Characterization of neonatal seizures by conventional EEG and single-channel EEG

Objective: To perform a detailed, contemporary temporal-spatial characterization of neonatal seizures (NS) and to compare conventional EEG (CEEG) to single-channel EEG for NS detection. Methods: Digitally recorded CEEGs were reviewed for NS characteristics (quantity, duration, location of onset, peak-to-peak amplitude). The presence and characteristics of each NS were simultaneously noted in a single, derived EEG channel (C₃ → C₄). Results: Eight hundred fifty-one seizures from 125 CEEGs recorded were analyzed. Mean seizure rate was 7.0 NS/h (range: 0.5-21). Mean seizure burden (percent time CEEG showed NS at any location) was 24.8% (range: 0.7-86.9). Seizure rate was only moderately correlated with seizure burden (Spearman coefficient = 0.58). Eighty-one percent of NS originated from central-temporal or midline vertex electrodes. Seventy-eight percent of NS appeared in the C₃ → C₄ channel. Conclusions: Accurate measurement of NS burden requires detailed temporal-spatial characterization. The theoretical ceiling of sensitivity for NS detection in the single EEG channel C₃ → C₄ is high. However, further processing the raw EEG in limited electrode arrays may reduce the sensitivity of NS detection. Significance: CEEG is the gold standard for NS detection. However, reduced montage EEG techniques are increasingly available. This detailed contemporary temporal-spatial characterization of NS evaluates the potential limitations of reduced montage techniques.