Humans navigate in complex environments with abundant information. However, it is unclear how the human brain involves specific mechanisms to extract meaningful features from high-dimensional information to guide adaptive decision making. Here, we focused on investigating the causal role of the lateral frontopolar cortex (FPl), an area uniquely evolved in the human brain, in decomposing high-dimensional choice information. This was achieved via three experiments that collectively involved transcranial magnetic stimulation (TMS), resting-state functional magnetic resonance imaging (fMRI), task-based fMRI, and computational modelling. First, we found that disrupting FPl using TMS with a continuous theta-burst stimulation (cTBS) protocol impaired decision making with high-dimensional, but not low-dimensional, information. Second, we developed a computational model that arbitrates between a multi-feature decomposition mechanism and a simple heuristic. This model aided explaining that the FPl-TMS effect was attributed to diminished capabilities in multi-feature decomposition. Finally, fMRI data revealed stronger intrinsic FPl signals were related to greater tendency of employing multi-feature decomposition. Together, our results suggest a causal role of FPl in extracting decision-related features from high-dimensional information.
Law et al. (Sun,) studied this question.