Dementia Sign Detection System Using Digital Twin

Abstract

Currently, the most common way to detect signs of dementia is an interview that mainly focuses on whether there is “cognitive function disorder.” However, detecting its signs more accurately, it is necessary to evaluate changes of behaviors related to “life function disorder.” Conventionally, an interview has also been the mainstream to determine whether there is “life function disorder.” It is conducted with an elderly person and family members living with the elderly. Therefore, in this study, we propose a system that detects signs of dementia so that we do not have to rely on interviews. The system is achieved by digitally transforming a subject's behaviors and configuring the subject's digital twin. We installed a communication robot and ambient sensors at each of their houses and, in addition, utilizing a wearable device. This enabled us to digitally transform their behaviors inside and outside houses and configure their digital twins. Using digital twins, we performed detection of signs of dementia targeting both “cognitive function disorder” and “life function disorder.” We evaluated the system in experiments conducted at Nagasaki University Hospital and an ordinary person's home in Nagasaki City. As the result, we confirmed that we were able to configure subjects' digital twins for detection of signs of dementia by comparing their actual behavioral histories.