Effects of bleaching and backgrounds on the flash response of the cone system.

1. Increment-threshold functions for flashed backgrounds were measured in the human fovea under several conditions: (1) during dark adaptation following full bleaches, (2) in the presence of steady adapting backgrounds and (3) 500 msec following extinction of adapting backgrounds.2. To prevent the intense flashed backgrounds from interfering with the course of dark adaptation the inter-trial interval was continuously increased during dark adaptation. This technique may prove generally useful for presenting suprathreshold stimuli during dark adaptation.3. All the increment-threshold functions measured during dark adaptation were found to be roughly shape invariant and continuously accelerating when plotted in log-log co-ordinates. Furthermore, in order to translate a function obtained at any given time into coincidence with a function obtained at any other time, it had to be translated vertically and horizontally the same number of log units. This is equivalent to adding or removing neutral density filters from in front of the eye.4. The increment-threshold functions obtained with steady adapting backgrounds were also continuously accelerating, but could not be brought into coincidence by equal vertical and horizontal translation. However, this became possible again if the adapting background was extinguished during presentation of the flashed background.5. These results contradict the equivalent-background hypothesis. None the less, they suggest that under present conditions the effects of bleaches and backgrounds may be similar except that steady backgrounds provide additional quanta which drive the visual system part of the way up its intensity-response function.6. The conclusions above were supported by applying a simple model based on the equation R = R(max). I(n) / (I(n) + I(1) (n)), which has frequently been used to describe the peak responses of retinal neurones to flashed stimuli. Virtually all of the data reported here were fitted by this simple model with R(max) held constant.7. The parameters estimated from the model imply that the flash responses measured in the present experiments differ in at least one fundamental way from receptor responses. Even after taking into account changes in the half saturation constant I(1), steady backgrounds were found to be much less effective than flashed backgrounds in driving the visual system up its intensity-response function. A subtractive inhibitory network prior to the non-linear stages responsible for threshold saturation could explain this result.