Jochen Braun: Stochastic dynamics of visual inference

Otto-von-Guericke University Magdeburg

Sensory inputs are volatile and unpredictable and a reliable inference of underlying causes is hard. Neural activity continuously accumulates and compares sensory evidence as perceptual decisions are reached. Psychophysical evidence from multistable perception, properly interpreted, elucidates many details of this stochastic dynamics. Astonishingly, this evidence implicates representations with multiple bistable elements (similar to an Ehrenfest process). I now show that this kind or representation provides clear functional benefits for volatile and unpredictable signals. It seems that bistable elements offer a biophysical realization of non-conventional estimators, which are known to outperform conventional estimators (such as sample mean) for small sample sizes. I conclude that a stochastic dynamics with multiple bistable elements has surprising explanatory power with regard to both cortical activity dynamics and to improving sensory function in a complex world.

Cao, Pastukhov, Mattia, Braun (2016) Collective activity of many bistable assemblies reproduces characteristic dynamics of multistable perception. J. Neurosci.,36: 6957-72.

Veliz-Cuba, Kilpatrick, Josic (2015) Stochastic models of evidence accumulation in changing environments. SIAM Review, 58: 264-89.

Catoni (2012) Challenging the empirical mean and empirical variance: a deviation study. Annal. Inst. Henri Poincare - Prob. Stat., 48 (4): 1148-85.

 

 

Part of the seminar series on "Current Topics in Computational Neuroscience"

 

Organized by

Tilo Schwalger / Margret Franke

Location

BCCN Berlin, lecture hall, Philippstr. 13 Haus 6, 10115 Berlin



Go back