Neural Effects Following Affect Recognition Training in Autism Spectrum Disorders

Background: One of the most consistent findings in the neuroscience of autism spectrum disorders (ASD) is hypoactivation of the fusiform gyrus (FG) during facial affect processing. In addition, reduced activation of the amygdala (AMG) has been associated with emotion perception alterations in ASD Objectives: In this study, we examined whether computer-aided basic facial affect recognition training using the «Frankfurt Test und Training of Facial Affect (FEFA)» is associated with increased activation of the FG and AMG as well as specific and generalized behavioral gains in high functioning ASD. Methods: Eight FEFA training sessions lasting one hour each are applied over a period of five to six weeks by experienced clinicians. BOLD-fMRI changes in the FG and AMG are registered pre-post applying an event-related facial emotion detection paradigm. Additionally, a battery of cognitive and clinical measures is assessed at baseline, post training and follow-up. Results: Preliminary findings in 13 trained versus 13 matched untrained participants aged 15 to 30 years (mean age: ~ 20.5 y, mean IQ: ~ 105) show enhanced BOLD-fMRI signal changes in the FG and AMG during implicit facial affect processing pre-post FEFA training. Behavioral facial affect recognition measures yield high effect size improvements for tasks being identical and similar to FEFA training material, but only minor effects for more general emotion recognition and social cognition tests as well as clinical measures. Conclusions: In contrast to a previous pilot study (Bölte et al., 2006), the present data indicate that circumscribed behavioral gains in basic facial affect recognition are indeed correlated with robust and expected neural activation changes in the FG and AMG. Nevertheless, on the behavioral level, of a lack of sufficient generalization of acquired affect processing skills was confirmed.