Meng-Fen Tsai and IBBME UofT

December 4, 2019 @ 12:30 pm – 1:00 pm
Red Seminar Room
Donnelly Building

Event Name: Graduate Seminar Series: Clinical Stream

Graduate Seminar Series for the Institute of Biomaterials and Biomedical Engineering (IBBME). This day is for clinical stream presenters.

Location: Red Seminar Room – Donnelly Building

Presentation Title: Monitoring Functional Hand Use of Stroke Survivors at Home Using Egocentric Video
Stroke survivors experience a significant impact on their quality of life after the onset of stroke. Upper limb function is an important determinant of independence after injury. Previous studies reveal that current hand function assessments, administered predominantly in a clinical setting, are not necessarily indicative of how stroke survivors truly use their hands in activities of daily living in their home environment. Wearable devices based on accelerometers have been proposed to capture daily upper limb use after stroke, but accelerometers do not document what tasks were performed, describe hand function, or reveal the performance of affected and unaffected hands during activities of daily living (ADLs). In this study, we will explore the use of egocentric cameras to record the daily activities of stroke survivors at home. Through automated analysis of the recorded videos, we can uniquely access information about the amount of functional hand use and the reliance on the affected and unaffected hands in bimanual ADLs.
In previous work, a machine learning-based algorithm was trained to identify hand-object interactions of individuals with cervical spinal cord injury, with bilateral hand impairment, and had promising results. In this study, we will apply similar and refined versions of this approach to data from 25 stroke survivors with unilateral hand impairments, collected both in a home simulation laboratory and in the homes of study participants. We will additionally extend the algorithms to classify the roles of hands of stroke survivors in ADLs to reveal the contribution of the affected hand. The output from our algorithm will be compared to clinical outcome measures to establish convergent validity. Furthermore, we will report the user experience with the egocentric camera of the participants after stroke.
This is the first study to monitor the hand use of stroke survivors at home via egocentric video, providing novel insights into their performance of upper limb ADLs. We expect that the work will have applications in quantifying the true impact of rehabilitation interventions, in stroke survivors as well as a wide variety of clinical populations suffering from hand function impairments.
Supervisor Name: Jose Zariffa
Year of Study: 2
Program of Study: PhD

Powered by