| Stroke is the third leading cause of death and second most frequent cause of disability in the United States. Stroke rehabilitation methods have been developed to induce the cortical reorganization and motor-relearning that leads to stroke recovery. In this thesis, we designed and developed an MR conditional upper extremity reach and grasp movement evaluation system for the stroke survivors to study their kinematic performances in reach and grasp movement and the relationship between kinematic metrics and the recovery level measured by clinical assessment methods. We also applied the system into the functional MRI experiments to identify the ability to study motor performance with the system inside the scanner and the reach, grasp and reach-to-grasp movements related brain activation patterns.;Our experiments demonstrate that our system is an MR conditional system in a 3.0 Tesla magnetic field. The system is able to measure the stroke survivors' reach and grasp movement in terms of grasp aperture and elbow joint angles. We used the Mann Whitney U test to examine the significant metrics in each tasks and principle component analysis to decide the major metrics that are associated with the outcome. Then we used the linear regression analysis to create the regression models between the recovery scores and the kinematic metrics. The regression models suggest that functional recovery for reach and/or grasp tasks is predictable with maximal and mean velocities, maximal movement, error in reach, grasp and reach-to-grasp tasks. We discovered that low functioning subjects generally showed smaller movement velocity, smaller maximal movement, larger error and longer time to peak velocity in reach, grasp and reach-to-grasp tasks. In addition to these metrics, time to maximal angle, time to target and time to peak velocity could also be used as additional metrics to help predict the recovery, assess robot-assisted therapy and optimize task-oriented rehabilitation strategy. We also applied the system into an fMRI case series and proved that we are able to capture the brain activation patterns after stroke with our system and experiment set up. |
