Dynamical models reveal anatomically reliable attractor landscapes embedded in resting-state brain networks.
Imaging neuroscience (Cambridge, Mass.) January 1, 2025 Ruiqi Chen, Matthew Singh, Todd S Braver et al.
Resting-state brain activity may reflect a nonlinear dynamical system with multiple attractors rather than noise-driven fluctuations around a single stable state. Whole-brain dynamical systems models built from individual resting-state fMRI recordings, using the MINDy framework, revealed a diverse taxonomy of attractor landscapes including multiple equilibria and limit cycles. When projected into anatomical space, these attractors mapped onto a limited set of canonical resting-state networks, such as the default mode network and frontoparietal control network, which were reliable at the individual level. Creating convex combinations of models induced bifurcations that recapitulated the full spectrum of found dynamics, suggesting the resting brain traverses diverse dynamics generating distinct but anatomically overlapping attractor landscapes.