Skip to main content
. Author manuscript; available in PMC: 2022 Oct 6.
Published in final edited form as: Neuroimage. 2021 Nov 17;245:118715. doi: 10.1101/2021.04.05.438429

Figure 2. Bland-Altman plots comparing the B-spline and warped Bayesian Linear Regression (BLR) models, using Image-Derived Phenotypes (IDPs).

Figure 2

The models are compared according to the following model selection criteria: the Mean Standardized Log Loss (MSLL) (A) and the Bayesian Information Criteria (BIC) (B). The green colour indicates a better fit for the warped model compared to the B-spline model. We also plotted a zoomed-in view of the model fit for two of the IDPs. On images C and D, we show the difference in absolute values of the skewness and kurtosis between the B-spline and warped model. A more positive value indicates that the B-spline model had a higher skewness or kurtosis than the warped model.