Fig. 4: Multifractal analysis primarily detects variations within neurons’ native circuits and is minimally influenced by adjacent network structures.

a Schematic of a network that comprises two interconnected subnetworks of spiking neurons. Subnetwork structures follow the same principles as in Fig. 2. Subnetworks receive distinct thalamic inputs. The connection probability between the subnetworks is 0.025. b, c Activation of one subnetwork by thalamic inputs reverberates to the other subnetwork through their connections. d, e Average multifractal spectrum of all excitatory cells in subnetwork 1 and subnetwork 2 for variations of thalamic input strength into the two subnetworks. The curves cluster based on input strength to the neurons' native subnetwork. f, g Quantified multifractal tokens of subnetwork 1 minimally vary with input strength to subnetwork 2 (f). Similarly, the multifractal spectrum of subnetwork 2 is maximally sensitive to input strength to subnetwork 2 and minimally influenced by input to subnetwork 1.