The neurobiological correlates of impaired insight in psychotic illness remain uncertain and may be confounded by factors such as illness progression and medication use. Our study consisted of two separate experiments. In the first experiment, we examined the association between measures of insight and regional brain volume in thirty-two patients with first-episode psychosis. In the second experiment, we looked at similar associations in thirty individuals with chronic schizophrenia. Detailed measures of symptom awareness and symptom attribution were obtained using the Scale to assess Unawareness of Mental Disorder. MRI scans were acquired and analysed using Statistical Non-Parametric Mapping for voxel-based analyses of grey matter maps. Regression models were used to assess the relationship between insight and grey matter volume in both the first-episode psychosis and the chronic schizophrenia experiments whilst controlling for potential confounds. In first-episode psychosis patients, symptom misattribution was associated with increased grey matter in the right and left caudate, right thalamus, left insula, putamen and cerebellum. In the chronic schizophrenia study, there were no significant associations between regional grey matter volume and measures of insight. These findings suggest that neuroplastic changes within subcortical and frontotemporal regions are associated with impaired insight in individuals during their first episode of psychosis.