Physical Society Colloquium

Interview for Faculty Position

Causal Entropy as a Measure of Temporal Relationships and Direction of Information Transfer in Neural Systems

Rhonda Dzakpasu

Department of Physics
University of Michigan

There are billions of neurons in the brain, each of which participates in the execution of various functions. How does the brain organize the operations of these fundamental units? What relationships exist, on a temporal scale, between neurons? Is there an ordering between the temporal patterns of neurons? In order to begin to address these questions, we have developed a novel analytical tool that measures temporal interdependencies between coupled neurons. The technique involves the real time monitoring of inter-event intervals between the coupled neurons. We demonstrate the feasibility of the measure on a mathematical model consisting of two, coupled non- identical Hindmarsh�Rose models of thalamo-cortical neurons. We show that the measure may be better than more conventional methods at detecting changes in asymmetrical temporal patterns. Finally, we demonstrate how the technique can be modified to study networks of coupled neurons and discuss the application of the measure in the analysis of experimental data.