Author(s) :

Jayant Namburi.

Volume/Issue :

Abstract :

Parkinson’s disease (PD), amyotrophic lateral sclerosis (ALS), Huntington’s disease (HD), and vascular dementia (VaD) all share a common feature of speech impairment. Recent advances in machine learning (ML) have allowed researchers to analyze speech patterns using acoustic features such as pitch variability, articulation rate, jitter, shimmer, and prosody. These measurable speech abnormalities can provide insight into neurological dysfunction and disease progression. This review evaluates the hypothesis that artificial intelligence (AI)-based machine learning of speech patterns can help recognize neurological impairment associated with PD, ALS, HD, and VaD by analyzing the pathophysiology, clinical symptoms, and speech characteristics of each disease, followed by an analysis of current speech-based ML approaches used in neurological research. Because most existing evidence comes from classifying speech in patients with an established diagnosis, this review distinguishes disease classification in diagnosed patients from genuine early-stage detection when interpreting the available evidence.