Abstract
Abstract: :
Purpose: Knowing where one is located within an environment requires the combination of egocentric information as well as knowledge of the scene’s spatial structure. In a recent mobility study we showed that persons with a loss in peripheral visual field utilized egocentric information to the same extent as fully sighted subjects but performed poorly in conditions requiring memory of the scene’s spatial structure. Without the ability to simultaneously view a wide spatial area people must rely on an internal representation of the scene. In this study we determined the visual field size at which navigation can be performed at normal speed (visual span). We also determined whether fast walkers make better use of an internal representation of the scene to guide gaze and navigation and, as such, have smaller visual spans. Methods: Fifteen fully sighted subjects walked to a target tree in a forest scene presented within an immersive virtual environment. A gaze–contingent display was used to restrict visual field diameter to 10°, 20°, or 40°. Presentation order of window size was counterbalanced across subjects, and 15 trials per size were tested with 45 forest scenes. Time to walk to the target was determined, and eye movement parameters were calculated. Subjects were classified into "fast" and "slow" walkers based on their baseline travel time relative to the median. Visual span was defined as the window size corresponding to a 20% increase in baseline time and was calculated from the mean log time–vs–size functions. Results: The visual span for fast walkers (23.4°) was significantly smaller than that of slow walkers (32.6°), p=0.02. As window size decreased, slow walkers made more fixations than fast walkers, p=0.01. With a 10° window, slow walkers made twice as many fixations as the fast walkers compared to only 1.4 times as many with a 40° window. Slow walkers and fast walkers did not differ in fixation duration but both groups showed an increase with decreasing window size, p=0.001. Average duration was 0.23 s and 0.30 s for window sizes of 40° and 10°, respectively. Conclusions: In navigation, people can tradeoff internal and external information about a scene’s spatial structure to guide where to look (sampling strategy) and where to walk. In situations where the visual field is obstructed, or when people have a loss in peripheral visual field, efficient navigation can be maintained by using memory of the scene instead of making multiple fixations. These findings have implications for the design of orientation and mobility rehabilitation strategies of the visually impaired.
Keywords: vision and action • low vision