Maternal psychological distress associates with alterations in resting‐state low‐frequency fluctuations and distal functional connectivity of the neonate medial prefrontal cortex

Abstract Prenatal stress exposure (PSE) has been observed to exert a programming effect on the developing infant brain, possibly with long‐lasting consequences on temperament, cognitive functions and the risk for developing psychiatric disorders. Several prior studies have revealed that PSE associates with alterations in neonate functional connectivity in the prefrontal regions and amygdala. In this study, we explored whether maternal psychological symptoms measured during the 24th gestational week had associations with neonate resting‐state network metrics. Twenty‐one neonates (nine female) underwent resting‐state fMRI scanning (mean gestation‐corrected age at scan 26.95 days) to assess fractional amplitude of low‐frequency fluctuation (fALFF) and regional homogeneity (ReHo). The ReHo/fALFF maps were used in multiple regression analysis to investigate whether maternal self‐reported anxiety and/or depressive symptoms associate with neonate functional brain features. Maternal psychological distress (composite score of depressive and anxiety symptoms) was positively associated with fALFF in the neonate medial prefrontal cortex (mPFC). Anxiety and depressive symptoms, assessed separately, exhibited similar but weaker associations. Post hoc seed‐based connectivity analyses further showed that distal connectivity of mPFC covaried with PSE. No associations were found between neonate ReHo and PSE. These results offer preliminary evidence that PSE may affect functional features of the developing brain during gestation.


Effects of additional independent variables on primary models (sensitivity analyses): a. Composite score and neonate birth weight
In this multiple regression model, neonate birth weight was set as a 4 th independent variable (IV) of no interest. Otherwise, we used prior default IVs: Neonate age at scanning (days), neonate sex and maternal pre-pregnancy BMI. The complete model thus consisted of the beforementioned IVs and composite score as the main explanatory variable (EV).
In this model, the effect observed in our main analysis of maternal composite score on neonate fALFF maps was reduced to statistical insignificance at p < 0.001 and p < 0.005 levels. No clusters that passed the multiple comparison statistical thresholding were detected.
Bivariate correlation analysis performed in SPSS revealed a significant correlation between neonate birth weight and composite score (rs = -0.606). Variance of inflation (VIF) analysis revealed no indications of multicollinearity (VIF = 1.479). Even though there is an established link between neonate birth weight and exposure to prenatal stress (1), there is little reason to believe that infant birth weight itself would be the driving factor for effects seen in neonate fALFF maps. Nevertheless, to test this possibility we generated an additional model with infant birth weight as the EV and with neonate age at scanning, sex, and maternal pre-pregnancy BMI as the IVs. No statistically significant effects at p < 0.001 or p < 0.005 FWE-corrected level were obtained in this model.
We conclude that the mitigation of results in our main model, when corrected for neonate birth weight, was caused by the high negative correlation between neonate birth weight and composite score.

b. Composite score and maternal age in years
Here, the multiple regression design consisted of four IVs of no interest: Neonate age at scanning (days), neonate sex, maternal pre-pregnancy BMI and maternal age in years. Composite score was set as the main EV.
The effects were reduced to statistical insignificance at p < 0.001 and p < 0.005 levels. The performed bivariate correlation analysis revealed that maternal age in years had a significant correlation with maternal pre-pregnancy BMI (rs = 0.570), but no significant correlation was observed between composite score and maternal age. VIF analysis showed no indications of multicollinearity (VIF = 1.549) in this model.
Maternal age has been established to associate with levels of mental distress during pregnancy (2). To test whether maternal age had an independent effect on neonate fALFF maps, we performed another analysis with neonate age at scanning, neonate sex and maternal pre-pregnancy BMI as IVs. In this model, maternal age was set as the main EV. Here, we found a statistically significant (at p < 0. Out of the five papers (3-7) investigating PSE effects on neonate rs-fMRI metrics, only one controlled for maternal age at beginning of pregnancy (4), with similar maternal age distribution as in our sample. They found no association between maternal age and neonate FC maps. Considering extant literature and the results of these sensitivity analyses, we cannot rule out that the effects of maternal psychological distress on offspring brain development might depend on maternal age.

Exclusion analysis (N=18)
To make sure the results of our main model were not influenced by exposure to illicit substances and/or alcohol, we performed an additional analysis, in which the exposed subjects were excluded.

a. Composite score model
In this model, we excluded the three subjects that were exposed to illicit substances (cannabis) and/or alcohol in utero, yielding a sample size of 18 subjects. Otherwise, identical design was used as in the main parametric model with neonate age at scanning, neonate sex and maternal pre-pregnancy BMI set as IVs. The composite score was set as the main EV. Statistical significance threshold was set to p < 0.001.
We obtained near-identical results with the exclusion analysis as with our main model with 21 subjects. A statistically significant effect was observed in the neonate mPFC (p < 0.001 uncorrected; p < 0.001 FWE-corrected; kE 796). Here, the cluster shape was slightly altered (supplementary materials, figure 2) and fragmented into two separate clusters.
Although minimally altered, this model yielded a highly comparable result to our main model with 21 subjects. In all cases, exposure to alcohol and/or illicit substances was mild. After consideration, we decided to include the three subjects exposed to alcohol for increased statistical power in our main model.

Non-parametric main model results (SnPM13)
To test the validity of the underlying assumptions in the main parametric model, we repeated the analysis using non-parametric permutation testing with the Statistical Non-Parametric Mapping software (SnPM13). As the parametric model, the non-parametric model had the following measures set as IVs: neonate age at scanning, neonate sex and maternal pre-pregnancy BMI. Composite score was set as the main EV. Statistical significance threshold was set to p < 0.001.
We found identical results as with the parametric model. The composite score -fALFF effect localized to the neonate mPFC (p < 0.001 uncorrected; p < 0.001 FWE-corrected; kE 794). No additional statistically significant clusters nor negative associations were observed.

Supplementary materials, table 1.
Correlation matrix for metrics involved in this study. rs = Spearman rank correlation coefficient. * and ** denote statistically significant correlation at p < 0.05 and p < 0.01 levels, respectively.

Maternal age in years
Neonate birth weight