Comparing Muscle Spindle Afferent Models

Abstract

Proprioception, the internal sense of where your body parts are relative to each other, is essential for many, particularly bimanual, daily activities. Unfortunately, because modern prosthetics lacks this sense it is often difficult or impossible to perform hand-eye coordination tasks with them, and thus upper extremity prosthetics can become a nuisance or burden amputees frequently abandon at home (Biddiss & Chau, 2007). Towards creating a naturally-functional prosthetic able to provide proprioception, this project aims to compare computational models of muscle spindle afferents, using experimental data, in hopes that the best computational model could later be used to predict what voltages should be provided to which afferent nerves in a residual limb.Two muscle spindle models were compared using experimentally-measured afferent and muscle length data digitally extracted from figures in 10 articles. Data was from cats and humans. Both models implement the same formulas in either MATLAB or Python, take muscle length as an input, and can provide primary and/or secondary afferent output. Comparing the experimentally-measured and the predicted afferents, the more recent, Python model was found to provide more accurate afferent output.

The code used in this project may be accessed here:

https://gitlab.com/bitSlayer29090/musclespindlemodelsanddata