1: \begin{abstract}

2: A mechanism for self-organization of the degree of connectivity in

3: model neural networks is studied. Network connectivity is regulated

4: locally on the basis of an order parameter of the global dynamics which

5: is estimated from an observable at the single synapse level.

6: This principle is studied in a two-dimensional neural network 

7: with randomly wired asymmetric weights. In this class of networks,  

8: network connectivity is closely related to

9: a phase transition between ordered and disordered dynamics.

10: A slow topology change is imposed on the network through a local

11: rewiring rule motivated by activity-dependent synaptic development:

12: Neighbor neurons whose activity is correlated, on average develop 

13: a new connection while uncorrelated neighbors tend to disconnect.

14: As a result, robust self-organization of the network towards the 

15: order disorder transition occurs. Convergence is independent of 

16: initial conditions, robust against thermal noise, and does not 

17: require fine tuning of parameters.

18: \medskip \\

19: PACS numbers:

20: 05.65.+b, %% self-organized systems

21: 64.60.Cn,  %% order-disorder transf, stat mech of model systems

22: 84.35.+i, %% Neural networks 

23: 87.18.Sn %% multicellular phenomena: nn

24: \medskip \\

25: Published in: Phys.\ Rev.\ E 67 (2003) 066118.  

26: \end{abstract}

27: