An increasing number of quantitative neuroimaging studies have been published over the last decade, carrying an unprecedented potential to further our understanding of the pathogenesis of mental diseases. MRI can be repeated safely, providing avenues for measuring maturational brain changes. Sophisticated methods also allow exquisite resolution in identifying the locus and nature of alterations. With the growing amount of neuroimaging studies, we are facing the risk to only enumerate all the altered regions, without much advances in understanding the mechanisms at play, and much clinical implications for affected individuals. Here, I emphasize the need to use image processing methods to target specific developmental processes, for a better understanding of the nature and timing of the neurodevelopmental disruption. I also elaborate on the need to understand the heterogeneity in mental diseases, towards the delineation of homogenous subgroups that rely on similar biological pathways, and have comparable prognosis or sensitivity to treatment.