Abstracts

Rationale: Patients diagnosed with childhood absence seizures often undergo short episodes of impaired attention and display 3 4 Hz spike and wave complexes on electroencephalography (EEG). Results from previous studies on neuroimaging indicate that EEG and behavior changes during absence seizures might be due to changes in cortical and subcortical attention networks. Here, we employ a data-driven method to analyze fMRI timecourses to identify networks responsible for impaired consciousness during absence seizures.Methods: Two task paradigms were used to measure attentional vigilance from patients with childhood absence epilepsy (CAE) during simultaneous EEG-fMRI recordings: (1) repetitive tapping task (RTT) where patients pressed a button to letters shown on a screen every second, and (2) continuous performance task (CPT) where patients pressed a button only when presented with the target letter ( X ).Results: fMRI percent change timecourses during peri-ictal periods (before, during, and after seizures) showed BOLD signal changes in attention networks over time. Variable task performance during seizures was observed, with several patients showing relatively spared performance. A data-driven approach applied to the percent change signals revealed three clusters of closely related fMRI signal changes: default mode regions, thalamus and primary sensory cortices, and frontal cortical areas. fMRI signal changes from these three regions were greater in seizures associated with poor performance (<25% correct) than in seizures with good performance (>75% correct).Conclusions: Our results suggest that higher fMRI signal changes from the three main time course clusters during seizures are correlated with poor performance in CAE patients. The data-driven approach used here may reveal specific networks tied to impaired performance and consciousness observed in CAE patients.