Stroke can lead to long-term impairment of upper-limb function, particularly in the chronic phase, where recovery is often limited. Brain-computer interface (BCI) systems combined with motor imagery (MI) are emerging as promising neurorehabilitation approaches. Providing real-time neurofeedback during MI may enhance motor recovery by promoting use-dependent neuroplasticity; however, evidence in individuals with chronic stroke remains limited.

This controlled pilot study aims to evaluate the feasibility and preliminary effects of an EEG-based BCI neurofeedback intervention combined with MI and physiotherapy on upper-limb motor function and brain activation patterns in individuals with chronic stroke. Participants are assigned to either an experimental group receiving MI with real-time EEG-based neurofeedback or a control group receiving MI with sham feedback, alongside a standardized four-week physiotherapy program.

Motor function is assessed using validated clinical measures, including the Action Research Arm Test, Fugl-Meyer Assessment, Motor Assessment Scale, and grip strength, with changes interpreted according to clinically meaningful thresholds. Neurophysiological changes are explored through EEG-based measures of brain symmetry and task-related functional MRI.

It is expected that participants receiving real neurofeedback will show greater improvements in upper-limb function and more adaptive brain activation patterns compared to those receiving sham feedback. Findings from this study will help determine the feasibility of this approach and inform the design of larger trials to evaluate its effectiveness in chronic stroke rehabilitation.

This study follows a double-blind design in which both participants and outcome assessors are unaware of group allocation, minimizing potential sources of bias. The inclusion of a sham feedback condition is intended to control for nonspecific effects related to task engagement, attention, and exposure to the intervention environment.

The neurofeedback protocol is based on real-time modulation of EEG-derived sensorimotor rhythms during motor imagery tasks. In the experimental group, feedback is contingent on participants' brain activity, enabling reinforcement of targeted neural patterns. In the control group, feedback is non-contingent but visually similar, ensuring maintenance of blinding.

To ensure intervention fidelity, the protocol is delivered over a fixed duration with standardized procedures across participants. The integration of the neurofeedback intervention within a structured rehabilitation schedule is designed to minimize variability and support consistency in delivery.

Outcome interpretation incorporates minimal clinically important difference (MCID) thresholds to distinguish clinically meaningful improvements. In addition, the combined use of EEG and task-related functional MRI provides complementary insight into neuroplasticity, capturing both temporal dynamics and spatial patterns of brain reorganization.

Feasibility is further evaluated through indicators such as adherence to the intervention protocol, participant retention, and the practicality of implementing EEG-based neurofeedback in a rehabilitation context. These elements are intended to inform protocol optimization and the design of future randomized controlled trials.

Inclusion Criteria:

• Adults aged ≥18 years
• First-ever cortical ischemic stroke affecting the left middle cerebral artery (MCA) territory
• ≥12 months since stroke onset
• Stroke diagnosis confirmed by structural neuroimaging (MRI or CT)
• Lesion localization confirmed based on clinical and radiological criteria

Exclusion Criteria:

• Complete hand plegia
• Presence of language or cognitive impairment, as assessed by the Montreal Cognitive Assessment (MoCA)
• Presence of attentional deficits
• Presence of visual impairments that may interfere with task performance
• Diagnosis of depression
• Pre-stroke dependence in activities of daily living
• Left-handedness, as assessed by the Edinburgh Handedness Inventory
• Inability to perform motor imagery tasks
• Contraindications to MRI (e.g., metallic implants, implanted electronic devices, or claustrophobia)

Individual participant data will not be shared due to privacy and confidentiality concerns. The dataset contains detailed neurophysiological and behavioral measures from a small sample of chronic stroke patients, which could potentially allow identification of participants. Data sharing was also not included in the informed consent, and ethical approvals were obtained based on participant privacy protections.

The study design includes double blinding, in which participants and outcome assessors are unaware of group allocation. This approach minimizes potential bias from expectation effects, observer influence, or differences in engagement between groups.