Looking versus seeing, peripheral versus central vision, and top-down feedback from higher to lower visual areas

Looking selects a fraction of visual inputs for the attentional bottleneck; seeing recognizes the objects in the selected inputs, typically brought by eye movements from peripheral to central visual field. Hence (Zhaoping 2014), peripheral and central vision are better at looking and seeing, respectively. Zhaoping (2017) hypothesized that, using analysis-by-synthesis for better object recognition (seeing), top-down feedback from higher cortical areas to the primary visual cortex (V1) is weaker or absent in peripheral visual field. Accordingly, peripheral vision is more easily fooled by feedforward inputs from V1, is more prone to illusions and vulnerable to crowding, and conveys mainly ensemble or summary input statistics. For example, V1 neurons respond to binocularly anti-correlated random-dot stereograms as if the input binocular disparities reversed their signs; peripheral but not central vision perceives the reversed depths in such stereograms (Zhaoping & Ackermann 2018)