The research proposed in this application aims to understand risk and protective factors that promote continuation and desistance of problematic substance use (SU) and antisocial behavior (ASB) that began in adolescence. We propose a fourth wave of follow-up, approximately 18 years after initial recruitment, of an extremely affected adolescent sample as they transition into middle adulthood; this is a developmental period when we expect a portion of these individuals to decrease or desist problematic SU and associated high-risk behaviors, while others will persist with the most serious, destructive behaviors leading to devastatingly high rates of morbidity and mortality. The aims of this proposal are to: Aim 1: Identify risk factors that predict level and change (growth or decline) in SU and ASB from adolescence to middle adulthood. a. We hypothesize that early age of onset, male sex, child maltreatment, neurocognitive deficits, and personality traits (behavioral undercontrol/impulsivity) will predict higher levels and more growth in SU and ASB. b. We hypothesize that genetic vulnerability as indexed by polygenic risks scores (PRS) will predict faster growth in SU and ASB that persists through later adulthood. c. We will explore mechanistic relationships; e.g., we hypothesize that the relationship between PRS and level and change of SU and ASB will be partially mediated by behavioral undercontrol/impulsivity. Aim 2: Identify protective factors associated with level and change in SU and ASB. a. We hypothesize that adopting adult prosocial roles (education, employment, marriage, parenting) will be associated with lessened growth in SU and ASB. b. We hypothesize that treatment will be associated with greater desistance of SU and ASB than incarceration. c. We will explore moderators of genetic vulnerability, specifically whether prosocial roles and treatment attenuate the effect of PRS on level of SU and ASB. Aim 3: Determine the extent to which findings are specific to our highly selected sample of individuals with early-onset SU and ASB or generalize more broadly by conducting comparative and joint analyses of data from our high-risk sample with similar longitudinal data from our currently funded study of twins, an unselected community-based sample. a. We hypothesize that risk and protective factors will operate similarly across the two samples, although we will have a greater magnitude of risk factors in the high-risk sample. b. We will confirm this hypothesis in joint analyses of the high-risk and community twin samples.