Motor Imagery and Motor Execution Based BCI in Stroke (BCI-MIME)

Effect of Motor Imagery and Motor Execution-based Brain Computer Interface on Motor Rehabilitation in Subacute Ischemic Stroke

About 50% of stroke patients are unable to live independently because of residual disability. Brain-computer interface (BCI) is based on closed-loop theory, which facilitates neurological remodeling by establishing a bridge between central and peripheral connections. Studies have confirmed that BCI real-time neurofeedback training system based on motor imagery alone can effectively improve patients' motor function. So, is the benefit greater if motor imagery is combined with motor execution? Current conclusions are mixed. In addition, previous studies and our preliminary study found that prefrontal Fp1 and Fp2 areas play an important role in motor recovery after stroke, and they are involved in motor imagery, motor execution, attention and other behavioral processes. Therefore, we designed a BCI training system based on motor imagery and motor execution with prefrontal electroencephalogram (EEG) signals as the modulatory target. This was a randomized placebo-controlled double-blinded clinical trial. Patients in the test group performed BCI-controlled upper extremity motor imagery + upper extremity pedaling training. The control group had the same equipment and training scenario, and patients were also asked to imagine the upper extremity pedaling movement with effort, and patients also wore EEG caps, but the EEG signals were only recorded without controlling the pedaling equipment. After 3 weeks of treatment, we observed the changes of motor and cognitive functions as well as fNIRS-related brain network characteristics in both groups.

The experimental group was trained with BCI-controlled pedaling rehabilitation training system. Patients wore EEG caps and were instructed to imagine upper limb pedaling movements. The greater the patients' movement intention, the higher the Mscore of movement intention index on the monitor and the faster the pedaling speed. In addition, the movements on the monitor are synchronized with the actual movements, and the system provides audio and text feedback according to the patient's performance.

In the control group, the training equipment and scenario were the same as in the experimental group, and the patients wore EEG caps and were also instructed to imagine upper limb pedaling movements. However, the system was changed to only record the EEG signal without controlling the pedaling equipment, and the Mscore score and pedaling speed displayed by the equipment in real time were pre-set data of the training performance of the previous pretest patients, i.e., sham neurofeedback.

We designed a motor imagery and motor execution-based BCI-controlled pedaling rehabilitation training system. Patients wore EEG caps and were instructed to imagine upper limb pedaling movements. The greater the patients' movement intention, the higher the Mscore of movement intention index on the monitor and the faster the pedaling speed

The training equipment was the same, but the program was changed to record only the EEG signal without controlling the pedaling equipment, and the Mscore score and pedaling speed displayed by the equipment in real time were pre-set training performance data of the previous pretest patients, i.e., sham neurofeedback.

The Berg Balance Scale contains 14 items. The score range is 0-56 points, the higher the score, the better the balance function.

The modified Barthel Index is for assessing activities of daily living. The score range is 0-100 points, the higher the score, the better the function.

P300 is an auditory evoked event related potential, it can be used to assess neural activity associated with cognitive processes. The latency is about 250ms-800ms.

Inclusion Criteria: Aged 35-79 years old; Patients with first ischemic stroke onset from 2 week to 3 months; Hemiplegia with upper limb strength grades 1-3; Consciousness, sitting balance level 1 or above, can cooperate with assessment and treatment; The patient or its authorized agent signs the informed consent form. Exclusion Criteria: Severely impaired cognition (MMSE<20), unable to pay attention to and understand screen information; Severe pain, spasticity and limited mobility of upper extremity.