Related Experiment Video
Updated: Sep 12, 2025
09:44
Neuro-rehabilitation Approach for Sudden Sensorineural Hearing Loss
Published on: January 25, 2016
19.3K
Uncovering Phenotypes in Sensorineural Hearing Loss: A Systematic Review of Unsupervised Machine Learning Approaches
Lilia Dimitrov1,2, Liam Barrett1,2, Aizaz Chaudhry3
1University College London Hospital (UCLH) Biomedical Research Centre (BRC) Hearing Theme, London, United Kingdom.
Ear and Hearing
|August 7, 2025
Summary
Unsupervised machine learning shows promise for classifying sensorineural hearing loss (SNHL) subtypes. However, current research quality is low, hindering definitive conclusions on model selection and subtype identification.
Area of Science:
- Audiology
- Artificial Intelligence
- Data Science
Background:
- Sensorineural hearing loss (SNHL) affects over 1.5 billion people globally.
- Accurate SNHL subtyping is crucial for developing personalized treatments.
- Unsupervised machine learning (AI) offers a potential solution for SNHL classification.
Purpose of the Study:
- To systematically review the application of unsupervised machine learning models in identifying SNHL subtypes.
- To synthesize existing literature on AI-driven SNHL classification methods.
Main Methods:
- Systematic literature search across multiple databases (MEDLINE, EMBASE, Scopus, etc.) and grey literature.
- Inclusion criteria: adult SNHL patients, unsupervised machine learning approach.
- Quality assessment using the APPRAISE-AI tool; narrative synthesis of results.
Main Results:
- Seven studies met the inclusion criteria; most were cohort studies.
- Four distinct unsupervised machine learning algorithms were identified.
- Identified subtypes varied (4 and 11), but low study quality prevents definitive conclusions.
Conclusions:
- Methodological improvements are needed to realize the potential of AI in SNHL subtyping.
- Future research should focus on justifying model selection, ensuring reproducibility, using high-quality data, and validating findings.
