Impact of Childhood Trauma on Executive Function in Adolescence—Mediating Functional Brain Networks and Prediction of High-Risk Drinking

Greater childhood trauma severity was associated with more executive dysfunction across all study years.

“Childhood trauma severity was significantly related to ED in all years.”

• Also stated in: § Effects of Childhood Trauma on Executive Dysfunction · plain

  “we found that CTQ severity predicted BRIEF global executive composite scores (β = .14, 95% confidence inerval [CI] = .07–.21, p < .001).”

At baseline, functional connectivity distributed across hub regions in the right anterior insula, right intraparietal sulcus, bilateral dorsal anterior cingulate cortex, and bilateral pre- and postcentral gyri mediated the association between childhood trauma and executive dysfunction.

“At baseline, distributed functional connectivity from hub regions in the bilateral dorsal anterior cingulate cortex, right anterior insula, right intraparietal sulcus, and bilateral pre- and postcentral gyri mediated the relationship between childhood trauma and ED.”

• Also stated in: § Mediation Between Childhood Trauma and Executive Dysfunction by Functional Connectivity · plain

  “As per our hypothesis, mediation hub nodes were located in the bilateral dACC, as well as in the right anterior insula, the right intraparietal sulcus, and bilateral pre- and postcentral gyri.”

High-risk drinking during follow-up years 1 through 4 was predicted with high accuracy from baseline trauma-affected functional networks that mediated executive dysfunction, along with age, childhood trauma severity, and executive dysfunction extent.

“Furthermore, high-risk drinking in follow-up years 1–4 could be predicted with high accuracy from the trauma-affected functional brain networks that mediated ED at baseline, together with age, childhood trauma severity, and extent of ED.”

• Also stated in: § Predicting High-Risk Drinking in Future Years · plain

  “This random forest model had an overall accuracy of 0.80; high-risk drinkers were predicted with a precision of 0.79, recall (or sensitivity) of 0.83, and F1 score of 0.81, and low-risk drinkers were predicted with a cumulative precision of 0.83, recall (or specificity) of 0.78, and F1 score of 0.80.”

Connectivity alterations within and between low-level network regions in pre- and post-central gyri and visual cortex affected high-level executive function.

“Alterations in connectivity strength within and between “low-level” network regions in the pre- and post-central gyri and in the visual cortex (Table 2) were found to affect high-level executive function.”

The childhood trauma effect was consistent across abuse and neglect subscales.

“We also found the effect to be consistent across CTQ subscales for CT forms of abuse (β = .13, 95% CI = .08–.18, p < .001) and neglect (β = .04, 95% CI = 0.01–.07, p = .006).”

Eleven hub seed regions and their connectivity strengths to distributed cortical regions qualified as neurobiological mediators at baseline.

“We identified 11 hub seed regions and their connectivity strengths to distributed cortical regions that qualified as neurobiological mediators at BL”

Functional connectivity from the left postcentral gyrus and left dorsal anterior cingulate cortex hub regions ranked among the top three most important features.

“Also, functional connectivity from the left postcentral gyrus and the left dACC (part of the CON) hub regions ranked in the top 3 most important features”

Lower absolute functional connectivity strengths were observed in high-risk versus low-risk drinkers for the postcentral gyrus and dorsal anterior cingulate cortex at follow-up years 1 through 4.

“lower absolute functional connectivity strengths were observed in high-risk versus low-risk drinkers (postcentral gyrus: 0.044 ± 0.015 vs. 0.049 ± 0.012; dACC: 0.000 ± 0.011 vs. 0.016 ± 0.008) at follow-up years Y1 to Y4.”

In a second model including major depressive disorder and anxiety disorders, childhood trauma severity significantly predicted executive dysfunction.

“CTQ was found to significantly predict BRIEF ED (β = .11, 95% CI = .04–.17, p = .002).”

In the NCANDA longitudinal neuroimaging sample, childhood trauma exposure was significantly associated with executive dysfunction during adolescence.

“Using data from the large NCANDA longitudinal neuroimaging study, we showed significant associations between CT exposure and ED during adolescence.”

Lower functional connectivity values were associated with greater executive function impairment.

“and lower connectivity values being associated with higher impairments in executive function.”

The predictive odds ratio of the random forest model was 17.31.

“The predictive odds ratio of this model was 17.31, suggesting that it has diagnostic value.”

The childhood trauma effect was consistent for both Behavior Rating Inventory of Executive Function indices: Behavioral Regulation Index and Metacognition Index.

“The effect of CTQ severity was consistent for both BRIEF indices of ED, that is, the Behavioral Regulation Index (β = .13, 95% CI = .07–.20, p < .001) and the Metacognition Index (β = .13, 95% CI = .06–.20, p < .001)”

The contributing region-to-region functional connections occurred within and between known functionally organized brain networks.

“These contributing region-to-region functional connections were found within and between known functionally organized brain networks.”

Functional connections of particular brain hubs mediate the link between childhood trauma and poor executive function.

“As a novel contribution, we demonstrated the brain region hubs whose functional connections mediate the relationship between CT and poor executive function.”

Models that used only behavioral data had substantially lower sensitivity than models that also included functional connectivity.

“Finally, we also found that all models based on behavioral data alone, which excluded functional connectivity data, had much poorer sensitivity than when functional connectivity was included (Supplemental Figure S1).”

When psychiatric disorder prevalence at baseline and the first three follow-up years was modeled, major depressive disorder and anxiety disorders significantly affected executive dysfunction.

“However, when additionally modeling prevalence of psychiatric disorders at BL and the first 3 years of follow-up assessments, we found that MDD and anxiety disorders significantly affect ED”

Greater childhood trauma severity corresponded to weaker functional connectivity.

“with trauma severity corresponding to weaker functional connectivity”

All other mediating hub region functional connections were within the top two-thirds of feature ranks, up to rank 15, including connections from the right anterior insula, bilateral postcentral gyri, left precentral gyrus, and bilateral dorsal anterior cingulate cortex.

“All other mediating hub region functional connections were within the top two-thirds highest feature ranks (up to rank 15), including mediating connection strengths from the right anterior insula, bilateral postcentral gyri, left precentral gyrus, and bilateral dACC region.”

The random forest algorithm using demographic, behavioral, and functional connectivity data showed the best classification results, with the highest combined sensitivity and specificity of all models.

“The random forest algorithm based on demographic, behavioral, and functional connectivity data showed the best classification results, with combined highest sensitivity and specificity of all models.”

The neurobiological mediators linking childhood trauma severity to executive dysfunction were not confined to cognitive control networks.

“Notably, we found that neurobiological mediators between CT severity and ED were not limited to cognitive control networks.”

Childhood trauma exposure was linked to lower executive function at all five assessment years.

“We found that exposure to CT was associated with lower executive function across all 5 assessment years.”

The association between childhood trauma and executive function was relevant only among adolescents without mental health diagnoses.

“and that associations between CT and executive function are only relevant in adolescents without mental health diagnoses.”

The feature importance of baseline high-risk drinking was equivalent to that of executive dysfunction.

“Feature importance of BL high-risk drinking (see “frequent binge BL” factor in Figure 3C) was equivalent to that of ED.”

Age was the most important classification feature, with older adolescents more likely to develop high-risk drinking.

“Relative feature importance scores extracted for the random forest ensemble learning model (Figure 3C) indicated that the factor of age was of highest classification relevance: older adolescents were more likely to develop a pattern of high-risk drinking.”

These functional connectivity features contribute to prediction beyond the effect of earlier drinking status, that is, baseline high-risk drinking.

“Notably, these functional connectivity features inform prediction beyond effects of drinking status at an earlier age, i.e., high-risk drinking at BL.”

High-risk drinkers significantly overlapped with adolescents meeting clinical criteria for alcohol abuse.

“although notably, we did find a significant overlap between high-risk drinkers and adolescents that fulfill clinical criteria of alcohol abuse.”

The study found consistent effects across the abuse and neglect categories of the CTQ.

“That we found consistent effects of abuse and neglect categories of the CTQ may suggest that multiple measures that are differentially sensitive to these categories need to be used to better tease apart their effects (29–32).”

Connectivity of the left dACC seed region within the CON was among the top three features for classifying high-risk drinkers in follow-up years Y1 through Y4.

“Particularly, functional connectivity of the left dACC seed region within the CON was found to be within the top 3 features important for classifying high-risk drinkers in follow-up years Y1 to Y4.”

Childhood trauma severity, related functional connectivity, and executive function can predict high-risk drinking in adolescents.

“We found that severity of CT, associated functional connectivity, and executive function can be used to predict high-risk drinking in adolescents.”

The study's results have relevance for forecasting alcohol use disorders.

“These findings are relevant for the prognosis of alcohol use disorders.”

Mediating brain connectivity was primarily distributed within and across hub regions in the salience network, cingulo-opercular network, and motor networks.

“Thus, mediating brain connectivity was primarily distributed within and across hub regions in the salience network, CON, and motor networks”

The high-risk drinking criterion showed significant overlap with alcohol use disorder individuals in each year.

“In the current sample, this high-risk drinking criterion showed significant overlap with AUD individuals in each year (Y1: χ2 = 13.75, p = .019; Y2: χ2 = 11.96, p = .013; Y3 χ2 = 10.59, p = .010).”

Prior research in clinical and nonclinical groups has linked neurodevelopmental changes caused by childhood trauma to executive function deficits.

“Previous research in clinical and nonclinical populations has established a link between the neurodevelopmental changes caused by CT and deficits in executive function”

• cites [11] Gould, 2012, The effects of child abuse and neglect on cognitive functioning in adulthood, J Psychiatr Res sourceNot checked yet

  Not checked in the free run.

• cites [12] Aas, 2014, A systematic review of cognitive function in first-episode psychosis, including a discussion on childhood trauma, stress, and inflammation, Front Psychiatry sourceNot checked yet

  Not checked in the free run.

Different forms of childhood trauma, abuse versus neglect, can have differential biological and neurofunctional impacts.

“different forms of CT, abuse versus neglect, can have differential biological and neurofunctional impacts (29–32)”

• cites [29] Sumner, 2019, Early experiences of threat, but not deprivation, are associated with accelerated biological aging in children and adolescents, Biol Psychiatry sourceNot checked yet

  Not checked in the free run.

• cites [30] Lambert, 2017, Differential associations of threat and deprivation with emotion regulation and cognitive control in adolescence, Dev Psychopathol sourceNot checked yet

  Not checked in the free run.

• cites [31] Sheridan, 2017, Dimensions of childhood adversity have distinct associations with neural systems underlying executive functioning, Dev Psychopathol sourceNot checked yet

  Not checked in the free run.

• cites [32] Gur, 2019, Burden of environmental adversity associated with psychopathology, maturation, and brain behavior parameters in youths, JAMA Psychiatry sourceNot checked yet

  Not checked in the free run.

In an international sample, the authors previously found that childhood trauma impairs the dACC's maturing connectivity within the CON.

“Furthermore, we recently found that the dACC’s maturing connectivity within the CON is hampered by CT in an international sample, which led to the hypothesis that functional connections from dACC will mediate the relationship between CT and executive function (27).”

• cites [27] Mishra, 2017, T13. Plasticity of cognitive control in adolescents with adverse childhood experiences, American College of Neuropsychopharmacology 56th Annual Meeting: Poster Session II, December 5, 2017. NeuropsychopharmacologyNot checked yet

  Not checked in the free run.

Childhood trauma is associated with weaker functional connectivity strength from the dorsal anterior cingulate cortex to the anterior insular cortex during adolescence.

“we recently found that CT is associated with weaker functional connectivity strength from the dACC to the anterior insular cortex, one of the core regions of the cingulo-opercular network (CON), during adolescence (27).”

• cites [27] Mishra, 2017, T13. Plasticity of cognitive control in adolescents with adverse childhood experiences, American College of Neuropsychopharmacology 56th Annual Meeting: Poster Session II, December 5, 2017. NeuropsychopharmacologyNot checked yet

  Not checked in the free run.

Resting-state functional brain networks, especially those supporting executive function, develop rapidly during typical adolescence.

“Resting-state functional brain networks, especially those that subserve executive function, are rapidly developing during typical adolescence (21–23)”

• cites [21] Uddin, 2010, Typical and atypical development of functional human brain networks: insights from resting-state fMRI, Front Syst Neurosci sourceNot checked yet

  Not checked in the free run.

• cites [22] Ordaz, 2013, Longitudinal growth curves of brain function underlying inhibitory control through adolescence, J Neurosci sourceNot checked yet

  Not checked in the free run.

• cites [23] Luna, 2015, An integrative model of the maturation of cognitive control, Annu Rev Neurosci sourceNot checked yet

  Not checked in the free run.

The random forest model has diagnostic value.

“suggesting that it has diagnostic value.”

Most findings on childhood trauma and alcohol outcomes come from adult population-based data with retrospective reports of childhood trauma.

“However, most findings are from adult population-based data with retrospective reports of CT (8,9).”

• cites [8] Just-Østergaard, 2017, Major life events and risk of slcohol use disorders: a prospective cohort study, Addiction sourceNot checked yet

  Not checked in the free run.

• cites [9] McEwen, 2015, Redefining neuroendocrinology: Stress, sex and cognitive and emotional regulation, J Endocrinol sourceNot checked yet

  Not checked in the free run.

Mental health status has a distinct moderating effect on trauma-related executive dysfunction.

“These findings highlight a distinct and moderating effect of mental health status on trauma-related ED.”

Behavioral studies support that dysregulated self-control can mediate the relationship between childhood trauma and alcohol use.

“behavioral studies have corroborated that dysregulated self-control can mediate the relationship between CT and alcohol use (16).”

• cites [16] Shin, 2019, Childhood maltreatment and alcohol use in young adulthood: the role of self-regulation processes, Addict Behav sourceNot checked yet

  Not checked in the free run.

Alcohol use disorder prevalence in the sample was low, at most 2.9%.

“A further limitation is the low prevalence of AUDs in the current sample (only up to 2.9%), demanding caution in a generalization of the results to the development of AUDs”

Executive function assessment relied on subjective self-report rather than objective performance measures.

“Regarding assessments of executive function, this study is limited by subjective self-report instead of objective performance measures.”

Adolescent high-risk drinking can be inferred from neurocognitive network impairments linked to trauma.

“This finding demonstrates that high-risk drinking in adolescents can be inferred from trauma-related impairments in underlying neurocognitive functional networks.”

• Also stated in: § DISCUSSION · boosted

  “Novel to this study, we found that prediction of high-risk drinking is critically informed by the trauma-affected brain networks that mediate poor executive function; this result underscores the value of functional brain connectivity in precision psychiatry.”

Because the CTQ is retrospective, the causal direction between childhood trauma and baseline executive function deficits remains unclear.

“While this allows assessment of long-term effects and developmental trajectories after the time of assessment, the cause and effect relationship with executive function deficits at study BL remains unclear; it is possible that children and adolescents with ED are more prone to experience trauma or to report trauma experience.”

The study extends earlier work that examined only subjective self-regulation behaviors as mechanisms.

“Our study thus extends prior work that has been limited to probing subjective self-regulation behaviors as underlying mechanisms (16).”

• cites [16] Shin, 2019, Childhood maltreatment and alcohol use in young adulthood: the role of self-regulation processes, Addict Behav sourceNot checked yet

  Not checked in the free run.

The findings emphasize the importance of distributed networks for cognitive and behavioral control and sensorimotor integration in predicting high-risk alcohol use.

“Taken together, our findings highlight the important role of distributed networks relevant for cognitive and behavioral control, as well as sensorimotor integration in the prediction of high-risk alcohol use behavior.”

By including brain function, the findings complement prior adolescent longitudinal studies that could not prospectively resolve the childhood trauma-excessive drinking relationship.

“By including brain function, our findings further complement prior longitudinal studies in adolescent samples that have not been able to prospectively resolve the relationship between CT and excessive drinking (58,59).”

• cites [58] Patten, 2015, Retrospective and prospectively assessed childhood adversity in association with major depression, alcohol consumption and painful conditions, Epidemiol Psychiatr Sci sourceNot checked yet

  Not checked in the free run.

• cites [59] Scott, 2012, Childhood maltreatment and DSM-IV adult mental disorders: Comparison of prospective and retrospective findings, Br J Psychiatry sourceNot checked yet

  Not checked in the free run.

Functional brain connectivity from a distributed set of hub regions may act as functional biomarkers that mediate the link between early-life adversity and executive behaviors.

“The present study demonstrates that functional brain connectivity from a distributed set of hub regions may serve as functional biomarkers mediating the relationship between early-life adversity and executive behaviors.”