Towards personalized and optimized fitting of cochlear implants

Area of Science:

• Neurotechnology
• Auditory Neuroscience
• Biomedical Engineering

Background:

• Cochlear implants (CI) restore hearing in sensorineural hearing loss by electrically stimulating the auditory nerve.
• Current CI fitting relies on subjective, time-limited clinical data, leading to significant inter-user variability in speech recognition.
• Major factors contributing to variability include auditory system malfunction, electrode-nerve activation selectivity, and lack of objective perceptual measures.

Purpose of the Study:

• To address the unexplained variability in speech recognition among cochlear implant users.
• To propose a novel approach for individualized CI device fitting using objective, patient-specific data.
• To develop machine-learning algorithms for personalized CI fitting based on auditory system characteristics.

Main Methods:

• Proposing a series of experiments to collect quantitative, reproducible, and reliable data.
• Characterizing three key processing levels: auditory system malfunction, electrode-to-auditory nerve (EL-AN) activation selectivity, and perceptual outcomes.
• Utilizing machine-learning algorithms to analyze collected data for personalized fitting.

Main Results:

• The proposed methodology aims to collect comprehensive objective data across three critical levels of CI user processing.
• Machine learning will be employed to derive personalized user characteristics and predict perceptual effects.
• This approach is expected to improve the accuracy and reliability of CI fitting.

Conclusions:

• Individualized CI fitting requires objective, patient-specific data encompassing auditory system function, EL-AN selectivity, and perceptual measures.
• The proposed experimental framework and machine learning approach offer a path towards optimizing CI performance.
• This research has the potential to significantly reduce speech recognition variability and enhance outcomes for CI users.