Answers are grounded in this paper and cite their sources.
Background Post-stroke cognitive impairment (PSCI) affects approximately 30–60% of stroke survivors, yet effective interventions remain limited. Brain–computer interface (BCI) technology has emerged as a promising neurorehabilitation tool, but its efficacy specifically for cognitive outcomes has not been systematically quantified. Objective To provide a focused quantitative synthesis of recent randomized controlled trials (RCTs, 2021–2026) evaluating the efficacy of BCI-based interventions on cognitive function in patients after stroke. Methods PubMed, the Cochrane Central Register of Controlled Trials (CENTRAL), Embase, and Web of Science were searched from January 2021 through May 31, 2026. RCTs comparing BCI-based interventions with conventional rehabilitation or sham control in stroke patients and reporting at least one standardized cognitive outcome were eligible. The primary outcome was global cognitive function measured with the Montreal Cognitive Assessment (MoCA). Pooled analyses used random-effects models with restricted maximum likelihood (REML) estimation of τ 2 and Hartung–Knapp–Sidik–Jonkman (HKSJ) confidence intervals and tests based on a t -distribution; Hedges’ g was the effect measure and 95% prediction intervals were computed. Risk of bias was assessed with the Cochrane RoB 2 tool and certainty of evidence with GRADE. Results Of 663 records identified, 544 unique records were screened and 40 full-text reports were assessed for eligibility; 7 RCTs involving 405 randomized participants were included. The primary global cognition analysis comprised two pairwise comparisons from one three-arm trial ( n = 89): the pooled effect was g = 0.92 (95% CI –0.08 to 1.92, p = 0.054), which did not reach statistical marginally significant. For attention (3 trials, n = 217), BCI-based interventions produced a large improvement ( g = 0.86, 95% CI 0.58 to 1.14, p < 0.001), which remained robust in a sensitivity analysis restricted to performance-based measures ( g = 0.86, 95% CI 0.43 to 1.29, p < 0.001). Single small trials suggested improvements in working memory ( g = 1.38, 95% CI 0.56 to 2.19) and processing speed ( g = 0.94, 95% CI 0.21–1.68). Certainty of evidence was moderate for attention, low for global cognition, and very low for the remaining domains. Conclusion In this focused synthesis of recent RCTs, BCI-based interventions show a promising short-term average benefit on attention in patients after stroke, whereas the effect on global cognition remains uncertain and did not reach statistical significance. Paradigm type, training dose, and timing of intervention should be regarded as hypotheses for future trials rather than clinical recommendations. Adequately powered, sham-controlled RCTs with standardized cognitive outcomes and longer follow-up are required. Systematic review registration https://www.crd.york.ac.uk/PROSPERO , identifier CRD420261415579.
No takes yet. Share an insight, caveat, or question.
Yang et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: