Complex motion selectivity in PMLS cortex following early lesions of primary visual cortex in the cat.

In the cat, the analysis of visual motion cues has generally been attributed to the posteromedial lateral suprasylvian cortex (PMLS) (Toyama et al., 1985; Rauschecker et al., 1987; Rauschecker, 1988; Kim et al., 1997). The responses of neurons in this area are not critically dependent on inputs from the primary visual cortex (VC), as lesions of VC leave neuronal response properties in PMLS relatively unchanged (Spear & Baumann, 1979; Spear, 1988; Guido et al., 1990b). However, previous studies have used a limited range of visual stimuli. In this study, we assessed whether neurons in PMLS cortex remained direction-selective to complex motion stimuli following a lesion of VC, particularly to complex random dot kinematograms (RDKs). Unilateral aspiration of VC was performed on post-natal days 7-9. Single unit extracellular recordings were performed one year later in the ipsilateral PMLS cortex. As in previous studies, a reduction in the percentage of direction selective neurons was observed with drifting sinewave gratings. We report a previously unobserved phenomenon with sinewave gratings, in which there is a greater modulation of firing rate at the temporal frequency of the stimulus in animals with a lesion of VC, suggesting an increased segregation of ON and OFF sub-regions. A significant portion of neurons in PMLS cortex were direction selective to simple (16/18) and complex (11/16) RDKs. However, the strength of direction selectivity to both stimuli was reduced as compared to normals. The data suggest that complex motion processing is still present, albeit reduced, in PMLS cortex despite the removal of VC input. The complex RDK motion selectivity is consistent with both geniculo-cortical and extra-geniculate thalamo-cortical pathways in residual direction encoding.

Simple and complex visual motion response properties in the anterior medial bank of the lateral suprasylvian cortex.

The cortical regions surrounding the suprasylvian sulcus have previously been associated with motion processing. Of the six areas originally described by Palmer et al. [J Comp Neurol 177 (1978) 237], the posteromedial lateral suprasylvian (PMLS) cortex has attracted the greatest attention. Very little physiological information is available concerning other suprasylvian visual areas, and in particular, the anteromedial lateral suprasylvian cortex (AMLS). Based on its cortical and sub-cortical connectivity patterns, the AMLS cortex is a likely candidate for higher-order motion processing in cat visual cortex. We have investigated this possibility by studying the receptive field sensitivity of AMLS neurons to complex motion stimuli. Neurons in AMLS cortex exhibited large (mean of 354 degrees (2)) and complex-like receptive fields, and most of them (74%) were classified as direction selective on the basis of their responses to sinusoidal drifting gratings. Most importantly, direction selectivity was present for complex motion stimuli. A subset of the neurons sampled (eight of 38 cells; 21%) exhibited pattern-motion selectivity in response to moving plaid patterns. The capacity of AMLS neurons to signal higher-order stimuli was further supported by their selectivity to moving complex random-dot kinematograms. Finally, 45% of 20 neurons were direction selective to a radial optic flow stimulus. Overall, these results suggest that AMLS cortex is involved in higher-order analyses of visual motion. It is possible that the AMLS cortex represents a region between PMLS and the anterior ectosylvian visual area in a functional hierarchy of areas involved in motion integration.

Higher-order motion processing in the pulvinar.

Thalamic nuclei have long been considered as passive relay stations for sensory signals en route to the cerebral cortex, where higher level processing occurs. In recent years, it has been proposed that thalamic nuclei may actively participate in the processing of specific information in conjunction with cortical areas. In support of this hypothesis, we recently discovered that neurons in the main extrageniculate visual nucleus, the pulvinar, exhibit higher-order visual properties that were, until now, only associated with higher-order cortical areas. Pulvinar neurons can indeed code the veridical direction of a moving plaid pattern, indicating that these cells can integrate ambiguous signals into a coherent percept. This finding as well as our demonstration that there are cortico-thalamo-cortical loops involved in complex motion analysis open promising avenues in unraveling the function of the pulvinar complex in normal vision.

Responses of neurons in the cat posteromedial lateral suprasylvian cortex to moving texture patterns.

The posteromedial lateral suprasylvian cortex represents a point of convergence between the geniculostriate and extrageniculostriate visual pathways. Given its purported role in motion analysis and the conflicting reports regarding the texture sensitivity of this area, we have investigated the response properties of cells in PMLS to moving texture patterns ("visual noise"). In contrast to previous reports, we have found that a large majority of cells (80.1%) responds to the motion of a texture pattern with sustained discharges. In general, responses to noise were more broadly tuned for direction compared to gratings; however, direction selectivity appeared more pronounced in response to noise. The majority of cells was selective for drift velocity of the noise pattern (mean optimal velocity: 26.7 degrees /s). Velocity tuning was comparable to that of its principal thalamic input, the lateral posterior pulvinar nucleus. In general, responsiveness of cells in the posteromedial lateral suprasylvian cortex increased with increasing texture element size, although some units were tuned to smaller element sizes than the largest presented. Finally, the magnitude of these noise responses was dependent on the area of the visual field stimulated. In general, a stimulus corresponding to roughly twice the size of the receptive field was required to elicit an equivalent half-maximal response to that for gratings. The results of this study indicate that the majority of cells in the posteromedial lateral suprasylvian cortex can be driven by the motion of a fine texture field, and highlight the importance of this area in motion analysis.

Motion integration in a thalamic visual nucleus.

Thalamic nuclei have long been regarded as passive relay stations for sensory information en route to higher level processing in the cerebral cortex. Recently, physiological and theoretical studies have reassessed the role of the thalamus and it has been proposed that thalamic nuclei may actively participate with cortical areas in processing specific information. In support of this idea, we now show that a subset of neurons in an extrageniculate visual nucleus, the lateral-posterior pulvinar complex, can signal the true direction of motion of a plaid pattern, indicating that thalamic cells can integrate different motion signals into a coherent moving percept. This is the first time that these computations have been found to occur outside the higher-order cortical areas. Our findings implicate extrageniculate cortico-thalamo-cortical loops in the dynamic processing of image motion, and, more generally, as basic computational modules involved in analysing specific features of complex visual scenes.

Spatial frequency processing in posteromedial lateral suprasylvian cortex does not depend on the projections from the striate-recipient zone of the cat's lateral posterior-pulvinar complex.

It is generally considered that the posteromedial part of the cat's lateral suprasylvian cortex is involved in the analysis of image motion. The main afferents of the posteromedial lateral suprasylvian cortex come from a direct retinogeniculate pathway and indirect retinotectal and retino-geniculo-cortical pathways. Removal of the primary visual cortex does not affect the spatial and temporal processing of suprasylvian cortex cells suggesting that these properties are derived from thalamic input. We have investigated the possibility that the striate-recipient zone of the lateral posterior nucleus-pulvinar complex may be responsible for the spatial (and temporal) frequency processing in posteromedial lateral suprasylvian cortex since these two regions establish strong bidirectional connections and share many visual properties. Experiments were done on anaesthetized normal adult cats. Visual responses in suprasylvian cortex were recorded before, during, and after the deactivation of the lateral part of the lateral posterior nucleus accomplished by the injection of lidocaine or GABA. Results can be summarized as follows. A total of 64 cells was tested. Out of this number, 11 units were affected by the deactivation of the lateral part of lateral posterior nucleus and one cell, by the blockade of pulvinar. For all cells, except one, the effect consisted in a global reduction of the evoked discharge rate suggesting that the thalamo-suprasylvian cortex projections are excitatory in nature. We did not find any significant differences in the optimal spatial frequency, nor in the width of the tuning function, whether the grating was presented at half- or saturation contrast. In addition, there were no significant differences between the low- and high cut-off spatial frequency values computed before and after the deactivation of the lateral posterior nucleus. No specific changes were observed in the contrast sensitivity function of the posteromedial lateral suprasylvian cortex cells. Similar results were observed with respect to the temporal frequency tuning functions. Deactivating the lateral posterior nucleus did not modify the direction selectivity nor the organization of the subregions of the lateral suprasylvian cortex "classical" receptive fields. The absence of strong changes in posteromedial lateral suprasylvian cortex cell response properties following the functional blockade of the lateral posterior nucleus suggests that the projections from this part of the thalamus are not essential to generate the spatial characteristics of most posteromedial lateral suprasylvian cortex receptive fields. These properties may be derived from other thalamic inputs (e.g., medial interlaminar nucleus) and/or from the intrinsic computation of the afferent signals within the lateral suprasylvian cortex. On the other hand, it is possible that the lateral posterior nucleus lateral suprasylvian cortex loop may be involved in other functions such as the analysis of complex motion as suggested by the findings from our and other groups.

Comparison of the responses to moving texture patterns of simple and complex cells in the cat's area 17.

1. Whether complex (C) cells are the only truly texture-sensitive units in the cat's primary visual cortex remains controversial. In view of the strong physiological significance of having putatively only one class of cells sensitive to visual noise in the striate cortex, we reinvestigated this issue. Sensitivities of simple (S) and C cells to noise were quantitatively studied and compared in order to clearly document the response properties of cells in the striate cortex to visual noise and to establish whether one can unequivocally segregate S from C cells on the basis of those specific properties. 2. Receptive fields were stimulated with all relevant stimuli, i.e., drifting sine-wave gratings, electronically generated noise pattern of 256 x 256 elements (ratio 1:1 of dark and light elements), and flashing and moving bars (both bright and dark). 3. A total of 60 S cells out of 85 (70.6%) and 90 C cells out of 101 (81.8%) responded to the motion of visual noise. Responses of most C cells were sustained, i.e., their discharge rate was maintained at a constant level throughout presentation of the stimulus. On the other hand, responses of the majority of S cells were characterized by several bursts of discharges. On average, optimal firing rates were greater for gratings than for noise. 4. For practically all cells, responses to noise varied as a function of direction of motion. The mean direction bandwidths were, respectively, 43 +/- 24 degrees and 48 +/- 23 degrees (mean +/- SD) for S and C cells. In both groups, neurons were more broadly tuned for the direction of noise than that of gratings (t-test, P < 0.001). We rarely observed bimodal tuning curves for noise, with each peak lying on either side of the orientation curve. These results could be expected if one considers texture stimuli not in the space domain (as dot patterns) but in the frequency domain, i.e., patterns containing all spatial frequencies and orientations. 5. In general, the direction indexes of S and C cells were similar whether they were stimulated by drifting noise or gratings. S cells had a slight tendency to be more direction selective for noise than for gratings. 6. For all S and C cells tested, responses to noise varied as a function of drift velocity. The mean optimal velocity was 12.9 and 10.2 degrees/s for S and C cells, respectively (t-test, P > 0.05). Most cells were band-pass with mean bandwidths of 2.2 and 2.7 octaves for S and C cells, respectively.(ABSTRACT TRUNCATED AT 400 WORDS)