Revolutionary Stroke Rehabilitation: Upper and Lower Extremity Training with BCI System
PublicationsUpper extremity training followed by lower extremity training with a Brain-Computer Interface rehabilitation system
Stroke Rehabilitation remains a significant global health concern, affecting mortality and disability rates despite advancements in prevention and treatment. Motor recovery after stroke is often inadequate, with many survivors facing upper and lower extremity impairments impacting daily activities. Brain-computer interfaces (BCIs) offer a promising avenue for motor rehabilitation, utilizing neural activity to control external devices. This study investigates the efficacy of a BCI system employing motor imagery, functional electrical stimulation (FES), and visual feedback using a 3D avatar in improving both upper and lower extremity function in stroke patients. Nineteen participants underwent consecutive BCI treatments, with assessments conducted before and after each treatment to evaluate functional outcomes.
Read the original article in frontiers
The study implemented a novel therapy approach for stroke patients utilizing Brain-Computer Interface (BCI) technology combined with Functional Electrical Stimulation (FES) and realistic 3D avatar visual feedback. Patients wore wireless EEG caps with 16 active electrodes covering the sensorimotor cortex during therapy sessions. For upper extremity (UE) therapy, patients sat with their forearms resting on a desk with surface FES electrodes attached to wrist extensors. During lower extremity (LE) therapy, patients sat with the affected leg slightly elevated, and FES electrodes were placed on wrist and foot dorsiflexors.
BCI-enhanced Motor Imagery therapy shows promising results in clinical assessments
During therapy sessions, patients engaged in Motor Imagery (MI) tasks, imagining dorsiflexion of the respective side. The BCI provided synchronous visual and proprioceptive feedback if the MI instruction matched the classified MI side. Therapy sessions consisted of three runs with 40 MI trials each, and EEG data were used to train the classifier. Clinical assessments included scales such as Fugl-Meyer Assessment (FMA), Barthel Index (BI), modified Ashworth Scale (MAS), 10-Meter Walk test (10MWT), and Timed Up and Go (TUG) test. Statistical analyses were performed to assess changes in clinical scales and BCI performance.
Study reveals key patient characteristics and treatment metrics in stroke rehabilitation
Baseline characteristics of patients showed a median age of 53.1 years, with a median time since stroke of 23.6 months. Most patients had ischemic stroke with various lesion locations. The median FMA-UE score before UE treatment was 19.0 points. Median time between treatments was 7.4 months, with a median walking speed of 1.2 m/s before LE treatment. Statistical analyses aimed to evaluate MI learning effects and the relationship between BCI performance during UE and LE treatments, with p-values corrected for multiplicity.
Improvements after BCI treatments for upper and lower extremities
The study observed improvements in motor function and daily living activities in stroke patients undergoing brain-computer interface (BCI) treatments for upper extremities (UE) and lower extremities (LE). For UE treatment, patients showed significant increases in Fugl-Meyer Assessment for Upper Extremity (FMA-UE) scores (p < 0.001), Barthel Index (BI) scores (p < 0.001), and reductions in wrist and finger spasticity (p < 0.001). Similarly, LE treatment resulted in improved gait speed (p = 0.001), BI scores (p = 0.049), and reduced ankle spasticity (p = 0.011).
Clinically relevant thresholds were reached for FMA-UE, BI, Modified Ashworth Scale (MAS), 10-Meter Walk Test (10MWT), and Timed Up and Go Test (TUG). There were no significant differences in BI and FMA-UE changes between UE and LE treatments, but combined treatment showed significant improvements in both FMA-UE (p = 0.002) and BI (p = 0.007). BCI performance improved during UE treatment sessions (p = 0.020) but not during LE treatment sessions (p = 0.102). Median BCI performance was higher in LE compared to UE treatment (p = 0.020), with moderate correlation between UE and LE BCI performances (ρ = 0.614, p = 0.020).
The study evaluated the effects of Brain-Computer Interface (BCI) treatments on upper extremity (UE) and lower extremity (LE) motor function in stroke patients. For UE treatment, the Fugl-Meyer Assessment for Upper Extremity (FMA-UE) was used, showing a significant average improvement of 4.2 points, with 22% average improvement. Most patients, despite severe impairment, demonstrated notable improvement. Additionally, Activities of Daily Living (ADLs) and wrist spasticity improved, even in the patient who didn’t show FMA-UE improvement. Similarly, LE treatment showed enhanced walking speed (10MWT) by 0.15 m/s on average, with 23% improvement. Patients also improved in ADLs and ankle spasticity decreased.
The study highlighted comparable improvements with previous BCI studies, although employing bilateral training and different feedback mechanisms. Despite severe impairment, patients showed motivation and significant progress in both UE and LE treatments. Moreover, the study revealed sustained improvements in motor function even after treatment cessation. Patients reported anecdotal evidence of improved daily activities, mobility, and quality of life.
Future studies aim to explore extended treatment sessions and variations in training intensity. Notably, patients demonstrated learning and improvement in MI performance over time, indicating intrinsic capabilities for BCI utilization. Although BCI performance was generally higher during LE treatment, it correlated with UE treatment performance, suggesting consistent patient-specific factors influencing BCI performance. Classification accuracy for BCI control surpassed significance thresholds, comparable to performance in healthy individuals, showcasing the feasibility and effectiveness of BCI interventions in stroke rehabilitation.
Study identifies limitations in upper extremity Brain-Computer Interface training research
The current study acknowledges certain limitations that should be addressed in future research. Although improvements following the upper extremity (UE) Brain-Computer Interface (BCI) training align with findings in existing literature, the lack of a separate UE BCI group as a control limits the ability to compare improvements achieved by patients undergoing both UE and lower extremity (LE) BCI training against those solely undergoing LE BCI training. Additionally, while the patient sample exhibits heterogeneity in stroke type and lesion location, its size remains small, necessitating larger cohorts for more comprehensive insights into observed improvements and their correlation with stroke characteristics.
Real-time monitoring reveals neuroplasticity’s role in functional recovery
Sebastián-Romagosa et al. (2020b) highlight BCI therapies’ unique potential to monitor brain activity during treatment, offering insights into neuroplasticity mechanisms driving functional improvements. In the current study, 19 stroke patients underwent BCI training involving motor imagery (MI), Functional Electrical Stimulation (FES), and a 3D avatar over 25 sessions each for upper and lower extremities. Real-time feedback was provided through FES and the avatar. Significant improvements in targeted extremities and daily activities were noted post-UE BCI treatment, including motor function enhancements and reductions in spasticity. Subsequent LE BCI treatment further improved ankle spasticity, mobility, balance, and walking speed, indicating additional benefits from sequential treatments.
Data accessibility is restricted due to privacy regulations, and requests to access datasets should be directed to the responsible party. Ethical approval was obtained, and participants provided informed consent. Funding for the study was provided by FFG, Austria.
