Abstract
Making eye-movements provides us with visual samples of the world that must be integrated to produce a coherent percept of the visual world. Transsaccadic Memory (TSM) is necessary for storing samples from previous fixations so they can be integrated with the current one. Visual search tasks place greater demands on TSM than change-detection since visual search often requires integration of multiple samples. Our goal was to estimate TSM capacity and investigate its relationship to visual short-term memory (VSTM) using two visual search tasks. Prior to the main experiment, participants completed a forced-choice detection task designed to estimate their psychometric function at each potential target location. Participants then completed a simulated or real saccade search task, requiring the localization of a target signal (Gabor) embedded in a field of 1/f filtered noise. Participants were presented with one, two or four fixation intervals and the target was present in each interval (redundancy condition) or only one (uncertainty condition). Trials were blocked by the number of intervals and target presence condition. In the real saccade task participants made a 5° saccade following each stimulus sample; in the simulated saccade task they maintained fixation while we simulated the dynamic transient produced by saccades. Performance was measured as target localization accuracy. We developed an ideal observer model of transsaccadic integration that takes into account the critical role of TSM. Our ideal observer model allowed us to place a lower bound on TSM capacity for each subject, quantified in terms of bits, a task independent unit of information. We found median capacity estimates of 7 bits for the simulated saccade task and 9.5 bits for the real saccade task. These estimates are somewhat higher than those previously found for VSTM, but they indicate that TSM capacity may play a limiting role in multiple fixation visual search.
Meeting abstract presented at VSS 2015