Abstracts

Rationale: The goal was to define the propagation dynamics during seizure activity relative to the default-state dynamics. Our hypothesis was that seizure dynamics is associated with a faster conductivity and steeper field gradient than default state is, as a result of hyper-synchrony predominant during seizures. We expected to find a consistent pattern of oscillation frequencies across subjects that best discriminate between seizure and default-state dynamics. Methods: We retrospectively analyzed ECoG data recorded from epilepsy patients (n>10), diagnosed with frequent complex partial seizures, candidates for resective surgeries. We compared the propagation dynamics between seizure activity and default-state activity. We divided the oscillations into three frequency bands: theta-alpha (1-25 Hz), high-beta (25-40 Hz) and gamma (40-80 Hz). We reconstructed the propagation dynamics of oscillations in each frequency band by mapping the relative phases in electrode space. From the slopes of propagation gradients between electrode pairs, we computed the average field-gradient (propagation speed), and compared those between default- and seizure-states from the same subject. For computing the phase we used (a) the Hilbert-transform, (b) the relative delay of oscillation peaks between electrodes, and (c) the cross-correlation of oscillations peaks between electrodes. The three different frequency bands and the three phase definitions with the two behavioral states provided 18 conditions. Results: The average propagation velocity was ~1 mm/ms during default-state, 6-9 times lower than expected (Nunez, 1995). More surprisingly, the seizure activity deviated from this dynamics in a frequency dependent fashion. While the theta-alpha component of seizures displayed an increasing phase asynchrony and slower propagation than during default-state, gamma oscillations displayed hyper-synchrony, faster and more anisotropic propagation across electrodes during seizures than default-state does. Conclusions: Our preliminary result suggests a frequency dependent propagation pattern during both default-state and seizure activity. Moreover, seizures affect the macroscopic propagation dynamics within cortical tissues in a frequency dependent fashion.