Cross-sectional network structure of alexithymia, empathy, anxiety, and depression in male inpatients with alcohol use disorder

Background Alcohol use disorder (AUD) frequently co-occurs with alexithymia, empathy-related difficulties, anxiety, and depression, but their conditional associations remain unclear in male inpatients with AUD. Methods We analyzed data from 385 male inpatients with AUD treated at a psychiatric specialty hospital in China. Alexithymia, empathy, anxiety, and depression were assessed with the Toronto Alexithymia Scale-20 (TAS-20), the Chinese Interpersonal Reactivity Index (IRI-C), the 14-item Hamilton Anxiety Rating Scale (HAM-A), and the 24-item Hamilton Depression Rating Scale (HAMD-24), respectively. We estimated a Gaussian graphical model and examined one-step expected influence (EI), one-step bridge expected influence (BEI), and bootstrap-based accuracy and stability. Results Difficulty identifying feelings (DIF) had the highest EI and exceeded empathic concern (EC), HAMD-24, and psychic anxiety (PSY) in bootstrap comparisons. Under the primary community assignment, HAMD-24, DIF, and PSY had relatively large BEI point estimates, with no bootstrap-supported pairwise differences among them. Externally oriented thinking (EOT) was negatively associated with perspective taking (PT), whereas personal distress (PD) was positively associated with PSY. For somatic anxiety (SOM), the retained between-community edges with PT and fantasy (FS) were weaker than the within-community edges with PSY and HAMD-24. Correlation stability coefficients for strength, EI, and BEI ranged from 0.595 to 0.673. Conclusions Conditional associations among alexithymia, empathy, anxiety, and depression varied across nodes in male inpatients with AUD. Under the primary community assignment, HAMD-24, DIF, and PSY had relatively large BEI point estimates, with no bootstrap-supported pairwise differences among them.