Using phase information to reveal the nature of event-related desynchronization.

It is common knowledge among brain researchers that certain events, such as motor imagery, can elicit event-related synchronization/desynchronization (ERD/ERS) of neurons in the brain, and that these phenomena manifest themselves in terms of band power changes in brain signals. There are, however, also important phase related changes in the brain signals associated with ERD/ERS which we investigate here. We quantify for the first time the “delta-phase” component of the Complex Band Power (CBP) features introduced in [G. Townsend, B. Graimann, G. Pfurtscheller, A comparison of common spatial patterns with complex band power features in a four-class BCI Experiment, IEEE Trans. Biomed. Eng. 53 (2006) 642–651] and [G. Townsend, Brain Computer Interfaces: Phase Features and Post Processing Issues, Ph.D. Thesis, 2006, p. 97]. Although this component, which makes explicit use of phase, is known from these earlier studies to provide improved classification results in a brain–computer interface (BCI), this phenomenon has never been investigated purely in the context of a physiological phenomenon. The importance of this study is to fill this gap in our knowledge by providing an analysis and quantification of this new feature so that an improved understanding of its nature will allow further exploitation of this feature to further improve BCI systems. Given that the primary use of BCIs is as an assistive technology for the disabled, the importance of this work to society cannot be overstated. Changes in band power in the context of ERD/ERS have been investigated in the past using a visualization tool called an ERD/ERS map [B. Graimann, Movement-Related Patterns in ECoG and EEG: Visualization and Detection, Ph.D. Thesis, 2003, p. 107]. We present and discuss here for the first time, an equivalent tool called the delta-phase map used to visualize the occurrence of delta-phase in different frequency bands and over the time courses of motor imagery trials. In this study, the delta-phase feature is shown to be proportional to frequency and is quantified and presented in that context for clarity. It was found that the most prominent ERD activity occurring in the frequency bands studied here are correlated to the activity captured in the delta-phase maps. The underlying nature of ERD/ERS is explored based on evidence provided by the delta-phase feature.