A Hybrid BCI and VR System to Promote Cortical Reorganization

What neurofeedback is and how it can be used for motor rehabilitation is the topic of our last paper, published by IEEE Transactions on Neural Systems and Rehabilitation Engineering. In this paper we look at the motor rehabilitation post-stroke issue not from the motor deficit point of view (consequence), but from the brain perspective (origin). Motor rehabilitation post-stroke is all about finding effective ways to stimulate the remaning healthy brain areas to "pick up" the functions lost by the damaged neural circuitry. Yet, we keep on developing novel motor training paradigms with the hope that they will drive in a more effective manner this brain process. Brain Computer Interface (BCI) technologies have been around already for some time, mostly targeting those that have no movement capacity left to interact with the world. Thus, BCI serves as the translator of "thoughts" into "commands" to the real world. However, the uses for BCI are numerous, and a case in point is neurofeedback. In a very simple manner, neurofeedback is the process of using neural activity as a feedback mechanism. This has been shown to be a successful technique in different areas such as the case of some attention disorders. In this work we explored the use of neurofeedback for post-stroke motor rehabilitation.

Our goal is to re-train the brain to be able recover as much as possible from stroke sequels. In this case, re-training means that neural circuits in our brain that were doing some task before now learn some of those lost capabilities. But, how can we make sure that we are giving the right stimuli to the brain? BCI enables us to measure brain activity in real-time and therefore assess what motor or mental activity engages our neural circuits most effectively. In our experiments, we have found that a technique known as "mental motor imagery" - the mental simulation of motor actions - is more efective at engaging brain areas than motor actions alone, suggesting that using metal simulations of motor actions is a valid motor training. Further, the brain responds "best" when mental imagery is combined with motor actions.

Brain activity maps for different motor training conditions.

Following up on this idea, we designed a neurofeedback paradigm for motor training. Our system implemented a mental imagery motor training task connecting a BCI to a computer game, in which the subject needs to imagine him/herself performing arm movements to control a virtual agent. Right now we are developing these concepts further in the RehabNet project, where among other objectives we aim at developing a low cost and personalized neurofeedback training for upper limb rehabilitation (http://neurorehabilitation.m-iti.org/rehabnet/). 

Virtual agent controlled by means of BCI.

Reference:

Bermudez i Badia, .S., Garcia, M.A., Samaha, H. and Verschure, PFMJ. Using a Hybrid Brain Computer Interface and Virtual Reality System to Monitor and Promote Cortical Reorganization through Motor Activity and Motor Imagery Training. IEEE transactions on neural systems and rehabilitation engineering: a publication of the IEEE Engineering in Medicine and Biology Society. In press.