Abstract
BACKGROUND: Excessive screen time among college students is increasingly prevalent and may impair executive functions, particularly working memory (WM). However, the behavioral and neural mechanisms remain unclear. METHODS: A total of 42 college students participated in the experiment and were assigned to either a high screen time group (HSTG) or a low screen time group (LSTG). Brain activity was measured with functional near-infrared spectroscopy (fNIRS) covering the frontal, temporal, and parietal regions during the 2-back working memory task. Group differences in behavioral performance (accuracy, reaction time, false alarms), task-related activation, functional connectivity, and graph-theoretical network were analyzed. RESULTS: LSTG participants demonstrated significantly higher accuracy and hit rates than those in HSTG, while no group differences were observed in reaction time or false alarm rate. Neuroimaging analyses revealed greater activation in bilateral dorsolateral prefrontal cortex (DLPFC-R: p< 0.001, DLPFC-L: p = 0.007) as well as premotor and supplementary motor cortex (PreM & SMC-R, p = 0.007) in LSTG. Functional connectivity was higher in LSTG at whole-brain (p = 0.047), intra-hemispheric (right: p = 0.022, left: p = 0.049), and inter-hemispheric levels (p = 0.033). Graph-theoretical results further indicated lower clustering coefficient (p = 0.040) and network density (p = 0.035) in HSTG, although global and local efficiency did not differ between groups. CONCLUSION: High screen exposure is linked to reduced working memory accuracy, weaker prefrontal engagement, and disrupted network organization, which suggests reliance on less efficient neural strategies. Screen use may thus represent a modifiable factor affecting cognition and brain networks.