Decoding neural activity from an intracortical implant in humans with tetraplegia

This work was part of a pilot clinical study that involved the placement of an intracortical implant in the primary motor cortex of human subjects. The study, the first of its kind, involved the implantation of a 96 electrode array in the human brain for up to a full year. The work described here involved two of the subjects in the study, one of which was a stroke victim, and the other had ALS. Both subjects were tetraplegic and also unable to speak. The intent of this work was to determine if intended/imagined movements would evoke neural activity that could be successfully decoded and allow the subjects to control various devices in their environment and communicate through a personal computer. Neural decoding algorithms were developed to interpret the intracortical activity in the primary motor cortex of both subjects. Adaptive support vector machine methods were developed and employed during this study. A large set of intended/imagined movements was successfully decoded and discriminated between with the first subject. Also, both imagined shoulder and wrist movements were decoded in the second subject. Finally, the algorithms developed in this work allowed, for the first time, control of a motorized wheelchair through intended/imagined movements in a human with an intracortical electrode array implant.