Hypotheses of pathophysiology in obsessive-compulsive disorder (OCD) have developed in parallel with advances in neuroimaging. Based on findings from early PET and SPECT studies evaluating cerebral blood flow and glucose metabolism, one theory proposed involvement of the orbitofrontal-striatum-thalamus loop (the "OCD loop"), which is relevant to the enforced learning and maintenance of OC symptoms. This OCD loop hypothesis has been revised in accordance with advances in neuroimaging techniques and the accumulation of findings. PET and SPECT molecular neuroimaging studies have provided the biological evidence to support the serotonin and dopamine hypotheses that were based on psychopharmacological findings. A symptom dimension hypothesis, based on symptomatology, has also been proposed. Neuroimaging has revealed the distinct neural basis of each symptom dimension. Neuroimaging has contributed to understanding the pathophysiology of OCD, and is expected to contribute to the development of treatment.