Is Emotional Allodynia in Fibromyalgia Reducible to Depression? A Preliminary Network Analysis in 149 Patients

Background/Objectives: Because affect is challenging to quantify in fibrom-yalgia, previous network analyses have primarily relied on symptom-severity measures, whereas emotion regulation and affective hypersensitivity have rarely been included. We examined where emotional allodynia, the tendency to respond with disproportion-ate distress to neutral or low-intensity interpersonal cues, sits in this network. Methods: Cross-sectional analysis of 149 consecutive outpatients with fibromyalgia (2016 Ameri-can College of Rheumatology criteria; 91.9% women; mean age 57.5 years). A regularized partial correlation network (EBICglasso: extended Bayesian information criterion graphical lasso; γ = 0.5, Spearman) was estimated over twelve nodes: emotional allo-dynia, pain catastrophizing, central sensitization, depression, state and trait anxiety, and six facets of emotion dysregulation. Accuracy was assessed with 5000 bootstrap replica-tions. Results: The correlation of emotional allodynia with depression (ρ = 0.502) was shrunk to exactly zero in the network and under every sensitivity analysis, and was near zero when the penalty was removed (ρ = -0.033). Its two strongest edges were with pain catastrophizing (0.190) and the Impulse facet of emotion dysregulation (0.184), non-zero in 98.9% and 99.6% of replications. It was among the less connected nodes, but, forming a community of one, all its connectivity crossed a domain boundary. Of that, 59.8% went to emotion regulation and 30.2% to pain, the largest share of any node outside either domain (77.0% and 48.1% of replications). Depression and anxiety received 10.0% (24.6% for catastrophizing). Conclusions: In this exploratory analysis, emotional allodynia oc-cupied a bridging position directed at emotion regulation and pain rather than at de-pression and anxiety, and was not reducible to depressive symptom severity once the other measures were held constant. It overlapped substantially with catastrophizing, and these data do not adjudicate their separability. These preliminary single-centre findings are hypothesis-generating and require replication.