Abstracts

Rationale: Because of the evidence suggesting distinct mechanisms underlying different seizure-onset patterns (SOP) in the electroencephalogram (EEG), we hypothesize that the interictal epileptic activity might show differences as well. We investigated the interictal signature of the two most common SOPs in mesiotemporal lobe epilepsy (mTLE), low-voltage fast activity (LVF) and high-amplitude low-frequency periodic spiking (PS). Methods: We searched our intracranial database for mTLE patients with either predominant LVF or PS. Rates of interictal epileptiform discharges (IEDs) and high frequency oscillations (HFOs) were computed. IEDs were characterized by the spike/polyspike amplitude, the amplitude of the post-spike slow wave, and the associated low-frequency (0.3-4 Hz) and high-frequency (>80 Hz) power increases. Results: Eighteen subjects (8 women; mean age, 36.7±12.1 years) were selected with uni- or bilateral mTLE and a predominant LVF (8 patients) or PS (10 patients) SOP. Figure 1 illustrates the inter-individual differences in IEDs of both SOPs providing representative examples of all investigated patients. Interictal spikes in LVF patients were less sharp as determined by the high-frequency power increase (LVF: median 10.5 dB, range -1.5–25.0 dB vs. PS: median 12.7 dB, range -8.5–28.1 dB, p=0.006), had a 1.2-times higher post-spike slow wave amplitude (p=0.02), and a 1.5-times larger low-frequency power content than IEDs in PS (p=0.007). IEDs in PS, however, co-occurred more frequently with fast ripples than ripples (ripple-IED vs. fast ripple IED co-occurrence difference: p=0.014). Figure 2 provides the time frequency plots of the IEDs averaged for the different SOPs. There was no difference in rates of IEDs, HFO rates, and spike amplitude. Conclusions: We demonstrated for the first time an association between the morphology present at seizure onset and the morphology of IEDs. Major differences comprised IED sharpness, post-IED slow wave amplitude and IED-HFO co-occurrence. These characteristics reflect the interictal expression of the different underlying pathophysiological mechanisms of seizure generation. Our data support the notion that antiepileptic drug therapy should be tailored to the respective underlying epileptic mechanism. Funding: This work is supported by the Austrian Science Fund (Schrödinger fellowship J3485-B24) and the Canadian Institutes of Health Research (grants MOP-102710 and FDN-143208 to J.G.). Figure legends. Figure 1. Representative examples of IEDs across the individual patients of the LVF (A) and PS (B) group. Note that for this illustration only IEDs of the most common IED subtype per patient were chosen. Figure 2. Time frequency plots of the IEDs averaged for the different SOPs: LVF (A), PS (B), difference between LVF and PS (C). Time zero corresponds to the peak of the spikes, with an automatic fine alignment to the local maxima closest to the visually placed mark. The white lines at the bottom of the plots indicate the intervals during which the high-frequency content (-50 to 50 ms) and low-frequency content (50 to 300 ms) were computed. The difference between the time frequency plots of LVF and PS patients revealed a significantly lower high-frequency power content as correlate for less sharp spikes in LVF compared to PS (see arrow at time zero), whereas the low-frequency power content as expression for the amplitude and the duration of the post-spike slow wave was significantly larger in LVF compared to PS (see arrow at time 0.2).