MUMBAI, India, Aug. 24 -- Intellectual Property India has published a patent application (202641098822 A) filed by Dr Potu Narayana; A. Sowjanya; Mr. Venkatram Vennam; J. Sowmya; and K. Shilpa on August 14, 2026, for A Machine Learning-Based Adaptive Multi-Modal Framework For Early Detection And Personalized Prediction Of Neurodegenerative Disorders Using Wearable And Clinical Data.
Inventors include Dr Potu Narayana; A. Sowjanya; Mr. Venkatram Vennam; J. Sowmya; and K. Shilpa.
The application for the patent was published on August 21, 2026, under issue no. 34/2026.
Abstract: . ABSTRACT OF THE INVENTION A machine learning-based adaptive multi-modal framework for early detection and personalized prediction of neurodegenerative disorders is disclosed. The framework integrates continuous wearable sensor data—including gait, activity, sleep, and autonomic signals—with clinical assessments and optional neuroimaging biomarkers through specialized modality-specific encoders and a hierarchical cross-modal attention fusion architecture. An adaptive personalization layer maintains patient-specific latent states that are continuously updated via meta-learning principles, enabling individualized risk trajectories and progression forecasts even with sparse prior data. Multi-task prediction heads simultaneously perform disease stage classification, conversion risk estimation, time-to- event prediction, and biomarker trajectory forecasting. Explainability modules generate feature attributions, counterfactual insights, and temporal importance rankings to support clinical interpretation. The system handles missing and irregularly sampled multi-modal data robustly and supports privacy-preserving continual learning. By transforming episodic clinical evaluation into continuous, data-driven, and personalized monitoring, the invention enables earlier detection of neurodegenerative processes and more precise, actionable prognostic information for clinicians and patients.
Disclaimer: Curated by HT Syndication.
