The brain causal network features in seizures provoked by COVID-19 using EEG

Seizures have been the most common neurological symptom in children with COVID-19, often occurring afebrile and distinct from febrile seizures in onset age, suggesting the unique mechanisms. Comparing these cases with healthy subjects and febrile seizures may help to identify neuroimaging biomarkers of COVID-provoked seizures. Here we collected EEG data from patients with febrile seizures unrelated to COVID-19 (FS-N), with seizures provoked by COVID-19 (PSP-COV) and healthy controls (HC) to construct causal brain networks using conditional Granger causality (CGC). Causal connections with group-difference were used for machine learning classification and to identify most affected areas. The results indicated that the CGC connections showed more significant group differences in the sleep state than those in the awake state. When classifying using connections with group-difference as features, the classification accuracy was 92.86% between FS-N and HC, 90.9% between PSP-COV and HC, and 84.62% between PSP-COV and FS-N. The top three discriminative connections between FS-N and PSP-COV were found during sleep. High-contribution connections featured positive out-in degrees in frontal and temporal lobes, whereas parietal and occipital lobes showed negative values. Results indicate sleep-specific EEG causal network disruption in PSP-COV, notably involving altered connections from the posterior temporal to parietal lobes. Additionally, features distinguishing FS-N and PSP-COV identified frontal and temporal lobes as sources and parietal and occipital lobes as targets in the EEG causal network, offering new insights into brain network mechanisms of seizures provoked by COVID-19.