Thread by @SlutzkyLab

Excited to announce a new paper out today in J NeuroEng and Rehab: rdcu.be/cQO3Z.
Myoelectric interface training enables targeted reduction in abnormal muscle co-activation. A collaboration with Jinsook Roh, Gang Seo, with help from Ameen Kishta & @emilymugler. 🧵

Abnormal muscle co-activation is a major cause of arm impairment after stroke. Our prior work (doi.org/10.1177%2F1545…) showed that gamified training with a myoelectric computer interface (MyoCI) could reduce this co-activation and reduce impairment in chronic stroke survivors.

The MyoCI training reduced co-activation between pairs of muscles. Here, we used muscle synergy analysis to assess whether MyoCI training was "breaking up" the abnormal synergies seen after stroke, changing their composition, or just affecting the pair of muscles trained.

We found no change in the number or composition of synergies.

Rather, the most co-activating muscles changed their within-synergy weights. That is, MyoCI training decoupled only the trained muscles, which seems to lead to reduced impairment. We developed a new disparity index to quantify the difference between muscle weights in a synergy.

The index (DI) increased after training the first (most abnormally co-activating) muscle pair.

This suggests that the CNS is capable of motor learning on a highly fractionated level, even after a stroke. It also confirms that MyoCI provides a targeted way to reduce abnormal co-activation. Now testing this in a RCT with a #wearable device: doi.org/10.1002/acn3.5…