An intelligent rehabilitation system for cognitive rehabilitation

Abstract

Prospective memory (PM) failure is one of the most disabling forms of cognitive impairment in people with acquired brain injury (ABI). To date, there is a paucity of research on effective treatment strategies for PM deficits. In the present study, an innovative PM treatment combining modern computer technologies with artificial intelligence (AI) and virtual reality (VR) was proposed. Combining the simulated training environment provided by VR and the individualized feedback provided by AI, an optimal training environment could thus be created for people with ABI. This "intelligent virtual reality prospective memory" (IVRPM) training programme could be a cost-effective means to offer PM training content which is considered ecologically valid for the living environment of people with ABI. In order to investigate the effectiveness of this novel IVRPM training programme, another non-AI version, a "virtual reality prospective memory" (VRPM) training programme was developed for comparison purposes. A single-blinded, pre-test/post-test randomized controlled trial was adopted. Fifty-six subjects were successfully recruited. They were randomly assigned to an IVRPM treatment group, a VRPM treatment group, or a control group. Subjects who were assigned to the two treatment groups were provided with ten sessions of training in the respective programmes. Subjects' PM performance, retrospective memory performance, frontal lobe functioning, level of community integration, and self efficacy were assessed before and after the training. The results showed that subjects in the IVRPM group performed significantly better in PM functioning, level of community integration and self-efficacy. It may suggest that the subjects in the IVRPM group not only learnt better, but also showed better transference of skills to the real environment. Finally, the findings showed that the AI system could help to establish subjects' self efficacy and hence it improved their perception of their level of community integration as compared with the control group. To conclude, the findings from the present study showed that the combination of AI and VR can be a feasible and successful treatment strategy for PM rehabilitation of people with ABI. Further application of the present training model to other groups of patients should be encouraged.