Fig. 2: Coordinated interactions of wave patterns in two interconnected cortical areas result in gamma burst and subspace-based interareal communication.

a Trajectories spanning 260 ms of the wave patterns in areas 1 (blue) and 2 (red) are shown within the same 2D space. Dashed circles mark synchronized events, and the green circle indicates the region of sampled neurons for On-Off states analysis in b–e. b Average firing rates of area 1 neurons (left panel) and area 2 neurons (right panel) sampled at the network center (green circle in a) during different states: S-Off (simultaneous-Off), S-On (simultaneous-On), 1-On (only area 1 in On states), 2-On (only area 2 in On states). Dots represent the average rates of individual random network realizations (n = 30). ***p < 10⁻¹³, two-sided paired t-test. Light and dark colors in each bar indicate the Off and On states, respectively, of the center region of each area. For example, the 1-On state is represented by a bar with the bottom half part dark blue/red and the top half part light blue/red. c Average phase locking value between MUA at the centers of areas 1 and 2 during S-Off (gray circle) and S-On states (black circle) within the frequency range of 30 to 120 Hz. d Average transfer entropy from area 1 to area 2 during S-Off (gray circle) and S-On states (black circle) at various time delays. e Average performance of predicting the spiking activity in the center of area 2 using the spiking activity in the center of area 1 through reduced-rank regression (circle) and the full model (triangle) during S-Off (gray) and S-On (black) states. All data in b–e are presented as averages across random network realizations (n = 30 realizations for b, c, and e; n = 60 realizations for d). Error bars denote ± 1 SEM. Source data can be found in Supplementary Data 1.