Purpose: Allodynia refers to pain evoked by physiologically innocuous stimuli. It is a disabling symptom of neuropathic pain following a lesion within the peripheral or central nervous system. In fact, two different pathophysiological mechanisms of cold allodynia (ie, hypersensitivity to innocuous cold) have been proposed. The peripheral sensitization of nociceptive neurons can produce cold allodynia, which can be induced experimentally by a topical application of menthol. An alternative mechanism involves reduced inhibition of central pain processing by innocuous cold stimuli. A model to induce the latter type of allodynia is the conduction block of peripheral A-fiber input. Patients and methods: In the presented study, functional MRI was used to analyze these two different experimental models of cold allodynia. In order to identify the underlying cerebral activation patterns of both mechanisms, the application of menthol and the induction of a mechanical A-fiber blockade were studied in healthy volunteers. Results: The block-induced cold allodynia caused significantly stronger activation of the medial polymodal pain processing pathway, including left medial thalamus, anterior cingulate cortex, and medial prefrontal cortex. In contrast, menthol-induced cold allodynia caused significantly stronger activity of the left lateral thalamus as well as the primary and secondary somatosensory cortices, key structures of the lateral discriminative pathway of pain processing. Mean pain intensity did not differ between both forms of cold allodynia. Conclusion: Experimental cold allodynia is mediated in different cerebral areas depending on the underlying pathophysiology. The activity pattern associated with block-induced allodynia confirms a fundamental integration between painful and non-painful temperature sensation, ie, the cold-induced inhibition of cold pain.