Associations between ventricular size and temperament at age 3 months observed in 912 infants*
| Ventricular assessment (per unit of SD) | Irritability or anger, unsettled–irregularity (SD)† | Orienting, alertness–responsiveness (SD)†‡ | ||||||
|---|---|---|---|---|---|---|---|---|
| β (95% CI)§ | p¶ | R2** | F†† | β (95% CI)§ | p¶ | R2** | F†† | |
| Prenatal | ||||||||
| Lateral ventricle in midpregnancy | −0.01 (−0.09 to 0.07) | 0.8 | 0.29 | 5.0 | −0.04 (−0.12 to 0.04) | 0.3 | 0.22 | 2.7 |
| Lateral ventricle in late pregnancy | −0.02 (−0.09 to 0.04) | 0.5 | 0.30 | 9.4 | 0.03 (−0.03 to 0.10) | 0.3 | 0.27 | 7.5 |
| Postnatal | ||||||||
| Lateral ventricles at age 6 weeks‡‡ | −0.08 (−0.16 to −0.00) | 0.05 | 0.30 | 6.8 | 0.10 (0.01 to 0.18) | 0.02 | 0.28 | 5.5 |
CI = confidence interval; SD = standard deviation.
↵* The number of infants varied per measurement of ventricular size: 504 in midpregnancy, 894 in late pregnancy and 580 at age 6 weeks.
↵† Higher scores on alertness–responsiveness indicated less difficult behaviour, whereas a higher score on unsettled–irregularity indicated more difficult behaviour..
↵‡ We used square root transformation for scores on this scale to satisfy the assumption of linearity.
↵§ Values are regression coefficients (and 95% CIs) from multivariable linear regression and reflect the difference in z-score per unit of standard deviation in the measurements of the atrial width in lateral ventricles or in the measurements postnatal ventricular volume associated with temperamental scores. Models are adjusted for age, sex, gestational duration, circumference of the head at the time of measurement, maternal age, maternal educational level and maternal psychopathology
↵¶ P value of ventricular size.
↵** Coefficient of multiple determination. This coefficient applies to the complete model.
↵†† F-statistic with 9 degrees of freedom in the numerator. Degrees of freedom in the denominator were 497, 882 and 570 for irritability or anger, and 498, 885 and 571 for orienting. These F-statistics and degrees of freedom apply to the complete models, including covariates.
↵‡‡ We used log transformation for ventricular volumes to normalize the distribution.