Abstracts

Rationale: The aim was to define epileptic network pathways during childhood absence seizures using a combined MEG and fMRI connectivity analysis. Methods: Magnetoencephalography (MEG) and combined EEG-functional magnetic resonance imaging (EEG-fMRI) were recorded in 7 participants with untreated childhood absence seizures. EEG was used to identify timing of absence seizures during fMRI and an event related independent component analysis (eICA) method identified region of blood oxygenation level dependent (BOLD response) correlating with the seizures. These regions identified were then parcellated to express the absence network in terms of functional nodes. These nodes were used as virtual sensor locations for a linearly constrained minimum variance (LCMV) beamformer analyses of absence seizures recorded using MEG in the same subjects. Group time frequency analysis (FFT) was used to identify the bandwidths with dominant power. After extracting source waveforms, the effective connectivity was estimated using a phase slope index (PSI) metric. Results: eICA of fMRI data identified regions similar to those previously reported (thalamus, frontal, precuneus, biparietal) (Figure 1). Thirty-four seizures were recorded during MEG and used for effective connectivity analysis. PSI at 3-4Hz and 13-30Hz showed connections within and between the parietal cortex, precuneus, and thalami. The main drivers of information during the seizures were the posterior cingulate, middle frontal gyrus, precuneus, occipital cortex, and the angular gyrus. The major receivers of information were thalamus, temporal cortex, precentral gyrus, precuneus, posterior cingulate, and occipital cortices. At higher frequencies (30-55Hz), connectivity tended to be within and between frontal regions with the major drivers of information flow present in the inferior frontal gyrus and occipital cortex (Figure 2). The major receivers of information at this higher frequency were in the thalamus and precentral gyrus. Inter-subject variability was noted and there was a difference in the proportion of thalamic connections in treatment responders (49% 26%, N=5) versus non-responders (28% 8%, N=2). Conclusions: fMRI-informed MEG analysis can be used to identify brain connectivity during generalized seizures, such as childhood absence seizures. We are hopeful that these types of connectivity patterns could be used in the future to explain the phenotypic diversity seen in epilepsy and predict treatment outcomes. Funding: Portions of this project were funded by a CURE "Taking Flight Award" (JRT) and a Procter Scholar Award (JRT).