Generalized Q-space imaging reveals complex muscle architecture in vivo

Muscle tissue architecture is difficult to probe in terms of biophysical interactions due to large-scale hierarchical composition, ranging from the molecular to the organ scale. To address this need, we developed a generalized Q-space imaging approach employing fundamental concepts of diffusion-weighted magnetic resonance imaging (DW-MRI). This approach considers the morphology and distribution of magnetic field gradients, and results in the generation of a probability distribution function (PDF) for diffusivity at the voxel scale to describe multi-cellular diffusional boundaries within muscle. The principal fiber populations identified within these PDFs were associated by streamlines to create multi-voxel tracts that depict the orientation of muscle fibers at the tissue scale. Computational simulations of Q-space paradigms, where the B-value was defined as a model input describing diffusion distance and time and where Q-space was considered to have either spherical or hemispherical morphology, were performed. These simulations resulted in a quantitative description of crossing and tracking angles, thereby delineating the geometric precision of the method. We applied this imaging approach in an MRI experiment of the human tongue in vivo, using a shell distribution (single B-value) or segmented distribution (spectrum of B-values) with directionality defined by Q-space morphology, and confirmed the capacity to image complex muscular interfaces in vivo. We propose that this method may be applicable for defining normal and abnormal muscular architectural phenotypes.