A portable neurofeedback system and EEG-analysis methods for evaluation

Abstract

Recently, there has been much interest on neurofeedback (NF; EEG biofeedback) based on self-regulation of the human electroencephalogram (EEG) by the subject. Many experimental protocols concerning specific EEG frequencies have been introduced and NF has been successfully applied to various clinical practices including epilepsy, attention deficit hyperactivity disorder, alcoholic abuse, etc. In addition to clinical application, NF has also contributed to research of the EEG such as the introduction of sensorimotor rhythm. In our previous works, a portable NF system had been developed, which consists of a portable EEG amplifier, a headband with dry electrodes placed on Fp1 and Fp2 sites, and various types of training programs such as car-racing, arrow-shooting, meditation, etc. And an EEG-analysis program based on time-frequency analysis had also been developed for evaluating performance of NF training. Especially, eight quotients, designated as brain quotients, play an important role in evaluation process, which characterize behaviors of EEG rhythms effectively. In this work, we extended the previous one to nonlinear dynamical analysis of the EEG for a better evaluation of the effect of NF training on brain functions. For this purpose, we calculated the correlation dimension (CD) and the Lyapunov exponent (LE) indicating the complexity and the level of chaos of the EEG, respectively. We obtained distributions of CD and LE for age groups from 10s to 60s and correlations between the nonlinear and the linear (brain quotients) coefficients. We also discussed reliability of our portable EEG measurement system by comparing with a generally used system. Finally, we suggested a proper index reflecting attention by considering event-related potential components, alpha rhythm, and sensorimotor rhythm.