OBJECTIVE ASSESSMENT OF VOCAL TREMOR

Area of Science:

• Neurology
• Speech Science
• Biomedical Engineering

Background:

• Early detection of neurodegenerative diseases is crucial for effective treatment planning.
• Speech changes are recognized as sensitive early indicators of neurological decline.
• Current clinical assessments of speech symptoms rely on subjective scales, limiting early detection accuracy.

Purpose of the Study:

• To develop an objective algorithm for assessing vocal tremor in neurological disorders.
• To overcome the limitations of subjective clinical scales in identifying subtle speech changes.
• To provide a reliable, low-dimensional feature set for classifying neurological conditions based on vocal characteristics.

Main Methods:

• Extraction and aggregation of acoustic features from energy and fundamental frequency profiles during sustained phonation.
• Analysis of average spectra from vocal recordings.
• Development of a classification algorithm utilizing a low-dimensional feature set.

Main Results:

• The algorithm successfully extracted relevant vocal features.
• The low-dimensional feature set demonstrated reliable classification between healthy controls and patients with amyotrophic lateral sclerosis (ALS).
• Classification accuracy was validated against perceptual ratings by speech-language pathologists.

Conclusions:

• The proposed algorithm offers an objective and quantitative method for assessing vocal tremor.
• This approach can aid in the early detection of neurological disorders like ALS through speech analysis.
• Objective vocal analysis presents a promising avenue for remote, hardware-independent neurological monitoring.