Table 3.
Adjusted associations between maternal antenatal IPV exposure and its interaction with neonatal sex on microstructural measures for white matter tracts.
| IPV Mean (SD) | Control Mean (SD) | Unstandardized IPV β (SE) | IPV P value | Effect size, Cohen’s d (95% CI) | Unstandardized IPV × sex β (SE) | IPV × sex P value | Partial eta-squared (95% CI) | |
|---|---|---|---|---|---|---|---|---|
| Uncinate fasciculus | ||||||||
| FA | 0.274 (0.028) | 0.282 (0.021) |
-0.016 (0.008) |
0.057 | 0.31 (−0.18 to 0.79) |
0.014 (0.012) |
0.241 | 0.02 (0–0.15) |
| MD | 1.958 (0.094) |
1.925 (0.088) |
0.095 (0.029) |
0.002 *δ | -0.37 (−0.85 to 0.12) |
-0.124 (0.042) |
0.004 *δ | 0.13 (0.01–0.30) |
| Fornix | ||||||||
| FA | 0.318 (0.028) |
0.328 (0.022) |
-0.016 (0.008) |
0.064 | 0.41 (−0.09 to 0.90) |
0.010 (0.012) |
0.414 | 0.012 (0–0.121) |
| MD | 1.923 (0.090) |
1.906 (0.069) |
0.043 (0.024) |
0.078 | -0.22 (−0.70 to 0.26) |
-0.037 (0.035) |
0.293 | 0.02 (0–0.13) |
| Cingulum | ||||||||
| FA | 0.232 (0.024) |
0.240 (0.024) |
-0.010 (0.009) |
0.273 | 0.33 (−0.15 to 0.81) |
0.005 (0.013) |
0.695 | 0.01 (0–0.08) |
| MD | 1.773 (0.067) |
1.757 (0.058) |
0.018 (0.021) |
0.387 | -0.253 (−0.73 to 0.23) |
0.004 (0.030) |
0.898 | 0.01 (0–0.08) |
| Corpus callosum | ||||||||
| FA | 0.299 (0.024) |
0.311 (0.017) |
-0.012 (0.008) |
0.135 | 0.60 (0.11–1.09) |
-0.002 (0.011) |
0.846 | 0.01 (0–0.06) |
| MD | 1.095 (0.042) |
1.079 (0.038) |
0.041 (0.013) |
0.002 *δ | -0.42 (−0.90–0.06) |
-0.049 (0.019) |
0.010 * | 0.11 (0.01–0.27) |
| Corticospinal tract | ||||||||
| FA | 0.327 (0.036) |
0.338 (0.032) |
-0.034 (0.012) |
0.008 * | 0.32 (−0.16 to 0.81) |
0.044 (0.018) |
0.016 * | 0.10 (0.01–0.26) |
| MD | 1.580 (0.083) |
1.558 (0.075) |
0.043 (0.024) |
0.074 | -0.29 (−0.77 to 0.20) |
-0.029 (0.034) |
0.403 | 0.01 (0–0.12) |
IPV exposure (0 =below threshold; 1 = above threshold), infant sex (male = 0; female = 1) and their interaction, as predictors of neonatal diffusion outcomes. All models are adjusted for household income, maternal HIV status, maternal depressive symptoms, antenatal smoking and alcohol consumption, duration of gestation, and infant birthweight. Positive IPV regression coefficients indicate that IPV exposure is associated with higher diffusion values for that region.
Notes: Diffusion coefficients are averaged across hemispheres (×10-3 mm2/s).
*Uncorrected p < 0.05.
δDTI comparisons to survive multiple comparison correction using the false discovery rate across the 5 regions and 2 metrics, which generated a corrected overall p-value of 0.005.
Abbreviations: β, unstandardized beta coefficient; CI, confidence interval; IPV, in utero exposure to intimate partner violence; MD, mean diffusivity; FA, fractional anisotropy.