diagram shows that the 50% of cases (13 couples) presents the minimal value of efficiency in the DD hyper-brain networks, the 11.6% (3 couples) in the CC hyper-brain networks, and the 19.2% (5 couples) in the TT hyper-brain networks. The remaining 19.2% (5 couples) exhibits the lowest efficiency in mixed-strategies (CD, CT and DT) hyper-brain networks. For any frequency band, the DD connectivity pattern has the lower efficiency with respect to the other five networks in approximately the 50% of the couples. Similarly, modularity and divisibility are maximal for DD strategies in about the 75% and 62% of the couples, respectively. The statistical significance of these percentages has been tested using a χ2 test, which gives p-values ranging from 0.0001 to 0.002. These results indicate that hyper-brain networks corresponding to DD have longer paths between ROIs (lower global efficiency) and a small number of links between the two brains (high divisibility), this number being much lower than expected in a random graph with the same number of nodes and links (high modularity). Conversely, as shown in the bottom panels of Figure 3, the efficiency is maximal for TT (resp. CC) in the 30% (resp. 34%) of couples, while the modularity
