Machine Learning-Assisted Diagnosis of Speech Disorders: A Review of Dysarthric Speech
Shaik Mulla Shabber, Mohan Bansal, Kodali Radha
Abstract
Speech disorders are commonly associated with neurodegenerative impairments in the motor system of the brain. Challenges to speech and communication are often experienced by those with conditions including Amyotrophic Lateral Sclerosis (ALS), Cerebral Palsy (CP), and Parkinson's Disease (PD). This review focuses on the machine learning applications for diagnosing speech disorders, with a particular emphasis on dysarthric speech. ALS, in particular, presents difficulties in speech production, spinal functions, respiration, and swallowing, making early diagnosis crucial for improving the patient's quality of life. Sustained vowel phonation emerges as a promising technique for distinguishing among ALS, PD, CP, and individuals in good health. This comprehensive review explores the various machine learning approaches utilized for diagnosing ALS, PD, and CP through speech analysis, highlighting their potential impact on future research.