A new AI-PROGNOSIS publication, “Rest Tremor and Dyskinesia Assessment in Parkinson’s Disease Using Wearable Sensors”, has been published in IEEE Internet of Things Magazine.

The study presents a deep learning-based Intelligent Internet of Medical Things (IIoMT) framework using wrist-worn accelerometer data to support three assessment tasks: rest tremor detection, rest tremor amplitude classification and dyskinesia detection.  

The proposed framework uses separate deep learning pipelines for tremor and dyskinesia assessment and transforms accelerometer data into symptom-specific outputs that can be aggregated into digital biomarkers.   

The approach was evaluated using data from the Michael J. Fox Foundation Levodopa Response Study and, for external validation of the rest tremor pipeline, a subset of the Verily Study Watch dataset. The study reports strong performance for rest tremor assessment and promising results for dyskinesia detection.

Read the publication: https://ieeexplore.ieee.org/document/11580313