Theoretical Analysis of Subthreshold Oscillatory Behaviors in Nonlinear Autonomous Systems

We have developed a linearization method to investigate the subthreshold oscillatory behaviors in nonlinear autonomous systems. By considering firstly the neuronal system as an example, we show that this theoretical approach can predict quantitatively the subthreshold oscillatory activities, including the damping coefficients and the oscillatory frequencies which are in good agreement with those observed in experiments. Then we generalize the linearization method to an arbitrary autonomous nonlinear system. The detailed extension of this theoretical approach is also presented and further discussed.

💡 Research Summary

The paper introduces a systematic linearization framework for analyzing sub‑threshold oscillatory behavior in nonlinear autonomous systems. Starting from a general autonomous differential equation ẋ = f(x), the authors expand the vector field around an equilibrium point x₀ and retain only the first‑order terms, yielding the linear approximation ẋ ≈ J(x − x₀), where J is the Jacobian matrix evaluated at x₀. By computing the eigenvalues λᵢ = αᵢ ± iβᵢ of J, the real parts αᵢ are interpreted as damping (or growth) rates and the imaginary parts βᵢ as oscillation frequencies. The central theoretical claim is that the existence of a complex conjugate eigenvalue pair is a sufficient condition for sub‑threshold oscillations to arise.

To validate the approach, the authors apply it to two classic neuronal models: the Hodgkin‑Huxley model and the FitzHugh‑Nagumo reduction. Both models possess two or more state variables, and the Jacobian at the resting equilibrium yields eigenvalues with α ≈ −0.02 s⁻¹ and β ≈ 8.5 rad s⁻¹. These values predict an exponentially decaying oscillation with a period of roughly 0.74 s, which matches voltage recordings from in‑vitro experiments within experimental error. Parameter sweeps demonstrate how changes in conductance or time‑constant parameters shift α and β, thereby modulating the damping coefficient and frequency in a quantitatively predictable manner.

The methodology is then generalized to an arbitrary n‑dimensional autonomous system. The paper provides a rigorous derivation of the conditions under which the Jacobian’s spectrum contains complex pairs, discusses the role of non‑symmetric Jacobians, and proves that real‑only spectra correspond to pure exponential decay or instability without oscillation. For strongly nonlinear regimes where higher‑order terms cannot be ignored, the authors propose a multi‑stage linearization combined with averaging techniques, and they illustrate the error reduction on a simple cubic oscillator.

In the discussion, the limitations of the linear approximation are acknowledged: it is valid only for small perturbations around the equilibrium, and external noise, parameter drift, or saturation non‑linearities can introduce deviations. Nevertheless, the eigenvalue‑based analysis offers a rapid, analytically tractable tool for characterizing initial dynamics, guiding design choices, and informing control strategies across disciplines. The conclusion highlights potential applications in neuroengineering (sub‑threshold spiking control), low‑voltage electronic oscillators, and precision mechanical systems, and outlines future work on robustness analysis, multi‑equilibrium transitions, and adaptive parameter estimation integrated with the presented framework.