Big bones mean big muscles: an MRI-based dataset of muscle-bone-body size relationships across 70 human muscles of the upper limb, trunk, and lower limb

Body sizes and shapes vary widely, even among healthy adults, resulting in diverse muscle sizes, strengths, and performance capacities. This study developed an artificially intelligent (AI) algorithm to segment individual muscles and bones from whole body MRI scans of 102 healthy adults (49 males, 53 females) aged 18-50 yr, generating three-dimensional (3-D) segmentations of 70 muscles and 13 bones spanning the upper limbs, trunk, and lower limbs. We quantified muscle volume, asymmetry, and fat fraction at whole body, regional, and individual-muscle levels, and examined how these properties correlate with body size and skeletal dimensions. Fat fraction and asymmetry varied across muscles and were generally similar between sexes; however, the distribution of muscle volume across the body differed between females and males. Across all predictors tested, total bone volume showed the strongest correlation with total muscle volume (<i>r</i>² = 0.85), followed by femur volume, height × mass, mass, height, and BMI. At the individual muscle level, the associated bone volume consistently explained more variance in muscle size than anthropometric predictors. Correlations between muscle volume and body-size parameters were significantly different between males and females, whereas bone-volume correlations showed no significant sex differences. These results suggest that skeletal dimensions-reflecting an individual's "frame size"-are stronger determinants of muscularity than body size metrics and explain the observed sex differences in muscle sizes. This work presents a comprehensive in vivo muscle-level dataset to date, introduces a novel framework for analyzing muscle-bone correlations, and provides reference data for applications from clinical diagnostics to athletic performance and musculoskeletal modeling.<b>NEW & NOTEWORTHY</b> This study presents the most comprehensive in vivo dataset of full-body muscle and bone volumes in healthy adults, showing that skeletal dimensions are the strongest predictors of muscularity, with height × mass emerging as the second-best predictor.