Background Brain computer interface (BCI) technology, as an innovative means to improve the limb function of stroke patients, has received extensive attention in recent years. However, there are few reports on the application of BCI technology in robot assisted stroke rehabilitation training and its impact on the rehabilitation effect of patients.

Objective To evaluate the clinical efficacy of robotic-assisted rehabilitation training for subacute stroke patients under the aegis of BCI technology and elucidate its influence on pertinent serum biomarkers.

Methods This study was a single-center, randomized, controlled clinical trial . Subacute stroke patients admitted to the First Affiliated Hospital of Air Force Medical University from March 2017 to August 2022 were selected. All the patients were randomly assigned into two groups using an envelope method: the control group received conventional rehabilitation training, while the BCI group underwent robot-assisted rehabilitation training with BCI technology assistance. The intervention period was 4 weeks. Motor function assessments and biochemical indicators were evaluated and compared between the two groups at baseline (T0), 1 week post-intervention (T1), 2 weeks post-intervention (T2), and 4 weeks post-intervention (T3).

Results A total of 57 patients participated in the study, with 27 cases in each group. Ultimately, the control group included 25 cases (20 males and 5 females, with an average age of 49.33 ± 10.51 years), and the BCI group included 24 cases (16 males and 8 females, with an average age of 46.98 ± 8.56 years). There were no statistically significant differences between the two groups (P＞0.05). At baseline, there were no statistically significant differences in the Loewenstein Occupational Therapy Cognitive Assessment (LOTCA) scores, Fugl-Meyer assessment for lower extremity (FMA-LE) scores, and Fugl-Meyer assessment for balance (FMA-B) scores between the two groups (P＞0.05). However, at the end of the intervention, scores in both groups improved compared to baseline, with statistically significant differences (P＜0.05). Between-group comparisons showed that the LOTCA score in the BCI group was significantly higher than that in the control group (81.56 ± 13.26 vs 73.56 ± 12.81, P=0.037), but the differences in FMA-LE (18.98 ± 5.16 vs 16.33 ± 7.05) and FMA-B scores (9.55 ± 2.98 vs 8.35 ± 3.56) were not statistically significant (P＞0.05). There were no statistically significant differences in neurotrophic factors (BDNF) between the groups at baseline and at the end of the intervention (P＞0.05). However, at the end of the intervention, both groups had significantly higher BDNF levels compared to baseline, and the change in the BCI group was significantly greater than that in the control group (16.90 ± 6.35 pg/mL vs 6.00 ± 2.51 pg/mL, P＜0.001).

Conclusion The integration of BCI-controlled robotic training with conventional rehabilitation protocols holds promise for the amelioration of cognitive functions, enhancement of lower limb motor capabilities, and elevation of BDNF levels in subacute stroke patients.