Attention decoded from brain activity of elderly hearing-impaired

fredag 17 apr 20
|

Kontakt

Jens Hjortkjær
Seniorforsker
DTU Sundhedsteknologi
30 29 16 90

Kontakt

Torsten Dau
Sektionsleder, Professor
DTU Sundhedsteknologi
45 25 39 77

Researchers from Hearing Systems at DTU Health Tech have developed techniques to decode attention from brain activity of elderly hearing impaired listeners. The results show promise for the development of future neuro-steered hearing instruments.

Hearing-impaired patients and hearing aid users suffer from difficulties understanding speech in noisy environments with many talkers. Normal-hearing listeners can navigate complex acoustic environments by focusing attention on one speaker and ignore background sounds. To restore this ability for patients with hearing loss, the hearing instrument must know what sounds the listener is trying to listen to.

In a new paper, Søren A. Fuglsang and colleagues show that attention can be decoded from only 10 seconds of EEG data recorded from the scalp of hearing-impaired listeners. Elderly listeners with hearing loss and normal-hearing controls were presented with competing talkers and asked to focus attention on one of them. By reconstructing the speech audio from the neural EEG recordings, the authors were able to identify an enhancement of the attended signals in both groups. The authors also report that hearing loss paradoxically enhances brain responses to sound, suggesting that the brain compensates for the loss of sensitivity at central stages of the auditory system.

Senior Researcher Jens Hjortkjær concludes: 'With these findings, we have increased our knowledge about the consequences of hearing loss in the brain, but also taken a step towards a potential new form of hearing aid technology.'

The results are published in a new paper in Journal of Neuroscience.

Acknowledgement

The work reported is part of the EU H2020 project COCOHA: Cognitive Control of a Hearing Aid, aiming to develop brain-steered hearing instruments that combine neural decoding and speech separation technology.

Nyheder og filtrering

Få besked om fremtidige nyheder, der matcher din filtrering.