The effect of brain–computer interface training on cognitive function in stroke patients: a systematic review and meta-analysis

This study aims to assess the therapeutic impact of BCI-based interventions on global and domain-specific cognitive functions (attention, memory, and executive function), and activities of daily living in stroke survivors. Furthermore, we seek to identify the potential moderating effects of feedback modes and BCI paradigms on the overall rehabilitative efficacy.

A systematic search of PubMed, Embase, Web of Science, the Cochrane Library, and CNKI databases was conducted to identify eligible randomized-controlled trials (RCTs). Meta-analyses were performed by pooling standardized mean differences (SMDs) to synthesize effect sizes. To explore sources of heterogeneity and the effects of potential moderators, subgroup analyses were conducted according to outcome measures, stroke phase, BCI paradigm, and feedback type.

Twelve studies were included. The meta-analysis demonstrated that BCI training significantly improved global cognitive function (SMD = 0.62, P < 0.00001), attention, and executive function, alongside enhanced activities of daily living performance. However, no significant improvement was observed in memory function. Subgroup analyses revealed that superior and more robust effects were associated with subacute patients, active BCI paradigms, and multimodal feedback (visual + auditory + proprioceptive).

BCI training is an effective intervention for post-stroke cognitive recovery. Early initiation of therapy and the integration of multimodal feedback appear to be critical factors for maximizing therapeutic outcomes.