Introduction: A better understanding of brain changes underlying functional gains with brain-computer interface (BCI) use in stroke rehabilitation is needed.
Hypothesis: BCI therapy will produce observable neuroplastic changes, and behavioral gains will relate to changes in diffusion tensor imaging (DTI) metrics of the corticospinal tracts (CSTs).
Methods: We obtained DTI and behavioral data from 16 stroke patients with upper extremity motor impairment before, during, and after BCI therapy. Additional DTI and behavior data were obtained from 9 control subjects at similar time intervals before receiving any BCI therapy. Behavioral measures included Stroke Impact Scale (SIS), Action Research Arm Test (ARAT), 9-Hole Peg Test (9HPT), and grip strength. We performed deterministic tractography of the ipsilesional CST (iCST) and contralesional CST (cCST) and extracted DTI metrics (fractional anisotropy, axial diffusivity, radial diffusivity, trace) from each. DTI and behavioral metrics were analyzed for changes from baseline between therapy and control groups. Individual changes in DTI metrics were analyzed for correlation with behavior changes during the therapy period.
Results: Group by time interactions were identified in DTI metrics of the iCST, with multiple metrics increasing with therapy. Changes in individual SIS Activities of Daily Living and 9HPT scores correlated with individual changes in both the iCST and cCST. Similarly, changes in SIS Hand Function correlated with cCST changes, and changes in grip strength correlated with iCST changes. Correlations between changes in ARAT scores and cCST changes trended to significance. Correlations significant at p<= 0.05; trending at 0.05
Conclusions: BCI therapy can induce structural neuroplastic changes during stroke recovery. These changes may be used to track both subjective and objective behavioral gains.