Sonification of EEG for Anomaly Detection

Abstract

The EEG is a technique which is utilized to monitor the electrical action of the brain. The communication between brain cells is through these electrical signals. An EEG is used to detect the possible issues which are associated with this electrical activity because normal activity makes decipherable patterns. One of the most important uses of EEG is to identify seizure, because seizure causes the abnormalities in the EEG wave forms. EEG is the one of basic investigative checks for seizure. It moreover plays a vital job in investigating other brain problems. EEG is the visualization technique of electrical action of brain, which is the simplest way to make information understandable to humans. An alternative way is sonification. We have done the sonification of EEG signal to differentiate between seizure and non seizure. Sonification gets the data as input and generated audio signals. It is the process in which data is converted into sound, which provides an auditory option instead of visually analyzing the data, which will be very helpful even for that listener which doesn’t have any training or knowledge about the seizure detection. We assumed that humans can easily distinguish among seizure and non seizure by listening the sound of EEG created using our algorithm. Our algorithm reads the data done the processing and creates and plays the created sound immediately. Real time EEG data is used for this system obtained from CHB-MIT website. We used data of single channels as well as of multiple channels and tested the potential of our algorithm. Our system will work effectively for real time EEG monitoring, neuro-feedback etc