Tracing Causal Paths from Prenatal Tobacco Exposure to Adolescent Brain and Behavioral Development

1790 

Abstract Submission 

Marvin Yan¹, Malick Abid², Erich Kummerfeld², Michael Kotlyar³, Jazmin Camchong⁴, Scott Sponheim^4,5, Bonnie Klimes-Dougan¹, David Bond⁶, Eric Rawls⁴

¹University of Minnesota, Department of Psychology, Minneapolis, MN, ²University of Minnesota, Institute for Health Informatics, Minneapolis, MN, ³University of Minnesota, Department of Experimental and Clinical Pharmacology, Minneapolis, MN, ⁴University of Minnesota, Department of Psychiatry and Behavioral Sciences, Minneapolis, MN, ⁵Minneapolis VA Health Care System, Minneapolis, MN, ⁶Johns Hopkins University, Department of Psychiatry and Behavioral Sciences, Baltimore, MD

Marvin Yan  
University of Minnesota, Department of Psychology
Minneapolis, MN

Malick Abid  
University of Minnesota, Institute for Health Informatics
Minneapolis, MN

Erich Kummerfeld  
University of Minnesota, Institute for Health Informatics
Minneapolis, MN

Michael Kotlyar  
University of Minnesota, Department of Experimental and Clinical Pharmacology
Minneapolis, MN

Jazmin Camchong  
University of Minnesota, Department of Psychiatry and Behavioral Sciences
Minneapolis, MN

Scott Sponheim, PhD  
University of Minnesota, Department of Psychiatry and Behavioral Sciences|Minneapolis VA Health Care System
Minneapolis, MN|Minneapolis, MN

Bonnie Klimes-Dougan  
University of Minnesota, Department of Psychology
Minneapolis, MN

David Bond  
Johns Hopkins University, Department of Psychiatry and Behavioral Sciences
Baltimore, MD

Eric Rawls, PhD  
University of Minnesota, Department of Psychiatry and Behavioral Sciences
Minneapolis, MN

In 2021, over 3 million adolescents in the United States were exposed to tobacco prenatally (prenatal tobacco exposure; PTE) (Osterman et al., 2023), a concern given the vital neural development that occurs during the fetal period. PTE is linked to altered neural communication and increased psychopathology (Indredavik et al., 2007; Muller et al., 2013), which is the most proximal risk factor for suicide-related behaviors including suicidical ideation (SI) and non-suicidal self-injury (NSSI). These influences are complicated to unravel, as PTE-related influences extend to physical health, cognitive ability, and intent to try tobacco later in life (future tobacco use; FTU) (Duko et al., 2021; Gonzalez et al., 2023). Existing research on PTE has studied these factors in isolation, but has not examined whether mediated causal paths from PTE to SI and NSSI exist. Traditional analytic techniques can only calculate associations between variables, while effective interventions require causality. Thus, causal discovery analysis (CDA), a statistical approach that uses machine learning to infer causal relationships between observed data, is necessary (Rawls et al., 2021). As part of a project investigating the impact of PTE on FTU, we used CDA to identify causal paths between PTE and neural, behavioral, and clinical factors to unravel the causal paths linking PTE to these factors, thereby illuminating potential intervention targets for NSSI and SI.

To investigate causal pathways between PTE, SI, and NSSI, we analyzed data from 8,884 adolescents (9-10 years old; 47.9% female), of which 1,142 (12.9%) experienced PTE, from the Adolescent Brain Cognitive Development Study baseline visit. We included measures assessing resting-state functional connectivity within established resting-state brain networks (RSFC), psychopathology, cognitive ability, physical health, PTE, and risk for FTU. A factor analysis was conducted for phenotypic data reduction, which excluded key variables of interest (PTE, SI, NSSI, RSFC).

PTE directly decreased RSFC within the cingulo-opercular (CO) network, which had indirect causal effects on outcome measures of interest. Specifically, decreases in CO RSFC causally increased self-reported psychopathology (SRP) (B=-.08), which then causally increased SI (B=.19) and NSSI (B=.15). As such, we can trace a causal path from PTE to psychopathology and suicide-linked behaviors, mediated through functional brain connectivity. PTE also indirectly caused increases in SI by causing increases in sleep problems (B=.09), parent-reported psychopathology (PRP) (B=.11), and impulsivity (B=.04), all of which then separately caused increases in SI (B=.08, .25, .10; i.e., three distinct causal paths). A direct causal relationship between SI and NSSI existed such that SI increased NSSI (B=.23); thus, all paths from PTE to SI also resulted in increased NSSI. Of note, in addition to their indirect causal paths to NSSI through SI, there were also direct causal paths that increased NSSI from impulsivity (B=.09), SRP (B=.15), and PRP (B=.01). Multiple causal paths from PTE also lead to risk for FTU, although due to the young age of the sample, this result should be considered exploratory. All results were adjusted such that p<.001.

This research illuminates how PTE can initiate a cascade of adverse developmental effects on adolescent brain connectivity, psychological health, and behavior. The findings underscore the impact of PTE on developing brain networks, particularly the salience network, as well as PTE's influences on impulsivity, cognitive control, and psychopathology. Our research implies that interventions focused on sleep problems, psychopathology, and impulsivity may causally reduce SI and NSSI. Finally, a brain-based intervention that increases salience network connectivity may be effective for reducing SI and NSSI. These insights advance our understanding of PTE's far-reaching consequences and pave the way for targeted interventions.

Early life, Adolescence, Aging ²

Connectivity (eg. functional, effective, structural)

fMRI Connectivity and Network Modeling ¹

Other Methods

Affective Disorders

Cognition

Computational Neuroscience

Data analysis

FUNCTIONAL MRI

Machine Learning

Modeling

Sleep

Statistical Methods