Principal Feature Analysis: A Multivariate Feature Selection Method for fMRI Data

Figure 1

Classification accuracies against the number of voxels for different feature selection methods. PFA: mean classification accuracy of PFA approach across 10 subjects; 𝑡-value: mean classification accuracy of 𝑡-statistics approach across 10 subjects; searchlight: mean classification accuracy of searchlight approach (spherical, two-voxel radius, comprising 33 voxels, and average absolute 𝑡-value) across 10 subjects. Error bars show the standard error of the mean accuracies across 10 subjects.