Michael A. Buice

Michael is a theoretical physicist with a background in non-equilibrium statistical mechanics.  His graduate work with Jack Cowan at the University of Chicago served to extend the Wilson-Cowan neural network equations to include the effects of fluctuations, so that they describe effects beyond mean-field approximations.  Following this, with Carson C. Chow at the National Institutes of Health in the Laboratory of Biological Modeling, he constructed a framework for describing the statistical mechanics of dynamical systems.  This framework enables a quantitative connection between probabilistic descriptions (which result in systems such as the Wilson-Cowan equations) and mechanistic ones.  An example is that one can generate an effective equation for a single neuron interacting with an unknown number of other neurons whose properties are not completely determined.  In his current position with the Fiete group he aims to explore the connections between information and dynamics, specifically as they regard the brain and nervous system.  A chief goal is to understand the role and mechanism for strong analog error correcting codes. 

Selected Publications:

Correlations, Fluctuations, and Stability in a Finite-size Network of Coupled Oscillators
Michael A. Buice, Carson C. Chow
Physical Review E, 76, 031118(2007)  arxiv:0704.1650

A Kinetic Theory of Coupled Oscillators
Eric J. Hildebrant, Michael A. Buice, Carson C. Chow
Physical Review Letters 98, 054101 (2007) arXiv:nlin/0612029

Field Theoretic Approach to Fluctuations in Neural Networks
Michael A. Buice, Jack D. Cowan
Physical Review E 75, 051919 (2007)