Les rencontres de STMS welcome Fotios Drakopoulos, a postdoctoral researcher who works with Prof. Nicholas Lesica at the Ear Institute at University College London (United Kingdom). He will present his work titled “Optimal hearing aid design through restoration of the neural code.”
Here is a Zoom link to join remotely:
ttps://us06web.zoom.us/j/83956412580?pwd=G97hbvoeM3S1jNnJuqevYB0wsbY9y0.1
Link to the meeting chat
https://us06web.zoom.us/launch/jc/83956412580
ID Meeting: 839 5641 2580
Password: 795365
Abstract :
Hearing loss introduces complex distortions in the neural coding of sound that current hearing aids fail to address. While most devices employ additional features (including AI-based) to compensate for the complex effects of hearing loss on auditory processing and address specific listening scenarios, their real-world utility for sounds such as speech in complex background noise or music remains limited.
We have taken a new approach to hearing aid design, operating from the perspective of the neural code - the brain activity patterns that underlie perception. If hearing is impaired, it is because the details of the neural code have become distorted. Here, we combine electrophysiology and deep learning to identify novel sound processing strategies to correct these distortions. We use large-scale intracranial recordings from the gerbil inferior colliculus (IC) to train deep neural network models of neural coding (ICNets) to serve as in silico surrogates for brains with normal and impaired hearing. We then use the ICNets to train another network (AidNet) to act as an optimal hearing aid, providing the individualized sound processing required to elicit normal neural activity in impaired brains.
We find that AidNet outperforms state-of-the-art hearing aid processing by a wide margin in correcting distortions in neural coding and deficits in simulated phoneme recognition. Much of this advantage was retained when AidNet's parameters were swapped across animals with similar hearing thresholds, suggesting that AidNet can compensate for complex effects of hearing loss without direct access to neural recordings. These results demonstrate that closed-loop optimization can identify novel and generalizable strategies for neural restoration, providing a foundation for the development of more effective hearing aids as well as other sensory devices and neurotechnologies. By allowing for flexible signal processing and considering the effects of hearing loss on the full neural code, these tools have the potential to address the key shortcomings of current hearing devices.
Biography:
Fotios Drakopoulos is a Postdoctoral Research Fellow with Prof Nicholas Lesica at the Ear Institute of University College London (United Kingdom). Throughout his career spanning both industry and academia, his research has been focussed on the application of advanced computational techniques to auditory neuroscience to improve real-world applications such as hearing aids. He received his PhD in Electrical Engineering from Ghent University (Belgium) in 2022, as part of the Hearing Technology Lab led by Prof Sarah Verhulst, and his MSc degree from the Electrical and Computer Engineering department at University of Patras (Greece) under the supervision of Prof John Mourjopoulos. His research interests include diverse audio-related topics such as audio signal processing, machine hearing, hearing loss, computational auditory modelling and psychoacoustics.
