Brain-Computer Interface-controlled Upper Limb Robotic System for Enhancing Daily Activities in Stroke Patients

Our protocol evaluates a BCI-controlled upper limb assistive robot for stroke rehabilitation using EEG and EOG signals to enhance bimanual function. It assesses functional improvements through BeBiTS, advancing assistive neuro rehabilitation.

This technique benefits patients with post-stroke hemiplegia by assisting paretic hand function. It could also assist patients with motor impairment due to spinal cord injuries or neurodegenerative diseases.

This technique bridges motor intention and execution, using EEG EOG, enabling stroke patients to control a robotic hand. It improves bimanual function, enhancing independence in daily life activities compared to conventional assistive approaches.

Patients may face difficulties if they are experiencing motor imagery training for the first time. Therefore, appropriate guidance and instructions are needed to ensure kinesthetic motor imagery.

DCI-controlled upper limb robot encompasses neuro rehabilitation components such as neuroplasticity and motor learning. This system can be extended to motor rehabilitation in the patients with stroke as well as cerebral palsy and neurodegenerative diseases.

[Instructor] To begin, provide all recruited patients with detailed information about the experimental procedure. Obtain signed informed consent from each participant. After completing the consent form, evaluate the 10 items of the BeBiTS assessment before training the BCI robot in a comfortable chair in front of a desk. Launch the brain-computer interface or BCI system. Place the cap on the patient's head and connect the amplifier. In the source module, select EegoModule, followed by impedance mode, and press start to activate the module. Observe the blue light indicating activation. Ensure the impedances are below 10 kilo ohms. Then press stop in the source module. Change the mode to EEG for data streaming. Press start, and check the signal quality. For EOG calibration, in the task module, set the number of cues. Instruct the participant to perform brief lateral eye movements, following the 10 arrows appearing on the screen. Check the result graph immediately after the training. For EEG calibration, select the EEG Calibration Task Module, and set the number of cues in the task module to five. In the feedback module, set laterality to the side of the robotic hand. Ensure display pacman is unselected. Now instruct the participant to imagine clenching their fist when the prompt, imagine making a fist, appears on the black screen, and then review the result graph. After the EOG and EEG training, set parameters for the specific target frequency of interest, reference value, and threshold, that distinguish the intention of making a fist. Using the configured parameters, proceed with feedback training using the pacman interface. With a USB dongle, connect an assistive robotic hand wirelessly to a computer. Next, have the participant wear the robot and perform the BeBiTS assessment. Wait for the white light on the screen indicating the ready state. Upon confirmation, instruct the participant to move their eyes to one side to change the light to green. When the green light appears, instruct them to imagine clenching their fist. Using the robot, assist the participant in clenching their fist and performing the task. After completing the task, instruct the participant to observe the red light on the screen. If the participant wants to open their hand, they can move their eyes to change the light color back to white. Finally, the patient is evaluated again, post-BeBiTS, using the BCI robot system. EOG values of a well-trained participant showed consistent trials, where the mean curve reached the threshold level, and their EEG results clearly differentiated between resting state and motor imagery. In contrast, EOG trials of the poorly trained participant were inconsistent, with the mean curve failing to meet the threshold level, and their EEG results lacked clear distinction between the resting state and motor imagery. Participants P1, P4, and P5 failed to perform most tasks during both pre and post-BeBiTS assessments. Participant P3 initially scored in the pre-BeBiTS assessment, but showed no scoring in the post-BeBiTS evaluation after inadequate training. Participants P2, and P6 to P8 showed improvement in some tasks during the post-BeBiTS assessments compared to the pre-BeBiTS evaluations.