Direct Temporal-Occipital Feedback Connections to Striate Cortex (V1) in the Macaque Monkey

Although there have been reports of sparse projections fromtemporal areas TE, TF, and even TH to area V1, it is generallybelieved that cortical afferents to V1 originate exclusivelyfrom prestriate areas. Injections of anterograde tracers inanterior occipital and temporal areas, however, consistentlyproduce labeled terminals in area V1. In order to confirm theseresults and display the full range of foci projecting to V1,we injected V1 in two monkeys with the retrograde tracer fastblue. Feedback connections were found, as expected, from severalprestriate areas (V2, V3, V4, and MT). These originate fromneurons in layers 3A and 6. Connections were also found fromseveral more distal regions, namely, areas TEO, TE, TF, TH,and from cortex in the occipitotemporal and superior temporal(STS) sulci. Filled neurons occurred in two small foci in thecaudal intraparietal sulcus. These more distal feedback connectionstend to originate only from layer 6. An additional injectionof the retrograde tracer diamidino yellow in area V2 of oneanimal revealed a similarly widespread network of feedback connections.In some areas (In the STS and in TEO), 10–15% of fluorescentneurons were double-labeled. These results indicate that feedback connections to early visualcortex derive from a widespread network of areas, includinglimbic-associated cortices. These connectional patterns testifyto the massive recursiveness of anatomical pathways. As thereare no reports of projections from V1 to anterior temporal cortices,our results also indicate that some cortical feedback connectionsmay not be strictly reciprocal.     

Areas PMLS and 21 a of Cat Visual Cortex: Two Functionally Distinct Areas

We have compared the receptive field properties of neurons recordedfrom visuotopically corresponding regions of area 21a and theposteromedial lateral suprasylvian area (PMLS) of cat visualcortex. In both areas, the great majority of neurons were orientation-selectiveand binocular, and their responses to moving contours were modulatedby simultaneous in-phase or anti-phase motion of large texturedbackground stimuli (‘visual noise’). However, despitethe great hodological similarity between the two areas, PMLSneurons had on average significantly higher peak discharge rates,exhibited substantially greater direction selectivity indices,and preferred substantially higher stimulus velocities thanarea 21a neurons. Furthermore, the majority of binocular neuronsin the PMLS area and in area 21a were dominated respectivelyby the contralateral and the ipsilateral eyes. Finally, while46% of PMLS neurons were excited by movement of visual noiseper Se. only 25% of area 21 a neurons could be excited by suchstimuli. We argue that the PMLS area, like its presumed primatehomologue the middle-temporal (MT) area, is mainly involvedin motion analysis. By contrast, area 21a appears to be involvedin pattern analysis rather than motion analysis. It is likelythat phylogenetically area 21a derives from the PMLS area.     

Connections of Inferior Temporal Areas TEO and TE with Parietal and Frontal Cortex in Macaque Monkeys

Inferior temporal cortex is perhaps the highest visual processingarea and much anatomical work has focused on its connectionswith other visual areas in temporal and occipital cortex. Herewe report connections of inferior temporal cortex with regionsin the frontal and parietal lobes. Inferior temporal areas TEOand TE were injected with WGA-HRP and ³H-AA, respectively, orvice versa, in 1-week-old infant and 3–4–year-oldadult monkeys (Macaca mulatta). The results indicated that whereasTEO has more extensive connections with parietal areas, TE hasmore extensive connections with prefrontal areas. Thus, in theintraparietal sulcus, area TEO is connected with areas LIPd,LIPv, and V3A, and with the as yet undefined region betweenLIPv and V3A, whereas the connections of TE are predominantlywith LIPd, and to a lesser extent with LIPv. In the prefrontalcortex, area TE is connected with areas 8 and 45 in the inferiorlimb of the anterior bank of the arcuate sulcus, with area 12on the inferior prefrontal convexity, and with areas 11 and13 on the orbital surface. By contrast, the connections of areaTEO are limited to areas 8, 45, and 12. Furthermore, withinprefrontal cortex, the projections from areas TEO and TE terminatein different layers in areas 8 and 45, such that those fromTEO terminate in all layers, whereas those from TE terminatein layers I and V/VI only. In contrast to the connections ofareas TEO and TE with various medical temporal-lobe and subcorticalstructures, which are immature in infant monkeys (Webster etal., 1991, 1993b), the connections with parietal and prefrontalareas appear adult-like as early as 1 week of age.     

Development of Horizontal Projections in Layer 2/3 of Ferret Visual Cortex

Pyramidal cells in layer 2/3 of cat striate cortex extend longaxons that form clustered projections linking iso-orientationcolumns. Using extracellular biocytin injections in brain slices,the formation of these projections was examined in the ferretto determine whether horizontal projections exhibit similarpatterns of development in the ferret and the cat, and to relatethe time course of horizontal projection formation to the onsetof patterned visual experience and orientation selectivity.Soon after the first appearance of axon collaterals in layer2/3. around postnatal day 22 (P22), pyramidal cell axons wereuniformly distributed and unbranched for up to 1 mm from thecell body. By P26, axons began to form secondary branches 1–2mm from the cell body, with little evidence for distinct clusters.The first indication of selective elaboration of secondary branchesand retraction of unbranched collaterals occurred around P28.By P34. patchy regions of axon branches emerged, though unbranchedcollaterals were still present, followed by distinct, adult-likeclusters by P45. Although the general pattern of horizontalprojection formation closely resembles that seen in the cat(Callaway and Katz, 1990), the ferret circuitry matures earlierthan that of the cat relative to the time of eye opening. Sinceeye opening in ferrets occurs between P30 and P32, this systemof orientation-specific patches begins to develop in the absenceof patterned visual input and when most cortical cells are notyet orientation selective, suggesting a prominent role for spontaneousactivity in initiating cluster formation. The refinement ofclustered connections, however, does occur synchronously withthe maturation of orientation-selective responses (Chapman andStryker, 1993).     

Specific Patterns of Intrinsic Connections between Representation Zones in the Rat Motor Cortex

The organization of intrinsic connections in rat motor cortexwas studied by combining microstimulation and tract-tracingtechniques. Maps of forelimb and vibrissal movements were constructedfrom the distribution of cortical sites from which movementswere evoked in response to intracortical microstimulation. Then,a single injection of a fluorescent dextran was placed intoeither a vibrissal or a wrist representation zone, or into aregion bordering these zones, resulting in anterograde labelingof long intrinsic, horizontal axons. Following injection intothe vibrissal area, axons were largely restricted to the whiskerrepresentation zone and to the border region with the forelimbrepresentation. Injections into a wrist zone labeled projectionslargely restricted to the forelimb area and to the border withthe vibrissal area. Injections into a border region labeleddense projections throughout most of the forelimb and vibrissalareas. These findings indicate that intrinsic axon collateralsin the motor cortex form specific and extensive connectionsamong representation zones related to movements of the samebody part. These connections may be involved in the coordinationof activity in different representation zones for the executionof complex movement patterns. The projection of axon collateralsinto border regions may be the anatomical substrate for therapid reorganization of motor cortical maps that occurs followingvarious experimental manipulations.     

Object and Spatial Visual Working Memory Activate Separate Neural Systems in Human Cortex

Human and nonhuman primate visual systems are divided into objectand spatial information processing pathways. In the macaque,it has been shown that these pathways project to separate areasin the frontal lobe and that the ventral and dorsal frontalareas are, respectively, involved in working memory for objectsand spatial locations. A positron emission tomography (PET)study was done to determine if a similar anatomical segregationexists in humans for object and spatial visual working memory.Face working memory demonstrated significant increases in regionalcerebral blood flow (rCBF), relative to location working memory,in fusiform, parahippocampal, inferior frontal, and anteriorcingulate cortices, and in right thalamus and midline cerebellum.Location working memory demonstrated significant increases inrCBF, relative to face working memory, in superior and inferiorparietal cortex, and in the superior frontal sulcus. Our resultsshow that the neural systems involved in working memory forfaces and for spatial location are functionally segregated,with different areas recruited in both extrastriate and frontalcortices for processing the two types of visual information.     

Prenatal Development of NADPH-diaphorase-Reactive Neurons in Human Frontal Cortex

Nicotinamide adenine dinucleotide phosphate-diaphorase (NADPH-d)histochemistry was used to study the morphology and developmentof neurons that metabolize nitric oxide (NO) in the frontalcortex of human fetuses aged from 13 weeks of gestation (13W)to term, to investigate whether the two distinct types of NOneuron described in the adult develop differently. Large, heavilystained, sparsely spiny, non-pyramidal neurons (Type I) developby 15W mainly in the subplate (SP) of the cortical Anlage. Theyachieve an adult-like pattern by 32W, distributed thoughoutthe cortex and subcortical white matter, but with the highestconcentration in the white matter. Small, lightly stained cells(Type II) develop later (32W) thoughout the cortex, but especiallyin layers II-IV. and increase in number to term. NADPH-d-positivedendrites and axons appear in the cortex and white matter by15W. They include thick, radially oriented, dendritic processesfrom Type I neurons in SP and CP Their arbors expand and maturebetween 17 and 28W. Fine horizontal axons are visible in layerI by 17W. Others develop in layers II-IV from 28W, and havereached a high degree of development by term. NADPH-d-positiveaxons in the cortex seem to have both intrinsic and extrinsicorigins. Thus the two types of NADPH-d neurons found in adultprimate, including human, cortex are reflected by differentdevelopmental forms prenatally. It is concluded that NO-metabolizingneurons in the human cortex may be involved in various aspectsof development, including morphological and functional maturation,and that the late-developing Type II neurons may represent acell line specific to primates, perhaps related to the developmentof their higher cortical activity and of potential importancein the pathophysiology of diseases of cognitive function.     

The Distribution of Callosal Connections Correlates with the Pattern of Cytochrome Oxidase Stripes in Visual Area V2 of Macaque Monkeys

In visual area V2 of monkeys, cytochrome oxidase (CD) histochemistryreveals a system of stripe-like subregions where densely labeledthick and thin stripes and pale interstripes can be recognized.Several lines of evidence suggest that CO stripe-like subregionsare associated with functional streams in the visual cortex.In the present study, the distribution of retrogradely labeledcallosal cells in V2 and the pattern of CO staining were correlatedusing tangential sections through the flattened cortex. Spectraland coherency analyses of the callosal and CO patterns wereperformed to assess quantitatively the degree of spatial correlationbetween these two patterns. The results showed that labeledcallosal cells accumulated along the V1/V2 border and in finger-likebands that protruded up to 7–8 mm into V2. These callosalbands were in register with thick and thin CO stripes, withrelatively few labeled callosal cells found in interstripe regions.This finding supports the notion that the distribution of callosalconnections in the visual cortex is dictated not only by thetopography of visual areas, but also by the arrangement of corticalfunctional streams. Further, these results extend to interhemisphericpathways the notion of functional specificity currently associatedmainly with some visual intrahemispheric pathways.     

Functional Segregation of Color and Motion Processing in the Human Visual Cortex: Clinical Evidence

Anatomical and physiological investigations indicate two majordistinct functional streams within the extrastriate visual cortexof the macaque monkey, and behavioral observations suggest thatthe ventral (occipitotemporal) pathway is the cornerstone forobject recognition whereas the dorsal (occipitoparietal) pathwayis primarily involved in visuospatial perception and visuomotorperformance. In the context of this dichotomy we conducted apsychophysical and neuropsychological study of visual perceptualabilities in two stroke patients, each with lesions involvingseveral extrastriate areas. Magnetic resonance imaging demonstratedbilateral lesions; in one patient (E.W.) the lesion involvesthe ventral medial portions of the occipital and temporal lobes,and in the other (A.F.) the lesion involves dorsally the occipital-parietalarea, including the region of the temporal-parietal-occipitaljunction. E.W. suffers from achromatopsia of central origin,prosopagnosia, visual agnosia, and alexia without agraphia.His depth and motion perception, including recognition of movingobjects, are normal. He has superior visual field loss bilaterally,and slightly impaired acuity, and complains that the world appearsin a deep twilight even on a sunny day. In contrast, A.F. showsspecific deficits of stereopsis, spatial localization, and severalaspects of motion perception. He is also impaired at recognizingobjects presented from unconventional views, but recognitionof prototypical views of objects, and color and form discriminationare normal, as is his ability to recognize faces. The anatomical characteristics of the lesions of these two patientspermit a direct experimental comparison of the effects of lesionsconfined to the parietal or temporal pathways. E.W.'s and A.F.'sperformance on the psychophysical and neuropsychological tasksdiscussed here supports the functional distinction between adorsal and a ventral extrastriate system but additionally suggeststhe existence of a pathway involved in identification-from-motionthat is separate from both the dorsal early motion/spatial analysispathway and the ventral color/static-form pathway.     

Human Extrastriate Visual Cortex and the Perception of Faces, Words, Numbers, and Colors

Electrophysiological correlates of the processing of visualinformation were studied in epileptic patients with electrodeschronically implanted on the surface of striate and extrastriatecortex. In separate experiments patients viewed faces, letterstrings (words and non-words), numbers, and control stimuli.A negative potential, N200, was evoked by faces, letter strings,and numbers, but not by the control stimuli. N200 was recordedbilaterally from discrete regions of the fusiform and inferiortemporal gyri. These category-specific face, letter-string,and number "modules" vary in location. In most cases there wasno overlap in the location of face and letter-string modules,suggesting a mosaic of functionally discrete regions. In somecases letter-string and number N200s were recorded from thesame location, suggesting that these modules may be less spatiallyand functionally discrete. Face N200-like potentials can berecorded from temporal scalp, allowing the possibility of studyingearly face processing in normal subjects. Longer-latency face-specificpotentials were recorded from the inferior surface of the anteriortemporal lobe. Potentials evoked by colored checkerboards wererecorded from a region of the fusiform gyrus posterior to thefusiform region from which category-specific N200s were recorded. These results suggest that there are several processing streamsin inferior extrastriate cortex. In addition to object recognitionsystems previously proposed for faces and words, our preliminaryresults suggest a separate system dealing with numbers. Postulatedsystems dealing with larger manipulable objects and animalshave not been detected.     

Prenatal Development of Parvalbumin Immunoreactivity in the Human Striate Cortex

In human primary visual cortex, parvalbumin (PV) is expressedby Cajal-Retzius cells in layer I by 20 weeks of gestation (20W),but its immunoreactivity is mostly lost by term. PV immunoreactivityin layers II–VI mainly develops later, from 26 to 34W,following an approximately ‘inside-outside’ sequencein a series of bands. PV-positive perikarya appear in layerV by 20W, but only in small numbers. They increase in numberand staining intensity by 26W. By 30W a band of densely labelledsomata and neuropil occupies layers IVC–VI. By 34W a second,less dense, band of cell bodies and neuropil appears in IVBand IVC     

Identification of Proteins Downregulated during the Postnatal Development of the Cat Visual Cortex

To identify proteins that play a role in the development ofthe mammalian visual cortex, we have used an immunosuppressionand rapid immunization strategy to generate monoclonal antibodiesto antigens that are present in area 17 of the cat during thepeak of cortical plasticity but are downregulated near the endof the plastic period. We report here the immunohistochemicaland immunobiochemical characterization of six monoclonal antibodiesthat identify antigens preferentially expressed in the cat visualcortex at 5 weeks of age. Monoclonal antibodies Cat-305 and Cat-306 detect three immunoreactiveelements that are not present at birth but are present at 5weeks. The majority of immunoreactivity is associated with apopulation of cells in the white matter that are absent at 15weeks of age. At both 5 and 15 weeks, a very small number ofneurons show intense immunoreactivity throughout all processes,resembling that achieved with a Golgi stain. In addition, adiffuse band of immunoreactivity in layer IV is largely restrictedto cortical areas 17 and 18. Cat-307 recognizes a 150 kDa soluble protein present in smallcytoplasmic inclusions. These cytoplasmic "dots" are presentin all layers, but are most prominent in layer V. Cat-307 immunoreactivityis present at birth and is completely downregulated by 15 weeks.Cat-104 and Cat-105 recognize a 200 kDa insoluble protein presentat birth and at 5 weeks, but markedly downregulated by 15 weeks.At birth, the white matter, subplate, and layer I are most denselylabeled, while at 5 weeks labeling is densest in layers II,III, and V. Cat-402 recognizes a number of high-molecular-weightantigens that are differentially expressed at 5 and 15 weeksof age. Stained non-neuronal cells that resemble protoplasmicastrocytes are present in all layers at both 5 and 15 weeks.At 5 weeks, but not at birth or 15 weeks, darkly immunoreactiveradial processes are observed that run through the full depthof the cortex. We show here that immunoreactivity for several different monoclonalantibodies is detected selectively during the period of maximaldevelopmental plasticity. The results demonstrate that the catvisual cortex at 5 weeks of age is molecularly distinct fromthe cortex at 15 weeks.     

Relationship between Orientation Domains, Cytochrome Oxidase Stripes, and Intrinsic Horizontal Connections in Squirrel Monkey Area V2

Area V2, the main target of primary visual cortex projections,is characterized by a striking functional and connectional compartmentalization.Many aspects of this organization are correlated to three setsof stripes (thick, thin, and pale) revealed by cytochrome oxidase(CO) staining. Several questions related to the physiologicalproperties of these compartments, their intrinsic connections,and points of similarity with area V1 modules are still unresolved.We have addressed some of these questions by combining the techniquesof optical imaging of intrinsic signals, tract tracing, andCO histochemistry in the same patches of areas V1 and V2 ofthe squirrel monkey. The following observations were made. Orientation domains: inarea V1 these are organized in narrow bands, while in area V2they form patches. In area V2, domain width and distance betweendomains are approximately double that found in area V1. Orientationand CO stripe organization: orientation tuning was organizedso that highly selective regions were centered on thick CO stripeswhile regions of broad orientation selectivity were centeredon thin CO stripes. However, the orientation domains appearedto ignore borders between thick and pale stripes. Intrinsicconnections: injections of the sensitive tracer biocytin intoarea V2 labeled a dense network of horizontally projecting fibersthat were organized in columnar patches. Patches were small(mean width, 211 µm; mean length, 342 µm) and thelabeling pattern extended over 4–5 mm. Axonal patchesand CO stripes: Axonal patches found were in all three stripecompartments. However, injections that straddled the bordersof thick/pale stripe compartments produced axonal projectionsthat tended to cluster around border regions. Axonal patchesand orientation domains: V2 injections produced labeling inV1 that appeared to be organized in narrow bands, reminiscentof orientation domain distribution in V1. Within area V2, axonalpatches targeted a wide range of orientation domains, but appearedto avoid domains having orthogonal orientation preference tothat found at the injection site. To conclude, our results show,on the one hand, a measure of functional specificity for theCO stripes and the intrinsic connections. On the other hand,they indicate additional substructures within area V2, whoseprecise relationship to the known compartmental organizationremains to be clarified.     

Numerical Relationships between Geniculocortical Afferents and Pyramidal Cell Modules in Cat Primary Visual Cortex

An analysis has been made of the quantitative data availableon the number of pyramidal cell modules of layer IV neurons,and of geniculocortical axons and their synapses in cat striatecortex. It is found that the convergence of geniculocorticalafferents upon any one pyramidal cell module is enormous, sincein any one location there is overlap between 360–540 X-axonsand 300–540 Y-axons. In total, the X- and Y-axonal arborsprovide some 1640 x 10⁶ synapses to area 17, which is equivalentto a ratio of 160–200 synapses per layer IV neuron. Thesevalues assume that geniculocortical terminals synapse only withthe spiny stellate cells of layer IV. The values are reducedto 100–125 per spiny stellate cell when account is takenof the synapses that involve the dendrites that enter layerIV from neurons with cell bodies in other layers. Since eachlayer IV neuron receives some 2500 asymmetric synapses, thismeans that only 5% of the total excitatory input to a layerIV neuron seems to be provided by the geniculocortical afferents.Further, if the boutons in the geniculocortical axonal arborsare distributed homogeneously across layer IV, each axon couldonly provide one synapse to about one in four of the layer IVneurons encompassed by its plexus. It may be, however, thatinstead of being spread evenly, boutons in individual arborsconverge upon individual neurons to supply a number of synapsesto them. But even so, it seems unlikely that any individualgeniculate axon could dominate the activity of a particularcortical neuron.     

Circuitry, Architecture, and Functional Dynamics of Visual Cortex

A fundamental understanding of the mechanisms of cortical processingrequires an examination of the relationships of cortical circuitry,functional architecture, and receptive field properties. Ultimately,this kind of analysis can be utilized to explore the neurobiologicalbasis of psychophysics and perception. At the outset our studieswere intended to account for the then known receptive fieldproperties of cortical cells in terms of their underlying circuitry,but surprisingly a good part of the cortical circuit appearedto he in violation of the principles of cortical architecture,and this led us to explore the possibility of new, more complexproperties of cortical cells. It has become increasingly possibleto relate the responsive specificity of cortical cells, andthe circuitry underlying this specificity, to the perceptualcapabilities of the visual system by performing analogone experimentson single cells and in human psychophysics.     

Optical imaging of contrast response in Macaque monkey V1 and V2

Our studies on brightness information processing in Macaque monkey visual cortex suggest that the thin stripes in the secondary visual area (V2) are preferentially activated by brightness stimuli (such as full field luminance modulation and illusory edge-induced brightness modulation). To further examine this possibility, we used intrinsic signal optical imaging to examine contrast response of different functional domains in primary and secondary visual areas (V1 and V2). Color and orientation stimuli were used to map functional domains in V1 (color domains, orientation domains) and V2 (thin stripes, thick/pale stripes). To examine contrast response, sinusoidal gratings at different contrasts and spatial frequencies were presented. We find that, consistent with previous studies, the optical signal increased systematically with contrast level. Unlike single-unit responses, optical signals for both color domains and orientation domains in V1 exhibit linear contrast response functions, thereby providing a large dynamic range for V1 contrast response. In contrast to domains in V1, domains in V2 exhibit nonlinear responses, characterized by high gain at low contrasts, saturating at a mid-high contrast levels. At high contrasts, thin stripes exhibit increasing response, whereas thick/pale stripes saturate, consistent with a strong parvocellular input to thin stripes. These findings suggest that, with respect to contrast encoding, thin stripes have a larger dynamic range than thick/pale stripes and further support a role for thin stripes in processing of brightness information.     

Mechanisms of Sensitivity Loss due to Visual Cortex Lesions in Humans and Macaques

This study represents the first use of noise masking and signal detection theory to examine mechanisms of visual loss after lesions of visual cortex. Noise-masked contrast thresholds were increased in 2 macaques and 2 humans at lesion-affected, compared with control, regions of their visual fields. Experiments suggested by the organization of visual cortex examined possible mechanisms of the visual loss. Two experiments tested the hypothesis that damage to feedback connections might eliminate the benefit of comparing test stimuli with remembered representations but neither could account for the sensitivity loss. The third experiment found that extrastriate lesions did increase the trial-to-trial variability of sensory decisions, suggesting this as one mechanism of sensitivity loss. In addition to clarifying mechanisms of lesion-induced contrast sensitivity loss, this study also showed that elevated contrast thresholds, that are subtle in the absence of external noise, became dramatic when measured with masking noise.