Presentation information

General Session

General Session » GS-10 AI application

[3F2-GS-10j] AI応用:QoL

Thu. Jun 10, 2021 11:00 AM - 12:40 PM Room F (GS room 1)

座長:水本 智也(フューチャー(株))

12:20 PM - 12:40 PM

[3F2-GS-10j-05] Prediction of functional independence measure (FIM) using machine learning and wearable data of stroke inpatients in a convalescent rehabilitation ward

〇Takayuki Ogasawara1, Kentaro Tanaka1, Masahiko Mukaino2, Yohei Otaka2, Masumi Yamaguchi1, Eiichi Saitoh2, Shingo Tsukada1 (1. NTT Corporation, 2. Fujita Health University)

Keywords:rehabilitation, stroke, wearable, Functional independence measure, convalescent period

This study aimed to estimate the score of motor FIM (Functional independence measure) of stroke inpatients, which is used as a clinical indicator in rehabilitation medicine, using wearable data recorded in a convalescent rehabilitation ward. We recorded the electrocardiogram and acceleration data of 192 stroke inpatients over a day. To estimate the score of motor FIM, we trained neural network using the wearable data and basic information of inpatients such as weight, height, sexuality and age, and then, performed five-fold cross validation. In the result, the coefficient of determination between estimated FIM by neural network and ground truth scored by therapists showed 0.73 and it was significant (p < 0.001), suggesting the possibility of estimation of motor function of stroke inpatients using activity record obtained with wearable devices.

Authentication for paper PDF access

A password is required to view paper PDFs. If you are a registered participant, please log on the site from Participant Log In.
You could view the PDF with entering the PDF viewing password bellow.