- Cochlear synaptopathy, or 'hidden hearing loss,' prevents patients from processing speech in noisy environments, yet often remains undetected by standard clinical audiograms.
- The EarDiTech research project has developed a novel diagnostic test that uses electrodes and computational models to measure synaptic response.
- Researchers have created a machine learning algorithm designed for hearing aids to optimize sound clarity for individual synaptic damage profiles.
- The team is now seeking commercial partnerships to bring the diagnostic hardware and software solutions to market.
Understanding Hidden Hearing Loss
- Traditional audiograms measure hair cell function but fail to assess synaptic health.
- Successful speech decoding requires multiple synapses per hair cell; when these connections degrade, patients struggle in noisy settings despite 'normal' test results.
- Professor Sarah Verhulst and Professor Ingeborg Dhooge identify synaptic loss as the primary structural damage, preceding hair cell degradation.
New Diagnostic and Therapeutic Tools
- The EarDiTech diagnostic test uses specifically designed audio stimuli to elicit a neural spike response, which is weaker in patients with synaptopathy.
- A compact, portable version of this diagnostic device has undergone successful clinical trials at Ghent University.
- The proposed machine learning-based software modifies sound to stimulate existing synapses effectively, rather than just increasing volume.
- The algorithm is optimized for low-power chips, making it compatible with future generations of hearing aids and consumer earbuds.
Future Outlook
- Developers are currently pursuing CE marking for the diagnostic device, a process expected to take 18 to 24 months.
- Clinical adoption of the diagnostic tool will provide a clearer data baseline regarding the prevalence and evolution of hidden hearing loss.
- The research team is actively engaging with hardware manufacturers to integrate their software solutions into commercial audio products.
This summary was generated by AI from the original article and may omit nuance or later updates. How everytldr works