Impaired social brain network for processing dynamic facial expressions in autism spectrum disorders

Abstract

Background: Impairment of social interaction via facial expressions represents a core clinical feature of autism spectrum disorders (ASD). However, the neural correlates of this dysfunction remain unidentified. Because this dysfunction is manifested in real-life situations, we hypothesized that the observation of dynamic, compared with static, facial expressions would reveal abnormal brain functioning in individuals with ASD.We presented dynamic and static facial expressions of fear and happiness to individuals with high-functioning ASD and to age- and sex-matched typically developing controls and recorded their brain activities using functional magnetic resonance imaging (fMRI).

Result: Regional analysis revealed reduced activation of several brain regions in the ASD group compared with controls in response to dynamic versus static facial expressions, including the middle temporal gyrus (MTG), fusiform gyrus, amygdala, medial prefrontal cortex, and inferior frontal gyrus (IFG). Dynamic causal modeling analyses revealed that bi-directional effective connectivity involving the primary visual cortex-MTG-IFG circuit was enhanced in response to dynamic as compared with static facial expressions in the control group. Group comparisons revealed that all these modulatory effects were weaker in the ASD group than in the control group.

Conclusions: These results suggest that weak activity and connectivity of the social brain network underlie the impairment in social interaction involving dynamic facial expressions in individuals with ASD.

Figure 1

Statistical parametric maps showing significant brain activation for dynamic versus static facial expressions. The control (CON) and autism spectrum disorders (ASD) groups are shown in the left and right panels, respectively. The areas of activation are rendered on spatially normalized brains (upper) and overlaid on the normalized anatomical MRI of one of the participants at the coronal section showing amygdala activation (lower). The cross hairs in the lower panels are centered on the activation focus of the left amygdala in the control group (x -26, y -6, z -16; t = 4.65; cluster size = 5368 mm³). An extent threshold of p < .05, corrected for multiple comparisons, with a height threshold of p < .01 (uncorrected) were used. L = Left hemisphere; R = Right hemisphere.

Figure 2

Brain activation for the significant interaction between group and presentation condition. Weaker activation was found in the autism spectrum disorders (ASD) group than in the control (CON) group for dynamic (DY) versus static (ST) expressions. A. Statistical parametric maps rendered on spatially normalized brains. A height threshold of p < .01 (uncorrected) was used without extent threshold restriction for display purposes. L = Left hemisphere; R = Right hemisphere. B. Statistical parametric maps of representative brain regions overlaid on the normalized anatomical MRI of one of the participants in this study. From left to right, the activation of the middle temporal gyrus (MTG; x 52, y -62, z 0; t = 5.08), fusiform gyrus (FG; x 40, y -58, z -14; t = 3.00), amygdala (AMY; x -28, y -4, z -18; t = 2.89), medial prefrontal cortex (MPFC; x 8, y 66, z 20; t = 3.87), and inferior frontal gyrus (IFG; x 48, y 26, z 8; t = 3.06) is shown. The statistical thresholds are the same as above. C. Mean parameter estimates (± SE) of brain regions corresponding to the above overlaid MRIs. The data were extracted at the sites of peaks. FE = Fear; HA = Happiness.

Figure 3

Models and results of dynamic causal modeling (DCM) regarding the mirror neuron system (MNS). A. Analyzed brain regions rendered on the spatially normalized brain. V1 = Primary visual cortex; MTG = Middle temporal gyrus; IFG = Inferior frontal gyrus. B. Analyzed models. Thin arrows indicate intrinsic connections between brain regions. Bold arrows indicate the modulatory effects of dynamic presentation. C. Mean coupling parameters (± SE) for the control (CON) and autism spectrum disorders (ASD) groups. Statistical comparisons showed that all parameters were significantly weaker in the ASD than in the control group (t-test, p < .05).