Abstracts

Rationale: While drug-resistant temporal lobe epilepsy (TLE) is primarily characterized by mesial temporal atrophy, there is evidence for amygdala enlargement (AE) unrelated to an expansive process. Previous studies were designed to assess amygdalar structural integrity in relation to emotional processing and affective disorders (Tebartz van Elst L, et al. Biological Psychiatry 1999;46:1614-23; Tebartz Van Elst L, et al. Brain 2002;125:140-9). On the other hand, studies directly evaluating the significance of AE in TLE have been either based on qualitative evaluation (Mitsueda-Ono T, et al. J Neurol Neurosurg Psychiatry 2011;82:652-7) or limited to patients with normal hippocampal volumetry (Bower SP, et al. J Neurol Neurosurg Psychiatry 2003;74:1245-9). Our objective was to localize regions of AE in TLE by means of MRI-based 3D surface-based shape modeling. In addition, we sought to investigate the relationship between AE and psychic epileptic phenomena as well as co-morbid psychiatric disease.Methods: Based on manual volumetry of mesiotemporal lobe structures (Bernasconi N, et al. Brain 2003;126:462-9) and a 2 SD cut-off from the mean volume of healthy controls, we classified TLE patients from our database (183 consecutive cases) into those with AE (TLE-AE, n=25; 14%) and those with normal amygdalar volume (TLE-NAV). TLE-AE patients were compared to a cohort of age- and sex-matched healthy controls (n=40), as well as a group of TLE-NAV (n=25) matched for sex, age, seizure onset, EEG lateralization, history of febrile convulsions and hippocampal atrophy. In each individual, we obtained surface-based measurements of local volumetric changes of the amygdala relative to a template model (Kim H, Besson P, et al. MICCAI 2008;11:1008-15). The vertex-wise analyses were thresholded using the false discovery rate procedure at FDR<0.05 to correct for multiple comparisons.Results: Regions of hypertrophy in TLE-AE were located mainly in the centromedial nuclear group (Figure). On the other hand, TLE-NAV presented with ipsilateral atrophy in the laterobasal nuclear group. The majority (72%) of TLE-AE had normal hippocampal volume (Table). Furthermore, the frequency of patients with favorable post-surgical outcome in this group was lower than in TLE-NAV (40% vs. 71%; p=0.09). Post-ictal psychosis was present in 16% of TLE-AE, whereas it was absent in TLE-NAV (p=0.12). After combining post-ictal psychosis, ictal fear and d j vu, we observed a trend for a greater frequency of these psychic epileptic phenomena in TLE-AE patients (56% vs. 28%; p=0.09). On the other hand, there were no differences in the prevalence of axis I or II disorders (24% vs. 40%; p=0.36).Conclusions: AE is present in 14% of all TLE patients, of whom the majority (72%) have normal hippocampal volumes. AE affects nuclear groups that are distinct from those that undergo atrophy. There appears to be an association between TLE-AE and poor surgical outcome, as well as psychic epileptic phenomena. The relationship between TLE-AE and psychiatric disorders remains less clear.