EXPLAINABLE AI-BASED SPEECH ANALYSIS FOR EARLY PARKINSON’S DISEASE DETECTION

Authors:

Dr. Kavitha A.S*, Shubha P, Vaishnavi Myakalmaddi, Bhavya G

Abstract—Parkinson’s disease is a progressive neurological disorder where early detection remains difficult, as traditional diagnostic methods mainly rely on motor symptoms that often appear at later stages. Recently, speech has gained attention as a non-invasive biomarker, since changes in vocal patterns can reflect underlying neurological impairments. However, many existing approaches depend on complex black-box models, making their predictions difficult to interpret and limiting their clinical applicability.In this work, we present an Explainable Artificial Intelligence (XAI)-based approach for early detection of Parkinson’s disease using speech signals. The system focuses on extracting meaningful acoustic features such as pitch, jitter, shimmer, and temporal variations, which are then analyzed using machine learning models. Unlike conventional methods, the proposed approach emphasizes interpretability by providing clear explanations for the model’s predictions.The experimental results indicate that speech-based analysis can effectively differentiate between early-stage Parkinson’s patients and healthy individuals, while also maintaining transparency in decision-making.Overall, the proposed method provides a practical, non-invasive, and interpretable solution for early screening, which can assist healthcare professionals in making more reliable and informed diagnostic decisions.