Voice disorder recognition using machine learning: a scoping review protocol

Area of Science:

• Speech and Hearing Sciences
• Biomedical Engineering
• Computer Science

Background:

• Machine learning (ML) algorithms demonstrate high accuracy in detecting voice disorders.
• ML holds potential for aiding clinicians in voice disorder analysis and treatment evaluation.
• Despite research, no ML algorithms are currently reliable enough for widespread clinical application.

Purpose of the Study:

• To identify critical issues impeding the clinical use of ML algorithms for voice disorder detection.
• To pinpoint standard audio tasks, acoustic features, processing algorithms, and environmental factors influencing ML efficacy.

Main Methods:

• A comprehensive literature search was conducted across seven major databases (Web of Science, Scopus, Compendex, CINAHL, Medline, IEEE Explore, Embase) from 2013 to 2023.
• Search strategy was refined with university library assistance for database-specific syntax.
• Data selection, extraction, and synthesis followed the 'Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for scoping reviews' (PRISMA-ScR) guidelines.

Main Results:

• This section is to be populated upon completion of the review.

Conclusions:

• This scoping review will provide insights into the challenges and limitations of applying ML in clinical voice disorder detection.
• Findings will guide future research to develop more robust and clinically viable ML solutions for voice pathology.