Epilepsy: Reorganization of the Brain and Interictal Activity

Epilepsy is undoubtedly a well-covered topic in computational neuroscience and beyond, yet important questions remain unsolved. At the BCF, we have combined a broad set of complementary expertises to expand epilepsy research towards new perspectives integrating anatomical, neurophysiological and computational tools. We mainly address focal mesial temporal lobe and neocortical epilepsies using experimental in vivo and in vitro mouse models, human data and tissue and simulations of spatially structured networks mimicking pathological conditions in the epileptic brain regions.

Fruitful collaboration of the labs of Carola A. Haas, Ulrich Egert, Jürgen Hennig, Arvind Kumar (now KTH Stockholm) and Andreas Schulze-Bonhage resulted in many publications providing new insights into the interaction of structural and functional changes in epilepsy, as well as clinical aspects of epilepsy diagnostics and treatment. Major obstacles for the treatment of epilepsy are the unpredictable early development of epilepsy in life and the seemingly random precipitation of seizures – in spite of extensive efforts to predict this.

We recently showed in a mouse model for the first time that the future progression of epilepsy can be foreseen early from structural changes in the hippocampus. Within less than one week, local reorganization of the hippo- campus indicates the severity of epilepsy that will develop subsequently. These indicators are accessible to magnetic resonance imaging, which could be applied in humans. Furthermore, in contrast to what degeneration in the hippocampus suggests, synaptic input to the dispersed dentate gyrus increased in later stages, indicating profound plasticity and changes of function. Analyses of local field potentials further show that even when detectable epileptiform activity is excluded, epilepsy leaves its fingerprint in the important theta rhythm. The frequency of this rhythm decreases drastically and its temporal coupling between the entorhinal cortex and the damaged regions change. In network simulations, we could show that even slight changes in connectivity between these regions could lead to such phase-shifts and networks incorporating the anatomical findings tend to produce spontaneous switches to epileptiform activity, suggesting a detrimental impact of ongoing network activity outside of epileptic events.

In detailed analyses of the spectrum of epileptiform events in our mice, however, we found that the prevalence of weak activity is anti-correlated with seizure susceptibility, i.e. high rates of mild activity indicate lower probability of seizures. Online analysis and intervention based on such sub-seizure activity could provide new approaches to the treatment of epilepsy. For example, we recently started to use optogenetic stimulation under closed-loop control with the aim to restore physiological theta activity and to reduce major seizures by evoking low-level events. In addition to the publications, our ndings led to several presentations at international conferences and laboratories.