The ethology of saccades: A non-cognitive model (Q1113849)

It is proposed that there is a low-level automatic component of visual scanning in which saccades are triggered probabilistically in time by non-foveal stimulus features. By ignoring cognitive factors and memory, a first-order Markov approach is taken, which is tractable for spatially homogeneous stimuli. The predicted distributions of saccade magnitude, fixation duration, fixation position, and how they are related to stimulus size are in good agreement with empirical observations. Ways in which this low-level process could be modulated by cognitive factors are discussed.