Fig. 1.

Functional and effective connectivity. A: Functional connectivity captures patterns of statistical dependence between regions of interest (ROIs) through the correlation of their fMRI time-series. Five ROIs, including subcortical nuclei (basal ganglia and thalamus) and primary sensory regions (dorsal and ventral somatosensory, primary auditory and primary visual cortex) showed increased functional connectivity in ASD compared to TD in our previous work (Cerliani et al. 2015). B: Effective connectivity models causal influences that one neural network exerts onto another. The figure depicts the connections we chose to model in our DCM analysis: bottom-up influence of subcortical nuclei on the primary sensory cortices; top-down influence of primary sensory regions on subcortical activity; and auto-connections, which in DCM model inhibitory connections of one neural system with itself (self-connections) and reflect its functional segregation – that is the sensitivity of a region to the influence of another modelled input (Zeidman et al., 2019a).