The impact of pre- and perinatal factors on psychopathology in adulthood

Predicting adolescent psychopathology from early life factors: A machine learning tutorial

Objective

The successful implementation and interpretation of machine learning (ML) models in epidemiological studies can be challenging without an extensive programming background. We provide a didactic example of machine learning for risk prediction in this study by determining whether early life factors could be useful for predicting adolescent psychopathology.

Methods

In total, 9643 adolescents ages 9-10 from the Adolescent Brain and Cognitive Development (ABCD) Study were included in ML analysis to predict high Child Behavior Checklist (CBCL) scores (i.e., t-scores ≥ 60). ML models were constructed using a series of predictor combinations (prenatal, family history, sociodemographic) across 5 different algorithms. We assessed ML performance through sensitivity, specificity, F1-score, and area under the curve (AUC) metrics.

Results

A total of 1267 adolescents (13.1 %) were found to have high CBCL scores. The best performing algorithms were elastic net and gradient boosted trees. The best performing elastic net models included prenatal and family history factors (Sensitivity 0.654, Specificity 0.713; AUC 0.742, F1-score 0.401) and prenatal, family, history, and sociodemographic factors (Sensitivity 0.668, Specificity 0.704; AUC 0.745, F1-score 0.402). Across all 5 ML algorithms, family history factors (e.g., either parent had nervous breakdowns, trouble holding jobs/fights/police encounters, and counseling for mental issues) and sociodemographic covariates (e.g., maternal age, child's sex, caregiver income and caregiver education) tended to be better predictors of adolescent psychopathology. The most important prenatal predictors were unplanned pregnancy, birth complications, and pregnancy complications.

Conclusion

Our results suggest that inclusion of prenatal, family history, and sociodemographic factors in ML models can generate moderately accurate predictions of adolescent psychopathology. Issues associated with model overfitting, hyperparameter tuning, and system seed setting should be considered throughout model training, testing, and validation. Future early risk predictions models may improve with the inclusion of additional relevant covariates.

Association of perinatal characteristics with biomarkers of stress and inflammation in young adults: An exploratory study

General peri- and postnatal characteristics may serve as markers linking pre- or early postnatal events to later health outcomes, which in turn are associated with altered stress- and immune system activity. Our exploratory study investigated whether A) the common perinatal measures "birth weight" and "birth mode" and B) the postnatal characteristics "breastfeeding" and "vaccination status" are associated with markers of stress systems - the hypothalamic-pituitary-adrenal (HPA) axis and autonomous nervous system (ANS) - and inflammation in healthy young adults (n = 68, females: 70.6 %, mean age: 24.21 years, SD = 4.38) exposed to psychosocial challenge, the 'Trier Social Stress Test' (TSST). Salivary cortisol, alpha-amylase (sAA) and plasma interleukin-6 (IL-6) were assessed before, during and after the TSST. Participants provided information on peri- and postnatal characteristics. Linear regressions were performed to determine whether peri-/postnatal variables predict basal and stress-response-related biomarker levels. Controlling for sex and sex hormone use as relevant confounders, we found a significant association between birth weight and cortisol recovery (p = 0.032), with higher birth weight predicting higher cortisol recovery values. There were no other significant associations between predictor and outcome variables. Our results show that, in healthy young adults of mixed gender, normal-ranged birth weight is related to the cortisol response to psychosocial stress, indicating a long-term association of this perinatal marker with HPA axis function. In contrast, birth weight was not associated with markers of the ANS stress response or inflammation in adulthood. Our results further suggest that the measures birth mode, duration of breastfeeding, and vaccination status at 4 months of age do not relate to markers of the inflammatory and stress systems in adulthood.

Experience of discrimination reported during pregnancy and infant's emerging effortful control

Discrimination reported during pregnancy is associated with poorer offspring emotional outcomes. Links with effortful control have yet to be examined. This study investigated whether pregnant individuals' reports of lifetime racial/ethnic discrimination and everyday discrimination (including but not specific to race/ethnicity) reported during pregnancy were associated with offspring emerging effortful control at 6 months of age. Pregnant individuals (N = 174) and their offspring (93 female infants) participated. During pregnancy, participants completed two discrimination measures: (1) lifetime experience of racial/ethnic discrimination, and (2) everyday discrimination (not specific to race/ethnicity). Parents completed the Infant Behavior Questionnaire-Revised when infants were 6 months old to assess orienting/regulation, a measure of emerging effortful control. Analyses were conducted in a subsample with racially/ethnically marginalized participants and then everyday discrimination analyses were repeated in the full sample. For racially/ethnically marginalized participants, greater everyday discrimination (β = -.27, p = .01) but not greater lifetime experience of racial/ethnic discrimination (β = -.21, p = .06) was associated with poorer infant emerging effortful control. In the full sample, greater everyday discrimination was associated with poorer infant emerging effortful control (β = -.24, p = .002). Greater perceived stress, but not depressive symptoms, at 2 months postnatal mediated the association between everyday discrimination and emerging effortful control. Further research should examine additional biological and behavioral mechanisms by which discrimination reported during pregnancy may affect offspring emerging effortful control.

Intergenerational risk and resilience pathways from discrimination and acculturative stress to infant mental health

Preconception and prenatal stress impact fetal and infant development, and women of color are disproportionately exposed to sociocultural stressors like discrimination and acculturative stress. However, few studies examine links between mothers' exposure to these stressors and offspring mental health, or possible mitigating factors. Using linear regression, we tested associations between prenatally assessed maternal acculturative stress and discrimination on infant negative emotionality among 113 Latinx/Hispanic, Asian American, Black, and Multiethnic mothers and their children. Additionally, we tested interactions between stressors and potential pre- and postnatal resilience-promoting factors: community cohesion, social support, communalism, and parenting self-efficacy. Discrimination and acculturative stress were related to more infant negative emotionality at approximately 12 months old (M = 12.6, SD = .75). In contrast, maternal report of parenting self-efficacy when infants were 6 months old was related to lower levels of infant negative emotionality. Further, higher levels of parenting self-efficacy mitigated the relation between acculturative stress and negative emotionality. Preconception and prenatal exposure to sociocultural stress may be a risk factor for poor offspring mental health. Maternal and child health researchers, policymakers, and practitioners should prioritize further understanding these relations, reducing exposure to sociocultural stressors, and promoting resilience.

Are perinatal measures associated with adolescent mental health? A retrospective exploration with original data from psychiatric cohorts

BACKGROUND

Perinatal markers of prenatal development are associated with offspring psychiatric symptoms. However, there is little research investigating the specificity of perinatal markers for the development of specific disorders. This study aimed to explore if perinatal markers are specifically associated with adolescent substance use disorder (SUDs).

METHODS

Adolescent participants from two study centers, one for SUD patients (n = 196) and one for general psychopathology (n = 307), were recruited for participation. Since the SUD participants presented with a number of comorbid disorders, we performed a 1-on-1 matching procedure, based on age, gender, and specific pattern of comorbid disorders. This procedure resulted in n = 51 participants from each group. From all participants and their mothers we recorded perinatal markers (mode of birth, weeks of completed pregnancy, birth weight, Apgar score after 5 min) as well as intelligence quotient (IQ). The SUD sample additionally filled out the Youth Safe Report (YSR) as well as the PQ-16 and the DUDIT. We aimed to distinguish the two groups (SUD sample vs. general psychiatric sample) based on the perinatal variables via a logistic regression analysis. Additionally, linear regressions were performed for the total group and the subgroups to assess the relationship between perinatal variables and IQ, YSR, DUDIT and PQ-16.

RESULTS

The perinatal variables were not able to predict group membership (X² [4] = 4.77, p = .312, Cox & Snell R² = 0.053). Odds ratios indicated a small increase in probability to belonging to the general psychiatric sample instead of the SUD sample if birth was completed via C-section. After Bonferroni-correction, the linear regression models showed no relation between perinatal markers and IQ (p = .60, R² = 0.068), YSR (p = .09, R² = 0.121), DUDIT (p = .65, R² = 0.020), and PQ-16 (p = .73, R² =0.021).

CONCLUSION

Perinatal markers were not able to distinguish SUD patients from patients with diverse psychopathologies. This pattern contradicts previous findings, perhaps because our chosen markers reflect general processes instead of specific mechanistic explanations. Future studies should take care to investigate specific prenatal markers and associate them with psychopathology on the symptom level.

Maternal smoking around birth may lower the protective effects of breastfeeding on anxiety, depression and neuroticism in adult offspring: a UK biobank study

We aim to explore the combined effects of the smoking and breastfeeding on offspring mental health outcomes. We used data from UK biobank (N = 342,846) to evaluate joint effect of breastfeeding and maternal smoke during pregnancy (MSDP) on seven adult offspring mental health outcomes (self-reported depression, depression score, self-reported anxiety, anxiety score, neuroticism score, self-harm, suicide). We stratified individuals to MSDP group and non-MSDP group as well as breastfeeding group and non-breastfeeding group. Multiple linear regression and logistic regressions analysis were performed between independent variables (MSDP or breastfeeding) and dependent variables separately (seven mental health outcomes) in each stratum. Effect estimates were expressed as β values and OR values. Sex, age, 10 principle components of population structure, smoking, alcohol use, and Townsend deprivation index were examined as covariates. At MSDP grouping level, coefficients (odds ratio [OR]) for association of breastfed as a baby with self-reported anxiety (category variable) were 0.87 (95%CI, (0.82-0.93), P = 1.74 × 10^-5) in the MSDP group and 0.83 (95%CI, (0.79-0.87), P = 2.76 × 10^-17) in the non-MSDP group. At breastfeeding grouping level, OR for association of MSDP and self-reported anxiety were 1.15 (95%CI, (1.10-1.20), P = 5.36 × 10^-11) in breastfeeding group and 1.12(95%CI, (1.06-1.20), P = 2.02 × 10^-4) in non-breastfeeding group. At MSDP grouping level, negatively associations were found for breastfeeding and anxiety score (continuable variable) in MSDP group (-0.04 SD change per SD change in MSDP, 95% CI, (- 0.06, - 0.02), P = 2.42 × 10^-3) and non-MSDP group (-0.06 SD change per SD change in MSDP, 95%CI, (- 0.07, - 0.04), P = 1.70 × 10^-11). At breastfeeding grouping level, positive association was found for MSDP and anxiety score in the breastfeeding group (0.07 SD change per SD change in MSDP, 95%CI, (0.06-0.09), P = 1.49 × 10^-20) and non-breastfeeding group (0.07 SD change per SD change in MSDP, 95%CI, (0.05-0.09), P = 7.19 × 10^-8). Compared with non-MSDP group, the protective effect (reflected by coefficients) of breastfeeding on anxiety in the MSDP decreased. Our preliminary study found MSDP may lower the protective effect of breastfeeding on the adult offspring anxiety, depression and neuroticism, providing useful recommendations for health care service via quitting smoking during pregnancy and encouraging prolonged breastfeeding.

Perinatal complications are associated with social anxiety: Indirect effects through temperament

The current investigation examines the relation between perinatal complications and social anxiety incorporating the potential indirect effect of child temperament. Participants were 149 children 9 to 12 years of age (M_age=9.97, SD_age=1.00) screened for behavioral inhibition (BI) and assessed for social anxiety symptoms using parent- and child-report. Participating families also reported on the presence of perinatal complications. Results indicated that children who experienced perinatal complications were higher in BI and social anxiety, compared to children who did not experience complications. Furthermore, there was an indirect effect between perinatal complications and social anxiety via BI. These findings provide further support for the established relation between perinatal complications and anxiety and demonstrate, for the first time, that this relation may be mediated by temperament, setting the stage for longitudinal analyses.

Does Prenatal Maternal Distress Contribute to Sex Differences in Child Psychopathology?

PURPOSE OF REVIEW

Prenatal maternal psychological distress is an established risk factor for the development of psychopathology in offspring. The purpose of this review is to evaluate whether sex differences in fetal responses to maternal distress contribute to sex differences in subsequent psychopathology.

RECENT FINDINGS

Male and female fetuses respond differently to stress signals. We review recent evidence that demonstrates a sex-specific pattern of association between prenatal maternal distress and pathways associated with risk for psychopathology including offspring hypothalamic pituitary adrenocortical (HPA) axis regulation, brain development, and negative emotionality. Prenatal maternal distress exerts sex-specific consequences on the fetus. These differences may contribute to the well-established sex differences in psychopathology and in particular to greater female vulnerability to develop internalizing problems.