A virtual training system based on human information processing model for improving mild cognitive impairment (MCI)

Abstract: In recent years, the incidence of age-related diseases has increased dramatically due to the increase in the average human lifespan. The most common age-related symptom is Mild Cognitive Impairment (MCI), which often represents the transition from healthy aging to dementia, a neurodegenerative disease that has hitherto been incurable. When people detect suspected cognitive decline, how should they respond in advance? Past studies have reported that short-term memory loss is the most predictive of dementia in neuropsychological tests, so strengthening memory training and encouraging consistent self-practice can slow the rate of cognitive deterioration into dementia. Maintaining memory is nothing more than exercising, learning new skills, cultivating interests, and participating in educational activities. However, people with cognitive degeneration often feel indifferent to activities and miss the critical period for cognitive recovery. With the advancement of technology and the upsurge of games driven by virtual reality (VR), it is hoped that this group of people will be willing to receive rehabilitation training under the appropriate situational design. There were some relevant research reports on applying VR game training to restore cognitive function degradation in the past. Still, the development and training of the overall game were rarely carried out in the cognitive model. In addition, the training effectiveness was mainly re-tested with a test questionnaire. Nevertheless, each detection tool has a "reasonable suspicion interval for cognitive deterioration" that will increase the precise assessment of training effectiveness. Therefore, this study aims to design a VR game training system based on the cognitive model. The significant differences between this training model and the game training developed in previous studies are:(1) The training system is built based on the four aspects of perception, cognition, action, and heuristics, not only emphasizing memory training or action response training.(2) The training required for the complete establishment of human information processing: Unlike other game training, it may only be a fragment of training and lacks the training of human message processing sequentially and coherently, thus limiting the effectiveness of the training.The study also applied the Receiver Operating Characteristic (ROC) evaluation method to evaluate the effect of training on cognitive function enhancement more accurately. The results show that if the traditional detection tool Mini-Mental State Examination (MMSE) is used for evaluation, the sensitivity is low, and the training effect cannot be specifically illustrated. However, when the ROC evaluation method is used, the indicator: AUC (Area Under Curve) analysis and Accuracy analysis have high sensitivity for training effect evaluation. In terms of the effectiveness of VR game training: if the participant's cognitive function is suspected to be degraded (but not dementia), the response effect is significantly improved after training, which has a particular effect on delaying dementia.

Keywords: Virtual training system, Mild cognitive impairment, Cognitive degeneration, Receiver operating characteristic curve

DOI: 10.54941/ahfe1003477

Cite this paper:

​