Stroke is the most common cause for physical disability and impairments to lower limb function remain one of its most debilitating symptom. Motor imagery (MI), as a safe, self-paced technique, has been shown to effectively facilitating the effects of motor practice. When combined with brain-computer interface (MI-BCI), it also demonstrates an improvement in stroke motor recovery. A feasibility trial was carried out to investigate the effect of MI-BCI neurofeedback in chronic hemiplegic lower limb rehabilitation. The neurophysiological correlates to clinical outcomes was also studied by using, transcranial magnetic stimulation (TMS).

Subjects (n=13) with more than 9 months post-stroke and Functional Ambulation Category 3–4 underwent 12 sessions of MI-BCI gait training, at a frequency of thrice a week. Subjects were instructed to perform a MI task whereby they imagined themselves walking properly with both legs. If the MI task is performed correctly as detected via electroencephalography acquisition, a pair of cartoon footprints in the monitor will be activated to walk forward. Each MI-BCI session includes 160 MI trials with resting interval every 40 trials. Timed up-to-go test and 10 meter walk test, as well as the resting motor threshold measured by TMS were performed before, after and 6 weeks after MI-BCI gait training.

It was shown that MI-BCI was safe and well tolerated by stroke subjects. Both walking speed and balance improved after MI-BCI gait training (Fig. 1). This was in line with an increase in the corticospinal activity in the contralesional M1 motor cortex (Fig. 2).

MI-BCI could improve mobility in chronic stroke patients with residual mobility impairment. The study also suggested that the contralesional motor cortex is involved in the recovery of mobility.