Motor Cortex Activation During Treatment May Predict Therapeutic Gains in Paretic Hand Function After Stroke
Background and Purpose— Functional brain imaging after stroke offers insight into motor network adaptations. This exploratory study examined whether motor cortical activation captured during arm-focused therapy can predict paretic hand functional gains.
Methods— Eight hemiparetic patients had serial functional MRI (fMRI) while performing a pinch task before, midway, and after 2 weeks of constraint-induced therapy. The Wolf Motor Function Test (WMFT) was performed before and after intervention.
Results— There was a linear reduction in ipsilateral (contralesional) primary motor (M1) activation (voxel counts) across time. The midpoint M1 Laterality Index anticipated post-therapeutic change in time to perform the WMFT. The change in ipsilateral M1 voxel count (pre- to mid-) correlated with the change in mean WMFT time (pre- to post-).
Conclusions— The relationship between brain activation during treatment and functional gains suggests a use for serial fMRI in predicting the success and optimal duration for a focused therapeutic intervention.
Functional MRI (fMRI) has revealed reorganization in the primary and secondary motor cortices during poststroke recovery and after therapeutic interventions.1 Few studies have explored the evolution of brain activation in relation to behavioral gains in a “one-to-one” correspondence during a specific rehabilitation intervention.2 This exploratory study examined whether the brain activation midway through a 2-week arm-focused intervention might capture adaptations induced by the initial week of training and, in turn, could be used to anticipate post-therapeutic behavioral changes in paretic hand function. If so, this brain–behavior correspondence may offer guidance to determine an optimal duration for task-specific therapy.2
Subjects and Methods
Eight patients with hemiparetic stroke (Fugl-Meyer [FM] motor score 33 to 62) participated. Inclusion criteria were >3 months after stroke, ability to perform the fMRI task, and a minimum of 10° of voluntary wrist and finger extension. Lesions varied in location, but all spared the hand motor representation (M1). No alternative therapy group was studied. Seven healthy volunteers were scanned twice to test the reproducibility of fMRI activation.
Physical Therapy and Functional Measure
All patients received constraint-induced therapy for 2 weeks as defined for the EXCITE trial.3 The Wolf Motor Function Test (WMFT)4 was performed before and after intervention. The behavioral outcome measure consisted of 6 dexterity items from the full 15-item WMFT (Lift Can; Lift Pencil; Lift Paper Clip; Stack Checkers; Flip Cards; Turn Key in Lock) that most directly captured fine motor control. The change in mean WMFT (mWMFT) time for the 6-item subset was correlated with that for the 15-item test (r=0.98), indicating reliability and validity for the subset. The pre-mWFT–post-mWFMT (absolute time) difference was used as a proxy for functional change in motor skill.
fMRI acquisition parameters were described previously.5 fMRI sessions were performed before intervention, midintervention, and after intervention, each with 4 30-s bouts of repetitive pinch alternating with 5 30-s rest periods. The pinch apparatus included a vertical plastic tube connected to a pressure transducer. The task required tube compression with the index and middle fingers against the thumb, creating enough pressure to match 50% of maximum, viewed through goggles as a target line, and paced by auditory cues at 75% maximum rate. These parameters were maintained constant across the 3 sessions. Practice before each fMRI session minimized unwanted movements and deviations from consistent task performance.
fMRI data were analyzed as described previously.5,6 Volumes related to head motion (>2 mm), and associated movements (visually identified from videotape) were excluded. Z statistic images were thresholded at Z>3.1, and significant clusters were defined atP<0.01 (corrected for multiple comparisons). Regions of interest (ROIs) were set in bilateral M1 and dorsal premotor (PMd) areas. Percentage signal change (% SC) and voxel counts (VCs) within each ROI were measured and a Laterality Index [LI=(contralateral−ipsilateral)/(contralateral+ipsilateral)] (contralateral and ipsilateral activation to the hand movement. LI ranges from −1 [all ipsilateral activation] to 1 [all contralateral activation]) was calculated using VC for each ROI. Linear Mixed Model was used for intersession comparisons of fMRI variables (% SC, VC), pinch pressure, and rate, separately. Individual linear regression analyses were performed between LI, VC (M1 and PMd; independent variable) pre-, mid-, and post- and the post-pre–mWMFT time difference (dependent variable). Pearson correlation coefficient analysis was used to assess the relationship between changes in fMRI measures and changes in mWMFT time. Preintervention fMRI from patients 5 and 6 was technically unusable.
Prefunctional to postfunctional gains (mWMFT) varied across patients, but the group 6-item time decreased for the paretic hand after therapy (P=0.03; Table).
No differences were detected in pinch pressure or rate across sessions (P>0.1). Intersession comparisons of M1 activation in healthy volunteers showed no differences (Table). Group analysis for the paretic hand showed a continuous reduction of VC in ipsilateral M1 (P=0.02; P=0.006 linear trend) across time (Table). No differences in M1 activation across time were found for the less-affected hand (P>0.1; data not shown). We observed 4 patterns of LI evolution for M1, including a progressive increase (patients 3, 4, and 7; Figure 1A), a midpoint-only increase (patient 8), a midpoint decrease (patient 1; Figure 1B), and nearly no change (patient 2). Among the 3 showing “progressive increase,” patients 3 and 4 (FM score 53 and 54, respectively) had either an increase in contralateral or a decease in ipsilateral M1 activation across time, whereas patient 7 (FM score 45) demonstrated a continuous reduction in bilateral M1 activation but more so ipsilaterally. The “midpoint decrease” in patient 1 (FM score 62), who was well recovered and showed the least functional improvement, was attributed to a pre- to mid- reduction in contralateral M1 activation. The “midpoint-only increase” in patient 8 (FM score 34), who showed the most functional improvement, resulted from a pre- to mid- decrease in ipsilateral M1 activation.
There was no correlation between post-pre change in mWMFT time and change in activation (VC or % SC) in M1 or PMd (ipsilateral or contralateral), except for that between post-pre change in mWMFT time and pre- to- mid- change in ipsilateral M1 activation (VC; r=0.82; P=0.05). The midpoint and postintervention LI for M1 and midpoint ipsilateral M1 VC, but not that for PMd, did predict the post-pre mWMFT time change (6-item; Figure 2).
The midpoint and postintervention LIs for M1 were found to predict paretic hand functional improvement. The midpoint ipsilateral activation (VC) was responsible for this brain–behavior correspondence (Figure 2B). Additionally, the ipsilateral VC (M1) evolved dynamically through progressive focusing during task-oriented training. Moreover, the reduction in ipsilateral M1 activation (VC) after the first week of training correlated with post-therapeutic behavioral improvements.
The small sample size and variability across patients in initial impairment level limit generalizability. Our ability to detect mirror movements from videotape was less accurate than other methods. However, the reproducibility of task performance across sessions in the patients and the similarity of task-related activation in healthy volunteers and in stroke patients for the less-affected hand suggest that these findings are most likely related to adaptive changes in brain activity with therapy and not to confounders such as differences in task performance or to undetected mirror movements. The results warrant further exploration in a larger study.
We selected M1 and PMd based on their reported role in stroke motor recovery.1,7 Other areas, such as ventral premotor and supplementary motor areas and cerebellum may also evolve with focused therapy.8 The predictive value of M1 activation and the correlates between changes in M1 activation and functional improvements, if replicated in patients across various lesion locations, degrees of impairment, and multiple behavioral assessments, has implications for using serial fMRI as a physiological indicator for “dose–response” interactions during a rehabilitative intervention.2 In addition, the relationships among initial impairment level, evolution of M1 activation, and functional gains may help establish the optimal duration of a rehabilitation intervention and lead to a better understanding of adaptive brain–behavior responses to therapy.
The study was supported by National Institutes of Health grants NS 45485 to C.J.W. and R24 HD 39629 (Western Medical Rehabilitation Research Network) to B.H.D. We thank the University of Southern California clinical team of Michelle Prettyman, Samantha Underwood, Chris Hahn, Janice Lin, and Jarugool Tretriluxana for patient recruitment and behavioral assessments, and Vikas Rao for postprocessing and analysis of fMRI data. The authors thank the generous support to the Ahmanson-Lovelace Brain Mapping Center and the National Center for Research Resources grants RR12169, RR13642, and RR08655.
- Received December 8, 2005.
- Revision received January 31, 2006.
- Accepted March 8, 2006.
Calautti C, Baron JC. Functional neuroimaging studies of motor recovery after stroke in adults: a review. Stroke. 2003; 34: 1553–1566.
Dobkin BH. Rehabilitation and functional neuroimaging dose-response trajectories for clinical trials. Neurorehabil Neural Repair. 2005; 19: 276–282.
Winstein CJ, Miller JP, Blanton S, Taub E, Uswatte G, Morris D, Nichols D, Wolf S. Methods for a multisite randomized trial to investigate the effect of constraint-induced movement therapy in improving upper extremity function among adults recovering from a cerebrovascular stroke. Neurorehabil Neural Repair. 2003; 17: 137–152.
Wolf SL, Thompson PA, Morris DM, Rose DK, Winstein CJ, Taub E, Giuliani C, Pearson SL. The EXCITE trial: attributes of the Wolf Motor Function Test in patients with subacute stroke. Neurorehabil Neural Repair. 2005; 19: 194–205.
Smith SM, Jenkinson M, Woolrich MW, Beckmann CF, Behrens TE, Johansen-Berg H, Bannister PR, De Luca M, Drobnjak I, Flitney DE, Niazy RK, Saunders J, Vickers J, Zhang Y, De Stefano N, Brady JM, Matthews PM. Advances in functional and structural MR image analysis and implementation as FSL. NeuroImage. 2004; 23 (suppl 1): S208–219.
Ward NS, Brown MM, Thompson AJ, Frackowiak RS. Neural correlates of motor recovery after stroke: a longitudinal fMRI study. Brain. 2003; 126: 2476–2496.
Johansen-Berg H, Dawes H, Guy C, Smith SM, Wade DT, Matthews PM. Correlation between motor improvements and altered fMRI activity after rehabilitative therapy. Brain. 2002; 125: 2731–2742.