Combining magnetic resonance imaging and genome-wide genetic mapping in epilepsy

Recent attempts to elucidate the genetic architecture of complex epilepsies have been limited by a variety of issues including phenotypic ambiguities, small sample sizes and restricted genetic scope. This thesis employed a diverse array of genetic mapping and MRI techniques to help improve the power of genetic mapping efforts and thus our understanding of the neurobiological factors contributing towards common forms of the disorder. As part of a collaborative meta-analysis of GWAS data on complex epilepsies, we contributed to the identification of loci at 2q24.3 (implicating the SCN1A gene) and 4p15.1 (implicating the PCDH7 gene) as risk factors for epilepsy in general, in addition to a third locus at 2p16.1 (implicating VRK2) as a risk factor for idiopathic generalised epilepsy. Genetic variants showing strong signals of association with epilepsy (from our meta-analysis) also significantly correlated with genetic variants showing strong signals of association with hippocampal volume (from an independent meta-analysis; ENIGMA2). Nominal ensembles of these hippocampal variants did not predict disease state (epilepsy patient/healthy control); suggesting that larger samples and more fine-grained phenotypes are necessary in order to detect any polygenic effect of ENIGMA2-Hippocampus variants on epilepsy predisposition. Turning our attention to phenotypic ambiguities in epilepsy, we employed diffusion MRI to a series of epilepsy patients with no abnormal findings on routine medical imaging (‘non-lesional’ epilepsy), their gender-matched, asymptomatic siblings, an independent group of epilepsy patients with mesial temporal sclerosis (‘lesional’ epilepsy) and healthy control participants. We identified distinct patterns of structural connectivity disruption in both the lesional and non-lesional epilepsy groups, in addition to supporting the role of heritable, co-segregating microstructural changes along limbic white matter tracts as intermediate phenotypes (endophenotypes) in non-lesional epilepsy patients and their unaffected siblings. This thesis has illustrated how multi-modal medical imaging and genetic mapping techniques can be employed as two distinct but complimentary approaches to elucidate novel genetic and neuroanatomical aspects of epilepsy. Future studies could combine these techniques in the large-scale collaborative application of imaging genetics in the disorder.