Purpose

Methods

Results

Conclusions

Adolescence is a period of dynamic change in nearly every domain. This developmental transition is characterized by dramatic alterations in hormone levels, remodeling of neural circuitry, and a number of behavioral changes including increased risk taking. In this dissertation, I present a multi-method program of research investigating the neurobiological contributors to adolescent risky decision-making in a sample of 14-18 year old adolescents (N=55). Taken together, findings suggest that the adolescent brain may be influenced by distinct hormones during specific aspects of decision making, and provide novel evidence for the possibility of a unique role for DHEA in cautious decision-making processes. Specifically, my research demonstrates that testosterone is related to trait measures in adolescents, replicating previous studies linking testosterone with sensation seeking and risk attitudes toward sexual behavior. During a novel laboratory paradigm, testosterone is associated with neural response preceding the selection of a risky choice, and DHEA with neural response preceding the selection of a cautious choice. Moreover, DHEA is associated with greater connectivity between the dorsolateral prefrontal cortex, a region implicated in regulatory processes, and regions of the brain including the ventral striatum, a region implicated in reward processing. In other words, adolescents with higher levels of DHEA exhibit greater frontostriatal connectivity when making decisions under conditions of risk. Finally, this research demonstrates that recruitment of neural circuitry during selection of cautious choice is related to individual differences in self-reported risky sexual behavior (i.e., lifetime condom use), providing support for the ecological validity of the laboratory task and relevance to behaviors important for public health.

A violation of expectations produces a prediction error, a learning signal by which organisms update current understanding and knowledge of expectations. The primary neural locus implicated in violations of expectations is the ventral striatum (VS), which undergoes significant development during adolescence, a period marked by sensitivity and responsivity to social feedback and reward. The VS demonstrates increased activation for reward and reward prediction errors in this age group compared to other age groups. Importantly, the VS is also recruited when adolescents take risks and are in the presence of their peers. While a few studies have examined social violations of expectations in adults, to our knowledge, none have examined social violations of expectations in adolescents. The goal of this dissertation was to implement novel, ecologically valid designs to determine whether neural responses to realistic social feedback from a friend was associated with a common risky behavior, substance use, in adolescents. We found that adolescents recruited the VS for positive social violations of expectations compared to negative violations of expectations, and were happiest when their social expectations were met (compared to positive or negative violations). We found that adolescents who recruited the VS more for violations of expectations reported greater substance use, while adolescents who recruited the VS less for violations of expectations reported lesser substance use, increased susceptibility to peer influence, and were more susceptible to peer influence in an experimental manipulation. Taken together, this body of work suggests adolescents who are more attuned to social cues from peers (and are likely to modify their behaviors in response) require smaller expectancy violations to experience reward; while those who engage in increased social risk taking require greater expectancy violations by comparison to experience reward. Implications of this work are discussed with regard to determining which adolescents are most likely to take greater social (substance use) risks, based on their VS response to social violations of expectations.