1988), the corticocortical neurons (Gilbert and Kelly1975; Bullier et al. 1984; Symonds and Rosenquist 1984; Ferrer et al. 1988; Einstein andFitzpatrick 1991) and the callosal neurons (Jacobson and Trojanowski 1974; Shatz 1977;Innocenti 1980; Segraves and Rosenquist 1982; Voigt et al. 1988; Buhl and Singer 1989)are all known to be present in layer 2/3. Detailed morphological examination based onGolgi studies has revealed several types of pyramidal neurons in the superficial layers(e.g., O’Leary 1941; Lorente de No 1949) and thus, it is not inconceivable that each typewill relate to a select pathway or function.In my thesis the relationship between the morphology of cells in the superficial layersand their projection target was investigated using retrograde tracing and intracellular dyeinjection. The local patch neurons and the callosal neurons in layer 2/3 of cat area 18 wereprelabeled by the retrograde transport of fluorescent dextrans and subsequentlyintracellularly injected with LY. The somatodendritic morphology of the two populations ofpyramids was compared to identify if the cells belonging to the two populations could beseperated based on morphological features alone. Double-labeling experiments were alsoundertaken to determine whether individual callosal neurons have local axon collateralswhich participate in local patch networks. In double-labeling experiments both local patch135cells and callosal cells were labeled in area 18 and the labeled regions in the superficiallayers were examined for evidence of double-labeling.4.2. SUMMARY OF THE FINDINGSMy study shows that the local patch networks arise from spiny pyramidal neuronsbelonging to 3 distinct morphological types, and at least 2 smooth, multipolar neuronaltypes in the superficial layers. Based on cell shape, the pyramidal cells could be classifiedinto modified, star and the standard pyramids. The most common local patch pyramidal celltype was the standard pyramid (80 of 94), but modified pyramids (9/94) and star pyramids(5/94) were also found (Thejomayen and Matsubara 1993). In agreement with previousstudies (Matsubara 1988; Albus et al. 1991; Matsubara and Boyd 1992) nonpyramidal cellswere also found among the cells labeled within the superficial layers of the local patches.These nonpyramidal cells clearly lacked spines, possesed large somatic areas and likelycorrespond to the basket cells described in the cat visual cortex (Peters and Regidor 1981;DeFelipe and Fairen 1982; Somogyi et al.1983; Martinet al. 1983). My study is one of thefirst to identify that these cell types participate in the local patch networks in area 18. Sinceit is not possible to fill more than the initial segment of the axon in fixed brain slices, it isdifficult to conclusively identify these nonpyramidal types further. However, recent studieshave shown that basket cells may be further categorized using the lectin from the plant Viciavillosa (VVA) (Naegele and Katz 1990). Thus, various categories of basket cells can nowbe studied in detail by using VVA in combination with retrograde labeling and LY injection.Recent studies of Kisvárday et al. (1993) show that the large basket cells in area 18 providedirect inhibition to certain pyramidal cells and facilitation to other pyramidal cells, byinhibiting other large basket cells (disinhibition). Kisvarday and Eysel (1993) on the basisof combined anatomical connections and physiological mapping provided evidencesupporting a multifunctional role for basket cells. They also provide evidence that the large136basket cells contain elements of both cross- and iso-orientation inhibition. Usingiontophoretic application of GABA and simultaneous recordings of single units at remotelocations, they have demonstated that basket cells can influence orientation selectivitywithin lateral distances of 500-600 Im by cross-orientation inhibition and directionselectivity over distances of 1-2.5 mm by iso-orientation inhibition (Eysel 1992).My study also shows that callosal projections in layer 2/3 arise mainly from amorphologically heterogenous population of standard pyramids. All the callosal pyramidalcells possesed spines. The most common cell type in the callosal population is the standardpyramid (86/98), but a few star pyramids (6/98) and fusiform pyramids (6/98) were alsofound. Included in the standard pyramidal category are pyramids with dendrites originatingfrom the upper portion of the soma (10/86), pyramids with bifurcated apical dendrites(7/86) and pyramids with short apical dendrites (4/86). As noted by others (Buhl andSinger 1989) these differences in the morphology do not seem to be related to the laminarlocation of the cell or its proximity to the 17/18 border.My study, in addition, reveals that the basilar dendritic trees of the local patchpyramidal neurons are organized along two main lines. Injections of prelabeled neurons intangential sections revealed pyramidal neurons with either elongated or circular dendriticfields. All of the elongated fields were oriented along the ML axis. Using Golgi technique,Colonnier (1964), studied the tangential organization of the dendrites of cat visual cortexand concluded that the preferred axis of dendritic elongation of layer 2/3 neurons is the APaxis, however, the population analysis used in my study revealed that the preferred axis ofelongation of the local patch cells is the ML axis. The apparent discrepancy suggests thatthe asymmetry seen in this study is specific for the local patch neurons, while the trendseen by Colonnier is present in a separate population of pyramidal cells. My study alsoreveals that in the tangential plane, the callosal cells possess circular basal dendritic fields.137Finally, my study, based on double-labeling experiments, show that the local patch andcallosal cells form separate populations but not separate patch networks in the superficiallayers of cat area 18.4.3. TECHNICAL CONSIDERATIONSThe primary goal of this study was the identification and the morphologicalcharacterization of neurons associated with the local patch and the callosal systems.Neuroanatomical staining techniques alone provide limited infonnation with respect to theprojection targets of labeled neurons. Conversely, most retrograde tracing techniques rarelyreveal the detailed morphology of projection neurons. Thus, when these techniques areused separately, it may be difficult to even discriminate local circuit neurons fromprojection neurons, let alone, address whether subsets of neurons are linked to multipleprojection systems. The combination of retrograde tracing and intracellular filling of theprelabeled neurons in fixed tissue is most appropriate to determine the morphology ofneurons with identified projection targets (Tauchi and Masland 1984; Buhl and Lübke1989). Due to the absence of electrophysiological criteria, intracellular staining in fixedtissue is done under visual control. Thus, prior to the injection procedure, neurons areprelabeled by retrograde tract-tracing. Subsequently, the detailed morphology of theprelabeled neurons may be revealed by intracellular dye injection. Currently, one of themost popular fluorescent intracellular markers is LY (Stewart 1981). It is a brilliantlyfluorescent dye which allows the visualization of the filled neurons without furtherhistochemical treatment in living and fixed tissue and due to its small size (MW= 457) itdiffuses more rapidly and is easily ejected from micropipettes. However, the disadvantageof fluorescent markers is that they are susceptible to fading and the time one has to view thecells for detailed analysis is limited. To obtain a highly defined permanant image from afluorescence-stained cell other techniques were tried. One of the techniques was the photo-138oxidization of the LY-filled cells. This technique was first described by Maranto (1982) andrequires irradiating a LY-filled cell with bright light in the presence of diaminobenzidine.The procedure results in the transformation of LY into an electron-dense, brown precipitatethat is visible in both the light and the electron microscope. However, this method requiresa brightly fluorescent cell for the photoconversion reaction to be effective and therefore thefiner, less bright neuronal elements are not fully photoconverted. In addition, thephotoconversion of LY is a time consuming procedure and is impractical for large amountsof tissue with filled neurons. Thus, the photoconversion technique was not suitable for thepresent study where large numbers of cells were studied and measurements, such asdendritic tree width, were required. It should be noted that the photoconversion techniquecan be used in studies which are mainly interested in the identification of the morphology ofthe cells, but does not require quantitative measurements. Another technique tried wasimmunohistochemistry using an antibody to LY. When 220 urn thick slices were used, theantibody against LY could not penetrate into the middle of the thick sections (even afterusing a detergent, Triton-X 100) resulting in partially immunoreacted neurons. Thus, forsuccessful immunoreaction, resectioning of the thick slices is required, and hence, serialreconstructions. I chose to use confocal microscopy in combination with retrograde tracingand intracelluar filling in fixed slices as it offered many advantages and allowed me to studya large number (n= 370) of neurons from intact, thick slices.4.3.1. ADVANTAGES OF USING FIXED SLICESUntil recently, it has been technically difficult to study the various forms of pyramidalcells systematically. The combination of retrograde tracing and intracellular dye injection inthe fixed brain slices has proven useful in dernarcating separate classes of projectionneurons in the visual cortex (e.g., Katz 1987; Hübener et al. 1990). Two of the advantagesof intracellular filling in fixed slice preparation are that, first a group of projection-specific139neurons may be studied, and second, unlike in vivo cell filling, the fixed slice methodallows the study of a large population of intracellularly injected neurons per animal. Thus,given the enormous morphological variability of pyramidal neurons I wanted to use apopulational approach to determine the relationship between specific pyramidal cell formsand their projection pathway and the fixed slice method was suitable for my study.The major limitation with the intracellular staining in fixed tissue is the poor filling ofaxonal collaterals (Buhl and Lübke 1989). In contrast, in living tissue the staining of axoncollaterals appears to be more complete (Katz 1987; Hübener et al. 1990). Anotherimportant concern in using fixed slices is the question of completeness of the dendritic fills.This concern was addressed by Hübener et al. (1990) who directly compared the live andfixed techniques and concluded that the fixed slice technique is appropriate if one is mainlyinterested in the basal dendritic morphology of cells.There are two more potential methodological concerns. First, the retrograde labelingtechnique may favour larger cells with larger diameter axons over smaller cells, when usedfor tracing callosal cell group. This is unlikely since in the callosal sample small cells werealways prelabeled with medium and large cells. Second, the truncation of dendrites in 220jim thick sections. Since the dendritic fields of many of the cells, particularly in the callosalsample are wider than 220 jim, it is possible that the dendrites exiting the section of somecells would have been truncated. This is the reason why the cells were analysed in coronal,sagittal and tangential planes in this study. In the tangential plane, the callosal cellspossessed circular basal dendritic fields and local patch cells possessed ML elongated fieldsand circular fields. For a circular cell, width taken in any plane should be nearly equal. Foran ML elongated cell the longest width is in the ML direction and measurements in thecoronal plane would be optimal for these cells. Thus, we think that measurements of140dendritic width taken in the coronal plane should provide the closest estimate of thedendritic field widths.4.3.2. ADVANTAGES OF CONFOCAL IMAGINGThe confocal microscope is an extremely useful tool that allows fluorescently labeledcells to be visualized in whole-mount preparations. This is particularly useful in studyingthe dendritic trees of neurons in three 3-dimensions. Optical serial sectioning of thickspecimens and reconstructing the images digitally provides access to information on therelationships of a cell and its processes intact within its tissue. This eliminates the potentialerror that physical sectioning and manual reconstruction may introduce. Specimens arescanned by a laser beam, momentarily focused by the objective on a single point in theimage plane. Light returning from this point passes through a small aperture to thephotodetector. Image-degrading light flare from nearby points misses the aperture and failsto reach the photodetector. In this way the confocal microscope rejects scattered light andout-of-focus blur, thus enabling the extraction of thin optical sections of high contrast andresolution from thick biological sections (Amos et al. 1987; Fine et al. 1988; Carlsson et al.1989; Shotton 1989). The successive images of the entire three-dimensional (3-D) dendritictree of an intracellularly injected neuron, for example, can be captured and compressed intoa two dimensional image (project image) with extreme accuracy. Thus, compared withtraditional methods, confocal microscopy has the following advantages: (1) a highresolution image can be obtained from an intact thick tissue, due to the substantialattenuation of stray light and of light from out-of-focus structures, (2) the process ofobtaining a permanant image is less time consuming, (3) the problem of fluorochromebleaching is minimal as the laser illumination system of a confocal microscope illuminatesonly a tiny fraction of the total observed area of the specimen at any one time and thusdramatically reduces bleaching effects in comparison to standard full-field epifluorescence141methods. Three-dimensional (3-D) reconstructions of neurons has proven useful in earlierstudies (Wallen et al. 1988; Brodin et al. 1988). In my study, I did not address the 3-Dorganization of dendritic trees. However, data stored on computer disks will allow future3-D analysis of the identified projection neurons studied here.4.4. COMPARISON OF THE INTRINSIC NEURONS AND EXTRINSICNEURONS IN LAYER 2/3When the local patch pyramidal cell types were compared to the callosal cell types itwas found that the star pyramids and the standard pyramids were found in both systems,but the modified pyramids of layer 2 which were present in the local patches were notpresent in the callosal projection. On the other hand, fusiform pyramids, standard pyramidswith dendrites originating from the upper portion of the cell body and pyramids with shortor bifurcated apical dendrites were only present in the callosal population. These findingsindicate that only the standard and star cell types are likely to participate in both projectionsystems.Comparisons of morphological types of pyramidal neurons across visual cortical areaswere possible by reviewing the literature. Noteworthy is the comparison between theresults of this study and those reported by Einstein and Fitzpatrick (1991). When themorphological types of pyramidal neurons found in my study in area 18 were compared tothe types reported for the corticocortical projection in area 17 (Einstein and Fitzpatrick1991), it was found that the modified pyramids of layer 2 and the standard pyramids oflayer 2/3 were present in both systems. However, the local patch population appeared tolack the inverted pyramids of layer 3 which were present in the corticocortical projections(Einstein and Fitzpatrick 1991). All other studies on corticocortical cells were focused onarea 17 neurons and thus, we were concerned about potential differences between areas142complicating our findings. For this reason, studies done in our lab were aimed atidentifying area 18 corticocortical cells in order to compare these projection populationswith the local patch neurons of area 18 directly (Matsubara et al., 1993). According to thestudies of Matsubara et al. (1993), the vast majority of corticocortical cells were pyramidalcells (99/105), but a small number of nonpyramidal (3/105) and fusiform cells (3/105)were also observed. The most common pyramidal cell type reported was the standardpyramid (88/99), but modified (7/99), star (2/99) and inverted pyramids (2/99) were alsopresent. Thus, the standard and modified pyramidal cell types of area 18 appear toparticipate in both the local patch and corticocortical systems in agreement with Einstein andFitzpatrick’s (1991) study. In addition, star pyramids (although relatively rare) also appearto participate in both local patch and corticocortical systems (Matsubara et al. 1993).All three projection populations contained standard pyramids, but of different sizeranges. Based on mean soma size, local patch cells form the smallest group, callosal cellsform the largest group and corticocortical cells were intermediate in size range. Quantitativecomparison of the basal dendritic field widths and the complexity of branching patternmeasured by the total number of basal dendritic branch points also showed similardistribution. On average, the local patch cells possesed smaller basal dendritic trees withless complex branching pattern, the callosal pyramids possesed larger basal dendritic treeswith more complex branching pattern and the corticocortical pyramids were againintermediate in dendritic field size and complexity.1434.5. DENDRITIC ORGANIZATION OF LOCAL AND CALLOSAL CELLS:POSSIBLE FUNCTIONAL IMPLICATIONSThe finding that local cells are the smallest group, callosal cells are the largest andcorticocortical cells are intermediate in somatic size range may simply reflect the fact thatcells which project the farthest have the largest axons and hence the biggest cell bodies andgreatest dendritic complexity. However, this does not appear to be true based on thefindings of my study. Cells of the same size belonging to separate pathways show pathwayspecific differences in the dendritic width and branch points. Alternatively, the differencesin the size and complexity of dendritic trees between the three populations may indicate thatcallosal pyramids may potentially sample more afferents compared to the corticocortical andlocal patch pyramids. This view is consistent with the findings reported in someparasympathetic and sympathetic ganglia (Purves and Hume 1981; Purves and Lichtman1985) that the dendritic arbor size and complexity are proportional to the number ofdifferent innervating axons. The studies of Purves and Hume (1981) and Purves andLichtman (1985) showed that the number of primary dendrites, total dendritic length, totalnumber of terminal dendritic branches and the number of dendritic branch points ofindividual neurons are all proportional to the number of inputs they receive. It was furthersuggested that, since all the ganglion cells studied innervate the same end organ, thedifferences in the geometry of the cells are related to the convergent innervation they receiveand not to the differences in their targets (Purves and Lichtman 1985). In addition, sincethe callosal projection contains a heterogenous population of pyramidal cell types, it maysample the outputs of many circuits within the superficial layers compared to the local patchcells. All of these factors may contribute to differences in the integrative properties ofneurons of the three systems. In an effort to correlate form and function of corticalneurons, several studies have looked at the relationship between the receptive fieldproperties of neurons and the size and shape of their dendritic trees, based on injections of144HRP into neurons whose receptive field properties have been determined. For example,Gilbert and Wiesel (1979) reported in cat visual cortex, that cells with larger dendriticarborizations have larger receptive fields. However, Martin and Whitteridge (1984b) didnot find any correlation between the size of the dendritic tree and the size of the receptivefield. Martin (1984) suggests that the size of the dendritic tree of a cell may only relate tothe number of synapses made onto the cell, and that the receptive field size may be relatedto the particular afferents synapsing on the dendritic tree rather than the size of the dendritictree.The morphology of the cell may also be influenced by the particular afferents synapsingon the cell. For example, the modified pyramids of layer 2 which were present in the localpatch sample, had their basal dendrites turned towards layer 1. These cells likely receivetheir major afferent input from layer 1. On the other hand, callosally projecting shortpyramids of layer 3, had apical dendrites that did not reach to layer 1. Thus, it is almostcertain that the short pyramidal cells do not receive afferent input from layer 1. Aspreviously discussed in section 1.3.4., the afferents originating from the Y- and W cells ofthe LGN terminate in different layers in area 18 (Boyd and Matsubara, in preparation).Thus, based on their dendritic patterns the modified and short pyramids may receive signalsfrom separate LGN pathways, for instance, the modified pyramid may receive signals fromlayer 1 W type axons and short pyramid may receive signals from Y type axons in lowerlayer 3. Thus, although the modified and short pyramids are both present in the superficiallayers, the combination of afferents they receive may be different.The efferent cells projecting to different cortical areas are also organized at differentlevels within the superficial layers in area 18. For example, in the superficial layers of area18, the cells projecting to PMLS are found deeper than the cells projecting to area 19(Symonds and Rosenquist 1984). Thus, it will be of interest to find out whether the145modified and short pyramids inturn contact different neurons in the superficial layers whichproject to separate hierarchies of visual areas. The rich connections of striate cortex is suchthat, one can be easily be convinced that all the layers are interconnected and all the cellscan receive input from the afferents directly or indirectly (Gilbert and Wiesel 1983; Lund1988). However, it should be recognized that certain anatomical circuits should stand outabove the rest to create particular functional properties of cortical neurons. In the abovediscussion, I took two examples, the modified and short pyramids, one from the localpatch system and the other from the callosal system to speculate on how differentmorphological types may participate in different circuits within the superficial layers. Thesame speculation can also be extended to different morphological types within the samesystem.The finding of elongated and circular dendritic fields of layer 2/3 pyramidal cells in thetangential plane is consistent with the results from the studies by Hübener and Bolz (1992),who also reported similarly organized dendritic trees of layer 2/3 pyramidal cells in primatevisual cortex. From their fixed slice studies, Hübener and Bolz (1992) further suggested arelationship between the dendritic organization and CO blobs. They reported that thedendrites of most (but not all) pyramidal cells close to CO blob borders tended to confinetheir dendrites to only one compartment, although Malach (1992), using Golgi staining,reported that the dendrites of CO blob cells freely crossed compartment borders. Recentimprovements in CO histochemistry have made it possible to visualize CO blobs in catareas 17 (Murphy et al, 1990; Dyck and Cynader, 1992) and 18 (Boyd and Matsubara,submitted), making it possible to study the relationship of CO blobs to the dendrites oflocal patch neurons in the cat. Given that nearly half of the local patch cells in this studypossesed mediolaterally elongated dendritic fields, it is difficult to imagine a strictavoidance of CO blob borders, unless the blobs themselves are mediolaterally elongatedand local patch neurons with elongated dendritic fields are centered within the blobs.146Alternatively, perhaps local patch neurons with circular dendritic fields are located in thecenters of CO blobs, while mediolaterally elongated local patch neurons associate withinterblob areas of area 18. These issues are presently being addressed in our laboratory.Recently, both Malach (1992) and Lund et al. (1993) have suggested corticalorganization schemes based on the relationship between the spread of dendrites and thedimensions of CO blobs or local patch width across areas and species respectively. Malach(1992), based on his studies in marmoset and squirrel monkeys, reported that the size ofthe pyramidal cell dendritic arbors is somewhat smaller, but roughly scaled equal to theblob size, to create a smooth transition of dendritic sampling of different proportions ofinputs from blob and interblob territories. Malach (1992) further suggests that under theconditions of free dendritic crossing of blob margins, the ratio of dendritic field width andthe blob size, becomes a critical factor in determining the level of blob/interbiob mixing andthis constant ratio is likely preserved across different species. Lund et al. (1993) reportedthat the average lateral spread of the basal dendritic fields of single pyramidal neurons inlayer 2/3, matches the mean local patch size in several cortical areas and species, despitedifferences in scale across these areas and species. They suggest that this architecturalfeature might be a fundamental aspect of cortical organization.The elongated dendritic fields of layer 2/3 local patch pyramidal cells may be related toaxonal field arborizations. Little is known about the number, direction and extent of theafferent inputs to the individual layer 2/3 neurons. While few studies have addressed thisissue directly, there is substantial indirect evidence for a preferred direction of axonal fieldsin cat visual cortex. First, the terminal arbors of geniculocortical Y axons are asymmetric,being two to four times longer anteroposteriorly than mediolaterally (Humphrey et al.1985). Y axons arborize in lower layer 3, in addition to layer 4 (LeVay and Gilbert 1976;Freund et al. 1985a) and make synaptic contacts with layer 3 pyramidal cells (Freund et al.1471985b). Second, in vivo filling of single cells in the superficial layers of cat visual cortexindicate that pyramidal cells themselves display axonal fields that extended greater distancesanteroposteriorly than mediolaterally (Kisvarday et al. 1986). This is consistent with earlierstudies on intrinsic connections based on anterograde degeneration suggesting that themajority of local intrinsic axons run anteroposteriorly (Creutzfeldt et al. 1975).Retrogradely labeled local patch cells are usually found at greater distances from theinjection site along the anteroposterior axis, again suggestive that the axons are longeralong this axis (Boyd and Matsubara 1991; Matsubara and Boyd 1992). Thus, it appearsthat both thalamic and intrinsic fibers course and arborize over larger cortical domains alongthe anteroposterior axis. Martin (1988), based on EM findings of HRP-filled geniculateaxons contacting Golgi stained cortical cells, concluded that a cortical cell is normallydriven by the convergent action of at least tens of geniculate axons (Martin 1988; Freund etal. 1985a). The mediolateral elongation of basal dendritic trees of layer 2/3 local patchpyramidal cells may ensure maximal convergence with many anteroposterior oriented axonterminal fields. Such an organizational feature has already been observed, in a moreextreme case, for the purkinje cells of the cerebellar cortex. Purkinje cells have very planardendritic trees, presumably for optimizing contacts with parallel fibers, which intersectthem at right angles. The parallel axons running right angles to the purkinje cell dendritesmay allow maximal divergence of each axon to many cells and maximal convergence ofmany axons onto each cell, in minimal space (Fox and Barnard 1957).4.6. FUTURE DIRECTIONSMy study shows that the local patch cells and the callosal cells form separatepopulations in the superficial layers of cat area 18. The finding that no cells participated inboth projections suggest that the callosal neurons probably possess vertically oriented axoncollaterals or axon collaterals which predominantly ramify within the vicinity of the cell,148compared to the local patch neurons which possess patchy, horizontal collaterals thatextend laterally covering many cortical columns. In this respect, it will be worthwhile tostudy the axonal morphology of these cells to examine whether they exhibit a clearprojection specific pattern. The technique of identification of cells of origin by retrogradetransport of fluorescent tracers followed by LY injection of prelabeled cells in living slices(Katz 1987), can be used to address this issue.My study also shows that the local patch cells possess either circular or elongated basaldendritic fields and the callosal cells possess circular fields in the tangential plane. Whatthis means, functionally, is a question of great interest. The fact that layer 2/3 neurons canpotentially receive a wide variety of afferents suggest that many different circuits may bepresent within a single localized region of area 18. But exactly which afferents contact thesetwo populations of layer 2/3 cells and in what combination is not known and cannot bedeterniined without the analysis of the presynaptic elements of the injected cells.Experimental design based on the method described for visualizing three consecutivecomponents of a neuronal network (Wouterlood et al. 1990), for example, anterogradelabeling of the afferents with PHA-L (Gerfen and Sawchenko 1984), in combination withretrograde tracing, and intracellular filling of prelabeled neurons with biocytin (Horihawaand Armstrong 1988), can be used to address this issue. Another technique, which ispresently under investigation in our lab, that yields information about the presynaptic inputsto cells is the retrograde transneuronal transport of the tetanus toxin C-fragment (TTC)(Manning et al. 1990; Matsubara et al. 1991).1494.7. CONCLUSIONMy thesis was designed to address two major questions about the neurons in thesuperficial layers of cat area 18. These are, 1) whether individual pyramidal cells belongingto the local patch or callosal systems can be distinguished based on somatodendriticmorphology, and 2) whether or not the cells in these two systems comprise separate,nonoverlapping populations.Do the intrinsic local patch neurons differ in dendritic morphology from theprojection pyramids?When the morphology of the local patch pyramids was compared with the callosal andcorticocortical projection pyramids in area 18, certain morphological types were found toparticipate in both the local patch system and the callosal or corticocortical projectionsystem. For example, the standard pyramids and star pyramids were found in both the localpatch and callosal systems. But the modified pyramids were only present in the localpatches. On the other hand, fusiform pyramids, pyramids with short or bifurcated apicaldendrites and pyramids with dendrites originating from the upper portion of the soma wereonly present in the callosal population. Similar comparisons between the morphology oflocal patch pyramids and the corticocortical pyramids revealed that the modified pyramids,the star pyramids, and the standard pyramids of layer 2/3 were found in both these systems(Matsubara et al. 1993). Thus, it seems possible that certain morphological types mayparticipate in local connections and in projection systems. Based on mean values of somasize, basal dendritic field width, and total number of basal dendritic branch points, theseclasses of cells were markedly different. However, the soma size and the basal dendriticfield width measurements showed an overlap in the 170-520 Im2 size range, i.e.,comparison of the measurements of local patch and callosal pyramids, for example, shows150that cells with soma size of 170-520 jim2 and dendritic field width of 180-340 jim, werepresent in both systems. But within this overlap range local patch cells can be distinguishedfrom callosal cells based on the total number of branch points, since the local patchpyramids generally have fewer basal dendritic branch points than the callosal pyramids.Small pyramids with soma size of 90-170 jim2 and dendritic field width of 70-180 jimwere only present in the local patch population. On the other extreme, large pyramids withsoma size of 520-640 jim2 and dendritic field width of 340-390 jim were only present inthe callosal population.Do local patch pyramids project to other cortical areas?The compariso.n of the morphology of the local patch, corticocortical and callosalneurons shows that certain morphological types participate in more than one system. Thisfinding raises the question of whether the overlapping cell types are the same neuronswhich participate in more than one system or whether they are separate neurons withsimilar morphological characteristics. The size ranges and the organization of the basaldendrites suggest that the three populations of cells may be separate. In order to addressthis issue double-labeling studies were undertaken. When both callosal and local patchpopulations in area 18 were labeled with two different fluorochromes, the two groups werefound in overlapping patches, but double-labeled cells were not found (Thejomayen et al.1992). Similar double-labeling studies revealed less than 5% of double-labeled cells in theregions of overlap of local patch and corticocortical cells (Matsubara et al. 1993). Thesefindings indicate that none of the local patch pyramids project callosally, while only a smallminority, less than 5% of the local patch neurons in area 18 also project corticocortically toarea 17. Earlier studies which examined the various projection populations in differentareas of cortex with the use of double-labeling technique (e.gs., Bullier et al. 1984; DeYoeand Van Essen 1985; Andersen et al. 1985; Segraves and Innocenti 1985) indicate that very151few cells participate in more than one ipsilateral corticocortical projection system and thatthe callosal and ipsilateral corticocortical projection populations are largely separate. Thefinding of my study adds to these studies by showing that the callosal pyramids are notonly separate from the ipsilaterally projecting corticocortical pyramids, but also are separatefrom the intrinsic local patch pyramids. The studies from our lab indicate that the ipsilateralcorticocortical projection pyramids are also largely separate from the intrinsic local patchpyramids. Less than 5% of the local patch neurons in area 18 participate in the ‘backwardipsilateral projection to area 17 (Matsubara et al. 1993). The contribution of local patchneurons of area 18 to ‘forward’ ipsilateral corticocortical projections remains to be beinvestigated. But the observations from our lab in area 17, on local patch neurons andneurons participating in ‘forward’ ipsilateral corticocortical projection to PMLS indicatesthat less than 1% of the local patch neurons project to PMLS. Thus, in general, a singlepyramidal cell in the superficial layers participates either in local connections or in an outputprojection to a single target. The assignment of different populations of neurons to provideinput to different targets is believed to reflect the extreme specificity of cortical connectionsby which functional separation and precise channeling of information is achieved in visualcortex (Bullier et al. 1984; Segraves and Innocenti 1985).Although the classical concept of pyramidal cells is that they are ‘projection’ neurons(Cajal 1911), subsequent Golgi studies reported some classes of pyramidal cells withoutaxonal projections into the white matter (O’Leary 1941; Lorente de No 1949; Sholl 1955;Lund 1973). From intracellular injections of HRP, Gilbert and Wiesel (1983), also havefound some pyramidal cells in the superficial layers of cat visual cortex with extensiveintrinsic axon collaterals, but with no descending axon entering white matter. Thesepyramidal cells with extensive local axon collaterals and no descending axons mayparticipate in local circuits exclusively and may represent the local patch neurons of thesuperficial layers. Katz (1987), based on the study of the morphology of layer 6 cells in152living cat visual cortical slices reported an unexpectedly large proportion (20%) of intrinsiclayer 6 cells with no efferent axon. Technical limitations imposed by Golgi and intracellularstaining is believed to be responsible for underestimating the population of the localintrinsic cells of the cortex (Katz 1987). Thus, in the living slice study of Katz (1987),where large numbers of cells are studied and good axonal filling is obtained, the largenumber of intrinsic pyramidal cells encountered may actually reflect the true frequency ofthe intrinsic neurons in visual cortex.Further studies are needed to determine the origin of the local patch pyramids andwhether they arise during development after axonal retraction from one of the transienttarget areas (Innocenti et al. 1986; Price and Ferrer 1993). The studies of Innocenti et al(1986) shows that about 15-20% of transitorily callosal neurons in area 17, laterparticipate, in ipsilateral corticocortical projections to areas 17 and 18. Tnnocenti et al.(1986) suggest that some tansitorily callosal neurons may become neurons whichparticipate in intrinsic local connections. In newborn kittens, cells in area 17 that project toarea 18 are uniformly distributed in the superficial layers. During postnatal weeks 2-3,some of these corticocortical connections are removed to generate a clustered adult-likepattern in the superficial layers (Price and Blakemore, 1985a). The formation of theseclusters is achieved by axonal elimination, without cell death in the superficial layers (Priceand Blakemore, 1985b). Price and Ferrer (1993) suggest that the transient corticocorticalcells after axonal elimination, may become neurons with exclusively intrinsic connections.Many questions emerge from the findings of this study. Why the local patch systemand the callosal and ipsilateral corticocortical systems form overlapping rather thaninterdigitating patch networks in visual cortex ? What would be the purpose of havingseparate sets of neurons consisting of a variety of morphological types, in each system?How the local patch neurons interact functionally with the spatially intermingled efferent153neuronal sets ? To investigate the nature of the interactions between the local patch neuronsand the efferent neurons destined to higher cortical areas, it is important to determine thepre- and post synaptic elements of the local patch cells. Since the input and output to thevisual cortex are segregated at various levels within the superficial layers, it is tempting tospeculate that local patch neurons at different levels may receive inputs from a combinationof different sets of afferents and may provide output to neurons which project to differentextrastriate areas. Recent studies of Lachica et al. (1993) in squirrel monkeys and bushbabies, using small injections of HRP and biocytin restricted to different sublayers of layer3, shows that these sublayers, 3A, 3B and 3C, receive different inputs and have verydifferent output patterns; 3A and 3C provide output to separate hierarchial areas whereas3B acts as an interneuronal layer projecting only to 3A. Thus, each sublayer appears tocontribute to different cortical circuits. It will be interesting to find out with the use of thesame technique, whether similar sublayers comparable to the primate sublayers of layer 3,exist in cat visual cortex and if so, what are the input and output of these sublayers.Clearly, further careful studies are required to obtain an insight into how information aboutdifferent visual attributes arriving via different pathways to the visual cortex are combinedthrough precise local circuits to provide a unified visual perception.154REFERENCESAbeles M, Goldstein MH Jr. (1970) Functional architecture in cat primary auditory cortex:Columnar organization and organization according to depth. J Neurophysiol 33: 172-187.Albus K (1979)‘4C-Deoxyglucose mapping of orientation subunits in the cat’s visualcortical areas. Exp Br Res 37:609-6 13.Albus K, Donate-Oliver F, Sanides D, Fries W (1981) The distribution of pontineprojection cells in visual and association cortex of the cat: An experimental study withhorseradish peroxidase. J Comp Neurol 201:175-189.Albus K, Wahie P, Lübke J, Matute C (1991) The contribution of GABA-ergic neurons tohorizontal intrinsic connections in upper layers of the cat’s striate cortex. Exp Brain Res85: 235-239.Amos WB, White JG, Fordham M (1987) Use of confocal imaging in the study ofbiological structures. Applied Optics 26(16): 3239-3243.Andersen P, Eccies JC, Voorhoeve PE (1963) Inhibitory synapses on somas of Purkinjecells in the cerebellum. Nature 199:655-656.Andersen P, Lomo T (1966) Mode of activation of hippocampal pyramidal cells byexcitatory synapses on dendrites. Exp Br Res 2:247-260.Andersen RA, Asanuma C, Cowan WM (1985) Callosal and Prefrontal Associationalprojecting cell populations in Area 7A of the Macaque Monkey: A study usingretrogradely transported Fluorescent dyes. I Comp Neurol 232: 443-455.Antonini A, Stryker MP (1993) Rapid Remodeling of Axonal Arbors in the visual cortex.Science 260:1819-1821.Barr ML (1979) Histology of the Cerebral Cortex. In The Human Nervous system. Thirdedition. pp.179-185.Berlucchi G (1972) Anatomical and physiological aspects of visual functions of the corpuscallosum. Brain Res 37: 37 1-392.Berlucchi G (1980) Recent advances in the analysis of neural substrates of interhemisphericcommunication. Satellite Symposium of the International Brain Research Organization,Pisa, Italy.Berman NE, Payne BR (1983) Alterations in connections of the corpus callosum followingconvergent and divergent strabismus. Brain Res 274:201-212.Blasdel GG, Yoshioka T, Levitt JB, Lund JS (1992) Correlation between patterns of lateralconnectivity and patterns of orientation preference in monkey striate cortex. Soc N Scabst 169.4.Boyd JD, Matsubara JA (1991) Intrinsic connections in cat visual cortex: a combinedanterograde and retrograde tracing study. Brain Res 560:207-215.155Boyd JD, Matsubara JA (1992) Segregated processing streams in cat visual cortex?Relationship of patchy connectivity to an extrastriate area, cytochrome oxidase stainingand local connections. Soc N Sc abst, 134.4.Boyd JD, Matsubara JA (1994) The tangential organization of callosal connectivity in thecat’s visual cortex. J Comp Neurol. SubmittedBoyd JD, Matsubara JA Laminar and Columnar patterns of geniculocortical projections incat: Relationship to cytochrome oxidase staining. Manuscript in preparation.Brodin L, Ericsson M, Mossberg K, Hökfelt T, Ohta Y, Griliner S (1988) Three-dimensional reconstruction of transmitter-identified central neurons by “en bloc”immunofluorescence histochemistry and confocal scanning microscopy. Exp Brain Res73: 441-446.Brodmann K (1909) Vergleichende Lokalisationslehre der Grosshirnrinde in IhrenPrinzipien dargestellt auf Grund des Zellenbaues. Barth, Leipzig. Cited in Kemper andGalaburda (1984).Buhl EH, LUbke 1 (1989) Intracellular lucifer yellow injection in fixed brain slicescombined with retrograde tracing, light and electron microscopy. Neuroscience28(l):3-16.Buhi EH, Singer W (1989) The callosal projection in cat visual cortex as revealed by acombination of retrograde tracing and intracellular injection. Exp Brain Res 75:470-476.Buhl EH, Schwerdtfeger WK, Germroth P (1990) Intracellular injection of neurons infixed brain tissue combined with other neuroanatomical techniques at the light andelectron microscopic level. In Handbook of Chemical Neuroanatomy. Vol.8: Analysisof Neuronal Microcircuits and Synaptic Interactions. pp. 273-304.Bullier J, Kennedy H, Salinger W (1984) Branching and laminar origin of projectionsbetween visual cortical areas in the cat. J Comp Neurol 228:329-341.Cajal SRy (1911) Histologie du systeme nerveux de l’homme et des vertebres. Trans by SAzoulay Paris: Maloine, vol 2.Callaway EM, Katz LC (1990) Emergence and refinement of clustered horizontalconnections in cat striate cortex. J. Neurosci 10:1134-1153.Callaway EM, Katz LC (1991) Effects of binocular deprivation on the development ofclustered horizontal connections in cat striate cortex. Proc Nati Acad Sci USA 88:745-749.Carlsson K, Wallen P, Brodin L (1989) Three-dimensional imaging of neurons byconfocal fluorescence microscopy. J Microscopy 155(1): 15-26.Colonnier M (1964) The tangential organization of the visual cortex. I Anat 98:327-344.Colonnier M (1968) Synaptic patterns on different cell types in the different laminae of catvisual cortex. Brain Res 9:268-287.156Creutzfeldt OD, Garey U, Kuroda R, Wolff J (1975) The distribution of degeneratingaxons after small lesions in the intact isolated visual cortex. Exp Br Res 27:419-440.Cusick CG, Gould HJ, Kaas JH (1984) Interhemispheric connections of visual cortex ofowl monkeys (Aotus trivirgatus), marmosets (Callithrix jacchus), and galagos (Galagocrassicaudatus). J Comp Neurol 230: 3 11-336.Cynader M, Gardner J, Dobbins A, Leporé F, Guillemot J-P (1986) Interhemisphericcommunication and binocular vision: function and developmental aspects. In: Leporé F,Ptito M, Jasper HH (eds) Two hemispheres - one brain: functions of the corpuscallosum. Alan R Liss, New York, pp 189-209.Defelipe J, Fairen A (1982) A type of basket cell in superficial layers of the cat visualcortex. A Golgi-electron microscopic study. Br Res 244: 9-16.Defelipe J, Hendry SHC, Jones EG, Schmechel D (1985) Variability in the terminations ofGAB Aergic chandelier cell axons on initial segments of pyramidal cell axons in themonkey sensory-motor cortex. J Comp Neurol 231:364-384.Deitch JS, Rubel EW (1984) Afferent influences on brain stem auditory nuclei of thechicken: Time course and specificity of dendritic atrophy following deafferentation. JComp Neurol 229: 66-79.Deitch JS, Rubel EW (1989) Rapid changes in ultrastructure during deafferentationinduced dendritic atrophy. J Comp Neurol 281: 234-258.DeYoe EA, Van Essen DC (1985) Segregation of efferent connections and receptive fieldproperties in visual area V2 of the macaque. Nature 317: 58-61.Dyck R, Cynader M (1992) Enzymes, ions and receptors distinguish novel,complementary columnar systems in developing cat visual cortex. Soc N Sc Abst,552.11.Einstein G, Fitzpatrick D (1991) Distribution and Morphology of Area 17 neurons thatProject to the Cat’s Extrastriate Cortex. J Comp Neurol 303:132-149.Eysel UT (1992) Lateral inhibitory interactions in areas 17 and 18 of the cat visual cortex.Progress Brain Res 90: 407-422.Fairén A, Valverde F (1980) A specialized type of neuron in the visual cortex of the cat. : AGolgi and electron microscopic study of the chandelier cells. J Comp Neurol 194:76 1-779.Fairén A, Defelipe I, Regidor 1 (1984) Nonpyramidal neurons: General account. In: TheCerebral Cortex. Vol 1, Cellular components of the Cerebral Cortex (Peters A andJones EG, eds.) Plenum Press, New York, pp. 201-253.Feldman ML, Peters A (1978) The forms of non-pyramidal neurons in the visual cortex ofthe rat. J Comp Neurol 179: 76 1-794.Ferrer IMR, Price DJ, Blakemore C (1988) The organization of corticocortical projectionsfrom area 17 to area 18 of the cat’s visual cortex. Proc R Soc Lond B 233:77-98.157Ferrer JMR, Kato N, Price DJ (1992) Organization of association projections from area 17to areas 18 and 19 and to suprasylvian areas in the cat’s visual cortex. J Comp Neurol316: 261-278.Fine A, Amos WB, Durbin RM, McNaughton PA (1988) Confocal microscopy:applications in neurobiology. TINS 11(8): 346-351.Fisken RA, Garey U, Powell TPS (1975) The intrinsic association and commissuralconnections of area 17 of the visual cortex. Phiilos. Trans. R. Soc. Lond Ser. B272:487-536.Fox CA, Barnard JD (1957) A quantitative study of the purkinje cell dendritic branchesand their relationship to afferent fibers. J Anat Lond 91: 299-308.Freund TF, Martin KAC, Smith AD, Somogyi P (1983) Glutamate decarboxylaseimmunoreactive terminals of Golgi-impregnated axoaxonic cells and of presumedbasket cells in synaptic contact with pyramidal neurons of the cat’s visual cortex. JComp Neurol 221:263-278.Freund TF, Martin KAC, Somogyi P, Whitteridge D (1985a) Innervation of cat visualareas 17 and 18 by physiologically identified X- and Y- type afferents. I. Arborizationpatterns and Quantitative distribution of postsynaptic elements. J Comp Neurol 242:263-274.Freund TF, Martin KAC, Somogyi P, Whitteridge D (1985b) Innervation of cat visualareas 17 and 18 by physiologically identified X- and Y- type afferents. II. Identificationof postsynaptic targets by GABA immunohistochemistry and Golgi impregnation. JComp Neurol 242: 275-29 1.Freund TF, Magloczky Z, Soltész I, Somogyi P (1986) Synaptic connections, axonal anddendritic patterns of neurons immunoreactive for cholecystokinin in the visual cortex ofthe cat. Neuroscience 19, 1133-1159.Gabbott PLA, Somogyi p (1986) Quantitative distribution of GABA-immunoreactiveneurons in the visual cortex (area 17) of the cat. Exp Brain Res 61: 323-331.Gerfen CR, Sawchenko PE (1984) An anterograde neuroanatomical tracing method thatshows the detailed morphology of neurons, their axons and terminals:Immunocytochemical localization of an axonally transported plant lectin, Phaseolusvulgaris lecoagglutinin (PHA-L). Brain Res 190: 219-238.Gilbert CD, Kelly JP (1975) The projections of cells in different layers of the cat’s visualcortex. J Comp Neurol 163:81-106.Gilbert CD, Wiesel TN (1979) Morphology and intracortical projections of functionallyidentified neurons in the cat visual cortex. Nature 280:120-125.Gilbert CD (1983) Microcircuitry of the visual cortex. Ann Rev Neurosci 6:217-247.Gilbert CD, Wiesel TN (1983) Clustered intrinsic connections in cat visual cortex. JNeurosci 3:1116-1133.Gilbert CD, Wiesel TN (1989) Columnar specificity of intrinsic horizontal andcorticocortical connections in cat visual cortex. I Neurosci 9:2432-2442.158Gray EG (1959) Axo-somatic and axo-dendritic synapses in the cerebral cortex: Anelectron microscopic study. J Anat 93:420-433.Greenough WT, Chang FF (1988) Dendritic pattern formation involves both orientedregression and oriented growth in the barrels of mouse somatosensory cortex. DevBrain res 43:148-152.Goidman-Rakic PS, Schwartz ME (1982) Interdigitation of contralateral and ipsilateralcolumnar projections to frontal association cortex in primates. Science 216: 755-757.Golgi C (1883) Recherches sur l’histologie des centres nerveux Arch Ital Biol 4:92-123.Hailman LE, Schofield BR, Lin CS (1988) Dendritic morphology and axon collaterals ofcorticotectal, corticopontine, and callosal neurons in layer V of primary visual cortex ofhooded rat. I Comp Neurol. 272: 149-160.Hamlyn LH (1963) An electron microscopic study of pyramidal neurons in the Ammon’shorn of the rabbit. J Anat 97:189-201.Hamori J, Szentagothai J (1965) The Purkinje cell baskets: Ultrastructure of an inhibitorysynapse. Acta Biol Hung 15:465-479.Harris RM, Woolsey TA (1979) Morphology of Golgi-impregnated neurons in mousecortical barrels following vibrissae damage at different post-natal ages. Brain Res161:143-149.Hendrickson AE (1985) Dots, stripes and columns in monkey visual cortex. TrendsNeurosci 8: 406-410.Henry GH, Salin PA, Bullier J (1991) Projections from Areas 18 and 19 to cat striatecortex: Divergence and Laminar Specificity. European J Neurosci 3: 186-200.Horihawa K, Armstrong WE (1988) A versatile means of labeling: Injection of biocytinand its detection with avidin conjugates. J Neuroscei. Meth 25: 1-11.Hornung JP, Garey U (1981) Ultrastructure of visual callosal neurons in cat identified byretrograde axonal transport of horseradish peroxidase. J Neurocytol 10:297-314.Horton JC, Hubel DH (1981) A regular patchy distribution of cytochrome oxidase stainingin primary visual cortex of the macaque monkey. Nature 292: 762-764.Hubel DH, Wiesel TN (1962) Receptive fields, binocular interaction and functionalarchitecture in the cat’s visual cortex. J Physiol (London) 165:559-568.Hubel DH, Wiesel TN (1969) Anatomical demonstration of columns in monkey striatecortex. Nature 22 1:747-750.Hubel DH, Wiesel TN (1977) Functional architecture of macaque monkey visual cortex.Proc Roy Soc (London) ser B 198:1-59.Hubel DH, Wiesel TN, LeVay S (1977) Plasticity of ocular dominance columns in monkeystriate cortex. Phil Trans Roy Soc B 278:377-409.159Hubel DH, Wiesel TN, Stryker PM (1978) Anatomical demonstration of orientationcolumns in macaque monkey. J Comp Neurol 177:361-380.Hübener M, Bolz J (1988) Morphology of identified projection neurons in layer 5 of ratvisual cortex. Neurosci Lett 94: 76-8 1.Hübener M, Schwarz C, Bolz J (1990) Morphological types of projection neurons in layer5 of cat visual cortex. J Comp Neurol 301:655-674.HUbener M, Bolz J (1992) Relationships between dendritic morphology and cytochromeoxidase compartments in monkey striate cortex. J Comp Neurol 324:67-80.Humphrey AL, Sur M, Uhlrich Di, Sherman SM (1985) Termination patterns ofindividual X- and Y- cell axons in the visual cortex of the cat: Projections to area 18, tothe 17/18 border region, and to both areas 17 and 18. J Comp Neurol 233:190-212.Innocenti GM, Fiore L (1976) Morphological correlates of visual field transformation in thecorpus callosum. Neurosci Lett 2:245-252.Innocenti GM, Fiore L, Caminiti R (1977) Exuberant projection into the corpus callosumfrom the visual cortex of newborn cats. Neurosci. Lett. 4:237-242.Innocenti GM, Frost DO (1979) Effects of visual experience on the maturation of theefferent system to the corpus callosum. Nature 280:231-234.Innocenti GM, Frost DO (1980) The postnatal development of visual callosal connectionsinthe absence of visual experience or eyes. Exp Br Res 39:365-375.Innocenti GM, Caminiti R (1980) Postnatal shaping of callosal connections from sensoryareas. Exp Brain Res 38:38 1-394.Innocenti GM (1980) The primary visual pathway through the corpus callosum:morphological and functional aspects in the cat. Arch. Ital. Biol. 118; 124-188.Innocenti GM (1981) Growth and reshaping of axons in the establishment of visual callosalconnections. Science 212: 824-827.Innocenti GM, Frost DO, Illes J (1985) Maturation of visual callosal connections invisually deprived kittens: A challenging critical period. J Neurosci 5:255-267.Innocenti GM (1986) General organization of callosal connections in the cerebral cortex. InJones EG and Peters A (eds): Cerebral Cortex, Vol. 5. New York: Plenum Press, pp.291-355.Innocenti GM, Clarke 5, Kraftsik R (1986) Interchange of callosal and associationprojections in the developing visual cortex. J Neurosci. 6(5): 1384-1409.Jacobson 5, Trojanowski JQ (1974) The cells of origin of the corpus callosum in rat, catand rhesus monkey. Br Res 74: 149-155.Jones EG (1975) Varieties and distribution of nonpyramidal cells in the somatic sensorycortex of the squirrel monkey. J Comp Neurol 160: 205-268.160Jones EG, Burton H, Porter R (1975) Commisural and corticocortical “columns” in thesomatic sensory cortex of primates. Science 190: 572-574.Jones EG, Wise SP (1977) Size, laminar and columnar distribution of efferent cells in thesensory-motor cortex of monkeys. J Comp Neurol 175:391-438.Jones EG (1981) Anatomy of cerebral cortex: Columnar input-output organization. InSchmitt FO, Worden FG, Adelman G and Dennis SG (eds): The Organization of theCerebral Cortex. Cambridge: MIT Press, pp. 199-236.Jones EG, Hendry SHC (1984) Basket cells. In: The Cerebral Cortex, Vol. 1, Cellularcomponents of the cerebral cortex (Peters A and Jones EG, eds). Plenum Press, NewYork, pp. 309-336.Jones EG (1986) Connectivity of the primate sensory-motor cortex. In: Cerebral Cortex,vol. 5. Sensory-motor areas and aspects of cortical connectivity, eds. Jones EG andPeters A. pp. New York: Plenum. 113-184.Jones EG (1988) What are the local circuits? In Neurobiology of Neocortex, Rakic P andSinger W eds. John Wiley and Sons Limited, pp. 137-152.Katz LC (1987) Local circuitry of identified projection neurons in cat visual cortex brainslices. J Neurosci 7(4):1223-1249.Katz LC, Gilbert CD Wiesel TN (1989) Local circuits and ocular dominance columns inmonkey striate cortex. J Neurosci 9(4): 1389-1399.Katz LC, Callaway EM (1992) Development of local circuits in mammalian visual cortex.Annu Rev Neurosci 15:31-56.Keller G, Innocenti GM (1981) Callosal connections of suprasylvian visual areas in the cat.Neuroscience 6: 703-7 12.Kelly JP, Van Essen DC (1974) Cell structure and function in the visual cortex of the cat.J Physiol 238:515-547.Kisvárday ZF, Martin KAC, Somogyi P, Whitteridge D (1983) The physiology,morphology and synaptology of basket cells in the cats visual cortex. J Physiology334: 21-22P.Kisvárday ZF, Martin KAC, Freund TF, Maglocsky ZS, Whitteridge D, Somogyi P(1986) Synaptic targets of HRP-filled layer III pyramidal cells in the cat striate cortex.Exp Br Res 64:541-552.Kisvárday ZF, Martin KAC, Friedlander MJ, Somogyi P (1987) Evidence for interlaminarinhibitory circuits in the striate cortex of the cat. I Comp Neurol 260: 1-19.Kisvárday ZF, Eysel UT (1992) Cellular organization of reciprocal patchy networks inlayer III of cat visual cortex (area 17). Neurosci 46 (2):275-286.Kisvárday ZF, Eysel UT (1993) Functional and structural topography of horizontalinhibitory connections in cat visual cortex. Europ J Neurosci 5:1558-1572.161Kisvárday ZF, Beaulieu C, Eysel UT (1993) Network of GABAergic large basket cells incat visual cortex (area 18): Implications for lateral inhibition. J Comp Neurol 327:398-415.Koester SE, O’Leary DDM (1992) Functional Classes of Cortical neurons DevelopDendritic Distinctions by Class-specific Sculpting of an Early Common Pattern. J ofNeuroscience 12(4): 1382-1393.Lachica EA, Beck PD, Casagrande VA (1993) Intrinsic connections of layer III of striatecortex in squirrel monkey and bush baby: Correlations of cytochrome oxidase. J CompNeurol 329:163-187.Leporé F, Guillemot J-P (1986) The role of the corpus callosum in midline fusion. In:Leporé F, Ptito M, Jasper HH (eds) Two hemispheres - one brain: functions of thecorpus callosum. Alan R Liss, New York, pp 211-229.LeVay S (1973) Synaptic patterns in the visual cortex of the cat and monkey. Electronmicroscopy of Golgi preparations. J Comp Neurol 150:53-86.LeVay 5, Gilbert CD (1976) Laminar patterns of geniculocortical projection in the cat.Brain Res 113:1-20.LeVay 5, Stryker M, Shatz C (1978) Ocular dominance columns and their development inlayer IV of the cat’s visual cortex: A quantitative study. J Comp Neurol 178:223-244.LeVay S (1988) The patchy intrinsic projections of visual cortex. Progr Br Res 75:147-161.Leventhal AG (1982) Morphology and distribution of retinal ganglion cells projecting todifferent layers of the dorsal lateral genicuate nucleus in normal and siamese cats. JNeurosci 2(8): 1024-1042.Lindsay RD. Scheibel AB (1974) Quantitative analysis of the dendritic branching pattern ofsmall pyramidal cells from adult rat somesthetic and visual cortex. Exp Neurol 45:424-434.Livingstone MS, Hubel DH (1984a) Anatomy and physiology of a color system in theprimate visual cortex. J Neurosci 4:309-356.Livingstone MS, Hubel DH (1984b) Specificity of intrinsic connections in primate primaryvisual cortex. J Neurosci 4: 2830-2835.Lorente de No (1933) Studies on the structure of the cerebral cortex. I Psychol Neurol45:382-438.Lorente de No (1949) Cerebral cortex: Architecture, intracortical connections, motorprojections. In Physiology of the Nervous System, 2nd ed. London:Oxford UniversityPress, pp.288-330.Löwel 5, Freeman B, Singer W (1987) Topographic organization of the orientation columnsystem in large flat-mounts of the cat visual cortex: A 2-deoxyglucose study. I CompNeurol 255: 401-415.162Luhmann HJ, Millan LM, Singer W (1986) Development of horizontal intrinsicconnections in cat striate cortex. Exp Brain Res 63:443-448.Lund JS (1973) Organization of neurons in the visual cortex, area 17, of the monkey(Macaca mulatta) J Comp Neurol 147: 455-496.Lund IS, Boothe RG (1975) Interlaminar connections and pyramidal neuron organizationin the visual cortex, area 17, of the macaque monkey. J Comp Neurol 159: 305-334.Lund JS, Lund RD. Hendrickson AE, Bunt AH, Fuchs AF (1975) The origin of efferentpathways from the primary visual cortex, area 17, of the macaque monkey as shown byretrograde transport of horseradish peroxidase. J Comp Neurol 164:287-304.Lund JS, Henry GH, MacQueen CL, Harvey AR (1979) Anatomical organization of theprimary visual cortex (Area 17) of the cat. A comparison with area 17 of the Macaquemonkey. J Comp Neurol 184: 599-618.Lund RD. Mitchell DE (1979) The effects of dark-rearing on visual callosal connections ofcats. Brain Res 167:172-175.Lund JS, Hendrickson AE, Ogren MP, Tobin EA (1981) Anatomical organization ofprimate visual cortex area VII. I Comp Neurol 202:19-45.Lund JS (1984) Spiny stellate neurons. In Cerebral Cortex, Vol.I:Cellular Components ofthe Cerebral Cortex. Peters A, Jones EG, eds, pp 255-308, Plenum, New York.Lund JS (1988) Anatomical organization of macaque monkey striate visual cortex. AnnRev Neurosci 11:253-288.Lund JS, Takashi Y, Levitt JB (1993) Comparison of intrinsic Connectivity in DifferentAreas of Macaque Monkey Cerebral Cortex. Cerebral Cortex 3(2):148-162.Malach R (1992) Dendritic sampling Across Processing Streams in Monkey Striate Cortex.J Comp Neurol 315:303-312.Manning KA, Erichsen JT, Evinger C (1990) Retrograde transneuronal transport propertiesof fragment C of tetanus toxin. Neuroscience 34 (1): 251-263.Maranto AR (1982) Neuronal mapping: A photooxidization technique reaction makesLucifer Yellow useful for electron microscopy. Science 2 17:953-955.Martin KAC, Somogyi P. Whitteridge D (1983) Physiological and morphologicalproperties of identified basket cells in the cat’s visual cortex. Exp Br Res 50:193-200.Martin KAC, Whitteridge D (1 984a) Form, function and intracortical projections of spinyneurons in the striate cortex of the cat. I Physiol (Lond) 353:463-504.Martin KAC, Whitteridge D (1984b) The relationship of receptive field properties to thedendritic shape of neurons in the cat striate cortex. I Physiol (Lond) 356:291-302.Martin KAC (1984) Neuronal circuits in cat striate cortex. In Cerebral Cortex, Vol 2. ed.Jones EG, Peters A. New york: Plenum Press 241-284.163Martin, KAC (1988) From single cells to simple circuits in the cerebral cortex. Quart I ExpPhysiol 73:637-702.Matsubara JA, Cynader MS, Swindale NV, Stryker MP (1985) Intrinsic projections withinvisual cortex: Evidence for orientation specific local connections. Proc Natl Acad SciUSA 82:935-939.Matsubara JA, Cynader MS, Swindale NV (1987) Anatomical properties and physiologicalcorrelates of the intrinsic connections in cat area 18. J Neurosci 7(5): 1428-1446.Matsubara JA (1988) Local, horizontal connections within area 18 of the cat. Prog in BrRes 75:163-171.Matsubara IA, Zhang I, Boyd ID (1991) Transneuronal transport of tetanus toxin C-fragment reveals second-order local circuits in cat visual cortex. Soc N Sc abst 48.15.Matsubara JA, Boyd ID (1992) Presence of GABA-immunoreactive neurons withinintracortical patches of area 18 of the cat. Br Res 583:161-170.Matsubara JA, Chase R, Zhang I, Thejomayen DM (1993) Local patch pyramidal neuronsare distinct from corticocortical and callosal cells in cat area 18. Soc N Sc Abst 397.5.Maunsell JHR, Van Essen DC (1983) The connections of the middle temporal visual area(MT) and their relationship to a cortical hierachy in the macaque monkey. J Neurosci 3:2563-2586.Meyer G, Albus K (1981) Spiny stellates as cells of origin of association fibers from area17 to area 18 in the cat’s neocortex. Brain Res 210:335-341.Mitchell DE, Blakemore C (1970) Binocular depth perception and the corpus callosum. VisRes 10:49-54.Mountcastle V (1957) Modality and topographic properties of single neurons of cat’ssomatic sensory cortex. J Neurophysiol 20:408-434.Murphy KM, VanSluyters RC, Jones DG (1990) Cytochrome oxidase activity in cat visualcortex: is it periodic? Soc N Sci abst, 130.3.Myers RE (1962) Transmission of visual information within and between the hemispheres.A behavioural study. In Mountcastle VB (ed): “Interhemispheric relations and cerebraldominance”. Baltimore: Johns Hopkins Press, pp 5 1-73.Naegele JR, Katz LC (1990) Cell surface molecules containing N-Acetylgalactosamine areassociated with basket cells and neurogliaform cells in cat visual cortex. J Neuroscil0(2):540-557.Nance DM, Burns J (1990) Fluorescent dextrans as sensitive anterograde neuroanatomicaltracers: applications and pitfalls. Br Res Bull 25:139-145.Newsome WT, Ailman JA (1980) Interhemispheric connections of visual cortex in the owlmonkey Aotus trivirgatus and the bushbaby, Galago senegalensis. J Comp Neurol194:209-233.Nolte 1 (1988) Cerebral cortex. In The Human Brain. pp.335-365 and pp. 366-389.164O’Leary JL (1941) Structure of the area striata of the cat. J Comp Neurol 75:131-161.Otsuka R, Hassler R (1962) Uber Aufbau und Gliederung der corticalen Sehsphare bei derKatze. Arch Psych Nevenkr 203:212-234.PayneBR, Elberger AJ, Berman N, Murphy EH (1980) Binocularity in the cat visual cortexis reduced by sectioning the corpus callosum. Science 207: 1097-1099.Peters A, Feldman ML, Saldanha J (1976) The projection of the lateral geniculate nucleusto area 17 of the rat visual cortex. II. Terminations upon neuronal perikarya anddendritic shafts. J Neurocytol 5:85-107.Peters A, Fairén A (1978) Smooth and sparsely-spined stellate cells in the visual cortex ofthe rat: A study using the combined Golgi-electron microscope technique. J CompNeurol 18 1:129-172.Peters A, Kimerer LM (1981) Bipolar neurons in rat isual cortex: A combined Golgielectron microscopic study. J Neurocytol 10:921-946.Peters A, Regidor J (1981) A reassesment of the forms of nonpyramidal neurons in area17 of the cat visual cortex. I Comp Neurol 203:685-716.Peters A (1984) Bipolar cells. In: The Cerebral Cortex, Vol. 1, Cellular components of theCerebral Cortex ( Peters A and Jones EG, eds.) Plenum Press, New York, pp 267-294.Peters A, Harriman KM (1988) Enigmatic bipolar cell of rat visual cortex. J Comp Neurol267:409-432.Price DJ, Blakemore C (1985a) The postnatal development of the association projectionfrom visual cortical area 17 to area 18 in the cat. I Neurosci 5:2443-2452.Price DI, Blakemore C (1985b) Regressive events in the postnatal development ofassociation projections in the visual cortex. Nature 316:721-724.Price DJ (1986) The postnatal development of clustered intrinsic connections in area 18 ofthe visual cortex in kittens. Dev Brain Res 24:1-2.Price DJ, Ferrer JMR (1993) The incidence of bifurcation among corticocorticalconnections from area 17 in the developing visual cortex of the cat. Eur I Neurosci5:223-231.Purves D, Hume RI (1981) The relation of postsynaptic geometry to the number ofpresynaptic axons that innervate autonomic ganglion cells. I Neurosci 1(5) :441-452.Purves D, Lichtman 1W (1985) Geometrical differences among homologous neurons inmammals. Science 228:298-302.Rockland KS, Pandya DN (1979) Laminar origins and terminations of cortical connectionsto the occipital lobe in the rhesus monkey. Brain Res 179:3-20.Rockland KS, Lund 15 (1982) Widespread periodic intrinsic connections in the tree shrewvisual cortex. Science 215: 1532-1534.165Rockland KS, Lund JS (1983) Intrinsic laminar lattice connections in primate visualcortex. J Comp Neurol 216:303-3 18.Rosenquist AC (1985) Connections of visual cortical areas in the cat. In Cerebral Cortex,Vol. 3, A. Peters and EG Jones, eds, pp. 8 1-117, Plenum, New York.Schmeud L, Kyriakidis K, Heimer L (1990) In vivo anterograde and retrograde axonaltransport of the fluorescent rhodamine-dextran-amine, Fluoro-Ruby, within the CNS.Brain Res 526: 127-134.Segraves MA, Rosenquist AC (1982) The distribution of the cells of origin of callosalprojections in cat visual cortex. J Neurosci 2:1079-108.Segraves MA, Rosenquist AC (1982b) Afferent and Efferent callosal connections ofretinotopically defined areas in cat cortex. J Neurosci 2(8); 1090-1107.Segraves MA, Innocenti G (1985) Comparison of the distributions of ipsilaterally andcontralaterally projecting corticocortical neurons in cat visual cortex using twofluorescent tracers. J Neurosci 5(8): 2107-2118.Sesma MA, Casagrande VA, Kaas JH (1984) Cortical connections of area 17 in treeshrews. J Comp Neurol 230:337-35 1.Shatz C (1977) Anatomy of interhemispheric connections in the visual system of Bostonsiamese and ordinary cats. J Comp Neurol 173:497-5 18.Shatz C, Lindstrom S, Wiesel TN (1977) The distribution of afferents representing theright and left eyes in the cat’s visual cortex. Brain Res 131:103-116.Shatz CJ, Stryker MP (1978) Ocular dominance in layer IV of the cat’s visual cortex andthe effects of monocular deprivation. J Physiol (London) 281:267-283.Shipp S, Grant 5 (1991) Organization of reciprocal connections between area 17 and thelateral suprasylvian area of cat visual cortex. Visual Neurosci 6: 339-355.Sherman SM (1985) Functional organization of the W-, X-, and Y-cell pathways in thecat:A review and hypothesis. In: Progress in Psychobiology and PhysiologicalPsychology (Sprague JM, Epstein AN ed), New York:Academic Press, pp 233-314.Sholl DA (1953) Dendritic organization in the visual and motor cortices of the cat. J Anat87:387-407.Sholl DA (1955) The organization of the visual cortex in the cat. J Anat (Lond) 89: 34-46.Shotton DM (1989) Confocal scanning optical microscopy and its applications forbiological specimens. J Cell Science 94, 175-206.Simons DJ, Woolsey TA (1984) Morphology of Golgi-Cox-impregnated barrel neurons inrat SmI cortex. J Comp Neurol 230:119-132.Smit GJ, Uylings HBM (1975) The morphometry of the branching pattern in dendrites ofthe visual cortex pyramidal cells. Brain Res 87:41-53.166Somogyi P (1977) A specific “axo-axonal” intemeuron in the visual cortex of the rat. BrainRes 136:345-350.Somogyi P (1978) The study of Golgi stained cells and of experimental degeneration underthe electron microscope: A direct method for the identification in the visual cortex ofthree successive links in a neuron chain. Neuroscience 3:167-180.Somogyi P. Hodgson Al, Smith AD (1979) An approach to tracing neuron networks in thecerebral cortex and basal ganglia: combination of Golgi staining, retrograde transport ofhorseradish peroxidase and anterograde degeneration of synaptic boutons in the samematerial. Neuroscience 4: 1805-1852.Somogyi P, Cowey A (1981) Combined Golgi and Electron microscopic study on thesynapses formed by double bouquet cells in the visual cortex of the cat and monkey. JComp Neurol 195:547-566.Somogyi P, Cowey A, Halasz N, Freund TF (1981) Vertical organization of neuronsaccumulating 3H-GABA in the visual cortex of the rhesus monkey. Neuroscience7:2577-2609.Somogyi P. Freund TF, Cowey A (1982) The axo-axonic interneuron in the cerebral cortexof the rat, cat and monkey. Neuroscience 7:2577-2609.Somogyi P, Kisvárday ZF, Martin KAC, Whitteridge D (1983) Synaptic connections ofmorphologically identified and physiologically characterized large basket cells in thestriate cortex of cat. Neuroscience 10:26 1-294.Somogyi P. Freund TF, Hodgson Al, Somogyi J, Beroukas D, Chubb 1W (1985)Identified axo-axonic cells are immunoreactive for GABA in the hippocampus andvisual cortex of the cat. Brain Res 332: 143-149.Somogyi P. Soltész 1(1986) Immunogold demonstration of GABA in synaptic terminals ofintracellularly recorded, horseradish peroxidase-fihled basket cells and clutch cells in thecat’s visual cortex. Neuroscience 19: 1051-1065.Sperry RW (1961) Cerebral organization and behaviour. Science 133:1749-1757.Steffen H, Van Der Loos H (1980) Early lesions of mouse vibrissal follicles: Theirinfluence on dendritic orientation in the developing barrelfield. Exp Brain Res 40:4 10-431.Stewart WW (1981) Lucifer dyes - highly fluorescent dyes for biological tracing. Nature292: 17-21.Symonds LL, Rosenquist AC (1984) Laminar origins of visual corticocortical connectionsin the cat. J Comp Neurol 229:39-47.Szentagothai J (1975) The “module-concept” in cerebral cortex architecture. Brain Res95:475-496.Tauchi M, Masland RH (1984) The shape and arrangement of the cholinergic neurons inthe rabbit retina. Proc R Soc Lond B 223: 101-119.167Tauchi M, Masland RH (1985) Local order among the dendrites of an amacrine cellpopulation. J Neurosci 5: 2494-2501.Thejomayen DM, Zhang J, Matsubara JA (1992) Local patch neurons and callosal neuronscomprise seperate, morphologically distinct populations in the upper layers of cat area18. Soc N Sc Abst 132.1.Thejomayen DM, Matsubara JA (1993) Confocal Microscopic Study of the DendriticOrganization of Patchy, Intrinsic Neurons in Area 18 of the Cat. Cerebral Cortex3(5):442-453.Tieman SB, Hirsch HVB (1982) Exposure to lines of only one orientation modifiesdendritic morphology of cells in the visual cortex of the cat. J Comp Neurol 211:353-362.Tigges J, Tigges M, Anschel S, Cross NA, Letbetter WD, McBride RL (1981) Areal andlaminar distribution of neurons interconnecting the central visual cortical areas 17, 18,19 and MT in squirrel monkey (Saimiri). J Comp Neurol 202:539-560.Valverde F (1968) Structural changes in the area striata of the mouse after enucleation. ExpBrain Res 5:274-292.Van der Loos H (1965) The ‘improperly oriented pyramidal cell in the cerebral cortex andits possible bearing on problems of neuronal growth and cell orientation. Bull JohnsHopkins Hosp 117:228-250.Van Essen DC, Newsome NT, Bixhy JL (1982) The pattern of interhemisphericconnections and its relationship to extrastriate visual areas in the macaque monkey. JNeurosci 2:265-283.Van Essen DC, Newsome NT, Maunsell JHR, Bixby JL (1986) The projections fromstriate coretex (Vi) to areas V2 and V3 in the macaque monkey: asymmetries, arealboundaries, and patchy connections. J Comp Neurol 244:451-480.Voigt T, LeVay 5, Stamnes MA (1988) Morphological and immunocytochemicalobservations on the visual callosal projections in the cat. J Comp Neurol 272:450-460.Wallén P, Carlsson K, Liljeborg A, Grillner S (1988) Three-dimensional reconstuction ofneurons in the lamprey spinal cord in whole-mount, using a confocal laser scanningmicroscope. I Neurosci Meth 24: 9 1-100.Welker WI (1976) Mapping the brain. Historical trends in functional localization. BrainBehav Evol 13:327-343.Wiesel TN, Hubel DH (1963) Single-cell responses in striate cortex of kittens deprived ofvision in one eye. J Neurophysiol 26: 1003-1017.Wiesel TN, Hubel DH (1965) Comparison of the effects of unilateral and bilateral eyeclosure on cortical unit responses in kittens. I Neurophysiol 28:1029-1040.Wong-Riley M (1979) Changes in the visual system of monocularly sutured or enucleatedcats demonstrable with cytochrome oxidase histochemistry. Brain Res 171: 11-28.168Woods JW, Radewan CW (1977) Kalman filtering in two-dimensions. IEEE Trans IT23:473-482.Woolsey CN (1964) Cortical localization as defined by evoked potential and electricalstimulation studies. In: Cerebral Localization and Organization (Schaltenbrand G andWoolsey CN, eds.) Univ of Wisconsin Press, Madison. pp. 17-26.Woolsey TA, Van der Loos H (1970) The structural organization of layer IV in thesomatosensory (SI) region of mouse cerebral cortex. Brain Res 17:205-242.Wouterlood FG, Jorritsma-Byham, Goede PH (1990) Combination of anterograde tracingwith Phaseolus vulgaris-leucoagglutinin, retrograde fluorescent tracing and fixed-sliceintracellular injection of Lucifer yellow. J Neurosci Meth 33: 207-217.Yoshioka T, Blasdel GG, Levitt JB, Lund JS (1992) Patterns of lateral connections inmacaque visual area Vi revealed by biocytin histochemistry and functional imaging.Soc N Sc abstract 134.13.169


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



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"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items