Purchase this article with an account.
Laurent Itti, Masatoshi Yoshida, David Berg, Takuro Ikeda, Rikako Kato, Kana Takaura, Tadashi Isa; Saliency-based guidance of gaze in monkeys with unilateral lesion of primary visual cortex. Journal of Vision 2008;8(6):992. doi: 10.1167/8.6.992.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
We investigate residual visually-guided behavior in monkeys after unilateral ablation of primary visual cortex (area V1), to unravel the specific contributions of V1 to salience computation in the primate brain. We analyze eye movements (108,458 saccades) of six macaque monkeys (three normals, three with unilateral V1 ablation) watching ∼54 minutes of television and other natural video stimuli (97,051 video frames). A computational model of saliency-based, bottom-up attention quantifies the monkeys' propensity to attend to salient targets in their normal vs. lesioned hemifields. To eliminate stimulus biases, we randomly presented all video clips twice, original and horizontally flipped. All monkeys are attracted towards salient stimuli, significantly above chance (assessed by simulating random saccades), for saccades directed both into normal and lesioned hemifields (t-tests, p[[lt]]0.0001 or better). We compute a saliency deficit score (SDS) comparing the extent to which saccades directed into one hemifield may be more or less saliency-guided than saccades directed into the other hemifield (SDS=0% would indicate no asymmetry, SDS=100% would indicate complete blindness in the lesioned hemifield). For the lesioned monkeys, SDS is significantly different from 0% (t-tests, p[[lt]]0.00001 or better) but surprisingly low (range 9% – 28%; average for normal monkeys: 4%), indicating a significant residual attraction towards salient targets even in the lesioned hemifields. Further, the recent history (up to 500ms before saccade onset) of saliency values around saccade targets reveals significantly increased salience just before saccades into the normal, but not the lesioned, hemifield (Bonferroni-corrected t-tests, p[[lt]]0.05 or better). Taken together, our results suggest that unilateral ablation of V1 does not abolish the natural tendency of monkeys to gaze towards salient targets during natural vision, although it significantly decreases the monkeys' ability to rapidly select targets in the lesioned hemifield that have recently become salient.
This PDF is available to Subscribers Only