Abstract
Humans can walk without vision to previewed targets (2.5-5 m distant) without large systematic error and with near perfect sensitivity to target distance, even when targets are briefly glimpsed. However, high levels of performance with brief viewing durations are observed primarily with floor-level targets, where the angular direction of the target provides a reliable source of information about distance. The present study determined the roles for visual cues (angular declination) and ocular cues associated with gaze direction in the extraction of distance to briefly viewed targets. In Experiment 1, an eyetracker was employed with a variable fixation point to ensure gaze was directed toward the target on every trial. The viewing duration was either brief (approximately 19 ms, based on a detection threshold procedure) and followed by a mask or participants saw only the mask. In the latter case, participants were led to believe the viewing duration was below threshold and were instructed to estimate distance to unseen floor-level targets based on the knowledge that they were looking directly at the object at the onset of the trial. Surprisingly, participants responses were more sensitive (slope = 1.06 versus 0.84) and less biased (-17% versus -22%) when they reported not seeing the object. In Experiment 2, distance estimates based purely on gaze direction were less precise but not less accurate than those observed in Experiment 1. Experiments 3 and 4 compared performance for briefly viewed targets as a function of target eccentricity at the moment of the glimpse. Peripheral targets were not less accurately localized than those that were directly fixated. The overall pattern of results suggests a dominant role for angular declination in the localization of briefly viewed targets regardless of gaze direction but also a role for ocular cues that does not depend on the availability of visual cues.
Meeting abstract presented at VSS 2013