Abstracts

Rationale: There has been considerable interest in using broadband intracranial EEG (iEEG) particularly high-frequency oscillation (HFO) and infraslow activity (ISA) in addition to conventional frequency (CFA) for localization of epileptic zone. Prior study reported that in lesional epilepsy, HFOs are linked to seizure onset zone and are independent of the pathology. The objective of the study was to evaluate if there is any difference in the rate and spatial distribution of HFO and ISA between lesional and non-lesional epilepsy. Methods: 9 patients who underwent intracranial monitoring with two-stage epilepsy surgery were selected and divided in 2 groups: (1) Lesional (L) (n= 6): lesion seen on MRI (n=3) or confirmed on pathology (n=3). (2) Non-lesional (NL) (n=3) with no lesion on MRI or pathology. iEEG with sampling rate of 1000 Hz for all the habitual seizures were analyzed under referential montage and CFA (1-70 Hz ,30 mm/second). Seizure onset was defined as earliest sustained rhythmic change in EEG clearly distinct from the background and accompanied by a habitual clinical seizure. HFO and ISA were marked if discrete HFO (≥70 Hz, 60mm/second timescale) or ISA (≤0.01 Hz, with deflection ≥ 0.5 mV, 10mm/second timescale) was seen 1 min prior or simultaneous to CFA. All channels with presence of HFO or ISA were recorded with earliest time of appearance of such activity. Analysis was performed for presence of HFO and ISA, temporal relation to CFA onset and their spatial distribution using chi square statistic. Results: We analyzed total of 30 seizures: 20 in group L and 10 in group NL. ISA was more frequent in NL (9/10 vs 11/20, p= 0.1), and also preceded the CFA seizure onset (9/9 vs 3/11 seizures, p=0.001). Spatially, ISA was more localized and involved the same electrodes as CFA onset in group L, while in the NL group, seizures showed more widespread involvement of electrodes, with an onset outside of CFA. For the HFO analysis, seizures from group L generated higher rate of HFOs as compared NL group (17/20 vs 4/10, p=0.03). Among the seizures from lesional group with HFO, only one of them preceded CFA onset, but involved the same electrodes as CFA. In the non-lesional group, 3 out of 6 (50%) seizures showed presence of HFO of which 2 were leading the CFA, but involved the same electrodes as CFA onset. Conclusions: Preliminary findings from our study suggest that HFO and ISA are differentially generated between lesional and non-lesional localization related epilepsy. In comparison to lesion group, ISA is more prevalent and tends to precede the CFA in patients with non-lesional epilepsy, which also involve area larger than what is defined with CFA. On the other hand, HFO are more prevalent in lesional epilepsy though they originate from the same area as CFA seizure onset. Though limited by small sample size, our study suggests that broadband EEG is might provide additional information in localization and defining the seizure onset zone. The underlying mechanisms for these observations need to be further investigated perhaps with a larger sample size.