Topographical Aspects of Intracortical Excitation and Inhibition Contributing to Orientation Specificity in Area 17 of the Cat Visual Cortex

Demonstrations of Spatiotemporal Integration and what they Tell us about the Visual System

Five sets of displays are presented on the journal website to be viewed in conjunction with the text. We concentrate on the factors that give rise to the integration and disruption of the direction of apparent motion in two-dimensional and three-dimensional space. In the first set of displays we examine what factors contribute to the integration and disruption of apparent motion in the Ramachandran/Anstis clustered bistable quartets. In the second set we examine what factors give rise to the perception of the direction of motion in rotating two-dimensional wheels and dots. In the third and fourth sets we examine how the depth cues of shading and disparity contribute to the perception of apparent motion of opaque displays, and to the perception of rotating unoccluded displays, respectively. In the fifth set we examine how the depth cue of motion parallax influences the perception of apparent motion. Throughout, we make inferences about the roles which various parallel pathways and cortical areas play in the perceptions produced by the displays shown.

Origins of feature selectivities and maps in the mammalian primary visual cortex

A common feature of the mammalian striate cortex is the arrangement of 'orientation domains' containing neurons preferring similar stimulus orientations. They are arranged as spokes of a pinwheel that converge at singularities known as 'pinwheel centers'. We propose that a cortical network of feedforward and intracortical lateral connections elaborates a full set of optimum orientations from geniculate inputs that show a bias to stimulus orientation and form a set of two or a small number of 'Cartesian' coordinates. Because each geniculate afferent carries signals only from one eye and its receptive field (RF) is either ON or OFF center, the network constructs also ocular dominance columns and a quasi-segregation of ON and OFF responses across the horizontal extent of the striate cortex.

Inhibition, spike threshold, and stimulus selectivity in primary visual cortex

Ever since Hubel and Wiesel described orientation selectivity in the visual cortex, the question of how precise selectivity emerges has been marked by considerable debate. There are essentially two views of how selectivity arises. Feed-forward models rely entirely on the organization of thalamocortical inputs. Feedback models rely on lateral inhibition to refine selectivity relative to a weak bias provided by thalamocortical inputs. The debate is driven by two divergent lines of evidence. On the one hand, many response properties appear to require lateral inhibition, including precise orientation and direction selectivity and crossorientation suppression. On the other hand, intracellular recordings have failed to find consistent evidence for lateral inhibition. Here we demonstrate a resolution to this paradox. Feed-forward models incorporating the intrinsic nonlinear properties of cortical neurons and feed-forward circuits (i.e., spike threshold, contrast saturation, and spike-rate rectification) can account for properties that have previously appeared to require lateral inhibition.

Cytophysiology of spiny stellate cells in the striate cortex and their role in the excitatory mechanisms of intracortical synaptic circulation

Spiny stellate cells (SS) are an exclusive category of cortical interneurons and are the major component of intracolumnar, lateral, and callosal excitation of pyramidal and non-pyramidal neurons in the neocortex. The axons of SS make contact with the apical dendrites of pyramidal neurons, forming cartridge-type en passant synapses with them. SS establish recurrent connections with inhibitory interneurons and other SS, and also form autapses contacts. SS subtypes were identified and descriptions of their structure, neurochemical specialization, and spatial organization in the human and animal neocortex are provided. The results of our own studies, along with published data, are used to form a critical analysis of current concepts of the histophysiology of recurrent excitatory and inhibitory neurocirculation in cortical modules.

Method of unconfounding orientation and direction tunings in neuronal response to moving bars and gratings

When an oriented bar or grating is drifted across the receptive field of a cortical neuron at various orientations, the tuning function reflects both, and thus confounds the orientation (ORI) and the direction-of-motion (DIR) selectivity of the cell. Since ORI (or DIR), by definition, has a period of 180(or 360) deg/cycle, a popular method for separating these two components, due to Wörgötter and Eysel [Biol. Cybern. 57, 349 (1987)], is to Fourier decompose the neuron's response along the angular direction and then identify the first and the second harmonic with DIR and ORI, respectively (the SDO method). Zhang [Biol. Cybern. 63, 135 (1990)] pointed out that this interpretation is misconceived--all odd harmonics (not just the first harmonic) reflect the DIR component, whereas all even harmonics (including the second harmonic) contain contributions from both DIR and ORI. Here, a simplified procedure is proposed to accomplish the goal of unconfounding ORI and DIR. We first construct the sum of all odd harmonics of the overall tuning curve, denoted ODDSUM, by calculating the difference in the neuronal response to opposite drifting directions. Then we construct ODDSUM+/ODDSUM/ and identify it with DIR (here . denotes the absolute value). Subtracting DIR, that is ODDSUM+ /ODDSUM/, from the overall tuning curve gives ORI. Our method ensures that (i) the reconstructed DIR contains only one, positive peak at the preferred direction and can have power in all harmonics, and (ii) the reconstructed ORI has two peaks separated by 180 degrees and has zero power for all odd harmonics. Using this procedure, we have unconfounded orientation and direction components for a considerable sample of macaque striate cortical cells, and compared the results with those obtained using Wörgötter and Eysel's SDO method. We found that whereas the estimate of the peak angle of ORI remains largely unaffected, Wörgötter and Eysel's method considerably overestimated the relative strength of ORI. To conclude, a simple method is provided for appropriately separating the orientation and directional tuning in a neuron's response that is confounded as a result of the use of drifting oriented stimuli.

Topographic plasticity in primary visual cortex is mediated by local corticocortical connections

The placement of monocular laser lesions in the adult cat retina produces a lesion projection zone (LPZ) in primary visual cortex (V1) in which the majority of neurons have a normally located receptive field (RF) for stimulation of the intact eye and an ectopically located RF (displaced to intact retina at the edge of the lesion) for stimulation of the lesioned eye. Animals that had such lesions for 14-85 d were studied under halothane and nitrous oxide anesthesia with conventional neurophysiological recording techniques and stimulation of moving light bars. Previous work suggested that a candidate source of input, which could account for the development of the ectopic RFs, was long-range horizontal connections within V1. The critical contribution of such input was examined by placing a pipette containing the neurotoxin kainic acid at a site in the normal V1 visual representation that overlapped with the ectopic RF recorded at a site within the LPZ. Continuation of well defined responses to stimulation of the intact eye served as a control against direct effects of the kainic acid at the LPZ recording site. In six of seven cases examined, kainic acid deactivation of neurons at the injection site blocked responsiveness to lesioned-eye stimulation at the ectopic RF for the LPZ recording site. We therefore conclude that long-range horizontal projections contribute to the dominant input underlying the capacity for retinal lesion-induced plasticity in V1.

Termination of the geniculocortical projection in the striate cortex of macaque monkey: a quantitative immunoelectron microscopic study

The goal of this present study was to derive a new estimate of the synaptic contribution of the dorsal lateral geniculate nucleus (dLGN) to the subdivisions of its main recipient layer, layer 4C, of striate cortex of macaque monkey. The projection from the dLGN and its terminal boutons within layer 4C were visualized by immunodetection of the calcium binding protein, parvalbumin (PV), which is expressed in relay cells of the dLGN. The proportion of asymmetric synapses formed by PV-positive boutons within the alpha and beta sublayers of 4C was estimated by using a nonbiased stereological counting method. The proportion of asymmetric synapses contributed by the PV-positive boutons to layer 4Calpha is 8.7%; to 4Cbeta is 6.9%. Assuming all the PV-positive asymmetric synapses derive from the dLGN relay cells, this gives a ratio of dLGN synapses per neuron of 192 in layer 4Calpha and 128 in layer 4Cbeta. Thus, the recurrent excitatory input from neighboring cortical neurons must play an important part in responses of the neurons lying at the input stage of the cortical circuit.

Comparison of the selectivity of postsynaptic potentials and spike responses in cat visual cortex

Intracellular recordings were made from neurons in the cat visual cortex (area 17) to compare the orientation and direction selectivities of the output of a cell with those of the input the cell receives. The input to a cell was estimated from the PSPs (postsynaptic potentials) evoked by visual stimulation, and the output estimated from the number of spikes generated during the same responses. For the whole sample, selectivity of the output of cells was significantly higher than selectivity of their input. Upon PSP to spike transformation, the selectivity index was, on average, doubled. However, the degree of the selectivity improvement in individual cells was very different, varying from cases in which highly selective output was created from a poorly selective input and thus selectivity was greatly improved, to little or no improvement in other neurons. The improvement of selectivity was not correlated with resting membrane potential, threshold for action potential generation, background discharge rate or amplitude of optimal PSP response. Further, no systematic difference was found between simple and complex cells in the input-output relations, indicating that the 'tip of the iceberg' effect on shaping the response selectivity was cell specific, but not cell type specific. This supports the notion that multiple mechanisms are responsible for generation of the response selectivity, and that the contribution of any particular mechanism may vary from one cell to the other. The heterogeneity of the input-output relations in visual cortical cells could indicate different functions of cells in the cortical network; some cells are creating selectivity de novo, the function of other neurons probably being repetition and amplification of the selected signal and arrangement of the output of a whole column.

Response latency of macaque area MT/V5 neurons and its relationship to stimulus parameters

A total of 310 MT/V5 single cells were tested in anesthetized, paralyzed macaque monkeys with moving random-dot stimuli. At optimum stimulus parameters, latencies ranged from 35 to 325 ms with a mean of 87+/-45 (SD) ms. By examining the relationship between latency and response levels, stimulus parameters, and stimulus selectivities, we attempted to isolate the contributions of these factors to latency and to identify delays representing intervening synapses (circuitry) and signal processing (flow of information through that circuitry). First, the relationship between stimulus parameters and latency was investigated by varying stimulus speed and direction for individual cells. Resulting changes in latencies were explainable in terms of response levels corresponding to how closely the actual stimulus matched the preferred stimulus of the cell. Second, the relationship between stimulus selectivity and latency across the population of cells was examined using the optimum speed and direction of each neuron. A weak tendency for cells tuned for slow speeds to have longer latencies was explainable by lower response rates among slower-tuned neurons. In contrast, sharper direction tuning was significantly associated with short latencies even after taking response rate into account, (P = 0.002, ANCOVA). Accordingly, even the first 10 ms of the population response fully demonstrates direction tuning. A third study, which examined the relationship between antagonistic surrounds and latency, revealed a significant association between the strength of the surround and the latency that was independent of response levels (P < 0.002, ANCOVA). Neurons having strong surrounds exhibited latencies averaging 20 ms longer than those with little or no surround influence, suggesting that neurons with surrounds represent a later stage in processing with one or more intervening synapses. The laminar distribution of latencies closely followed the average surround antagonism in each layer, increasing with distance from input layer IV but precisely mirroring response levels, which were highest near the input layer and gradually decreased with distance from input layer IV. Layer II proved the exception with unexpectedly shorter latencies (P< 0.02, ANOVA) yet showing only modest response levels. The short latency and lack of strong direction tuning in layer II is consistent with input from the superior colliculus. Finally, experiments with static stimuli showed that latency does not vary with response rate for such stimuli, suggesting a fundamentally different mode of processing than that for a moving stimulus.

The contribution of intracortical inhibition to dynamics of orientation tuning in cat striate cortex neurons

Orientation tuning of some neurons in cat visual cortex (area 17) revealed successive shifts of the preferred orientation and widening of tuning in time during the first 150 ms after onset of a flashing light bar. The mechanisms of these dynamics and the possible role of intracortical inhibition are still under discussion. In this study we analysed the dynamics using the time slice method before and during blockade of GABAergic inhibition by microiontophoretic application of bicuculline and observed two main types of neuronal behaviour. The first group of neurons (39 of 68 units or 57.4%) with relatively sharp tuning and absence or relatively small shifts of preferred orientation under control conditions increased or developed this shift during bicuculline application. Changes in tuning were observed between 30 and 150 ms after stimulus onset when inhibition was blocked. Neurons of the second group (29 units or 42.6% of cases) displayed pronounced shifts of preferred orientation under control conditions which was typically diminished or lost during blockade of inhibition. The results indicate different contributions of intracortical inhibition to different neurons distinguishing by stability or time dependence of their orientation preference during normal response generation. In one group of striate cells orientation tuning was kept narrow and constant in time by intracortical inhibition, while in another group orientation tuning dynamics are induced by inhibitory mechanisms.

Orientation tuning and receptive field structure in cat striate neurons during local blockade of intracortical inhibition

The contribution of intracortical inhibition to orientation tuning in the cat striate cortex (area 17) was studied by responses to flashing light bars of different orientations and lengths in 68 single-units before and during microiontophoretical application of bicuculline, a GABAA antagonist, A three-fold increase in the background activity (13.3 +/- 1.3 vs 4.4 +/- 0.5 imp/s) and 4.4-fold increase in the maximal discharge frequency (264.4 +/- 22.3 vs 60.6 +/- 5.3 imp/s) was found in 96.0% of the cells studied during microiontophoresis. In most units all characteristics of orientation tuning significantly changed during application of bicuculline: i) tuning width increased in 76.3% of cells from 52.7 +/- 2.8 degrees in control to 85.2 +/- 4.6 degrees, ii) tuning selectivity diminished in 63.6% of cells by a factor of 1.5, and iii) tuning quality dropped in 68.5% of cases by a factor of 2.5. The threshold ejection current of bicuculline for widening of tuning was in 2/3 of the cells in the range from +10 to +40 nA (+31.0 +/- 4.5 nA) and the maximum effect was obtained in 3/4 of units with +30(-) + 100 nA (+67.1 +/- 6.0 nA). Unmasking of additional excitatory inputs to the studied cells due to blockade of the inputs from inhibitory interneurons in considered as the main mechanism of the described bicuculline effects.

Pyramidal and nonpyramidal callosal cells in the striate cortex of the adult rat

The aim of this study has been to determine the neuronal types (pyramidal and nonpyramidal) within the rat's visual cortex, which project through the corpus callosum. To this end, the morphology and laminar distribution of callosal cells have been investigated by combining Diamidino Yellow retrograde tracing with intracellular injection of Lucifer Yellow in slightly fixed tissue slices. The visual callosal projection arises from pyramidal cells of diverse morphology in layers II to VIb, as well as from several modified pyramids located mainly in layers II, IV (star pyramids) and VIb (horizontal or inverted pyramids and related forms of spiny stellate cells). Our results indicate that in rats, as in other mammals, several types of nonpyramidal neurons also contribute to the contralateral projection. Bitufted cells in layers II-III and V were found to project contralaterally. Moreover, a spine-free layer V cell and a sparsely spiny multipolar neuron of layer IV were also labeled. In both stellate cells, partial axonal labeling reveals that these callosal cells display a local axonal arborization. Finally, our results of retrograde transport with Diamidino Yellow and with another sensitive retrograde tracer, the beta subunit of the cholera toxin, demonstrate for the first time that the two main neuronal types of layer I participate in the callosal projection. In layer I, several small horizontal cells of the inner half of layer I and a large subpial cell displaying long radiating dendrites were also injected. The latter cell may correspond to the Cajal-Retzius cell of the adult rat. In spite of the important differences in the organization of the visual system between rodents and cats, the callosal projection in both mammals is composed of a large variety of pyramidal cells and several nonpyramidal neurons. This high morphological diversity suggests that the callosal projection is much more physiologically complex than the extracortical efferents of the visual cortex, resembling other cortico-cortical connections. The roles that the different callosal cells may play in the processing of visual information are discussed in relation to the known functions of the corpus callosum.

Inactivation of the infragranular striate cortex broadens orientation tuning of supragranular visual neurons in the cat

Intracortical inhibition is believed to enhance the orientation tuning of striate cortical neurons, but the origin of this inhibition is unclear. To examine the possible influence of ascending inhibitory projections from the infragranular layers of striate cortex on the orientation selectivity of neurons in the supragranular layers, we measured the spatiotemporal response properties of 32 supragranular neurons in the cat before, during, and after neural activity in the infragranular layers beneath the recorded cells was inactivated by iontophoretic administration of GABA. During GABA iontophoresis, the orientation tuning bandwidth of 15 (46.9%) supragranular neurons broadened as a result of increases in response amplitude to stimuli oriented about +/- 20 degrees away from the preferred stimulus angle. The mean (+/- SD) baseline orientation tuning bandwidth (half width at half height) of these neurons was 13.08 +/- 2.3 degrees. Their mean tuning bandwidth during inactivation of the infragranular layers increased to 19.59 +/- 2.54 degrees, an increase of 49.7%. The mean percentage increase in orientation tuning bandwidth of the individual neurons was 47.4%. Four neurons exhibited symmetrical changes in their orientation tuning functions, while 11 neurons displayed asymmetrical changes. The change in form of the orientation tuning functions appeared to depend on the relative vertical alignment of the recorded neuron and the infragranular region of inactivation. Neurons located in close vertical register with the inactivated infragranular tissue exhibited symmetric changes in their orientation tuning functions. The neurons exhibiting asymmetric changes in their orientation tuning functions were located just outside the vertical register. Eight of these 11 neurons also demonstrated a mean shift of 6.67 +/- 5.77 degrees in their preferred stimulus orientation. The magnitude of change in the orientation tuning functions increased as the delivery of GABA was prolonged. Responses returned to normal approximately 30 min after the delivery of GABA was discontinued. We conclude that inhibitory projections from neurons within the infragranular layers of striate cortex in cats can enhance the orientation selectivity of supragranular striate cortical neurons.

Stimulus dependence of orientation and direction sensitivity of cat LGNd relay cells without cortical inputs: a comparison with area 17 cells

The cortical contribution to the orientation and direction sensitivity of LGNd relay cells was investigated by recording the responses of relay cells to drifting sinusoidal gratings of varying spatial frequencies, moving bars, and moving spots in cats in which the visual cortex (areas 17, 18, 19, and LS) was ablated. For comparison, the spatial-frequency dependence of orientation and direction tuning of striate cortical cells was investigated employing the same quantitative techniques used to test LGNd cells. There are no significant differences in the orientation and direction tuning to relay cells in the LGNd of normal and decorticate cats. The orientation and direction sensitivities of cortical cells are dependent on stimulus parameters in a fashion qualitatively similar to that of LGNd cells. The differences in the spatial-frequency bandwidths of LGNd cells and cortical cells may explain many of their differences in orientation and direction tuning. Although factors beyond narrowness of spatial-frequency tuning must exist to account for the much stronger orientation and direction preferences of cells in area 17 when compared to LGNd cells, the evidence suggests that the orientation and direction biases present in the afferents to the visual cortex may contribute to the orientation and direction selectivities found in cortical cells.

Laminar distribution and neuronal targets of GABAergic axon terminals in cat primary auditory cortex (AI)

The form, density, and neuronal targets of presumptive axon terminals (puncta) that were immunoreactive for gamma-aminobutyric acid (GABA) or its synthesizing enzyme, glutamic acid decarboxylase (GAD), were studied in cat primary auditory cortex (AI) in the light microscope. High-resolution, plastic-embedded material and frozen sections were used. The chief results were: 1) There was a three-tiered numerical distribution of puncta, with the highest density in layer Ia, an intermediate number in layers Ib-IVb, and the lowest concentration in layers V and VI, respectively. 2) Each layer had a particular arrangement: layer I puncta were fine and granular (less than 1 micron in diameter), endings in layers II-IV were coarser and more globular (larger than 1 micron), and layer V and VI puncta were mixed in size and predominantly small. 3) The form and density of puncta in every layer were distinctive. 4) Immunonegative neurons received, in general, many more axosomatic puncta than immunopositive cells, with the exception of the large multipolar (presumptive basket) cells, which invariably had many puncta in layers II-VI. 5) The number of puncta on the perikarya of GABAergic neurons was sometimes related to the number of puncta in the layer, and in other instances it was independent of the layer. Thus, while layer V had a proportion of GABAergic neurons similar to layer IV, it had only a fraction of the number of puncta; perhaps the intrinsic projections of supragranular GABAergic cells are directed toward layer IV, as those of infragranular GABAergic neurons may be. Since puncta are believed to be the light microscopic correlate of synaptic terminals, they can suggest how inhibitory circuits are organized. Even within an area, the laminar puncta patterns may reflect different inhibitory arrangements. Thus, in layer I the fine, granular endings could contact preferentially the distal dendrites of pyramidal cells in deeper layers. The remoteness of such terminals from the spike initiation zone contrasts with the many puncta on all pyramidal cell perikarya and the large globular endings on basket cell somata. Basket cells might receive feed-forward disinhibition, pyramidal cells feed-forward inhibition, and GABAergic non-basket cells would be the target of only sparse inhibitory axosomatic input. Such arrangements imply that the actions of GABA on AI neurons are neither singular nor simple and that the architectonic locus, laminar position, and morphological identity of a particular neuron must be integrated for a more refined view of its role in cortical circuitry.

Morphology of callosal axons interconnecting areas 17 and 18 of the cat

Seventeen callosally projecting axons originating near the border between areas 17 and 18 in adult cats were anterogradely labelled with biocytin and reconstructed in 3-D from serial sections. All axons terminated near the contralateral 17/18 border. However, they differed in their diameter, tangential and radial distributions, and overall geometry of terminal arbors. Diameters of reconstructed axons ranged between 0.45 and 2.25 microns. Most of the axons terminated in multiple terminal columns scattered over several square millimetres of cortex. Thus in general callosal connections are not organized according to simple, point-to-point spatial mapping rules. Usually terminal boutons were more numerous in supragranular layers; some were also found in infragranular layers, none in layer IV. However, a few axons were distributed only or mainly in layer IV, others included this layer in their termination. Thus, different callosal axons may selectively activate distinct cell populations. The geometry of terminal arbors defined two types of architecture, which were sometimes represented in the same axon: parallel architecture was characterized by branches of considerable length which supplied different columns or converged onto the same column; serial architecture was characterized by a tangentially running trunk or main branch with radial collaterals to the cortex. These architectures may relate to temporal aspects of inter-hemispheric interactions. In conclusion, communication between corresponding areas of the two hemispheres appears to use channels with different morphological and probably functional properties.

Quantification of directional and orientational selectivities of visual neurons to moving stimuli

Directional and orientational components usually coexist and are mixed in the cell's overall responses when moving optical stimuli are used to study the response characteristics of visual neurons. While these two properties were quantified with all the previous methods for data analysis, their effects could not be efficiently separated from each other, and thus the analyses were imperfect. In this paper, theoretical evidence and examples are provided to show the defects of the old methods. In order to separate the two components completely, we propose to apply optimal regression analysis with the sine-cosine function series as the fundamental variables. Based on this separation, we defined the orientational selectivity as variation of response strength with orientation and performed integration and averaging to quantify the two properties [cf. Eqs. (5) and (6)]. The present method has the advantages of completeness and accuracy, and can detect some details which would have been missed by other methods. An explanation of the intrinsic implications of the method and our comprehension of directional and orientational selectivities and preferred direction and orientation are also given.

Dynamics of orientation coding in area V1 of the awake primate

To investigate the importance of feedback loops in visual information processing, we have analyzed the dynamic aspects of neuronal responses to oriented gratings in cortical area V1 of the awake primate. If recurrent feedback is important in generating orientation selectivity, the initial part of the neuronal response should be relatively poorly selective, and full orientation selectivity should only appear after a delay. Thus, by examining the dynamics of the neuronal responses it should be possible to assess the importance of feedback processes in the development of orientation selectivity. The results were base on a sample of 259 cells recorded in two monkeys, of which 89% were visually responsive. Of these, approximately two-thirds were orientation selective. Response latency varied considerably between neurons, ranging from a minimum of 41 ms to over 150 ms, although most had latencies of 50-70 ms. Orientation tuning (defined as the bandwidth at half-height) ranged from 16 deg to over 90 deg, with a mean value of around 55 deg. By examining the selectivity of these different neurons by 10-ms time slices, starting at the onset of the neuronal response, we found that the orientation selectivity of virtually every neuron was fully developed at the very start of the neuronal response. Indeed, many neurons showed a marked tendency to respond at somewhat longer latencies to stimuli that were nonoptimally oriented, with the result that orientation selectivity was highest at the very start of the neuronal response. Furthermore, there was no evidence that the neurons with the shortest onset latencies were less selective. Such evidence is inconsistent with the hypothesis that recurrent intracortical feedback plays an important role in the generation of orientation selectivity. Instead, we suggest that orientation selectivity is primarily generated using feedforward mechanisms, including feedforward inhibition. Such a strategy has the advantage of allowing orientation to be computed rapidly, and avoids the initially poorly selective neuronal responses that characterize processing involving recurrent loops.