Existence of Millisecond-order Stable States in Time-Varying Phase Synchronization Measure in EEG Signals

TL;DR

This study introduces a new measure for analyzing EEG phase synchronization, revealing stable millisecond-scale states called synchrostates that are consistent across trials and may reflect cognitive processing dynamics.

Contribution

The paper presents the discovery of millisecond-order stable synchrostates in EEG signals and demonstrates their potential as a new measure of brain connectivity stability.

Findings

Synchrostates are well-defined, stable for a few milliseconds during face perception tasks.

These states are consistently observed across multiple trials.

Switching sequences of synchrostates reflect underlying cognitive dynamics.

Abstract

In this paper, we have developed a new measure of understanding the temporal evolution of phase synchronization for EEG signals using cross-electrode information. From this measure it is found that there exists a small number of well-defined phase-synchronized states, each of which is stable for few milliseconds during the execution of a face perception task. We termed these quasi-stable states as synchrostates. We used k-means clustering algorithms to estimate the optimal number of synchrostates from 100 trials of EEG signals over 128 channels. Our results show that these synchrostates exist consistently in all the different trials. It is also found that from the onset of the stimulus, switching between these synchrostates results in well-behaved temporal sequence with repeatability which may be indicative of the dynamics of the cognitive process underlying that task. Therefore these synchrostates and their temporal switching sequences may be used as a new measure of the stability of phase synchrony and information exchange between different regions of a human brain.