Language-related task-fMRI patterns and restorative online brain–computer interfaces in post-stroke aphasia: a protocol for a systematic review with voxel-based meta-analysis and narrative synthesis

Background Post-stroke aphasia (PSA) can restrict everyday communication, social participation and access to rehabilitation. Functional magnetic resonance imaging (fMRI) during language tasks (task-fMRI) identifies group-level regional activation patterns across distributed language networks, whereas restorative online brain–computer interfaces (BCIs) adjust feedback or training progression contingent on real-time neural signals. Here, restorative denotes a prespecified rehabilitation purpose rather than demonstrated therapeutic benefit. We will examine both evidence streams within language production and comprehension. Methods This protocol describes a systematic review incorporating voxel-based meta-analysis and a focused narrative BCI synthesis. Searches will cover PubMed, Embase via Embase.com, Web of Science Core Collection and APA PsycINFO via Ovid, adding IEEE Xplore for the BCI component. The primary task-fMRI outcome is spatial convergence of positive blood-oxygen-level-dependent responses for task > baseline during production in adults with PSA; comprehension serves as a secondary model. Spatial pooling requires at least 10 independent inputs per domain, accepting compatible whole-brain maps, complete signed coordinate sets, or confirmed whole-brain reports of no suprathreshold peaks in either direction. Domains below this minimum will be described without spatial pooling or vote counting. Eligible BCI studies must have a prespecified restorative aim and evaluate adults with PSA over at least two training sessions featuring online neural control, reporting a longitudinal neural outcome, an independent pre–post language or functional-communication outcome measured without online BCI assistance, or both. BCI implementation and outcomes will still be synthesized narratively when regional classification is not possible. Conclusion The review addresses three predefined objectives: identifying consistent task-fMRI convergence in PSA, describing trained neural features and reported clinical associations in restorative BCIs, and relating reported cortical BCI targets to peaks from the corrected task-fMRI meta-analysis for the corresponding language domain. Trail registration: CRD420261457995.