Synergistic delay and strength tuning for logical operations in a chaotic neuron

An improved FitzHugh–Nagumo neuron model is proposed, incorporating electromagnetic induction via a flux-controlled memristor and dual autaptic feedback with tunable strengths and delays. This brain-inspired unit can reliably perform AND and OR logic operations under chaotic driving, serving as a tunable nonlinear computing element. Systematic numerical analysis reveals that the synergistic interplay between autaptic delays and strengths plays a key role in achieving optimal logical stochastic resonance. Specifically, the system achieves near-perfect success rates within specific parameter regions, where a compensatory effect enables delay asymmetry to balance synaptic strength mismatch. Furthermore, at lower coupling strengths, a multiple logical resonance phenomenon is observed, in which the success probability exhibits multiple peaks as a function of synaptic delay. This work not only elucidates the dynamical mechanism underlying reliable neuronal logic operation but also provides a co-design principle for tuning neuromorphic circuits, with potential implications for low-power, high-flexibility computing hardware.