muscle activity and Non-negative Matrix Factorization, NMF

Main Forum
1

Dear experts, I have a question that has been troubling me for a long time: the extraction of muscle synergies has traditionally relied on the processing of surface electromyography (sEMG) signals combined with matrix decomposition algorithms. Can muscle activity simulated by Anybody be used for muscle synergy analysis?Additionally, after using Anybody simulation to calculate muscle forces, what can I observe? What insights can these force values provide me—for example, when an injury might occur? Or does it tell me how much force the muscle exerts per second to complete a movement?

2

Hi @shanhai,

sEMG and muscle activity are not all that different. But importantly, they are not the same also. sEMG measures the electrical activity in the muscles and nerves, and it is well known to relate to usage of muscles. Muscle activity from AnyBody is the relative mechanical load on the muscle (muscle force / muscle strength). Muscle activity can therefore also be a good indicator of muscle synergy. The analysis and the results are an outcome of the muscle recruitment criterion. You can read more about the muscle recruitment problem in our tutorials and some selected publications.

The muscle force from AnyBody tells you just the force in the muscle. The calculation is based on the input of motion and external force data. In other words, it is the muscle force required to balance the requirements of doing a motion and satisfying the external forces acting on the human model. With muscle activity, you get an additional insight into the relative load on the muscle. It is expressed as a ratio, where 1 means the muscle is the maximum force possible. There are also some additional outputs like metabolic energy and fatigue model.

AnyBody can't tell you exactly when an injury might occur. You can of course compare two simulations and then say that the one with lower muscle activity has a lower injury risk. Alternatively, you can look into some other references like the NIOSH recommended limits for spinal loads. In general, there are other factors such as repetition and recovery that come into play. You will have to extract data from AnyBody and post process it with all the relevant factors to understand the injury risk.

Best regards,
Dave