Special Issue on Reliable Mechanisms for Translational Applications.

The reliability of neuroscienti fi c measurements is critical to the translation of neuroscienti fi c advances into clinical applications. In recent years, researchers have uncovered substantial limitations in the reliability of many neuroscience, social science, and psychology fi ndings. Shortcomings in the reliability of scienti fi c research have been prominently featured in both scienti fi c publications and the wider press (1,2). Unreliable scienti fi c fi ndings can be identi fi ed by further research, but this involves diverting efforts away from the truly promising research directions. Unreliable or incorrect research results can have long-lasting and pernicious effects on the state of knowledge (3). These fi ndings have subsequently galvanized efforts into understanding the source of unreliable fi ndings and into addressing them, resulting in important changes to sci-enti fi c procedures that aim to ensure improved replicability. This special issue of Biological Psychiatry: Cognitive Neuroscience and Neuroimaging presents the state of the art in a series of articles covering advances in our understanding of reliability relevant to mental health neuroscience research. The topics range from novel methodologies to a focus on analytical and speci fi c issues in particular settings. First, Botvinik-Nezer and Wager (4) focus on reproducibility of neuroimaging, but their insights and lessons apply much more broadly to the fi eld. They describe novel tools and practices to improve reproducibility, i.e., the ability to identify the same set of results using the same analysis methods on the same data. The fact that this is frequently not possible is a major reminder of the challenges ahead and points to the necessity of improving reporting standards, code and data sharing practices, management of computing environments, and analytic fl exibility. They point to a novel form of “ doing ” science, the Psychological Science Accelerator, whereby a global network of laboratories coordinates data collection for democratically selected studies (5), and a novel form of “ doing ” analyses involving diverse analytical approaches involving a multiverse of