Advance restless leg syndrome monitoring with deep learning

Periodic Limb Movements (PLMs) are frequently observed in patients with Restless Legs Syndrome (RLS). While electromyography (EMG) of leg muscles is traditionally used to quantify the motor symptom burden of RLS, wearable trackers may cause discomfort to patients. This study investigates the efficacy of pressure-sensing mat data in detecting PLMs, offering a non-invasive alternative for monitoring limb movements during sleep. Our approach utilizes a pressure-sensing mat that captures subtle changes in pressure distribution, providing a comfortable method for continuous monitoring. We collected a comprehensive dataset comprising 153.5 hours of synchronized pressure mat and EMG recordings from 21 patients. Ground truth labels were derived from concurrent EMG data, ensuring reliable annotations for PLM detection. We also propose to finetune a deep learning model, 3D-ResNet18 with pretrained Kinetics700k weights, as a baseline for this dataset that predicts PLMs from pressure mat data. Our approach offers a comfortable alternative to EMG-based detection, opening new avenues for continuous home-based monitoring.

Name thesis:degree_name

M.S.

Level thesis:degree_level

Thesis

Discipline thesis:degree_discipline

Computer Science

Grantor

University of Illinois at Urbana-Champaign

Year dc:date

2024

Author dc:creator

• Yu, Hang

Contributors dc:contributor

• Wang, Yuxiong

• Restless Leg Syndrome
• Deep Learning
• Transfer Learning

Statement dc:rights

• Copyright 2024 Hang Yu

Language dc:language

en, eng

Handle dc:identifier

https://hdl.handle.net/2142/127243

OAI record oai:identifier

https://www.ideals.illinois.edu/oai-pmh?verb=GetRecord&metadataPrefix=etdms&identifier=oai%3Awww.ideals.illinois.edu%3A2142%2F127243