PhD Dissertation Proposal Defense: Yunda Liu, Advancing Objective Assessment of Physical and Behavioral Phenotypes Using Wearable and Mobile Technologies

Content

Neurological disorders and brain injuries are often accompanied by physical  and behavioral phenotypes that severely limit patients' ability to perform activities of daily living and maintain social connections. Current assessments of these symptoms rely on subjective and visual observations by  trained clinicians, which are limited by assessors’ professional experience. While patient-reported outcomes offer a more accessible  alternative, they are subject to recall bias and personal perception, further complicating the accuracy of assessments.

To address these limitations, wearable sensors and mobile technologies are  increasingly used outside clinical settings, offering the potential for convenient, objective, and continuous evaluation. This dissertation focuses on advancing wearable technology-based models for assessing physical and behavioral phenotypes in patients with neurological conditions.

First, we examine the kinematic differences between Huntington’s disease chorea and Parkinson’s disease choreic levodopa-induced dyskinesia. We leverage movement decomposition techniques and divide the inertial data of involuntary movements into sub-movements. A combination of unsupervised and supervised machine learning algorithms are employed to automatically select data features extracted from sub-movements and distinguish the two types of involuntary choreic movements.

Second, we investigate the concurrent validity and reliability of two  movement segmentation approaches widely used to assess the upper-limb motor function of stroke survivors. Acceleration time-series from wrist movements are decomposed into movement segments using each segmentation approach. Reliable features are extracted from the movement segments and supervised regression-models are trained to establish concurrent validity against  existing clinical measures.

Finally, the dissertation extends the application of wearable and mobile devices beyond motor deficit assessment to detect social isolation in stroke survivors. This dissertation aims to advance the frequent and reliable monitoring of social health, ultimately facilitating real-time detection and intervention strategies.