Brain and Behavior Correlates of Arm Rehabilitation

PURPOSE OF THE STUDY AND BACKGROUND

1. Specific Aims and Purpose of The Study: Overall, this proposal is to use three-dimensional movement analysis and magnetic resonance (fMRI) neuroimaging techniques to examine the brain activity and motor behavior changes associated with constraint-induced (CI) therapy for patients with sub-acute stroke. This project is a companion study to the recently funded multi-center randomized clinical trial (HD37606-01), Extremity Constraint-Induced Therapy Evaluation (EXCITE). This project complements EXCITE by probing the mechanisms underlying the clinical intervention in stroke rehabilitation. However, since enrollment for EXCITE has been completed, we will be enrolling a new set of participants.
2. SPECIFIC AIMS:
3. Determine the relationship between lesion size and location as assessed using 3-D volumetric MRI and functional outcome (i.e, motor control and functional use) as a consequence of CI therapy.
4. Determine the effects of CI therapy on the control of reaching and grasping actions as assessed using behavioral kinematics.
5. Determine the effects of CI therapy on the neural substrate (sensorimotor network) as assessed using fMRI during goal-directed aiming and grasping tasks (performed as activation tasks in the scanner).
6. Determine the persistence and or stability of these changes in, motor control, neural substrate, and functional use of the upper extremity in stroke with CI therapy 6 months later.

A.1: Specific working hypotheses for each aim are:

For Specific Aim # 1, Hypothesis 1: There will be a predictable relationship between lesion location, size and functional outcome from CI therapy. Those with specific sensorimotor cortex (MI, SI) lesions will show a poorer response than those with lesions confined to the subcortical regions (e.g., posterior limb of internal capsule, subcortical white matter, or basis pontis, paramedian pontine areas) controlled for lesion size. One working assumption underlying this hypothesis is that there will be a differential effect of CI therapy with subcortical vs cortical sensorimotor area lesions. Hypothesis 2: The differences between groups (treated vs untreated) will be greater in those with subcortical lesions than in those with cortical sensorimotor area lesions, controlled for lesion size, at the post-test/intervention time point.

For Specific Aim # 2, Hypothesis 1: CI therapy will benefit the control of reaching and grasping actions through changes in (arm) transport duration, timing of peak aperture, peak transport velocity, and smoothness of both the reach and the grasp component for grasping different sized objects and reaching different distances. Hypothesis # 2 is that the change in the motor control of reaching and grasping actions from baseline to post-test/intervention time point will be greater in the treated compared to the untreated group. Hypothesis # 3 is that these motor control changes will be correlated with changes in real-world use of the upper extremity as indexed by the Motor Activity Log (MAL) and Wolf Motor Function Test (WMFT).

For Specific Aim # 3, Hypothesis 1: CI therapy will result in a decrease in ipsilateral hemisphere activation and an expanded extent and or increased intensity of specific regions of interest (ROIs) in the contralateral (affected) sensorimotor network with CI therapy. Specifically, and as it relates to the planning of more complex hand actions (i.e., precision grasp), specific ROIs that will show change in intensity and/or extent include primary motor cortex (MI), primary sensory cortex (S1), supplementary motor area (SMA), and premotor area (PM). Hypothesis # 2 is that the change in ROI activation from baseline to post-test/intervention time point will be greater in the treated compared to the untreated group. Hypothesis 3: The presence of an expanded extent and/or more intense activation in these contralateral hemisphere ROIs will be correlated with changes in real-world use of the upper extremity as indexed by the MAL and WMFT.

For Specific Aim # 4, Hypothesis 1: Persistence of these changes in reach-to-grasp motor control (behavioral kinematics), expanded extent/intensity of sensorimotor network ROIs (fMRI) and functional use (MAL and WMFT) will be demonstrated through comparison of the treated and untreated (delayed-intervention) groups at the 6-month follow-up point. It is possible that the differences between groups at the 6-month point may be greater than at the post-test/intervention point particularly if functional use continues to improve in the treated group.

1. Background and Significance: Stroke is the leading cause of disability among American adults. Each year in the U.S., approximately 750,000 people suffer strokes and of those, nearly 400,000 survive with some level of neurological impairment and disability. The cumulative total of living, stroke-affected Americans is nearly 3 million. Thirty billion dollars annually is the estimated burden from stroke-related disability. Despite the enormity of these statistics, there have been almost no scientific studies of the effectiveness of post-stroke rehabilitation interventions. As a result, the recently published Clinical practice guidelines for post-stroke rehabilitation (Gresham et al., 1995) have been forced to rely primarily on expert opinion and not on well-controlled scientific studies.

A new therapeutic approach to the rehabilitation of movement after stroke, termed constraint-induced (CI) movement therapy, has been derived from basic research with monkeys given somatosensory deafferentation (Taub, 1980). CI movement therapy consists of a family of therapies; their common element is that they induce stroke patients to greatly increase the use of an affected arm and hand for many hours a day over a period of 10 to 14 consecutive days. The signature intervention involves motor restriction of the less affected arm in a sling or mitt and training of the affected arm. The therapies result in large changes in amount of use of the affected arm in the activities of daily living outside of the clinic that have persisted for the 2 years measured to date (Morris et al., 1997; Taub \& Wolf, 1997; Taub et al., 1994; Taub et al., 1993; Wolf et al., 1989). Previous research has never resulted in any adverse effects from the restraint on the "good arm/hand".

There is mounting evidence for cortical reorganization that accompanies recovery of motor control after stroke, and recently, similar changes were demonstrated in connection with CI therapy in chronic stroke (Liepert et al., 2000). There appears to be a relationship between lesion size and location and the nature of this reorganization as well as the time course of this reorganization (Azari \& Seitz, 2000). The fact that therapy (forced use) can influence this reorganization as revealed by the sub-human primate work (Nudo et al., 1996a,b) suggests that there would be differences in the change in cortical activation patterns for reaching and grasping tasks from baseline to post-test/intervention time points between the treated and untreated subjects in our proposed project. Further, and related to the neural network for the control of uni-manual reaching and grasping actions, ipsilateral (or bilateral) hemisphere activation has been shown to be a corollary of task complexity, motor set, and excessive effort (Winstein et al., 1997; Chen et al., 1997a, 1997b; Beltramello et al., 1998; Doyon et al.,1998; Winstein et al., 2000). Motor set effects have been demonstrated most consistently in the motor control literature by comparing predictable and unpredictable order conditions. In an unpredictable condition, motor set is based on a 'best guess' plan and default parameters are selected based on the range of possible responses (see Winstein et al., 2000, for motor set effects in triggered grasp responses). By contrast, under predictable task conditions, a more precise and tuned plan is enabled as the exact response can be planned in advance (Velicki et al., 2000). As such, we hypothesize that as reaching and grasping actions improve with practice, 'task complexity' will be reduced, motor set effects will be enhanced (i.e., the task is planned and performed more easily, with less cognitive demand, Lee et al., 1994), and there should be a concomitant reduction in the extent of ipsilateral (bilateral) neural activity for affected limb movements. At the same time, and consistent with use-dependent cortical reorganization, as skill (learning) improves, there should be an expansion or increased intensity of activation in the appropriate contralateral neural networks of reaching and grasping actions, particularly in regions of interest including M1, S1, SMA and PM areas. These arguments provide the logic for Specific Aim # 3 to determine the changes, as a consequence of CI therapy, in the sensorimotor cortical network through event-related fMRI during specific reaching and grasping tasks, and our associated hypotheses.

By logical extension, we suggest that the effects of CI therapy on the control of upper extremity reaching and grasping actions can be indexed at both the level of cortical neural activation patterns (fMRI) and behavioral kinematics and in relation to well known psychophysical laws of motor learning and control. Further, we suggest that the changes in performance as reflected by physiological brain activation and motor control kinematics can be related to changes in the use of the affected upper extremity as indexed by MAL and WMFT. This companion project is innovative in that it represents the first attempt to correlate the results of psychophysical studies of the 'control' of reaching and grasping actions with neuroimaging (fMRI) studies designed to highlight the 'central motor set' effect (Winstein et al., 2000) using a repeated measures design (3 repeated tests) with a between group factor (treated and untreated-control group). This study complements current research in the exploration of brain activity changes using different motor activation tasks and circumstances that are geared to lend science to the examination of this clinically based therapy.

With this proposal, we plan to assess the changes in neural activation (cortical reorganization) and motor control of reaching and grasping actions that pertain directly to CI therapy compared to that associated with standard and customary care. These experiments will be done on a separate day from the baseline, post-test, and 6-month follow-up tests, but they will be scheduled within 1 week of those tests. This design allows us to explore the physiological neural circuits, and motor control variables mediating the expected behavioral change associated with CI therapy (treated group) and standard and customary care (delayed intervention group) and characterize the persistence of that change in a group of sub-acute stroke survivors who have relatively small lesions and some ability to move the wrist and fingers on the hemiparetic side.