Fig. 6.

Post-hoc pairwise comparison between all four FPNs. a Group-level Bayesian predictions (i.e., Gaussian process regression) based on pairwise differences in BOLD between each pair of FPNs (i.e., Components (Comp) from the meta-analysis). b To explore if the Gaussian process regression resulted in sensible predictions, we also plotted the mean differences in BOLD between each pair of FPNs across all subjects for each cell in the task space. Data is pooled from both Experiment 1 and Experiment 3. In the bottom row, we show results for the same target measure we used in Experiment 3: difference in BOLD between each single FPN and the mean of the remaining three FPNs. Our results demonstrate the existence of a complex many-to-many mapping between cognitive tasks and FPNs as not a single task, but instead a functional profile across many tasks (i.e., the whole task space) is unique to the dFPN and vFPN