Women with catamenial epilepsy often experience increased seizure burden near the time of ovulation (periovulatory) or menstruation (perimenstrual). To date, a rodent model of chronic temporal lobe epilepsy (TLE) that exhibits similar endogenous fluctuations in seizures has not been identified. Here, we investigated whether seizure burden changes with the estrous cycle in the intrahippocampal kainic acid (IHKA) mouse model of TLE. Adult female IHKA mice and saline-injected controls were implanted with EEG electrodes in the ipsilateral hippocampus. At one and two months post-injection, 24/7 video-EEG recordings were collected and estrous cycle stage was assessed daily. Seizures were detected using a custom convolutional neural network machine learning process. Seizure burden was compared within each mouse between diestrus and combined proestrus and estrus days (pro/estrus) at two months post-injection. IHKA mice showed higher seizure burden on pro/estrus compared with diestrus, characterized by increased time in seizures and longer seizure duration. When all IHKA mice were included, no group differences were observed in seizure frequency or EEG power. However, increased baseline seizure burden on diestrus was correlated with larger cycle-associated differences, and when analyses were restricted to mice that showed the severe epilepsy typical of the IHKA model, increased seizure frequency on pro/estrus was also revealed. Controls showed no differences in EEG parameters with cycle stage. These results suggest that the stages of proestrus and estrus are associated with higher seizure burden in IHKA mice. The IHKA model may thus recapitulate at least some aspects of reproductive cycle-associated seizure clustering.
Keywords: Convolutional neural network; EEG; Epilepsy; Estrous cycle; Hippocampus; Kainic acid.
Copyright © 2020 Elsevier Inc. All rights reserved.