Back

A deep learning framework for understanding cochlear implants

Banerjee, A.; Saddler, M. R.; Arenberg, J. G.; McDermott, J. H.

2025-07-21 neuroscience
10.1101/2025.07.16.665227 bioRxiv
Show abstract

Sensory prostheses replace dysfunctional sensory organs with electrical stimulation but currently fail to restore normal perception. Outcomes may be limited by stimulation strategies, neural degeneration, or suboptimal decoding by the brain. We propose a deep learning framework to evaluate these issues by estimating best-case outcomes with task-optimized decoders operating on simulated prosthetic input. We applied the framework to cochlear implants - the standard treatment for deafness - by training artificial neural networks to recognize and localize sounds using simulated auditory nerve input. The resulting models exhibited speech recognition and sound localization that was worse than that of normal hearing listeners, and on par with the best human cochlear implant users, with similar results across the three main stimulation strategies in current use. Speech recognition depended heavily on the extent of decoder optimization for implant input, with lesser influence from other factors. The results identify performance limits of current devices and demonstrate a model-guided approach for understanding the limitations and potential of sensory prostheses.

Matching journals

The top 5 journals account for 50% of the predicted probability mass.

1
Scientific Reports
○ 3612 papers in training set
Top 0.5%
19.2%
Similar papers in this journal
2
Journal of Neural Engineering
○ 221 papers in training set
Top 0.3%
13.4%
Similar papers in this journal
3
PLOS Computational Biology
● 1863 papers in training set
Top 4%
8.2%
Similar papers in this journal
4
Frontiers in Neuroscience
○ 256 papers in training set
Top 0.3%
6.5%
Similar papers in this journal
5
Nature Communications
○ 5641 papers in training set
Top 26%
5.7%
Similar papers in this journal
50% of probability mass above
6
Nature Human Behaviour
○ 95 papers in training set
Top 0.4%
4.2%
7
eLife
● 5828 papers in training set
Top 29%
4.2%
8
PLOS ONE
● 5266 papers in training set
Top 35%
3.3%
9
Trends in Hearing
○ 15 papers in training set
Top 0.1%
3.3%
10
The Journal of the Acoustical Society of America
● 35 papers in training set
Top 0.2%
2.9%
11
Nature Machine Intelligence
○ 70 papers in training set
Top 1.0%
2.7%
12
Science Advances
● 1243 papers in training set
Top 14%
2.5%
13
Communications Biology
○ 993 papers in training set
Top 10%
2.2%
14
iScience
○ 1154 papers in training set
Top 15%
2.0%
15
Proceedings of the National Academy of Sciences
● 2444 papers in training set
Top 33%
1.2%
16
NeuroImage
○ 903 papers in training set
Top 6%
1.0%
17
Neuromodulation: Technology at the Neural Interface
○ 14 papers in training set
Top 0.3%
0.9%
18
The Journal of Neuroscience
● 1025 papers in training set
Top 10%
0.9%
19
Journal of the Association for Research in Otolaryngology
○ 15 papers in training set
Top 0.2%
0.9%
20
IEEE Transactions on Biomedical Engineering
● 40 papers in training set
Top 0.9%
0.9%
21
Scientific Data
○ 209 papers in training set
Top 3%
0.6%
22
Hearing Research
○ 54 papers in training set
Top 0.4%
0.6%
23
Neural Computation
● 39 papers in training set
Top 0.9%
0.5%
24
Journal of Neurophysiology
● 302 papers in training set
Top 4%
0.5%
25
Ear & Hearing
○ 17 papers in training set
Top 0.2%
0.5%
26
eneuro
● 439 papers in training set
Top 9%
0.5%
27
Brain Stimulation
○ 125 papers in training set
Top 1%
0.5%

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.