Developmental research posits that childhood family experiences shape the capacity for trust-based relationships. This study uses longitudinal data from 570 healthy adolescents (aged 14–25) to characterize normative developmental trajectories of trust behavior and individual differences. Previous cross-sectional findings from this cohort indicated age-related increases in trust. This study extends these findings by investigating the longitudinal development of trust, exploring the underlying cognitive mechanisms, and examining the impact of past family experiences on individual trust development. The study hypothesizes that childhood family experiences, particularly adversity, shape the intra-individual development of trust during adolescence. The researchers also investigated the relationship between individual differences in trust and the development of peer relationships, testing whether trust acts as a resilience factor, particularly for individuals with a history of adversity. Trust, defined as a willingness to be vulnerable to another's actions, involves a risk-benefit trade-off. Economic games are commonly used to quantify trust, enabling the measurement of individual differences in this trade-off. Computational modeling of these games helps identify core cognitive mechanisms, including social preferences, planning, and an understanding of others' mental states. A key parameter in the model, 'irritability,' captures retaliatory impulses due to perceived unfairness. Previous research using cross-sectional designs has yielded inconsistent findings on age-related changes in trust during adolescence. This study addresses this limitation using a longitudinal design to capture individual-level change and explore the moderating role of early family experiences on trust development.

The literature review highlights the longstanding hypothesis in developmental psychopathology that early childhood family experiences significantly influence the development of healthy social relationships later in life. The importance of trust as a key contributor to other-oriented behaviors, particularly during adolescence when stable relationships are established, is emphasized. Economic games, a common tool for measuring trust, allow quantification of the risk-benefit trade-off involved in trusting another person. These studies have identified neural correlates of trust and computational models that suggest key cognitive mechanisms underlying trust behavior, including social preference factors like risk aversion and theory of mind. However, the literature on age-related changes in trust during adolescence is inconsistent, with some studies reporting age-related increases while others find no change or even decreases. The inconsistent findings highlight the need for longitudinal studies to capture individual differences and developmental trajectories of trust behavior. The authors emphasize the importance of considering individual variability in adolescent social development, particularly concerning the impact of early social experiences that can cascade across developmental stages. Early family experiences, including parent-child interactions, are viewed as crucial for developing the ability to form and maintain healthy social relationships later in life.

The study utilized longitudinal data from the Neuroscience in Psychiatry Network (NSPN) project, including a multi-round trust game and self-reported measures of family adversity and peer relationships. The trust game involved 10 rounds where participants (investors) decided how much of an initial endowment to transfer to another player (trustee), whose repayment was determined by a computer algorithm. Three trust measures were derived: unconditional trust (first-round investment), mean trust (across all rounds), and reciprocity. Participants also completed questionnaires at three time points to assess the quality of their peer relationships using the Cambridge Friendship Questionnaire (CFQ). A composite measure of childhood family adversity was created through principal component analysis of self-report questionnaires, including the Alabama Parenting Questionnaire (APQ) and the Measure of Parenting Styles (MOPS). Linear mixed-effects models were employed to analyze the longitudinal development of trust, considering cross-sectional age, longitudinal development, and interactions with family adversity. A validated computational model of the trust game was used to probe the cognitive mechanisms underlying trust development. This model included parameters for social risk aversion, theory of mind, planning, guilt, irritability, irritation awareness, and stochasticity. Mixed-effects models were used to analyze the effects of age, development, and family adversity on each model parameter. Further analyses investigated the relationship between baseline trust (round 1 investment) and longitudinal development of peer relation quality, considering family adversity as a moderator.

The study found a significant increase in trust from adolescence to young adulthood, both cross-sectionally and longitudinally. Computational modeling indicated that this increase was primarily due to a decrease in social risk aversion. Family adversity significantly interacted with longitudinal development on trust behavior, attenuating the typical developmental gain in trust. Participants with higher family adversity showed a flatter trajectory of trust development compared to those with lower adversity. Computational modeling revealed that higher family adversity was associated with higher 'irritability' parameter estimates, indicating a greater propensity for retaliatory behavior in response to perceived unfairness. This type of retaliation was characterized by a 'mentalization breakdown,' lacking planning, theory of mind, and concern for fairness. A priori trust (round 1 investment) at baseline significantly predicted the quality of peer relationships over a 3-year period. This effect was particularly strong for adolescents who reported higher levels of family adversity, indicating that initial trust served as a resilience factor in the context of adverse childhood experiences. Importantly, there was no significant association found between non-social risk-taking and age, suggesting the age-related changes in risk aversion were specific to social contexts.

The findings support the hypothesis that trust develops significantly from adolescence to young adulthood, primarily driven by decreasing social risk aversion. The interaction between family adversity and longitudinal development of trust highlights the importance of early social experiences in shaping later social behavior. Family adversity appears to interfere with the typical developmental trajectory of trust, potentially through mechanisms related to heightened irritability and a tendency towards retaliatory behavior in social interactions. The finding that initial trust acts as a resilience factor in those with higher family adversity is particularly notable. This suggests that an optimistic starting point in trust, even in the face of past negative experiences, may support the formation of positive relationships in adolescence. The specificity of social risk aversion, as opposed to general risk aversion, suggests that this developmental change might reflect a specific adaptation in the adolescent social domain.

This longitudinal study demonstrates a developmental increase in trust from adolescence to young adulthood driven by decreasing social risk aversion. Childhood family adversity significantly attenuates this gain, potentially through increased irritability and impaired mentalization. Importantly, unconditional trust acts as a resilience factor, particularly for individuals with higher family adversity, promoting positive peer relationships. Future research should investigate the role of trust in different social contexts and explore interventions to promote trust development in adolescents facing adversity.

The study relied on retrospective self-reported measures of family adversity, which are subject to biases. The majority of the sample reported low levels of negative family experiences, which might limit the generalizability of the findings and explain the modest effect sizes. Future studies with larger samples and diverse adversity levels are needed.