Abstracts

Rationale: The EEG characteristics of the postictal state have not been extensively investigated, apart from descriptions of diffuse attenuation following some generalized seizures. Using combined microelectrode and intracranial (subdural grid) EEG recordings, we investigated postictal electrophysiology at multiple time and spatial scales. Methods: Seven patients with complex partial or secondarily generalized seizures undergoing epilepsy surgery evaluation were simultaneously implanted with adjacent microelectrode arrays, limited to either frontal or temporal lobes. Single unit action potentials were extracted using a 300 Hz to 5 kHz bandpass filter, and standard voltage threshold. Burst firing was detected based on time points when the instantaneous population firing rate surpassed 3 standard deviations from the mean firing rate. Gaussian curves were fitted to each detected burst to assess burst duration and intensity across the neuronal population. Results: In 4 of 7 patients and 7 of 11 seizures, we detected a burst firing pattern that began within a few seconds of seizure termination, and continued for as long as three hours. The bursting consisted of disorganized neuronal firing limited to a window of less than 1 second, occurring initially at a rate of under 0.5 Hz and increasing to 1-2 Hz before ending abruptly. Extensive interictal data samples were reviewed to confirm that the pattern was present only after seizures. The simultaneous subdural EEG recording did not show any correlate with the bursts upon visual review, but filtering (90th order FIR bandpass) revealed a high gamma (80-150 Hz) correlate. The high gamma correlate did not appear in every channel, but was seen across a wide region within the coverage area, with small, variable temporal delays between channels. Conclusions: It appears that postictal neuronal bursting is commonly present following focal seizures with impaired awareness, and may contribute to postictal cognitive impairment. Interestingly, it was not detected in the two patients with brief focal motor seizures, suggesting that this pattern is less likely present after seizures that are brief and do not impair awareness.  We also speculate that postictal lateralizing periodic discharges, commonly seen after status epilepticus, may represent an extreme form of this pattern in which the neuronal bursting generates a strong synaptic output. The specific etiology of the burst firing remains unknown, but from our data it appears unlikely to be generated from the seizure onset zone. We conclude that the landscape of postictal electrographic changes is complex, and that further study in this area is needed. Funding: NIH NINDS R01 NS084142