Presentation Title
Using Machine Learning and Wearables to Classify Multiple Sclerosis Fall Risk
Abstract
Introduction: In the United States, there are nearly 1 million people living with Multiple Sclerosis. Multiple Sclerosis (MS) is a neurological disorder commonly characterized by fatigue, walking difficulties, numbness or tingling, and several other symptoms. These symptoms increase the risk of the MS population falling; which can prove detrimental for disease progression. Therefore, it is vital to provide patients with an accurate assessment of their fall risk and alert them when a fall is likely. Recent developments in wearable technologies has shown promise in using inertial mass units (IMUs) to quantify gait over extended periods of time, thus, allowing data to be gathered at home and in daily life routines. The purpose of this project is to develop an algorithm that determines when patients with MS are walking and determine whether the patient is classified as a faller or non-faller. This is an important step for ultimately creating a model that can predict fall risk in real time.
Materials and Methods: Data were collected from 29 subjects; each performing several daily life activities in the laboratory environment and from a thigh and chest mounted IMU at home for 48 hours. This data is analyzed using a support vector machine (SVM) classifier and bouts of walking are extracted. Following the identification of walking, a deep learning model will be used to group the walking of fallers and non-fallers.
Expected Results: Using the SVM classifier and deep learning model, patients with MS will be identified as a faller or non-faller based on walking IMU data.
Conclusion: The preliminary results obtained in this study suggest that it is possible to create a model that can distinguish between an MS patient classified as a faller or non-faller. This is a vital step in the process of creating a real-time fall risk model.
Primary Faculty Mentor Name
Ryan McGinnis
Graduate Student Mentors
Reed Gurchiek, Lindsey Tulipani
Faculty/Staff Collaborators
Reed Gurchiek (Graduate Student Mentor)
Status
Undergraduate
Student College
College of Engineering and Mathematical Sciences
Program/Major
Biomedical Engineering
Primary Research Category
Engineering & Physical Sciences
Secondary Research Category
Health Sciences
Using Machine Learning and Wearables to Classify Multiple Sclerosis Fall Risk
Introduction: In the United States, there are nearly 1 million people living with Multiple Sclerosis. Multiple Sclerosis (MS) is a neurological disorder commonly characterized by fatigue, walking difficulties, numbness or tingling, and several other symptoms. These symptoms increase the risk of the MS population falling; which can prove detrimental for disease progression. Therefore, it is vital to provide patients with an accurate assessment of their fall risk and alert them when a fall is likely. Recent developments in wearable technologies has shown promise in using inertial mass units (IMUs) to quantify gait over extended periods of time, thus, allowing data to be gathered at home and in daily life routines. The purpose of this project is to develop an algorithm that determines when patients with MS are walking and determine whether the patient is classified as a faller or non-faller. This is an important step for ultimately creating a model that can predict fall risk in real time.
Materials and Methods: Data were collected from 29 subjects; each performing several daily life activities in the laboratory environment and from a thigh and chest mounted IMU at home for 48 hours. This data is analyzed using a support vector machine (SVM) classifier and bouts of walking are extracted. Following the identification of walking, a deep learning model will be used to group the walking of fallers and non-fallers.
Expected Results: Using the SVM classifier and deep learning model, patients with MS will be identified as a faller or non-faller based on walking IMU data.
Conclusion: The preliminary results obtained in this study suggest that it is possible to create a model that can distinguish between an MS patient classified as a faller or non-faller. This is a vital step in the process of creating a real-time fall risk model.