Exploring VR as Therapy

 

 

 

Exploring the use of VR as a therapy tool for children with motor impairments

This project is part of an ongoing study to investigate the use of virtual reality technology as a method of providing occupational therapy. Specifically, the focus of this project has been to target children with motor impairments such as cerebral palsy. People afflicted with such disorders typically have one limb which is significantly less functional than the other. While the child’s natural inclination is to use their dominant arm to complete therapeutic tasks, the goal of therapy is to force them to utilize the less functional limb to increase mobility. In traditional therapy, the child’s dominant arm is restrained, often leading to frustration and unwillingness to participate.

Virtual reality technology provides a means to immerse the child in a world that will entice them to interact by using their less functional limb. Specifically, a cognitively interesting VR-based game could offer an entertainment incentive that traditional therapy lacks. We have designed a VR game intended for children between the ages of five and twelve, which tracks only one arm. Thus, the child is not restrained, but can only interact with the world with the arm that the therapist chooses.

Our virtual reality set-up includes a stereoscopic head mounted display and a wired Ascension magnetic tracker placed on the subject’s non-dominant hand. For display, we used the VRStation software made available from Sandia National Laboratories, running on an SGI Onyx Graphics Server. The simulation that controls the game is driven from a Windows-based machine using VRSim, a virtual reality simulation engine developed at Ithaca College.

For this experiment, the subjects were children between five and twelve with no motor impairments. The virtual game placed the subject in a small room, with a series of five different colored virtual blocks placed at arm’s length directly in front of them. A block would spin, signaling the child to touch it. Upon completion of this task, the block would turn into an animated toy, and another block would begin to spin. The entire game consisted of six cycles of five blocks each. We also designed a real world equivalent game which placed the five blocks at their exact locations in the virtual world, affixed to board.

The experiment goals included an evaluation of the level of cognitive interest (fun factor) of the VR game, an evaluation of the effect of playing the game on proprioception and cyber-sickness, and a comparison of arm movement between the virtual environment and its real world equivalent. The subject took a proprioception pointing test at the beginning of the experiment and at the end, after being immersed in the virtual environment. After the first proprioception test, the subject completed the virtual block game and the real world block game, the order of which was determined on an alternating basis. The subject filled out questionnaires immediately prior to and upon completion of the virtual game.

 Software was developed to allow us to visualize the raw positional data obtained from the hand sensor as a line graph in 3D space. It was intended to allow therapists to view and analyze the recorded data from the experiment, by allowing the placement of waypoints that defined the start (hand in lap) and end (hand touching block) of each reach. From these waypoints, accuracy scores are calculated for each reach (based on the ratio of distance between the two points and the actual length of the hand’s path.) The software also allows one to see the different types of reaching motions of each child and the differences between reaching for the real blocks versus the virtual blocks, which is quite significant.

Statistical analysis of the proprioception data across all subjects showed no statistical significant evidence of the virtual reality hardware’s effect on proprioception. Statistics also showed no effect on nausea, ability to see, or ability to walk. Children in the age range of five to eight found the game the most fun, while older children aged ten through twelve seemed less interested. The average accuracy score of the real block pointing test was 93%, while the score for the virtual block test were 79%. Statistically, there was no variation in the accuracy scores of the individual blocks, with the exception of a slight increase of the center block during virtual tests, which is a result we could not account for. Examination of cycle times indicates that the children exhibited learning of the task. Between the second and fifth cycles, the virtual test an average improvement of 2.1 seconds was observed; likewise, there was an average improvement of 1.2 seconds in the real tests. Only 5.5% of the children took longer on the fifth real test, and only 4.4% took longer on the fifth virtual test.