The clinical effects of brain-computer interface with robot on upper-limb function for post-stroke rehabilitation: a meta-analysis and systematic review

PURPOSE

Many recent clinical studies have suggested that the combination of brain-computer interfaces (BCIs) can induce neurological recovery and improvement in motor function. In this review, we performed a systematic review and meta-analysis to evaluate the clinical effects of BCI-robot systems.

METHODS

The articles published from January 2010 to December 2020 have been searched by using the databases (EMBASE, PubMed, CINAHL, EBSCO, Web of Science and manual search). The single-group studies were qualitatively described, and only the controlled-trial studies were included for the meta-analysis. The mean difference (MD) of Fugl-Meyer Assessment (FMA) scores were pooled and the random-effects model method was used to perform the meta-analysis. The PRISMA criteria were followed in current review.

RESULTS

A total of 897 records were identified, eight single-group studies and 11 controlled-trial studies were included in our review. The systematic analysis indicated that the BCI-robot systems had a significant improvement on motor function recovery. The meta-analysis showed there were no statistic differences between BCI-robot groups and robot groups, neither in the immediate effects nor long-term effects (p > 0.05).

CONCLUSION

The use of BCI-robot systems has significant improvement on the motor function recovery of hemiparetic upper-limb, and there is a sustaining effect. The meta-analysis showed no statistical difference between the experimental group (BCI-robot) and the control group (robot). However, there are a few shortcomings in the experimental design of existing studies, more clinical trials need to be conducted, and the experimental design needs to be more rigorous.Implications for RehabilitationIn this review, we evaluated the clinical effects of brain-computer interface with robot on upper-limb function for post-stroke rehabilitation. After we screened the databases, 19 articles were included in this review. These articles all clinical trial research, they all used non-invasive brain-computer interfaces and upper-limb robot.We conducted the systematic review with nine articles, the result indicated that the BCI-robot system had a significant improvement on motor function recovery. Eleven articles were included for the meta-analysis, the result showed there were no statistic differences between BCI-robot groups and robot groups, neither in the immediate effects nor long-term effects.We thought the result of meta-analysis which showed no statistic difference was probably caused by the heterogenicity of clinical trial designs of these articles.We thought the BCI-robot systems are promising strategies for post-stroke rehabilitation. And we gave several suggestions for further research: (1) The experimental design should be more rigorous, and describe the experimental designs in detail, especially the control group intervention, to make the experiment replicability. (2) New evaluation criteria need to be established, more objective assessment such as biomechanical assessment, fMRI should be utilised as the primary outcome. (3) More clinical studies with larger sample size, novel external devices, and BCI systems need to be conducted to investigate the differences between BCI-robot system and other interventions. (4) Further research could shift the focus to the patients who are in subacute stage, to explore if the early BCI training can make a positive impact on cerebral cortical recovery.

Effects of High-Frequency Repetitive Transcranial Magnetic Stimulation Combined with Motor Learning on Motor Function and Grip Force of the Upper Limbs and Activities of Daily Living in Patients with a Subacute Stroke

Functional paralysis of the upper extremities occurs in >70% of all patients after having a stroke, and >60% showed decreased hand dexterity. A total of 30 patients with a subacute stroke were randomly allocated to either high-frequency repetitive transcranial magnetic stimulation combined with motor learning (n = 14) or sham repetitive transcranial magnetic stimulation combined with motor learning (n = 16). High-frequency repetitive transcranial magnetic stimulation combined with the motor learning group was conducted for 20 min (10 min of high-frequency repetitive transcranial magnetic stimulation and 10 min of motor learning) three times a week for 4 weeks. The sham repetitive transcranial magnetic stimulation combined with the motor learning group received 12 20-min sessions (10 min of sham repetitive transcranial magnetic stimulation and 10 min of motor learning). This was held three times a week for 4 weeks. Upper-limb function (Fugl-Meyer Assessment of the Upper Limbs) and upper-limb dexterity (box and block tests) concerning upper-limb motor function and grip force (hand grip dynamometer), and activities of daily living (Korean version of the modified Barthel index), were measured pre- and post-intervention. In both groups, there were significant improvements in the upper-limb motor function, grip force, and activities of daily living (p < 0.05). Regarding grip force, the high-frequency repetitive transcranial magnetic stimulation combined with the motor learning group improved significantly compared to the sham repetitive transcranial magnetic stimulation combined with the motor learning group (p < 0.05). However, except for grip force, there were no significant differences in the upper-limb motor function or activities of daily living between the groups. These findings suggest that high-frequency repetitive transcranial magnetic stimulation combined with motor learning is more likely to improve grip force than motor learning alone.

Effects of different frequencies of repetitive transcranial magnetic stimulation on the recovery of upper limb motor dysfunction in patients with subacute cerebral infarction

Studies have confirmed that low-frequency repetitive transcranial magnetic stimulation can decrease the activity of cortical neurons, and high-frequency repetitive transcranial magnetic stimulation can increase the excitability of cortical neurons. However, there are few studies concerning the use of different frequencies of repetitive transcranial magnetic stimulation on the recovery of upper-limb motor function after cerebral infarction. We hypothesized that different frequencies of repetitive transcranial magnetic stimulation in patients with cerebral infarction would produce different effects on the recovery of upper-limb motor function. This study enrolled 127 patients with upper-limb dysfunction during the subacute phase of cerebral infarction. These patients were randomly assigned to three groups. The low-frequency group comprised 42 patients who were treated with 1 Hz repetitive transcranial magnetic stimulation on the contralateral hemisphere primary motor cortex (M1). The high-frequency group comprised 43 patients who were treated with 10 Hz repetitive transcranial magnetic stimulation on ipsilateral M1. Finally, the sham group comprised 42 patients who were treated with 10 Hz of false stimulation on ipsilateral M1. A total of 135 seconds of stimulation was applied in the sham group and high-frequency group. At 2 weeks after treatment, cortical latency of motor-evoked potentials and central motor conduction time were significantly lower compared with before treatment. Moreover, motor function scores were significantly improved. The above indices for the low- and high-frequency groups were significantly different compared with the sham group. However, there was no significant difference between the low- and high-frequency groups. The results show that low- and high-frequency repetitive transcranial magnetic stimulation can similarly improve upper-limb motor function in patients with cerebral infarction.

Comparison of three tools to measure improvements in upper-limb function with poststroke therapy

BACKGROUND

Functional ability is regularly monitored poststroke to assess improvement and the efficacy of clinical trials. The balance between implementation times and sensitivity has led to multidomain tools that aim to assess upper-limb function comprehensively.

OBJECTIVE

This study implemented 3 common multidomain tools to investigate their suitability across a broad spectrum of movement ability after stroke.

METHODS

Forty-nine hemiparetic patients (18 females), aged 22 to 83 years and 24.7 ± 39.2 months poststroke, were assessed before and after a 14-day upper-limb rehabilitation program of Wii-based Movement Therapy. Assessments included the upper-limb motor subscale of the Fugl-Meyer Assessment (F-M), the Wolf Motor Function Test (WMFT), and the Motor Assessment Scale (MAS) upper-limb sections 6 to 8. The MAS was analyzed both with and without the hierarchical system. Patients were stratified with low, moderate, or high motor-function.

RESULTS

Upper-limb function improved significantly for the pooled cohort for all assessments (P < .001), although ceiling effects were evident for the F-M, floor effects for the WMFT, and both floor and ceiling effects for MAS. When analyzed by stratified subgroup these improvements were significant for all groups with the F-M, for the moderate and high motor-function groups with both the WMFT and the MAS scored without hierarchical system, but only for the high motor-function group with the hierarchically scored MAS.

CONCLUSION

These results suggest that no single test is suitable for measuring function and improvement across the spectrum of poststroke upper-limb dysfunction and that assessment tool selection should be based on the level of residual motor-function of individual patients.