Fig. 2.

First and second column represent effective connectivity without and with GSR – using the longitudinal datasets – within three resting state networks: (A) DMN, (B) SMR, and (C) SAL. The color of the squares indicates the MAP connection strength for the respective connection (all in linear scale). Thresholding for inference used a posterior probability of 90%. Values in smaller fonts represent connections with a posterior probability smaller than 90%. Right columns show the (natural) log Bayes factors of the alternative hypothesis of an effect of GSR versus the null hypothesis of no effect (see methods for a more detailed description). Anatomical Abbreviations: SMA ​= ​supplementary motor area, l/r SMR ​= ​left/right somatomotor region, ACC ​= ​anterior cingulate gyrus, l/r INS: left/right insula, l/r MFG: left/right middle frontal gyrus, PRC ​= ​precuneus, l/r IPC ​= ​inferior parietal cortex, mPFC ​= ​medial prefrontal cortex. *In this and the next figure, the Bayes factors are based on the first level (subject-specific) PEB estimates of connectivity – not on the second level (between subject) PEB estimates of group means.