The present preliminary investigation has

The present preliminary investigation has limitations. First, sample size was relatively small for testing the effects of Otamixaban on behavior. Whereas the first studies linking 5-HTTLPR genotype to the amygdala found effects with sample sizes as modest as 23 (Fallgatter et al., 1999) and two independent groups of 14 participants each (Hariri et al., 2002), these results have replicated in a larger sample (N=92; Hariri et al., 2005) as well as by independent groups (Bertolino et al., 2005; Canli et al., 2005; Dannlowski et al., 2007). It is possible that imaging the effect of endophenotypes such as neural activation or connectivity may be more sensitive in detecting the effects of genotype than directly relating genotype to self-reported behavioral measures, as stated by Canli and Lesch (2007). Meta-analyses such as Murphy et al. (2013) and Flint and Munafò (2006), however, highlight the need for replication studies and collaboration, as no single study has been able to gather enough data to be sufficiently powered to reliably demonstrate effects. As such, preliminary studies with modest effect sizes serve the purpose of offering routes of investigation for future replication within and between research groups. Second, the sphere chosen as the seed was selected based on the result of a previous 5-HTTLPR/ASD study conducted by our group (Wiggins et al., 2014a). This is a limitation because artifacts (e.g. scanner, motion) that could have affected Wiggins et al. (2014a) may have affected the present investigation as well. Amygdala dropout in data from certain participants made us unable to conduct PPI analyses using a larger, independent seed (structural left amygdala). Nevertheless, the present study corrected for possible motion artifacts by censoring volumes with more than 1mm framewise displacement; a step that may help to avoid shared false positive results due to motion. Third, our sample differed in age, gender, and pubertal development among the four diagnosis-by-genotype subgroups. To control for this, we tested whether including age, gender, and puberty as a covariate affected our group analysis. Results continued to show that the ASD group with low expressing genotypes had significantly greater amygdala-sACC connectivity than ASD higher expressing genotypes and both TD genotype groups. Fourth, our sample’s ethnic background consisted, largely, of Caucasian individuals (see Supplementary materials: Table 1). This may be a limitation given differences in processing other-race faces (Golby et al., 2001). However, the results of a Caucasian-only analysis yielded the same interaction pattern around a 6mm sphere located in BA 25. Fifth, developmental transitions from childhood into adolescence have been shown to be important components in determining whether individuals with ASD display amygdala over-connectivity or under-connectivity in comparison to TD individuals (Nomi and Uddin, 2015). Analyses that consider the effects of genotype on connectivity across development will require a larger sample. Such an approach would provide a more complete understanding of neural changes occurring during this important developmental transition. Sixth, Hardy–Weinberg Equilibrium was not met in the TD sample. Whereas equilibrium was met in our ASD sample, our data were collected using convenience sampling and are not representative of a population with an expected distribution of 5-HTTLPR alleles. Participants included in this test were those who provided data for genotyping and successfully completed our fMRI scan. However, in analyses including all participants who provided data for genotyping (scanned and not-scanned), both groups meet Hardy–Weinberg Equilibrium. These data are preliminary and these limitations should be considered in replication attempts. Future research may wish to investigate participant interpretation of the emotions presented. Our construal of the participants’ reaction in response to happy faces (i.e. an aversion to happy faces) was not tested. Future studies may add to the present results by testing the relationship between distress to social contact and happy faces in ASD using various sources. This could be assessed through skin conductance response to test for physiological arousal and eye tracking to test for avoidance of gaze to the eyes. Moreover, polymorphisms do not operate alone in their effect on brain function. Researchers may benefit from examining the additive effect of genotype variants along the same metabolic pathway; a method used by Hernandez et al. (2016) in their investigation on the oxytocin receptor. This would give us a more informed picture of genetic influences on particular brain networks. Finally, a diffusion tensor imaging (DTI) study would inform our understanding of structural connectivity between the amygdala and the prefrontal cortex and its relationship to 5-HTTLPR genotype. Such research may highlight the relevance of the present finding to the differences in cortical connection size and number observed between ASD and TD individuals linked to serotonin (Casanova et al., 2002; Janušonis et al., 2004).