| The response to an abrupt change in stimulus configuration is known as the step response. We have examined the step responses of retinal ganglion cells, recorded extracellularly from the cat retina. Step responses were elicited by abrupt contrast reversals of stationary sinusoidal grating stimuli, and filtered peristimulus time histograms of the responses were examined. They typically begin with a sudden increase or decrease in firing rate, depending on the configuration of the grating relative to the cell's receptive field.; Positive step responses began with an abrupt increase in firing rate that peaked between 50-90 msec following the contrast reversal. Subsequently, firing rate diminished more gradually. However, in many Y-cells and Off-center X-cells this decline was sometimes interrupted by a secondary rise in firing rate that created a secondary peak or hump in the response profile. Thus, the profiles of many of the observed step responses had two peaks in their step response profiles, an early-peak and a late-peak. Although prior investigators have noted this two-peaked appearance, they did not provide any information about the underlying mechanisms. The purpose of this study was to determine the mechanisms responsible for these early- and late-peaks in the response profile.; There were three major findings of this study. (1) The late- and the early-peaks are both present at spatial frequencies that are above the resolution of the receptive field surround. (2) The early-peak is produced by a non-rectifying mechanism, while the late-peak is produced by a rectifying mechanism. (3) The early-peak is produced by a mechanism whose output is dependent on spatial phase, but the late-peak is produced by a mechanism whose output is phase invariant.; These results strongly suggest that the early-peak is produced by the linear center mechanism of the cell, whereas the late-peak is produced by the nonlinear rectifying mechanisms that have been previously described for Y-cells. These results are of particular interest to modelers of retinal ganglion cell dynamics, as they show that models must consider the nonlinear contribution to be accurate. |
