Answers are grounded in this paper and cite their sources.
Introduction Chaotic intermittency, the irregular alternation of laminar (quiescent or regular-spiking) phases and chaotic bursts, is a generic route to complexity in single-neuron models. Its statistics are fixed by the reinjection probability density (RPD), i.e., by how trajectories return to the laminar region. This work applies the M-function reinjection formalism to a discrete-time and a continuous-time neuron model. Methods The chaotic Rulkov map is analyzed for Type-I intermittency under two laminar thresholds, c = 0.1 and c = 0.01, and over a scan of the laminar half-width. The continuous-time Izhikevich neuron model is reduced, through a Poincare section at the spiking threshold for the slow variable, to a one-dimensional return map whose second iterate is studied near a subcritical period-doubling bifurcation; bifurcation diagrams and Lyapunov exponents (with the saltation matrix for the reset) locate the intermittent regime as a function of the control parameter d. In both models the RPD, the laminar-length density and the mean laminar length are obtained from the numerical M-function and compared with the analytical predictions. Results For c = 0.1 the Rulkov RPD and laminar-length density follow the classical single-branch theory, whereas for c = 0.01 reinjection becomes discontinuous and requires a piecewise formulation; the scan over the laminar half-width shows that the continuous and discontinuous regimes are the two limits of a single, parameter-free law fixed by the geometry of the map. In the Izhikevich model the second iterate displays the local cubic tangency characteristic of Type-III intermittency; the M-function of the F^2 reinjection data is piecewise linear with two branches separated by one breakpoint at the accumulation locus, from which the RPD is derived as a piecewise power law. Numerical and analytical RPDs agree closely for several values of d near the bifurcation, and the mean laminar length saturates instead of diverging at the bifurcation. Discussion The laminar threshold and parameter drifts dictate the statistical distribution of quiescent and bursting phases, and the M-function methodology bridges discrete maps and continuous-time neuron models through Poincare section analysis.
No takes yet. Share an insight, caveat, or question.
Colman et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: