Computational modeling of fixation durations in scene viewing

Eye-movement control during scene viewing can be represented as a series of individual decisions about where and when to move the eyes. Here, we propose a computational model to account for variations in fixation durations. First, we assume an autonomous (i.e., random) saccade timer; timing signals are modeled as random-walk diffusion processes. Second, difficulties at the levels of visual and cognitive processing can inhibit, and thus modulate, the timer. Third, saccade programming is completed in two stages: an initial, labile stage that is subject to cancellation, and an ensuing, nonlabile stage. These model assumptions are consistent with current evidence concerning basic oculomotor control. The model was successfully tested on experimental data. We conclude (1) that fixation durations are sensitive to moment-to-moment processing demands and (2) that viable models of gaze control in scene perception should be able to account for variations in fixation durations.