BACKGROUND: Specific phobia (SP) is a prevalent mental disorder for which exposure-based treatments are the most effective. Little is known about the intrinsic functional connectivity of SP and its modification by treatment. While previous studies were limited to a priori-defined brain regions, we used connectome-wide analyses to capture the full extent of altered functional connectivity. METHODS: We used functional magnetic resonance imaging in combination with hypothesis-free, data-driven functional connectivity multivariate pattern analysis (fc-MVPA) to identify differences between 72 individuals with SP and a nonphobic control group (CG) ( = 82). The SP group then received a one-session exposure treatment and was scanned again 9 weeks later on average. RESULTS: fc-MVPA identified the largest differences between the SP group and CG in sensorimotor regions, with lower connectivity to temporal nodes of the default network and anticorrelations with the frontoparietal network in the SP group compared with the CG. Stronger connectivity in the pretreatment compared with the posttreatment condition was evident in visual regions, while anticorrelations with the frontoparietal network were reduced. Post hoc comparisons showed that the connectivity strengths of the SP group after treatment between almost all identified nodes of both contrasts (SP vs. CG and pre vs. post) were comparable to those of the CG at baseline. CONCLUSIONS: Given the known functions of the identified networks, it is possible that the changes in connectivity measured after treatment indicate improved action control, enabled by more accurate prediction of the environment and stronger coupling of perceptual and action regions with higher-order control regions.