Activation Level (HbO concentration)
For both stroke and healthy groups, MT activated both sides of the motor cortex, but greater activation was observed on the ipsilateral hemisphere to the robotic glove, which was contralateral to the hand that performed the movement (Fig.
8). RT primarily activated the motor cortex in the hemisphere contralateral to the robotic glove, while RMT induced activation of both sides (Fig.
8) with similar magnitude. RMT induced the strongest activation on contralateral channels (Fig.
8). Particularly, on contralateral side of RMT, channel 5 and 8 showed the highest activation (Fig.
8), which are located around C4 (or C3) that is known to be the area responsible for hand movement [
36]. MT induced activation on both sides, but stronger activation on ipsilateral hemisphere to the robotic glove, which is contralateral to the hand executing the pinch movements. RT showed primary activation on the contralateral hemisphere but its magnitude was smaller than in RMT.
The group-level contrast is shown in Fig.
9. Significantly greater activation on the contralateral hemisphere was shown for stroke survivors in RMT condition compared to RT condition even though both conditions involved the robotic glove moving the hand. Specifically, RMT induced greater activation on the contralateral primary motor cortex (channel 3) and contralateral somatosensory cortex (channel 10) compared to RT (Stroke RMT-RT, Fig.
9a). This effect was not observed in the healthy group. As expected due to the presence of contralateral hand moment, greater motor cortex activation was observed in both stroke and healthy groups on the hemisphere ipsilateral to the robotic glove in RMT compared to RT. The healthy group showed significantly greater activation on the contralateral primary motor cortex in RMT condition compared to MT condition (channels 6, 8, and 18; Fig.
9a) whereas contralateral activation magnitude was greater RMT compared to MT in stroke group but it did not reach significance at the group level (0.1 < p < 0.05). MT induced significantly smaller activation on contralateral primary motor cortex (channel 8) compared to RT in healthy group, while no significant differences were shown in the cortical activation in the contralateral hemisphere between MT and RT in the stroke group. For the ipsilateral hemisphere, MT showed significantly larger activation than RT because of the voluntary movement of the contralateral hand (unaffected hand for stroke subjects and nondominant hand for healthy subjects).
The cortical activation showed significant differences between subject groups (stroke vs. healthy) in RMT condition on the ipsilateral premotor cortex (channel 45) and contralateral primary motor cortex (channel 3 and 18), while no significant difference was observed between subject group in RT and MT condition (Fig.
9b).
Additional analysis was conducted at the individual level in participants post-stroke to see how training conditions affect the brain (Fig.
10) for each individual. P4 was excluded because there were no channels that satisfy the significant channel selection conditions. Only one participant (P5) had significant activation the contralateral hemisphere to the affected hand during MT whereas all four showed significant activation on the ipsilateral hemisphere that was contralateral to the hand being moved. P1, P3 and P5 showed significant contralateral hemisphere activation during RMT whereas P2 retained a similar pattern as MT with only significant activation on the ipsilateral hemisphere to the robotic glove for RMT. P5 showed almost the same activation on the contralateral side for MT and RMT. For RT, P2 and P3 showed no significant activation on both contralateral and ipsilateral hemispheres. P1 and P5 had activated channels on the contralateral hemisphere for RT, but the magnitude and number of activated channels were less than in RMT.