Abstracts

Rationale: Intracranial EEG (iEEG) used for monitoring patients with intractable epilepsy for seizure focus localization before surgical intervention are limited by temporal and spatial resolution, and hence sometimes fail to satisfactorily localize a seizure onset zone. The use of micro-electrode arrays and hybrid macro/micro depths have localized electrographic activity and pathological high-frequency oscillations to submillimeter scale [1,2]. Here, we analyze micro iEEG ( iEEG) recorded from PMT 4X4 grid and Adtech hybrid depth electrodes with the following objectives: i) To seek signatures helpful in localizing a seizure onset which might not be visible in the iEEG lower frequency (<30 Hz) bands used for clinical evaluation; ii) Study seizure propagation on micro scales.Methods: i) The recorded iEEG was decomposed into its spectral contents using short time Fourier transform (STFT) over a 1 s window with 50% overlap and its logarithmic ratio with a baseline was compared with that of the neighboring iEEG channels and/or the ones that clinically detected the first seizure onset. ii) The 20-50 Hz band of each of the 16 iEEG electrode was divided by that for a baseline signal. The time instant when the calculated ratio for each electrode exceeds a threshold value (K) for 3 consecutive time points were used to determine the order in which the ictal activity propagates across the microgrid.Results: The analysis in i) and ii) were applied to a data set from a patient implanted with both microgrids and hybrid depth as shown in Figure1. i) The clinical onset occurred at left fronto- temporal macro1 (LFT1), the topmost spectrogram in Fig1(b). The following subplots correspond to iEEG recorded from hybrid hippocampal depth electrodes 2, 3 and adjacent micro contacts 6, 8. The vertical black line indicates the instant of spectral power rise in the low frequency band in LFT1 and it can be seen that there is a chirp like spectral power rise in depth micro 8, 1-2 seconds before that in LFT1. However, the adjacent depth macro contacts did not show this early seizure onset. ii) The same seizure as in Fig1(b) recorded by a PMT micro grid (brown dot in Fig1(a)) shows a microflow pattern as in Fig1(c). Clinically, the seizure propagated across the LFT grid from 1-2-11-5 and then to 29,30 (Fig 1(a)). However, the microflow seems to be in a direction opposite to that expected. Conclusions: From i) it seems like there was some micro seizure activity which was not detected by the macro contacts or the seizure activity in other parts was reflected as gamma power rise in the hybrid depth micros. ii) Suggests that the ictal propagation pattern on micro scales might not be predicted by that observed on larger spatial scales. Hence, we conclude that study of iEEG recordings of ictal activity is important for seizure localization for a precise targeting which will help improve surgical outcomes as well as the efficacy of implantable anti-epileptic devices. References: [1]Stead et al. (2010) Brain 133: 2789 2797; [2]Schevon et al. (2012) Nature Communications 3:1060