Fig. 6.

Comparison of estimation methods. a Connections estimated using the conventional cross-correlation method, the jittering method, and our GLMCC method, in reference to the true connectivity of the synthetic data (used in Fig. 3). For the GLMCC and the true connectivity, the size of each square is proportional to the PSP amplitude, while for the first two methods, the estimated connections are represented in equal size, because they do not estimate the PSP. b Neuronal connections estimated from spike trains recorded from the hippocampus of a rat (used in Fig. 5d). Source data are provided as a Source Data file