Network analysis of anxiety, depression, and resilience among students at the University of Rajshahi, Bangladesh

Abstract

Anxiety and depression are common and often co-occurring mental health problems among university students, while resilience is frequently inversely associated with psychological distress. In Bangladesh, these constructs have mostly been studied in isolation, and their interrelationships at the symptom level remain poorly understood. This study aimed to examine the network structure of anxiety, depression, and resilience among Rajshahi university students in Bangladesh. This cross-sectional study included 423 university students with complete data in Bangladesh. Anxiety, depression, and resilience were assessed using the Generalized Anxiety Disorder-7 (GAD-7), Patient Health Questionnaire-9 (PHQ-9), and Connor–Davidson Resilience Scale-10 (CD-RISC-10), respectively. A regularized Gaussian graphical model was estimated using the graphical least absolute shrinkage and selection operator with Extended Bayesian Information Criterion model selection. Bridge centrality indices were calculated to identify symptoms and resilience components with comparatively strong cross-community connectivity across three a priori scale-based domain which were anxiety, depression, and resilience. Network robustness was assessed using bootstrap procedures. The network was visualized using three pre-specified scale-based domains representing anxiety, depression, and resilience. Anxiety and depression symptoms were positively interconnected, whereas resilience nodes showed predominantly negative connections with psychological distress. Depressed mood (PHQ2) had the highest positive bridge expected influence, followed by irritability (GAD6) and feeling nervous (GAD1), suggesting that these symptoms showed the strongest positive cross-community associations. In contrast, resilience items such as bouncing back after hardship (R5) and achieving goals despite obstacles (R6) showed the most negative bridge expected influence values, indicating inverse cross-community associations. Robustness analyses supported the stability of the estimated network, with correlation stability coefficients of 0.671 for bridge expected influence and 0.593 for node strength. This study identified an interconnected pattern of conditional associations among anxiety, depression, and resilience. Depressed mood and irritability showed strong positive cross-community connectivity, whereas specific resilience components showed negative cross-community associations with psychological distress. These findings are hypothesis-generating and may help prioritize symptoms and resilience components for further investigation in longitudinal studies among university students.