V-VSS, June 1-2

Eye tracking evidence for V1 Saliency Hypothesis from an anomalous visual search behavior

Poster Presentation 83.547: Thursday, June 2, 2022, 6:00 – 8:00 am EDT, Video Chat
Session: Poster Session 4

Times are being displayed in EDT timezone (Florida time): Tuesday, August 9, 1:45 am EDT America/New_York.
To see the V-VSS schedule in your timezone, Log In and set your timezone.

Search Abstracts | VSS Talk Sessions | VSS Poster Sessions | V-VSS Talk Sessions | V-VSS Poster Sessions

Junhao Liang1,2, Severin Maher1,2, Li Zhaoping1,2; 1Eberhard Karls University of Tübingen, 2Max Planck Institute for Biological Cybernetics

Typically, humans find a target among uniformly oriented non-targets more quickly when this target is perpendicular, rather than parallel, to the non-targets. V1 Saliency Hypothesis (V1SH), which states that the primary visual cortex (V1) creates a saliency map to attract attention exogenously, predicts a special case in which exactly the opposite holds. Each visual item contains two equally-sized disks. Its orientation is that of the line connecting the disk centers. The target is a homo-pair of disks of the same color, black or white. Each non-target is a hetero-pair of one black and one white disk. The target is 45 degrees clockwise or counterclockwise from vertical, and parallel or perpendicular to uniformly oriented non-targets. With its on/off subfield of a receptive field activated by a white/black disk, a V1 neuron is more activated by a target/non-target parallel/perpendicular to its preferred orientation. Since a V1 neuron is more strongly inhibited by nearby V1 neurons tuned to similar orientations, it is more activated by a target parallel, rather than perpendicular, to the non-targets by being less suppressed, making the parallel target more salient according to V1SH. Our subjects had to press a button to report the homo-pair target’s location as quickly as possible. Report reaction times (RTs) were significantly shorter for targets parallel to the non-targets. Splitting each RT into components, RT_gaze and RT_lapse, for durations before and after the gaze reaching the target, this parallel advantage was only caused by RT_gaze for attentional shifts. In RT_gaze, the parallel advantage is stronger among trials in which the target-reaching saccade has a larger saccadic amplitude, consistent with the idea that saliency is stronger in peripheral vision. Our findings support the idea that the neural substrate for saliency is in V1 rather than higher brain areas which signal figure orientations of symbolic objects.

Acknowledgements: This work is supported by fundings from the University of Tübingen, Max Planck Society, and the Deutsche Forschungsgemeinschaft (DFG) with Projektnummer 465358224.