Abstracts

Rationale: Recent trials of chronic EEG in humans showed that epilepsy is a cyclical disorder of the brain with rhythms at multiple time-scales: circadian, multi-day (multidien) or even seasonal. Continuously monitoring fluctuations in interictal epileptiform activity (IEA) helps determine seizure timing. We asked whether such rhythms would also exist in animal models of epilepsy. Methods: We analyzed chronic EEG data (>30 days) in two models of temporal lobe epilepsy in male rats (kainate or pilocarpine i.p.) and determined hourly spiking rate and seizure timing. A subset of animals were equipped with abdominal sensors for ECG, accelerometry and temperature. We used a wavelet analysis to resolve component fluctuations in spiking rate or other modalities, determine their periodicity and instantaneous phase. Using circular statistics, we calculated the average phase at which clinical seizures occurred. We report results as the phase-locking value, an index of phase clustering and statistics of the Rayleigh test. Results: We confirmed circadian rhythms of IEA, but also unraveled compound multidien rhythms of IEA with periodicity of about 3 and 5-7 days. In a figure, we show one representative example of raw spike rate, with an underlying six-day rhythm. Seizures tend to occur in the rising phase of this rhythm on average (PLV 0.7, p<0.001). In eight out of 10 animals, seizures clustered at preferential phases of underlying multidien rhythms with PLV ranging from 0.32 to 0.87 (p = or < 0.01, Rayleigh test). Fluctuations were not synchronous across animals, but were reflected in peripheral measurements (accelerometry, ECG, temperature), suggesting an endogenous generator. Conclusions: Our results show that circadian and multidien modulation of epileptic brain activity exist in rat TLE models, with striking resemblance to human data. In epilepsy, across species, unknown systemic factors modulate seizure timing in cyclical patterns. Funding: None