Classifying Hypnotizable Groups Using EEG Weighted Regional Frequency

Abstract

Determination of hypnotizability is important, before prescribing any hypnotic treatment. Existing methods for measuring the level of hypnotic susceptibility are subjective, with some problems. In this study, a feature based on EEG weighted regional frequency was introduced, which can characterize the level of the subject's hypnotizability objectively. The ability of this feature for making a signi cant di erence between three hypnotizable groups at the end of hypnotic suggestion was shown using statistical analyses. This feature was calculated based on the empirical mode decomposition method and the Hilbert transform. The EEG signals that were used in this study were recorded during hypnotic suggestion from 32 subjects. A K-nearest neighborhood-based classi er was designed for classi cation of the hypnotizable groups. The performance of the classi er was validated using the leave-one-out method, which showed the mean error of 3.13% in determination of the subject's hypnotic susceptibility level. This evaluation and obtaining the error were done by comparing the new method's results with the score of hypnotizability that was determined for each subject, using the subjective Waterloo-Stanford criterion. The new method, as opposed to common subjective clinical methods, represents a real time and objective procedure for determining hypnotic susceptibility.