Extending Local Canonical Correlation Analysis to Handle General Linear Contrasts for fMRI Data
Figure 4
Detection performance of different data analysis methods showing the area under the ROC curve (AUR), integrated over , using pseudoreal data. The AURs for the contrast V-F of the visual paradigm (“VIS 0.55” and “VIS 0.65”) and the contrast E-C of the memory paradigm (“MEM 0.55” and “MEM 0.65”) are shown for the low noise case () and the high noise case (), respectively. The cCCA-RG achieves the greatest AUR values in all cases.