Abstract
Long-term physical disabilities impose life-changing challenges on millions of people around the world. Many are caused by neurological injuries, such as stroke, leaving the survivors with permanent motor impairment. Assessing neuromuscular capacity after a neurological injury is critical to selecting targeted therapy and maximizing the recovery of motor function. Measurements of electrophysiological activity have shown promise as an investigation tool for neurological capacity and function, but further research is needed to identify biomarkers that are most informative in understanding motor impairment. The research presented in this dissertation aims to develop analytical technics and expertise to interpret electrophysiological signals, such as electroencephalography (EEG) and electromyography (EMG), for the purpose of studying functional aspects of human movement and sensation and improving stroke assessment and therapy. This work was completed to support robotic applications for assessment and therapy in the Assistive Robotics Lab at the University of Idaho, and included collaboration with a multidisciplinary team.
Acquiring and processing EEG and EMG data requires significant expertise to ensure the resulting biomarkers are scientifically valid; a description of this expertise is presented in chapter 2. Moreover, the review of the state of the art presented in chapter 3 states the importance of biomarkers’ reliability for these to be used in clinical settings.
In chapter 4, this theoretical knowledge is put to practice by performing multidomain analysis of stroke survivors’ data, resulting in methods for understanding stroke-related impairment. Some of this analysis tested the predictive power of proposed and existing EEG metrics in relation to stroke survivors’ response to robotic-assisted therapy.
Also investigated were the neural correlates of proprioception during active bilateral and unilateral movements, and the adaptation of this brain activity due to proprioceptive training. This study allowed the identification of a neurological feature not only related to proprioceptive processing but also adapted after proprioceptive training. The feature shows potential as a biomarker as it could be used to quantify altered proprioceptive neural processing in skill and movement disorders. The pertinent research is presented in chapter 5.
All these applications show the validity of EEG-derived metrics to aid the assessment of upper-extremity impairment, such as the one resulting from stroke lesions. Electrophysiological data were also used to study functional aspects of sensation and identified features that correlate to aspects of proprioceptive uncertainty in both healthy and stroke groups.