Abstracts

Rationale: Idiopathic generalised epilepsies have a complex inheritance pattern and are predominantly polygenic (Epilepsia 2005;46(S10):7-14). Previous studies support the genetic basis of both qualitative non-epileptiform and epileptiform EEG traits, with an increased prevalence of epileptiform abnormalities in unaffected relatives of patients with epilepsy compared with controls (Acta Neurol Scand. 1996;93(1):9-13). These characteristic abnormalities or endophenotypes in clinically unaffected relatives may represent a genetic liability to develop symptoms. Characterising such traits could be of considerable use both for understanding seizure pathology, and for identifying individuals at increased risk. Given that epilepsy is characterised by abnormalities in the synchronisation of neuronal activity (Neuron 2006; 52(1):155-68), we sought to identify altered synchronisation properties and network topology in a group of patients with idiopathic generalised epilepsy and their first degree relatives using quantitative methods. Methods: We recorded resting, eyes-closed EEG data from 15 patients with idiopathic generalised epilepsy (IGE), 19 of their unaffected relatives, and 24 healthy controls. The phase locking factor (PLF) was calculated for each pair of channels in several different frequency bands, and compared between groups. Additionally we transformed this phase locking data into a series of unweighted graphs, and assessed, for each network, basic facets of network topology including the clustering coefficient and path distance. We then compared these network properties between groups. Results: IGE patients differed significantly from controls in mean PLF (p=0.025, averaging across all frequency bands) and mean clustering coefficient (p =0.33) with a strong trend towards significance for mean path length (p=0.073). In each case clinically unaffected relatives showed a similar pattern of difference from controls. This was significant in the case of mean clustering coefficient (p=0.047) and showed a strong trend towards significance in the case of of mean PLF (p = 0.063, averaging across all frequency bands). These differences were more pronounced at higher frequencies. Conclusions: Our data strongly suggests that IGE is associated with significant changes to the phase synchronisation properties and network topology of resting electrical activity. They also provide evidence that such changes are found in clinically unaffected relatives of patients with IGE, suggesting that such abnormalities may by characteristic of those with a high liability to develop IGE.