Best Practices
Design Variables
Compared to the other model personalization tools, NCP is a much simpler optimization. The design variables are the synergy activations and synergy vector weights.
Rather than solving every time point independently for each synergy activation, the optimization adjusts spline points for the synergy activations. This formulation reduces the number of design variables and ensures that the synergy activations are continuous and smooth.
The synergy vectors for all muscles are design variables in NCP as well. To ensure the uniqueness of the solution, synergy vectors are normalized so that the max weight in each vector has a magnitude of 1.
Each muscle synergy has two components: the time-varying synergy activation that represents timing of muscle activations, and the time-invariant synergy vector that represents coordination of muscle activations. The synergy vector is composed of multiple synergy weights - One weight for each muscle in that synergy.
Choosing a Number of Synergies
At their core, muscle synergies are a mathematical decomposition of muscle activations. The more synergies you have, the better the synergies will represent the original data. There are diminishing returns as more synergies are added, however. For a dataset with 16 EMG signals per leg, such as the dataset used in Hammond et al., 2025, 5-6 synergies can typically represent more than 95% of the original data. Adding more synergies beyond that runs the risk of overfitting your experimental data which can make predicting new motions more difficult. To avoid overfitting our data, it is important to use a minimum set of synergies that represent the experimental data well.
The measure we use to quantify how well muscle synergies reproduce the original activations is called percentage of variance accounted for (%VAF, or just VAF), represented by the equation:
Where x and x' are the experimental EMG data and the reconstructed EMG data, respectively. For modeling with the NMSM Pipeline, we recommend that your synergies have 90-95% VAF.
You can obtain an initial guess for the number of muscle synergies to use by conducting a VAF analysis on your EMG data before running NCP. Note however, that NCP will almost always produce a lower VAF than a pure VAF analysis on EMG data because NCP is tracking both muscle activations and joint moments.
Evaluating NCP Results
The two main considerations for evaluating an NCP result are the muscle activation tracking quality and the joint moment tracking quality. The best NCP solutions will closely track both quantities. NCP results can be influenced by the quality of your prerequisite MTP run. For example, if the MTP run struggled with moment matching, the NCP run will likely also struggle to match the moments well.
Muscle activation tracking quality is automatically evaluated using two metrics:
VAF. A larger VAF indicates better muscle activation tracking quality. For modeling with the NMSM Pipeline, we generally recommend .
Worst individual root mean squared error (RMSE). This metric evaluates which muscle is tracked worst by the muscle synergies, where a lower value indicates better muscle activation tracking quality.
Joint moment tracking quality is evaluated primarily with RMSE, where a lower RMSE indicates better joint moment tracking quality.
Bilateral Symmetry
Unlike MTP, NCP allows personalization of multiple legs (or any other synergy set) simultaneously. If you are personalizing multiple synergy sets, you have the option to enforce bilateral symmetry. Bilateral symmetry enforces that synergy vectors must be symmetric between symmetry sets. Ie, the synergy vectors for right synergy 1 and left synergy 1 must be equal to each other. Using bilateral symmetry allows you to compare muscle synergies between legs by forcing the same muscles to be in matching synergies between legs. Enforcing bilateral symmetry generally slows down the NCP run and results in a lower VAF because the optimization is much more constrained. An example that uses bilateral symmetry is described in Hammond et al., 2025.