Related Experiment Video
Updated: May 30, 2025
06:04
Systematic Hearing Performance Evaluation Process for Adolescents with Cochlear Implantation at Early Ages
Published on: March 24, 2023
330
Machine-Learning Predictions of Cochlear Implant Functional Outcomes: A Systematic Review
Jonathan T Mo1, Davis S Chong1, Cynthia Sun1
1University of California, Davis School of Medicine, Sacramento, California, USA.
Ear and Hearing
|January 29, 2025
Summary
Machine learning (ML) models show promise in predicting cochlear implant (CI) user outcomes by analyzing complex data. However, improved reporting and validation are needed for clinical use.
Area of Science:
- Medical Informatics
- Biomedical Engineering
- Auditory Neuroscience
Background:
- Predicting cochlear implant (CI) user outcomes is complex due to individual variability.
- Machine learning (ML) offers advanced analytical capabilities for these predictions.
Purpose of the Study:
- To systematically review ML models predicting CI functional outcomes (sound perception and production).
- To analyze ML model strengths, weaknesses, and key predictive features.
- To suggest future directions for ML in CI research and clinical practice.
Main Methods:
- Systematic literature search across multiple databases (inception-September 2024).
- Inclusion of 16 studies with 5058 pediatric and adult CI users.
- Extraction of data on participant, CI characteristics, ML models, and performance metrics.
Main Results:
- ML models predicted sound production, perception, and language outcomes.
- Key predictors included demographic, audiological, imaging, and patient-reported data.
- Best model accuracies ranged from 71.0% to 98.83%, with diverse ML approaches used.
Conclusions:
- ML models demonstrate high predictive performance for CI outcomes.
- Inadequate reporting and unclear overfitting limit current clinical applicability.
- Standardized reporting and robust validation are crucial for future clinical adoption.
