Abstract
Successful interaction with complex scenes requires the ability to switch efficiently from one visual behavior (searching for an item) to another (memorizing location of an item). Though it is established that task-set influences the location and timing of eye movements during scene perception, the impact of switching task-sets on visual behavior is less clear. For instance, whereas manual responses are associated with large performance costs when switching tasks, oculomotor responses often are not associated with a residual cost and, in fact, can even show a paradoxical switch-benefit with sufficient preparation (Hunt & Klein, 2002). The purpose of the present study was to examine the influence of task-switching on saccade amplitude and fixation duration while viewing natural scenes. Scenes were colored photographs of forests and mountains. Participants were cued to perform discrimination (mountain/forest) and search (man-made-object present/absent) tasks in a mixed-task block, with either a 300 or 900 ms preparation interval. Switch-cost asymmetries were observed in the RT and saccade amplitude data such that large switch (and residual) costs were observed for the discrimination task, whereas no switch costs were observed for the search task (a switch benefit was observed for both outcomes at the 900 ms interval). Switch costs on fixation duration were observed for both tasks. However, whereas there was a residual cost associated with discrimination, there was not a residual cost associated with search. Finally, a multilevel multivariate difference score analysis indicated that fixation duration and saccade amplitude significantly predicted RT performance such that longer durations and smaller amplitudes were associated with better performance on the discrimination task, whereas the opposite was true for the search task. Interestingly, the magnitude of the amplitude effect on performance was larger than the fixation duration effect. Taken together, these results suggest that switch costs may be offset by improved saccade control.
Meeting abstract presented at VSS 2013