Fig. 3From: On the statistical performance of Granger-causal connectivity estimatorsThe patterns (in this and all the figures of similar kind that follow) containing subplots with variables in columns representing the sources and the target structures in rows. Each subplot possesses boxplots of the distribution of GCT −log10(p value) for 1000 Monte Carlo simulations over different record lengths K = {100, 200, 500, 1000, 2000, 5000, 10000} along the x-axis of each subplot. Since \(\alpha =0.01\), values above 2 (dashed-line) indicate rejection of the null hypothesis of connectivity absence. Red crosses indicate p value distribution outliers, and those above dashed-line represent false positives (FPs) for nonexisting connections. (Color figure online)Back to article page