NOISE-INDUCED FIRST-ORDER PHASE TRANSITIONS IN CHAOTIC BRAIN ACTIVITY

Publisher Website
Google Scholar
Abstract
Brain electrical activity in animals during normal behaviors has aperiodic wave forms suggesting its origin in chaotic dynamics. Attempts at finding experimental proofs using low-dimensional, deterministic chaotic models have not succeeded. The assumptions of autonomy, stationarity,and noise-free operation that are needed to define these at tractors and their embedding dimensions have been shown not to hold for brains, because numerical estimates of correlation dimensions and Lyapunov exponents have failed to converge to normative values. Analysis of EEGs from sensory cortices show that a very different model applies to brains, which is more closely related to lasers than models of twist–flip maps and reaction–diffusion systems. Neurons interact with each other through channels with nonlinear, amplitude-dependent gains, by which they form fields of white noise. When the strengths of interaction are increased by input from receptors, the neurons interact more strongly and lose their high degrees of freedom. The macroscopic constraint on their activity appears as an order parameter in the form of spatially coherent, indeterministic chaos.