Comparative Analysis of Machine Learning Approaches for the Detection of Parkinson’s Disease: A Systematic Review of Data Modalities, Methodologies, and Clinical Challenges

Abstract

Parkinson’s disease (PD) is a progressive neurodegenerative disorder characterized by the degeneration of dopaminergic neurons in the substantia nigra, leading to motor and non-motor impairments. Early and accurate diagnosis remains challenging due to symptoms heterogeneity and reliance on subjective clinical scales such as the Unified Parkinson’s Disease Rating Scale (UPDRS). In recent years, machine learning (ML) and deep learning (DL) approaches have emerged as promising tools for automated, objective, and non-invasive PD detection using multimodal biomedical data. This paper presents a comprehensive and systematic review of ML-based approaches for PD detection, covering data modalities including speech and acoustic signals, wearable and kinematic sensors, handwriting and drawing patterns, and neuroimaging data such as MRI, PET, SPECT, and EEG. We analyze preprocessing techniques, feature engineering strategies, classification and regression models, and evaluation metrics employed in the literature. A detailed comparative study of landmark works highlights the strengths and limitations of traditional ML models and modern DL architectures. The review further discusses key challenges such as data scarcity, class imbalance, lack of generalization, and the black-box nature of deep models, emphasizing the need for explainable AI in clinical adoption. Finally, future research directions focusing on multimodal fusion, longitudinal analysis, and clinically interpretable models are outlined.