Purchase this article with an account.
Xilin Zhang, Li Zhaoping, Fang Fang; Neural activity in V1 creates the saliency map. Journal of Vision 2011;11(11):190. doi: https://doi.org/10.1167/11.11.190.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Zhaoping proposed (Li, 1999, 2002) that pre-attentive computational mechanisms in the primary visual cortex create a bottom-up saliency map. We performed psychophysical, ERP and fMRI experiments to test her theory. In these experiments, stimulus images have a regular Manhatten grid of 15 × 29 low-contrast bars, presented in the lower visual field. All bars were identically oriented except for a target region of 2 × 2 bars with another orientation in either the lower-left or the lower-right quadrant. There were four possible orientation contrasts between the target and the background −7.5, 15, 30 and 90°. To avoid top-down influences, each stimulus was presented for only 50 ms and was followed by a high-contrast mask, which rendered the whole stimulus invisible to subjects (confirmed by a forced-choice test). In the psychophysical experiment, the Posner cueing paradigm was adopted to measure the spatial cueing effect of the invisible target on an orientation discrimination task. We found that the cueing effect (the saliency of the invisible target) increased as the orientation contrast increased, but saturated at 30°, which can be predicted by Li Zhaopingâ's V1 model. In the ERP experiment, we measured the amplitude of C1 to the invisible stimulus with various orientation contrasts. C1 is the earliest visual evoked potential component and is generated in V1. In the fMRI experiment, we measured the BOLD response in V1 to the invisible target with various orientation contrasts. Both the amplitude of C1 and the BOLD response in V1 were proportional to the orientation contrast. More interestingly, they were significantly correlated with the cueing effect across subjects. Thus, the overall results support that, in human subjects, neural activities in V1 could represent the bottom-up saliency map.
This PDF is available to Subscribers Only