1: \begin{abstract}

2:   In a network of dynamical systems, concurrent synchronization is a

3:   regime where multiple groups of fully synchronized elements

4:   coexist. In the brain, concurrent synchronization may occur at

5:   several scales, with multiple ``rhythms'' interacting and functional

6:   assemblies combining neural oscillators of many different

7:   types. Mathematically, stable concurrent synchronization corresponds

8:   to convergence to a flow-invariant linear subspace of the global

9:   state space. We derive a general condition for such convergence to

10:   occur globally and exponentially. We also show that, under mild

11:   conditions, global convergence to a concurrently synchronized regime

12:   is preserved under basic system combinations such as negative

13:   feedback or hierarchies, so that stable concurrently synchronized

14:   aggregates of arbitrary size can be constructed. Robustnesss of

15:   stable concurrent synchronization to variations in individual

16:   dynamics is also quantified. Simple applications of these results to

17:   classical questions in systems neuroscience and robotics are

18:   discussed.

19: \end{abstract}

20: