Fourth, the most impacted circuits in our study GSK1210151A included the very regions that exhibit the greatest MET expression in the developing neocortex, including circuits that subserve processing of socially

relevant information. And lastly, measures of structural and functional circuitry correlated with symptom severity in the expected direction, although this correlation was driven by the fact that MET risk genotype was associated with both increased symptom severity and alterations in brain circuitry. These findings highlight a key principle that is consistent with the concept of endophenotypes ( Gottesman and Gould, 2003), whereby a functional risk allele predisposing to a disorder will have a larger impact on disorder-relevant phenotypes (i.e., relevant to the function of the gene) than the disorder itself. Thus, the present data suggest that taking into account MET risk genotype will serve as a sound strategy to stratify individuals with ASD and gain insight into the neurobiological bases of the functional heterogeneity that characterizes

ASD ( Figure 4). In our analyses, we first focused on functional activation patterns in response to the passive observation of emotional facial expressions in a large sample of 66 ASD and 78 TD subjects. The high expression of MET in ventral temporal cortex, including the amygdala and fusiform gyrus, prompted us to test whether the “C” risk allele might impact activity in these regions in response to socially CDK inhibitor relevant and affect-laden stimuli. While early studies of emotional face processing documented amygdala and fusiform hypoactivation in ASD (Baron-Cohen et al., 2000; Critchley et al., 2000; Schultz et al., 2000), later studies that better controlled for eye gaze (such as a fixation cross that directs gaze

at the eyes, similar to the one used in the present study) found either no differences or hyperactivation in these regions (Hadjikhani et al., 2004; Pierce et al., 2004; Dalton et al., 2005; Monk et al., 2010). Here, we found that MET risk genotype was associated with hyperactivation of below amygdala and striatum, as well as the relatively unexpected finding of reduced deactivation in temporal and midline neocortex. These latter areas comprise circuits that have the highest MET expression in developing humans and monkeys ( Judson et al., 2011a; Mukamel et al., 2011). In whole-brain analyses comparing TD and ASD groups, we also found evidence for reduced deactivation in temporal and DMN regions in ASD subjects, although there were no significant differences in the amygdala and regions of occipital fusiform gyrus corresponding to the fusiform face area. Overall, the MET risk group and ASD subjects (particularly the intermediate-risk group) showed less deactivation in multiple cortical and subcortical regions.