A Computational Dual-Process Model of Fixation-Duration Control in Natural Scene Viewing

Abstract Research on eye-movement control during natural scene viewing has investigated the degree to which duration of individual fixations can be immediately adjusted ongoing visual-cognitive processing demands. Results from several studies using fixation-contingent quality paradigm suggest that timing adapts stimulus changes occur a fixation-to-fixation basis. Analysis fixation-duration distributions revealed saccade-contingent degradations and enhancements have two qualitatively distinct types influence. The surprise effect begins early in fixation is tied surprising visual events such as unexpected changes. encoding difficulties occurs relatively late within fixation. Here, we formalize an existing descriptive account these effects (referred dual-process account) by stochastic simulations. In computational model, related influences are implemented time-dependent rate at saccade programming completed critical fixations. model was tested data experiments luminance image either decreased or increased selected (Walshe & Nuthmann, Vision , 100, 38–46, 2014). A counterfactual method used remove components identify their specific influence distributions. results provides good for luminance-change studies. We describe how simulations generalized explain diverse set experimental results.