Abstracts

Rationale: 25% of adults with newly diagnosed focal epilepsy (NDfE) have an MRI-identifiable lesion [Van Paesschen et al. 1997, Neurology, 49(3):753-57] but all patients have spontaneous onset of seizures. Pathological mechanisms contributing to seizure onset in the remaining 75% are unknown. Diagnostic assessment of MRI or EEG provide limited insights into underlying mechanisms. MRI studies of hippocampal and cerebellar volumes have failed to reveal atrophy in patients relative to controls [Liu et al., 2002, Annals of Neurology, 52, 573-80]. Epilepsy is a brain network disorder [Berg and Scheffer 2011, Epilepsia 52(6):1058-62.]. We therefore employed diffusion MRI (dMRI) connectome analysis using Network Based Statistics (NBS) to identify brain alterations in NDfE. Methods: 20 adults with nonlesional NDfE (mean age=33, SD=11, 11 female) and 29 controls (mean age=32, SD=11, 16 female) were studied. Focal epilepsy was diagnosed by expert epileptologists based on ILAE classifications [Fisher et al. 2017, Epilepsia, 58:522-30] and seizure semiology. We acquired isotropic 1mm T1-weighted (T1-w) and 3.1mm multiple-shell dMRI with b-values of 1000 and 2000 s/mm2, 60 gradients each and a b0 image. T1-w Freesurfer segmentations were computed and manually corrected. ENIGMA steps [http://enigma.ini.usc.edu/protocols/dti-protocols/] were followed for dMRI artefact correction, while DSI-studio was used for upsamling the dMRI to isotropic 1.55mm and performing deterministic tractography. We performed unpaired t-tests within NBS and T-score thresholds were set to 1.5-4.0 to investigate significant effects in count, mean length, fractional/quantitative anisotropy (FA/QA), mean/axial/radial diffusivity (MD/AD/RD) and isotropy (ISO). Results: In our network of 82 cortical and subcortical regions (nodes) and corresponding connections (edges), patients showed increased AD, MD and RD across multiple NBS thresholds relative to controls (Figure 1), but no alterations in the other dMRI connectome metrics. All significant networks found in AD, MD and RD (p<0.05) largely showed the same altered edges between nodes. AD and MD represented subsets of significant edges found in MD and RD respectively. The most significant edges (T>3.1) included those between left posterior cingulate and right lateral orbito-frontal nodes with T = 3.99 for AD, T = 4.25 for MD and T = 4.3 for RD (Figure 1). In addition to this, altered MD connections were observed between left rostral middle-frontal and posteriorcingulate (T = 3.26) and inferior-temporal (T = 3.22) nodes (Figure 2A) and in RD between left pericalcerine and right parsopercularis (T = 3.92), left rostral middle-frontal and post-central (T = 3.19) and left pericalcerine and rostral middle-frontal (T = 3.12) nodes (Figure 2B). Conclusions: Our results indicate that patients with NDfE have widespread structural network alterations at time of diagnosis. Therefore, network alterations found in patients with longstanding focal refractory epilepsy may not necessarily emerge as a consequence of long-term epilepsy. Increased AD/MD/RD but unchanged FA may indicate increased inter-axonal space due to dysmyelination or dystrophy. Connectomics may have the potential to identify underlying pathological mechanisms in NDfE at time of diagnosis. Funding: BAKK: Epilepsy Action postgraduate research bursary (2014-2015)BAKK & SSK: Epilepsy Research UK (Grant number 1085)SSK: UK Medical Research Council (Grant numbers MR/S00355X/1 & MR/K023152/1)