Our hearing is a complex system. Sound is encoded to a neural signal in our inner ear and transmitted via the auditory pathway to our primary auditory cortex in the brain. We can think of this part of our hearing as lower-level processing of the sound. The electrophysiological signals that we can measure from this part of our brain is basically driven by the physical attributes of the sound and is largely independent of our perception and attention. Attention, on the other hand, is a result of higher-order cognitive processes in our brain.
An auditory event is a short sound that is perceptually distinct from the surrounding sounds. It could be a bell ringing, a drum sound from an orchestra, or words in speech. The attention process in our brain depends on whether we are attending the auditory event.
In this research project we investigate if we can detect auditory attention based on EEG. Our approach is to identify auditory event in the sound stimuli and look for corresponding attention responses in the EEG.
The aim of a hearing aid is to compensate for hearing loss. To do so, the hearing aid should make the part of the sound signal that the user is interested in listening to audible and suppress ambient noise. Thus, a fundamental problem is to identify which parts of a complex sound environment a user perceives as noise and which parts the user wants to listen to.
Two listeners in the same sound environment are not necessarily interested in the same sound sources. This clearly shows that it is not possible to identify which parts of the sound environment the listener finds interesting based on the audio signal alone - we need somehow to decode the listeners auditory attention.
Ear-EEG is a technology for recording of electrophysiological signals from electrodes placed in or around the ear. The ultimate goal of this project is to decode auditory attention based on ear-EEG and thereby pave the way for better hearing aids in the future.
The project is funded by the William Demant Foundation grant number 20-2673.
