Objective: The understanding of childhood absence epilepsy (CAE) has been revolutionized over the past decade, but the biological mechanisms responsible for variable treatment outcomes are unknown. Our purpose in this prospective observational study was to determine how pretreatment ictal network pathways, defined using a combined electroencephalography (EEG)–functional magnetic resonance imaging (EEG?fMRI) and magnetoencephalography (MEG) effective connectivity analysis, were related to treatment response.
Methods: Sixteen children with newly diagnosed and drug?naive CAE had 31 typical absence seizures during EEG?fMRI and 74 during MEG. The spatial extent of the pretreatment ictal network was defined using fMRI hemodynamic response with an event?related independent component analysis (eICA). This spatially defined pretreatment ictal network supplied prior information for MEG?effective connectivity analysis calculated using phase slope index (PSI). Treatment outcome was assessed 2 years following diagnosis and dichotomized to ethosuximide (ETX)–treatment responders (N = 11) or nonresponders (N = 5). Effective connectivity of the pretreatment ictal network was compared to the treatment response.
Results: Patterns of pretreatment connectivity demonstrated strongest connections in the thalamus and posterior brain regions (parietal, posterior cingulate, angular gyrus, precuneus, and occipital) at delta frequencies and the frontal cortices at gamma frequencies (P < .05). ETX treatment nonresponders had pretreatment connectivity, which was decreased in the precuneus region and increased in the frontal cortex compared to ETX responders (P < .05).
Significance: Pretreatment ictal connectivity differences in children with childhood absence epilepsy were associated with response to antiepileptic treatment. This is a possible mechanism for the variable treatment response seen in patients sharing the same epilepsy syndrome.