Cailey Tennyson¹, Jeremy Wolfe^1,2; ¹Brigham and Women's Hospital, ²Harvard Medical School

Observers can process about 4-6 items in the vicinity of a single fixation. In the present experiment, observers moved their eyes to a fixation point that appeared at different locations on a screen. After 300 msec, a ring of 8 stimuli was flashed around the point and participants had to indicate the orientation of a target object using a 4AFC decision. Exposure duration was staircased (3-up, 1-down) to produce ~25% errors. In prior work, more errors fell on the vertical meridian than horizontal. Of more interest, individual patterns of errors showed significant idiosyncratic deviations from this average pattern. Previous studies involved targets defined by shape – TvL, hammers among tools. In the present work, the target was defined solely by color; blue rectangles among distractors of yellow and red. Observers indicated the 4AFC orientation of the blue rectangle. Participants ran 400 trials for each task, alternating tasks in 200 trial blocks. Using a TvL task, we replicated past findings with most observers showing idiosyncratic deviations from the average result. These idiosyncratic patterns were consistent between sessions with an average r=0.71 correlation. Interestingly, results for the color task were quite different. The average pattern of errors was different from the TvL pattern with most errors in the lower left. For individuals, color search produced different error patterns from their TvL search, with only r=.12 correlation between the two tasks. For each observer, the two sessions of color search were not well correlated with each other. Finally, there was no evidence of idiosyncratic patterns in the color task. The differences between tasks might be related to task difficulty. The staircased threshold for the color task was much shorter than for the TvL task, indicating that the color task was easier. Alternatively, patterns of attentional deployment might differ between tasks involving different features.

Acknowledgements: NEI EY17001