High-order synchronization in identical neurons with asymmetric pulse coupling
Abstract: The phenomenon of high-order ($p/q$) synchronization, induced by \textit{two different} frequencies in the system, is well-known and studied extensively in forced oscillators including neurons and to a lesser extent in coupled oscillators. Their frequencies are locked such that for every $p$ cycles of one oscillator there are $q$ cycles of the other. We demonstrate this phenomenon in a pair of coupled neurons having \textit{identical} frequencies but \textit{asymmetric} coupling. Specifically, we focus on an excitatory(E)-inhibitory(I) neuron pair where such an asymmetry is naturally present even with equal reciprocal synaptic strengths $(g)$ and inverse time constant $(\alpha)$. We thoroughly investigate the asymmetric coupling-induced $p/q$ frequency locking structure in $(g,\alpha)$ parameter space through simulations and analysis. Simulations display quasiperiodicity, devil staircase, a novel Farey arrangement of spike sequences, and presence of reducible and irreducible $p/q$ regions. We introduce an analytical method, based on event-driven maps, to determine the existence and stability of any spike sequence of the two neurons in a $p/q$ frequency-locked state. Specifically, this method successfully deals with non-smooth bifurcations and we could utilize it to obtain solutions for the case of identical E-I neuron pair under arbitrary coupling strength. In contrast to the so-called Arnold tongues, the $p/q$ regions obtained here are not structure-less. Instead they have their own internal bifurcation structure with varying levels of complexity. Intra-sequence and inter-sequence multistability, involving spike sequences of same $p/q$ state, are found. Additionally, multistability also arises by overlap of $p/q$ with $p'/q'$. The boundaries of both reducible and irreducible $p/q$ regions are defined by saddle node and non-smooth grazing bifurcations of various types.
Paper Prompts
Sign up for free to create and run prompts on this paper using GPT-5.
Top Community Prompts
Collections
Sign up for free to add this paper to one or more collections.