Skip to main content
Figure 1 | BMC Neuroscience

Figure 1

From: Chaos in heterogeneous neural networks: I. The critical transition point

Figure 1

(A) Example spectra of connectivity matrices (gray) with r>1 (top, indicated in blue and purple respectively) and r<1 (bottom). The average synaptic gain (red) does not give the correct boundary of the spectrum and would predict opposite behavior. The matrix G c ( i ) d ( j ) is indicated by the color plots (top), and activity of representative neurons from the two groups (bottom) of each example. (B) The activity of a readout unit during spontaneous activity, a FORCE learning epoch, and post learning for neurogenic and homogeneous subcritical networks. The neurogenic network quickly matches the target signal (gray) and robustly reproduces it. (C) The learnability of an ensemble of neurogenic networks as a function of the hyperexcitability and new neuron fraction coincides with contour lines of r (white).

Back to article page