Fig. 5

Plausible mechanisms supporting neural dynamics criticality in a representative stroke patient. Analysis of neural activity patterns, \(\langle S_1\rangle /N\) and \(\langle S_2\rangle\), of a stroke participant. In blue dashed line we show the corresponding control’s group average, while the shaded area corresponds to one standard deviation. For the patient, \(t_1\) and \(t_2\) correspond to 3 months and 12 months post-stroke. The critical transition appear as a trade-off between balanced excitability across brain regions (i.e., homeostatic plasticity) and overall brain excitation, as regulated by the \(r_2\) parameter. In the low \(r_2\) regime, we have \(r_2=0.36\), while in the high regime, \(r_2=0.7\). In low regimes of brain excitability (low \(r_2\)), unbalanced connectivity weights fail to support a critical state at both time-steps (a-c). Normalizing network excitability brings the dynamics closer to criticality at \(t_2\) but not at \(t_1\) (b-d). Conversely, increasing brain excitability (high \(r_2\)) drives the unbalanced network toward criticality, as evidenced by a peak in \(\langle S_2\rangle\) at \(t_2\) (e-g). The balanced network in the high \(r_2\) regime shows a sharper transition at \(t_2\) compared to the low \(r_2\) regime (f-h).