Cerebrospinal fluid TNF-α, regional degree centrality, and depressive symptoms after herpes simplex encephalitis: a cross-sectional observational study

Abstract

Background Herpes simplex encephalitis (HSE) survivors frequently experience neuropsychiatric sequelae, including depressive symptoms, but the biological correlates underlying post-HSE depression remain incompletely understood. Neuroinflammation and altered functional brain organization have been implicated in depression; however, their relationships after HSE have not been characterized. This study investigated the associations among cerebrospinal fluid (CSF) tumor necrosis factor-α (TNF-α), resting-state functional brain connectivity, and depressive symptoms in HSE survivors. Methods This two-center cross-sectional observational study included adults with HSE and healthy controls recruited between June 2023 and June 2025. Participants were classified into healthy controls (HC), HSE without depression (HSE-ND), and HSE with depression (HSE-D) groups. Depressive symptoms were assessed using the 17-item Hamilton Depression Rating Scale (HAMD-17). CSF inflammatory cytokines were measured, and resting-state functional MRI was performed. CSF TNF-α was prespecified as the primary cytokine for the imaging, hierarchical regression, and exploratory mediation analyses; the other cytokines were evaluated in secondary group-comparison and symptom-correlation analyses. Voxel-wise degree centrality (DC) analysis was used to identify functional network alterations associated with post-HSE depression. Associations among CSF TNF-α, regional DC values, and depressive symptom severity were further evaluated using correlation, regression, and exploratory mediation analyses. Results A total of 74 participants were included (24 HC, 27 HSE-ND, and 23 HSE-D). CSF TNF-α concentrations were higher in HSE-D than in HC and HSE-ND and were positively associated with HAMD-17 scores in the HSE cohort. The primary whole-brain voxel-wise contrast identified candidate clusters with lower DC in the right amygdala, right insula, left fusiform gyrus, and left postcentral gyrus. In secondary regional comparisons, HSE-D versus HSE-ND differences remained significant in the right amygdala and left postcentral gyrus but not in the left fusiform gyrus or right insula. CSF TNF-α was inversely associated with DC in the right amygdala, right insula, and left postcentral gyrus. Regional DC values were negatively correlated with depressive symptom severity. Exploratory hierarchical models showed greater in-sample explained variance after adding CSF TNF-α and regional DC to clinical covariates, but Model 3 was not a validated independent-predictor model. Standardized indirect-effect point estimates were 0.149 for the right amygdala, 0.081 for the left fusiform gyrus, 0.163 for the right insula, and 0.091 for the left postcentral gyrus. These exploratory estimates require validation. Conclusion In HSE survivors, elevated CSF TNF-α and altered functional brain network organization were associated with depressive symptoms. The regional pattern is exploratory and should not be interpreted as specific to a primary depressive disorder, independent of neurological sequelae, or causal. Longitudinal studies are needed to determine temporal relationships, compartmental inflammatory correlates, and clinical relevance.

The authors' abstract, as published at the source. Frontiers in Neurology, 2026 · DOI ↗