Figure 4

Individual ROIs and dynamic causal modeling. (A) Regions of Interest. The individual ROIs projected on a 3D rendering of the left hemisphere, with different colors for the Precuneus (PreC, in magenta), the left fusiform gyrus (FusG, orange), the left hippocampus/parahippocampus (HC/PHC, cyan) and the ventro-medial prefrontal cortex (vmPCF, green). The ROIs comprised 8 mm-radius spheres defined on the basis of a combination of functional and anatomical criteria (see Methods section). These ROIs were used for the analyses of effective connectivity (DCM: see panel B). The 3D rendering was produced using BrainNet⁵³. (B) Connectivity modulation by Rem-responses. Results of the analyses of effective connectivity using Dynamic Causal Modeling. The panel shows the model that best explained our data according to Bayesian Model Selection (see also Supplementary Figure S2, “Mod_2”). The winning model included modulatory influences of the Rem condition on the connectivity of the medial temporal cortex (HC/PHC) both with the precuneus and the ventro-medial prefrontal cortex. By contrast, the connections between the fusiform gyrus (FusG, the input node of the model) and the PreC/vmPFC were unaffected (gray arrows). Bayesian Parameter Averaging of the winning model showed that the Rem condition had opposite patterns of positive/negative modulation on the HC/PHC-PreC and HC/PHC-vmPFC connections (see red/blue arrows: posterior probability > 99%), as well as highlighting the effect of Rem of the FusG-HC/PHC connectivity. The images were made with Statistical Parametric Mapping software SPM12 (Wellcome Department of Imaging Neuroscience, University College London, UK; http://www.fil.ion.ucl.ac.uk/spm).