Research highlight

Neural prosthetic cursor control achieves highest human performance to date

Nature Medicine

September 29, 2015

A neural prosthetic system that decodes brain-derived neuronal signals to enable people with tetraplegia to control the movement of a computer cursor is described in a paper published online this week in Nature Medicine. The neural cursor control demonstrated in two people with amyotrophic lateral sclerosis (ALS) represents the highest performance achieved in humans to date.

For people who have become paralyzed as a result of stroke, spinal cord injury or neurodegenerative disease, the ability to re-engage healthy brain circuits that control voluntary movement has the potential to improve quality of life. Previous proof-of-concept clinical trials have demonstrated that neural activity can be decoded to enable control of computer cursors and robotic limbs. However, better neural prosthesis performance is required before an attempt can be made to transition these technologies into widespread clinical use.

Jaimie Henderson and colleagues report the clinical translation of a neural prosthetic system, developed in nonhuman primate studies, for use by two people with ALS as part of a multi-site pilot clinical trial.

These two patients had previously undergone surgery to implant microelectrode arrays into their motor cortices to record neural activity corresponding to imagined finger movements. The array was coupled to a neural prosthetic system with an improved decoding algorithm to translate neural activity into two-dimensional movement commands. The patients showed faster and more precise neural cursor control than that obtained in a prior study. Furthermore, an experiment that alternated, without the participants' knowledge, between a previous version of the decoding algorithm and the improved version developed here revealed that the new algorithm results in both objective and subjective improvements in task performance.

Further work is required to determine how other factors, including the different causes of motor impairment or other variation between individuals, may affect the performance of this system.

doi: 10.1038/nm.3953

Return to research highlights

PrivacyMark System