Brain Connectivity for the Diagnosis of Epileptic Encephalopathy with Spike Wave Activation in Sleep (EE-SWAS)...

This diagnostic accuracy study demonstrates the potential of traditional feature-based and brain connectivity indicators to differentiate patients with epileptic encephalopathy with spike-wave activation in sleep (EE-SWAS) from those without neurocognitive impairment along the spectrum of self-limited focal epilepsy with centro-temporal spikes (SeLECTS) in sleep and awake electroencephalography (EEG). Both types of indicators were highly effective, with the feature-based sleep spike-wave index achieving the best univariate performance (AUC-ROC: 94.78%). Connectivity features showed significant discriminatory power, with the characteristic path length, clustering coefficient, and strength achieving AUC-ROC values exceeding 90% in awake EEG. Combining feature-based and connectivity indicators in multivariate models further improved classification, reaching a top AUC-ROC of 96.59%. These findings show the potential of brain connectivity as a biomarker for EE-SWAS, complementing traditional EEG approaches and enhancing diagnostic capabilities by being quick, objective, and automated. Furthermore, it offers valuable insights into the connectivity mechanisms underlying neurocognitive regression in SeLECTS, which appears to emerge from an "over-connectivity" hampering normal functioning.Clinical Relevance- This work proposes fully automated indicators for diagnosing the EE-SWAS.