The emotional landscape of fibromyalgia: a supervised machine learning approach

The Imbalance of Threat and Soothing Systems model of fibromyalgia proposes that affect (dys)regulation, indexed by an overactive threat system and an underdeveloped soothing system, may contribute to fibromyalgia. This model has not been tested empirically. Individuals with fibromyalgia and controls (N = 2416) completed measures of affect regulation and symptomatology. At group-level, participants with fibromyalgia reported greater intensity of threat-related emotions (e.g., anger, anxiety, and disgust) and lower drive- (e.g., pleasure, enthusiasm) and soothing-related emotions (e.g., contentment, relaxation). Supervised machine learning using affect regulation system scores and discrete emotions distinguished fibromyalgia from controls with good discrimination (AUC = 0.74–0.83). Threat-related emotions contributed most to fibromyalgia classification, whereas soothing-related emotions contributed most to control classification. These findings align with the model and highlight affective processes as clinically relevant correlates and potential treatment targets in fibromyalgia.