Thomas Hermann's research
on Sonification, Data Mining & Ambient Intelligence
bar shadow

eeg sonification

This website delivers information about a very promising application of auditory display: the sonification of electroencephalography data.

News & Events

Why sonification of eeg data?

EEG, the measurements of surface potential differences between electrodes attached to certain positions on the human scalp, is a well established techniques to gain insight into brain activity and to study cerebral processes, from deseases to cognitive processing. The data, however, is of a type which is notoriously difficult to understand: it is spatially indexed multi-channel time series with lots of noise, and, due to the cerebral functioning, a coding where spectrally resolved activity is relevant for meaningful interpretation. There is a sensor which is good at coping with temporally structured, spectrally organized, spatially indexed, and noisy data in an excellent fashion: the human ear and listening system. It is worth an intensive trial to challenge our listening skills with the task of extracting the same (or even more) information from eeg signals than we typically do by visial inspection.
However, although eeg data can directly be played as sound (a technique called audification), sonification enables to render more differentiated acoustic shapes from eeg data to better match pecularities of our listening sense. The author develops since severals years (mainly in cooperation with Gerold Baier) techniques for sonification of eeg, which we plan to explain and demonstrate on this website in short time.

Towards a new "Auditory Medicine" and Listening culture

Although the above motivation was specifically taylored to eeg, it can be generally seen that many data types in biomedical research are of similar type: structured in many regards, rich, complex, noisy, unknown semantics. Particularly researchers in biomedicine and physisians are open-minded to the application of listening to characterize deseases (think of the stethoscope and auscultation), and - thank to the recent advances in data mining, sound rendering, and human-computer interfaces, we are now able to make new advances towards "intelligent stethoscopes", that are versatile in converting the data into sound so that specific tasks are ideally supported by exploiting our listening skills.

What is the tradition of listening in medicine?

Where to learn more about the field

A good starting point (besides studying ICAD papers on this matter) is to follow our Tutorial on EEG sonification and sonification of other biomedical data, held at ICAD 2006, London, which we have prepared as pdf slides and mp3 files.