Correspondence Between Effective Connections in the Stop-Signal Task and Microstructural Correlations

Response inhibition is considered to involve the fronto-basal ganglia circuit including the inferior frontal gyrus (IFG), pre-supplementary motor area (preSMA)/SMA, subthalamic nucleus (STN), and the motor cortices, but it remains unclear whether there exists a correspondence between the anatomical and effective connections between these regions. We defined regions of interest (ROI) based on the results of our previous study, and subsequently used diffusion tensor imaging (DTI), especially probabilistic fiber tractography, for the identification of white matter tracts of interest. Accordingly, we extracted the fractional anisotropy (FA) from the tracts of interest and applied data-driven hierarchical clustering to examine whether a specific pattern exists in white matter tracts. We found three clusters in the fronto-basal ganglia circuits: (1) the IFG-SMA and IFG- STN; (2) the dorsolateral prefrontal cortex (DLPFC)-caudate and caudate-STN and caudate-IFG; and (3) the SMA-STN. Further investigation with pairwise linear inter-tract FA correlations revealed that there were significant correlations between specific pairs: (1) the DLPFC-caudate and caudate-IFG; (2) the caudate-IFG and IFG-SMA; (3) the IFG-SMA and SMA-STN; (4) the IFG-SMA and caudate-SMA; (5) the IFG-SMA and IFG-STN; (6) the SMA-STN and caudate-STN; (7) the SMA-STN and IFG-STN; and (8) the caudate-STN and IFG-STN. The combination of results from hierarchical clustering and microstructural correlations showed that probabilistic tractography infers effective connectivity: i.e., the DLPFC-caudate-IFG-SMA-STN pathway. Our results revealed that specific clusters in the fronto-basal ganglia circuit and certain pairs of white matter tracts with significant correlations predict the effective pathways (hyper-direct and indirect pathways) in response inhibition.