Cross-correlation analysis of neuronal activities.

Nerve impulses are generally regarded as spike train and analyzed by use of various kinds of so-called time series analyses. Cross-correlation analysis is used to reveal temporal and/or spatial relationships between more than two spike trains in the neuronal circuit, in which two neurons are synaptically connected or two neurons are independent but receive a common input. Three main characteristics of discharge dependency between two neurons become clear from the cross-correlation histogram: 1) the direction of dependency; 2) the latency of interaction; and 3) the type of functional connections between the neurons. We reviewed the cross-correlation analysis from these points of view: 1) what is cross-correlation analysis; 2) what are primary and secondary effects; 3) what kinds of functional information can be obtained from the shape of the primary effect; 4) how sensitively cross-correlation can detect neuronal interaction; 5) how to express quantitatively the primary effect; 6) what kinds of extension technique are available from cross-correlation; and 7) research trends using cross-correlation analysis.