Abstracts

Rationale: Functional magnetic resonance (fMRI) technique that has been used most frequently is the BOLD (Blood Oxygen Level Dependent) method. Temporal resolution is high in EEG but lacks spatial resolution which is superior with fMRI. Complementary advantages of combined EEG and fMRI (EEG-fMRI) can identify activation and deactivation in BOLD contrast related to interictal epileptiform discharges (IEDs). To date majority of EEG-fMRI reported studies have been in focal and idiopathic generalized epilepsies (IGE) but not in secondary generalized epilepsy (SGE). The archetypical SGE is Lennox Gastaut syndrome (LGS). The study was undertaken to probe the extent of epileptic network dysfunction in SGE in patients with LGS and determine whether filling in the prevailing gap would lead to understand some of mechanisms of SGE with tonic or atonic seizures Methods: The study was approved by the Austin Health Human Research Ethics Committee and informed consent was obtained in the 13 SGE patients with LGS .They had continuous EEG-fMRI in a 3T GE magnet. The total EEG-fMRI scan time was 2.5 hours. Functional images were acquired using multi-slice gradient-recalled echo-planar imaging sequences. For pre-processing and data analysis we used in-house iBrain™ software. The hemodynamic response function to the spiking was modelled using an event-related design and estimated using Statistical Parametric Mapping (SPM2) software. Results: We recruited 13 subjects with LGS (5 women, mean 35 years range 25-52; median age at seizure onset was 3 years). All 13 patients had tonic or atonic and atypical absence seizures, confirmed on video EEG monitoring in the majority. MRI was abnormal in 9 (69%) subjects. Generalized paroxysmal fast activity (PFA) and generalized slow spike wave (GSW) were the 2 main IEDs in our cohort of LGS. In 12 (92%) patients 21 IED events were : 6 PFA, 7 GSW, 6 sharp waves and 2 polyspike wave discharges were identified. The cortical and subcortical BOLD signals are shown in table 1. Illustrative BOLD responses associated with PFA and GSW in 2 patients (#1,#5) are shown in figure 1. Both patients had PFA and GSW IEDs. In PFA the cortical BOLD is positive but with GSW the cortical BOLD signals are biphasic. Caudate, thalami and brainstem are also involved in PFA and GSW (table 1). Conclusions: In all patients in LGS with SGE there was a cortical BOLD response in 100% of IEDs. The PFA showed intense positive diffuse cortical BOLD signal in the cortex and subcortical structures. The GSW also showed diffuse cortical but the BOLD responses were both positive and negative. The intense positive only and biphasic cortical BOLD signals appear to be IED event specific. The GSW pattern in LGS resembles IGE. The most salient features in our study were the subcortical thalamic (77%) and brainstem (62%) BOLD signals and majority were positive. We propose that the cerebral cortex is the network driver for seizure propagation in SGE to thalamus and brainstem and that the mechanism for tonic or atonic seizures is due to brainstem recruitment supported by animal and PET studies in LGS.