Evaluating the sensitivity to threat and affiliative reward (STAR) model in relation to the development of conduct problems and callous-unemotional traits across early adolescence

Yael Paz,Emily R Perkins,Olivier Colins,Samantha Perlstein,Nicholas J Wagner,Samuel W Hawes,Amy Byrd,Essi Viding,Rebecca Waller

DOI: https://doi.org/10.1111/jcpp.13976

Abstract:Background: The Sensitivity to Threat and Affiliative Reward (STAR) model proposes low threat sensitivity and low affiliation as risk factors for callous-unemotional (CU) traits. Preliminary evidence for the STAR model comes from work in early childhood. However, studies are needed that explore the STAR dimensions in late childhood and adolescence when severe conduct problems (CP) emerge. Moreover, it is unclear how variability across the full spectrum of threat sensitivity and affiliation gives rise to different forms of psychopathology beyond CU traits. Methods: The current study addressed these gaps using parent- and child-reported data from three waves and a sub-study of the Adolescent Brain Cognitive Development Study® of 11,878 youth (48% female; ages 9-12). Results: Consistent with the STAR model, low threat sensitivity and low affiliation were independently related to CU traits across informants and time. Moreover, there was significant interaction between the STAR dimensions, such that children with lower sensitivity to threat and lower affiliation had higher parent-reported CU traits. Unlike CU traits, children with higher threat sensitivity had higher parent-reported CP and anxiety. Finally, children with lower affiliation had higher parent-reported CP, anxiety, and depression. Results largely replicated across informants and time, and sensitivity analysis revealed similar findings in children with and without DSM-5 defined CP. Conclusions: Results support the STAR model hypotheses as they pertain to CU traits and delineate threat sensitivity and affiliation as independent transdiagnostic risk factors for different types of psychopathology. Future research is needed to develop fuller and more reliable and valid measures of affiliation and threat sensitivity across multiple assessment modalities.