Abstract
Purpose: :
Little is known about how the visual deficits due to the presence of a central scotoma affect higher-level visual processing. In this study, we assessed the impact of a simulated scotoma on the perception of biological motion, a higher-level visual task that involves inferring human actions from motion.
Methods: :
Observers (N=4) with normal vision watched short movie clips (1.3 s) of a walker rendered as a pointlight (walker made up of a few clusters of discrete joints), skeleton, contour or silhouette figure embedded in a field (7.1° x 10.6°) of dynamic white Gaussian luminance noise. A dynamic circular window (0, 7.1°, 9.4° diameter), containing pixels at the mean background luminance, was overlaid on the walker image to simulate the effect of a scotoma. The position of the simulated scotoma was stationary, or changed several times during a trial following a Poisson distribution with a mean "fixation" duration of 50 or 500 ms to simulate the oculomotor instability of patients with central scotoma. The maximum displacement of the simulated scotoma between "fixation" was 2.4° x 3.5° horizontally and vertically, or 3.2° x 4.8°. On each trial, observers indicated the presence or absence of the walker (detection task), or the walking direction -- rightward or leftward (discrimination task). Stimulus contrast threshold corresponding to 75% correct performance was measured for the two tasks and for the various simulated scotoma conditions.
Results: :
Averaged across the four observers, contrast thresholds for detecting the presence of the walker and for discriminating the walking direction were very similar, regardless of the size and the amount of displacement of the simulated scotoma. Thresholds were similar for the four renditions of the walker image and the various simulated scotoma conditions, except when the simulated scotoma was stationary, or for the 9.4° diameter simulated scotoma with a 500 ms averaged "fixation" period, when significant threshold elevations were observed.
Conclusions: :
The remarkably good performance in perceiving biological motion in the presence of a simulated scotoma (except for the most difficult conditions) suggests that human observers require very little information in the stimulus to perform the task (e.g. as long as the feet of the walker were present). The finding suggests that the missing stimulus information within a scotoma may not be sufficient to cause any deficits in the ability to perceive biological motion for patients with central scotomas.
Keywords: shape, form, contour, object perception • motion-2D • low vision