Abstracts

Rationale: Absence epilepsy is a common childhood disorder featuring frequent cortical spike-wave seizures associated with behavioral arrest. The uniquely stereotyped EEG signature is assumed to reflect the global synchrony of neurons firing in unison among all cortical layers, disrupting complex patterns of synaptic integration. However, network activity of cortical neuronal populations in superficial layers has never been analyzed during seizures compared with the interictal state. Methods: Using the calcium indicator Gcamp6 with in vivo 2-photon calcium imaging and simultaneous electrocorticography, we examined the collective temporal activity profiles of layer 2/3 neurons and neuropil during spike-wave seizure activity in the visual cortex of 9 stargazer mice, a monogenic absence epilepsy model. The calcium signal is an indirect but well-established surrogate measure of electrical activity in cortical neurons. Results: With a meanSEM of 30.23.5 neurons and 13.51.2 patches of neuropil per dataset, we show that both neurons and neuropil in superficial layers of visual cortex are only sparsely and non-uniformly coupled to EEG synchronization. Unexpectedly, the majority of neurons (79.93.6%, meanSEM) were significantly less active in the ictal compared to the interictal state. Only a small proportion of cells were significantly more active during seizures (5.32.0 %). In contrast, 99.3%0.6% of neuropil patches had significantly reduced activity during the ictal state with only 1 patch of neuropil showing no significant change. When aligning the activity from all neurons and neuropil with the onset of seizures, the ictally suppressed neurons and neuropil show a significant but gradual reduction in activity several seconds before the first spike of the seizure. When aligned to the last seizure spike, activity returned to baseline within one second in the neuropil, while neurons lagged behind by 2-5 seconds. Furthermore, seizure-participation dynamics within this upper layer pool are not static, but fluctuate on a flexible (seconds to days) time scale within and across seizure episodes, indicating a loose and temporally transient community structure. Pairwise correlation analysis of calcium activity revealed a surprising lack of synchrony across layer 2/3 neurons and neuropil patches during seizures. The mean pair-wise correlation coefficient during seizures was 0.070.02 (meanSEM) for neurons and 0.480.02 for neuropil patches versus 0.140.02 and 0.760.04 respectively in the interictal state (p < 0.05, Wilcoxon matched-pairs signed rank test). This reduction in synchronous firing was independent of the overall reduction in activity (r2 = 0.0002). Conclusions: Our results demonstrate that the conventional view of cortical neuronal ensemble behavior during a generalized spike-wave seizure as a fixed hypersynchronous unit overlooks a distinct separation of superficial cortical networks from stereotyped pathophysiological thalamocortical oscillations. This complex portrait of dynamic neuronal participation has major implications for deciphering cortical network excitability during hyperactive EEG states. Funding: We acknowledge funding from: NINDS R21 NS088457 (JM and SS), NINDS K08 NS096029-01 (AM); and NINDS NS29709 (JLN).