Abstracts

This abstract has been invited to present during the Basic Science Poster Highlights poster session

Rationale: Absence seizures (AS) are characterized by abrupt onsets of synchronous spike wave discharges. Assigning AS causes to localized brain regions or firing patterns such as deep layer cortical hyperexcitability or corticothalamic post inhibitory rebounding have yielded conflicting results over necessary and/or sufficient structures. This may however obscure roles of interactions between regions which are embedded in recurrent corticothalamic loops with interdependent firing patterns. We thus sought to uncover dynamics between multiple regions which may reconcile apparent discrepancies in AS mechanisms.

Methods: We used electrocorticogram (Ecog) recordings (16 sites) in Scn8a^+/- and Hcn2^EA/EA mouse models of absence and the machine learning algorithm: the sparse identification of nonlinear dynamics (SINDy) to identify governing equations for AS generation. Silicon probe recordings (1152 sites) in Scn8a^+/- mice were used to identify neuronal correlates. We then tested AS induction by these populations alone or in concert via optogenetic stimulation.

Results: Using SINDy, we find Ecog AS dynamics are well fit by equations for damped harmonic oscillators with quadratic and cubic interregional coupling, predicting 81 and 74% of variance in each model. Importantly, amplitudes of harmonic oscillations grow in response to specific drive frequencies, providing a mechanism for large amplitude oscillations observed in AS. We find oscillation amplitude growth at AS onset corresponds to shifts to a common frequency between regions. Silicon probe recordings reveal individual neurons exhibit similar firing frequency shifts within and between populations, most notably in the somatosensory cortex (SC), ventral basal (VB), thalamic reticular (TRN), and the posteromedial (PO) thalamic nuclei, the latter showing the earliest shift. Given these regions are interconnected somatosensory components, we asked if regional optogenetic manipulation in isolation or in concert could trigger ASs. We find stimulation in PO of inhibitory axons (predominantly TRN inputs) or excitatory SC axonal input triggers ASs. Furthermore, direct activation of PO or SC also triggered ASs. We find stimulation within SC and inhibitory axon stimulation to PO to be dependent upon the relative (frequency dependent) phases, showing AS induction is enhanced by multiregional interactions.

Conclusions: Here we find a simple model of AS generation derived from experimental data. We show neuronal firing rate frequency matching within and between regions coincides with AS generation. Both SC and PO stimulation can generate ASs and cortical input interacts with phasic inhibition in PO to modulate AS induction with a maximal effect when inputs are 3pi/4 radians out of phase. We predict similar phase dependence among other somatosensory components beyond those tested here directly. These results identify multiple targets for treatment, suggesting exclusive focus on a single site would unnecessarily hinder alternative treatment option development.

Funding: Work supported by (Epilepsy training grant #NS07288 and F32#NS123009) to JMH and (R01 #NS34774) to JRH.