Automatic analysis of cochlear response using electrocochleography signals during cochlear implant surgery

Area of Science:

• Oto-neurology
• Biomedical Engineering
• Surgical Technology

Background:

• Cochlear implants (CIs) restore hearing but risk damaging natural hearing, impacting speech and music perception.
• Preserving residual hearing is crucial for optimal outcomes and patient satisfaction with cochlear implantation.
• Electrocochleography (ECochG) monitors cochlear nerve function during surgery to detect potential trauma.

Purpose of the Study:

• To develop and validate an automated method for analyzing ECochG responses during CI surgery.
• To assess the accuracy of automated ECochG analysis in detecting intra-operative trauma compared to human experts.
• To enhance patient safety and the clinical utility of ECochG in preserving natural hearing.

Main Methods:

• Utilized historical patient data for training and validation of the automated analysis algorithm.
• Compared the performance of the automated ECochG analysis against assessments made by human experts.
• Focused on identifying changes in ECochG responses indicative of cochlear nerve trauma.

Main Results:

• The automated method achieved high accuracy in detecting intra-operative trauma.
• Sensitivity reached approximately 94%, and specificity reached approximately 95% when compared to human expert analysis.
• Demonstrated the feasibility of real-time analysis of cochlear responses.

Conclusions:

• Automated ECochG analysis is a highly accurate and reliable tool for detecting cochlear nerve trauma during CI surgery.
• This technology has the potential to significantly improve patient safety by enabling early intervention to preserve natural hearing.
• Automation is expected to increase the scalability and accessibility of ECochG monitoring in clinical practice.