poster
Multi-method analysis of three-dimensional techniques for the reconstruction of muscle architecture
The architectural properties of skeletal muscle provide an anatomical estimate of their functional capabilities. Gross dissection has long been the gold standard for exploring these variables, yet this technique has several inescapable drawbacks: being both irreversibly destructive and requiring the removal of a muscle from its original context for analysis thereby losing valuable spatial data. To counter these hurdles, several novel digital methodologies have emerged. Among these, diffusible iodine-based contrast-enhanced computed tomography (DiceCT) has been widely adopted to allow the mapping of individual fascicles in situ - enabling analysis of both their orientation and compression (i.e., tortuosity). Here, we present a cross-methodological analysis of four independent techniques for the reconstruction of muscle architectural variables: traditional gross dissection, manual segmentation of fascicles from DiceCT datastacks, and two distinct automated workflows (i.e., using Amira’s XFiber and the R package GoodFibes) using these same digital datasets. We applied these methods to a sample of eight mammals, ranging in body size from 0.1 to 10kg. First, we ran sensitivity analyses on our automated techniques to produce a single replicable workflow that could be applied across all taxa. Using this, we then calculated generally strong convergence in fascicle lengths between gross dissection, manual segmentation, and XFiber, while GoodFibes yielded consistently longer fascicles than would be expected. Both automated techniques also capture less fascicular tortuosity than manual fascicle segmentation. We conclude that automated methods represent a promising workflow to both enhance the speed of fascicular analysis and reduce user subjectivity and potential bias.