ABSTRACT The co-occurrence of substance use and suicidal thoughts and behaviors (STBs; suicidal ideation, suicide attempt) in adolescents is both common and deadly. Adolescents who report comorbid substance use and suicidality are at high risk of suicide. Unfortunately, clinicians’ ability to predict the emergence of both remains near-chance, with the current standard of care relying on non-biological measures, such as patient self-report and clinical impression. In this proposal, we leverage machine learning to better understand the predictive power of neural health on the emergence of substance use, suicidality, and their co-occurrence, using data from the Adolescent Brain Cognitive Development (ABCD) study (n=11,878). We will examine a set of structural brain and neurocognitive features of the sample at three-time points: Baseline (ages 9-10), Year 2 (ages 11-12), and Year 4 (ages 13- 14). Our primary aim is to identify structural brain and neurocognitive features associated with longitudinal changes in (1) substance use, (2) suicidality, and (3) comorbid substance use and suicidality. Very little research has examined neural factors that may precede suicidality, and no research has examined how these factors might predict comorbid substance use and suicidality in a large, longitudinal sample of preadolescents and adolescents. The training gained by completing this research with guidance from a diverse team of mentors will pave the way for career independence as an interdisciplinary clinical scientist with expertise in machine learning and neuroscience. Finally, the results of this study will suggest avenues for neuroscience-informed assessments and the development of an adjunctive cognitive remediation intervention targeting modifiable neurobiological risk factors associated with co-occurring substance use and suicidality for preadolescents and adolescents. This study has the potential for high public health impact and promise to help improve the quality of life for preadolescents and adolescents at high risk for suicide.

NARRATIVE This CTSA UL1 diversity supplement proposal aims to address two of the most vexing and interconnected public health problems facing adolescents today–substance use and suicidality. An iterative and mentor- engaged process will be used to better understand the predictive power of neural and neurobehavioral features on the emergence of substance use and suicidality by leveraging machine learning and applied to longitudinal data being collected as part of the landmark Adolescent Brain Cognitive Development study. This proposal is relevant to the part of NIH’s mission that pertains to yielding highly translatable research in the development of novel interventional approaches for reducing adverse outcomes associated with comorbid substance use and suicidality in adolescence.