Quantum Augmented Multimodal Parkinson’s Disease Detection and Rehabitation Using Hand-Drawing and Voice Biomarkers

Parkinson’s disease (PD) is a major health challenge because it gradually damages the nervous system and still lacks simple, reliable, and early methods for diagnosis. Conventional diagnostic techniques often rely on subjective clinical assessments or unimodal biomarkers, resulting in limited accuracy, poor generalization, and delayed intervention. To overcome these limitations, this work introduces a Quantum Augmented Multimodal Framework for Parkinson’s Disease Detection Using Hand-Drawing and Voice Biomarkers with Adaptive Digital Rehabilitation Recommendation. The proposed approach integrates multimodal biomarker analysis through three major innovations: Hand-Drawing based (Motor Signal) Feature Extraction via Hybrid Spatial-Frequency Motor Embeddings (HSFME) for precise motor variability modeling, Voice Signal based Feature Extraction via Temporal-Spectral Voice Embedding Framework (TSVEF) for comprehensive vocal feature representation, and a Quantum-Classical Diagnostic Classifier (QCDC) that leverages variational quantum circuits (VQC) for enhanced feature entanglement and discriminative learning. Implemented in Python, the proposed framework achieves 99.2% accuracy and 97.63% precision, outperforming state-of-the-art baselines. The system further includes an actor-critic-based recommendation module for adaptive rehabilitation, establishing a reliable, interpretable, and technology-driven paradigm for early Parkinson’s diagnosis and patient-specific digital therapy guidance.