Measurement accuracy in mechanobiology: A unifying statistical framework for testing cellular forces.

Abstract

Mechanobiology is gaining traction as it reveals the fundamental role of physical forces and mechanical stress in biological function. Because physiologically relevant experiments are often inaccessible to direct physical probes, forces and stress at the microscale are increasingly estimated via multistep, image-based inverse problems such as Traction Force Microscopy. However, these measurements typically lack essential statistical descriptors such as error bars, CI, or <i>P</i>-values, limiting their reliability in experimental science. We present a single-step reconstruction framework that unifies a broad class of image-based inverse methods in mechanobiology under a general formulation that enables uncertainty quantification. This includes the visualization of high-dimensional credible regions, as well as an original formalization of abstract experimental questions into hypothesis tests. Overall, our work systematizes the development of new measurement techniques and contributes rigor and interpretability to image-based quantification in biophysical systems.