Action potentials generated by neurons in the cerebral cortex eventually give rise to conscious sensations. Understanding this process requires both a description of what information is represented in the activity of single neurons, and a description of the mechanism by which that representation is generated. In natural viewing, image changes almost invariably occur simultaneously in the two eyes. In order to exploit this information it is necessary that information from the two eyes be conducted to the cerebral cortex in a way that preserves the relative timing of retinal events. This may be difficult to achieve far from the vertical meridian, since one eye will receive stimulation in the temporal hemiretina while the other eye is stimulated in the nasal hemiretina. Earlier studies (using monocular stimuli) have suggested that conduction may be different for the two halves of the retina. The impact of naso-temporal lag on disparity processing was examined in 48 disparity selective V1 neurons recorded from left hemispheres of two awake fixating monkeys (so all stimuli were presented in the right hemifield). An interocular delay was introduced into dynamic random dot stereograms by delaying or advancing the sequence of dot patterns shown to one eye. Delays of 14ms (1 video frame) generally produced a substantial attenuation of disparity selectivity. A systematic asymmetry was observed: when the left eyes image (temporal retina) was delayed relative to the right eye (nasal retina), the attenuation was greater than for delays of the same magnitude but in the opposite direction. To quantify this, the data for each cell (a minimum of 7 disparities at each of 5 different delays) were fit with a single two dimensional Gabor function. The location of the center of the Gaussian envelope was used to estimate naso-temporal lag. 40/48 neurons preferred negative delays (left eye stimulus preceding right), with a mean value of -3.3ms (+- 0.7ms SEM, p < 0.0001). Stereoacuity thresholds were measured in one monkey and two human observers, over a range of interocular delays. For stimuli presented in the right hemifield, thresholds were systematically higher when the left eye stimulus lagged the right eye. The difference was compatible with a naso-temporal lag similar in size to that observed physiologically. For stimuli presented in the left hemifield lower thresholds were observed when the right eye stimulus lagged the left. Thus it appears that conduction delays (probably in the retina) lead to a naso-temporal difference in the timing of activity reaching the cortex. Disparity selective neurons fail to compensate for this delay. This failure is also reflected in psychophysical measures of stereoacuity. When stimuli are moving, these delays can give rise to false sensations of depth (the Pulfrich effect). Earlier studies in cats suggested that this phenomenon is primarily the result of joint encoding of motion and depth by cortical neurons. We investigated the response of neurons in V1 of the awake monkey to random dot stereograms with both spatial and temporal disparity. Three quantitative differences between monkey and cat emerged: 1. Temporal integration is much shorter in the monkey. In monkey V1, disparity tuning is rapidly disrupted as interocular time-lags are introduced. We fit a Gaussian to the amplitude of disparity tuning as a function of interocular delay; the mean SD was 15ms. This is much shorter that the intergration times (approx 100ms) seen in cat neurons. 2. Cells tuned both to motion and disparity are much rarer in the monkey. 3. Joint motion-disparity tuning is correlated with direction selectivity. Direction selectivity was measured monocularly with drifting gratings for 33 cells. 9/16 direction selective cells showed significant tilt, while only 1/17 non direction selective cells showed this. Thus the high incidence of joint motion-disparity encoding reported in cats probably reflects the higher proportion of direction selectivity in cat visual cortex. Thus, only a minority of V1 neurons in the monkey exhibit joint encoding of motion and disparity. These neurons probably contribute to the Pulfrich effect, but in its usual form the percept is more likely dominated by the consequences of the stimulus geometry.
Showing the most recent 10 out of 24 publications
