Nearly 75% of mental health conditions begin before the age of 25 years. The disability associated with mental health problems is well established, with depression as the leading cause worldwide. In youth and emerging adults, the top causes of disability in Organisation for Economic Co-operation and Development countries are primarily disorders of mental health, including those that typically present with a first episode of psychosis, such as schizophrenia and bipolar disorder. A clinically useful prediction algorithm that is able to inform functioning (a major determinant of disability) in early phases of these disorders might be tremendously useful.The study by Koutsouleris et al 1 in this issue of JAMA Psychiatry is a bold, brave, yet pragmatic attempt to assess the accuracy and utility of predictive models in youth and emerging adults based on clinical data, neuroimaging data, or their combination for both role functioning and social functioning. In any multicenter or multisite study, prospective rather than retrospective efforts go a long way toward design, quality, data aggregation, and general rigor, informing results and conclusions. The success of studies combining clinical, cognitive, and imaging data from the general population, such as the Philadelphia Neurodevelopmental Cohort and Adolescent Brain Cognitive Development studies, as well as those drawing from severe mental illness populations, such as Bipolar and Schizophrenia Network for Intermediate Phenotypes and Social Processes Initiative in Neurobiology of the Schizophrenia(s) studies, all share the common thread of prospective planning, preparation, and solid execution. The present study is a similar example of such efforts in a population at risk for or experiencing early signs of mental illness.The authors studied 116 participants at clinical high risk (CHR) (ie, those in the psychosis prodrome phase) and 120 participants with recent-onset depression (ROD) aged 15 to 40 years as well as 176 healthy control participants. 1 Their work represents a herculean effort of planning and execution, with data collection at 7 academic early recognition services across 5 European countries. Scanning and neurocognitive data acquisition occurred at months 0 and 9, while clinical assessments were conducted every 3 months over an 18-month period, collectively constituting a longitudinal observational study design. Efforts were made to demonstrate interrater reliability on key outcome measures of global functioning roles and social scales. Magnetic resonance imaging (MRI) harmonization was conducted using a minimal approach, but decisions were clearly defined. Magnetic resonance imaging pro-