Abstract
An important question in scene perception research is the degree to which individual fixation durations reflect perceptual and cognitive processing. The present study examined this issue by measuring the effect of moment to moment changes in the spatial frequency content of presented scenes on fixation duration. Eye movements were monitored while subjects performed a free-viewing memorization task. During saccades prior to critical fixations, scenes changed to low-pass filtered or unfiltered versions of the original image. The image reverted back to its original unfiltered form during the saccade following the critical fixations. Results from a linear mixed effects model showed significantly longer fixations as the changed image became more degraded. The unfiltered image produced significantly shorter fixation durations than any of the filtered images. These results are the first to show an immediate and monotonic relationship between fixation duration and the spatial frequency content of the scene during that fixation, and are consistent with previous research showing a similar effect of luminance on fixation durations (Henderson, Nuthmann & Luke, in press). We outline how these results are consistent with the predictions of the CRISP model of saccade generation in scenes (Nuthmann, Smith, Engbert & Henderson, 2010), which proposes that difficulty in moment-by-moment visual and cognitive processing of the scene modulates fixation durations.
Meeting abstract presented at VSS 2013