Observing touch is known to activate regions of the somatosensory cortex but the interpretation of this finding is controversial (e.g. does it reflect the simulated action of touching or the simulated reception of touch?). For most people, observing touch is not linked to reported experiences of feeling touch but in some people it is (mirror-touch synaesthetes). We conducted an fMRI study in which participants (mirror-touch synaesthetes, controls) watched movies of stimuli (face, dummy, object) being touched or approached. In addition we examined whether mirror touch synaesthesia is associated with local changes of grey and white matter volume in the brain using VBM (voxel-based morphometry). Both synaesthetes and controls activated the somatosensory system (primary and secondary somatosensory cortices, SI and SII) when viewing touch, and the same regions were activated (by a separate localiser) when feeling touch - i.e. there is a mirror system for touch. However, when comparing the two groups, we found evidence that SII seems to play a particular important role in mirror-touch synaesthesia: in synaesthetes, but not in controls, posterior SII was active for watching touch to a face (in addition to SI and posterior temporal lobe); activi ty in SII correlated with subjective intensity measures of mirror-touch synaesthesia (taken outside the scanner), and we observed an increase in grey matter volume within the SII of the synaesthetes' brains. In addition, the synaesthetes showed hypo-activity when watching touch to a dummy in posterior SII. We conclude that the secondary somatosensory cortex has a key role in this form of synaesthesia.