Abstracts

Rationale: Juvenile Myoclonic Epilepsy (JME) is characterised by generalised tonic-clonic seizures and myoclonic jerks, sometimes triggered by cognitive effort. Previous imaging work has identified altered fronto-parietal and thalamo-cortical functional and structural connections in patients with JME during a working memory task1,2. In the absence of a cognitive task, we investigated whole-brain functional connectivity in patients with JME and healthy controls using a graph network approach applied to resting-state fMRI.Methods: 25 JME and 21 control participants underwent a resting-state fMRI scan and a high-resolution structural scan as part of a larger study1,2. 90 brain regions were defined from the Automated Anatomical Labelling atlas, and a mean timecourse was obtained for each region by averaging timeseries from all voxels within the region. Correlation coefficients were computed between each pair of regions to obtain a 90x90 adjacency matrix per subject. Functional networks were thresholded at the mean plus 1 standard deviation of each subject’s connectivity matrix and binarised. Average network metrics – degree centrality (DC), clustering coefficient (CC) and path length (PL) – were compared between groups using two-sample t-tests. Graph metrics related to the relative importance (or “hubness”) of a node in a network, DC and eigenvector centrality (EVC), were calculated and compared using two-sample permutation t-tests and p-values from the empirical distributions thresholded at p<0.05. Network edge strengths were also compared using permutation t-tests thresholded at p<0.001.Results: Averaged network metrics did not differ between groups. Networks in both groups had higher CC than and similar PL to random networks, indicating a small-world topology of brain networks in both groups. The JME group had higher DC and EVC in several frontal and parietal regions (Table 1) including the right superior frontal region (DC: p=0.007), while reduced DC and EVC were found in the thalamus (EVC: p=0.009) and temporal pole. Comparing correlation strengths between regions, increased correlation was found in patients with JME between the right superior frontal and the right inferior temporal regions (p=0.0004), while reduced correlation was found between bilateral thalami and the left occipital cortex (left: p=0.0009; right: p=0.0007).Conclusions: The results show widespread alterations in the brain network hubs of patients with JME compared to controls, predominantly in, but not limited to, fronto-parietal regions and the thalamus. These findings suggest an over-connected fronto-parietal network and an under-connected thalamus in patients with JME. The results from this whole-brain network analysis have some similarities with previous findings from the task-based fMRI study in the same cohort, as well as some differences, suggesting that brain network alterations are present during the inter-ictal resting state and are modulated by a cognitive task. References: 1. Vollmar, C. et al. Brain 2011; 134:1710-1719. 2. O'Muircheartaigh, J. et al. Brain 2012; 135:3635-3644.