Identifying Mild Cognitive Impairment by Using Human–Robot Interactions

Article type: Research Article

Authors: Chang, Yu-Ling^{a; b; c; d; *} | Luo, Di-Hua^a | Huang, Tsung-Ren^{a; b; d} | Goh, Joshua O.S.^{a; b; d; e} | Yeh, Su-Ling^{a; b; d; e} | Fu, Li-Chen^{f; g; h}

Affiliations: [a] Department of Psychology, College of Science, National Taiwan University, Taipei, Taiwan | [b] Neurobiology and Cognitive Science Center, National Taiwan University, Taipei, Taiwan | [c] Department of Neurology, National Taiwan University Hospital, College of Medicine, National Taiwan University, Taipei, Taiwan | [d] Center for Artificial Intelligence and Advanced Robotics, National Taiwan University, Taipei, Taiwan | [e] Graduate Institute of Brain and Mind Sciences, College of Medicine, National Taiwan University, Taipei, Taiwan | [f] Department of Computer Science and Information Engineering, National Taiwan University, Taipei, Taiwan | [g] Department of Electrical Engineering, National Taiwan University, Taipei, Taiwan | [h] MOST Joint Research Center for AI Technology and All Vista Healthcare, Taipei, Taiwan

Correspondence: [*] Correspondence to: Yu-Ling Chang, PhD, Department of Psychology, National Taiwan University, No. 1, Sec. 4, Roosevelt Rd, Taipei 10617, Taiwan. Tel.: +886 2 33663105; Fax: +886 2 23629909; E-mail: [email protected]; ORCID: 0000-0003-2851-3652.

Abstract: Background: Mild cognitive impairment (MCI), which is common in older adults, is a risk factor for dementia. Rapidly growing health care demand associated with global population aging has spurred the development of new digital tools for the assessment of cognitive performance in older adults. Objective: To overcome methodological drawbacks of previous studies (e.g., use of potentially imprecise screening tools that fail to include patients with MCI), this study investigated the feasibility of assessing multiple cognitive functions in older adults with and without MCI by using a social robot. Methods: This study included 33 older adults with or without MCI and 33 healthy young adults. We examined the utility of five robotic cognitive tests focused on language, episodic memory, prospective memory, and aspects of executive function to classify age-associated cognitive changes versus MCI. Standardized neuropsychological tests were collected to validate robotic test performance. Results: The assessment was well received by all participants. Robotic tests assessing delayed episodic memory, prospective memory, and aspects of executive function were optimal for differentiating between older adults with and without MCI, whereas the global cognitive test (i.e., Mini-Mental State Examination) failed to capture such subtle cognitive differences among older adults. Furthermore, robot-administered tests demonstrated sound ability to predict the results of standardized cognitive tests, even after adjustment for demographic variables and global cognitive status. Conclusion: Overall, our results suggest the human–robot interaction approach is feasible for MCI identification. Incorporating additional cognitive test measures might improve the stability and reliability of such robot-assisted MCI diagnoses.

Keywords: Cognitive assessment, dementia, health care, human–robot interaction, mild cognitive impairment, older adults

DOI: 10.3233/JAD-215015

Journal: Journal of Alzheimer's Disease, vol. 85, no. 3, pp. 1129-1142, 2022