Approximate entropy is a method which provides a model independent nonlinear measure (the index ApEn) of the “regularity” of the process generating a time-series. In recent years, ApEn has been vigorously employed in the study of several biological signals, but only a few applications in the analysis of vocal disorders have been proposed. Here, we investigate the potential usefulness of ApEn in the study of electroglottography and microphone signals in normal and dysphonic subjects. Results show that statistically significant ApEn differences between the two groups can be found, more easily detectable in the microphone signal case.

Index Terms. chaos, time-series, signal processing, vocal disorders