Abstracts

Rationale: How does a focal seizure start? Focal MTL seizures are traditionally thought to begin locally and then progressively recruit brain areas as they spread and generalize. Here, we show evidence that the start of a focal seizure may involve the contribution of multiple brain areas, supporting the idea that epilepsy seizures may be mediated by the malformed circulation of activity within a recurrently connected brain network.

Methods: Mice (n=8) with unilateral virally transduced CaMKII-ChR2 in the hippocampal CA1 were optogenetically stimulated (10Hz, 5ms pulses). Optogenetic stimulation simulates local hypersynchronous activity which induces focal to bilateral tonic-clonic seizures in the animals. We examined the network-wide EEG activity of the seizure induction process with electrodes implanted in the CA1 injection site, DG, entorhinal cortex, thalamus, and motor cortex ipsilateral to the viral transduction, as well as in the corresponding contralateral sites. The EEG of ictogenesis was examined using the Pulsogram described in our previous publication (Mueller et al., 2023). The Pulsogram is constructed with peri-event EEG snippets phase-locked to each optogenetic stimulation pulse.

Results: Optogenetic stimulation initially evoked activity that is phase-locked to the pulsogram (i.e., to each stimulation pulse). In time, seizures were induced and characterized by EEG activity unsynchronized to the pulsogram. Puslograms identified multiple stages in the optogenetic-evoked EEG activity leading up to the seizure (Fig. 1). Dynamical EEG changes before seizure onset in channels away from the ChR2 expression, particularly in the contralateral CA1/DG channels, indicate that they are not simply entrained to the optogenetic evoked response from ipsilateral CA1, but play a role in shaping the ictogenesis buildup. These dynamics feature the emergence of secondary discharges (2⁰DsC) 20-50ms from the pulse onset as a necessary activity pattern before seizure onset. 2⁰DsC activity remained phase-locked to the optogenetic pulse, but the latency of the peak 2⁰DsC undulated away and towards the pulse onset over time. Ictogenesis begins with increasing 2⁰DsC latency suggesting increasing engagement of network inhibition. Latency changed over the course of several seconds with shifts of up to 8ms (Fig. 2). Preceding the seizure onset, 2⁰DsC were often observed with decreasing latency, indicating a progressive increase in network excitation. Latency dynamics were characteristic of seizures from the same animal but can differ substantially between animals.

Conclusions: Our results point to a network-based ictogenesis process for MTL seizures involving the contralateral hippocampus. Changes in 2⁰DsC latency during the ictogenesis process characterize the dynamics of the excitatory/inhibitory balance involved in initiating a seizure. The exact process differed between animals and may have depended on the exact set of neurons optogenetically activated and the relative location of the EEG electrodes.

Funding: None