Subliminal beauty engages the brain's valuation circuits

Neuroeconomic models propose that the anterior ventral striatum (aVS) and ventromedial prefrontal cortex (vmPFC) are key regions for computing subjective value (SV) signals that guide choice. However, the role of conscious awareness in this process remains debated. Here we examined whether SV can be automatically computed in these regions without conscious awareness. In an fMRI experiment, participants viewed faces that varied in attractiveness under three conditions: (i) suppressed from awareness via continuous flash suppression (CFS), (ii) clearly visible without suppression, or (iii) absent (background only) with CFS. Participants reported trial-wise facial identity (objective) and visibility (subjective) measures of facial awareness. In a post-fMRI session, they rated the attractiveness of each face as a measure of SV. When faces were seen, task performance (d) exceeded chance and responses were faster than in absent trials. When faces were unseen, performance was at chance but responses were slower than absent trials. Neurally, seen and unseen faces elicited greater neural signal than absent trials and showed similar neural patterns in the fusiform face area (FFA). Critically, neural signal in vmPFC correlated with SV for both seen and unseen faces, with similar neural patterns. In aVS, SV-related signal was only observed for unseen faces. Furthermore, mean face-related signal in FFA correlated with SV-related signals in aVS and vmPFC for unseen faces. These findings demonstrate that SV can be automatically computed in aVS and vmPFC without conscious awareness, suggesting a neural pathway by which subliminal information can influence value-based choice.