Unveiling common psychological characteristics of proneness to aggression and general psychopathology in a large community youth cohort

Aggression is frequently observed in individuals with various psychiatric disorders, exceeding its typical association with externalizing disorders. Conversely, youth prone to aggression often experience compromised mental health, indicating a complex interplay. Prior research highlights the association between aggression and overall psychopathology, suggesting aggression might be a marker of broader mental health issues rather than a symptom of specific disorders. While transdiagnostic factors like impulsivity and emotion regulation are implicated, a comprehensive investigation incorporating diverse psychological characteristics in a nonclinical youth population is lacking. This study aims to address this gap by examining the relationship between general psychopathology, proneness to aggression, and multiple contributing factors in a large Hong Kong youth cohort, utilizing a data-driven approach to identify shared characteristics and understand the dynamic relationships between these variables. The study hypothesizes a robust association between general psychopathology and proneness to aggression, with overlapping features predicted across models for both outcomes. Furthermore, the study anticipates that impulsivity, cognitive control, and personality traits will emerge as shared factors linking both constructs.

Existing research demonstrates a strong link between aggression and various psychiatric conditions, including depression, bipolar disorder, and schizophrenia. Studies have shown that psychiatric diagnoses are risk factors for aggression, and youth prone to aggression exhibit increased internalizing symptoms like anxiety. Aggression's association with the overall number of lifetime psychiatric disorders suggests it’s a general marker of psychopathology rather than a symptom of specific disorders. Dimensional models highlight the strong association between internalizing symptoms and proneness to aggression. Transdiagnostic factors such as impulsivity, cognitive control, emotion regulation, and personality traits have been linked to both general psychopathology and proneness to aggression in youth. Neuroimaging studies reveal overlapping brain abnormalities in areas related to cognitive control, social functioning, and emotion processing. While these findings suggest shared underlying mechanisms, a systematic examination of these relationships, controlling for the interplay of variables, remains absent. The adolescent period presents a crucial developmental stage where both psychopathology and aggression are prevalent, making it a critical time to investigate these associations. Understanding the common factors could inform preventative and interventional strategies, reducing the risk of aggression and promoting better mental well-being.

This secondary data analysis utilized data from the Hong Kong Youth Epidemiological Study (HKYES) cohort (n=2184, aged 15-24). Proneness to aggression was assessed using the Buss-Perry Aggression Questionnaire (BPAQ), and general psychopathology was defined as the first principal component of various symptom scales (depression, anxiety, etc.). A total of 230 features (sociodemographics, cognitive functions, lifestyle, well-being, psychological characteristics) were included after data preprocessing. The dataset was split into discovery (75%) and holdout (25%) samples. LASSO regression models were trained on the discovery dataset to predict general psychopathology and proneness to aggression. The performance of these models was then evaluated on the holdout dataset using metrics such as RMSE, MAE, accuracy (correlation between empirical and predicted scores), and R². The procedure was repeated 100 times with random subsampling, and only features with 95% confidence intervals not crossing zero were considered significant. Cross-prediction analysis was conducted to assess the shared information between the two outcomes. To investigate the dynamic relationships between the features and outcomes, a Gaussian Graphical Model (GGM) was constructed using the overlapping significant features (after factor analysis). The GGM identified partial correlations between variables, visualized as a network graph, and stability of the network was assessed using bootstrapping.

A robust positive correlation was found between general psychopathology and proneness to aggression in both the discovery (r = 0.56) and holdout (r = 0.60) datasets. LASSO models showed good predictive accuracy for both outcomes (general psychopathology: r = 0.793; proneness to aggression: r = 0.676). Cross-prediction analyses demonstrated substantial shared information between the two outcomes: the model for proneness to aggression predicted general psychopathology (r = 0.717), and the model for general psychopathology predicted proneness to aggression (r = 0.606). The GGM revealed that impulsivity and isolation (loneliness) were directly associated with both general psychopathology and proneness to aggression. Bootstrapping analysis confirmed the stability of these associations. The impulsivity factor had high centrality in the network, suggesting its crucial role in both constructs.

The study's findings confirm a strong association between general psychopathology and proneness to aggression in a large community youth sample. The high predictive accuracy of cross-prediction models strongly suggests overlapping underlying mechanisms. The identification of impulsivity and isolation as key factors linking both constructs provides valuable insights for developing prevention and intervention strategies. Impulsivity, a prominent characteristic of adolescence, contributes to risky behavior and poor self-regulation, increasing vulnerability to both psychopathology and aggression. Social isolation also plays a crucial role, exacerbating aggression and mental health challenges. Targeting these factors could lead to more effective interventions compared to focusing solely on specific diagnoses or psychopathology dimensions. These findings highlight the importance of considering the interplay between general psychopathology and proneness to aggression in interventions.

This study provides robust evidence for the strong association between general psychopathology and proneness to aggression in youth. Impulsivity and isolation emerge as key shared psychological characteristics, offering valuable targets for preventive and interventional strategies. Future research should explore longitudinal designs to establish causal relationships and investigate the generalizability of these findings to clinical populations. The integration of both shared and unique characteristics of aggression and psychopathology will lead to more tailored and effective interventions.

While the study included a wide range of features, some risk factors might be uncaptured. The cross-sectional design limits causal inferences. The use of LASSO might select only a few highly correlated variables, potentially excluding others. Although data transformations were applied, some variables showed only approximate normality. The lack of an external validation sample limits generalizability, although a robust holdout sample was used for validation. Future studies should address these limitations to enhance the generalizability and further refine the understanding of these complex relationships.