Abstracts

Rationale: Lennox-Gastaut syndrome (LGS) is a severe form of childhood-onset epilepsy characterised by drug-resistant generalized tonic seizures, characteristic electroencephalogram (EEG) patterns, and cognitive impairment. The electroclinical phenotype arises from diverse aetiologies, including structural lesions and genetic abnormalities. Fluorodeoxyglucose Positron Emission Tomography (18FDG-PET) during the interictal state often shows reduced brain metabolism in brain networks involved in the epileptic process. We aimed to show the cortical regions involved in the epileptic process of LGS by comparing brain metabolism in a group of patients with LGS to a group of epilepsy controls.  Methods: We compared 18FDG-PET metabolism in 18 patients with LGS (mean age=26 years) to the uninvolved hemispheres in 39 age- and sex-matched control patients with temporal lobe epilepsy (21 with left temporal lobe epilepsy [L-TLE] and 18 with right TLE [R-TLE]; mean age=28 years). Each patient and control 18FDG-PET scan was first split into left and right brain hemispheres. Each hemisphere was then intensity normalised to the mean signal within cortical white matter; co-registered to the patient's T1-weighted MRI; spatially warped and re-sampled to Montreal Neurological Institute space (2mm3 voxels); converted to z-score units by subtracting the image mean and dividing by the image standard deviation; and spatially smoothed at 10mm. Non-parametric, permutation-based, two-sample t-tests were used to compare the left and right hemispheres of the LGS patients to the 'unaffected' hemispheres of the epilepsy control patients (i.e., left hemisphere in LGS vs left hemisphere in R-TLE; right hemisphere in LGS patients vs. right hemisphere in L-TLE). Results: In both left and right hemispheres, group analysis found significant hypo-metabolism (LGS relative to epilepsy controls; p<0.025, uncorrected) in midline and lateral fronto-parietal cortex, including middle and superior frontal gyri, posterior cingulate, precuneus, and angular gyri (Fig. 1). Hypo-metabolism was also found in the anterior insula and anterior cingulate/paracingulate cortex. These cortical areas correspond to the location of the Executive Control Network (ECN; dorso-lateral prefrontal, inferior parietal cortex), the Anterior Salience Network (ASN; anterior cingulate, anterior insula, and lateral frontal cortex) and Default Mode Network (DMN; precuneus, mesial frontal cortex). Some hyper-metabolism was also noted in peri-central cortex. Conclusions: Patients with LGS have areas of 18FDG-PET hypo-metabolism that closely resemble the areas of cortical activation we have previously reported using simultaneous EEG-functional MRI of interictal generalised paroxysmal fast activity (