Fig. 7From: A deep generative adversarial network capturing complex spiral waves in disinhibited circuits of the cerebral cortexGenerative adversarial network trained to capture snapshots of spatial activity. A Architecture of the GAN model including both a generator and discriminator network. “conv.”: convolution operator. B Performance of the discriminator and generator networks. C Snapshots generated by the network after training. D Distribution of eigenvalues across 1000 snapshots generated by the network. E Center of mass across all snapshots. F Pairwise correlations decreased with spatial distance across the GAN snapshotsBack to article page