Automated Vocal Fatigue Screening in Professional Voice Users: Development and Occupational Validation of an...

Area of Science:

• Occupational Health
• Speech Science
• Biomedical Engineering

Background:

• Professional voice users have a higher risk of vocal injury due to sustained voice use.
• Current scalable tools for workplace vocal health surveillance are limited.

Purpose of the Study:

• To develop and validate an automated vocal fatigue monitoring system for occupational health.
• To provide a speaker-independent assessment tool for vocal load.

Main Methods:

• Developed a speaker-independent deep learning system trained on diverse speech samples.
• Validated the system against the Vocal Fatigue Index (VFI) in 27 participants.
• Assessed temporal sensitivity with twice-daily recordings from 8 teachers over 5 workdays.

Main Results:

• Automated scores strongly correlated with VFI (r = 0.942).
• Fatigue scores significantly increased across work shifts (d = 1.00) and correlated with speaking duration (r = 0.544).
• Teachers reported substantially higher vocal fatigue than the general population (d = 3.21).

Conclusions:

• The automated system shows promising validation for shift-level screening of occupational vocal load.
• The speaker-independent assessment is suitable for workplace health surveillance.
• This technology can aid in vocal injury prevention programs for professional voice users.