CIRCADIAN RHYTHMIC CRITICAL DYNAMICS ARE DISTURBED IN A MOUSE MODEL OF TEMPORAL LOBE EPILEPSY

Many temporal lobe epilepsy (TLE) patients still fail to achieve adequate seizure control, and endure severe cognitive and memory symptoms. Analyses of non-linear dynamics of neural activity, which stably persist close to criticality, have recently provided insight into information processing in the brain, which is optimal at the “edge-of-chaos” and appears disrupted in epilepsy. However, the timecourse of dynamical changes across epileptogenesis and chronic epilepsy, along with their relation to cognitive symptoms, remains unclear.
Using the intraamygdala kainic acid model of TLE, we perform continuous recordings of electrocorticography and electromyography in animals both before, during, and after the onset of seizures. In a separate cohort, we perform neuropixels recordings from the dorsal and ventral hippocampi at set timepoints throughout the development of epilepsy. To probe critical dynamics of the brain, we use a combination of irregular resampling auto-spectral analysis (IRASA), to characterise the ECoG aperiodic component, and the 0-1 chaos test.
At baseline, animals displayed a robust circadian periodicity to neural dynamics, with nighttime activity exhibiting a shallower 1/f exponent and increased chaoticity. This daily rhythm was lost in animals with chronic epilepsy, in whom the day:night ratio of the 1/f exponent was not significantly different from unity. Interestingly, this loss of circadian rhythm persisted beyond ictal activity by several days.
Overall our data indicate that neural dynamics are disrupted for an extended period of time after seizures, which is captured well by analyses of criticality, and could provide an explanation for post-ictal confusion and cognitive impairment.