Fig 1.
Group 80x1 analysis activation maps for the different contrasts.
Activation maps for the contrasts a) f > h, b) f > l, and c) f > o, thresholded at p<0.05 FWE at voxel level, are rendered on the lateral and medial surface of the SPM standard brain template. Areas which are driven by enhanced activation and decreased deactivation to faces are shown in red/yellow and blue/light blue, respectively.
Fig 2.
Group 80x1 analysis conjunction and per subject averaged beta estimates per condition and region.
The results of the conjunction analysis f > h ∩ f > l ∩ f > o, thresholded at p<0.05 FWE at voxel level, are shown on lateral and medial surface of the SPM standard brain template as well as transversal slices (z = -18, top, and z = 12, bottom). Brain areas which are driven by enhanced activation and decreased deactivation to faces are shown in red/yellow and blue/light blue, respectively. Extracted beta estimates (mean ± standard error of the mean, a.u. = arbitrary units) from right fusiform cortex (a), right amygdala (b), and the inferior part of the right pSTS/TPJ (c) which resulted from enhanced activation are shown in the upper panels. Correspondingly, responses from the superior part of the pSTS/TPJ (d), PC/pCC (e) and MPFC (f) which resulted from decreased deactivation are presented in the lower panels.
Fig 3.
FFA-1 and FFA-2 in different contrasts and single conditions in group 80x1.
Activation maps for f > h, f > l, f > o projected onto a transversal slice (z = -18) of the averaged normalized brain of the study participants (upper panels). Activation peaks are marked with white arrows. Cytoarchitectonic maps for FG-2 (blue areas) and FG-4 (green areas) are shown in the lower central panel. Extracted beta estimates (mean ± standard error of the mean, a.u. = arbitrary units) obtained from left and right FG-2 (blue bars) as well as FG-4 (green bars) are shown in the left and right lower panels, respectively. Significantly (p<0.001) smaller responses to houses, landscapes, and objects as compared to the face condition are denoted by asterisks.
Table 1.
Areas showing significantly stronger activation resp.
significantly weaker deactivation to faces than to all other conditions (conjunction analysis f>h and f>o and f>l) in group 80x1.
Table 2.
Average size (mean ± standard error of the mean) of individual activation peaks in mm3 across subjects in right and left FG2 and FG4 for each of the three control conditions.
Table 3.
Number of peaks across subjects in the 80x1 group.
Fig 4.
Right FFA peaks per subject across session in group 5x5.
In each row, the individual conjunctions f >h ∩ f >o ∩ f >s of sessions 1 to 5, masked by the right FG ROI are shown in an exemplary sagittal slice. The consistent peaks are marked by yellow arrows.
Table 4.
Percentage of overlapping voxels within the ROI across sessions in group 5x5.
Values are given in % +/- standard error of the mean and number of included subjects in brackets.
Table 5.
Peak reproducibility within the 5x5 group.
Table 6.
SVM classification accuracy in %, with classification based on smoothed (unsmoothed) data within the 5x5 group.
Table 7.
Simple hyperplane classification accuracy in %, with classification based on smoothed (unsmoothed) data.