Brainmachine interfaces (BMIs) hold a lot of promise for restoring some level of motor function to patients with neuronal disease or injury. Current BMI approaches fall into two broad categoriesthose that decode discrete properties of limb movement (such as movement direction and movement intent) and those that decode continuous variables (such as position and velocity). However, to enable the prosthetic devices to be useful for common everyday tasks, precise control of the forces applied by the end-point of the prosthesis (e.g., the hand) is also essential. Here, we used linear regression and Kalman filter methods to show that neural activity recorded from the motor cortex of the monkey during movements in a force field can be used to decode the end-point forces applied by the subject successfully and with high fidelity. Furthermore, the models exhibit some generalization to novel task conditions. We also demonstrate how the simultaneous prediction of kinematics and kinetics can be easily achieved using the same framework, without any degradation in decoding quality. Our results represent a useful extension of the current BMI technology, making dynamic control of a prosthetic device a distinct possibility in the near future.
|Original language||English (US)|
|Number of pages||9|
|Journal||IEEE Transactions on Neural Systems and Rehabilitation Engineering|
|State||Published - Jun 2009|
Bibliographical noteFunding Information:
Manuscript received June 17, 2008; revised January 14, 2009; accepted February 10, 2009. First published June 02, 2009; current version published July 06, 2009. This work was supported in part by a Merit Review grant from the Department of Veterans Affairs, in part by the National Institutes of Health under Grant NS42278, and in part by a Doctoral Dissertation Fellowship from the Graduate School, University of Minnesota.
- Dynamic control
- Force fields
- Motor cortex
- Neural prostheses