文献类型：期刊文献

英文题名：Wearable Technology-Enhanced Assessment of the Effect of Interpersonal Trust on Adolescents' Physical Activity Behavior: The Mediating Role of Family Capital

作者：Li, Baixia[1]; Zhang, Zhiling[2]; Chen, Yonghong[3]; Zhang, Bo[4]; Chen, Maoshui[4]; Zeng, Xiangyi[4]; Li, Qunfeng[5]

机构：[1] School of Sports Science and Engineering, East China University of Science and Technology, Shanghai, Xuhui, China; [2] East China University of Science and Technology, Shanghai, China; [3] Zhu Hai No.13 Middle School, Guangdong, Zhuhai, China; [4] Dept Orthoped Spinal Surg 2, Guangdong Prov Hosp Chinese Med, Guangdong, Guangzhou, China; [5] Jinan University, Zhuhai Campus, Guangdong, Zhuhai, China

年份：2026

起止页码：470

外文期刊名：Proceedings of 2025 2nd International Conference on Sports Technology and Performance Analysis, ICSTPA 2025

收录：EI(收录号：20261720581313)

语种：英文

外文关键词：Behavioral research - Data aggregation - Data reduction - Noise abatement - Pattern recognition - Pipeline processing systems - Pipelines - Quality control - Signal to noise ratio - Social sciences computing - Wear of materials - Wearable computers

摘要：Objective: Accurate measurement of physical activity in adolescents has long been challenging due to the limitations of self-report methods. Recent advances in wearable accelerometer technology offer objective, continuous, and high-resolution data collection capabilities that can revolutionize behavioral research. This study presents a comprehensive computational framework for processing and analyzing accelerometer data in adolescent populations, demonstrating its application in investigating psychosocial determinants of physical activity behavior.Methods: We deployed ActiGraph wGT3X-BT triaxial accelerometers on 358 secondary school students for 7-day continuous monitoring. Raw acceleration data were processed using a multi-stage computational pipeline: (1) Choi et al. algorithm for automated non-wear time detection with 90-minute window and 2-minute spike tolerance; (2) Butterworth low-pass filtering at 0.25 Hz cutoff for noise reduction; (3) 60-second epoch aggregation for computational efficiency; (4) Evenson cut-points for intensity classification; and (5) quality control criteria requiring ≥4 valid days with ≥10 hours wear time. After processing, 326 participants were included in analysis. We then applied this objective measurement framework to test hypotheses about interpersonal trust, family capital, and physical activity using structural equation modeling.Results: The computational pipeline successfully processed 2,282 person-days of accelerometer data with high reliability. Technical validation showed: (1) non-wear detection accuracy of 94.3%; (2) signal-to-noise ratio improvement of 18.2 dB after filtering; (3) epoch aggregation reduced data volume by 98.3% while preserving activity patterns. Applying this framework to behavioral analysis revealed: (1) interpersonal trust had a significant direct effect on objectively-measured physical activity (β=0.20, p ? 2025 Copyright held by the owner/author(s).