Abstract
The location of a stationary flashed object appears shifted in the direction of nearby motion (flash drag effect; FDE). Previous studies have reported the FDE with a temporal tuning as broad as hundreds of milliseconds for various stimulus configurations (Durant & Johnston, 2004; Shim & Cavanagh, 2005; Fukiage, Whitney & Murakami, 2011). Contrary to this notion, we demonstrate that the FDE has a relatively narrow temporal tuning. We systematically varied the stimulus onset asynchrony (SOA) between the flash and the moving stimulus. If the perceived position of the flash is affected by the presence of motion within a broad temporal window located before and after the flash onset time, the FDE should be broadly observed well before and well after the period within which the moving stimulus was actually displayed. We presented a vertically drifting sinusoidal grating for 500 ms, and briefly flashed a horizontal bar next to it at various SOAs. We asked observers to judge the vertical offset between the flash and the central fixation cross in two-alternative forced choice, and determined the physical offset required to establish the subjective alignment as the magnitude of the FDE at each SOA. We found that the FDE began to occur approximately 100 ms before the onset of the moving stimulus and sharply increased with increasing SOA, and that the FDE started to decrease approximately 100 ms before the disappearance of the moving stimulus and quickly vanished thereafter. Our results suggest that under a certain condition, the position of the flash is calculated in reference to visual events that occur only within a time window as narrow as 100 ms. To reconcile apparent controversy between these results and early views, multiple levels of position processing and differences in task strategy are discussed in relation to object localization from impoverished visual inputs.
Meeting abstract presented at VSS 2012