Paper Reading #3: Reducing Compensatory Motions in Video Games for Stroke Rehabilitation

Introduction

In their paper "Reducing Compensatory Motions in Video Games for Stroke Rehabilitation" Gazihan Alankus and Caitlin Kelleher explored and developed a way to help people who have had strokes rehabilitate by using a video game that could be played at home. It would be used specifically to help reduce compensation for lack of motion in the torso. This paper was presented at CHI 2012 in May 2012 in Austin, TX. Gazihan Alankus is from both Washington University in St. Louis and Izmir University of Economics in Turkey. His interests reside with video games for rehabilitiation, virtual characters, and sensor networks. Caitlin Kelleher has varied interests within HCI. Currently she is working on a project that allows children to learn computer programming.

Summary

Alankus and Kelleher conducted their research in two steps. They first designed a way to detect torso compensation and a way to evaluate such compensation. Then they designed a game which can be used as a theraputic tool to help patients get feedback on their personal conscious or unconscious compensation for lack of motion in the torso.

For the first step, they designed a harness using Nintendo Wii remotes to help track a participant's torso movements. Then they analyzed data from the movements of the participants to evaluate how compensation data was collected. They took into account the location of the Wii remotes, the movements that users make when doing shoulder abduction and adduction exercises, and the range of error in comparison to a motion capture system.

Figure 1: The harness created for this study

In the second step, the authors designed a game that would have users perform therapeutic exercises and would create situations in which users would have to naturally compensate. In this side-scrolling type game, the user has to control a floating hot air balloon. They have move their bodies to control the vertical position of the balloon to catch parachute jumpers. This causes the user to do shoulder abduction and adduction exercises. They calibrated the game for each user by including an seamless in-game method, making movements much more matched between calibration and actual performance. They placed the jumpers at both motion range extremes to create therapeutically beneficial motions and used obstacles to promote controlled motions. They used a point system and visual and audio ques to shape rewards and punishments for the users to help condition them to use less compensatory movements. They modified, improved, and tested the game in iterative steps.

Related Work Not Referenced in the Paper

The authors appropriately talked about related work. They divided there related works into three sections, each addressing a different aspect of their work. Their work is not very novel. Using technology (and games) to aid in rehabilitation is not new. However, it helps supplement the work that has already been done. Several other related works are listed below.

1. McNeill, M.D.J., D.K. Charles, et al. "Evaluating user experiences in rehabilitation games." Journal of Assistive Technologies. 6.3 (2012): 173-181. Web. 5 Sep. 2012.

McNeill, et al. analyzed the game-based systems that aided in rehabilitation and discussed how current work could be improved. Alankus and Kelleher could probably look at their work in order to improve their game and harness.

2.

Evaluation

In order to evaluate the possible success of the game, the authors created five versions of the game:

1. The balloon does not tilt, so there is no compensation feedback
2. The balloon tilts only
3. Punishments and rewards are provided depending on the amount of compensation
4. Uses punishment and reward events with an adaptive algorithm
5. Uses punishment and reward events with variable ratio feedback

The authors used a subjective qualitative questionnaire and interview to see how the users felt about the game. They used the compensation data from the Wii remotes and the game to quantitatively analyze actual success in modifying compensation behavior. There was not a significant difference between the first two versions of the game nor between the second and third versions of the game. The third version was able to reduce relative compensation. However, when punishments and rewards were given, then there was a notable reduction in compensation. They also found that players were willing to play the game and seemed to enjoy it.

Discussion

I think this is a very interresting thing that the authors looked at. I have read about how video games can help people learn, but rehabilitation is another perspective that I did not think about. I think this work is novel and significant. At the same time, since this is novel, I think more time and research must be done to refine how the game can be used to effectively help patients recover compared to once-a-week sessions in physical therapy. I think the evaluation was appropriate, but it, too, must be refined to further minimize error. A possible way to do this is to use some sort of sensor or device that is less bulky than the Wii remote. I hope that these authors continue this work and that others follow in their footsteps to extend to other types of rehabilitation methods.

Reference

Alankus, Gazihan, and Caitlin Kelleher. "Reducing Compensatory Motions in Video Games for Stroke Rehabilitation." CHI '12 Proceedings of the 2012 ACM annual conference on Human Factors in Computing Systems. (2012): 2049-2058. Web. 4 Sep. 2012.