Towards Personalized Auditory Models: Predicting Individual Sensorineural Hearing-Loss Profiles From Recorded Human Auditory Physiology.
Sarineh Keshishzadeh, Markus Garrett, Sarah Verhulst
PMID 33526004WHAT IT FOUND
A computational model using otoacoustic emissions and brainstem responses predicted hidden nerve damage in hearing-impaired listeners.
The method achieved 83.81% accuracy in simulations but cannot yet be validated against actual human tissue, so it is not ready for clinical use.
Key findings
01The study developed a method to predict cochlear synaptopathy (hidden hearing loss) by combining computer models of the inner ear with recorded auditory brainstem responses and otoacoustic emissions.
02Using distortion product otoacoustic emission thresholds to personalize the model yielded higher prediction accuracy (83.81%) than using standard audiogram thresholds (68.57%).
03The accuracy of these predictions is based on comparison with computer simulations, not actual histological verification, because human temporal bone validation is impossible.
STILL TO COME
How it was doneWhat they foundWhat it means for SLPs
Read the rest of this summary
You get three full summaries a month, free, and we do not ask for a card. Search, the TL;DRs and your library stay unlimited either way.
What it does not show
The 'accuracy' reported is how well the model predicted its own simulations, not how well it predicted actual nerve damage in living humans, which cannot be measured directly. The study included only 35 subjects for the main analysis, which is a small sample for developing a diagnostic algorithm. The model could not simulate hearing loss greater than 35 dB, which limits its applicability to patients with severe high-frequency loss. The validation group consisted only of young normal-hearing listeners, so the method has not been tested on hearing-impaired individuals using different equipment.
Declared interests
The authors declared no potential conflicts of interest. The work was funded by the European Research Council and the DFG Cluster of Excellence.
The easy way to misread this
Do not interpret the 83.81% accuracy as evidence that this method can diagnose hidden hearing loss in your patients. The accuracy reflects how well the computer model matched its own simulations, not how well it matched actual biological nerve damage, which remains unverified in humans.