The Development of Amygdala-Fusiform Structural Connectivity and Face Processing in High Functioning Autism

Background: Individuals with autism spectrum disorders (ASD) show impairments in face emotion processing, which may be related to differences in visual attention (less eye-contact than typically developing (TD) individuals (Klin et al. 2002)). Individual differences in gaze behavior have been associated with alterations in amygdala and fusiform gyrus (FG) in adults with ASD (Dalton et al., 2004), and decreased attention to the eyes predicts less ability to recognize fear and increased social anxiety (Corden et al., 2007). Given the strong bidirectional connections between these two neural regions, we assessed the integrity of these structural connections as they related to face expression processing. Moreover, we administered a laboratory-based intervention that encouraged eye-contact and evaluated its effect on face expression processing. Objectives: To examine the development of amygdala-fusiform structural connectivity as it relates to face emotion processing in ASD and to examine the effect of a laboratory-based intervention on face emotion processing. Methods: Reaction times (RT) to facial expressions (angry, fear, happy, neutral) were collected from children, adolescents, and adults with and without ASD. Two separate blocks were presented that either directed visual attention to the eye region (EYE) or allowed for natural viewing (NATURAL). In order to direct attention to the eye region, participants were instructed to identify the location of a cue placed in the eyes of the face stimulus. Additionally, the integrity of white matter tracks between the amygdala and fusiform gyrus was assessed using diffusion tensor imaging (DTI) with fractional anisotropy (FA) as the index. Results: Under natural viewing conditions [NATURAL], TD adults were slower when responding to fear faces relative to other expressions, consistent with our previous work (Hare, Tottenham et al., 2005). In contrast, the ASD group showed no evidence of this slowing to fear. There were no group differences in adolescents and children. Slower RTs to fear faces were associated with stronger connectivity between amygdala and FG. The eye-contact intervention [EYE] was effective in that, when making eye-contact, the ASD adults performed like the TD adults under natural viewing conditions (i.e., showed evidence of slowing to fear faces). FA values showed group differences in developmental change, where FA values were higher in TD than in ASD individuals during adolescence, while there were no group differences in adulthood. Conclusions: These findings suggest, as others have (Dalton, et al., 2004; Klin et al., 2002), that face processing impairments in ASD may be the result of decreased eye-contact. The current study extends these findings and shows that a laboratory-based eye-contact intervention may equate performance across TD and ASD adults. Moreover, individual differences in face expression processing may be the result of decreased connectivity between the amygdala and FG. The development in connectivity seemed delayed in the ASD group, where the group differences disappear by adulthood.