Abstract
Abstract: :
Purpose: It is often assumed that transient channels can yield little information about slowly changing stimuli. We tested this hypothesis on cells of the MC-pathway and showed it to be false. Methods: We measured responses of macaque retinal ganglion cells to moving stimuli and analyzed the variability of response locus. Results: With moving bars and edges, peak rate of MC-cell responses increases rapidly with velocity, but the number of impulses per response decreases. From neurometric analysis, these two factors are shown to both contribute to positional accuracy inherent in the retinal signal, so that between 0.5 to 4-8 deg/sec MC-cells deliver a consistently accurate spatial signal. The accuracy of responses of PC-cells deteriorates very rapidly with velocity. A similar analysis was performed for gratings. Each cycle of response was fourier analyzed. The variance of the amplitude and phase of response vectors give an estimation of noise in the response. MC-cell signal-to-noise ratio was highest at low temporal frequencies; with increasing temporal frequency, response variance increased more rapidly than response amplitude, so that signal-to-noise ratio decreased. For bars, edges and gratings, cellular signal-to-noise ratios were matched well to human vernier performance with the appropriate targets. Conclusion: Transient MC-cells have high signal-to-noise ratios and can yield accurate spatial information at low target velocities. Since the results correlate with psychophysical data, it appears that this information is centrally utilized. In detection of flicker, psychophysical thresholds correlate with peak cellular firing rates (Lee et al., JOSA A, 7, 2223-36, 1970) but in the current data psychophysical thresholds correlate with cellular signal-to-noise ratio. There is thus an interesting task-dependent difference in the way cell behavior correlates with psychophysical performance.
Keywords: 586 spatial vision • 415 ganglion cells • 620 visual acuity