Anxiety disorders and depression are highly prevalent, debilitating, and commonly comorbid mental disorders. Epidemiological, symptomatic, and pathogenetic perspectives suggest a particularly strong relationship between generalized anxiety disorder (GAD) and major depressive disorder (MDD), leading to a continuous debate about their nosological and neurobiological uniqueness [1]. Traditional case control studies in either MDD or GAD patients suggest shared dysregulations in emotional and cognitive domains and underlying amygdala-frontal circuits [1, 2], while initial studies have reported disorder-specific dysregulations in the insular and ventrolateral prefrontal cortex (vlPFC)[3]. During recent years, dysregulations in social processes, such as emotional empathy, have gained increasing attention as transdiagnostic etiological and diagnostic factors for internalizing disorders, including both depression and anxiety [4]. Expression of emotional empathy is dependent on the integrity of the anterior insula (AI) and adjacent ventral frontal regions, but while initial evidence suggests that empathic experience is impaired in MDD [5], common and distinct neural alterations in this domain between MDD and GAD and whether these vary between physical and affective pain observation have not been examined.

To this end, the present neuroimaging study examined neural empathic reactivity and everyday empathic experience in unmedicated, treatment-naïve first-episode GAD (n= 35) and MDD (n= 37) patients and healthy controls (HC, n= 35) by means of a validated blocked-design emotional (pain) empathy fMRI paradigm [6] employing visually presented physical and affective pain stimuli, as well as corresponding control stimuli (Fig. 1 a). To account for potential cognitive alterations during the symptomatic state, participants were asked to passively view the stimuli. Following initial quality assessments, 30 sociodemographically matched-individuals per group entered the final analysis (online suppl. Fig. S1 and Table S1; for all online suppl. material, see www. karger. com/doi/10.1159/000504180). Categorical diagnostics were conducted during hospital admission and independently confirmed by a standardized clinical interview according to DSM IV criteria. In addition, GAD and MDD symptom load was dimensionally assessed using validated self-report scales (Penn State Worry Questionnaire, PSWQ; Beck Depression Inventory II, BDI-II), indicating that depressive symptom load was higher in MDD compared to both, GAD patients and HC, while GAD symptom load was higher in both patient groups relative to HC (all ps Bonferroni-corrected< 0.001) but comparable between patient groups (p> 0.1). MRI data collection, preprocessing, and modeling adhered to evaluated standard protocols (online suppl. material). The main interaction contrast of interest ([affective pain> affective control]>[physical pain> physical control]) was subjected to a voxel-wise mixed ANOVA, including group as between-subject factor (FWE-p cluster= 0.05, whole-brain). Results of this categorial strategy revealed a significant interaction effect of diagnosis in a cluster predominantly located in the right dorsal anterior insula (rdAI) spreading into the adjacent vlPFC (FWE-p cluster= 0.023, k= 119). Post hoc analyses demonstrated that compared to HC, MDD patients exhibited exaggerated neural reactivity during affective pain yet attenuated reactivity during physical pain observation in this region, whereas GAD patients did not (Fig. 1 b). Based on the contribution of both the insula and right amygdala to emotional empathy [7], functional connectivity alterations between these regions were …