Integrate-and-fire models with an almost periodic input function

TL;DR

This paper studies leaky integrate-and-fire models driven by almost periodic functions, analyzing firing maps and their properties, and establishing conditions for their well-definedness and periodicity, with implications for neural dynamics.

Contribution

It introduces new conditions for firing maps in LIF models with almost periodic inputs and explores their periodicity and continuity properties, extending previous results to discontinuous inputs.

Findings

Firing maps are well-defined and uniformly continuous under certain conditions.

The displacement map is uniformly almost periodic for Stepanov almost periodic inputs.

Displacement maps may be continuous but not almost periodic for μ-almost periodic inputs.

Abstract

We investigate leaky integrate-and-fire models (LIF models for short) driven by Stepanov and $\mu$-almost periodic functions. Special attention is paid to the properties of a firing map and its displacement, which give information about the spiking behaviour of the system under consideration. We provide conditions under which such maps are well-defined for every $t \in \mathbb R$ and are uniformly continuous. Moreover, we show that the LIF model with a Stepanov almost periodic input has a uniformly almost periodic displacement map. We also show that in the case of a $\mu$-almost periodic drive it may happen that the displacement map corresponding to the LIF model is uniformly continuous, but is not $\mu$-almost periodic (and thus cannot be Stepanov or uniformly almost periodic). By allowing discontinuous inputs, we generalize some results of previous papers, showing, for example, that the firing rate for the LIF model with a Stepanov almost periodic drive exists and is unique. This is a starting point for the investigation of the dynamics of almost-periodically driven integrate-and-fire systems. The work provides also some contributions to the theory of Stepanov- and $\mu$-almost periodic functions.