*********************************
There is now a CONTENT FREEZE for Mercury while we switch to a new platform. It began on Friday, March 10 at 6pm and will end on Wednesday, March 15 at noon. No new content can be created during this time, but all material in the system as of the beginning of the freeze will be migrated to the new platform, including users and groups. Functionally the new site is identical to the old one. webteam@gatech.edu
*********************************
Title: Combining wearables and nearables for patient state analysis
Date: August 9th, 2022
Time: 3pm (EDT)
Location: https://gatech.zoom.us/j/99493423780
Pradyumna Byappanahalli Suresha
Machine Learning PhD Student
School of Electrical and Computer Engineering
Georgia Institute of Technology
Committee
Prof. Gari D. Clifford (Advisor)
Prof. David V. Anderson
Prof. Omer T. Inan
Prof. Thomas Ploetz
Prof. Ali Bahrami Rad
Abstract
Recently, ambient patient monitoring using wearable and nearable sensors is becoming more prevalent, especially in the neurodegenerative (Rett syndrome) and sleep disorder (Obstructive sleep apnea) populations. While wearables capture localized physiological data such as pulse rate, wrist acceleration and brain signals, nearables record global passive data including body movements, ambient sound and environmental variables. Together, wearables and nearables provide a more comprehensive understanding of the patient state.
The processing of data captured from wearables and nearables have multiple challenges including handling missing data, time synchronization between sensors and developing data fusion techniques for multimodal analysis. The research described in this thesis addresses these issues while working on data captured in the wild. First, we describe a Rett syndrome severity estimator using a wearable biosensor and uncover physio-motor biomarkers. Second, we present the applications of an edge computing and ambient data capture system for home and clinical environments. Finally, we describe a transfer learning and multimodal data fusion based sleep-wake detector for a mixed-disorder elderly population. We show that combining data from wearables and nearables improves the performance of the sleep-wake detection in terms of the F1-score and the Cohen’s kappa compared to the unimodal
