Characterizing the precise dynamical balance of visual perception

• Applicant: Professor Jochen Braun
• Subject Area: Cognitive, Systems and Behavioural Neurobiology ; General, Cognitive and Mathematical Psychology ; Bioinformatics and Theoretical Biology ; Sensory and Behavioural Biology
• Term: from 2013 to 2017
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 247495463

The proposed project aims to confirm and extend a novel approach to multi-stable perception. Superficially, the dynamics of multi-stable perception is highly diverse in different observers and in different situations but, fundamentally, it is consistent with a precise balance of competition, adaptation and noise, as our published preliminary work has shown. Several predictions flow from this observation, which the proposed project will test.The project combines detailed psychophysical observations of multi-stable perception (with more statistical measures than in other studies) in normal adults, in adolescents, and anorexia nervosa patients, with computational analyses to infer the precise dynamical operating regime of each individual. Specifically, we wish to test our central working hypothesis (precise balance of competition, adaptation, and noise) in four project parts:Part A: Do experimental manipulations of this balance shift the dynamical operating regime into the predicted direction?Part B: Does perceptual stability increase with decreasing sensitivity to input modulations (and vice versa), as computational theories of perception predict?Part C: Do the results of the computational analysis generalize to other mathematical formulations of the dynamical model?Part D: Is the dynamical operating regime of diagnostic value, that is, does it reveal meaningful differences in the dynamical balance of different individuals?In fact, there are good theoretical reasons to expect visual perception (like any other process based on statistical inference) to operate in a precise dynamical balance. The larger significance of the proposed projects is threefold:(i) It will establish whether or not the observed balance really is the balance predicted by theory, namely, a trade-off between the stability and the sensitivity of perceptual inference.(ii) It will establish whether or not the observed dynamical balance is generated by a multi-stable attractor dynamics (balance of competition, adaptation, and noise), rather than by some other mechanism generating itinerant dynamics.(iii) It will establish whether or not the precise dynamical balance of individual observers is of diagnostic value, either over the course of development or in the context of neurological disorders.In conclusion, the proposed project will establish a novel, quantitative, and empirical method to test and refine important theoretical ideas such as the Bayesian brain hypothesis [Knill and Pouget, 2004] or the free-energy principle [Friston, 2010].

• DFG Programme: Research Grants