Fig 1. Comparison of the distance to criticality between the hemisphere containing the epileptic focus and the contralateral hemisphere.
a Branching process approximation of activity propagation in the brain. Both excitatory and inhibitory connections are summarized to an effective excitation. Depending on the branching parameter m, dynamics are stable (m < 1), unstable (m > 1) or critical (m = 1). b, c Hypothetical scenarios of seizure generation. The epileptic focus might, in general, operate closer to criticality and consequently enter the supercritical regime in an unpredictable manner because of small fluctuations (scenario A). Alternatively, dynamics might systematically drift towards instability before seizure onset. If that drift is sufficiently slow, seizures might be predictable (scenario B). d The branching parameter of neuronal activity in MTL across recordings and patients shows no significant difference between ipsilateral and contralateral hemisphere (mixed effect ANOVA with patientID as random effect). e Estimated in ipsilateral and contralateral hemisphere for multiple recordings of two exemplary patients. While several patients showed a consistent difference between hemispheres across recordings, this difference is not predicted by the location of the epileptic focus (orange: reference recording, red: pre-seizure recording, see S2 Fig for all patients). f Comparison of the branching parameter in ipsilateral and contralateral hemisphere for different subregions of the MTL, (hippocampus (H), amygdala (A), parahippocampal cortex (PHC) and entorhinal cortex (EC)). The hippocampus shows a just significant difference between ipsilateral and contralateral hemisphere, with a smaller m in the ipsilateral hemisphere. The other subregions show no significant difference (p < 0.05/4 required after Bonferroni correction). Box plots show median and quartiles, while whiskers extend to the rest of the distribution, except for outliers (points beyond 1.5 x interquartile range).