Predicting inattention and hyperactivity-impulsivity trajectories using random forest classification in the Adolescent Brain Cognitive Development Study
Abstract
Introduction: Despite well-established genetic and environmental risk factors for attention deficit–hyperactivity disorder (ADHD), it remains unclear why children show different developmental trajectories from mild to severe levels. This study aimed to investigate distinct developmental trajectories of inattention and hyperactivity-impulsivity scores and identify predictors of these trajectories from childhood to adolescence in a community-based sample. Methods: Data were drawn from the longitudinal Adolescent Brain and Cognitive Development dataset (n = 11,875) with yearly measures from baseline to wave 3 (ages 9-10 to 12-13). Growth mixture models were applied to the inattention and hyperactivity-impulsivity items of the ADHD subscale of the Child Behavior Checklist to identify trajectories. Random forest classification models were then trained to predict these trajectories using familial, demographic, behavioral and genetic variables, previously linked to ADHD. Logistic regression analyses were conducted to determine the direction of associations. Results: Three trajectories were identified for inattention (non-affected, moderate, and persistent-high), and four for hyperactivity-impulsivity (non-affected, worsening, improving, and persistent-high). Key predictors for both domains included male sex and higher externalizing problems, family conflict, and polygenic risk scores of ADHD, distinguishing affected from non-affected children. Additional distinguishing factors were less secondary caregiver acceptance (inattention) and lower family income (hyperactivity-impulsivity). Further, a persistent-high inattention trajectory was associated with negative school experiences and more internalizing problems, while worsening and persistent-high hyperactivity-impulsivity trajectories were linked to slower pubertal development. Conclusion: This study supports the distinction between inattention and hyperactivity-impulsivity trajectories, which may be shaped by domain-specific factors. Nevertheless, at its core, both ADHD domains appear to be associated with shared factors, including biological and environmental, in the distinction between non-affected youth and those with ADHD symptomatology.
Related
- The paradox of high IQ in ADHD: dissociable effects on inattention and impulsivity - insights from behavior, cognition & brain in 5138 children.
- Distinct affective and impulsivity-related profiles are associated with facets of state boredom in adults with attention-deficit hyperactivity disorder
- Impulsivity is the principal correlate of problematic internet use in young adults with attention-deficit/hyperactivity disorder: a cross-sectional clinical study
- Co-development, initial evaluation and implementation of a digital health intervention (SmartADHD) for young people (aged 16-30) with attention deficit hyperactivity disorder, including optimisation for sustainable delivery via health and social care settings: protocol for a mixed-methods study.
- Familial risk for attention-deficit/hyperactivity disorder and related behaviors in toddlerhood are preliminarily associated with altered frontoparietal development in infancy measured by functional near-infrared spectroscopy
BibTeX
@misc{sonmez2026predicting,
title = {{Predicting inattention and hyperactivity-impulsivity trajectories using random forest classification in the Adolescent Brain Cognitive Development Study}},
author = {Özde Sönmez and Hyun Ruisch and Jan Buitelaar and Christiane Thiel and Andrea Hildebrandt and Andrea Dietrich and Pieter Hoekstra},
howpublished = {PsyArXiv},
year = {2026},
month = sep,
doi = {10.31234/osf.io/bqdtk_v1},
url = {https://osf.io/preprints/psyarxiv/bqdtk_v1/},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Predicting inattention and hyperactivity-impulsivity trajectories using random forest classification in the Adolescent Brain Cognitive Development Study
AU - Özde Sönmez
AU - Hyun Ruisch
AU - Jan Buitelaar
AU - Christiane Thiel
AU - Andrea Hildebrandt
AU - Andrea Dietrich
AU - Pieter Hoekstra
PY - 2026
DA - 2026/09/10
PB - PsyArXiv
DO - 10.31234/osf.io/bqdtk_v1
UR - https://osf.io/preprints/psyarxiv/bqdtk_v1/
AB - Introduction: Despite well-established genetic and environmental risk factors for attention deficit–hyperactivity disorder (ADHD), it remains unclear why children show different developmental trajectories from mild to severe levels. This study aimed to investigate distinct developmental trajectories of inattention and hyperactivity-impulsivity scores and identify predictors of these trajectories from childhood to adolescence in a community-based sample. Methods: Data were drawn from the longitudinal Adolescent Brain and Cognitive Development dataset (n = 11,875) with yearly measures from baseline to wave 3 (ages 9-10 to 12-13). Growth mixture models were applied to the inattention and hyperactivity-impulsivity items of the ADHD subscale of the Child Behavior Checklist to identify trajectories. Random forest classification models were then trained to predict these trajectories using familial, demographic, behavioral and genetic variables, previously linked to ADHD. Logistic regression analyses were conducted to determine the direction of associations. Results: Three trajectories were identified for inattention (non-affected, moderate, and persistent-high), and four for hyperactivity-impulsivity (non-affected, worsening, improving, and persistent-high). Key predictors for both domains included male sex and higher externalizing problems, family conflict, and polygenic risk scores of ADHD, distinguishing affected from non-affected children. Additional distinguishing factors were less secondary caregiver acceptance (inattention) and lower family income (hyperactivity-impulsivity). Further, a persistent-high inattention trajectory was associated with negative school experiences and more internalizing problems, while worsening and persistent-high hyperactivity-impulsivity trajectories were linked to slower pubertal development. Conclusion: This study supports the distinction between inattention and hyperactivity-impulsivity trajectories, which may be shaped by domain-specific factors. Nevertheless, at its core, both ADHD domains appear to be associated with shared factors, including biological and environmental, in the distinction between non-affected youth and those with ADHD symptomatology.
ER -