Associations with childhood aggressive behavior using multiple raters : co-occurrence, antisocial personality disorder prediction, and biomarkers

Problems of aggressive behavior affect as many as one in every six children and are associated with negative outcomes for not only the individual themselves, but also their family, friends, and community. Aggressive behavior includes a variety of different behaviors (e.g., yelling, hitting, bullying) and has been notoriously difficult to treat. In assessing aggression, researchers and clinicians have repeatedly been recommended to utilize reports from multiple informants (e.g., parents, teachers, the child him/herself) in order to obtain the most complete picture of the problem. In studying aggressive behavior, early research was initially focused on males only and severe outcomes, but now investigates gender similarities and differences and a broad range of behaviors related to aggression. While much has been learned from research thus far, new biological mechanisms and detailed phenotypic information are still important to continue clarifying the heterogenous nature of aggression and to improve ideas for personalized treatment. Thus, this thesis aimed to contribute to those efforts. Study I and II showed that aggressive behavior (as rated by all raters) often co-occurred with other externalizing behaviors and low prosocial behavior, and also co-occurred with internalizing problems but not as often. Patterns were similar across cohorts and genders, however, parents indicated more co-occurrence with internalizing problems and less co-occurrence with other externalizing behaviors than teachers did. Study III showed teacher and self ratings were able to predict ASPD, both in separate models and when both ratings were in the same model. Additionally, the direct aggression subtype (e.g., hitting, yelling) was able to predict ASPD well, for both genders. Furthermore, when the co-occurring influence of hyperactivity was removed from the aggression ratings (using a residual aggression variable with hyperactivity co-occurrence removed), aggressive behavior was still able to predict ASPD. Study IV showed the ketone body 3-hydroxybutyrate to be negatively associated with aggressive behavior in initial analyses. In more detailed modeling, nearly all raters of aggression showed the same trend with 3-hydroxybutyrate, including in fully adjusted models. In a model including both teacher and self ratings, 3-hydroxybutyrate was significantly associated with both aggressive behavior ratings. A replication dataset of young adult Dutch twins (N=960) showed support for the association found in FinnTwin12, however, the issue of whether there are gender differences of the association of 3-hydroxybutyrate with aggressive behavior remains to be clarified by future research. These findings help to clarify the co-occurrence of aggressive behavior with other behaviors across raters and countries, to show how common the co-occurrence is and that it should be taken into consideration when studying aggressive behavior, including from (epi)genetic or biological perspectives. Additionally, aggressive behavior, in particular direct aggression, can inform future ASPD risk, and obtaining behavior data from teachers and the child are of high importance. Furthermore, the new association of 3-hydroxybutyrate with aggressive behavior suggests new biological pathways to investigate to improve our understanding of aggressive behavior, including potential treatments. This thesis provides refinements to the aggressive behavior phenotype, new avenues for aggression biology investigations, and ideas for where to improve or personalize treatment options.