Abstracts

Rationale: Significance of ictal pathological High Frequency Oscillations (HFOs in 80-200 Hz) using MEG in focal epilepsy has been little known. Our aim was to study its spatial, spectral characteristics and evolution of HFOs from preictal to ictal period and identification of seizure onset zone.Methods: Twenty patients (M: F=13:7; Mean age=23±8.2) with drug resistant focal epilepsy had ictal MEG recording (sampling rate of 2000Hz) as a part of pre-surgical evaluation. After eliminating MEG channels and segments of data with artefacts, ictal onset was identified using simultaneously acquired EEG after band passing at 1-70 Hz. Data was then filtered at 80-200 Hz band pass using Hann filter, ‘ictal’ (complete seizure activity from onset till offset) and ‘pre-ictal’ activity (data preceding the ictal onset with same duration as the ictal activity) was viewed at 1second per page for characterizing morphological patterns of HFOs (ripples) and calculating the power of the ripples (fT2/Hz) in these periods. Time frequency decomposition was then performed with Short Time Fast Fourier Transform (STFFT) on the same segments with 16 to 32 ms (ms=millisecond) temporal resolution and 50% overlap between each FFT bin. Spectral power mapping and HFOs source localization was then performed with sLORETA (Standardized Low Resolution Electromagnetic Tomography) on a realistic head model on the patient MRIResults: MEG identified ripples in both periods pre-ictal (n=13; 65%) and ictal (n=19; 95% patients), dominating mainly during the ictal period. Though highly variable, the morphological characteristics of identified ripples were: 4 to 6 oscillations with peak amplitude ranging from -2.2 pT to +1.93 pT (pT=pico Tesla; 1pico =10 -12), lasting for 20-40 milliseconds; the corresponding STFFT showed rise in spectral power between 80-140 Hz frequencies across the patients. The ratio of power of ripples during ictal period was 61.3% higher than the pre-ictal period (P <0.003) across all the patients. Ictal onset HFOs source localization showed concordance with corresponding lesion (50%), sub-lobar level (30%), lobar level (10%) and discordance (10%).Conclusions: This study reveals that there is a definite temporal increase in pre-ictal to ictal cortical hyper excitability as evident by the evolution of HFOs in these periods. HFOs were predominantly higher during ictal activity and its localization (seizure onset zone) had a very good spatial concordance with the lesional (epileptogenic) zone in 80% of the patients. Thus, ictal HFOs can detect the seizure onset zone with minimal spatial error. To avoid spurious HFOs identification and localization, we performed morphological characteristics, spatial topography, spectral power and source level analysis.