Subtle facial expressions may cause "core impressions" of other people, i.e. a feeling of like or dislike witch is affected by facial cues that is not explicitly and consciously recognized. In the present investigation, we were interested in how the neuropeptide oxytocin affects recognition of these subtle facial expressions. Participants received oxytocin or placebo, and viewed static and dynamic "hybrid" faces that showed a facial expression (happiness, anger, fear, sadness) only in the lowest spatial frequency (cycles/image), which was blended with the same face's neutral expression in the rest of the bandwidth (cycles/image). Two tasks was used as measures, a core impression task where participants was asked to rate "hybrid" images indicating friendliness, and a emotional labeling task, where participants was asked to choose the emotion they believed the "hybrid" images presented. We expected rates of friendliness to be higher after oxytocin administration versus placebo, especially for the hybrids containing low passed happy expressions. Further, we expected a higher "hit rate" on the emotional labeling task after oxytocin administration. Contrary to the hypothesis, ratings of friendliness did not increase after oxytocin administration. In the labeling task, oxytocin did not increase hit rate of emotional expressions in the lowest spatial frequency. Future implications are discussed.