Abstract
It has been more than 30 years since magnetic resonance imaging (MRI) was applied to clinical research on psychiatric disorders. From case‐control studies with a few cases, MRI research for psychiatric disorders has progressed to multi‐site studies and meta‐analysis. Currently, the research trend is cross‐disease analysis and normative modeling based on harmonization techniques and non‐linear statistics, handling large‐scale data from multi sites. Under these considerations, machine learning classification and subtyping are relatively easy compared to other research field such as genetics. Since MRI‐based classification and subtypes become popular, a multi‐layer database is needed to utilize the classification. Due to the recent advances in MRI measurement and analysis techniques, it has become more difficult for novice researchers to tackle all analytical steps in clinical MRI research. This situation is also seen in other research fields like genetics. As multi‐layer data continue to progress in both quality and quantity, biological psychiatrists will be limited in the areas they can handle. However, they should organize multi‐disciplinary teams and show the right way to clinical psychiatry.
