Related Experiment Video
Updated: Oct 2, 2026
A Machine Learning Approach to Design an Efficient Selective Screening of Mild Cognitive Impairment
Published on: January 11, 2020
Association Between Smartphone-Based IMU Turning Performance and Screening-Defined Mild Cognitive Impairment in Older
Masayuki Hoshi1, Koei Ishida2, Ayane Nakamaru3
1Department of Physical Therapy, School of Health Sciences, Fukushima Medical University, Fukushima 960-8516, Japan.
Abstract:
This study used a cognitive screening instrument rather than a clinical diagnosis to define screening-defined mild cognitive impairment (screening-defined MCI) in community-dwelling older adults. Smartphone-based instrumented Timed Up and Go (iTUG) assessment using inertial measurement unit (IMU) sensors enables quantitative assessment of individual movement components. This cross-sectional study investigated the association between Turn 1 time and screening-defined MCI in 318 community-dwelling older adults aged ≥65 years. Cognitive function was assessed using the Japanese version of the Montreal Cognitive Assessment (MoCA-J); 177 participants with scores ≤ 25 were classified as having screening-defined MCI, and 141 with scores ≥ 26 as without screening-defined MCI. Turn 1 time was significantly longer in participants with screening-defined MCI than in those without screening-defined MCI (median [IQR], 1.40 [1.20-1.50] vs. 1.20 [1.10-1.30] s; p < 0.001). Receiver operating characteristic analysis yielded an AUC of 0.694 (95% CI, 0.637-0.751), and the preliminary, data-derived cutoff maximizing the Youden index was 1.35 s (sensitivity, 0.525; specificity, 0.780). After adjustment for age and sex, each 0.1 s increase in Turn 1 time was associated with higher odds of screening-defined MCI (OR, 1.322; 95% CI, 1.176-1.487; p < 0.001), while a Turn 1 time ≥ 1.35 s was associated with approximately threefold higher odds of screening-defined MCI (OR, 3.108; 95% CI, 1.832-5.271; p < 0.001). The association remained significant after additional adjustment for maximum walking speed and psychological distress (OR per 0.1 s increase, 1.286; 95% CI, 1.135-1.457; p < 0.001). Smartphone-based iTUG assessment of turning may provide complementary mobility-related information alongside established cognitive screening measures as part of a broader assessment of community-dwelling older adults. The 1.35 s cutoff should be considered a preliminary, data-derived threshold from an exploratory analysis, should not be used for clinical decision-making, and requires validation in an independent sample before any clinical application can be considered.
