S109 Magnetoencephalographic-based brain–machine interface robotic hand for controlling sensorimotor cortical plasticity and phantom limb pain

Abstract

Objectives Phantom limb pain is neuropathic pain after amputation of a limb and partial or complete deafferentation such as brachial plexus root avulsion. The underlying cause of this pain has been attributed to maladaptive plasticity of the sensorimotor cortex. It has been suggested that experimental reorganization would affect pain, especially if it results in functional restoration. We tested the hypothesis that restoration of hand motor function using a brain–machine interface (BMI) based on magnetoencephalographic (MEG) signals will normalize maladapted cortical representation and relieve pain. Methods This study included 10 phantom limb patients (9 brachial plexus root avulsion and 1 amputee). MEG signals during movements of the phantom hand or intact hand were used to train the decoder inferring movements of each hand. The robotic hand was controlled by the decoder. Patients controlled the robotic hand by moving the phantom hand. The training effects were compared among trainings with the phantom decoder, real hand decoder, and random decoder in a randomized cross-over trial. Results BMI training with the phantom decoder increased the decoding accuracy of phantom hand movements and pain. In contrast, BMI training with the intact hand decoder reduced accuracy and pain. Discussion It was suggested that BMI training to modulate the motor representation of phantom hand controlled pain. The sensorimotor cortical plasticity might induce pain. Conclusions and Significance Phantom limb pain was controlled by BMI training to induce sensorimotor cortical plasticity.