Abstracts

Rationale: The neonatal period carries significantly high risk for seizures in the human lifespan. Seizures in the neonate are different electrographically and clinically from those in older children; they have no characteristic morphology (even within the same patient) except for rhythmicity, spikes or sharp waves may not be present, and they may remain localized with little tendency to evolve(1). The latter features are reminiscent of afterdischarge (AD) and lateralized periodic discharge (LPD) that are spatially 'fixed' and minimally-evolving epileptic rhythms in the adult brain. We have shown previously (2,3) that the synchronous waveforms of AD and LPD arise from a stimulus- or disease-associated process of coalescence ('spectral condensation') of the background EEG rhythms. We hypothesized that neonatal seizures also arose from a similar process. Methods: Power spectra of EEG epochs of duration 10 seconds, sampled between 200-500 Hz and band-pass filtered between 1.6-70 Hz, from five neonates with seizures were computed. Results from twenty (N=20) discrete EEG seizures were compared against 100 control epochs. Line length of the area-normalized spectral profile was used as a measure of the 'peakedness' of the spectra, and statistical comparison drawn between seizure and baseline spectra. Results: In all instances, visual analysis confirmed that the peaks of the seizure power spectra arose from within the frequency mix of the baseline control spectra, suggesting that the dynamic process of spectral condensation was operative (Figure). Line lengths of seizure spectra were significantly different from baseline spectra (p < 0.05, 2-sample t test). Different EEG seizure types ?" across patients, and within the same patient ?" differed only in the positions of the peaks and their harmonics, which were often minor. Conclusions: (i) Neonatal seizures arise from the same phenomenological mechanisms that underlie ADs and LPDs ?" the coalescence of the baseline frequency mix of the baseline EEG into one or more dominant modes and their harmonics, (ii) the balance between excitation and inhibition amongst the various modes may determine which of them dominate in a particular seizure, (iii) the variety of morphology of neonatal seizures are due to minor variations in the selection of dominant modes and the power attributed to their harmonic overtones, (iv) new lines of therapy ?" pharmacological or neuromodulatory - directed to the neural circuits that generate the key dominant modes(4) are suggested. References: 1. Clancy RR & Mizrahi EM (2006). In Wyllie, E (Ed). The Treatment of Epilepsy. 2. Kalamangalam GP, Tandon N & Slater JD (2014). Clin Neurophys 125 (7): 1324-1328. 3. Kalamangalam GP & Slater JD (2015). J Clin Neurophys 32(4); 331-340. 4. Buszaki G (2006). Rhythms of the brain. Figure Caption: Spectral condensation underlying neonatal seizures; all data from a single patient with multiple seizures. x-axis: Hz, y-axis: arbitrary units. A1: Power spectrum of a 10-s baseline (seizure-free) EEG epoch showing a mix of frequencies. A2, A3: Spectra from a seizure: specific spectral components (arrows) from within the baseline mix attain dominance. The initial part of the seizure had a slightly different EEG morphology (not shown) from the later part, reflected in the slight differences in the relative power of the seizure spectral peaks between A2 and A3. B1, B2, B3; C1, C2, C3: Similar figures from other segments of the same EEG trace showing baseline spectra (B1, C1) and seizure spectra (B2, C2, C3). B3 is a post-ictal trace following seizure B2; the third harmonic of the main (fundamental) peak has disappeared and the fundamental itself is seen to 'spread out' as the condensation process reverses. Funding: NINDS award K23NS079900 to GPK.