(A) Stimuli used to rate each of the emotional facial expressions on the dimensions of valence, intensity and attractiveness. (B) Mean ratings for each of the facial expressions according to emotion and pupil size (64 to 180% left to right) (1) Positive/negative rating on a 0–100 absolute scale. Small pupils in expressions of sadness are rated as significantly more negative (asterisks represent repeated-measures ANOVA F(3, 90) = 4.340, P = 0.007, Contrasts, 64 vs 100% F(1, 30) = 5.481, P = 0.026, 64 vs 180% F(1, 30) = 9.311, P = 0.005, 80 vs 180% F(1, 30) = 5.377, P = 0.027) than those with larger pupils; (2) Emotional intensity rating on a 0–100 scale. Sad faces with small pupils are rated as significantly more intense (asterisks represent repeated-measures ANOVA contrast 64 vs 180% F(1, 30) = 4.575, P = 0.041) than those with larger pupils; (3) Attractiveness rating on a 0–100 scale. Pupil size had no effect on attractiveness ratings when comparing combined male and female responses.

(A) Brain regions showing a significant correlation with linearly decreasing pupil size in the context of expressions of sadness. All regions shown are significant at the P ≤ 0.001 uncorrected. (B) Percentage signal change for each region shown above plotted against emotional expression. Decreasing pupil size effects a significantly greater percentage signal change to sad than other facial expressions in all regions shown.

Subject's own mean pupillary response to observed pupil size in emotional expressions. (A) Mean pupil response across all subjects to a 500 ms stimulus presentation, illustrating the pupillary light response beginning approximately 200 ms after stimulus onset and peaking 200 ms after stimulus offset, followed by a gradual return to baseline. (B) Subject's mean pupil size in the 500 ms window following maximal pupillary constriction for neutral, happy, sad and angry facial expressions. Pupil size is plotted in response to observed pupil areas 64, 80, 100 and 180% of the original image (from left to right). Observers own pupil size was significantly smaller when viewing sad faces with small pupils than when viewing those with larger pupils [repeated-measures ANOVA, main effect pupil size, F(3, 24) = 5.04, P = 0.008*]. Post hoc contrasts comparing 64% (P = 0.002), 80% (P = 0.005) and 100% (P = 0.049) pupil areas with 180% images were also significant. There was no main effect of observed pupil size for the other emotional expressions [repeated-measures ANOVA, F(3, 24) = 0.746 Neutral, P = 0.525, F(3, 24) = 0.568, P = 0.641 Happy, F(3, 24) = 0.475, P = 0.703 Angry]. The horizontal line indicates subjects mean pupil size across all trials.

(A) Mid-brain regions showing a significant correlation with linearly increasing pupil size in the context of expressions of sadness. Both regions shown are significant at P ≤ 0.001 uncorrected. All activations are shown overlaid on T1 canonical brain slices. (B) Percentage signal change for the right and left mid-brain regions plotted against emotional expression. Increasing pupil size effects a significantly greater percentage signal change in sad facial expressions than the other emotional expressions in both mid-brain regions shown.

(A) Mid-brain region showing a significant correlation between BOLD response to linearly increasing pupil size in sad expressions and individual's sensitivity to pupillary contagion. Coordinates demonstrate that this area lies within and between the mid-brain regions shown in Figure 4A. (B) Correlation between activity in the peak voxel within this cluster and subjects’ individual indices of sensitivity to pupillary contagion.