In this article we suggest a working definition for the concept of "neuropsychotherapy" encompassing three areas: the identification of mediators and targets of psychotherapeutic effects, the determination of new therapeutic routes using neurotechnology, and the design of psychotherapeutic interventions based on neuroscientific knowledge. We review neuroimaging studies of the psychotherapy of depression and discuss some of the methodological limitations inherent in functional Magnetic Resonance Imaging (fMRI) and common fallacies perpetrated in interpreting fMRI studies. Finally, we discuss some neuroethical issues related to this new and active field of research. In sum, we argue that neuroscience harbors great potential and add-on value for improving psychotherapeutic practice and research if applied properly.