歩行情報を用いた軽度認知障害の予測と認知症予防に関する研究

抄録

In Japan today, the number of patients with dementia is increasing year by year due to the aging society with low birthrate, and it has become a social problem. In the case of mild cognitive impairment (MCI), half of the patients recover normally after treatment. The physical characteristics of patients with MCI include a gait that is characterized by a gait that is unsteady and with a gait that is unsteady. In this study, we derived a model to discriminate cognitive function from waist acceleration/angular velocity information during walking. The need for nursing care is increasing significantly with the aging of the population and the increase in the number of patients with dementia, so prevention of dementia as well as early detection is important. Since it has been reported that a low percentage of REM sleep is associated with decreased brain function and increased risk of dementia, getting adequate REM sleep is an effective means of preventing dementia. Another objective of this study is to derive a model to estimate the quality of REM sleep from daily exercise intensity.

In an MCI discrimination experiment, acceleration/angular velocity sensors were attached to the waist of elderly people, and they were made to walk naturally outdoors. Parameters for estimating cognitive function were extracted from the measured data, and a discriminant model was derived using binomial logistic regression analysis. For REM sleep estimation, wearable terminals were attached to elderly people, and their sleep status and exercise intensity were measured for one month. Using these data, a multinomial logistic regression analysis was conducted to derive an estimation model. As a result, a significant model that could estimate the probability of MCI was obtained. The accuracy of the model was high at 84.7%.

In the REM sleep level estimation, a significant model was obtained to estimate whether the REM sleep level was "adequate" or "deficient" according to the average walking speed. On the other hand, it is not possible to significantly estimate whether the REM sleep level is above or below "adequate," so it is necessary to increase the number of exercise intensity parameters and the number of subjects and reanalyze the results. It is also necessary to consider a new approach to the analysis method, such as multinomial logistic regression analysis with REM sleep level as a nominal variable.