Abstract
Goal: Normalization models provide a formal description of contextual modulation in the visual system. Normalization has been widely applied to contextual interactions across space, but how normalization mechanisms may operate over time remains an important area for investigation. Here we investigated temporal normalization across hundreds of milliseconds, a timescale implicated in temporal interactions across successive stimuli such as the attentional blink and temporal crowding. Normalization models include suppressive interactions between stimuli, with stronger suppression from higher contrast stimuli. We psychophysically tested for such interactions in the temporal domain using an orientation discrimination task with two sequential stimuli that independently varied in contrast. Method: Participants viewed a sequence of two Gabor stimuli (T1 and T2) presented in the lower right quadrant of the screen (5.7° eccentricity). A readiness cue indicated the stimulus timing, and a response cue indicated the target, instructing participants to report the tilt of either T1 or T2. Stimuli were presented with a 100 ms duration and a 250 ms stimulus onset asynchrony. Contrasts were independently manipulated so that each target was either high (64%) or low (4%) contrast. Perceptual sensitivity (d’) for orientation discrimination was calculated for each target (T1 and T2) as a function of both target and non-target contrast. Results and conclusions: Perceptual sensitivity to target orientation was influenced by both target and non-target contrasts. As expected, perceptual sensitivity increased with target contrast for both T1 and T2. Conversely, perceptual sensitivity decreased with higher non-target contrast for both T1 and T2 when the targets were low contrast and for T1 only when the targets were high contrast. The results show strong influences of temporally separated non-targets on perceptual sensitivity for targets, consistent with normalization across a timescale of hundreds of milliseconds.