Abstract
The frontal eye field (FEF) and the inferior frontal junction (IFJ) are two prefrontal structures crucially involved in attention, working memory, and cognitive control, and neuroimaging and non-invasive brain stimulation evidence suggest that their roles in these functions can be dissociated according to their specialization in the control of spatial versus non-spatial processing. We hypothesized that their structural connectivity fingerprints may underlie their specialized roles in top-down processes within the prefrontal cortex. In particular, we predicted predominant structural connectivity of the FEF with the dorsal visual pathway compared to the IFJ, and, vice versa, predominant structural connectivity of the IFJ with the ventral visual pathway compared to the FEF. An activation likelihood estimation fMRI meta-analysis was performed to accurately infer the localization of the FEF and IFJ in standard space. We included only saccadic functional localizers for the FEF sample, and covert attention, n-back, and task-switching paradigms for the IFJ sample (n = 21 studies for both samples, respectively). The resulting peak coordinates were employed as seeds for surface-based probabilistic tractography in each subject’s native space using 3T diffusion MRI data from 56 subjects and the multimodal parcellation released by the Human Connectome Project. Our results show that globally the connectivity likelihood of the bilateral FEF with regions in the dorsal visual pathway is higher than IFJ, whereas the connectivity likelihood of the bilateral IFJ with regions in the ventral visual pathway is higher than FEF. Critically, these results were replicated even when the seed-to-target average distance was accounted for, and taken together are consistent with the anatomical connectivity patterns of the macaque dorsolateral (i.e., FEF) and ventrolateral prefrontal cortex as inferred from invasive tract-tracing methods. The structural connectivity fingerprints of the FEF and IFJ may allow these regions to efficiently bias spatial versus non-spatial selection and other top-down processes.