Listening to the Mind Listening: Sonification of the Coherence Matrix and Power Spectrum of EEG Signals

In this paper, we present our strategy for sonification of EEG signals as part of the ‘listening to the mind listening’ project [1]. It is generally thought that similarities between EEG signals recorded at two distinct points on the scalp indicate that the brain is performing the same task at these two points. A high correlation coefficient can be used as an indicator of such similarities. Using EEG data supplied by project coordinators, we have created a coherence matrix representing levels of similarity between 26 electrode signals. A new coherence matrix was calculated for every one second time frame of the EEG data. In order to focus on the true significance of the data some pre-processing was applied on the original EEG data before this matrix was created. Our sonification approach then aimed at representing the correlation coefficients of the coherence matrix. From the the EEG data we also extracted and sonified the level and position of delta, theta, alpha and beta waves which indicate various states of brain activity (e.g. alertness, sleep, relaxation, etc.). Alpha waves (7-13Hz) were most prominent in the data. Alpha activity corresponds to a relaxed state suggesting that the subject was most probably in a stationary sitting position with eyes closed while the EEG recording was being made. We were also able to detect other brain activity in the delta, theta and beta frequency bands and locate regions of this activity. Our sonification combined activity in these frequency bands with other aspects of the coherence matrix. We decided to use only the 26 first electrode signals in order to focus directly on activity within the brain. We first explain the need for data pre-processing. We then briefly overview our sonification software architecture, and finally present our sonification strategy for the coherence matrix and brain activity in the four low frequency bands.