Abstract
The perceived position of briefly presented visual stimuli can be markedly affected by nearby motion signals. Here we investigate the effective temporal window of this interaction by measuring shifts in the perceived position of stimuli presented at various times before, during and after a fixed period of motion. Observers were required to judge the vertical alignment of two target patches (1-dimensional Gaussian blobs; width= 1 deg; σvertical = 0.33 deg; horizontal separation = 5 deg; duration = 20ms) presented adjacent to two vertical strips of sinusoidal grating (1 cycle/deg, width = 1 deg, height = 25.6 deg, separation = 1 deg) that drifted in opposing directions (up/down) at 5 deg/s. Results reveal a consistent and distinctive time course for periods of inducing motion ranging from 2 to 8 seconds. Perceived position begins to be pulled in the direction of motion for target onsets 100–200ms prior to the commencement of motion. This positional shift increases in magnitude as a function of temporal proximity, reaching maximum where target and motion onset are synchronous. Thereafter, the effect reduces by ∼50% over the next 800ms, then remains constant for the majority of target onsets spanning the period of motion. Near the offset of motion, the magnitude of positional shifts dissipates to zero over a period of ∼500ms and positional shifts in the opposite (aftereffect) direction are observed. This pattern of results is incompatible with existing accounts, which posit that shifts in perceived position reflect the simple accrual of motion signals over a fixed temporal interval.
This work was supported by the Wellcome Trust.