Figure 3.

(a,d,g) Sum of the state variables as the bifurcation parameter c changes over time. (b,e,h) Largest eigenvalue σ₁ of the covariance matrix estimated using a moving window as the bifurcation parameter c changes over time. (c,f,i) Largest eigenvalue of the covariance matrix over the Euclidean norm of a vector consisting of all the eigenvalues $\frac{{\sigma }_1}{\sqrt{{\sigma }_1^2+\mathrm{...}+{\sigma }_n^2}}$ estimated using a moving window as the bifurcation parameter c changes over time.