Neurobiological models of obsessive-compulsive disorder (OCD) have emphasized the mediation of clinical symptoms by aberrant frontal-striatal circuits. A recent neurobiological model for normal human emotion perception has proposed distinct ventral and dorsal processing systems corresponding with differential levels of emotion perception. This construct appears also relevant for elucidating the pathophysiology of OCD, being an anxiety disorder. In the present paper, we present a comprehensive review of the neuroimaging literature in OCD, with the aim of critically appraising the current status of imaging-based knowledge on neurophysiological factors in the pathogenesis of OCD, along with the methodological possibilities and limitations of each imaging paradigm for OCD research. Moreover, we intend to relate these experimental findings to the frontal-striatal model and to the model of normal emotion perception. Results of imaging studies point to the need for adjustments in current models, with respect to the involvement of frontalstriatal circuits in OCD, as a ventral-dorsal dissociation may be distinguished. Moreover, data from pediatric OCD patient samples have commenced to shed light on developmental characteristics of OCD. Furthermore, longitudinal designs, in combination with pre-post treatment comparisons, are of great value for assessing the state-trait duality. Finally, the use of multimodal study designs holds great promise for the near future.