Putting things at work: the design of the Rehabilitation Gaming System

One can have a clear idea of what the components for an ideal motor rehabilitation system could be, but... what is the recipe to put all of these ingredients together? In the case of motor deficits after a brain lesion, we know we must find a way to stimulate our brain so it can relearn and rewire its remaining neural circuits in order to achieve the desired motor function by means of alternative neural pathways. We do know that motor training is a necessary component for brain rewiring and motor recovery. We do know that motivational and engagement aspects can be crucial for positive treatment outcomes. However, it is not yet known what the best way to promote brain rewiring is, or what the optimal motor training is, or even how to design a “universally” engaging training. The easy one is that we know that praising and constructive feedback are essential for learning, that is, positive reinforcement should be one of the fundamental pilars for a successful motor training.

Given that from the beggining we know that even the most basic aspects are not yet understood... How can we go about building a system that puts so many unknowns together and still makes sense? Our approach was the following: “keep things simple”. It is obvious no one can come up with the ideal solution for such a complex system if we do not even know how to solve its parts. Thus, we decided to reduce the complexity of our system to the bare ingredients that we could understand, despite knowing that there is huge room for exploration and improvement. It was not reasonable to assume that putting together the fanciest aspects of gaming, current opinion on brain mechanism of recovery and the ultimate physical therapy can provide us with clear insights on what are the key elements for motor recovery are. In such a case, it would be impossible to understand the how the brain would respond to each of the fancy aspects of our game, distinguish its positive and negative contributions, and understand the interplay of the many things that would be at play at once. Therefore, we based the development of the Rehabilitation Gaming System (RGS) on few but well established principles and testable hypotheses:

1.  Our action recognition system (Mirror Neuron System) responds both to actions performed by others and actions performed by us. Hence, the RGS uses virtual reality to map simple physical arm movements into meaningful, task oriented manipulation of virtual objects by means of virtual arms. This virtual reality environment presents the brain with a more complex motor repertoire than reality can, enabling patients with low mobility to accomplish complex tasks with their virtual arms that couldn’t otherwise. The premise here is that this computer generated visual input can activate the action recognition system to a larger extend.

2.     It is known that moderate levels of stress can improve performance as compared to high and low stress situations. For this reason, we studied the effect of game parameters with both healthy and stroke patients. This allowed us to assess and formulate mathematically the effect of changing any parameter of our game in patient performance. This allowed us to adaptively change all parameters during training to set the patient always to a performance level around 70%. Ensuring a 70% performance avoids frustration and allows always a 30% of improvement to be made by the patient, independent of his/her level of difficulty achieved in the training game. This results in a training that continuously adapts it difficulty to challenge patients at their right level.

3.  Positive reinforcement and augmented feedback on performance is important for learning. Virtual arms displaying successful actions, an adaptive difficulty system that avoids a too difficult game, a point system that allows patients to self monitor their improvements, and the use of rewarding sounds and animations on success is used to inform and praise patients.

Based on these premises, we developed the RGS, a computer mediated gaming system for motor rehabilitation that makes use of state of the art interactive technologies – data gloves and a camera based tracking system for arm and hand movement tracking - and virtual reality and gaming aspects for augmented visual and performance feedback – Torque Game Engine – to deliver a personalized training. The game is quite simple. It consists basically of a pair of virtual arms that copy the movement of the patient’s physical arms, and have to be controlled to grasp objects that approach towards the patient. The speed of the objects, position and time interval between them defines the difficulty of the training task. We decided to go start with a task that mostly trains grasping and range of movement since those are very important components for the performance of activities of daily living. We deployed this system in 2 randomized controlled longitudinal studies (with patients in both the acute and chronic phase) over the last 3 years with positive results as compared to controls. In addition, we are using functional imaging techniques (fMRI) to evauate the effect of training with the RGS on brain reorganization. These results will be coming out soon.

My ultimate objective is to be able to reduce the technological complexity and cost of systems such as the RGS to be able to provide a home based alternative or an extended care for patients once discharged from the rehabilitation facilities. These systems can sustain and improve their recovery – and thus quality of life – and reduce transportation costs for treatment.

At the moment, my collegues in Barcelona are distributing a free version of the RGS for evaluation at http://rgs-project.upf.edu/. If you want to check it, feel free to go for it. In the near future I will be posting about newer systems I'm working on that I will make freeley available to gather feedback. So, keep on cheking my posts!

References:

Cameirao, M.S., Bermúdez i Badia, S., Duarte Oller, E., and Verschure, P.F.M.J. Neurorehabilitation using the Virtual Reality based Rehabilitation Gaming System: Methodology, design, psychometrics, usability and validation. J Neuroeng Rehabil. 2010; 7: 48. (http://www.ncbi.nlm.nih.gov/pubmed/20860808)

Cameirão, M.S., Bermúdez i Badia, S., Duarte Oller, E., Verschure, P.F.M.J.  The Rehabilitation Gaming System: a Review. Studies in health technology and informatics. 145, 65-83. (http://www.ncbi.nlm.nih.gov/pubmed/19592787)