Using Machine Learning to Identify Predictors of Maternal and Infant Hair Cortisol Concentration Before and During the COVID-19 Pandemic.

Using machine learning to identify predictors of maternal and infant hair cortisol concentration before and during the COVID-19 pandemic - PMC Skip to main content An official website of the United States government Here's how you know Here's how you know Search Log in Primary site navigation Search Log in Search PMC Full-Text Archive Search in PMC Journal List User Guide As a library, NLM provides access to scientific literature. Inclusion in an NLM database does not imply endorsement of, or agreement with, the contents by NLM or the National Institutes of Health. Learn more: PMC Disclaimer | PMC Copyright Notice Stress Health . Author manuscript; available in PMC: 2026 Jun 1. Published in final edited form as: Stress Health. 2025 Jun;41(3):e70051. doi: 10.1002/smi.70051 Using machine learning to identify predictors of maternal and infant hair cortisol concentration before and during the COVID-19 pandemic Gabriel A León 1 , Alyssa R Morris 2 , Pia Sellery 3 , Darby E Saxbe 1 Author information Copyright and License information 1 Department of Psychology, University of Southern California 2 Semel Institute for Neuroscience and Human Behavior, University of California, Los Angeles 3 Department of Psychology and Neuroscience, University of Colorado Boulder ✉ Correspondence: Correspondence concerning this article should be addressed to Gabriel A. León, Department of Psychology, University of Southern California, Los Angeles, CA, 90089. gleon@usc.edu . PMC Copyright notice PMCID: PMC12857001  NIHMSID: NIHMS2096943  PMID: 40444444 The publisher's version of this article is available at Stress Health Abstract Hair cortisol concentration (HCC) has been theorized to reflect chronic stress, and maternal and infant HCC may be correlated due to shared genetic, physiological, behavioral, and environmental factors, such as stressful life circumstances. The current study examines HCC as a retrospective indicator of hypothalamic pituitary adrenal (HPA) axis activity in mothers and infants in the context of a major ecological stressor, the COVID-19 pandemic. First, we will compare HCC across two cohorts of mothers and their infants at six months postpartum. One cohort was recruited before the COVID-19 pandemic ( N = 154; 77 mothers, 77 infants) and another cohort was recruited during the first wave of COVID-19 lockdowns in the United States ( N = 120; 60 mothers, 60 infants). Next, we will apply machine learning to identify indicators of psychological stress that best predict maternal and infant HCC across these two cohorts. Our set of predictors will include pre- and postpartum measures of maternal perceived stress, parenting stress, and depressive symptoms. Finally, we will test for within-dyad covariation in mother-infant HCC and investigate whether covariation changes with respect to mothers’ psychological stress or their experience of the pandemic. Our findings will inform research on hair cortisol as a measure of psychological stress across the peripartum window, particularly in the context of large-scale stressors. Cortisol, a glucocorticoid hormone, reflects the activity of the hypothalamic pituitary adrenal (HPA) axis – a neuroendocrine pathway that regulates the physiological stress response (McEwen, 2004). Hair cortisol concentration (HCC) is a non-invasive biomarker of HPA axis activity that may link psychosocial experience to biological stress during sensitive periods of development such as the transition to parenthood ( Dickerson & Kemeny, 2004 ; Kirshbaum et al., 2009 ). However, research on the relationship between HCC and psychosocial stress has yielded many mixed or inconclusive findings, warranting more data-driven investigations. Systematic reviews suggest HCC may correlate more strongly with external, observable stressors compared to subjective, self-reported measures of distress ( Stalder et al., 2017 ). The COVID-19 pandemic is an example of a large-scale stressor that has been shown to amplify distress for new parents ( Mollard et al., 2021 ) and thus may be a particularly promising window within which to investigate HPA axis activity via HCC. The current study, a registered report, will investigate HCC in two cohorts of mothers and their infants assessed at six months postpartum: one recruited between 2013–2018, before the COVID-19 pandemic, and another recruited in spring 2020, during the first wave of COVID-19 lockdowns. We will compare HCC levels across the two cohorts. Using machine learning techniques (regularized regression, random forests, gradient boosting machines), we will identify indicators of prenatal and postpartum psychological distress most predictive of maternal and infant HCC. Finally, we will assess if mother-infant HCC covariation varies as a function of maternal psychological stress and pandemic exposure. Hair Cortisol Concentration Hair cortisol assays provide a non-invasive method for assessing cortisol secretion over time by measuring cortisol diffused into the hair shaft during growth ( Kirschbaum et al., 2009 ). Individuals vary, but on average, hair grows about 1cm per month, suggesting that 3cm of hair taken nearest to the scalp would index systemic cortisol secretion in the prior three months ( Greff et al., 2019 ; Kalliokoski et al., 2019 ; Stalder et al., 2017 ). HCC has been found to be moderately correlated with salivary, plasma, and urine samples in humans and non-human primates ( Kapoor et al., 2018 ; Short et al., 2016 ; Solorzano et al., 2023 ; Stalder et al., 2017 ; Sugaya et al., 2020 ; van Ockenburg et al., 2016 ). During pregnancy, circulatory cortisol peaks during the third trimester before returning to pre-pregnancy levels in the weeks following childbirth – making HCC a valid indicator of mothers’ HPA axis activity by six months postpartum ( Howland et al., 2017 ). Perinatal Hair Cortisol and Psychological Distress Studies on perinatal HCC have generated mixed results regarding the association between psychological distress and HCC. A meta-analysis of 28 independent studies finds a positive but non-significant association between psychological distress and women’s perinatal HCC ( Khoury et al., 2023 ). Of nine studies published in the last decade on postpartum HCC specifically, only three evidence a positive association between maternal HCC and symptoms of depression, anxiety, or perceived stress (e.g., Broeks et al., 2021 ; King et al., 2022 ; Nyström‐Hansen et al., 2019 ). These mixed results may be attributable to many causes including differences in physical activity, sleep patterns, metabolic activity, alterations of one’s circadian rhythm (e.g., due to shiftwork), and hormonal fluctuations in women ( Greff et al., 2019 ; Klusmann et al., 2022 ; Manenschijin et al., 2011 ; Stalder et al., 2017 ; Staufenbiel et al., 2015 ). Even fewer studies have assessed the influence of maternal distress on infant HCC ( Karl et al., 2023 ; Liu et al., 2016 ). Of note, no studies to date have tested the impact of parenting stress or division of caregiving responsibilities on postpartum HCC in mothers and infants despite the salience of parenting-specific stress during the transition to parenthood ( Thomason et al., 2014 ; Yavorsky et al., 2015 ). One study finds that mothers’ daily hassles positively predicted HCC during pregnancy ( Mustonen et al., 2019 ), but no studies have replicated this finding in the postpartum period. Our study fills this gap by directly comparing the effect of perceived stress, depressive symptoms, parenting stress, and caregiving responsibilities on postpartum HCC. Hair Cortisol and the COVID-19 Pandemic Work from our laboratory and others suggests that mothers of infants showed greater psychological distress during the pandemic compared to non-pandemic samples ( King et al., 2022 ; León et al., 2025 ; Morris & Saxbe, 2022 ). Therefore, we conceptualize the COVID-19 pandemic as a highly stressful, disruptive event that may predict greater HCC in mothers and infants from our pandemic cohort ( Bruinhof et al., 2022 ; Doan et al., 2022 ). Although we consider the pandemic to be a large-scale objective stressor, its specific effects on health may vary based on contextual factors such as individuals’ socioeconomic status (SES), health concerns, social isolation, and disruptions to daily routines. Such factors may intersect to mitigate or magnify psychological stress or HPA axis activity ( Bruinhof et al., 2022 ). Our study capitalizes on data-driven machine learning to identify associations between maternal psychological stress and HCC during the COVID-19 pandemic, while considering intersecting effects of different types of stressors measured pre- and postnatally (e.g., mothers’ subjective stress and mental health). Furthermore, we also consider the impact of contextual factors such as maternal age, education, ethnicity, socioeconomic status, and time of measurement. Covariation in Mother-Infant Hair Cortisol Concentration Researchers studying biobehavioral synchrony in mother-infant dyads posit numerous pathways by which mothers may influence infant stress, including fetal programming during gestation, hormone transmission via breast milk, and behavioral exchanges of gaze, touch, and vocalizations (Feldman, 2012). Physiological synchrony has been observed in face-to-face mother-infant interactions through covariation of the autonomic nervous system ( Water et al., 2014 ; 2017 ) and HPA axis activity measured via salivary cortisol ( Tarullo et al., 2017 ). Because HCC indexes basal HPA axis activity across weeks and months, it is more difficult to measure “synchrony” in mother-infant HPA axis activity using this HCC compared to salivary cortisol, which fluctuates hourly ( Davis et al., 2018 ; e.g., Tarullo et al., 2017 ). Furthermore, given that HCC has been found to have heritability estimates as high as ~70% ( Ritschel et al., 2017 ), much of the covariation in mother-infant HCC is likely explainable by genetic influences. Indeed, multiple studies find positive associations between infant and maternal HCC in the postpartum period ( Broeks et al., 2021 ; Stoye et al., 2021 ), including a study of 241 mother-infant dyads at twelve-month postpartum ( Galbally et al., 2019 ). However, some covariation in HCC among mothers and their children may be due to shared environmental influences, given that hair cortisol concentration shows relatively weak intra-individual reliability over long durations of time ( r = .32 over a two-year period; Ritschel et al., 2017 ). For example, a study of 422 twin pairs found that additive genetic factors explain 39% of variability in HCC, whereas nonshared environmental factors explained the remaining 61% of variance ( Cantave et al., 2022 ). Another sample of 1070 twin pairs found that additive genetic factors explain approximately 50% of the variance in HCC ( Tucker-Drob et al., 2017 ). In stressful environments, mothers might show greater covariation with infants ( Davis et al., 2018 ; Feldman, 2012). It could be that highly stressed mothers are sending a “stronger signal” than normal to their infants, which might cause infants to attune more strongly to their mothers’ physiology. In support of this idea, Liu and colleagues find maternal and infant HCC to be positively correlated in a high-risk, low SES sample from Brazil, but not in a middle-class U.S. sample ( Liu et al., 2016 , 2017 ). Given that we analyze data collected from two independent cohorts, one collected before and one collected after the onset of the pandemic, we aim to test for differences in mother-infant HCC covariation across both cohorts as a function of mothers’ stress and caregiving responsibilities. Supervised Machine Learning As previously mentioned, meta-analyses exploring the association between HCC and psychological stress find that effects are difficult to replicate and null findings are common ( Stalder et al., 2017 ). Most cross-sectional studies assessing HCC across the peripartum window use statistical tests like analysis of variance (ANOVA) and ordinary least squares (OLS) regression which are ill-equipped for exploring nonlinear associations and interactions among many collinear predictors ( Fife & D’Onofrio, 2023 ; Marceau et al., 2020 ). Supervised machine learning techniques such as regularized regression, random forests, and gradient boosting machines outperform linear regression on most tasks and retain many of its benefits (i.e., null hypothesis testing) while also offering solutions for interrogating interactions, variable importance, and partial effects among many collinear predictors ( Fife & D’Onofrio, 2023 ; Henninger et al., 2023 ; James et al., 2021 ). The capacity to explore interactions and nonlinear associations using graphical techniques like partial dependence plots help to make machine learning algorithms more interpretable – addressing a major criticism that these techniques are “black boxes.” Currently, we find only one other study that applied machine learning to study the correlates of HCC, and these authors were successful in identifying which specific periods of abuse and neglect during development were most predictive of HCC in adulthood ( Schalinski et al., 2019 ). Extending this approach, we hope to compare the performance of multiple machine learning algorithms to find a model of best fit and identify the indicators of psychological stress that are most predictive of HCC in mothers and infants across cohorts. Present Study The present study examines HCC within mother-infant dyads and compares two independent cohorts – one cohort recruited prior to the onset of the 2020 COVID-19 pandemic ( N = 77) and another cohort of mothers who were pregnant during the first wave of the pandemic lockdowns in spring 2020 ( N = 60). As mentioned, we have already found increases in prenatal psychological stress and postpartum depressive symptoms in our cohort of mothers from the first wave of the pandemic ( León et al., 2025 ; Morris & Saxbe, 2023 ). The present study builds upon these prior investigations. First, we explore the main effect of exposure to the COVID-19 pandemic on maternal and infant HCC across the two cohorts assessed at six months postpartum. Of note, participants were not randomly assessed to the pandemic exposure group. Therefore, we do not infer direct causal associations between pandemic exposure and HCC. Next, we will use machine learning to identify indicators of prenatal and postpartum psychological stress that are most predictive of maternal and infant HCC, and we will test if these effects vary as a function of pandemic exposure. Predictors will include maternal perceived stress, parenting stress, caregiving responsibilities, and depressive symptoms. We include measures collected during pregnancy, at three months postpartum, and at six months postpartum to evaluate if psychological stress measured closer in time to hair sampling is more strongly associated to HCC. Finally, we will test for covariation in mother-infant HCC and assess if covariation varies across cohorts and across levels of mother’s stress. We will test the hypotheses outlined below: H1: Maternal and infant HCC will be greater in the pandemic cohort compared to the non-pandemic cohort. H2: Across both cohorts, maternal perceived stress, parenting stress, caregiving responsibilities, and depressive symptoms will predict greater maternal and infant HCC We will test whether the effects of perceived stress, parenting stress, caregiving responsibilities, and depressive symptoms on maternal and infant HCC differ by pandemic cohort with no a priori hypotheses regarding the direction of these effects. H3: Maternal and infant HCC will be positively correlated across both cohorts – indicating positive maternal-infant HCC covariation. Maternal-infant HCC covariation will be greater in the pandemic cohort compared to the non-pandemic cohort. We will test for differences in maternal-infant HCC with respect to mothers’ perceived stress, parenting stress, caregiving responsibilities , and depressive symptoms , but we do not posit any a priori hypotheses about the direction of these effects. Method Participants Pre-pandemic Cohort The first cohort was recruited as a part of a multiyear longitudinal study of the biopsychosocial correlates of first-time parenthood (see Saxbe et al., 2019 ; 2023 ). To meet eligibility criteria, participants needed to be cohabiting, opposite-sex couples, expecting their first child, anticipating a singleton birth, and planning on cohabiting after the birth of their child. Participants also needed to be at least 18 years old and speak English. Participants were recruited through flyers posted in obstetricians’ offices and community health centers, social media posts on parenting-related forums, and word of mouth. The current study analyzes survey data collected from mothers in this cohort during two in-laboratory visits, one that occurred during mid-to-late pregnancy (20 – 36 weeks) and one that occurred at six months postpartum. We will also analyze survey data collected remotely, via an online questionnaire battery, at three months postpartum. The present study analyses HCC extracted at the six-month postpartum visit from both mothers and infants. Informed consent was received from all couples, and procedures were approved by the University of Southern California’s Institutional Review Board (IRB). Of the 100 mothers and 100 infants who participated in the larger study, thirteen families did not provide hair samples and ten families provided hair samples after the onset of COVID-19 pandemic lockdowns in March 2020. We dropped these observations from the final analyses. We also dropped mother-infant dyads from analyses if the HCC value being used as the outcome (either mothers or infants) was missing. The machine learning algorithms we aim to employ accommodate missing values of predictors, but not missing outcomes. Therefore, our final sample of mothers and infants included in at least one model is N = 154 (77 mothers; 77 infants). Pandemic Cohort The second cohort was recruited via social media in spring 2020 for an online study of the effects of COVID-19 pandemic lockdowns on prenatal mental health (see Morris et al., 2020 ). Data collection commenced on April 6th, 2020, the same week that “sheltering in place” behavior peaked in the U.S. (Schaul et al., 2020). Participants provided consent and completed questionnaires through the Qualtrics platform. We followed up with participants at three, six, and 12 months postpartum, with the timing of follow-up contact based on infant due dates (as provided prenatally) and/or actual birth dates (as provided in the first postpartum survey). At all timepoints, participants responded to questionnaires assessing relationship quality and mental health. The study design and questionnaire battery were modeled after the battery used in the pre-pandemic study in anticipation of comparing the two cohorts. All study procedures were IRB approved, and participants were compensated for participation (see Morris & Saxbe, 2022 ; 2023 ). Responses were carefully screened for potentially fraudulent participation, as detailed in Morris et al. (2022) . We contacted participants directly to clarify discrepancies in cases of suspected fraud and dropped participants whose discrepancies were not resolved. A total of 760 unique responses were received via Qualtrics, and 615 of those responses met all of our verification checks. Of those, 56 respondents provided both maternal and infant hair samples, two provided maternal hair samples only, and two provided infant samples only. Therefore, our final sample of mothers and infants included in at least one model is N = 120 (60 mothers; 60 infants). Measures Hair Cortisol Concentration In both pre-pandemic and pandemic cohorts, approximately 50–100 strands of hair were located near the lower posterior vertex of the scalp were tied together using a slipknot and clipped from mother and infants, respectively. Hair samples were covered with a thin layer of aluminum foil to prevent degradation by UV-light and then stored in individually sealed plastic bags at room temperature. Hair sampling was conducted by research assistants in the lab for the pre-pandemic cohort and self-sampled from home and sent back via postal mail in the pandemic cohort. We sent detailed instructions with photographs to participants in the pandemic cohort – the same instructions that research assistants used in the lab to collect hair samples. Participants did not provide information concerning hair care, hair type, or exposure to hair dyes. For both cohorts, cortisol assays were performed on the 2 cm of hair nearest to the scalp (approximating hair growth over the previous couple of months) by the TU Dresden laboratory led by Dr. Clemens Kirschbaum ( Kirschbaum et al., 2009 ). Samples were analyzed using liquid chromatography-tandem mass spectrometry (LC-MS/MS) protocol with established validity in other studies ( Gao et al., 2013 ). Demographic information In both samples, we collected data on maternal age, ethnicity, education, and household income (see Table 1 ). Each of these demographic factors will be included in analyses. The dates each participant provided their survey responses and hair samples are depicted in Figure 1 . All respondents in the pre-pandemic sample were living in the Los Angeles, CA, USA metropolitan area during the time of data collection. Respondents in the pandemic sample were surveyed from across the United States, with the largest portion (21.67%) being from California (see Figure S4 and Table S1 of the Supplementary Materials ). Table 1. Demographic Information Variable Both Samples Pandemic Pre-Pandemic Household Income, n(%) <$25,000 3 (2.20%) 2 (3.30%) 1 (1.30%) $25,000 - $50,000 11 (8.0%) 4 (6.70%) 7 (9.10%) $50,000 - $75,000 12 (8.80%) 9 (15.0% 3 (3.90%) $75,000 - $100,000 15 (10.90%) 7 (11.70%) 8 (10.40%) $100,000 - $125,000 9 (6.60%) 6 (10.0%) 3 (3.90%) >$125,000 42 (30.70%) 31 (51.70%) 11 (14.30%) Decline 45 (32.80%) 1 (1.70%) * 44 (57.10%) Education, n(%) High School/GED 3 (2.20%) 2 (3.30%) 1 (1.30%) Some College 9 (6.60%) 1 (1.70%) 8 (10.40%) Associate’s 4 (2.20%) 2 (3.30%) 2 (2.60%) Bachelor’s 41 (29.90%) 12 (20.0%) 29 (37.70%) Master’s 48 (35.0%) 21 (35.0%) 27 (35.10%) Professional/Doctoral 30 (21.90%) 22 (36.70%) 8 (10.40%) Decline 2 (1.50%) 0 (0.0%) 2 (2.60%) Ethnicity, n (%) AI, AN 1 (0.70%) 0 (0.0%) 1 (1.30%) Asian 18 (13.10%) 5 (8.30%) 13 (16.90%) Black 4 (2.90%) 0 (0.0%) 4 (5.20%) Latinx 24 (17.50%) 6 (10.0%) 18 (23.40%) White 83 (60.60%) 49 (81.70%) 34 (44.20%) Other 5 (3.60%) 0 (0.0%) 5 (6.50%) Decline 2 (1.50%) 0 (0.0%) 2 (2.60%) Parity, n(%) Primiparous 113 (82.50%) 36 (60%) 77 (100%) Multiparious 24 (17.50%) 24 (40%) 0 (0.0%) Infant Sex Female 54 (39.40%) 25 (41.70%) 29 (37.70%) Male 63 (46.0%) 25 (41.70%) 38 (49.40%) Decline 20 (14.60%) 10 (16.70%) 10 (13.0%) Open in a new tab Note. * = Household income began being collected partway through the pre-pandemic study; therefore, most entries are missing. AI, AN = American Indian, Alaska Native. Primiparous = First-time mother. Figure 1. Timeline of Data Collection in the Pandemic Sample. Open in a new tab Note. Pandemic data collection began on April 6 th , 2020. All prepandemic participants who provided hair samples after shelter-in-place ordinances began in California, USA (i.e., March 19 th , 2020) were omitted from the analysis. Perceived Stress Perceived stress was measured using the Perceived Stress Scale (PSS), a well-validated 14-item scale of stress in the last month ( Cohen et al., 1994 ). Participants are asked to report their level of agreement with statements concerning their stress on a 5-point Likert-type scale ranging from “1” = Never to “5” = Very Often. Items include in the last month, how often have you felt nervous or stressed or felt that you were unable to control the important things in your life. The PSS was administered during pregnancy, at three months postpartum, and at six months postpartum. Reliability estimates were good for all three timepoints in the pandemic cohort (α = .87, .92, and .89, respectively) and the prepandemic cohort (α = .80, .90, and .87, respectively). Parenting Stress and Caregiving Responsibilities Parenting stress was measured using two separate scales – the Parenting Stress Scale ( Berry & Jones, 1995 ) and the Parenting Stress Index – Short Form (PSI-SF; Abidin, 1995 ). The Parenting Stress Scale was used at the three-month postpartum wave, whereas the PSI-SF was used at the six-month postpartum wave. Additionally, we included two separate items to assess caregiving responsibilities. The first item was how much time do you spend alone with your baby as the main caregiver on a 5-point Likert scale ranging from “1” = Less than 25% of the time to “5” = All or most (at least 75%) of the time . This item was asked for both weekdays and weekends and we computed a weighted average of the two by multiplying the weekdays score by 5/7 and the weekends score by 2/7 and taking the sum. The second item concerning parenting responsibilities asked compared to your partner, what would you estimate is your share of overall infant care responsibilities, and participants responded on a 5-point Likert-type scale ranging from “1” = I do less than 20% and my partner does more than 80% and “5” = I do more than 80% and my partner does less than 20%. Parenting Stress Scale. The Parenting Stress Scale is an 18-item measure of parenting stress that asks participants to rate their agreement with each item on a 5-point Likert-scale with scoring ranging from “1” = Strongly Disagree to “5” = Strongly Agree ( Berry & Jones, 1995 ). Items include the major source of stress in my life is my child and I feel overwhelmed by the responsibilities of being a parent. This measure was administered at three months postpartum and demonstrated good reliability in our pandemic (α = .89) and prepandemic cohorts (α = .85). Parenting Stress Index. The PSI-SF is a 36-item questionnaire comprised of that assesses stress in the parent-child relationship ( Abidin, 1995 ). Respondents rate their agreement with each statement using a 5-point Likert-type scale. Consistent with the literature, we summed all items to generate composite score of parenting stress. The PSI-SF was only measured at six months postpartum and demonstrated strong reliability in the pandemic (α = .94) and prepandemic (α = .93) cohorts. Depressive Symptoms Depressive symptoms were measured using the Beck Depression Inventory-II (BDI), a well-validated 21-item questionnaire that measures the severity of depression symptoms including sadness, pessimism, and loss of pleasure in the last two weeks ( Beck et al., 1996 ). Participants rate their agreement with each item on a 4-point Likert type-scale. In compliance with our university’s (IRB) guidelines, we exclude one question concerning suicidal ideation. This measure was administered during pregnancy, at three months postpartum, and at six months postpartum. Reliability estimates were good for all three timepoints in the pandemic cohort (α = .91, .93, and .87, respectively) and the prepandemic cohort (α = .84, .91, and .90, respectively). Data Analysis Regularized Regression, Random Forest, and Gradient Boosting Machines All analyses will be performed using R version 4.4.3 (R Core Team, 2025). We will use the ‘h2o’ package version 3.44.0.3 to fit all machine learning models ( Fryda et al., 2023 ). The ‘iml’ package version 0.11.1 ( Molnar et al., 2018 ) will be used to derive permutation variable importance measures and interaction metrics (i.e., Friedman’s H-statistic; Friedman et al., 2008 ), partial dependence plots, and two-way interaction plots. Randomization testing will be used to construct null distributions and derive test statistics and empirical p -values for all variance importance measures and H-statistics ( Altmann et al., 2010 ). No analyses have been conducted with our collected dataset and will not begin until the conditional acceptance of this registered report. Prior to analysis, we intend to truncate HCC values ± 3SD from the mean for mothers’ and infants’, respectively, to control for extreme outliers. Next, we will log-transform both mothers’ and infants’ HCC due to previous work that has employed the same transformation to control positive skew in HCC data (e.g., Tucker-Drob et al., 2017 ). Normalizing our data is expected to increase model performance by decreasing sparseness in the data ( James et al., 2021 ). Supplementary Material Supinfo NIHMS2096943-supplement-Supinfo.docx (788.9KB, docx) Table 2. Descriptive Statistics Variable Both Samples Pandemic Pre-Pandemic n M SD n M SD n M SD * Maternal HCC, Mean (SD) 129 16.88 24.03 56 12.87 24.34 73 19.95 23.50 * Infant HCC, Mean (SD) 128 748.43 1544.13 57 1230.55 2079.73 71 361.38 724.17 Maternal HCC (Log), Mean (SD) 129 2.24 1.07 56 1.74 1.15 73 2.61 0.84 Infant HCC (Log), Mean (SD) 128 4.65 2.13 57 4.70 2.58 71 4.61 1.69 Maternal Age, Mean (SD) 136 33.45 4.70 60 36.03 3.99 86 31.41 4.21 Infant Age, Weeks 130 27.75 2.44 53 26.62 0.63 87 28.54 2.89 Infant Birthweight, lbs. 132 6.92 1.07 55 7.00 1.26 87 6.86 0.92 Perceived Stress (PN) 135 3.35 0.30 59 3.36 0.29 86 3.34 0.30 Perceived Stress (3PP) 120 3.23 0.24 54 3.25 0.25 76 3.21 0.24 Perceived Stress (6PP) 130 3.24 0.31 53 3.23 0.34 87 3.24 0.28 Depressive Symptoms (PN) 135 0.57 0.34 59 0.66 0.39 86 0.49 0.26 Depressive Symptoms (3PP) 121 0.54 0.39 55 0.60 0.43 76 0.49 0.35 Depressive Symptoms (6PP) 130 0.55 0.35 53 0.54 0.34 87 0.56 0.36 Parenting Stress (3PP) 120 3.42 0.31 54 3.46 0.31 76 3.39 0.31 Parenting Stress (6PP) 130 4.21 0.49 53 0.54 0.34 87 4.28 0.43 Open in a new tab Note. HCC = Hair cortisol concentration; PN = Prenatal; 3PP = Three months postpartum; 6PP = Six months postpartum. Lbs. = Pounds; Log = Logarithmic transformation * = Maternal and infant HCC have been truncated all values >3SD above the mean. Funding: This research was funded by awards from the National Science Foundation (NSF) (Career Award #1552452) and the National Institutes of Health (NIH-NICHD R01 HD104801) to Darby Saxbe, and the National Institutes of Health (NIH) National Institute of Child Health and Human Development (NICHD) Research Supplement to Promote Diversity in Health-Related Research (PA-21–071) and Predoctoral Ford Fellowship awarded by the National Academies of Sciences, Engineering, and Medicine to Gabriel León. Additional support was provided by the University of Southern California (USC) Center for the Changing Families: Small Grants for COVID-19, the USC Zumberge Special Solicitation: Epidemic & Virus-Related Research and Development Award, and the Social Science Research Council Award. Footnotes Ethical Approval: All study procedures were approved by the USC Institutional Review Board and all human subjects completed informed consent. Conflicts of interest: The authors declare no conflict of interest regarding authorship or publication of this work. Data Availability: Following Stage 1 in principle acceptance, we intend to register our approved protocol on the Open Science Framework and will make publicly available the materials, methods, and data necessary to reproduce our reported results. References Abidin RR (1995). Parenting Stress Index, Third Edition: Professional Manual. Odessa, FL: Psychological Assessment Resources, Inc. [ Google Scholar ] Altmann A, Tolosi L, Sander O, & Lengauer T (2010). Permutation importance: A corrected feature importance measure. Bioinformatics, 26(10), 1340–1347. doi: 10.1093/bioinformatics/btq134 [ DOI ] [ PubMed ] [ Google Scholar ] Beck A, Steer R, Brown G (1996). Manual for the Beck Depression Inventory-II. Psychological Corporation, San Antonio, TX. [ Google Scholar ] Berry JO, & Jones WH (1995). The Parental Stress Scale: Initial psychometric evidence. Journal of Social and Personal Relationships, 12(3), 463–472. [ Google Scholar ] Breiman L (2001). Random forests. Machine Learning, 45, 5– 32. 10.1023/a:1010933404324. [ DOI ] [ Google Scholar ] Broeks CW, Choenni V, Kok R, van der Voorn B, de Kruijff I, van den Akker ELT, van Rossum EFC, Hoogendijk WJG, Hillegers MHJ, Kamperman AM, & Lambregtse-Van den Berg MP (2021). BJPsych Open, 7(1), 1–9. doi: 10.1192/bjo.2020.159 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Bruinhof N, Vacaru SV, van den Heuvel MI, de Weerth C, & Beijers R (2022). Prenatal hair cortisol concentrations during the COVID-19 outbreak: Associations with maternal psychological stress and infant temperament. Psychoneuroendocrinology, 144, 105863. 10.1016/j.psyneuen.2022.105863 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Cantave CY, Brendgen M, Lupien S, Dionne G, Vitaro F, Boivin M, & Ouellet-Morin I (2022). Association between the timing of family socioeconomic deprivation and adolescence hair cortisol among adolescent twins: A study of the genetic and environmental processes involved. Psychoneuroendocrinology, 144, 105889. 10.1016/j.psyneuen.2022.105889 [ DOI ] [ PubMed ] [ Google Scholar ] Center for Disease Control (July, 2024). CDC Museum COVID-19 Timeline. https://www.cdc.gov/museum/timeline/covid19.html Cohen S, Kamarck T, & Mermelstein R (1994). Perceived stress scale. Measuring Stress: A Guide for Health and Social Scientists, 10, 1–2. [ Google Scholar ] Davis M, West K, Blims J, Morelen D & Suveg C (2018). A systematic review of parent-child synchrony: It is more than skin deep. Developmental Psychobiology, 60, 674–691. DOI: 10.1002/dev.21743 [ DOI ] [ PubMed ] [ Google Scholar ] Dickerson SS, & Kemeny ME (2004). Acute stressors and cortisol responses: a theoretical integration and synthesis of laboratory research. Psychological Bulletin, 130(3), 355–391. 10.1037/0033-2909.130.3.355 [ DOI ] [ PubMed ] [ Google Scholar ] Doan SN, Ding M, Burniston AB, Smiley PA, Chow CM, & Liu CH (2022). Changes in maternal depression and children’s behavior problems: Investigating the role of COVID-19-related stressors, hair cortisol, and dehydroepiandrosterone. Clinical Psychological Science, 10(6), 1098–1110. 10.1177/21677026221076845 [ DOI ] [ Google Scholar ] Fife DA & D’Onofrio J (2023). Common, uncommon, and novel applications of random forest in psychological research. Behavioral Research Methods, 55, 2447–2466. 10.3758/s13428-022-01901-9 [ DOI ] [ PubMed ] [ Google Scholar ] Friedman H (2001). Greedy function approximation: A gradient boosting machine. Annals of Statistics. 1189–1232. [ Google Scholar ] Friedman J, Bogdan H, & Popescu E (2008). Predictive learning via rule ensembles.The Annals of Applied Statistics 2(3), 916–954. [ Google Scholar ] Fryda T, LeDell E, Gill N, Aiello S, Fu A, Candel A, … & Wong W. (2023). h2o: R Interface for the ‘H2O’ Scalable Machine Learning Platform. R package version 3.44.0.3. https://github.com/h2oai/h2o-3 [ Google Scholar ] Galbally M, van Rossum EFC, Watson SJ, de Kloet ER, & Lewis AJ (2019). Trans-generational stress regulation: Mother-infant cortisol and maternal mental health across the perinatal period. Psychoneuroendocrinology, 109, 104374. 10.1016/j.psyneuen.2019.104374 [ DOI ] [ PubMed ] [ Google Scholar ] Gao W, Stalder T, Foley P, Rauh M, Deng H, & Kirschbaum C (2013). Quantitative analysis of steroid hormones in human hair using a column-switching LC–APCI–MS/MS assay. Journal of Chromatography B, 928, 1–8. 10.1016/j.jchromb.2013.03.008 [ DOI ] [ PubMed ] [ Google Scholar ] Goldstein A, Kapelner A, Bleich J, & Pitkin E (2015). Peeking inside the black box: Visualizing statistical learning with plots of individual conditional expectation. Journal of Computational and Graphical Statistics, 24(1), 44–65. [ Google Scholar ] Greenwell B (2018). Pdp: Partial dependence plots. https://CRAN.R-project.org/package=pdp . [ Google Scholar ] Greff MJ, Levine JM, Abuzgaia AM, Elzagallaai AA, Rieder MJ, & van Uum SH (2019). Hair cortisol analysis: An update on methodological considerations and clinical applications. Clinical Biochemistry, 63, 1–9. 10.1016/j.clinbiochem.2018.09.010 [ DOI ] [ PubMed ] [ Google Scholar ] Henninger M, Debelak R, Rothacher Y, & Strobl C (2023). Interpretable machine learning for psychological research: Opportunities and pitfalls. Psychological Methods. Advance online publication. 10.1037/met0000560 [ DOI ] [ PubMed ] [ Google Scholar ] Howland MA, Sandman CA, & Glynn LM (2017). Developmental origins of the human hypothalamic-pituitary-adrenal axis. Expert Review of Endocrinology & Metabolism, 12(5), 321–339. 10.1080/17446651.2017.1356222 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] James G, Witten D, Hastie T, & Tibshirani R (2021). An introduction to statistical learning: With applications in R (2nd Ed.). New York: Springer. [ Google Scholar ] James G, Witten D, Hastie T, & Tibshirani R (2021). An introduction to statistical learning: With applications in R (2nd Ed.). New York: Springer. [ Google Scholar ] Kalliokoski O, Jellestad FK, & Murison R (2019). A systematic review of studies utilizing hair glucocorticoids as a measure of stress suggests the marker is more appropriate for quantifying short-term stressors. Scientific Reports, 9(1), 11997. 10.1038/s41598-019-48517-2 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Kapoor A, Schultz-Darken N, & Ziegler TE (2018). Radiolabel validation of cortisol in the hair of rhesus monkeys. Psychoneuroendocrinology, 97, 190–195. 10.1016/j.psyneuen.2018.07.022 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Karl M, Huth V, Schalicke S, Muller-Stark C, Weise V, Mack JT, Kirschbaum C, Weidner K, Garthus-Niegel S, & Steudte-Schmiedgen S (2023). The association between maternal symptoms of depression and hair glucocorticoids in infants across the perinatal period. Psychoneuroendocrinology, 147(105952), 1–10. 10.1016/j.psyneuen.2022.105952 [ DOI ] [ PubMed ] [ Google Scholar ] Khoury JE, Giles L, Kaur H, Johnson D, Gonzalez A, & Atkinson L (2023). Associations between psychological distress and hair cortisol during pregnancy and the early postpartum: A meta-anlysis. Psychoneuroendocrinology, 147(105969). 10.1016/j.psyneuen.2022.105969 [ DOI ] [ PubMed ] [ Google Scholar ] King LS, Humphreys KL, Cole DA, & Gotlib IH (2022). Hair cortisol concentration across the peripartum period: Documenting changes and associations with depressive symptoms and recent adversity. Comprehensive Psychoneuroendocrinology, 9, 100102. 10.1016/j.cpnec.2021.100102 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Kirschbaum C, Tietze A, Skoluda N, & Dettenborn L (2009). Hair as a retrospective calendar of cortisol production—increased cortisol incorporation into hair in the third trimester of pregnancy. Psychoneuroendocrinology, 34(1), 32–37. [ DOI ] [ PubMed ] [ Google Scholar ] Klusmann H, Schulze L, Engel S, Bücklein E, Daehn D, Lozza-Fiacco S, ... & Schumacher S. (2022). HPA axis activity across the menstrual cycle-a systematic review and meta-analysis of longitudinal studies. Frontiers in Neuroendocrinology, 66, 100998. [ DOI ] [ PubMed ] [ Google Scholar ] Laurent H (2017). Early calibration of the HPA axis by maternal psychopathology. Psychoneuroendocrinology, 78, 177–184. 10.1016/j.psyneuen.2017.01.034 [ DOI ] [ PubMed ] [ Google Scholar ] León GA, Waizman Y, Cardenás S, Aviv L, Newsome P, Vaccaro A, … & Saxbe DE. (2025). Trajectories of mothers’ postpartum depressive symptoms following COVID-19 pandemic lockdowns and the protective role of relationship quality. Journal of Clinical Science and Psychopathology. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Liu CH, Fink G, Brentani H, & Brentani A (2017). An assessment of hair cortisol among postpartum Brazilian mothers and infants from a high-risk community in Sao Paulo: Intra-individual stability and association in mother-infant dyads. Developmental Psychobiology, 59, 916–926. DOI: 10.1002/dev.21557 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Liu CH, Snidman N, Leonard A, Meyer J, & Tronick E (2016). Intra-individual stability and developmental change in hair cortisol among postpartum mothers and infants: Implications for understanding chronic stress. Developmental Psychbiology, 58(4), 509–518. doi: 10.1002/dev.21394. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Manenschijn L, van Kruysbergen RGPM, de Jong FH, Koper JW, van Rossum EFC (2011). Shift work at young age is associated with elevated long-term cortisol levels and body mass index. The Journal of Clinical Endocrinology & Metabolism, 11(1), E1862–E1865. 10.1210/jc.2011-1551 [ DOI ] [ PubMed ] [ Google Scholar ] Marceau K, Wang W, Robertson O, & Shirtcliff EA (2020). A systematic review of hair cortisol during pregnancy: Reference ranges and methodological considerations. Psychoneuroendocrinology, 122, 104904. 10.1016/j.psyneuen.2020.104904 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] McEwen BS (2004). Protection and damage from acute and chronic stress: Allostasis and allostatic overload and relevance to the pathophysiology of psychiatric disorders. Annals of the New York Academy of Sciences, 1032(1), 1–7. [ DOI ] [ PubMed ] [ Google Scholar ] McEwen BS (2004). Protection and damage from acute and chronic stress: Allostasis and allostatic overload and relevance to the pathophysiology of psychiatric disorders. Annals of the New York Academy of Sciences, 1032(1), 1–7. [ DOI ] [ PubMed ] [ Google Scholar ] Mollard E, Kupzyk K, & Moore T (2021). Postpartum stress and protective factors in women who gave birth in the United States during the COVID-19 pandemic. Women’s Health, 17, 1 – 10. 10.1177/17455065211042190 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Molnar C, Bischl B, Casalicchio G (2018). iml: An R package for Interpretable Machine Learning. Journal of Open Source Software, 3(26), 786. doi: 10.21105/joss.00786 [ DOI ] [ Google Scholar ] Morris AR & Saxbe DE (2022). Mental health and prenatal bonding in pregnant women during the COVID-19 pandemic: Evidence for heightened risk compared with a prepandemic sample. Clinical Psychological Science, 10(5), 846–855. Doi: 10.1177/21677026211049430 [ DOI ] [ Google Scholar ] Morris AR & Saxbe DE (2023). Differences in infant negative affectivity during the COVID-19 pandemic. Infant Mental Health Journal, 44, 466–479. Doi: 10.1002/imhj.22061 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Morris A, Traube D, Lakshmanan A West A., & Saxbe D. (2020). Expecting A Baby During the COVID-19 Pandemic: Mental Health Concerns in Pregnant Women Warrant New Treatment Approaches. Research Brief, University of Southern California Center for the Changing Family. [ Google Scholar ] Mustonen P, Karlsson L, Kataja E, Scheinin NM, Kortesluoma S, Coimbra B, Rodrigues AJ, Sousa N, & Karlsson H (2019). Maternal prenatal hair cortisol is associated with prenatal depressive symptom trajectories. Psychoneuroendocrinology, 109, 104383. 10.1016/j.psyneuen.2019.104383 [ DOI ] [ PubMed ] [ Google Scholar ] Nyström‐Hansen M, Andersen MS, Khoury JE, Davidsen K, Gumley A, Lyons-Ruth K, MacBeth A, & Harder S (2019). Hair cortisol in the perinatal period mediates associations between maternal adversity and disrupted maternal interaction in early infancy. Developmental Psychobiology, 61, 543–556. DOI: 10.1002/dev.21833 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Ritschel L, Streit F, Zhu G, McAloney K, Frank J, Couvy-Duchesne B, … Ritschel M (2017). Hair cortisol in twins: Heritability and genetic overlap with psychological variables and stress-system genes. Scientific Reports, 7(15351), 1–15. DOI: 10.1038/s41598-017-11852-3 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Romero-Gonzalez B, Caparros-Gonzalez RA, Gonzalez-Perez R, Delgado-Puertas P, & Peralta-Ramirez MI (2018). Newborn infants’ hair cortisol levels reflect chronic maternal stress during pregnancy. PloS ONE, 13(7), e0200279. 10.1371/journal.pone.0200279 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Saxbe D, Khaled M, Horton KT, & Mendez AJ (2019). Maternal prenatal plasma oxytocin is positively associated with prenatal psychological symptoms, but method of immunoassay extraction may affect results. Biological Psychology, 147, 107718. 10.1016/j.biopsycho.2019.107718 [ DOI ] [ PubMed ] [ Google Scholar ] Saxbe D, Martínez‐Garcia M, Cardenas SI, Waizman Y, & Carmona S (2023). Changes in left hippocampal volume in first‐time fathers: Associations with oxytocin, testosterone, and adaptation to parenthood. Journal of Neuroendocrinology, 35(7), e13270. 10.1111/jne.13270 [ DOI ] [ PubMed ] [ Google Scholar ] Schalinski I, Teicher MH, & Rockstroh B (2019). Early neglect is a key determinant of adult hair cortisol concentration and is associated with increased vulnerability to trauma in a transdiagnostic sample. Psychoneuroendocrinology, 108, 35–42. 10.1016/j.psyneuen.2019.06.007 [ DOI ] [ PubMed ] [ Google Scholar ] Short SJ, Stalder T, Marceau K, Entringer S, Moog NK, Shirtcliff EA, Wadhwa PD, Buss C (2016). Correspondence between hair cortisol concentrations and 30-day integrated daily salivary and weekly urinary cortisol measures. Psychoneuroendocrinology, 71, 12–18. 10.1016/j.psyneuen.2016.05.007 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Solorzano CS, Serwinski B, Grano C, & Steptoe A (2023). Longitudinal association between saliva and hair cortisol concentration: A systematic comparison. Psychoneuroendocrinology, 156, 106340. 10.1016/j.psyneuen.2023.106340 [ DOI ] [ PubMed ] [ Google Scholar ] Stalder T, Steudte-Schmiedgen S, Alexander N, Klucken T, Vater A, Wichmann S, Kirschbaum C, & Miller R (2017). Stress-related and basic determinants of hair cortisol in humans: A meta-analysis. Psychoneuroendocrinology, 77, 261–274. 10.1016/j.psyneuen.2016.12.017 [ DOI ] [ PubMed ] [ Google Scholar ] Staufenbiel SM, Penninx BWJH, de Rijke YB, van den Akker ELT & van Rossum EFC (2015). Determinants of hair cortisol and hair cortisone concentrations in adults. Psychoneuroendocrinology, 60, 182–194. 10.1016/j.psyneuen.2015.06.011 [ DOI ] [ PubMed ] [ Google Scholar ] Stoye DQ, Sullivan G, Galdi P, Kirschbaum C, Lamb GJ, Black GS, Evans MJ, Boardman JP, & Reynolds RM (2021). Perinatal determinants of neonatal hair glucocorticoid concentrations. Psychoneuroendocrinology, 128, 105223. 10.1016/j.psyneuen.2021.105223 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Sugaya N, Izawa S, Ogawa N, Shirotsuki K, & Nomura S (2020). Association between hair cortisol and diurnal basal cortisol levels: A 30-day validation study. Psychoneuroendocrinology, 116, 104650. 10.1016/j.psyneuen.2020.104650 [ DOI ] [ PubMed ] [ Google Scholar ] Tarullo AR, John AMS, & Meyer JS (2017). Chronic stress in the mother-infant dyad: Maternal hair cortisol, infant salivary cortisol and interactional synchrony. Infant Behavior and Development, 47, 92–102. 10.1016/j.infbeh.2017.03.007 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Thomason E, Volling BL, Flynn HA, McDonough SC, Marcus SM, Lopez JF, & Vazquez DM (2014). Parenting stress and depressive symptoms in postpartum mothers: Bidirectional or unidirectional effects?. Infant Behavior & Development, 37, 406–415. 10.1016/j.infbeh.2014.05.009 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Tucker-Drob EM, Grotzinger AD, Briley DA, Engelhardt LE, Mann FD, Patterson M, Kirschbaum C, Adam EK, Church JA, Tackett JL, Harden KP (2017). Genetic influences on hormonal markers of chronic hypothalalic-pituitary-adrenal function in human hair. Psychological Medicine, 47, 1389–1401. doi: 10.1017/S0033291716003068 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] van Ockenburg SL, Schenk HM, van der Veen A, van Rossum EFC, Kema IP, Rosmalen JGM (2016). The relationship between 63 days of 24-h urinary free cortisol and hair cortisol levels in 10 healthy individuals. Psychoneuroendocrinology, 73, 142–147. 10.1016/j.psyneuen.2016.07.220 [ DOI ] [ PubMed ] [ Google Scholar ] Venables WN & Ripley BD (2002) Modern Applied Statistics with S. Fourth Edition. Springer, New York. ISBN 0–387-95457–0 [ Google Scholar ] Waters SF, West TV, & Mendes WB (2014). Stress contagion: Physiological covariation between mothers and infants. Psychological Science, 25(4), 934–942. 10.1177/0956797613518 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Waters SF, West TV, Karnilowicz HR, & Mendes WB (2017). Affect contagion between mothers and infants: Examining valence and touch. Journal of Experimental Psychology: General, 146(7), 1043–1051. 10.1037/xge0000322 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Young C, Roberts R, & Ward L (2019). Application of resilience theories in the transition to parenthood: A scoping review. Journal of Reproductive and Infant Psychology, 37(2), 139–160. 10.1080/02646838.2018.1540860 [ DOI ] [ PubMed ] [ Google Scholar ] Yavorsky JE, Kamp Dush CM, & Schoppe-Sullivan S (2015). The production of inequality: The gender division of labor across the transition to parenthood. Journal of Marriage and Family, 77, 662–679. DOI: 10.1111/jomf.12189 [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] Associated Data This section collects any data citations, data availability statements, or supplementary materials included in this article. Supplementary Materials Supinfo NIHMS2096943-supplement-Supinfo.docx (788.9KB, docx) Data Availability Statement Following Stage 1 in principle acceptance, we intend to register our approved protocol on the Open Science Framework and will make publicly available the materials, methods, and data necessary to reproduce our reported results. ACTIONS View on publisher site PDF (973.1 KB) Cite Collections Permalink RESOURCES Similar articles Cited by other articles Links to NCBI Databases Follow NCBI NCBI on X (formerly known as Twitter) NCBI on Facebook NCBI on LinkedIn NCBI on GitHub NCBI RSS feed Connect with NLM NLM on X (formerly known as Twitter) NLM on Facebook NLM on YouTube National Library of Medicine 8600 Rockville Pike Bethesda, MD 20894 Web Policies FOIA HHS Vulnerability Disclosure Help Accessibility Careers NLM NIH HHS USA.gov Back to Top

Related documents

• Mental Health Prediction Using Machine Learning#404893
• A unified machine learning framework for gestational diabetes mellitus diagnosis#377980
• Machine learning-optimized perinatal depression screening: Maximum impact, minimal burden#682334
• Predicting Psychiatric Readmission of Children and Adolescents Using Machine Learning.#388741
• Machine Learning-Based Prediction of Antepartum Depression Symptoms: A Prospective Cohort Study.#615941