Figure 1.
An example set of four processed face images (faces not from this study).
The images are converted to greyscale and normalised with respect to visual area and average brightness. They are roughly matched in terms of spatial frequency and intensity contrast. The faces are all of the same sex, the expressions are similar and a vertically aligned full face image has been selected in each case. An individual set of four such faces was presented to each subject. One of the faces was of a person hated by that particular subject, the other three faces were known to the subject, but were of a neutral relationship, neither loved nor hated.
Figure 2.
Activations for the contrast All Faces>Baseline.
Reported probabilities at voxel level are corrected family-wise for multiple comparisons over the whole brain volume.
Figure 3.
Activation for the contrast Hated faces>Neutral faces.
The reported voxel-level probability has been corrected family-wise for multiple comparisons over the whole brain volume.
Figure 4.
Clusters of activation for the contrast Hated face>Neutral faces.
The statistical threshold was set at p≤0.05 at the cluster level, corrected for multiple comparisons, with an underlying voxel-level threshold of p≤0.00025, as displayed.
Figure 5.
Voxels covarying with hate questionnaire score.
A search volume of 5,225 voxels was defined using the t-statistic for the effect Hated face>Neutral faces with the statistical threshold set at p≤0.01 (uncorrected). Within this search volume voxels were identified where the effect Hated face>Neutral faces covaried with the hate questionnaire score. The voxel-level statistical threshold was set at p≤0.05, family-wise corrected for multiple comparisons within the search volume . The graphs in the right hand column plot the parameter estimate of Hated face>Neutral faces against the questionnaire score at each voxel.
Figure 6.
Clusters of activation for the contrast Neutral faces>Hated faces.
The statistical threshold was set at p≤0.05 at the cluster level, corrected for multiple comparisons, with an underlying voxel-level threshold of p≤0.00025, as displayed.