Abstracts

Rationale: Despite the cessation of ictal neuronal firing following temporal lobe seizures, patients still have a high chance of experiencing loss of normal memory or consciousness. This phenomenon suggests a post-ictal brain status during which multiple brain networks involved in information processing are still affected by seizures. Previous intracranial EEG (iEEG) studies have suggested the correlation of cortical slow waves with seizure-induced loss of consciousness (Englot et al., 2010). In the current study, we propose to use dense-array EEG to image the distribution of postictal slow-wave (PIS) and to investigate the relationship of the PIS with consciousness phenomena in epilepsy. Methods: EEG recordings were obtained from four patients with epilepsy during dense-array EEG monitoring. An approximately 1 min post-ictal period was segmented. Seizures of the same type in the same patients were concatenated. We then applied independent component analysis (ICA) to separate signals into independent components (ICs). The ICs of artifactual origin or ICs that could not be explained by dipolar sources were removed. We clustered IC maps of all the patients using the K-means method. The clusters centroids were localized using the source imaging technique onto a standard Coline MRI brain. Results: Six clusters were identified. The within cluster correlation was greater than 0.8, suggesting similar PIS networks across the investigated cohort. Three clusters were localized to the ipsilateral temporal lobe (IT), the ipsilateral posterior temporal lobe (IPT) and the contralateral temporal lobe (CT), respectively. The IT component appeared in all of the patients (100%) and had more than 30% of the total delta-band power. The localization of this component was concordant with surgically removed regions. The PIS in the IPT (in 3 patients, 75%) and CT (in 2 patients, 50%) had much smaller power than the IT. The second group exhibited midline frontal and parietal lobe distribution. The three components were present in 3 (75%), 2 (50%) and 3 patients (75%), respectively. The three clusters were localized to the medial prefrontal cortex with spread to the cingulate gyrus and precuneus. These clusters were highly co-localized with previously-reported seizure-induced cerebral blood flow (CBF) decrease in the default mode network (Blumenfeld et al., 2009). The common appearance of the PIS in this network of patient group may explain the loss of consciousness, another commonality across patients.Conclusions: We have investigated the distribution of PIS networks in a small group of epilepsy. The significant PIS in the ipsilateral temporal lobe suggests the localization value of the PIS for the delineation of the epileptogenic brain. In addition, the results convey the importance of the default mode regions in maintaining normal consciousness. The common appearance of the PIS in this network may explain the loss of consciousness, another commonality across patients. The co-localization of the EEG-PIS signal and the previously reported CBF change suggest the neuronal origin of the altered consciousness in temporal lobe epilepsy.