Neurofeedback reveals mechanism of phantom limb pain to reduce pain

Phantom limb pain is an intractable pain for which no effective treatment has been establish ed. The pain has been attributed to abnormal plastic changes in the sensory motor cortex corresponding to the deafferented body part. Some feedback therapy such as mirror therapy have been applied to modify the abnormal cortical changes, although it is not been unveiled how to change the corresponding sensory motor cortex to reduce pain. We have applied neural decoding to magnetoencephalography (MEG) to extract motor information of the upper limb, and realized a Brain–Computer Interface (BCI) that allows patients to operate a prosthetic hand as if they were moving a phantom limb. In addition, we have demonstrated that neurofeedback (NF) training to control the BCI induced plastic changes in the patient’s sensorimotor cortex and changes in the pain. Actually, the training to attenuate the motor representation of the phantom limb reduced the pain. In addition, we evaluated the efficacy of the NF training by a blinded crossover trial of training with three consecutive days. Twelve patients were trained to control the phantom limb images, that were controlled through BCI. After three days NF trainings, the pain assessed with the Visual Analogue Scale (VAS) was significantly reduced for five days. Furthermore, the pain reduction was associated with the attenuation of the motor representation of phantom limb. These results suggest that the residual motor representations of phantom limb cause the phantom limb pain. We have demonstrated that the NF training elucidates the pathogenesis of chronic pain and develops a new treatment.