Audiovisual Stimulation of Human Brain. Linear and Nonlinear Measures.
Svorad Štolc, Anna Krakovská, Michal Teplan · 2003
One of the most fundamental problems in brain signal analysis is an absence of reliable measures of certain brain states. For purpose of study and detection of such states (especially the state of the mental relaxation which is still not exactly defined) we decided to investigate the influence of commercially well-known mind machines (light & sound machines) on subject’s brain. In this paper we discuss changes in common linear and novel nonlinear measures applied on series of EEG signal recorded before and after mind machine sessions. We will namely discuss linear correlation, mutual information content, spectral entropy, spectral edge frequency and relative power of alpha, beta, delta and theta band. 1 Experiment description The experiment consisted of 25 series (days) of 20 minutes training with the mind machine in a darkened Faraday chamber. There were 6 adult volunteers participating in the experiment. The mind machine was supposed to train subjects ’ brain to relax more efficiently using light and sound stimulation. The mind machine affected audiovisual nervous sensors by decreasing frequency stimulation from 18 Hz to 2 Hz and back to 18 Hz at the end of each session (Fig. 1). Before and after each session the EEG signal was recorded from selected 6 channels according to International 10-20 system: F3C3, F4C4, C3P3, C4P4, P3P1, P4O2. The reference electrode was placed at CZ point and ground electrode at FPZ point. Each EEG record was 3 minute long what gives 90000 samples at the 500 Hz sampling frequency. All records with obvious artefacts or sleep occurrences were excluded from further analysis.