INTRODUCTION: Surprise (i.e., errors in outcome prediction) drives reinforcement learning. Animal studies point to the critical role of the centromedial subdivision (CMA) of the amygdala in signaling reinforcement-related surprise. Little is known as to the role of the CMA in this process in humans, as the problem of functional organization of the human amygdala was undertaken by only few studies. AIM(S): The goal of this study was to investigate the role of the human amygdala subdivisions in signalling surprise during reinforcement learning. METHOD(S): We used a Pavlovian conditioning task. The task was composed of two trial types: aversive and neutral, in which small amounts of aversive (0.4 M NaCl) and neutral (25 mM KCl and 2.5 mM NaHCO3) gustatory stimuli (liquids) were provided to participants. In the beginning of each trial, participants were presented with two visual cues: one associated with a high probability (on 70% of occasions) and the other with a low probability (on 30% of occasions) of obtaining gustatory stimulus. After a few seconds, one of visual cues disappeared and the subjects’ task was to indicate whether the remaining cue forerun liquid delivery. In an fMRI study, we compared the amygdala activity during: 1) reinforcement-related surprise (unexpected vs. expected delivery of aversive stimuli), and 2) surprise not related with the reinforcement (unexpected vs. expected delivery of neutral stimuli). RESULTS: We found the right CMA activation during reinforcement-related surprise, whereas surprise not related with the reinforcement did not activate any amygdala subdivision. CONCLUSIONS: The results showed selective involvement of the CMA in signalling reinforcement-related surprise in humans. Moreover, they prove that investigation of the amygdala at the level of distinct subdivisions using fMRI in humans should be valuable direction for future studies. This work was supported by a grant from the Polish National Science Centre based on decision number DEC-2014/15/B/HS6/03658. FINANCIAL SUPPORT: Polish National Science Centre grant No. DEC-2014/15/B/HS6/03658.