Auditory Display of Brain Oscillatory Activity with Electroencephalography

Auditory display is a relatively young field that has nonetheless been applied to many different types of research, including data exploration, musical composition and performance and for studying medical conditions such as epilepsy. Sonification, a type of auditory display which can generally be defined as the transformation of data into sound, has been applied to biofeedback systems, computer assisted collaborative work, and astronomy, to name just a few. Despite the considerable amount of recent work involving sonification and biofeedback, it is not clear what types of sonification strategies or mappings are most effective for aiding the perception of brain and body states. This master thesis aims to help answer this question by developing and testing a sonification engine capable of a range of sonification techniques, from audification-the simplest, most direct technique-to more arbitrary or musical mappings. In all, three techniques are presented: a relatively simple and direct approach, a second technique where events or features of the data modulate features of the output sound, and a third technique that could be described as more musical or arbitrary. This sonification system is built using Pure Data, an open-source graphical computing environment. Using the Emotiv EPOC Brain-computer interface headset and a complex chain of signal acquisition and processing software, we tested our sonification system in an experiment featuring 14 subjects. Each subject was brought into a quiet room and fitted with the EPOC device. They were then exposed to a total of 12 different sonification sessions, half of which were in real-time and half of which were placebos. The order of sonifications and placebos was randomized for each subject. Data was gathered using a questionnaire and the recording of physiological data taken directly from the BCI device. Both types of data were used to compare real-time techniques to placebos as well as to one another. While no significant findings were reported with regards to a our sonifications having a greater effect on subjects' state of relaxation, the data do conclude that our parameter-mapping sonification technique is the most relaxing, and thus a good candidate for further study.