Amygdala volume and psychiatric burden in focal epilepsy: PET evidence for a glial substrate

Abstract Temporal lobe epilepsy is of heterogeneous pathology. While many patients present with hippocampal sclerosis and ipsilateral mesiotemporal atrophy, a subgroup of patients without human-detectable sclerosis via structural MRI are characterized by ipsilateral amygdala enlargement and worse postsurgical outcomes. A bidirectional relationship between temporal lobe epilepsy and depression is well-established, with amygdala enlargement and neuroimmune alterations increasingly implicated as substrates for both. In this study, we interrogate the link between amygdala volume and psychiatric burden in temporal lobe epilepsy and evaluate gliosis as a possible substrate. We hypothesize that gliosis drives ipsilateral amygdala volume changes that underlie psychiatric burden in patients with no hippocampal sclerosis. To test these relationships, we analyzed two independent cross-sectional cohorts: 1) A Discovery cohort (43 patients, 80 controls) with structural MRI and robust psychiatric measures, and 2) A Mechanistic pilot cohort (10 patients, 9 controls) with both structural MRI and [18F]FEPPA PET/MR, an in vivo measure of gliosis, to conceptually replicate and extend findings from the Discovery cohort. Amygdala volumes were assessed with FreeSurfer. Psychiatric burden in the Discovery cohort was assessed with the Depressive Problems subscale from the Achenbach Adult Self-Report. Regional gliosis in the Mechanistic cohort was quantified using [18F]FEPPA PET, and an exploratory measure of psychiatric burden for this cohort was obtained via medical chart review. In the Discovery cohort, the group without hippocampal sclerosis showed significant ipsilateral enlargement of the whole amygdala, basolateral, and central subnuclei compared to controls and the group with sclerosis, and a group-dependent association between decreased amygdala volume and increased depressive symptoms. In the Mechanistic cohort, increased [18F]FEPPA binding was associated with reduced amygdala volume. Adjusting for this effect unmasked significant amygdala enlargement in the group without hippocampal sclerosis. Exploratory analysis revealed a possible central subnucleus-specific relationship between gliosis, volume, and psychiatric burden in patients without sclerosis such that amygdala volume was positively correlated with psychiatric burden specifically in patients with a greater degree of gliosis. These data associate amygdala volume alterations with psychiatric burden in patients without sclerosis and implicate gliosis as a potential mechanistic substrate. Findings support a conceptual shift towards understanding amygdala enlargement not as a static structural phenotype, but as a dynamic, glia-associated state that may confer risk for psychiatric illness. Consequently, patients without hippocampal sclerosis may derive a distinct dual benefit from therapeutic strategies targeted at underlying neuroimmune function.