Dynamic limbic–executive decoupling underpins the reciprocal link between drug‐resistant state and depression in temporal lobe epilepsy

OBJECTIVE: Depression is the most common psychiatric comorbidity in temporal lobe epilepsy. It is not merely a psychological reaction to seizures; accumulating evidence supports a bidirectional relationship in which mood dysregulation may also precede drug resistance. Here, we tested whether dynamic limbic-executive integration provides an in vivo circuit-level substrate for this reciprocity. METHODS: We cross-sectionally studied 82 patients with temporal lobe epilepsy (33 drug-resistant, 49 drug-responsive) and 50 healthy controls. The drug-resistant state was defined according to International League Against Epilepsy (ILAE) criteria (failure of two appropriate and tolerated anti-seizure medication regimens), and depression severity was indexed by Self-Rating Depression Scale scores. Using resting-state functional magnetic resonance imaging (rs-fMRI), we quantified time-varying large-scale network coupling to characterize limbic interactions with the broader connectome. We then identified specific patterns of limbic disintegration and tested their clinical relevance with mediation models, asking whether limbic integration with executive systems statistically linked drug-resistant state and depression severity bidirectionally. Finally, we spatially correlated network effects with normative positron emission tomography (PET)-derived neurotransmitter receptor/transporter maps and Allen Human Brain Atlas gene-expression profiles, followed by pathway enrichment analyses. RESULTS: The drug-resistant state was characterized by dynamic decoupling of the limbic network from dorsal attention and frontoparietal systems. Reduced limbic-dorsal attention integration tracked depression severity and statistically mediated both the pathway from the drug-resistant state to depression (indirect = 2.18, 95% confidence interval [CI] = 0.20-4.54) and the reverse pathway from depression to the drug-resistant state (indirect = 0.022, 95% CI = 0.006-0.049). The spatial signature of limbic-executive disintegration co-localized with higher serotonin (5-HT) 1B receptor (5-HT1B) and lower dopamine D2 receptor distributions and was transcriptomically enriched for cellular stress adaptation and cell-fate pathways, including translation reinitiation, telomerase activity, and tumor necrosis factor-related apoptosis-inducing ligand (TRAIL)-mediated apoptotic signaling. SIGNIFICANCE: These convergent multi-scale results propose a dynamic limbic-executive disintegration model of the reciprocity between seizure intractability and depression. Our findings reframe comorbidity as coupled systems-level dysfunction and motivate integrated stratification and target prioritization for dual-burden-focused neuromodulatory intervention.