Purchase this article with an account.
Michael Dorr, Peter Bex; Peri-Saccadic Visual Sensitivity while Freely-Viewing Natural Movies. Journal of Vision 2011;11(11):517. doi: https://doi.org/10.1167/11.11.517.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
We now have a relatively good understanding of contrast sensitivity for narrowband gratings in observers who maintain stable fixation. However, in everyday vision, humans make several eye movements per second, and the time-varying input to the visual system comprises a very broad distribution of spatio-temporal frequencies, orientations, and intensities. We used a real-time gaze-contingent display to locally change contrast in high-resolution naturalistic (blu-ray) videos during free viewing. Every video frame was decomposed into its spatial frequency bands on a Laplacian pyramid in less than 1 ms; eye movements were recorded at 1000 Hz and local spatially narrowband contrast modifications in retinal coordinates were applied at the screen refresh rate (120 Hz) and a typical latency of 14–22 ms.
In video watching sessions lasting about 10 minutes and aproximately every 2 s, energy in the 1.5–3 cyc/deg band was increased for 600 ms in a 2 × 2 deg spatio-temporal raised Gaussian windowed patch at one out of four locations, 2 deg away from the point of regard.
The spatio-temporal structure at test locations and the latency between stimulus and saccade onsets were analyzed offline. Localization performance overall was low compared with gratings; was best for stationary eyes and worst when a saccade was initialized c. 500 ms after target onset (“late”).
This effect could only be partially explained by differences in stimulus contrast and motion that modulated sensitivity and were correlated with eye movement activity. A second experiment replicated the retinal input of the first experiment by shifting the movie on screen according to the previously recorded gaze, while subjects maintained central fixation. Even though the retinal input was the same, simulated “late” saccades had a less detrimental effect.
Contrast sensitivity is profoundly changed by eye movements and natural contexts and is thus poorly characterized by measurements obtained with traditional methods.
This PDF is available to Subscribers Only