Abstracts

Rationale: Seizures are common in critically ill neonates. Their recommended diagnosis is via full array conventional electroencephalograms (EEGs). A precise measure of the dose of neonatal seizures is needed to extend our understanding of their injurious capabilities. Seizures are usually quantified solely by their “temporal” characteristics: frequency (number of seizures per hour) or burden (the percentage of time that each hour of EEG demonstrates seizures in any brain location). However, no study has specifically incorporated seizures’ spatial distributions. This study compares two “temporal-spatial” metrics of neonatal seizure severity by measuring seizure burden on a region-by-region basis.Methods: Two unique 5-channel EEG montages were generated (montages A and B, see image) from the full array neonatal EEG. All recommended neonatal electrode positions were included in both montages but with no electrode redundancy. A convenience sample of full array neonatal EEGs with seizures was reviewed by consensus by two readers. For each EEG, seizure severity was quantified in three ways: (i) seizure frequency, (ii) seizure burden, and (iii) seizure burden on a region-by-region basis for the 5 locations of montages A and B. The resultant values were compared by Spearman’s rank coefficient to determine correlation. Internal validity was estimated by re-measuring the same channel twice independently (channel 5, see image).Results: Ten neonatal EEGs with seizures were evaluated. Seizure frequencies ranged from 1 to 14 seizures per hour and seizure burdens ranged from 1.6% to 54.5%. There was a strong positive relationship between seizure frequency and burden in this sample (Pearson’s coefficient = 0.769). A total of 18,268 seconds of seizure across 116 seizures were captured in the full array conventional EEG montage. Montages A and B both yielded similar values of seizure burden on a region-by-region basis (Pearson’s coefficient = 0.9916, r2 = 0.983). There was a strong correlation between both novel measures and seizure burden as defined by total duration of EEG record seizing (Montage A vs conventional measure: Pearson’s coefficient = 0.9681, r2 = 0.937; Montage B vs conventional measure: Pearson’s coefficient = 0.9505, r2 = 0.904).Conclusions: There was no significant difference in the measurement of seizure burden on a region-by-region basis comparing montages A and B. Further study using only one montage will examine if seizure severity is adequately assessed using only the convention measure of seizure burden.