NeuroImage 11, Number 5,2000, Part 2 of 2 Parts 1 D Ekk@ DISORDERS - NEUROLOGY Event-related fMR1 of Monitored Movements in Stroke Patients and Healthy Controls Jennifer Newton*, Kyung Peck?, Steven Butterworth+, Andrew Peters?, Penny Gowiandt, Alan Sunderland* *Division of Stroke Medicine, tMagnetic Resonance Centre, $Division of Neurology, University of Nottingham, NG7 2RD, UK Introduction There are many proposed mechanisms for the recovery of motor function after stroke’. Neuroimaging studies have provided important insights into these mechanisms. However, most neuroimaging studies of subjects recovered from hemiparetic stroke have employed block designs with PET or fMRI with subjects instructed to carry out fine finger movements for prolonged periods. This study uses an event-related paradigm with wrist extension to reduce the physical demands during testing on stroke patients and to enable inclusion of patients with partial hand recovery. In addition, the amplitude of the motor output was monitored in both hands during testing. Methods We studied four patients with a first completed ischemic stroke. Five healthy subjects were also tested. All patients had right hemisphere damage. Two patients had lesions of the white matter underlying the primary motor cortex, whereas the other two had cortical damage including at least part of the primary motor cortex. At the time of admission, all patients presented with hemiparesis with severe impairment/complete loss of hand function. At the time of the fMRI scans, 6-8 months after stroke, hand function had returned at least to the extent of controlled wrist extension. All fMRI measurements were performed on a 3T scanner with a linear RF coil, using an EPI sequence (TE=30ms, 3mm in plane resolution, 9mm slice thickness). Ten contiguous coronal images were acquired every 1.87 seconds. A pressure bulb was positioned in light contact with the back of each hand to allow compression by near-isometric wrist extension. The pressure produced an electrical signal which was sampled every 187 msec during the experiment and recorded using a PC. Feedback of the amplitude of the pressure to the subject was via a thermometer-type display viewed through prismatic glasses. The experiment was conducted in two blocks: one for movements of each hand. Each trial began when the subject was cued by the appearance of two arrows on the display that indicated the required force amplitude to be generated. Subjects were asked to extend their wrist until the pressure level on the display reached the target arrows and then relax, allowing the pressure level to return to baseline. There were thirty trials per block, each lasting 18.7 seconds. There were two target force levels, at 10% and 20% of the maximum voluntary contraction measured for that hand, which alternated between trials. The fMRI data was then motion-corrected, globally normalised, spatially and temporally filtered. Using a box-car waveform (modelled on a movement duration of approx. 1.8s) convolved with a gamma variate function, correlation maps were produced and thresholded at p<O.OOl uncorrected. Clusters of 5 or more significantly activated pixels were localised following co-registration with an anatomical image. ReSUltS All control subjects showed clusters of activation in the contralateral primary motor cortex (PMC) and the supplementary motor area (SMA). Four of these subjects also showed activation of lateral premotor cortex and supramarginal gyms activation, whereas only three showed activation of the superior parietal lobule. However, three patients showed activation of the contralesional PMC during movements of the recovered hand. Since motor output of both hands was monitored and movements of the unaffected limb were not observed in any of the patients during movements of the recovered hand, synkinesia cannot account for these results. Bilateral activation of the primary motor cortex was only seen in one of these patients, where spared motor cortex adjacent to the lesion was activated. A fourth patient showed activation of the ipsilesional lateral premotor cortex for movements of the paretic hand, in the absence of PMC activation in either hemisphere. Overall, the stroke patients showed recruitment of a larger motor network during movements of the recovered upper limb, consisting of contralesional primary motor cortex, ipsilesional lateral premotor and supplementary motor areas. Conchsions Our data supports several of the proposed mechanisms of recovery of motor function after stroke. Movement of the recovered hand after damage to the PMC and its underlying white matter appears to depend on combinations of activation of(i) undamaged PMC adjacent to the lesion, (ii) the contralesional PM@, presumably based on uncrossed descending tracts, or (iii) non-primary motor areas’, via their direct corticospinal connections and corticocortical connections. References 1. Steinberg BA & Augustine JR, Brain Research Reviews (1997), 25:125-132. 2. Cao Y et al, Stroke (1998), 29:112-122. 3. Seitz RJ et al, Archives of Neurology (1998), 55:1081-1088. s119