Skip to main content
. 2019 Jan 10;126(3):717–729. doi: 10.1152/japplphysiol.00714.2018

Fig. 6.

Fig. 6.

Root mean squared error (RMSE) for the best-performing linear regression models with k = [1,2, . . . 8] predictors. The best model for each k was considered the model with the lowest RMSE out of all possible models containing k predictors. Adding additional signals improves predictive performance, but with diminishing return.