Library

Notes: Abstract  Two parallel visual systems, the magnocellular (M) and parvocellular (P) pathways, originate from different types of retinal ganglion cells, and are known to be segregated in different portions of the pregeniculate visual pathways. Their relative contribution to two main cortical streams, dorsal and ventral, is still under discussion, but it is reasonable to suppose that selective damage to the M or P subcortical system might interfere with specific aspects of processing within one or the other cortical system. Using two different apparent-motion tasks, we compared the performance of patients affected by compression of the ventral part of the pregeniculate visual pathways with that of normal controls. In the first task, observers detected small displacements of a low-contrast vertical bar, while in the second task they estimated the visible persistence of moving dots. In the first task, patients were impaired with parafoveal displays, especially in the temporal portion of the visual field. In the second task, patients showed reduced suppression of visible persistence at long, but not at short, exposure durations. Three considerations support the hypothesis that these results represent a selective impairment of the M system. First, M axons are more likely to suffer from compression, particularly in the case of a mass growing from below since they are known to occupy a ventral subpial position in the optic chiasm and tract. Second, the performance of patients with a ventral compression is consistent with the characteristics of the response properties of P ganglion cells, which have previously been shown to exhibit elevated and unmodulated thresholds for displacement detection in the macaque monkey. Third, such patients are less sensitive to the inhibitory signals that suppress visible persistence, which probably originate in the M system.

Notes: Summary Using permanently implanted electrodes in squirrel monkeys and macaques, transmission through the lateral geniculate nucleus (LGN) was assayed from the amplitude of potentials evoked in optic radiation by an electrical pulse applied to optic tract. Averaging of either individually or machine selected potentials, elicited at 0.3, 1.0, 20 or 50 Hz, in all cases showed a decrease in transmission ranging from 5–60 % in the period after saccadic eye movements made ad libitum. The suppression was greater in a patterned visual environment than in diffuse illumination, which in turn was greater than that occurring following saccades in the dark. Demonstration of the effect in darkness always required data averaging and never exceeded 20%. The effect was consistently greater in the magnocellular than parvocellular component. Suppression was often abruptly terminated and replaced by a facilitation of 5–15% about 100 msec after saccade detection. Comparable effects were observed for excitability of striate cortex tested by a stimulus pulse applied to optic radiation. In addition, sharply demarcated potentials inherently arising in LGN and striate cortex were found in association with saccades made even in total darkness. Neglecting a possible but dubious contribution from eye muscle proprioceptors, the experiments establish the existence of a centrally originating modulation of visual processing at both LGN and striate cortex in relation to saccadic eye movement in primates. This modulation may partially underlie the phenomenon of “saccadic suppression” and hasten the acquisition of a meaningful visual sample immediately following an ocular saccade. It remains uncertain as to how it may relate to similar or greater effects accompanying changes in alertness, or to fluctuations of unknown origin occurring sometimes semirhythmically at 0.05–0.03 Hz (Fig. 7).

Notes: Summary The responses of neurones in laminae A and A1 of the cat lateral geniculate nucleus to moving stimuli were investigated at different background luminances. Moving bright slits, dark bars and edges were employed; the contrast of stimuli against the background was held constant. Background intensities varied from 10−3 to 102 td. Responses as stimuli passed across the centres of LGN receptive fields became stronger with increasing levels of light adaptation up to 10−1–101 td and then remained constant. Responses as stimuli passed through surround regions altered qualitatively with adaptation level, generally increasing in strength and complexity with background luminance. As a bright slit for on-centre cells or dark bar for off-centre cells left the surround, in almost all units a strong secondary peak could be elicited by an appropriate selection of the adaptation conditions. Many features of the responses to moving stimuli could not be predicted from the responses to stationary stimuli under different adaptation conditions described in the previous paper.

Notes: Summary The responses of neurones in the lateral geniculate nucleus (LGN) were investigated in anaesthetised rhesus monkeys. A new classification for cells in the parvocellular layers (PCL) is proposed, based on their spectral response curve and their response to white stimuli: (A) narrow-band, short wavelength (NS) excited cells, activity suppressed by white stimuli; (B) wide-band, short-wavelength (WS) excited cells, excited by white stimuli; (C) wide-band, long-wavelength (WL) excited cells, (D) narrow-band, long-wavelength (NL) excited cells, activity suppressed by white stimuli; (E) light suppressed (LI) cells, activity suppressed by all wavelengths, usually with some concealed excitatory input at extreme short or long wavelengths. Responses to moving bars and to spots of various diameters (area response curves) were determined for various wavelengths. It was found that the receptive fields from which wavelength-dependent excitatory or suppressive effects could be elicited are concentrically superimposed. The spectral responsiveness of the excitatory inputs to individual cell types corresponds to the absorption curves of single cones (S-, M- or L-cone for NS, WS and WL cells respectively), the spectral distribution of the suppressive mechanisms of all cells was panchromatic and approximately fitted to a sum of all cones. The excitatory input to NL-cells cannot be related to any of the known cone absorption curves, and a simple (L-M) subtraction model is questioned. Neurones in the magnocellular layers (MCL) can be divided into on- and off-centre cells as in the cat's LGN and give qualitatively similar responses over the whole spectrum. In contrast to the tonic responses of PCL cells, MCL cells respond phasically to chromatic and white flashed spots, even with the smallest stimuli. Implications of these findings for colour processing in the LGN are discussed.

Notes: Summary Steadily illuminated surrounds, remote from the receptive field centre, are shown to affect the responses of primate visual cells. Intensity-response curves of cells of the macaque lateral geniculate nucleus were measured using a successive contrast paradigm where chromatic or achromatic stimuli were presented in alternation with a white adaptation field of constant luminance. Adding white surround annuli around stimuli and adaptation field shifted the intensity-response curves to higher intensity ranges. Since response curves can be non-monotonic, this remote surround effect can result in an increase or decrease in responsiveness (facilitation or suppression) dependent on stimulus intensity. Steady surrounds, remote from the receptive field centre, thus control cell sensitivity and responses by means of simultaneous contrast.

Notes: Summary The receptive fields of LGN cells were investigated with stationary light and dark spot and annulus stimuli. Stimulus size and background intensity were varied while stimulus/background contrast was kept constant. The speed of dark adaptation varied considerably from cell to cell. Dark adaptation made responses more sustained in all neurones and eliminated the oscillatory on-responses evoked under some conditions in the light-adapted cells. Dark adaptation led also to a disappearance of early phasic inhibition in on-responses, and increased response rise time and latency. The power of surround responses to inhibit centre responses decreased slightly at low levels of light adaptation in LGN cells but much less than in retinal ganglion cells. Some other traces of changing retinal surround effects also appeared in the LGN on dark adaptation. For example, the functional size of receptive fields increased at low levels of illuminance as has been observed in retinal ganglion cells and the receptive fields as estimated from response peaks were larger than those estimated from sustained components.

Notes: Summary Response patterns to complex visual stimuli were further analysed. Patterns were correlated with linear or non-linear components of the stimulus pattern at various wavelengths. Resulting correlograms revealed the spatial and spectral structure of receptive fields; they showed peaks or troughs according to whether that wavelength was associated with an increase or decrease in cell firing. Spectral response curves as derived from linear correlograms were similar to those reported for monochromatic stimuli. Variability in responsiveness and crossover wavelengths was high between parvocellular layer (PCL) cells even of the same class. Spatial differences between excitatory and suppressive receptive field components, i.e. a centre-surround organisation, are not apparent in linear correlograms from PCL cells. In this respect, spectral response curves do not qualitatively change with stimulus size. Correlation in time and the derivation of impulse functions showed that, even in magnocellular layer (MCL) cells, responses to luminance steps are of mainly temporal origin and due to a transient component in the response. A description of cell responses based on linear processing accounted well for the response patterns obtained in our experiments. Of various non-linear interactions investigated, only some kind of non-linear spectral differentiation provided an improvement in the description of cell responses. This improvement, however, was only minor and not present in all cells.

Notes: Summary We recorded from single neurons in the parvocellular layers of the lateral geniculate body of anesthetized monkeys. Spectral response curves of parvocellular neurons depended on the luminance ratio between the chromatic stimuli and achromatic background. From response/intensity curves, we determined the relative luminance between a coloured and an achromatic (white) light at which a given cell became non-responsive (critical luminance ratio, CLR). The spectral dependence of the CLRs of narrow (N) and wide band (W) cells with opponent receptor input showed characteristic differences. The activity of W-cells increased with luminance increase of a white light and of a coloured light in the specific spectral region of the cell (yellow-red for the long wave length sensitive WL-, and yellow-green-blue for the short wave length sensitive WS-cells), while N-cells were activated by their specific spectral light (blue for NS-cells, red for NL-cells) and by a luminance decrease of achromatic white. N-cells discriminate best between their characteristic colour and white at luminance ratios below their respective CLR, while W-cells distinguish best between a light of their characteristic colour and white at chromatic/ achromatic luminance ratios above their respective CLR. Yellow sensitive W-cells with a narrow spectral sensitivity peaking around 570 nm and with only a small or no response to white light, could enable distinction between white and yellow of similar luminance. The findings are consistent with the opponency model of spectrally sensitive cells in the LGB. We discuss their implications for colour coding by parvocellular cells. N- and W-cells appear to behave complementary with respect to luminance information (N-cells may be compared to the cat's off-cells, W-cells to on-cells). S- and L-cells are complementary with respect to colour. The yellow sensitive WM-cells are critical for the discrimination of yellow and white, while cells with excitatory cone input from blue and red cones (W-SL-cells) may aid the perception of purple. The fact that, at different relative luminance ratios between a chromatic stimulus and a white background, the whole family of parvocellular cells is involved differently in coding for colour, may explain the different appearance of colours against a white background at different luminance ratios and the perception of induced colours.

Notes: Summary We have recorded from 661 single neurons in the foveal and parafoveal region of area 17 of the awake trained macaque monkey. The functional properties of 538 cells were investigated in detail, with flashed and moving stimuli of varying form and colour. Irrespective of their functional properties such determined, each neuron was also tested with a 2×2° square of various luminance and colour. This was done in order to get an idea how such a simple stimulus is represented by the activities of neurons in area 17. Most of the neurons showed response preference for certain aspects of visual stimuli. We have distinguished the following functional groups: 1. Sustained spectrally selective neurons (21%). These cells respond with tonic discharges to light of their optimal wavelength, and their spectral selectivity corresponded to that of opponent parvocellular cells of the lateral geniculate body. 44% of these cells were excited selectively by long, 23% by middle and 33% by short wavelength light. When slowly moving the 2×2° square of their preferred wavelength across the receptive field, discharge rate remained elevated, as long as the stimulus covered the RF and with little contour enhancement. The majority of the sustained spectrally sensitive cells responded equally well or better to large than to small (1.0°) stimuli, 17.5% were less activated and few of them completely suppressed by larger stimuli. Such cells were poorly orientation sensitive. Only three cells with weak double opponency could be identified (2.7% of this group). 2. Broadband contour (18%) and 3. Panchromatic contour cells (41%). Most neurons of these two groups were strongly activated by spots (1°) centered on their RF. They showed a short phasic response to contrast borders and most of them responded to luminance contrasts, including contrast reversal and colour contrasts equated for luminance. The broadband contour cells showed a slight wavelength preference with only weak or without any opponent suppression, the panchromatic contour neurons did not show any wavelength selectivity. Most showed orientation or direction sensitivity, but very sharp orientation selectivity was less common in spectrally biassed than in panchromatic contour cells (see Fig. 11). They responded tonically to gratings of optimal orientation and therefore may play a role also for cortical representation of textures. 22% of a restricted sample of panchromatic contour cells (or 9% of all cells) were hypercomplex. 4. Light inhibited cells. 7% of all cells were inhibited by small and large light stimuli of any wavelength centered on their receptive field, and were tonically activated by darkspots or contrasts, comparable to the light inhibited cells of the parvocellular lateral geniculate layers. 5. Neurons without consistant visual responses (11%). These neurons could not be driven by any of our visual stimuli. They were usually found in the upper cortical layers. 61 cells were tested for monocular vs. binocular input. 96% were excited from both eyes with various degrees of ocular dominance, but more binocular cells were contralaterally than ipsilaterally dominated (43 and 22%, respectively). Binocular cells showed qualitatively the same functional properties from both eyes, including spectral selectivity if there was any. Binocular summation varied between cells and was in the average 0.7, probably due to interocular inhibition. Some columnar grouping of cells with similar response properties as defined above was found in vertical penetrations, but “mixed” penetrations were common. Spectrally selective cells with the same spectral preference or light inhibited cells often were found close to each other and in the same penetration, but also often mixed with other cells excited by parvocellular input. This spatial organization is consistant with a columnar segregation of cells excited predominantly by one type of parvocellular afferents on the one hand, and contour cells with a mixed excitatory and a strong inhibitory input, on the other hand, but also indicates a considerable mixing and overlap of functional inputs into any axis perpendicular to the cortical surface. The functional organization of area 17 is compared with that of the lateral geniculate body and the prelunate visual area (V4) as investigated with the same methods and by the same laboratory.

Notes: Summary Neurones recorded during penetrations through cat area 17 as near parallel to the radial fibre bundles as possible have been quantitatively tested as to their optimal orientation. Optimal orientation within any one penetration was similar though considerable variability was observed. Histological reconstruction and other considerations showed that this variability could not be attributed to poor penetration angle or limitations of the microelectrode technique. These results confirm that neurones with similar optimal orientations are found in all cortical layers at one cortical locus, but it is difficult to reconcile the variability observed with a mosaic-like distribution of orientation across the cortical surface. The findings are consistent, however, with the assumption of a continuous distribution of orientation sensitivity across the cortical surface with considerable superimposed scatter.