LMP, sociodemographic variables, pregnancy complications, and delivery characteristics are controlled for (Alexander et al., 1992).

Although some of these postnatal gestational age measures are extensively used to estimate gestational age at birth, first- and second-trimester ultrasounds are far more accurate at estimating gestational age (Alexander et al., 1990, 1992; Mitchell, 1979; Wariyar et al., 1997). In a comparison study, Wariyar et al. (1997) found that ultrasound before 20 weeks of gestation was the most accurate (95% confidence interval = ±9 days, whereas the 95% confidence interval = ±17 days for postnatal methods). For preterm infants with gestational ages of less than 30 weeks, an ultrasound performed before 20 weeks of gestation was more accurate than an ultrasound performed at or after 20 weeks of gestation for determination of gestational age at birth (95% confidence intervals = ±9 days and ±15 days, respectively), the New Ballard Score (95% confidence interval = ±24 days), and the Dubowitz gestational age assessment (95% confidence interval = ±34 days).

The difficulty of using postnatal measures of degree of maturation of external physical characteristics and neurological muscle tone to estimate gestational age at birth highlights the difference between pregnancy duration and degree of maturation (Allen, 2005a). Although conceptually the use of the words “gestational age” implies a time interval, duration of pregnancy, the measurement of gestational age has historically involved either measures of fetal or infant size, or measures of degree of infant maturation. Since degree of fetal maturation plays an important role in infant mortality and morbidity rates, and may play a role in the signaling mechanisms for the normal initiation of labor at term, clarity in how gestational age is defined and determined is essential for understanding the mechanisms leading to preterm birth.

Neuromaturational changes in brain structural and functional development have been noted in preterm infants. These changes can be detected by detailed neurological examination, neuroimaging (especially cranial ultrasound), electroencephalography (EEG), amplitude-integrated EEG (a-EEG), electroretinography and neurophysiological measures of conduction time after auditory, visual, or tactile stimulation (Allen, 2005a; Amiel-Tison and Gosselin, 2001; Burdjalov et al., 2003; Finnstrom, 1972; Henderson-Smart et al., 1985; Kesson et al., 1985; Klimach and Cooke, 1988; Leaf et al., 1995; Miller et al., 1983; Olischar et al., 2004a,b). Prenatal ultrasounds have detected sonographic landmarks of normal fetal cortical development, which is important to know for the prenatal detection of fetal brain malformations (Perri et al., 2005).