Learning From Broken Brains:

Robotic Stroke Therapy and Human Movement Control


Neville Hogan

Department of Mechanical Engineering

Department of Brain and Cognitive Sciences

Massachusetts Institute of Technology



Neurologic injury is a major and growing source of health-care cost in the US and worldwide, affording a new opportunity for robotics: improved treatment and reduced cost may be achieved using appropriately human-compatible robots. Results to date are promising: robotic treatment of stroke has a lasting benefit and yields twice the impairment reduction of human-administered therapy.


Using robots to enhance the performance of brain-injured humans affords a unique opportunity to study how human brains learn and control movement. This presentation will review striking evidence that although neural computation is massively parallel and continuous, human movement control is fundamentally serial and intermittent. Furthermore, despite their grace and coordination, movement speed is surprisingly difficult for humans to control. Some implications for brain-machine interfaces and human




Friday, January 25, 2002

3:30 - 5:00 p.m.

1500 EECS