Quantifying cognitive state from EEG using phase synchrony

Abstract

Phase synchrony is a powerful amplitudeindependent measure that quantifies linear and nonlinear dynamics between non-stationary signals. It has been widely used in a variety of disciplines including neural science and cognitive psychology. Current time-varying phase estimation uses either the Hilbert transform or the complex wavelet transform of the signals. This paper exploits the concept of phase synchrony as a mean to discriminate face processing from the processing of a simple control stimulus. Dependencies between channel locations were assessed for two separate conditions elicited by distinct pictures (representing a human face and a Gabor patch), both flickering at a rate of 17.5 Hz. Statistical analysis is performed using the Kolmogorov-Smirnov test. Moreover, the phase synchrony measure used is compared with a measure of association that has been previously applied in the same context: the generalized measure of association (GMA). Results show that although phase synchrony works well in revealing regions of high synchronization, and therefore achieves an acceptable level of discriminability, this comes at the expense of sacrificing time resolution.