The field of Parkinson's disease diagnosis and assessment is moving towards the development of more accurate and reliable methods for early detection and monitoring. Researchers are exploring various modalities, including gait analysis, keystroke dynamics, facial expressions, and hand-drawn patterns, to identify biomarkers for the disease. Multimodal approaches are being proposed to integrate complementary information from different data sources, and techniques such as multi-objective optimization and attention-based feature fusion are being used to improve performance. Additionally, there is a focus on addressing class imbalance and improving cross-patient generalization. Notable papers in this area include: Towards Relaxed Multimodal Inputs for Gait-based Parkinson's Disease Assessment, which proposes a framework for relaxed multimodal inputs and achieves state-of-the-art performance. Cross-dataset Multivariate Time-series Model for Parkinson's Diagnosis via Keyboard Dynamics, which demonstrates the potential of keystroke dynamics as a reliable digital biomarker for PD. SAFE-D: A Spatiotemporal Detection Framework for Abnormal Driving Among Parkinson's Disease-like Drivers, which investigates driving behavior and proposes a framework for detecting Parkinson-related behavioral anomalies. Facial Expression-based Parkinson's Disease Severity Diagnosis via Feature Fusion and Adaptive Class Balancing, which integrates multiple facial expression features and mitigates class imbalance. Improving Cross-Patient Generalization in Parkinson's Disease Detection through Chunk-Based Analysis of Hand-Drawn Patterns, which proposes a chunk-based analysis approach to improve cross-patient generalization.