By Jacopo Epifanio.

In Greek mythology, the chimera is a monster composed of several parts of different animals: it resembles a lion in the forepart, a goat in the middle, and a dragon behind.

In the context of dynamical systems, chimera states are intriguing phenomena of coexistence between synchronous and asynchronous motion. They have firstly been observed in networks of coupled identical oscillators, which segregate into two distinct groups: a high-coherence group and a low-coherence one. It is this very presence of partial synchronization in systems of identical oscillators to make chimeras good models for studying real-world complex systems, such as ecological networks, social interactions and, especially, the brain. Indeed, they are excellent frameworks for interesting brain conditions such as unihemispheric sleep in animals, but also neurological disorders like Parkinson’s, Alzheimer’s and epilepsy.

In particular, the synchronization-desynchronization pattern observed at the beginning of epileptic seizures strongly resembles the one observed at the collapse of a chimera state into a fully synchronized state. However, partial synchronization is important not only within networks but also across networks, therefore chimeras have also been investigated in multi-layer settings. It has been found that coupling with another layer can stabilize, promote or suppress partial synchronization within a layer, thereby providing a model for the propagation of seizures from its onset zone to the rest of the brain in focal epilepsy.

My project relies precisely on this background: by replacing the first layer of a chimera state network with those signals chimeras are supposed to represent, i.e., the phases of EEG recordings, we expect the signals from the focal area of the brain to synchronize better with the chimeras they drive. This could provide us with a new technique to identify the seizure onset zone (SOZ) in focal epilepsy patients. Indeed, a significant portion of such patients are resistant to anti-epileptic drugs, thus requiring SOZ resection to be seizure-free. For this reason, the localization of the SOZ is of crucial importance for them, and it is done by physicians who analyze hours-long electroencephalography (EEG) recordings. Nevertheless, SOZ localization is a time-consuming and challenging task, which often results in non-optimal surgical resections. Hence, the method we just introduced could facilitate and support the work of physicians in identifying focal signals.

Work: The entrainment power of the external environment on chimera states: a computational stochastic model. Jacopo Epifanio, Ralph Gregor Andrzejak ​​​​.

Poster presented at Cosyne 2024.