AI-Based Smart Tremor Monitoring System for Parkinson’s Disease Detection and Management

Abstract

Parkinson’s disease is a progressive neurologic, motor function-affecting disorder, one of the most common, measurable symptoms of which is tremor. Early detection and continuous monitoring of the severity of tremors is important for decision-making about treatment and for long-term care of patients. This paper presents a tremor analysis framework built using the Long Short-Term Memory network (LSTM) for the prediction of Parkinson’s disease severity using data on patient tremors. The proposed system does a preprocessing process of data, sequential pattern learning and severity classification using multiple stages such as healthy, mild, moderate and severe. Long Short-Term Memory (LSTM) networks have been used due to the ability of this type of neural network to detect temporal patterns in tremor patterns more effectively than conventional machine learning techniques. In addition to prediction, the framework has an interactive monitoring dashboard, patient history tracking, and automated report generation. The proposed system can act as a support tool for intelligent screening, evaluation, and continuous monitoring of Parkinson’s disease in a health care environment.