Deep grey matter growth predicts neurodevelopmental outcomes in very preterm children

NeuroImage 111 (2015) 360–368

Contents lists available at ScienceDirect

NeuroImage journal homepage: www.elsevier.com/locate/ynimg

Deep grey matter growth predicts neurodevelopmental outcomes in very preterm children Julia M. Young a,b,f,⁎, Tamara L. Powell a, Benjamin R. Morgan a, Dallas Card a, Wayne Lee a, Mary Lou Smith b,d,f, John G. Sled c,e, Margot J. Taylor a,b,f a

Department of Diagnostic Imaging, Hospital for Sick Children, Toronto, Ontario, Canada, M5G 1X8 Program in Neurosciences & Mental Health, Hospital for Sick Children, Toronto, Ontario, Canada, M5G 1X8 Program in Physiology & Experimental Medicine, Hospital for Sick Children, Toronto, Ontario, Canada, M5G 1X8 d Department of Psychology, Hospital for Sick Children, Toronto, Ontario, Canada, M5G 1X8 e Department of Medical Biophysics, University of Toronto, Toronto, Ontario, Canada, M5G 1L7 f Department of Psychology, University of Toronto, Toronto, Ontario, Canada, M5S 3G3 b c

a r t i c l e

i n f o

Article history: Accepted 13 February 2015 Available online 21 February 2015 Keywords: Magnetic resonance imaging Preterm birth Neonate Basal ganglia Cognitive outcome

a b s t r a c t We evaluated whether the volume and growth rate of critical brain structures measured by MRI in the first weeks of life following very preterm (b32/40 weeks) birth could predict subsequent neurodevelopmental outcomes at 4 years of age. A significant proportion of children born very prematurely have cognitive deficits, but these problems are often only detected at early school age. Structural T2-weighted magnetic resonance images were acquired in 96 very preterm neonates scanned within 2 weeks of birth and 70 of these at term-equivalent age. An automated 3D image analysis procedure was used to measure the volume of selected brain structures across all scans and time points. At 4 years of age, 53 children returned for neuropsychological assessments evaluating IQ, language and visual motor integration. Associations with maternal education and perinatal measures were also explored. Multiple regression analyses revealed that growth of the caudate and globus pallidus between preterm birth and term-equivalent age predicted visual motor integration scores after controlling for sex and gestational age. Further associations were found between caudate and putamen growth with IQ and language scores. Analyses at either preterm or term-equivalent age only found associations between normalized deep grey matter growth and visual motor integration scores at term-equivalent age. Maternal education levels were associated with measures of IQ and language, but not visual motor integration. Thalamic growth was additionally linked with perinatal measures and presence of white matter lesions. These results highlight deep grey matter growth rates as promising biomarkers of long-term outcomes following very preterm birth, and contribute to our understanding of the brain–behaviour relations in these children. © 2015 Elsevier Inc. All rights reserved.

Introduction An estimated 15 million preterm births occurred worldwide in 2010, with about 10% at less than 32 weeks gestation (Blencowe et al., 2012). These infants are born during the third trimester of human gestation, a period of accelerated brain growth that coincides with a critical window when dendritic and axonal arborization, synaptogenesis and myelination occur (Lenroot and Giedd, 2006). Foundational thalamocortical networks are consolidated that further establish cortical and basal ganglia connectivity with widespread cerebral networks (Kostović and Jovanov-Milosević, 2006). Neural components key to these networks, such as cortical and deep grey matter projection

⁎ Corresponding author at: Diagnostic Imaging, Hospital for Sick Children, 555 University Avenue, Toronto, Ontario, Canada, M5G 1X8. E-mail address: [email protected] (J.M. Young).

http://dx.doi.org/10.1016/j.neuroimage.2015.02.030 1053-8119/© 2015 Elsevier Inc. All rights reserved.

neurons, subplate neurons and oligodendrocyte precursors, are potentially the most vulnerable during this period, especially if very preterm birth is associated with white matter injury (WMI) and illness (Back et al., 2001; Ferriero and Miller, 2010; McQuillen et al., 2003). As deep grey matter structures are implicated in a wide range of cognitive functions (Arsalidou et al., 2013), their development is fundamental to normal cognition. Modern health care interventions have greatly improved survival rates of very preterm-born infants in developed countries, yet the rates of subsequent comorbid neurodevelopmental impairments have not improved; while 40–50% are indistinguishable at school-age from term-born children, at least 50% of very preterm-born children experience cognitive, language and/or motor skill deficits (Marlow, 2004; Saigal and Doyle, 2008). Perinatal clinical measures alone have failed to explain long-term developmental outcomes (Hart et al., 2008), and current explanations include a role for interacting environmental factors such as maternal education and biological factors such as WMI

J.M. Young et al. / NeuroImage 111 (2015) 360–368

361

and cortical dysmaturation detected by magnetic resonance imaging (MRI). Longitudinal study designs provide more potential to understand developmental outcomes in contrast to cross-sectional designs. For example, serial imaging of abnormal maturation of white matter microstructure and metabolism within the basal nuclei in conjunction with WMI beginning in the neonatal period is associated with adverse outcomes in very preterm-born infants (Chau et al., 2013). Dynamic changes in cortical thickness and growth in preterm and typically developing children and adolescents have also demonstrated connections with cognitive ability (Kapellou et al., 2006; Rathbone et al., 2011; Shaw et al., 2006; Sowell et al., 2004). The relation between longitudinal, neonatal structural brain maturation, however, has not been established. Previous cross-sectional MRI studies of deep grey matter structural development in preterms are limited and vary by analysis technique. The thalamus and lentiform nucleus at term-equivalent age display reduced growth, exacerbated by the presence of WMI (Ball et al., 2012; Boardman et al., 2006; Lin et al., 2001; Srinivasan et al., 2007). Wholebrain analyses similarly found reductions of the deep grey matter directly associated with disability and developmental outcome at infancy, yet only indirectly associated with cognitive function in later childhood and adolescence (Boardman et al., 2010; Inder et al., 2005; Kesler et al., 2004; Nosarti et al., 2008; Peterson et al., 2000). Specific examination of the thalamus, caudate and hippocampi in later childhood and adolescence of very preterm-born infants, however, found correlations between thalamic and caudate volumes with verbal fluency and intelligence (Giménez et al., 2006; Abernethy et al., 2004). While these cross-sectional studies have investigated aspects of deep grey matter volumes in relation to neurodevelopmental abilities, there remains a gap in our understanding of how these associations evolve from birth in very preterm-born infants. The present longitudinal study examined whether growth of the caudate, putamen, globus pallidus, thalamus, and total brain during the crucial third trimester of rapid brain growth predicted neurodevelopmental outcomes at 4 years of age. The contribution of perinatal clinical factors and maternal education with deep grey matter development and outcome measures was also investigated. We hypothesized that the maturation of the deep grey matter structures over the preterm period would predict cognitive outcomes at 4 years of age, and that these early weeks of maturation would prove to be a critical developmental window influencing cognitive outcomes in very preterm born children.

Manual segmentation of the whole caudate, putamen, globus pallidus, internal capsule and thalamus was performed on two 3D average images by one primary rater (JY) with the aid of a brain atlas (Harsberger et al., 2006) as shown in Figs. 1A–B. Two independent raters, a neuroradiologist and neurosurgeon, also manually segmented the selected structures. Intraclass correlations were calculated and averaged between the primary rater and two independent raters to assess the accuracy of the segmentation (caudate: 0.88, putamen: 0.99, globus pallidus: 0.85, thalamus: 0.99). The reference segmentation created by the primary rater was then used to automatically segment and compute the volume of the brain structures on each MRI using the previously derived spatial transformations (Fig. 2). These segmentations were inspected visually for accuracy. Only two caudate and one thalamus segmentation at the preterm time point were determined to be unreliable and excluded from further analysis. Total brain volume, including the cerebellum and excluding ventricles, was also manually drawn on the average images and propagated back to individual scans using the same method previously described. For those with enlarged ventricles, additional manual editing was performed to ensure that only brain regions were included to acquire total brain volume measures.

Materials and Methods

Table 1 Perinatal characteristics at very preterm birth.

Participants One hundred and five very preterm neonates (median age at birth in weeks: 28.6; range: 24.43–32.86; 55 males and 50 females) were recruited from the neonatal intensive care unit at the Hospital for Sick Children in Toronto. Neonates with any known chromosomal or major congenital abnormalities were excluded from recruitment. All families signed an informed consent agreeing to MRI scans, access to medical records and follow-up participation. The study protocol was approved by the Hospital for Sick Children research ethics board. Each very preterm neonate underwent an MRI within 2 weeks of birth (median age in weeks: 30.14, range: 25.1–34.86) while swaddled and lying flat during natural sleep. Following the first scan, five neonates died before term-equivalent age and four had gross motion artefact and anatomical abnormalities and were excluded from subsequent analyses. At term-equivalent age, 70 infants were scanned again (median age in weeks: 42; range: 36.57–46.43). At four years of age, 53 children returned (median age in years: 4.2; range: 4.02–4.91; median gestational age: 28.86; range: 26.29–31.14; 30 males and 23 females) for a comprehensive neuropsychological assessment.

MRI data MRI scans were performed on a 1.5 T GE Signa Excite HD Scanner (GE Medical Systems, Milwaukee, WI, USA) using an MR-compatible incubator and neonatal head coil (AIR Inc., Cleveland, OH, USA). Axial T2weighted and T1-weighted images were acquired (repetition time/echo time: 4000/145 and 23/4 ms; field of view: 128 and 128 mm; resolution: 1 × 1 × 1 mm; 90 and 110 slices; scan time: 4.16 and 5.39 min). Two paediatric neuroradiologists with extensive experience in neonatal imaging evaluated clinical images of each neonate's first scan independently. Images were evaluated for the grade of germinal matrix haemorrhage (GMH) and non-cystic white matter lesions. Incidence numbers are reported in Table 1 for infants with both preterm and term-equivalent scans. White matter lesions were further graded for mild to moderate and severe levels of injury by an expert neurologist (Miller et al., 2003). Two neonates had white matter injury consistent with periventricular leukomalacia (PVL).

MRI segmentation

Characteristic

Mean (SD) or number (%)

Gestational age (weeks) Antenatal steroids Intrauterine growth restriction Caesarean-section delivery Multiple births Males Birth Weight (g) Head circumference (cm) Apgar score at 5 min Resuscitation required (CPR) CRIB II Endotracheal tube days Oxygen administration days Patent ductus arteriosus (treated) Sepsis (culture positive) Meningitis Necrotizing enterocolitis (stages 2 & 3) Bronchopulmonary dysplasia GMH (grades 1–2) GMH (grades 3–4) White matter lesions

28.84 (1.78) 49 (75%) 12 (19%) 39 (60%) 10 (7%) 35 (54%) 1162.5 (263.5) 25.9 (2.0) 7.3 (1.8) 7 (11%) 6.6 (2.5) 12.7 (16.5) 19.6 (28.3) 16 (25%) 23 (35%) 5 (8%) 7 (11%) 15 (23%) 13(20%) 13 (20%) 22 (34%)

Characteristics are reported for 65 infants with longitudinal scans. CRIB II — Clinical Risk Index for Babies; GMH — germinal matrix haemorrhage.

362

J.M. Young et al. / NeuroImage 111 (2015) 360–368

Fig. 1. (A) The first two columns depict an average of 96 very preterm images scanned within 2 weeks of birth with and without deep grey matter segmentations. The third column depicts a combination of the average template (in cyan) and an individual image (in grey) from the fourth column to visualize the congruence of an individual registered to the average template. (B) Similar to (A) the first two columns depict an average of 70 very preterm images scanned at term-equivalent age. The internal capsule was additionally segmented to provide anatomical boundaries.

Fig. 2. Individual anatomical images and deep grey matter segmentations in four different ages across preterm and term-equivalent age. Images are scaled relative to each other.

J.M. Young et al. / NeuroImage 111 (2015) 360–368

Ninety-six T2-weighted images acquired shortly after preterm birth and 70 T2-weighted images acquired at term-equivalent age were coregistered using MICe-build-model (https://wiki.mouseimaging.ca/ display/MICePub/MICe-build-model; Lerch et al., 2011), a software for automated groupwise non-linear alignment of 3D images. This coregistration procedure yielded average images derived from all of the preterm and term-equivalent scans respectively, as well as a set of spatial transformations linking the average images to each individual scan. With the aid of these spatial transformations, individual brain structures outlined in 3D on the average images were propagated to individual source images for the purpose of computing the volumes of brain structures across the whole cohort. The alignment between individual images and the average templates were visually inspected and found to be consistently accurate for the subcortical structures and total brain. One infant's registration at term-equivalent age, however, was unsuccessful and excluded from further analyses bringing the total number of infants with preterm and term scans to 65. T2-weighted average images were chosen over T1-weighted average images for manual segmentation, due to better contrast between structures at this age. Neuropsychological assessments At four years of age, children underwent a neurodevelopmental assessment. Intelligence quotients (IQ) were determined by the Wechsler Preschool and Primary Scales of Intelligence — Third Edition (WPPSI-III) (Wechsler, 2002) using Canadian norms. Three different indices of cognitive abilities were obtained: Verbal IQ (VIQ), Performance IQ (PIQ), and Processing Speed (PSQ). The comprehensive indices of these subtests comprise the Full Scale IQ (FSIQ). Overall language ability was determined by the Clinical Evaluation of Language Fundamentals — Preschool, Second Edition (CELF-Pre-2) (Semel et al., 2004), which measures receptive and expressive language, language content and structure, and yields a core language summary score. Visual motor integration was assessed by the Beery–Buktenica Test of Visual Motor Integration (VMI) (Beery et al., 2010). Supplemental VMI tests of visual perception and motor coordination were also administered. Performance on these measures is shown in Table 2. Maternal education Levels of maternal education were obtained for children with 4-year assessments. Each education level (grade school, high school, nonuniversity post-secondary school, university, and post-graduate school) was ranked on a scale from 1–5. The percentage of mothers who completed the associated education levels was as follows: grade school

Table 2 Neuropsychological assessments at 4 years of age. Assessment

Subtests

N

Mean (SD)

Range

WPPSI-III

Full Scale IQ Performance IQ Processing Speed Verbal IQ Core language Receptive language Expressive language Language content Language structure Visual motor integration Visual perception Motor coordination

53 53 43 53 48 45 45 45 45 53 50 50

94.3 (15.2) 94.7 (12.9) 92.3 (18.7) 99.0 (16.9) 95.1 (15.7) 95.2 (16.5) 96.7 (18.2) 97.4 (15.8) 94.4 (18.2) 98.9 (11.8) 91.9 (19.8) 85.4 (15.4)

62–123 65–131 63–121 62–133 57–121 61–127 53–123 59–124 50–121 65–121 46–136 58–126

CELF-Pre-2

VMI

Neuropsychological measures are reported for a total of 53 very preterm children who returned for a follow-up visit. WPPSI-III: Wechsler Preschool and Primary Scales of Intelligence — 3rd Edition. CELF-Pre-2: Clinical Evaluation of Language Fundamentals — Preschool-2nd Edition; VMI — Beery–Buktenica Test of Visual Motor Integration. Numbers of assessments vary due to children's inability to complete every assessment.

363

(1.9%), high school (15.1%), non-university post-secondary school (24.5%), university (45.3%) and post-graduate school (13.2%).

Perinatal data Perinatal demographic, clinical and radiological characteristics obtained at birth are shown in Table 1 for 65 infants with both preterm and term-equivalent scans. Measures of illness severity were calculated for Apgar scores at 5 min and the Clinical Risk Index for Babies (CRIB-II). Medical interventions were noted such as the use of resuscitation (CPR), antenatal steroids, caesarean-section delivery, and days of endotracheal tube and oxygen administration. Incidences of intrauterine growth restriction (IUGR), patent ductus arteriosus, sepsis, meningitis, necrotizing enterocolitis, bronchopulmonary dysplasia, germinal matrix haemorrhage, and white matter lesions were also noted. Statistical analyses All statistical analyses were conducted using R version 3.1.2 and SPSS version 20.0 (SPSS Inc., Chicago, IL.). Linear mixed effects models were performed for summed left and right cross-sectional deep grey matter structures with age across preterm and term-equivalent scans to determine developmental trends. The equation for this model is as follows: lme(structure ~ scan age, random = subject). Total brain volume was also examined in this manner. Growth rates used in subsequent analyses were calculated for each brain structure based on longitudinal imaging data by dividing the difference in volumes by the difference in ages between preterm and term-equivalent scans. Linear regressions were then performed to determine associations between gestational age and growth rates. Our primary hypothesis was tested using a multiple hierarchical regression model. Composite scores of the WPPSI-III, CELF-Pre-2, and VMI were entered as dependent variables into separate regression models. Sex and gestational age were treated as covariates and entered into the model first. Growth rates of the total brain were treated as independent variables. An omnibus test of all the deep grey matter structures was performed and treated as independent variables and entered as separate models for each dependent variable. Post-hoc analyses for the deep grey matter structures using the same hierarchical regression model determined the predictive ability of each individual structure with neuropsychological measures. The equation for this model, where bn indicates the coefficients, is as follows: composite score = intercept + b1(sex) + b2(gestational age) + b3(structure volume). Values of P b 0.05 for the change in F-statistic based upon the multiple hierarchical regression analyses were considered statistically significant. Multiple comparisons were addressed using an omnibus test to help control for Familywise Type 1 error (Levin et al., 1994). To test whether growth rates were more predictive of neuropsychological measures than cross-sectional neuroanatomical volumes, hierarchical regressions were repeated for the preterm and term-equivalent time points separately. These analyses were performed both with absolute volumes and volumes normalized by total brain volume. Furthermore, to test whether levels of maternal education predicted neuropsychological measures better than deep grey matter growth measures, additional hierarchical regressions were performed treating levels of maternal education as an independent variable, individually and in combination with neuroanatomical growth measures. Sex and gestational age were also treated as covariates. Exploratory analyses of neuroanatomical growth measures and neuropsychological measures with perinatal clinical measures were performed. Pearson correlations, Spearman Rank-Order correlations and non-parametric Mann–Whitney U tests were used for continuous, scaled and non-normally distributed and dichotomous measures, respectively. Post-hoc tests of white matter lesion severity were also

364

J.M. Young et al. / NeuroImage 111 (2015) 360–368

performed. Values of P b 0.01 were considered significant for the exploratory analyses.

Table 3 Multiple hierarchical regression statistics. Measures

Results Neuroanatomical structures are associated with age at scan Dynamic linear growth of the deep grey matter and total brain was apparent between preterm and term-equivalent ages, representative of the marked changes in brain development during this period, equivalent to the third trimester of pregnancy. The caudate (t(62) = 48.560, P b 0.001), putamen (t(64) = 59.545, P b 0.001), globus pallidus (t(64) = 49.145, P b 0.001), thalamus (t(63) = 53.450, P b 0.001), and total brain (t(64) = 76.725, P b 0.001) volumes were all highly associated with age at scan (Fig. 3). Growth measures of these structures over this period were obtained from 65 infants with both preterm and term-equivalent scans. Average (SD) growth of the caudate was 116.13 (21.52) mm3/week, the putamen 206.93 (30.99) mm3/week, the globus pallidus 70.99 (13.05) mm3/week, the thalamus 418.59 (71.04) mm3/week, and the total brain 19,338.34 (2308.05) mm3/week. The association between growth measures and gestational ages was not significant for the caudate (R2 = 0.00008, P = 0.94), putamen (R2 = 0.004, P = 0.61), globus pallidus (R2 = 0.00008, P = 0.94), and thalamus (R2 = 0.03, P = 0.17), yet was significant for the total brain (R2 = 0.127, P = 0.004).

Neuroanatomical growth and 4-year outcomes Longitudinal deep grey matter and total brain growth were available for 45 children with 4-year neuropsychological measures. Growth of all of the deep grey matter structures significantly predicted measures of VMI (P = 0.02) after controlling for sex and gestational age from the omnibus tests as shown in Table 3. Post-hoc analyses revealed caudate and globus pallidus growth to be significant individual predictors of VMI. Although omnibus tests were insignificant for FSIQ and core language scores, caudate growth and putamen growth were individually associated with FSIQ and core language scores. Thalamic growth

Total brain All DGM measures Caudate Thalamus Globus pallidus Putamen

FSIQ

Core language

VMI

df

F

P-value df

F

P-value df

F

P-value

1,41 4,36 1,39 1,40 1,41 1,41

0.066 1.799 4.827 1.746 3.271 1.790

0.799 0.150 0.034⁎ 0.194 0.078 0.188

0.102 2.054 3.884 2.813 3.866 5.224

0.751 0.109 0.056 0.102 0.056 0.028⁎

0.253 3.247 4.365 1.095 7.859 1.961

0.618 0.023⁎ 0.043⁎ 0.302 0.008⁎

1,39 4,34 1,37 1,38 1,39 1,39

1,41 4,36 1,39 1,40 1,41 1,41

0. 169

Analyses are reported for 45 children with longitudinal DGM (deep grey matter), total brain and neuropsychological measures. Statistics represent the change in F-statistic and significant value after controlling for sex and gestational age. All DGM measures denote all deep grey matter growth measures entered into the model together. Individual measures such as the caudate, thalamus, globus pallidus, and putamen denote post-hoc tests performed. Degrees of freedom vary due to children who were unable to complete every assessment. ⁎ Significant values of P b 0.05.

was not a significant predictor of any neuropsychological measure (Figs. 4A–D). Neither cross-sectional deep grey matter volumes at the preterm (n = 49) nor term (n = 49) time point predicted FSIQ (F(4,40) = 0.661, P = 0.62; F(4,40) = 1.677, P = 0.17 respectively), core language (F(4,37) = 0.952, P = 0.45; F(4,36) = 2.022, P = 0.112), or VMI scores (F(4,40) = 0.823, P = 0.52; F(4,40) = 0.681, P = 0.61). This was also true for preterm and term cross-sectional total brain volumes with FSIQ (F(1,45) = 0.641, P = 0.43; F(1,41) = 1.243, P = 0.27), core language (F(1,42) = 2.005, P = 0.16; F(1,39) = 0.480, P = 0.493) and VMI scores (F(1,45) = 0.907, P = 0.35; F(1,41) = .019, P = 0.89). No associations were found between cross-sectional normalized deep grey matter volumes and the neuropsychological measures at the preterm time point. At term, however, there were significant associations found between normalized deep grey matter volumes and VMI (F(4,36) = 3.252, P = 0.022) driven by the caudate (F(1,39) = 6.840, P = 0.013). Including maternal education levels with deep grey matter growth measures (n = 45) in the model was only significant with VMI scores (F(5,35) = 3.241, P = 0.02). Levels of maternal education tested alone resulted in significant associations with FSIQ (F(1,41) = 5.502, P = 0.02) and core language (F(1,39) = 4.804, P = 0.03), but not VMI (F(1,41) = 4.194, P = 0.05).

Associations with clinical factors

Fig. 3. Scatter plot depicting linear growth of deep grey matter volumes (mm3) across preterm and term-equivalent ages (weeks). Included are 164 thalamus (green circles), 165 putamen (yellow circles), 163 caudate (red circles), and 165 globus pallidus (blue circles) volumetric measures of very preterm infants. Thin lines connect individual preterm and term-equivalent data points; thick lines depict linear regression lines across all individuals for each structure.

Total brain growth was most correlated to perinatal clinical variables such as gestational age, birth weight, and head circumference compared to other structures (Table 4). Higher CRIB II scores, an index of illness severity (Parry et al., 2003), and greater number of days of oxygen administration resulted in significantly less total brain growth. Thalamic growth was also significantly reduced with higher CRIB II scores, evidence of lung disease (bronchopulmonary dysplasia) and presence of white matter lesions. Post-hoc Mann–Whitney U tests determined that 15 infants with moderate and severe white matter lesions drove the association (U = 166, Z = − 2.7, P = 0.007) (Figs. 5A–B). No other structures were associated with white matter lesions. Intrauterine growth restriction was significantly associated with putamen growth. No other perinatal clinical or radiological measures were significantly associated with any deep grey matter growth. Core language scores were associated with IUGR, birth weight and CRIB II scores. Full Scale IQ was only associated with IUGR. No other perinatal clinical and radiological measures were associated with neuropsychological measures (Table 5).

J.M. Young et al. / NeuroImage 111 (2015) 360–368

365

Fig. 4. Residuals of post-hoc tests are plotted to illustrate partial correlations of neuropsychological measures and structural growth (mm3/weeks) after controlling for sex and gestational age. (A, B) Association of caudate and globus pallidus growth with VMI. (C, D) Association of putamen and caudate growth with core language and FSIQ.

Table 4 Perinatal characteristics with neuroanatomical growth. Characteristic

Gestational age (weeks) Antenatal steroids Intrauterine growth restriction Caesarean-section delivery Birth weight (g) Head circumference (cm) Apgar score at 5 min Resuscitation required (CPR) CRIB II Endotracheal tube days Oxygen administration days Patent ductus arteriosus (treated) Sepsis (culture positive) Meningitis Necrotizing enterocolitis (stages 2–3) Bronchopulmonary dysplasia GMH (grades 1–2) GMH (grades 3–4) White matter lesions

Total brain

Thalamus

Caudate

Globus pallidus

Putamen

rp,rs, U/Z

rp,rs, U/Z

rp,rs, U/Z

rp,rs, U/Z

rp,rs, U/Z

0.357⁎ 325/−1.020 304/−0.237 451/−0.750 0.445⁎ 0.405⁎

0.172 331/−0.822 263/−.0843 368/−1.722 0.189 0.133 0.135 163/−0.785 −0.38⁎

−0.009 279/−1.532 237/−1.208 333/−2.066 0.137 0.09 0.144 162/−0.744 −0.112 0.071 −0.08 361/−0.113 449/−0.029 132/−0.271 176/−0.219 271/−1.436 253/−1.223 214/−1.304 368/−1.197

0.099 169/−0.720 −0.357⁎ −0.258 −0.321⁎ 466/−0.458 404/−1.084 140/−0.188 122/−1.923 220/−2.413 210/−2.099 284/−0.008 383/−1.248

−0.259 −0.244 308/−1.074 467/−0.063 133/−0.305 160/−0.656 159/−3.304⁎ 194/−2.294 279/−0.568 265/−2.785⁎

−0.009 365/−0.411 242/−1.285 475/−0.428 0.069 0.035 0.08 185/−0.381 −0.092 0.144 0.213 368/−0.248 315/−2.305 129/−0.463 165/−0.612 348/−0.420 324/−0.230 308/−0.492 338/−1.872

−0.065 375/−0.259 127/−3.229⁎ 430/−1.031 0.133 0.066 0.043 177/−0.55 −0.032 0.098 0.038 351/−0.512 408/−1.029 111/−0.913 176/−0.367 366/−0.140 275/−1.033 294/−0.722 397/−1.054

Analyses are shown for 65 very preterm neonates with longitudinal DGM measures, total brain measures, perinatal clinical data, and radiological data. Pearson correlations (rp), Spearman Rank Order correlations (rs), or Mann–Whitney U tests (U/Z) were performed for continuous, scaled, or dichotomous measures. ⁎ Significant values of P b 0.01 are considered statistically significant.

366

J.M. Young et al. / NeuroImage 111 (2015) 360–368

Fig. 5. (A) Boxplot depicting thalamic growth between groups with and without white matter injury (WMI). (B) Scatterplot of cross-sectional thalamus volumes by scan age shows a gradient decline in thalamus volume with lesion severity. Blue diamonds, green squares and red circles respectively represent no, mild, and moderate to severe white matter lesions.

Discussion We found longitudinal growth of key deep grey matter structures between preterm and term-equivalent age predicted long-term developmental outcomes in very preterm-born children. These dramatic age-related volumetric changes reflect the extensive maturation of the deep grey matter that is initiated during the third trimester. Growth of these structures, particularly the caudate and globus pallidus, was related to visual motor integration abilities in early childhood. The present study highlighted longitudinal changes in structural growth, providing novel insight into the rapid brain development immediately following very preterm birth. Prior cross-sectional studies have found reductions in the size of the thalamus and basal ganglia in preterms at term-equivalent age (Boardman et al., 2006), childhood (Lax et al., 2013; Peterson et al., 2000), adolescence and young adulthood (Bjuland et al., 2014; Cheong et al., 2013; Nosarti et al., 2008).

Table 5 Perinatal characteristics with outcome measures. Characteristic

Gestational age (weeks) Antenatal steroids Intrauterine growth restriction Caesarean-section delivery Birth weight (g) Head circumference (cm) Apgar score at 5 min Resuscitation required (CPR) CRIB II Endotracheal tube days Oxygen administration days Patent ductus arteriosus (treated) Sepsis (culture positive) Meningitis Necrotizing enterocolitis (stages 2–3) Bronchopulmonary dysplasia GMH (grades 1–2) GMH (grades 3–4) White matter lesions

FSIQ

Core language VMI

rp,rs, U/Z

rp,rs, U/Z

rp,rs, U/Z

0.080 0.288 185/−1.869 175.5/−0.843 88.5/−2.595⁎ 62.5/−2.700⁎ 347/−0.018 232.5/−1.108 0.222 0.485⁎ 0.131 0.298 −0.073 0.152 60.5/−1.263 85.5/−0.093 −0.209 −0.414⁎

0.083 275/−0.051 132/−1.567 311/−0.662 0.212 0.045 −0.008 54/−1.485 −0.151 0.027 0.039 0.128 −0.139 −0.185 0.040 339.5/−0.027 211.5/−1.351 295.5/−0.823 312.5/−0.195 238.5/−0.672 266.5/−1.050 51.5/−1.529 55.5/−1.164 48/−1.651 92/−2.012 85/−1.505 158/−0.265 181.5/−0.762 133/−1.124 210.5/−0.103 224/−0.154 189/−0.025 200/−0.681 171/−1.001 125.5/−1.639 166/−1.116 246.5/−1.533 259/−0.097 306/−0.441

Analyses are shown for 53 very preterm children with neuropsychological, perinatal clinical, and radiological data. Pearson correlations (rp), Spearman Rank Order correlations (rs), or Mann–Whitney U tests (U/Z) were performed for continuous, scaled, or dichotomous measures. ⁎ Significant values of P b 0.01 are considered statistically significant.

Our current findings suggest that the onset of deep grey matter dysmaturation likely occurs during the preterm period and before term-equivalent age, as reduced volumes have already been reported at term. Absolute cross-sectional volumetric measures at the preterm and term-equivalent time point were not associated with outcome measures within our cohort. Only at term-equivalent age, an association was found between normalized deep grey matter growth and visual motor integration abilities. More importantly, by measuring the change in volume between two time points, we identified informative associations between longitudinal deep grey matter growth and neurodevelopmental outcome. Infants with poorer outcomes experienced slower growth of the caudate, putamen and globus pallidus. Longitudinal changes provided more predictive information of neurodevelopmental trajectories than cross-sectional data alone. Several reports have examined the role of early brain growth in typically developing and very preterm infants. Robust volumetric brain growth immediately following birth has been shown to be driven primarily by grey matter relative to white matter in typically developing termborn infants (Knickmeyer et al., 2008; Gilmore et al., 2007). Furthermore, brain growth in a sample of term-born and term-equivalent pretermborn infants showed differential patterns of regional growth marked by a deceleration of the rate of growth already evident at birth and continuing through the first three months of life (Holland et al., 2014). In relation to neurodevelopmental outcomes, extremely preterm-born infants serially imaged until term-equivalent age experienced cortical surface area expansion but not cerebral volume growth to be proportional to neurodevelopmental outcome measures at 2 and 6 years of age (Kapellou et al., 2006; Rathbone et al., 2011). This may be because there are more developmental increases in cortical surface area due to cortical folding compared to volume during the neonatal period. Similarly, our findings of total brain volume growth rates between preterm and termequivalent age did not yield similar associations with outcome measures. In contrast, caudate growth was associated with VMI and FSIQ outcomes, which supports its developmental role in cognitive function. As a part of the striatum and basal ganglia, the caudate is an integral structure within the corticostriatal network, which undergoes agedependent functional connectivity changes (Greene et al., 2014). Three distinct functional pathways comprising the limbic, sensorimotor and associative corticostriatal networks are formed by topographical inputs from the cortex to the striatum, projecting back through the globus pallidus, substantia nigra and thalamus (Grahn et al., 2008). A few studies have investigated relations between caudate structure and cognition in preterm-born adolescents and adults. Reduced caudate volumes correlated with IQ, hyperactivity and working memory in very preterm

J.M. Young et al. / NeuroImage 111 (2015) 360–368

born children, adolescents and young adults (Abernethy et al., 2004; Bjuland et al., 2014; Nosarti et al., 2005), while altered functional activation in the caudate of very preterm-born adolescents and young adults compared to controls correlated with visuo-perceptual learning and executive function tasks (Narberhaus et al., 2009; Nosarti et al., 2009). Our findings, that caudate growth predicts VMI and FSIQ in the developing brain, support motor and cognitive-related caudate functions, featuring its importance for predicting outcomes in very preterm-born children. The putamen and globus pallidus exhibited contrasting associations with neuropsychological measures. The lentiform nucleus, comprising both the putamen and globus pallidus was reported to be smaller in preterms at term-equivalent age (Srinivasan et al., 2007). Our analyses determined that putamen growth was only marginally informative in predicting core language scores while the growth of the globus pallidus demonstrated a robust association with VMI scores. This is possibly because the globus pallidus serves as the primary output following striatal inputs from the outer cortex. Circuitries with inputs from the frontal and occipital cortices specific to regions of the caudate and globus pallidus (Grahn et al., 2008) are likely to be involved in VMI scores, as they require the ability to coordinate both visual and motor systems. We found that thalamic growth was the most rapid compared to the other deep grey matter structures yet it was not associated with neuropsychological outcome measures. This finding was surprising considering prior studies, which have demonstrated relations between thalamic volumes and cognitive measures in preterm-born children and adolescents (Boardman et al., 2010; Cheong et al., 2013; Giménez et al., 2006; Zubiaurre-Elorza et al., 2012), although often presented in conjunction with WMI. Studies of typically developing children through to young adulthood have also found correlations of the thalamus with FSIQ, VIQ and syntactic and semantic language (Frangou et al., 2004; Wahl et al., 2008; Xie et al., 2012). Importantly, none of these studies investigated thalamic growth when thalamocortical connections are beginning to form. The functional associations with thalamic growth during this time may be dominated more by concurrent developmental progressions rather than long-term outcomes. Early connections from the thalamus to the cortex are critically important during preterm development. Most thalamic neurons appear in the second trimester, although studies support a second, later migration of neurons to dorsal regions of the thalamus (Letinic and Rakic, 2001). The vulnerability of the dorsal medial nucleus of the thalamus to WMI associated with preterm birth may be due to a disruption of this event (Volpe, 2009). We confirmed that thalamic growth was adversely affected by the presence of WMI as moderate and severe WMI impacted thalamic growth. In addition to WMI, the thalamus was most affected by associations with the CRIB II (Parry et al., 2003), and prolonged oxygen requirement. None of the other deep grey matter structures were affected by WMI in our cohort, possibly due to the majority of infants in our series having focal white matter lesions rather than more severe PVL. Further studies are needed to define the relative contributions of different forms of WMI versus primary insults. Perinatal clinical measures obtained shortly after birth were marginally associated with outcome measures, which is an important concern for clinician's ability to predict outcomes. Studies have reported that slower postnatal growth, as reflected by measures such as weight gain, is an indicator of delayed cortical growth (Vinall et al., 2013) and worse neurodevelopmental outcomes (Belfort et al., 2011). While our current study did not explore changes in such measures, we did find correlations between FSIQ and IUGR in addition to core language scores and IUGR, birth weight and the CRIB II. Maternal education was related to intelligence and core language scores in the very preterm-born children. When maternal education and neurodevelopmental outcomes were tested alone and in combination with deep grey matter growth measures, however, the strength of the prediction for VMI scores was not stronger than for deep grey matter growth alone. One report found that maternal education was a stronger predictor of cognitive function than neuroanatomical or

367

perinatal clinical factors in preterm children around 8 years of age (Kesler et al., 2004). In a large series of typically developing children and adolescents, parental education was found to be highly correlated with verbal, performance and Full Scale IQ, but did not mediate effects between total brain volume and IQ measures (Lange et al., 2010). Levels of maternal education are associated with environmental and genetic factors that can serve as a potential protective factor against the adversity of very preterm birth. Such interacting factors include enriched environments, better nutrition and inherited predispositions that progressively influence brain development and intelligence over time (Isaacs et al., 2008; Tonga and Thompson, 2005). There are some limitations to consider with our study. Follow-up at term-equivalent age and 4 years of age was a subset of the original cohort due to loss of contact or disinterest in continuing participation. In addition, our semi-automatic segmentation method was not formally quantified for accuracy as precise manual tissue delineation within individual scans across the preterm period is very challenging. The coregistration method was also not appropriate for detailed cortical analysis, however, due to stark variations in cortical morphology across the preterm age range. This is because the co-registration method works on the assumption that all structures are present, whereas the cortex develops new sulci and gyri during this period, such that the cortical segmentation that is produced is only approximate. Improved automated segmentation of the neonatal brain has been recently demonstrated with potential to address this problem (Gousias et al., 2013; Oishi et al., 2012). The contribution of white matter was not explored in the present study, yet also undergoes significant changes during the preterm period, such as the creation of thalamocortical and early corticocortical connections (Kostović and Jovanov-Milosević, 2006). Through early childhood following the neonatal period, white matter continues to develop rapidly with grey matter and plays an integral role in cognitive function (Dubois et al., 2014; Pandit et al., 2013). Further characterization of both grey and white matter at the time of neuropsychological testing in our cohort would be beneficial to understand this relation. Our study is the first to characterize longitudinal growth of the deep grey matter and total brain in children born very preterm and identify the relative contributions of specific structures to long-term neurodevelopmental outcomes. In particular, growth of the caudate appears to play a greater role in cognitive and visual motor function than the other structures. This may be due to its timing of coordinated growth and earlier maturation relative to other brain regions, as deep grey matter structures have earlier developmental trajectories than cortical regions. Early brain dysmaturation of the caudate, putamen and globus pallidus in relation to neurodevelopmental outcomes identifies the early growth of these structures as promising biomarkers of long-term outcomes and emphasizes the preterm period as a critical window for development. Acknowledgments We thank all of the families who have participated in the study. We are most grateful for the MRI technicians, Tammy Rayner, Ruth Weiss and Gary Detzler, for their imaging expertise. We thank Dr. Charles Raybaud, Dr. Manohar Shroff, Dr. Hilary Whyte and Dr. Aideen Moore for their invaluable ongoing involvement in the study. We also thank Dr. Steven Miller for his contributions on the WMI assessments. We thank Angela Thompson for her clinical support and Drs. Divyata Hingwala and George Ibrahim for their help in validating the anatomical segmentations. All phases of this study were supported by a Canadian Institutes of Health Research Grant, MOP-84399. The authors declare no competing financial interests. References Abernethy, L.J., Cooke, R.W.I., Foulder-Hughes, L., 2004. Caudate and hippocampal volumes, intelligence, and motor impairment in 7-year-old children who were born preterm. Pediatr. Res. 55, 884–893.

368

J.M. Young et al. / NeuroImage 111 (2015) 360–368

Arsalidou, M., Duerden, E.G., Taylor, M.J., 2013. The centre of the brain: topographical model of motor, cognitive, affective, and somatosensory functions of the basal ganglia. Hum. Brain Mapp. 34, 3031–3054. Back, S.A., Lui, N.L., Borenstein, N.S., Levine, J.M., Volpe, J.J., Kinney, H.C., 2001. Late oligodendrocyte progenitors coincide with the developmental window of vulnerability for human perinatal white matter injury. J. Neurosci. 21, 1302–1312. Ball, G., Boardman, J.P., Rueckert, D., Aljabar, P., Arichi, T., Merchant, N., Gousias, I.S., Edwards, A.D., Counsell, S.J., 2012. The effect of preterm birth on thalamic and cortical development. Cereb. Cortex 22, 1016–1024. Beery, K., Buktenica, N., Beery, N., 2010. Beery–Buktenica Test of Visual Motor Integration. The Psychological Corporation, San Antonio, TX. Belfort, M.B., Rifas-Shiman, S.L., Sullivan, T., Collins, C.T., McPhee, A.J., Ryan, P., Kleinman, K.P., Gilman, M.W., Gibson, R.A., Makrides, M., 2011. Infant growth before and after term: effects on neurodevelopment in preterm infants. Pediatrics 128, e899–e906. Bjuland, K.J., Rimol, L.M., Løhaugen, G.C.C., Skranes, J., 2014. Brain volumes and cognitive function in very-low-birth-weight (VLBW) young adults. Eur. J. Paediatr. Neurol. 18, 578–590. Blencowe, H., Cousens, S., Oestergaard, M.Z., Chou, D., Moller, A.B., Narwal, R., Adler, A., Garcia, C.V., Rohde, S., Say, L., Lawn, J.E., 2012. National, regional, and worldwide estimates of preterm birth rates in the year 2010 with time trends since 1990 for selected countries: a systematic analysis and implications. Lancet 379, 2162–2172. Boardman, J.P., Counsell, S.J., Rueckert, D., Kapellou, O., Bhatia, K.K., Aljabar, P., Hajnal, J., Allsop, J.M., Rutherford, M.A., Edwards, A.D., 2006. Abnormal deep grey matter development following preterm birth detected using deformation-based morphometry. Neuroimage 32, 70–78. Boardman, J.P., Craven, C., Valappil, S., Counsell, S.J., Dyet, L.E., Rueckert, D., Aljabar, P., Rutherford, M.A., Chew, A.T.M., Allsop, J.M., Cowan, F., Edwards, A.D., 2010. A common neonatal image phenotype predicts adverse neurodevelopmental outcome in children born preterm. Neuroimage 52, 409–414. Chau, V., Synnes, A., Grunau, R.E., Poskitt, K.J., Brant, R., Miller, S.P., 2013. Abnormal brain maturation in preterm neonates associated with adverse developmental outcomes. Neurology 81, 2082–2089. Cheong, J.L.Y., Anderson, P.J., Roberts, G., Burnett, A.C., Lee, K.J., Thompson, D.K., Molloy, C., Wilson-Chind, M., Connelly, A., Seal, M.L., Wood, S.J., Doyle, L.W., 2013. Contribution of brain size to IQ and educational underperformance in extremely preterm adolescents. PLoS One 8, e77475. Dubois, J., Dehaene-Lambertz, G., Kulikova, S., Poupon, C., Hüppi, P.S., Hertz-Pannier, L., 2014. The early development of brain white matter: a review of imaging studies in fetuses, newborns and infants. Neuroscience 276, 48–71. Ferriero, D.M., Miller, S.P., 2010. Imaging selective vulnerability in the developing nervous system. J. Anat. 217, 429–435. Frangou, S., Chitins, X., Williams, S.C.R., 2004. Mapping IQ and gray matter density in healthy young people. Neuroimage 23, 800–805. Gilmore, J.H., Lin, W., Prastawa, M.W., Looney, C.B., Vetsa, S.K., Knickmeyer, R.C., Evans, D.D., Smith, J.K., Hamer, R.M., Lieberman, J.A., Gerig, G., 2007. Regional gray matter growth, sexual dimorphism, and cerebral asymmetry in the neonatal brain. J. Neurosci. 27, 1255–1260. Giménez, M., Junqué, C., Narberhaus, A., Botet, F., Bargallo, N., Mercader, J.M., 2006. Correlations of thalamic reductions with verbal fluency impairment in those born prematurely. Neuroreport 17, 463–466. Gousias, I.S., Hammers, A., Counsell, S.J., Srinivasan, L., Rutherford, M.A., Heckemann, R.A., Hajnal, J.V., Rueckert, D., Edwards, A.D., 2013. Magnetic resonance imaging of the newborn brain: automatic segmentation of brain images into 50 anatomical regions. PLoS One 8, e59990. Grahn, J.A., Parkinson, J.A., Owen, A.M., 2008. The cognitive functions of the caudate nucleus. Prog. Neurobiol. 86, 141–155. Greene, D.J., Laumann, T.O., Dubis, J.W., Ihnen, S.K., Neta, M., Power, J.D., Pruett, J.R., Black, K.J., Schlaggar, B.L., 2014. Developmental changes in the organization of functional connections between the basal ganglia and cerebral cortex. J. Neurosci. 34, 5842–5854. Harsberger, H.R., Osborn, A.G., Ross, J.S., Macdonald, A., 2006. Diagnostic and Surgical Imaging Anatomy. 1st ed. Amirsys, Salt Lake City, UT. Hart, A.R., Whitby, E.W., Griffiths, P.D., Smith, M.F., 2008. Magnetic resonance imaging and developmental outcome following preterm birth: review of current evidence. Dev. Med. Child Neurol. 50, 655–663. Holland, D., Chang, L., Ernst, T.M., Curran, M., Buchthal, S.D., Alicata, D., Skranes, J., Johansen, H., Hernandez, A., Yamakawa, R., Kuperman, J.M., Dale, A.M., 2014. Structural growth trajectories and rates of change in the first 3 months of infant brain development. JAMA Neurol. 71, 1266–1274. Inder, T.E., Warfield, S.K., Wang, H., Huppi, P.S., Volpe, J.J., 2005. Abnormal cerebral structure is present at term in premature infants. Pediatrics 115, 286–294. Isaacs, E.B., Gadian, D.G., Sabatini, S., Chong, W.K., Quinn, B.T., Fischl, B.R., Lucas, A., 2008. The effect of early human diet on caudate volumes and IQ. Pediatr. Res. 63, 308–314. Kapellou, O., Counsell, S.J., Kennea, N., Dyet, L., Saeed, N., Stark, J., Maalouf, E., Dugan, P., Ajayi-Obe, M., Hajnal, J., Allsop, J.M., Boardman, J., Rutherford, M.A., Cowan, F., Edwards, A.D., 2006. Abnormal cortical development after premature birth shown by altered allometric scaling of brain growth. PLoS Med. 3, e265. Kesler, S.R., Ment, L.R., Vohr, B., Pajot, S.K., Schneider, K.C., Katz, K.H., Ebbitt, T.B., Duncan, C.C., Makuch, R.W., Reiss, A.L., 2004. Volumetric analysis of regional cerebral development in preterm children. Pediatr. Neurol. 31, 318–325. Knickmeyer, R.C., Gouttard, S., Kang, C., Evans, D., Wilber, K., Smith, J.K., Hamer, R.M., Lin, W., Gerig, G., Gilmore, J.H., 2008. A structural MRI study of human brain development from birth to 2 years. J. Neurosci. 28, 12176–12182. Kostović, I., Jovanov-Milosević, N., 2006. The development of cerebral connections during the first 20–45 weeks' gestation. Semin. Fetal Neonatal Med. 11, 415–422.

Lange, N., Froimowitz, M.P., Bigler, E.D., Lainhart, J.E., 2010. Associations between IQ, total and regional brain volumes, and demography in a large normative sample of healthy children and adolescents. Dev. Neuropsychol. 35, 296–317. Lax, I.D., Duerden, E.G., Lin, S.Y., Chakravarty, M.M., Donner, E.J., Lerch, J.P., Taylor, M.J., 2013. Neuroanatomical consequences of very preterm birth in middle childhood. Brain Struct. Funct. 218, 575–585. Lenroot, R.K., Giedd, J.N., 2006. Brain development in children and adolescents: insights from anatomical magnetic resonance imaging. Neurosci. Biobehav. Rev. 30, 718–729. Lerch, J.P., Sled, J.G., Henkelman, R.M., 2011. MRI phenotyping of genetically altered mice. Methods Mol. Biol. 711, 349–361. Letinic, K., Rakic, P., 2001. Telencephalic origin of human thalamic GABAergic neurons. Nat. Neurosci. 4, 931–936. Levin, J.R., Serlin, R.C., Seaman, M.A., 1994. A controlled, powerful multiple comparison strategy for several situations. Psychol. Bull. 115, 153–159. Lin, Y., Okumura, A., Hayakawa, F., Kato, T., Kuno, K., Watanabe, K., 2001. Quantitative evaluation of the thalami and basal ganglia in infants with periventricular leukomalacia. Dev. Med. Child Neurol. 43, 481–485. Marlow, N., 2004. Neurocognitive outcome after very preterm birth. Arch. Dis. Child. 89, F224–F228. McQuillen, P.S., Sheldon, R.A., Shatz, C.J., Ferriero, D.M., 2003. Selective vulnerability of subplate neurons after early neonatal hypoxia-ischemia. J. Neurosci. 23, 3308–3315. Miller, S.P., Cozzio, C.C., Goldstein, R.B., Ferriero, D.M., Partridge, J.C., Vigneron, D.B., Barkovich, A.J., 2003. Comparing the diagnosis of white matter injury in premature newborns with serial MR imaging and transfontanel ultrasonography findings. ANJ. Am. J. Neuroradiol. 24, 1661–1669. Narberhaus, A., Lawrence, E., Allin, M.P., Walshe, M., McGuire, P., Rifkin, L., Murray, R., Nosarti, C., 2009. Neural substrates of visual paired associates in young adults with a history of very preterm birth: alterations in fronto-parieto-occipital networks and caudate nucleus. Neuroimage 47, 1884–1893. Nosarti, C., Allin, M.P., Frangou, S., Rifkin, L., Murray, R.M., 2005. Hyperactivity in adolescents born very preterm is associated with decreased caudate volume. Biol. Psychiatry 57, 661–666. Nosarti, C., Giouroukou, E., Healy, E., Rifkin, L., Walshe, M., Reichenberg, A., Chitnis, X., Williams, S.C.R., Murray, R.M., 2008. Grey and white matter distribution in very preterm adolescents mediates neurodevelopmental outcome. Brain 131, 205–217. Nosarti, C., Shergill, S.S., Allin, M.P., Walshe, M., Rifkin, L., Murray, R.M., McGuire, P.K., 2009. Neural substrates of letter fluency processing in young adults who were born very preterm: alterations in frontal and striatal regions. Neuroimage 47, 1904–1913. Oishi, K., Mori, S., Donohue, P.K., Ernst, T., Anderson, L., Buchthal, S., Faria, A., Jiang, H., Li, X., Miller, M.I., Chang, L., 2012. Multi-contrast human neonatal brain atlas: application to normal neonate development analysis. Neuroimage 56, 8–20. Pandit, A.S., Ball, G., Edwards, A.D., Counsell, S.J., 2013. Diffusion magnetic resonance imaging in preterm brain injury. Neuroradiology 55, 65–95. Parry, G., Tucker, J., Tarnow-Mordi, W., 2003. CRIB II: an update of the clinical risk index for babies score. Lancet 361, 1789–1791. Peterson, B.S., Vohr, B., Staib, L.H., Cannistraci, C.J., Dolberg, A., Schneider, K.C., Katz, K.H., Westerveld, M., Sparrow, S., Anderson, A.W., Duncan, C.C., Makuch, R.W., Gore, J.C., Ment, L.R., 2000. Regional brain volume abnormalities and long-term cognitive outcome in preterm infants. JAMA 284, 1939–1947. Rathbone, R., Counsell, S.J., Kapellou, O., Dyet, L., Kennea, N., Hajnal, J., Allsop, J.M., Cowan, F., Edwards, A.D., 2011. Perinatal cortical growth and childhood neurocognitive abilities. Neurology 77, 1510–1517. Saigal, S., Doyle, L.W., 2008. An overview of mortality and sequelae of preterm birth from infancy to adulthood. Lancet 371, 261–269. Semel, E., Wiig, E., Secord, W., 2004. Clinical Evaluation of Language Fundamentals — Preschool. 2nd ed. The Psychological Corporation, San Antonio, TX. Shaw, P., Greenstein, D., Lerch, J., Clasen, L., Lenroot, R., Gogtay, N., Evans, A., Rapoport, J., Giedd, J., 2006. Intellectual ability and cortical development in children and adolescents. Nature 440, 676–679. Sowell, E.R., Thompson, P.M., Leonard, C.M., Welcome, S.E., Kan, E., Toga, A.W., 2004. Longitudinal mapping of cortical thickness and brain growth in normal children. J. Neurosci. 24, 8223–8231. Srinivasan, L., Dutta, R., Counsell, S.J., Allsop, J.M., Boardman, J.P., Rutherford, M.A., Edwards, A.D., 2007. Quantification of deep gray matter in preterm infants at termequivalent age using manual volumetry of 3-tesla magnetic resonance images. Pediatrics 119, 759–765. Tonga, A.W., Thompson, P.M., 2005. Genetics of bran structure and intelligence. Annu. Rev. Neurosci. 28, 1–23. Vinall, J., Grunau, R.E., Brant, R., Chau, V., Poskitt, K.J., Synnes, A.R., Miller, S.P., 2013. Slower postnatal growth is associated with delayed cerebral cortical maturation in preterm newborns. Sci. Transl. Med. 5, 168ra8. Volpe, J.J., 2009. Brain injury in premature infants: a complex amalgam of destructive and developmental disturbances. Lancet Neurol. 8, 110–124. Wahl, M., Marzinzik, F., Friederici, A.D., Hahne, A., Kupsch, A., Schneider, G.H., Saddy, D., Curio, G., Kostermann, F., 2008. The human thalamus processes syntactic and semantic language violations. Neuron 59, 695–707. Wechsler, D., 2002. Wechsler Preschool and Primary Scales of Intelligence. 3rd ed. The Psychological Corporation, San Antonio, TX. Xie, Y., Chen, Y.A., De Bellis, M.D., 2012. The relationship of age, gender, and IQ with the brainstem and thalamus in healthy children and adolescents: a magnetic resonance imaging volumetric study. J. Child Neurol. 27, 325–331. Zubiaurre-Elorza, L., Soria-Pastor, S., Junqué, C., Fernandez-Espejo, D., Segarra, D., Bargallo, N., Romano-Berindoague, C., Macaya, A., 2012. Thalamic changes in a preterm sample with periventricular leukomalacia: correlation with white-matter integrity and cognitive outcome at school age. Pediatr. Res. 71, 354–360.