Machine learning models for predicting cognitive decline based on long-term longitudinal data

基于长期纵向数据预测认知能力下降的机器学习模型

• 批准号: 22K10074
• 负责人: 高橋 利士
• 金额: $ 2.66万
• 依托单位: Osaka University
• 依托单位国家: 日本
• 项目类别: Grant-in-Aid for Scientific Research (C)
• 财政年份: 2022
• 资助国家: 日本
• 起止时间: 2022-04-01 至 2025-03-31
• 项目状态: 未结题
• 来源: https://kaken.nii.ac.jp/grant/KAKENHI-PROJECT-22K10074/
• 关键词: 高齢者; 認知症; 機械学習

本研究は，身体的因子，社会的因子，心理的因子のデータに，口腔因子（残存歯数や歯周病の状態など）や口腔機能（咬合力や唾液分泌機能など）のデータを加えた包括的なデータから，認知機能低下を機械学習により予測するモデルを作成することを目的とした．計画初年度の本年度は，認知機能低下を予測するモデルを作成するにあたり，認知機能低下に関わる因子の間にある因果関係を探索することを目的として，LinGAM（Linear Non-Gauussian Acyclic Model）モデルを作成した．本研究の対象者は，2012年および2013年に健康長寿研究のベースライン調査に参加した70歳と80歳の地域在住高齢者の中で，6年後の追跡調査時にも参加した920名（70歳:548名（男性271名，女性277名），80歳:372名（男性184名，女性188名））とした．本研究の対象者の中で，6年後に認知機能が低下したものは155名(25.5%)であった．70歳男性群の場合，口腔内因子の中では残存歯数がGOHAI，GDS5を通じて認知機能の低下に間接的な因果関係がある結果となった．70歳女性群の場合，口腔内因子の中では，残存歯数がWHO5を通じて認知機能低下に間接的な因果関係があり，歯周病の状態が直接的に因果関係がある結果となった．80歳男性群の場合，口腔内因子の中では，認知機能低下に直接的，あるいは間接的な因果関係がある因子はなかった．80歳女性群の場合，口腔内因子の中では，残存歯数が認知機能低下に間接的な因果関係がある結果となった．

In this study, physical factors, social factors, psychological factors, oral factors (residual weekly disease symptoms) oral mechanical capacity (bite force saliva secretion machine function) to increase the number of people who are involved in this study, the mechanical science of the machine is known that the mechanical science of the machine will lead to the goal of the disease at the beginning of the year. Know that the machine is able to reduce the health and longevity of the people who study the health and longevity of the population in 2012, the health and longevity study of 2013, the participants of the health and longevity study in 2012, the participants of the health and longevity study in 2012, the participants of the health and longevity study in 2012, the participants of the health and longevity study in 2013, the participants of the health and longevity study in 2012, the participants of the health and longevity study in 2013, the participants of the health and longevity study in 2012, the participants of the health and longevity study in 2012, the participants of the health and longevity study in 2012, the participants of the health and longevity study in 2012, the participants of the health and longevity study in 2012, the participants of the health and longevity study in 2012, the study of health and longevity in 2012, the study of health and longevity in 2013, the study of health and longevity in Six years later, 920 participants (70: 548 (male 271, female 277), 80: 372 (male 184, female 188) were enrolled in the study. The subjects in this study were in the middle of the study. After 6 years, there were 155 (25.5%) men who knew that the machine was low. 70 men were well matched, and there was a residual GOHAI in the oral factor. The results of the causality test of the low-risk cause-and-effect response of the GDS5 were similar to that of the 70-year-old female group, the oral factor, the residual factor, the direct cause-and-effect response of the disease, and the oral factor, respectively. 80-year-old men were in the same group. It is known that the cause and effect of the machine is low, the causal factor is affected. 80% of the female population is in the same group, the factor in the oral cavity is affected, and the residual number is known to be low.