Humanoid robot teleoperation reminiscence group therapy for older adults with dementia: A controlled trial study
XiangYu Liu MA*, Ryuji Yamazaki PhD, Jie Yang MA, Hiroko Kase PhD
Full text PDF 
( Download count: 246)
AbstractBackground: According to the Ministry of Health, Labor, and Welfare of Japan, there will be a shortage of 270,000 nursing staff in Japan by 2025. The Japanese government has also proposed using intelligent technology to alleviate this problem. Introducing robots into nursing facilities is an essential solution to the labor shortage.
Objective: This study attempts to find the differences between humanoid robots in replacing caregivers through linguistic and emotional comparisons with a controlled trial. The actual use of the humanoid robot Pepper for dementia in older adult care facilities was conducted. In addition to the common scales and linguistic analysis, an interdisciplinary approach was used to increase the dimension of facial expression recognition to better represent the differences in communication between elderly people with dementia and different objects(human or robot).
Method: We conducted a controlled trial for 8 participants into two groups, G1 (average age=85.75, SD±3.7) and G2 (average age=89, SD±2.73). G1 was divided into face-to-face Reminiscence Group Therapy (RGT) and G2 into robot-teleoperated RGT for six sessions. The interventional experiment lasted from December 4, 2020, to January 12, 2021. The differences in Mini-Mental State Examination (MMSE), Neuropsychiatric Inventory-Nursing Home Version (NPI-NH), and Dementia Behavior Disturbance (DBD) scales between the two groups before and after the intervention experiment was adopted. Linguistic analysis (including sentence-ending particles and Entropy) using independent sample t-test. Results of the trend using a facial emotional recognition tool Emo-Rec App was introduced to compare differences.
Results: The two intervention methods have no obvious differences in MMSE (t(8)=0.156, p=0.881), NPI-NH(t (8)=0.380, p=0.717), and DBD(t(8)=-1.001, p=0.355) scales. Utterance(t(8)=2.434,p=0.022) and sentence-ending particle “ne” (t(8)=2.313,p=0.029) existed statistical difference. Mood changes of the humanoid robot group have improved intuitively (Emo-Rec point 10.387 to 8.531 on average).
Conclusion: The linguistic results for the teleoperated humanoid robot group were somewhat inferior. On the opposite, the mood of the robot group was improved by a better trend.Keywords: humanoid robot, teleoperation, dementia, RGT, nursing home, nursing technology
XiangYu Liu MA*, Ryuji Yamazaki PhD, Jie Yang MA, Hiroko Kase PhD (2022). Humanoid robot teleoperation reminiscence group therapy for older adults with dementia: A controlled trial study. Gerontechnology, 21(1), 1-12
https://doi.org/10.4017/gt.2023.21.1.794.02