Long-term recordings in patients and animal models reveal circadian and multidien (multiple days) rhythmicity in epilepsy. The underlying mechanisms of these cycles are not known. We are interested in predicting the onset of epileptic seizures over the scale of days. To this end, we performed chronic telemetric and single-unit recordings in epileptic rats (pilocarpine model). The primary focus of our investigation lies in the study of large network discharges known as interictal spikes (IS), closely associated with epileptic seizures.As previously shown, in our recordings the number of seizures and the number of IS follow a multidien rhythm. Our data suggest show that multidien modulation in IS may primarily be a consequence of the fluctuations in IS that occur during sleep. Moreover, we observe that IS during sleep and wakefulness exhibit different waveforms and have different firing patterns: while wakefulness IS regularly fire about every 5 seconds, sleep IS are more irregular. Finally, single-unit recordings in the hippocampus reveal opposite neuronal responses to IS during sleep and wakefulness: while neurons are excited by IS during sleep, they are inhibited by them during wakefulness. Neural assemblies show a similar response to neurons. Ongoing analysis are aimed at identifying how network synchronization in the form of cell assemblies’ formation and activity may predict epilepsy cycles.