Georgia Institute of Technology College of Sciences

Understanding the computations performed by neuronal circuits requires characterizing the strength and dynamics of the connections between individual neurons. This characterization is typically achieved by measuring the correlation in the activity of two neurons through the computation of a cross-correlogram or one its variants. We have developed a new measure for studying connectivity in neuronal circuits based on information theory, the incremental mutual information (IMI). IMI improves on correlation in several important ways: 1) IMI removes any requirement or assumption that the interactions between neurons is linear, 2) IMI enables interactions that reflect the connection between neurons to be differentiated from statistical dependencies caused by other sources (e.g. shared inputs or intrinsic cellular or network mechanisms), and 3) for the study of early sen- sory systems, IMI does not require that the external stimulus have any specific properties, nor does it require responses to repeated trials of identical stimulation. We describe the theory of IMI and demonstrate its utility on simulated data and experimental recordings from the visual system.