A new method for relating behavior to neuronal activity in performing monkeys

Abstract

A new method is presented for the analysis and interpretation of data from single neuronal recording in behaving monkeys. Most current methods relate changes in neuronal activity to behavior by constructing rasters and histograms centered on specific previously defined behavioral events whose relevance to neuronal activity cannot be determined beforehand. As a consequence false conclusions may result. The new method aligns rasters and histograms of neuronal activity relative to patterns inherent in the neuronal activity irrespective of the animal's behavior. The behavior is then related to the neuronal activity. This frees interpretation from a priori notions as to the relevant behavior. The method also determines the variability of temporal latencies between changes in neuronal activity and any behavioral event on an individual trial by trial basis. With the assumption that the latencies will be least variable between the change in neuronal activity and the most relevant behavioral event, the relatedness of neuronal activity changes to behavior can be tested so that physiological conclusions can be inferred. The same method and subsequent analyses can be applied to electromyographic activity.