Smartphone and the brain: stress and self-control mediate the association between the connectome-based predictive modeling of fMRI brain network and problematic smartphone use
Bijie Tie,Tianyuan Zhang,Miao He,Li Geng,Qiuyang Feng,Cheng Liu,Xuyang Wang,Yunhong Wang,Dingyue Tian,Yixin Gao,Pengcheng Wang,Wenjing Yang,Jiang Qiu
DOI: https://doi.org/10.1016/j.chb.2024.108531
IF: 9.9
2024-12-08
Computers in Human Behavior
Abstract:Background Although neuroimaging patterns linked to problematic smartphone use (PSU) are increasingly understood, studies utilizing whole-brain machine learning to identify connectome-based neuromarkers are lacking. Additionally, however the I-PACE model has identified affective (e.g., stress) and cognitive (e.g., self-control) as key contributors to PSU, the neuroscientific basis of these factors remains underexplored. This study employed connectome-based predictive modeling (CPM) to examine how distributed brain networks influence PSU and to investigate the mediating roles of stress and self-control. Methods We analyzed functional MRI and behavioral data from 403 participants (mean age, 19.37 SD = 1.24; 111 males). CPM with leave-one-out cross-validation was used to identify functional networks predictive of PSU. Additionally, results were subjected to ten-fold cross-validation. The predictive ability of the identified networks was validated using two datasets (dataset 1: n = 320; dataset 2: n = 115). Mediation analysis explored the roles of stress and self-control between CPM results and PSU. Results Connectivity predictive of PSU primarily involved connections between the frontal-parietal and the salience, motor/sensory, and visual networks, as well as connections between the motor/sensory and visual networks. The negative network connections associated with PSU, identified in one sample, was successfully generalized to predict PSU in validation datasets. Significant findings included the single mediative effect of stress and the serial mediative effect of both stress and self-control. Conclusions These findings confirm that distributed brain networks are predictive of individual PSU and highlight the need to consider both affective and cognitive factors in understanding and addressing PSU.
psychology, multidisciplinary, experimental