What can "drag & drop" tell? Detecting mild cognitive impairment by hand motor function assessment under dual-task paradigm. Issue 145 (January 2021)

• Record Type: Journal Article
• Authors: Zhang, Yingwei; Chen, Yiqiang; Yu, Hanchao; Lv, Zeping; Yang, Xiaodong; Hu, Chunyu; Zhang, Tengxiang
• Is Part Of: International journal of human-computer studies. Issue 145(2021)
• Publication Date: 2021-01
• Subjects: Mild cognitive impairment -- Hand motor function -- Dual-task -- Touchscreen; Human-machine systems -- Periodicals; Systems engineering -- Periodicals; Human engineering -- Periodicals; Human engineering; Human-machine systems; Systems engineering; Periodicals; Electronic journals; 004.019
• DOI: 10.1016/j.ijhcs.2020.102547
• Languages: English
• ISSNs: 1071-5819

Abstract

Highlights: CogSYS, a MCI detection system, including various cognition assessment tasks, was designed, optimized, and developed in this study. Through ML-based auxiliary diagnostic method and statistic-based analysis, cogSYS provides comparable accuracy and significant insights. We discuss the influences of "drag & drop" on cortical activation, demonstrating the consistency of designed goals and the involved brain areas. Abstract: Early diagnosis of mild cognitive impairment (MCI) is critical for reducing the incidence of serious neurodegenerative diseases. However, current diagnostic solutions, such as biomarkers and cognitive screening tests, are expensive, time-consuming, or not user-friendly. In this study, we designed cogSYS to explore the feasibility and accuracy of detecting MCI through one-minute hand motor function assessment. Inspired by a clinically validated screening test, cogSYS contains a series of touchscreen-based colorful block "drag & drop" tasks, including four unilateral gross manual dexterity single-tasks and three language function related dual-tasks for each single-task. We study how to design and optimize these interactive tasks, and evaluate them through three user studies (i.e., effectiveness verification, cortical activation analysis, and user experience feedback). Experimental results show that cogSYS can detect MCI effectively, with a mean accuracy of 82.4%. Furthermore, by statistical comparison on features, we discover that the velocity and time-based features of failure circumstance are the most effective among all the features. These discoveries can provide insights for follow-up research and clinical applications. Cortical activation analysis shows that bilateral prefrontal cortices, bilateral motor cortices, and the occipital lobe are involved in "drag & drop" tasks, proving the effectiveness of cogSYS in specific cognitive functions. User experience feedback shows that the subjects evaluate highly to cogSYS and provide valuable information for further improvements.