IBANS mini-symposium presentation

To be presented at the Institute of Behavioral and Neural Sciences mini-symposium, 3rd October 2014. [Details]

What do statistics tell us about perception?
David Hunter, Paul Hibbard

Psychophysical experiments can tell us what the visual system perceives, physiological measurements can tells us how the early visual system processes information, but neither can explain why the visual system works that why. The idea that sighted organisms are attuned to regularities in the visual world has been around since(Barlow 1961). Simple rules governing local neural processing have had great success in explaining the functioning of early visual processing, decorrelation in the retina(Atick and Redlich 1992; Shlens, Field et al. 2006), and energy efficient coding in the V1 of the visual cortex(Olshausen and Field 1996). Surveys of the information content of visual scenes can give us valuable insight into the information that is available to the visual system. Using the problem of binocular depth as an example we found substantial qualitative agreement between statistical models and physiological measurement. We also found a substantial bias toward features of zero phy sical disparity, but different shapes (phases) and receptive fields in each eye. These results may provide the first evidence that the standard energy model of binocular disparity(Fleet, Wagner et al. 1996) is statistically efficient.