ABSTRACT Parkinson’s disease (PD) is a devastating, progressive neurodegenerative brain disease, with no known cure. The disease afflicts 10 million people worldwide - ~1.5 million in the U.S. alone, and ~50k-60k new cases are diagnosed annually. Risk factors or interventions are extremely hard to evaluate as we lack objective metrics of how PD affects the brain. The vast global availability of brain imaging has led to several promising metrics to gauge PD progression in the brain - structural changes in the basal ganglia and motor cortex, abnormalities in neural connectivity seen with diffusion MRI (dMRI), and disruptions of the brain’s functional synchrony across regions, seen with resting state functional MRI (fMRI). Despite these findings, factors that affect disease severity are difficult to discover, as most imaging studies of PD scan <100 patients. Most PD research is conducted in isolated cohorts from the US and Europe, limiting worldwide generalizability. Factors that affect PD progression are hard to verify, leading to a crisis of reproducibility. Responding to NIH’s call for more reproducible studies, here we launch ENIGMA’s Worldwide Parkinson’s Initiative. ENIGMA recently published the largest neuroimaging studies of schizophrenia, bipolar disorder, major depression, epilepsy, and autism spectrum disorder. With ENIGMA’s globally coordinated, highly powered consortium approach, we plan to overcome the crisis of small studies with poor power and reproducibility. Pooling anatomic, diffusion and resting state functional MRI metrics from 21 deeply assessed international cohorts - from the US, Brazil, Taiwan, New Zealand, the Netherlands, Italy, Switzerland, South Africa, China, and Russia - we ask: How does the illness affect the brain's structure, neural connectivity, and functional synchrony? What imaging biomarkers track disease progression and consistently predict clinical outcomes? Do genetic risk loci for PD help predict brain decline? What PD subtypes, or clusters, can imaging identify? Combining multimodal data from 2,307 patients and 1,264 controls, we will thoroughly evaluate predictors and brain biomarkers in PD. Our aims are to: (1) Evaluate and rank structural, diffusion, and resting state functional MRI biomarkers of PD worldwide; (2) Evaluate the added value of polygenic risk scores (PRS) in predicting PD brain biomarkers; (3) Predict future functional decline in PD with machine learning, multi-modal imaging and genomics. We will use genetic data and baseline clinical variables from PD patients and healthy controls across our cohorts to construct an ensemble of models to predict the annual rate of change in combined scores from the Movement Disorder Society—Unified Parkinson's Disease Rating Scale parts II and III. We will rank the best predictors of decline, and assess how robust they are internationally. By better modeling variance in patient outcomes, our multimodal predictive model will empower PD clinical trials by ran...