• social phobia;
  • shyness;
  • facial expressions of emotions;
  • ERP;
  • fMRI



Cross-sectional studies report biased reactivity to facial expressions among shy children, anxious adolescents, and adults with social anxiety disorder (SAD). It remains unknown whether cerebral reactivity to facial expressions can predict longitudinally the development of SAD in adolescence and characterize the degree of social anxiety among the general population of adolescents.


In a longitudinal study of 21 general population volunteers characterized for behavioral and genetic variables, N400 event-related potentials, and 3-Tesla fMRI activations in response to happy/neutral/angry expressions were acquired at age 8–9 and 14–15, respectively.


By stepwise regression, N400 amplitudes acquired at age 8–9 predicted the number of DSM-IV SAD symptoms at age 14–15, with the sole, significant (P = .018) contribution of the “anger” condition. Factorial ANOVA revealed increased (Voxel-Level P(FWE) range: .02–.0001) bilateral fMRI activations of several brain areas, including the amygdala, in response to facial expressions compared to a fixation cross. The number of symptoms of DSM-IV SAD was positively correlated with left amygdala response to angry (P(FWE) = .036) and neutral (P(FWE) = .025) facial expressions. Factorial ANOVA revealed that the 5-HTTLPR –S allele was associated with heightened left amygdala response to anger (P(FWE) = .05).


Cerebral reactivity to facial expressions, anger especially, measured at different developmental stages by different techniques is associated with adolescence SAD. The 5-HTTLPR genotype affects the neural processing of interpersonal affective stimuli during development.