Sensory representation by non-equilibrium flows of neural activity at the mesoscale of cortex

A minimal non-equilibrium model of sensory cortex links spontaneous on–off toggling of coupled cortical columns to sudden large-scale activity-flow patterns that represent sensory inputs and perceptions.

• Why it matters: Existing sensory theories often assume equilibrium dynamics, but perception shows spontaneous “reversals” in perceived appearance across visual, auditory, and tactile stimulation.
• What they did: The authors use a Maximum Caliber (maximal path entropy) variational framework with dynamic Hamiltonian functions, constraining a two-area cortical-column model by empirical reversal behavior.
• The result: Effective representation depends on small-number fluctuations of toggling columns, and the framework yields microscale spiking predictions while implying perceptions can be rapidly revised.

AI summary of the abstract · may contain errors · not medical advice · read the paper.