Abstracts

Rationale: Slow spike and wave (SSW) of Lennox-Gastaut Syndrome (LGS) and generalised spike and wave (GSW) of genetic generalised epilepsy (GGE) show some similarity on scalp EEG. However the associated clinical syndromes show marked differences, with intellectual impairment, interictal paroxysmal fast activity, and medication resistant tonic seizures features of LGS. In contrast GGE patients tend to have a more benign course, usually respond well to anti-convulsants, and do not experience tonic seizures. These phenotypic differences suggest there may be fundamental differences in the generators of spike and wave activity between these conditions. We compared EEG-fMRI activation maps and the time course of signal change in key cortical and subcortical regions during SSW of LGS and GSW of GGE.Methods: Clinical i) LGS - 11 subjects (mean age 34.1 yrs) who had SSW discharges during EEG-fMRI were identified from a larger study of secondary generalised epilepsy; ii) GGE - 23 studies (21 subjects; mean age 12.3 yrs; 16 CAE, 4 JAE, 3 JME) who had GSW discharges during scanning. EEG-fMRI For each subject, up to 60 minutes of blood oxygen level dependent (BOLD) fMRI data (3T GE Signa LX; EPI, TR 3.2 s, TE 40 ms, 3.4mm3 voxels) were acquired during continuous EEG. Timings of EEG discharges were used as regressors in an event-related fMRI analysis to model BOLD signal change associated with discharges, for each group (SPM8, iBrain Analysis Toolbox). Peri-event BOLD signal time-courses were estimated across each group from regions of interest (ROIs) placed in: i) Attention (medial prefrontal cortex + intraparietal sulcus) and REST / default mode network (DMN) (posterior cingulate + lateral parietal) areas, ii) primary cortical structures, iii) thalamic nuclei and iv) pons.Results: LGS - 3272 SSW discharges (2004 secs); GGE - 451 GSW discharges (1325 secs) of epileptic activity contributed to analysis. GGE subjects showed the recognised pattern of deactivation of posterior cingulate, precuneus and lateral parietal lobe, with activation in the thalamus and mesial frontal region. LGS subjects appeared to show only deactivation in the mesial occipital region, although this was a poor representation of the widespread changes occurring. Time course analysis showed fMRI signal trending upwards in attention and REST networks leading up to the discharge, falling abruptly at or immediately after the SSW or GSW event. Primary cortical regions were relatively uninvolved during GSW, but showed an upward drift in activity prior to SSW, then a precipitous drop. The pons showed reduced fMRI activity following GSW, but a biphasic response with SSW. Conclusions: The REST/default mode network seems to behave similarly during SSW and GSW. Attentional areas also show similarities. The major differences appear to be in primary cortical areas, which show little change during GSW, but prominent deactivation during SSW. The pons is deactivated during GSW but has rising activity prior to SSW.