1: \begin{abstract}

2: {\bf Networks in nature do not act in isolation but instead exchange

3: information, and depend on each other to function

4: properly \cite{little,rosato,Buldyrev2010}.  An incipient theory of

5: Networks of Networks have shown that connected random networks may

6: very easily result in abrupt failures

7: \cite{Buldyrev2010,raissa,raissa2,gao}. This theoretical finding

8: bares an intrinsic paradox~\cite{ref2,ref3}: If natural systems organize in

9: interconnected networks, how can they be so stable?  Here we provide a

10: solution to this conundrum, showing that the stability of a system of

11: networks relies on the relation between the internal structure of a

12: network and its pattern of connections to other networks.

13: Specifically, we demonstrate that if network inter-connections are

14: provided by hubs of the network and if there is a moderate degree of

15: convergence of inter-network connection the systems of network are

16: stable and robust to failure. We test this theoretical prediction in

17: two independent experiments of functional brain networks (in task- and

18: resting states) which show that brain networks are connected with a

19: topology that maximizes stability according to the theory.  }

20: \end{abstract}

21: