Figure 8.

Robustness of relationship between SMI anticorrelation and synaptic self-excitation in silico. A, Single-trial simulation of a three-module network with strong self-excitations. Upstream and downstream modules are as in Figure 5, A and C, respectively. One of two downstream modules (black) was set in the Up state injecting an additional external current. Its inhibitory synaptic input from an upstream module determined SDT whenever a presynaptic SUT occurs. Raster plots show spikes emitted by a subset of excitatory neurons in the two downstream modules (dark and light gray, respectively). Solid curves indicate firing rates of the two whole pools. B, SMIs for different configurations of in silico UFPs and network composition. UFPs were reconstructed using different fractions of SUT and SDT neurons from downstream modules (represented by different gray shadings). Two network configurations were tested in which the SDT module had strong (top left, diamond) and weak (top right, circle) synaptic self-excitation (parameters as in Fig. 5, C and D, respectively). Symbols indicate average across 10 simulations with randomly extracted synaptic connectivity. Vertical and horizontal (not visible) bars show SDs.