Computational characteristics of interictal EEG as objective markers of epileptic spasms

• DOI: 10.1016/j.eplepsyres.2021.106704
• Publication Date: 2021-06-25T20:18:56Z
• Authors: Rachel J. Smith; Derek K. Hu; Daniel W. Shrey; Rajsekar Rajaraman; Shaun A. Hussain; Beth A. Lopour
• Keywords: Approximate entropy; Sample entropy

Favorable neurodevelopmental outcomes in epileptic spasms (ES) are tied to early diagnosis and prompt treatment, but uncertainty the identification of disease can delay this process. Therefore, we investigated five categories computational electroencephalographic (EEG) measures as markers ES.We measured 1) amplitude, 2) power spectra, 3) Shannon entropy permutation entropy, 4) long-range temporal correlations, via detrended fluctuation analysis (DFA) 5) functional connectivity using cross-correlation phase lag index (PLI). EEG data were analyzed from ES patients (n = 40 patients) healthy controls 20 subjects), with multiple blinded measurements during wakefulness sleep for each patient.In patients, amplitude was significantly higher all electrodes when compared controls. lower than control subjects. The DFA intercept values subjects, while exponent not different between groups. networks stronger based on both PLI. Significance statistical tests p < 0.05, adjusted comparisons Benjamini-Hochberg procedure appropriate. Finally, logistic regression, a multi-attribute classifier derived that accurately distinguished cases (area under curve 0.96).Computational features successfully distinguish large, study.These objective markers, combination other clinical factors, may speed treatment disease, thereby improving long-term outcomes.