Abstracts

Rationale: Significance of Ictal pathological High Frequency Oscillations (HFOs in 80-200 Hz) using MEG in focal epilepsy has not been studied. We aimed to study the occurrence of HFOs, its spatial and spectral characteristics during the preictal and ictal phases. Methods: MEG (306 channels) recording at 2000Hz sampling rate and analysis of ictal record was performed in 5 patients (M: F= 1:4; age: 23.13 ± 6 years) with drug-resistant focal epilepsy. Segments with artefacts were eliminated. The ictal onset was identified visually at band pass of 0.3-70 Hz; topographical analysis and source localisation was carried out using s-LORETA (standardized Low Resolution Electromagnetic Topography) on boundary element head model. Data was then filtered with a band pass of 80-200 Hz using Hann filter and viewed at 1second epoch/page to study HFOs (ripples). Morphological patterns of HFOs were identified. The time-frequency analysis was performed using Short Time Fast Fourier Transform (STFFT) in 80-250 Hz bandwidth. To understand the evolution of HFOs from the preictal to ictal period, power of ripples (fT²) was calculated during these 2 time periods over a fixed time window that varied among patients. Spectral power mapping was computed for entire frequency bandwidth (80-200 Hz) during ictus. To avoid spurious HFOs, morphological characteristics of HFOs, dipolar topography, rise in spectral power and spectral power mapping was performed. Results: Source localisation at the ictal onset was at lateral temporal cortex -3; right superior parietal cortex -1; and left frontal lobe-1patient. MEG identified ripples in both the phases- pre-ictal (n=3) and ictal (n=5 patients), dominating during the ictal phase. The morphological characteristics of ripples were: 4 to 6 oscillations with peak amplitude of -2.2 to +1.93 pico Tesla (1pico =10 ^-12), lasting for 20-40 milliseconds; the corresponding STFFT showed rise in spectral power between 80-140 Hz frequency across the patients. The ratio of power of ripples between ictal to pre-ictal phase was >70% (at significant P value <0.05) in all 5 patients, which explains the evolution of HFOs. There was a good concordance between the ictal onset source localisation and the ripple spectral power mapping in the brain regions. Conclusions: This is the first comprehensive study involving MEG which analysed the occurrence, morphological and spectral characteristics of pathological ictal HFOs. Ripples were predominant during the ictus compared to the pre-ictal phase. Identification of ictal onset might be possible by analysing spectral power changes in ripple bandwidth. Analysis of ictal HFOs could allow us to non-invasively localize the seizure onset zone and understand its evolution.