Abstracts

Rationale: Idiopathic generalised epilepsies (IGE) have a complex inheritance pattern. Endophenotypes are subclinical phenotypes that have a simpler genetic basis than the overall phenotype, and may represent a genetic liability to develop symptoms. Characterising such traits could be of considerable use both for understanding seizure pathology, and identifying individuals at increased risk. Although structural imaging is qualitatively normal in IGE, quantitative imaging studies have reported focal abnormalities in patients with IGE compared to controls with reduced volumes of the subcortical nuclei such as thalamus, caudate, globus pallidus and putamen (Epilepsia 2005 46(10):1642-5, J. Neurol. Neurosurg. Psychiatry 2006 77(4):489-94, Neurology 2006 67(4):683-6). There are no published studies of imaging in relatives of patients with IGE. In this study we sought to identify whether patients with IGE and their first degree relatives had smaller subcortical nuclei, using FIRST (FSL 4.14, FMRIB http://www.fmrib.ox.ac.uk/fsl/first/), (J. Magn Reson. Imaging 2011 33(4):817-21, Neuroimage 2011 56(3):907-22 ) an automated model based segmentation/registration tool which uses a training data set of manually segmented images to investigate shape and volume of subcortical grey matter nuclei (thalamus, putamen, globus pallidus, hippocampus, caudate and amygdala). Methods: 30 patients with IGE, 36 unaffected first degree relatives and 40 controls were scanned using a 3T scanner. A 3-d isotropic high resolution inversion recovery prepared spoiled gradient-recalled echo (SPGR) scan was obtained in the coronal plane (1.1mm3 voxels, TE 2.8ms, TR 7 ms, Inversion time 450ms). Each axially reoriented coronal SPGR image was processed using FIRST. FIRST gives outputs of total volume for each subcortical structure. ANOVA was used to investigate group differences between the three groups with age and total brain volume (SPM 8, VBM toolbox) entered as covariates. In order to correct for multiple comparisons, a Bonferroni correction factor of 8 was used. For those measures where it appeared that patients and relatives both differed in a similar way from controls, a 2-tailed test for linear trends was carried out. Results: Patients and relatives showed significantly lower volumes of right thalamus and right caudate compared with controls (right thalamus: F=5.51, df 2:101, p=0.005, right caudate: F=4.95, df 2:101, p=0.006). See Figure 1. There were significant linear trends for right thalamus (p=0.038), right caudate (p=0.006), left caudate (p=0.006), right (p=0.012) and left putamen (p=0.011) with relatives falling between patients and controls for all these regions. Conclusions: Our data support the hypothesis that IGE is associated with reduced volumes of subcortical nuclei, consistent with the theory that abnormal cortico-subcortical connections underlie the pathophysiology of IGE. Our data also provides evidence that such changes are found in clinically unaffected relatives of patients with IGE, suggesting that such abnormalities are a heritable endophenotype.