In a recent article published in JAMA Network Open, researchers investigated the role of genetics in the phenotypic association between screen time and psychiatric problems using a large population-based cohort study of preadolescents.
Specifically, they modeled genetic information using Gsens, a method previously used in prediction studies but rarely in causation studies to quantify genetic confounding by integrating information from both polygenic risk scores (PRSs) and single-nucleotide variant (SNV)- and twin-based heritability.
The extent to which genetic variations influence the complex association between screen time and children’s mental health, particularly internalizing and attention problems, remains unknown. Scientists speculate that genes affect screen time through neurodevelopmental pathways by altering the genetic expression of the central nervous system.
As genetic confounding may have a noncausal but direct influence on the association between child screen time and psychiatric problems, its evaluation has public health implications.
About the study
In the present study, researchers used the Adolescent Brain Cognitive Development (ABCD) study release 4.0 genotype data of 4,262 children with European ancestry aged nine to 11 years.
After requisite quality control and imputation, they extracted 6,833,710 genetic variants to compute genome-wide PRSs using a Bayesian scoring method. PRSs is the weighted sum score indicating the risk of some specific diseases due to an individual’s genotypes.
The team used samples of GWASs examining children with specific PRSs, such as attention-deficit/hyperactivity disorder (ADHD) (n = 55 374), depression (n = 500 199), and who watched television in leisure time (n = 365 236), all relevant proxy phenotypes for screen time.
Next, they asked children and their parents to fill in a 14-item questionnaire at baseline, which provided measures of daily screen time ranging from zero to four hours or more. At one-year follow-up, parents also completed the Achenbach Child Behavior Checklist for their six to 18-year-olds.
The researchers assessed attention and internalizing problems using the 10-item attention problem subscale (score range: 0-20) and the combined scores on other subscales (score range: 0-64), where higher scores indicated higher severity.
Several confounders were assessed in this study, including age, sex, and study site. Additionally, they adjusted study models for family income, highest parental education, and maternal psychopathological disorder as confounders and principal components (top 10) for residual confounding.
In primary statistical analyses, the team examined and quantified the associations between child-reported screen time and parent-reported attention or internalizing problems using linear regressions. They also quantified genetic confounding for these associations using the Gsens framework.
Three structural equation models used PRSs for the exposure and outcomes and modeled SNV- and twin-based heritability and the PRSs, respectively. The first one adjusted for genetic confounding, the second generated a lower-bound genetic confounding estimate, and the third delineated its upper bound.
These analyses standardized PRSs, child-and-parent-reported screen time, and pediatric psychiatric problems to mean zero and standard deviation (SD) 1 to facilitate comparisons.
This study followed the STREGA reporting guideline.
There were 4,262 children in this study, of which 2,269 were males with an average (SD) age of 9.9 years. The analysis found that screen time was associated with attention and internalizing problems (β = 0.10 and 0.03 SD), in agreement with prior research.
There was specificity in associations between PRSs and their corresponding traits. Of all, the television time PRS exhibited the highest association with child screen time and factored in both attention and internalizing problems.
Associations were also detected between other PRSs, such as ADHD PRS was associated with attention problems, and depression PRS was associated with internalizing problems.
The association of PRSs with cross-traits suggested horizontal pleiotropy of the genetic variants (shared genetic risk factors) and possible genetic confounding. Genetic confounding accounted for 42.7% of the association between child screen time and internalizing problems when using estimates of PRSs and SNV-based heritability, while it fully elucidated associations with both internalizing and attention problems when using PRSs and twin-based heritability estimates.
Notably, despite wide variations in the sample population, quality control thresholds, and quantification methods, SNV- and twin-based heritability estimates of this study’s analytic sample were comparable to previous studies.
Overall, genetic factors highly confounded the association between screen time and attention problems; conversely, their effect was relatively small on the association between screen time and internalizing problems. Likely, environmental factors (e.g., parenting practices) confounded residual associations.
Although associations between child screen time and mental health issues are complex, many policymakers and scientists view these as modifiable. So, parents should continue to prevent children from using electronic devices for prolonged periods. It could also help curb the adverse effects of excessive screen time on physical activity levels and academics of children.
To conclude, the study underscores the need to consider genetic factors in socio-behavioral research examining modifiable risk factors for mental health in young children and adolescents.