Pregnancy entails a U-shaped trajectory in human brain structure linked to hormones and maternal attachment

Abstract

Growing evidence places the gestational period as a unique moment of heightened neuroplasticity in adult life. In this longitudinal study spanning pre, during, and post pregnancy, we unveil a U-shaped trajectory in gray matter (GM) volume, which dips in late pregnancy and partially recovers during postpartum. These changes are most prominent in brain regions associated with the Default Mode and Frontoparietal Network. The U-shaped trajectory is predominantly linked to gestational factors, as it only presents in gestational mothers and correlates with fluctuations in estrogens over time. Finally, the mother's mental health status mediates the relationship between postpartum GM volume recovery and maternal attachment at 6 months postpartum. This research sheds light on the complex interplay between hormones, brain development, and behavior during the transition to motherhood. It addresses a significant knowledge gap in the neuroscience of human pregnancy and opens new possibilities for interventions aimed at enhancing maternal health and well-being.

Fig. 1. Longitudinal cortical gray matter volume trajectory across pregnancy and postpartum (N = 179).

Longitudinal changes were derived from the group x session² fixed effect term of the adjusted linear mixed effect model: Cortical GM Volume ∼ Group + Session + Session² + Group*Session + Group*Session² + eTIV + Age + Euler + Inter-session Interval + (1 ∣ participant ID). A Mean percentage of cortical GM volume change per group and session in relation to the baseline (i.e., Session 1 - Pre-pregnancy Session). Error bars correspond to 95% confidence intervals around the sample’s mean at each experimental group and session, respectively. The gray shading corresponds to the pregnancy period. B Vertex-wise signed effect size maps (η_p²) of the group x session² interaction (q value < 0.05). Indicating larger quadratic effects (red) and smaller quadratic effects (blue) in gestational mothers than in nulliparous or non-gestational mothers. Signed effect size maps were projected to the inflated fsaverage template provided by the FreeSurfer software. Colors were collapsed to ±0.14, which indicated a large effect size. Gest, gestational mothers; nGest, non-gestational mothers; Null, nulliparous women.

Fig. 2. Cortical gray matter volume trajectories based on the functional location of the changes in gestational mothers (N = 179).

The different trajectories were assessed by parcellating the brain in Yeo’s seven large-scale functional brain networks. A Cortical parcellation of Yeo’s seven large-scale functional brain networks. B Mean percentage of cortical GM volume change per group and session and functional location in relation to the baseline (i.e., Session 1 - Pre-pregnancy Session). The gray shading corresponds to the pregnancy period. C Vertex-wise signed effect size maps (η_p²) of the group x session² interaction (q value < 0.05 and η_p² > 0.06 to capture medium to big effect sizes). Indicating larger quadratic effects (red) and smaller quadratic effects (blue) in gestational mothers than in nulliparous. Signed effect size maps were projected to the inflated fsaverage template provided by the FreeSurfer software. Colors were collapsed to ±0.14, which indicated a large effect size. Gest, gestational mothers; Null, nulliparous women. D Spin test for the signed effect sizes of the vertex-wise Group [gestational mothers vs nulliparous women]*Session² interaction in cortical volume within the seven large-scale functional brain networks. Black horizontal bars represent the observed values and the violin plots reflect the null distributions obtained using 1000 spin-permutations of the maps. The exact one-tailed p values are reported when p < 0.05. No multiple comparisons corrections were applied. The black dot on the center of the boxplot represents the median, the box encloses the lower and upper quartiles, and the whiskers extend to the minimum and maximum values within a range of 1.5 times the interquartile range.

Fig. 3. Longitudinal correlation between estrogens and cortical gray matter (GM) volume in gestational mothers (N = 100).

From our sample of 127 gestational mothers, 100 provided urine samples at sessions 1, 2, 3, and 4 and were therefore included in these analyses. A On the left Y-axis, there is the global percentage of cortical GM volume change in each session in relation to the baseline (i.e., Session 1 - Pre-pregnancy Session). On the right Y-axis, the relative levels in estriol 3-sulfate and estrone sulfate per session relative to the maximum change (% of max) are represented. Mean values are represented as circles and their 95%-confidence intervals as error bars. B Two-sided correlations between the quadratic parameter coefficients of GM volume change (Y-axis, extracted from the adjusted model: cortical GM volume change ∼ Session + Session² + (Session²∣ participant ID)) and the two steroid concentrations (X-axes, extracted from the adjusted model: Steroid concentration ∼ Session + Session² + (Session²∣ participant ID)). P-values were corrected within each metabolite family using FDR. The green lines and the gray shaded area represent the least squares regression lines and the 95% confidence intervals around the smooth line, respectively. Spearman’s R, Spearman’s correlation coefficient; p, uncorrected p-value; and q, hierarchical FDR corrected p-value.

Fig. 4. Longitudinal correlation between cortical gray matter volume, maternal attachment, and maternal well-being in gestational mothers (N = 98).

From our sample of 127 gestational mothers, 98 completed the MRI session at 6 months postpartum, and were therefore included in these analyses (see Fig. S7 for the dropout scheme). A Two-sided Pearson’s correlation between the percentage of cortical GM volume recovery during postpartum—from the 34th week of pregnancy to 6 months postpartum- (X-axis) and absence of hostility at 6 months postpartum (Y-axis). P-values are corrected for multiple testing, using FDR. The orange line and the gray shaded area represent the least squares regression line and the 95% confidence intervals around the sample’s mean at each experimental session, respectively. Pearson’s R, Pearson’s correlation coefficient; p, uncorrected p-value; and q, False-Discovery-Rate corrected p-value. B Path diagram of the mediation model between the percentage of cortical gray matter volume recovery (% GMV) during postpartum—from the 34th week of pregnancy to 6 months postpartum -, maternal well-being, and absence of hostility at 6 months postpartum. P-values used for the path diagram are uncorrected p-values. Numbers represent the coefficient estimates, asterisks indicate the significance of each pair of associations (*, p < 0.05; **, p < 0.01; ***, p < 0.001), and the positive symbol (+) indicates the positive association between each pair of variables.