Skip to main content
. Author manuscript; available in PMC: 2020 Nov 15.
Published in final edited form as: Neuroimage. 2019 Aug 13;202:116081. doi: 10.1016/j.neuroimage.2019.116081

Figure 4.

Figure 4.

Dice coefficient (i.e. spatial overlap), sensitivity and specificity values of single-trial activation maps obtained with each of the analysis methods (GLM-IM, 1E-SPFM, ME-SPFM and MEICA+ME-SPFM) for each of the experimental conditions. Each box depicts the median and the mean (dashed line) across all trials. Its width is defined by the interquartile range (IQR) defined by the 1st quartile (Q1) and 3rd quartile (Q3) quartiles. The whiskers define the 1.5xIQR below Q1 and 1.5xIQR above Q3. TASK-LEVEL activation maps thresholded at FDR-corrected q ≤ 0.05 (TASKq05) and only including voxels with positive activation were used as reference maps, shown on the left for a representative dataset. The GLM-IM and 1E-SPFM activation maps were computed from the E02, OC and DN (i.e. MEICA and OC) preprocessed datasets, and GLM-IM maps were obtained at thresholds FDR-corrected q ≤ 0.05 (IMq05), as well as uncorrected p ≤ 0.05 (IMp05) and p ≤ 0.001 (IMp001).