Abstract
Purpose: :
The natural courses of most debilitating retinal diseases do not usually demarcate the visual field into seeing and unseeing areas. A transition from relative to absolute scotoma is common, though the spatial extent of the transition may vary greatly depending on disease pathology. The effects of the edge profile of a central scotoma on visual search performance and on adaptation to foveal input deprivation were investigated using gaze contingent display simulation.
Methods: :
A head-mounted, 500Hz eye tracker was used to simulate a 10 deg circular central scotoma. Alpha blending technique was used to create two scotoma profiles, one with a sharp change from seeing to unseeing (SCS) and the other with a gradual, Gaussian transition (GCS). The half-height diameter of the GCS was equal to the diameter of SCS. Ten normally-sighted subjects with no known oculomotor abnormalities, 5 for each type of scotoma profile, practiced a search task (finding an "O" target among "C" distractors) at 3 set sizes (1, 8 and 32) for 11 sessions (162 trials per session). Search reaction time (RT) and eye movement data were collected.
Results: :
When first exposed to the simulated scotomas, subjects took 1.5 times longer to complete search and made 1 time more fixations than without scotoma, though the initial impacts of scotomas were similar on both groups. Repeated measures ANOVA with Session as the within-subjects variable and Scotoma Profile (SCS/GCS), Set Size (1, 8, 32) and Trial Type (positive/negative) as between-subjects variables showed the following highly significant differences between the two scotoma profiles. SCS searched faster and made fewer fixations than GCS (RTs 4990 vs 5296 ms; numbers of fixations 15.35 vs 17.69). SCS made faster and larger saccades than GCS (velocities 253 vs 223 deg/sec; amplitudes 6.27 vs 5.85 deg,). Saccade and fixation durations were not significantly different between the two groups (saccade durations 42.8 vs 43.3 ms; fixation durations 339 vs 328 ms). Search performance of SCS improved with practice and saturated after 6-7 sessions while GCS underwent a slower improvement and showed no sign of saturation at the 11th session. The difference in adaptation was significant.
Conclusions: :
A more abrupt transition from seeing to unseeing may introduce highly noticeable visual dynamics when it moves across visual boundaries of objects, providing salient information about the location and extent of the unseeing region and helping guiding more purposeful eye movements. A gradual transition, corresponding to an absolute scotoma surrounded by extended relative scotoma, may not offer such information. This lack of awareness of the scotoma may underlie difficulties some low vision patients experiencing in rehabilitation.
Keywords: low vision • visual search • eye movements