Non-invasive assessment of human neurotransmitter function is a highly valuable tool in clinical research. Despite the current interest in task-based pharmacological MRI (phMRI) for the assessment of neural correlates of serotonin (5-HT) function, test–retest reliability of this technique has not yet been established. Using a placebo-controlled crossover design, we aimed to examine the repeatability of task-related phMRI with a single dose of oral citalopram in twelve healthy female subjects. Since we were interested in the drug's effect on neural correlates of 5-HT related cognitive processes, a sensorimotor and an emotional face processing paradigm were used. For both paradigms, we found no significant effects of the oral citalopram challenge on task-positive brain activity with whole-brain analysis. With ROI-based analysis, there was a small effect of the challenge related to emotional processing in the amygdala, but this effect could not be reproduced between sessions. We did however find reproducible effects of the challenge on task-negative BOLD-responses, particularly in the medial frontal cortex and paracingulate gyrus. In conclusion, our data shows that a single oral dose of citalopram does not reliably affect emotional processing and sensorimotor activity, but does influence task-negative processes in the frontal cortex. This latter finding validates previous studies indicating a role for 5-HT in suppression of the task-negative network during goal-directed behavior.