Meta-analyses are useful to summarize the exponential amount of inconsistent and conflicting neuroimaging data. However, they are usually separately conducted for each different neuroimaging modality, preventing the multimodal integration of different imaging findings in a given neuropsychiatric disorder. Here, we describe an innovative method to meta-analytically combine the results of different imaging modalities, such as structural and functional paradigms. The method accounts for the presence of noise in the estimation of the p-values, and can be easily applied to any meta-analytical software. We hope that with this advanced imaging tool, researchers will be able to provide more complete multimodal pictures of the brain regions affected in different neuropsychiatric disorders.