Abstracts

Rationale: Childhood absence seizures are characterized by brief impairments of consciousness with bilaterally synchronous 3 Hz spike and wave discharges on EEG. High frequency oscillations (HFOs) (80-500Hz) are thought to be promising biomarkers of epileptic tissue and they have been clearly linked to the seizure onset zone in patients with intractable focal epilepsies. This goal of this study was to evaluate whether HFOs occur during childhood absence seizures and to what areas does that activity localize. The corticothalamic system plays a major role in the underlying pathophysiology of CAE. The purpose of this study was to use magnetoencephalography (MEG) recordings during childhood absence seizures to complete time-frequency analyses of cortical and thalamic activity. Methods: Children, aged 8 to 12 years old, with newly diagnosed and untreated absence seizures were recruited for the study. MEG recordings were conducted on a 275 channel CTF magnetometer and completed in 10 minute time blocks. Time-frequency analysis using short time fast Fourier transform (STFFT) was completed during absence seizures for multiple frequency bandwidths: 1-20Hz, 20-70Hz, 70-150Hz, and 150-300Hz. Source localization for each of these frequency bandwidths was then completed using a sLORETA algorithm for sources identified with independent component analysis (ICA). The first spike-wave complex of the seizure was compared to a spike-wave complex occurring 3 seconds later but during the same seizure. Epochs of 25 milliseconds were analyzed starting 50 milliseconds prior to spike of interest, for a total time interval of 300 milliseconds. Results: Four children, aged 8 to 12 years old, with newly diagnosed and untreated childhood absence seizures were recruited and thirteen absence seizures occurred during MEG recording. Time-frequency analysis with STFFT showed significant power density in the 1-20Hz, 20-70Hz, and 70-150Hz bandwidths (Figure 1A). Several components were identified with ICA between the 1-20Hz, 20-70Hz, and 70-150Hz but no components above the noise level were identified between 150-300Hz. Source localization using a sLORETA algorithm demonstrated consistent localization preferentially in the parietal region for activity at 1-20Hz and localization preferentially to the lateral frontal region at 20-70Hz (example, Figure 1B). Whenever a component occurred between 70-150Hz it localized only to a unifocal area in the lateral frontal region. Conclusions: Using MEG, we have been able to detect focal ictal HFOs in untreated children with absence seizures. Spatial localization of activity at lower frequency bandwidths (1-20Hz) tended to localize to the parietal regions whereas activity at higher bandwidths (20-70Hz, 70-150Hz) localized to a unifocal area in the lateral frontal region. Further studies will help to determine whether areas of HFOs are a critical component of the network responsible for generating childhood absence seizures and whether their presence correlates with treatment response or prognosis.