The statistical relationship between the "entropy" of a neuronal signal and its variability

Abstract

We used a serial order nonparametric analysis technique to analyze trains of neuronal action potential intervals in terms of classical information theory. We observed a marked correlation between the information theory descriptor, entropy, and several common measures of variability (median interval length, range of interval length, and skewness of the probability density function). The correlation between variability and entropy was accounted for in the following decreasing order: range of intervals, median interval, and skewness. These data suggest a significant relationship between the signal and its variability and entropy, when entropy is calculated by our serial-order pattern detection method.