Statistical plan for ReBlast pilot study. This figure depicts the primary statistical analyses proposed in this study: 1) a stepwise method to reduce the data into a meaningful set of biomarkers that distinguish the groups and then, assessing their relationship with GBEV scores (left panel); and 2) an unsupervised learning approach to reduce the data into meaningful clusters (representing groups of individuals with similar phenotypes) and then, evaluating their association with GBEV scores (right panel). Of note, the scatter plot in the left panel and the clusters in the right panel are hypothetical, not actual results. BEH, behavioral; BLO, blood; COG, cognitive; DMN, default mode network; FA, fractional anisotropy; fALFF, fractional amplitude of low-frequency fluctuations; GBEV, Generalized Blast Exposure Value; GBEVa, participant group with GBEV values above the median; GBEVb, participant group with GBEV values below the median; GFAP, glial fibrillary acidic protein; MCP, middle cerebellar peduncle; NEU, neuroimaging; NfL, neurofilament light chain; PHY, physical; rsFC, resting-state functional connectivity.