Skip to main content
. 2014 Nov 26;34(48):15997–16008. doi: 10.1523/JNEUROSCI.1676-14.2014

Figure 3.

Figure 3.

MVPA analysis procedure. Top, Valence-labeled voxel patterns (from a single ROI) used to train a linear support vector machine (SVM). Middle, Learned voxel weights used to predict valence of unlabeled test data (voxel patterns not used for training). Bottom, Cross-validation schemes for testing for stimulus-specific and stimulus-independent emotion representations.