Rapid auditory readout tracks nerve loss in gerbils
A machine-learning model trained on gerbil synapse counts separated age groups above chance when tested on 56 human listeners.
bioRxiv
In Mongolian gerbils with histologically verified cochlear neural degeneration, researchers evaluated dynamic amplitude-modulated envelope following responses that sweep the modulation spectrum in a single brief stimulus. Cochlear neural degeneration involves the loss of synapses between inner hair cells and auditory nerve fibers, which standard hearing tests miss. The test revealed selective deficits at fast modulation rates without elevating hearing thresholds. A machine-learning classifier distinguished young from middle-aged gerbils with high accuracy, relying on power near 400 to 500 Hz that tracked physical synapse counts. When applied without retraining to 56 human listeners, the gerbil-trained classifier separated age groups above chance.
Why it matters
Synapse loss between sensory hair cells and auditory nerves is a subtle hallmark of inner ear aging that conventional audiograms overlook. Developing rapid, objective electrophysiological readouts could allow clinicians to monitor age-related auditory decline and assess interventions designed to protect cochlear synapses.
Caveats
The findings are reported in a preprint and have not yet undergone peer review. In addition, the human evaluation was limited to 56 listeners, and the classifier separated human age groups above chance rather than directly measuring histopathology in people.
The paper
Carnegie Mellon University · University of Pittsburgh
bioRxiv · 30 Sep 2026 · Preprint, not peer-reviewed