Early Human Development 90 (2014) 443–450
Contents lists available at ScienceDirect
Early Human Development journal homepage: www.elsevier.com/locate/earlhumdev
Delay of cortical thinning in very preterm born children Ines Mürner-Lavanchy a,b,c,e, Maja Steinlin a,c, Mathias Nelle d, Christian Rummel b, Walter J. Perrig c,e, Gerhard Schroth b, Regula Everts a,b,c,⁎ a
Division of Neuropediatrics, Development and Rehabilitation, Children's University Hospital, Inselspital, Bern, Switzerland University Institute of Diagnostic and Interventional Neuroradiology, University Hospital, Inselspital, Bern, Switzerland Centre for Cognition, Learning and Memory, University of Bern, Bern, Switzerland d Division of Neonatology, Children's University Hospital, Inselspital, Bern, Switzerland e Institute of Psychology, University of Bern, Bern, Switzerland b c
a r t i c l e
i n f o
Article history: Received 11 December 2013 Received in revised form 2 May 2014 Accepted 29 May 2014 Keywords: Cortical thickness Very preterm Very low birth weight Children Development Delay
a b s t r a c t Background: Cortical gray matter thinning occurs during childhood due to pruning of inefficient synaptic connections and an increase in myelination. Preterms show alterations in brain structure, with prolonged maturation of the frontal lobes, smaller cortical volumes and reduced white matter volume. These findings give rise to the question if there is a differential influence of age on cortical thinning in preterms compared to controls. Aims: To investigate the relationship between age and cortical thinning in school-aged preterms compared to controls. Study design and outcome measures: The automated surface reconstruction software FreeSurfer was applied to obtain measurements of cortical thickness based on T1-weighted MRI images. Subjects: Forty-one preterms (b32 weeks gestational age and/or b 1500 g birth weight) and 30 controls were included in the study (7–12 years). Results: In preterms, age correlated negatively with cortical thickness in right frontal, parietal and inferior temporal regions. Furthermore, young preterms showed a thicker cortex compared to old preterms in bilateral frontal, parietal and temporal regions. In controls, age was not associated with cortical thickness. Conclusion: In preterms, cortical thinning still seems to occur between the age of 7 and 12 years, mainly in frontal and parietal areas whereas in controls, a substantial part of cortical thinning appears to be completed before they reach the age of 7 years. These data indicate slower cortical thinning in preterms than in controls. © 2014 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Important maturational events take place in the third trimester of pregnancy [1]. Since very preterm born children are born in this crucial phase, they are particularly vulnerable to primary injuries and secondary maturational difficulties. Various methods have been used to identify structural alterations in the preterm brain. Studies using magnetic resonance imaging (MRI) have reported lower cortical gray and white matter volumes, lower cerebellar volumes as well as smaller corpus callosum and hippocampus size in preterm born infants [2,3], children [4–7] and adolescents [8,9]. Furthermore, a smaller cortical surface area was observed in 7–10 year-old preterm children when compared to agematched term born controls [7]. A voxel-based morphometry study found atypical gray and white matter distribution in preterm adolescents aged 14–15 years [10]. Diffusion tensor imaging studies detected alterations in white matter fiber tract organization throughout the brain, ⁎ Corresponding author at: Division of Neuropediatrics, Development and Rehabilitation, Children's University Hospital, Inselspital, 3010 Bern, Switzerland. Tel.: +41 31 632 41 30. E-mail address: [email protected] (R. Everts).
http://dx.doi.org/10.1016/j.earlhumdev.2014.05.013 0378-3782/© 2014 Elsevier Ireland Ltd. All rights reserved.
suggesting differences in structural connectivity [11] as well as widespread microstructural white matter abnormalities in very preterm born children compared to same-aged term born controls [12]. The structural alterations in very preterm born children raise the question if the process of cortical thinning differs between preterms and controls. Cortical thickness is thought to be an indicator of the number of neurons per cortical column (groups of neurons which connect the six horizontal layers of the neocortex vertically) as well as glial support and dendritic arborization [13]. From early childhood to adolescence, decrease of cortical thickness co-occurs with the pruning of dispensable neurons and synapses [14]. This process leads to more efficient synaptic connections. The normal developmental cortical thinning does not occur simultaneously over the whole cortex. Findings concerning synaptic density, which is indirectly related to cortical thickness, suggest that during the course of development a synaptic loss occurs first in primary sensory and motor regions and later in multimodal association areas [15,16]. More recent studies confirmed this pattern using neuroimaging of cortical thinning in healthy children [14,17], others found contradictory results. Correspondingly, a longitudinal MRI study found not only regional specific thinning but also thickening in circumscribed perisylvian language relevant regions in 45 healthy
444
I. Mürner-Lavanchy et al. / Early Human Development 90 (2014) 443–450
children between 5 and 11 years of age [18]. Since the process of cortical thinning varies over different cerebral regions, it is important to examine cortical thickness over the whole cortex and on a regional level. A cross-sectional study investigated the development of cortical thinning in children with very low birth weight and term born controls at 18–22 months and 3–4 years of age [19]. At both age groups, children with very low birth weight displayed greater mean cortical thickness compared to control children. Although the difference did not reach statistical significance, the authors suggest that their results point to a delay in normal cortical thinning following prematurity, up to the age of four years. Correspondingly, a longitudinal study comparing 55 preterm children (born with 600 to 1250 g birth weight) with 20 term born control children at 8 and 12 years of age, identified different gray and white matter volume changes over time between the study groups: In preterms, gray matter volumes decreased and white matter volumes increased less than in controls, which suggests a different gray and white matter development in preterms than in controls [20]. To the authors' knowledge, no study has investigated the effect of age on global and regional cortical thinning in a sample of school-aged very preterm born children in a cross-sectional study design. Based on the existing findings, we hypothesize that age is differentially associated with cortical thinning in preterms and controls between 7 and 12 years of age. We further assume that the normal developmental cortical thinning is delayed in very preterm born children compared to their term born peers, even in our sample of relatively healthy very preterm born children. 2. Methods This study reports on a subset of data from the NEMO (NEuropsychology and meMOry) research project at the Children's University Hospital in Bern, Switzerland. The NEMO project examines cognitive development in very preterm born children including behavioral and neuroimaging data. The study protocol was approved by the local ethics committee. All children and caregivers provided informed written consent to the research and publication of the results prior to participation, consistent with the Code of Ethics of the World Medical Association (Declaration of Helsinki). 2.1. Very preterm group Medical reports of all very preterm born children (b 32 weeks of gestation and/or b 1500 g birth weight) born between 1998 and 2003 at the Children's University Hospital Bern, Switzerland were reviewed. Native German speakers aged 7 to 12 years with normal neonatal ultrasound, no or mild periventricular leukomalacia (grade I or II), no or mild neonatal cerebral lesions (hemorrhage grade I), no chronic illness, no pervasive developmental disorders, and Full-scale IQ N 85 in the neuropsychological follow-ups were included. A total of 247 children fulfilled the inclusion criteria and were contacted by letter out of which 75 children agreed to participate in the study. Fifty-five very preterm born children completed the MRI assessment. Fourteen preterms had to be excluded (technical problems n = 4, movement n = 10) resulting in 41 preterms for inclusion in the analysis. 2.2. Control group Term born controls from the same year cohorts were recruited by means of announcements in the hospital. Forty-two children completed the MRI examination, twelve children were excluded (technical problems n = 2 and movement, n = 10), resulting in 30 controls included in the analysis. Handedness was enquired by telephone interview prior to the first assessment. Socioeconomic status (SES) was defined as mother's and father's education level at the time of the neuropsychological assessment (no graduation = 1, college = 2, college of higher education =
3, university degree = 4). Full-scale IQ was assessed using the short form of the German version of the ‘Wechsler Intelligence Scale for Children, Fourth Edition’ [21]. 2.3. MR imaging Children underwent a one-hour MRI assessment at the Department of Diagnostic and Interventional Neuroradiology, University Hospital Bern. 2.3.1. Image acquisition MRI was performed on a Verio3-T whole body scanner (Siemens Erlangen, Germany) equipped with a 40 mT/m gradient system and a CP standard head coil (12 channels). The scanner was equipped with the Syngo MR 2002B (VA17) software. Anatomical imaging was obtained using a T1-weighted, 3D-MPRAGE sequence (TR 2300 ms, TE 2.98 ms, TI 900 ms, 0 mm gap, FoV 256, 1 mm voxel resolution, 160 contiguous sagittal slices) with an acquisition duration of 5.21 min, recommended by ADNI (http://www.adni-info.org/). 2.3.2. Image analysis The FreeSurfer software package (version 5.1.0, http://surfer.nmr. mgh.harvard.edu) was used for an automated cortical reconstruction of the T1-weighted images. The method used to create a threedimensional cortical surface model of cortical thickness using intensity and continuity information has been previously described in detail [22]. Briefly, the automated processing included removal of non-brain tissue, Talairach transformation, intensity normalization, tessellation of the gray matter/white matter boundary, topology correction and surface deformation to detect gray matter/white matter and gray matter/cerebrospinal fluid boundaries. The resulting representation of cortical thickness was measured as the shortest distance between tissue boundaries (gray matter/white matter and gray matter/cerebrospinal fluid). Thickness measures were mapped to the inflated surface of the reconstructed brain in order to allow for visualization of data across the entire cortical surface without being obscured by cortical folding. To ensure the accuracy of the automated segmentation, each scan was reviewed to check the delineation of gray and white matter differentiation. Where necessary, pial surface correction and/or white matter corrections were made according to the FreeSurfer guidelines. Small pial surface corrections (i.e. over 3–5 slices) were made in 13 preterms and 15 controls and a white matter correction was made in one control child. Morphologically homologous cortical locations were accurately matched across subjects by morphing each reconstructed brain to an average spherical surface representation which optimally aligned sulcal and gyral features across subjects while minimizing metric distortion [23]. To reduce noiseinduced variations and registration errors in measurements, a full width half maximum (FWHM) Gaussian blurring kernel of 15 mm was applied to smooth the thickness estimates. The applied method has been methodologically evaluated and has been applied in various settings, showing reliability even across different scanner platforms in terms of spatial localization and cortical thickness results [24]. 2.3.3. Statistics Pearson's chi square test (IBM SPSS Statistics 21.0) was used to examine group differences for categorical data. Unpaired two-sided t-tests were used to calculate group differences for continuous, normally distributed data. Cohen's delta coefficient d served as a measure of effect size, with r = 0.20 representing a small, r = 0.50 a medium and r = 0.80 a large effect. In order to compare mean global cortical thickness of both groups over the whole hemispheres, unpaired two-sided t-tests were used within SPSS. Associations of mean global cortical thickness with age were calculated for each group with Pearson correlations (data was distributed normally), p b .05 was considered statistically significant. Using the high-resolution surface-based averaging techniques of the FreeSurfer software, thickness maps were averaged within both groups
I. Mürner-Lavanchy et al. / Early Human Development 90 (2014) 443–450
and compared to each other [22]. Regional thickness difference maps were generated using t-tests between groups at each vertex using a general linear model approach as implemented by the FreeSurfer QDEC (query, design, estimate, contrast) application. The linear model was created with the DODS (different offset, different slope) method. A false discovery rate of p b .05 was applied to the difference maps to correct for multiple comparisons. The resulting difference maps show statistically significant differences of cortical thickness at each vertex. Simple linear regression was used in QDEC to analyze age effects on regional cortical thickness within each group. 3. Results 3.1. Sample characteristics Mean gestational age of the very preterm born children was 30.1 weeks (SD = 2.0, range = 25.7–32.0) and mean birth weight was 1294.0 gram (SD = 386.5, range = 570.0–2060.0). Groups were comparable with regard to age (Mpreterms = 9.9 years, SD = 1.6, range = 7.1–12.8, Mcontrols = 9.8 years, SD = 1.7, range = 7.6–12.9; t(69) = −.197, p = .844), sex (preterms 23 girls, 18 boys; controls 17 girls, 13 boys; χ2(1) = .002, p = .962) and handedness (preterms 34 right, 3 ambidexter, 4 left; controls 24 right, 2 ambidexter, 4 left; χ2(2) = .225, p = .893). Mothers' SES differed significantly between groups, with mothers of the control group holding higher educational degrees than mothers of the preterm group (SESmother χ2(2) = 11.357, p = .003; SESfather χ2(3) = 5.914, p = .116). Full-scale IQ differed significantly between groups, with preterm children showing a lower Full-scale IQ than controls (Mpreterms = 102.6, SD = 9.8, Mcontrols = 109.5, SD = 8.1; t(69) = 3.175, p = .002, d = 0.767). Full-scale IQ did not correlate with age in preterms and controls. Correspondingly, young and old preterms and young and old controls did not differ with regard to IQ. 3.2. Age effects on cortical thickness 3.2.1. Correlation of age and cortical thickness In preterms, mean global cortical thickness correlated negatively with age in the left hemisphere and right hemisphere (left r(41) = − .380, p = .014; right r(41) = − .394, p = .011). In controls, mean global cortical thickness did not correlate significantly with age in the left and
445
right hemispheres (left r(41) = −.191, p = .311; right r(41) = −.258, p = .168). On a regional level, the preterm group showed significant negative correlations between cortical thickness and age in right frontal and parietal regions, representing cortical thinning with age in these areas (Fig. 1 blue areas, the regions are listed in detail in Table 1). In controls, no correlations between regional cortical thickness and age were found. Cortical thickness did not correlate with SES and Full-scale IQ in very preterm born and control children. To rule out any possibility that sex, SES or Full-scale IQ influences the results, all analyses were additionally computed with correction for these potentially confounding variables. These additional analyses did not yield significantly different results. Therefore, all following analyses are presented without correction for sex, SES or Full-scale IQ. To further investigate the age effect on cortical thickness in preterms and controls, young and old age groups were built by separating the whole group by a median split (21 young preterms, age range = 7.1–9.9 years; 20 old preterms, age range = 10–12.8 years; 15 young controls, age range = 7.6–9.3 years; 15 old controls, age range = 9.4– 12.9 years). The young and old age groups did not differ significantly with regard to gestational age, birth weight, sex and SES.
3.2.2. Young vs. old preterms and young vs. old controls In preterms, mean global cortical thickness differed between young and old preterms in the left and right hemispheres with young preterms showing a thicker cortex than old preterms (left hemisphere: Myoung = 2.91 mm, SD = 0.09, Mold = 2.79 mm, SD = 0.12; t(39) = 3.485, p = .001, d = 1.131; right hemisphere: Myoung = 2.89 mm, SD = 0.09, Mold = 2.78 mm, SD = 0.12; t(39) = 3.363, p = .002, d = 1.037). In controls, mean global cortical thickness differed between young and old controls neither in the left hemisphere nor in the right hemisphere (left hemisphere: Myoung = 2.83 mm, SD = 0.09, Mold = 2.78 mm, SD = 0.09; t(28) = 1.468, p = .153, d = 0.556; right hemisphere: Myoung = 2.81 mm, SD = 0.09, Mold = 2.75 mm, SD = 0.08; t(28) = 1.899, p = .068, d = 0.705). On a regional level, there were significant cortical thickness differences between the young and old preterms in left parietal and frontal regions and right medial frontal and parietal areas, with old preterms showing a thinner cortex than young preterms (Fig. 2, Table 2). In controls, no differences in regional cortical thickness were found between young and old children.
Fig. 1. Association between cortical thickness and age in preterms. Dark to light blue areas represent negative correlations with age. The color scale represents the dynamic range of statistical change in p-values (FDR corrected; the statistical map is overlaid on the reconstructed surface with inflated brains, dark gray = sulci, light gray = gyri).
446
I. Mürner-Lavanchy et al. / Early Human Development 90 (2014) 443–450
Table 1 Anatomical areas with significant correlations (p b .05) between age and cortical thickness in preterms as determined by the cortical thickness maps. Anatomical areas
Hemisphere
Brain lobe
r
p value
Superior frontal gyrus Middle frontal gyrus Postcentral gyrus Inferior temporal sulcus Precuneus Anterior cingulate gyrus Lingual gyrus
R R R R R R R
Frontal Frontal Frontal Temporal Parietal Frontal Occipital
−0.488 −0.433 −0.396 −0.421 −0.428 −0.318 −0.382
.001 .002 .005 .003 .003 .021 .007
One-sided Pearson's correlations. L = left, R = right.
3.2.3. Young preterms vs. young controls and old preterms vs. old controls In the young age groups, global cortical thickness differed significantly between preterms and controls in the left hemisphere and in the right hemisphere (right hemisphere: Mpreterm = 2.91 mm, SD = 0.09, Mcontrol = 2.83 mm, SD = 0.09; t(34) = − 2.592, p = .014, d = 0.889; right hemisphere: Mpreterm = 2.89 mm, SD = 0.09, Mcontrol = 2.81 mm, SD = 0.09; t(34) = −2.510, p = .017, d = 0.889). In the old age groups, global cortical thickness did not differ in the left and right hemispheres (left: t(34) = − 0.274, p = .786, d = 0.094; right: t(34) = − 0.807, p = .425, d = 0.311). Neither young preterms and young controls nor old preterms and old controls differed significantly with regard to age. On a regional level young preterms had thicker cortex areas than young controls in small areas of the left frontal and right medial frontal and parietal regions (Fig. 3, Table 3). In the old age groups, no regional cortical thickness differences between preterms and controls were found.
3.3. Group differences in cortical thickness Mean global cortical thickness of the right hemisphere differed significantly between preterms and controls, with preterms having a thicker cortex than controls (Mpreterms = 2.84 mm, SD = 0.11, Mcontrols = 2.79 mm, SD = 0.09; t(69) = −2.123, p = .037, d = 0.498). Mean global cortical thickness in the left hemisphere did not differ significantly between preterms and controls (Mpreterms = 2.85 mm, SD = 0.12, Mcontrols = 2.80 mm, SD = 0.09; t(69) = −1.736, p = .087, d = 0.471). On a regional level, thicker cortex in preterms was found mainly in the left medial and superior frontal cortex (Fig. 4, red areas). The preterm group had a significantly thinner cortex than the control group in the left temporal region (Fig. 4, blue areas). In the right hemisphere, thicker cortex in preterms than in controls was found mainly in the parietal and medial frontal regions. However, introducing age as a covariate in this analysis, there remained no significant differences in cortical thickness between preterms and controls. 3.4. Sex differences On a global level, girls of the preterm group had a thicker cortex than preterm boys in the left hemisphere (t(39) = 2.203, p = .034, d = 0.690). No other sex-related differences in global or regional cortical thickness were found in preterms and controls. 4. Discussion To the authors' knowledge, no study has investigated the effect of age on regional cortical thinning in a sample of school-aged very
Fig. 2. Differences in cortical thickness between young and old preterms. Areas with significant differences are shown in color, red to yellow representing a thicker cortex in the young compared to the old children. The color scale represents the dynamic range of statistical difference in p-values (FDR corrected; the statistical map is overlaid on the reconstructed surface with inflated brains, dark gray = sulci, light gray = gyri).
I. Mürner-Lavanchy et al. / Early Human Development 90 (2014) 443–450
447
Table 2 Anatomical areas with significant differences (p b .05) in mean cortical thickness (in mm) between young and old preterms as determined by the cortical thickness maps. Anatomical areas
Hemisphere
Brain lobe
Young preterms (n = 21)
Old preterms (n = 20)
Difference in thickness (%)a
p value
Precuneus Superior frontal gyrus Middle frontal gyrus Pars opercularis Superior frontal gyrus Precuneus Pars orbitalis Lingual gyrus
L L L L R R R R
Parietal Frontal Frontal Frontal Frontal Parietal Frontal Occipital
2.88 3.13 2.69 2.92 2.97 2.85 3.14 2.57
2.75 2.95 2.58 2.82 2.78 2.71 2.88 2.47
−4.51 −5.75 −4.10 −3.43 −6.40 −4.91 −8.28 −3.89
.004 .001 .027 .032 b.001 .004 .001 .018
Two-sided t-test. L = left, R = right. a A negative value indicates that cortex was thicker in the young preterms than in old preterms.
preterm born and term born children so far. We assumed differential age influences on cortical thickness in very preterm and control children and hypothesized that developmental cortical thinning is delayed as a result of prematurity. In preterms, there was global and regional cortical thinning with age in the frontal and parietal areas, whereas in controls, cortical thickness was not associated with age. Up to date, it is unclear whether preterm children exhibit persistent atypical cortical development or whether they show a delay in development, and catch up with their peers with increasing age. In our preterm sample, age was significantly associated with cortical thickness. The older the children, the thinner their cortex mainly in the frontal and parietal regions, indicating that cortical thinning still occurred between the ages of 7 to 12 years in preterm children but not in controls. A study examining cortical thickness in 18- to 21-year-old adults found no
global differences between adults with very low birth weight (b1500 g) and controls but found specific regions of thicker and thinner cortex in the adults with very low birth weight [25]. The authors conclude that persistent cortical thickness deviations exist in preterm born adults. In the present study, however, we examined prematurely born children with a mean birth weight of approximately 1300 g (ranging from 570 g up to 2060 g) and therefore representing a sample with lower risk for neurodevelopmental problems. Additionally, the children examined in this study had no or minimal complications at birth, an IQ in the normal range and moderate to high socioeconomic status. Since not only perinatal variables but also environmental and social factors play a role in cortical development [26], we suggest that the preterm children in our sample are more likely to catch up with normal development across childhood than children with more severe neonatal problems.
Fig. 3. Differences in cortical thickness between young preterms and young controls. Areas with significant differences are shown in color, red to yellow representing a thicker cortex in the young preterm compared to the young controls. The color scale represents the dynamic range of statistical difference in p-values (FDR corrected; the statistical map is overlaid on the reconstructed surface with inflated brains, dark gray = sulci, light gray = gyri).
448
I. Mürner-Lavanchy et al. / Early Human Development 90 (2014) 443–450
Fig. 4. Differences in cortical thickness between preterms and controls. Areas with significant differences are shown in color, red to yellow representing a thicker cortex, dark to light blue representing a thinner cortex in the preterm than in the control group. The color scale represents the dynamic range of statistical difference in p-values (FDR corrected; the statistical map is overlaid on the reconstructed surface with inflated brains, dark gray = sulci, light gray = gyri).
Pruning – as reflected in cortical thinning – may underlie the emergence of functional specialization [27,28]. This functional specialization has been shown on the basis of cognitive data on executive functions previously published using the same study sample. These data indicate a developmental catch-up rather than an ongoing deficit in three of the core executive functions (shifting, working memory and inhibition) in very preterm born children between the ages of 7 to 12 years [29]. The present study is in line with this idea, suggesting that a delay in cortical thinning following prematurity is already present in early childhood, proceeds throughout middle childhood and is caught up on in the beginning of adolescence. It is unknown whether preterms follow the same developmental trajectories as term born children, only proceeding somewhat slower, or if development occurs through different developmental trajectories that serve to compensate structural brain alterations. In our sample of preterm children, the development of frontal and parietal areas is
found to be associated with age, with older preterms showing thinner cortex in these areas than young preterms. In the normal structural development of the brain, primary sensory and motor areas are known to mature first, followed by parietal and frontal higher-order areas [15,30]. We assume that in our sample, thinning has already taken place in early maturing brain regions. In our preterms, parietal and frontal regions still seemed to be in the process of thinning, up to the age of 12, when we no longer observed differences between preterms and controls. Consequently, our data led us to assume that the normal process of cortical thinning is somewhat delayed, but occurs under the same developmental trajectories as in term born controls. Two of the possible reasons for the delay of cortical thinning lie in the early premature development. A first possible reason for the delay of cortical thinning in premature children might be the early exposure to the extrauterine environment affecting brain structure through disturbance of the neuronal proliferation and migration taking place
Table 3 Anatomical areas with significant differences (p b .05) in mean cortical thickness (in mm) between young controls (n = 15) and young preterms (n = 21) as determined by the cortical thickness maps. Anatomical areas
Hemisphere
Brain lobe
Young controls (n = 15)
Young preterms (n = 21)
Difference in thickness (%)a
p value
Lateral orbitofrontal cortex Precentral gyrus Superior frontal gyrus Medial orbitofrontal gyrus Medial orbitofrontal cortex Superior parietal lobe Inferior temporal gyrus
L L L L R R R
Frontal Frontal Frontal Frontal Frontal Parietal Temporal
3.11 2.62 2.98 2.74 2.68 2.43 3.33
3.25 2.72 3.13 2.94 2.87 2.57 3.47
−4.50 −3.82 −5.03 −7.30 −7.09 −5.76 −4.20
.014 .010 .001 .008 .016 b.001 .015
Two-sided t-test. L = left, R = right. a A negative value indicates that cortex was thicker in the young preterms than in young controls.
I. Mürner-Lavanchy et al. / Early Human Development 90 (2014) 443–450
between the 24th and 40th weeks of gestation. During this time, axons from the thalamus, corpus callosum and other cortical sites synapse on subplate neurons and then enter the cortex, forming connections in the deep cortical plate. Because of its complexity and rapidity, this process is particularly liable to exogenous and endogenous events and hence the process is at risk in children born very preterm [31]. A second factor possibly associated with a delay of cortical thinning might be the preand postnatal administration of steroids to premature babies [32]. Steroid treatment during critical periods of brain development has been shown to impair myelination and brain cell division [33]. In animal studies, steroid injections during the first weeks of life enhanced the maturation of cholinergic neurons [34] and reduced the concentrations of microtubule-associated proteins, which are involved during late neurogenesis [35]. Thus, disturbances in neuronal development due to steroid treatment might be a reason for delay in cortical thinning in our preterm sample, given that 31 of the examined preterms received prenatal and four preterms additionally received postnatal steroid treatment. Taken together, exposure to physical and psychological stressors in early life might contribute to alterations in cortical thinning. The analyses showed that preterm and control children had global and regional differences in cortical thickness. However, when age was considered as a covariate, differences between the groups disappeared. Consequently, the differences in preterms and controls were mainly caused by the influence of age, indicating that differences in cortical thickness were not a result of the prematurity per se. Several studies [7,36,37] found differences in cortical thickness between preterms and controls. An MRI study found cortical thickness differences in preterm children without significant post-natal medical problems and control children in the age range of 8 to 10 years, without focusing on the particular influence of age [7]. According to our results, the catch-up of cortical thinning in children without major neonatal complications seems to take place around the ages of 10 to 12 years. Consequently, the children in the above-mentioned study might not yet have caught up with their term born peers [7]. In adolescents with very low birth weight (b1500 g) areas of thinner and thicker cortex in the preterm children compared to controls were found at the age of 15 [36,37]. It is possible that the lower birth weight of these children modifies cortical development to such an extent that persistent cortical deviations are the consequence. In the control children of the present study, there was no linear association between age and cortical thickness nor did cortical thickness differ between young and old control children on a global or regional level. A possible interpretation of these data is that the process of cortical thinning in the control children continues with such a minimal intensity that changes in cortical thickness cannot be detected with the MRI-based assessment used in the present study. There is reason to assume that even in our sample of control children, cortical thinning has not yet come to completion, since normal cortical thinning is thought to continue up to adolescence [30]. Indeed, gray matter thinning in the right frontal and bilateral parieto-occipital regions has been found in term born children between 5 and 11 years of age [18]. The voxel-based method used by these authors might have been more sensitive to subtle changes in cortical thickness than the method used in the present study. 4.1. Limitations As mentioned earlier, the sample examined in this study had no or minimal brain lesions at birth, no or minimal neurodevelopmental impairment and an IQ above 85. It is therefore not representative of the total of very preterm born children with more severe neonatal complications such as brain lesions and their more detrimental consequences for structural and functional outcome [4]. Moreover, as an indirect measure of cortical thickness, the surface based reconstruction does not yield the same morphological information as histology. Still, cortical thickness measurements based on surface reconstruction of
449
MRI images are useful markers of cortical morphology, because they allow in-vivo measurements of cortical changes. Finally, the results of this study are based on cross-sectional data, whereas longitudinal studies are required to give detailed insight into the process of cortical thinning across childhood. 5. Conclusion Our study reports differential influences of age on cortical thickness in preterms when compared to controls. In preterms, global and regional cortical thickness in the frontal and parietal areas decreased with age, whereas in controls, cortical thickness was not associated with age. These results point to a developmental delay of cortical thinning likely caused by prematurity. Our data suggest that this developmental delay is caught up on at the beginning of adolescence. Further studies are required to investigate the detailed trajectories of developmental cortical thinning and how they relate to functional outcome in very preterm born children. Conflict of interest None of the authors have any conflict of interest. Acknowledgments The authors gratefully acknowledge the contributions of Caroline Benninger for her help with scanning the children and Martin Zbinden for his ongoing technical support. This work was funded by project grants and fellowships from the Swiss National Science Foundation (PZ00P1_126309, IZK0Z1_137130/1 and PZ00P1_ 143173). References [1] Volpe JJ. Brain injury in premature infants: a complex amalgam of destructive and developmental disturbances. Lancet Neurol 2009;8(1):110–24. [2] Counsell SJ, Boardman JP. Differential brain growth in the infant born preterm: current knowledge and future developments from brain imaging. Semin Fetal Neonatal Med 2005;10(5):403–10. [3] Peterson BS, Vohr B, Staib LH, Cannistraci CJ, Dolberg A, Schneider KC, et al. Regional brain volume abnormalities and long-term cognitive outcome in preterm infants. JAMA 2000;284(15):1939–47. [4] Kesler SR, Ment LR, Vohr B, Pajot SK, Schneider KC, Katz KH, et al. Volumetric analysis of regional cerebral development in preterm children. Pediatr Neurol 2004;31(5):318–25. [5] Nosarti C, Rushe TM, Woodruff PW, Stewart AL, Rifkin L, Murray RM. Corpus callosum size and very preterm birth: relationship to neuropsychological outcome. Brain 2004;127(Pt 9):2080–9. [6] Isaacs EB, Edmonds CJ, Chong WK, Lucas A, Morley R, Gadian DG. Brain morphometry and IQ measurements in preterm children. Brain 2004;127(Pt 12):2595–607. [7] Lax ID, Duerden EG, Lin SY, Mallar Chakravarty M, Donner EJ, Lerch JP, et al. Neuroanatomical consequences of very preterm birth in middle childhood. Brain Struct Funct 2013;218(2):575–85. [8] Gimenez M, Junque C, Narberhaus A, Bargallo N, Botet F, Mercader JM. White matter volume and concentration reductions in adolescents with history of very preterm birth: a voxel-based morphometry study. NeuroImage 2006;32(4):1485–98. [9] Nosarti C, Al-Asady MH, Frangou S, Stewart AL, Rifkin L, Murray RM. Adolescents who were born very preterm have decreased brain volumes. Brain 2002;125(Pt 7):1616–23. [10] Nosarti C, Giouroukou E, Healy E, Rifkin L, Walshe M, Reichenberg A, et al. Grey and white matter distribution in very preterm adolescents mediates neurodevelopmental outcome. Brain 2008;131(Pt 1):205–17. [11] Constable RT, Ment LR, Vohr BR, Kesler SR, Fulbright RK, Lacadie C, et al. Prematurely born children demonstrate white matter microstructural differences at 12 years of age, relative to term control subjects: an investigation of group and gender effects. Pediatrics 2008;121(2):306–16. [12] Counsell SJ, Allsop JM, Harrison MC, Larkman DJ, Kennea NL, Kapellou O, et al. Diffusion-weighted imaging of the brain in preterm infants with focal and diffuse white matter abnormality. Pediatrics 2003;112(1 Pt 1):1–7. [13] la Fougere C, Grant S, Kostikov A, Schirrmacher R, Gravel P, Schipper HM, et al. Where in-vivo imaging meets cytoarchitectonics: the relationship between cortical thickness and neuronal density measured with high-resolution [18F]flumazenilPET. NeuroImage 2011;56(3):951–60. [14] Shaw P, Kabani NJ, Lerch JP, Eckstrand K, Lenroot R, Gogtay N, et al. Neurodevelopmental trajectories of the human cerebral cortex. J Neurosci Off J Soc Neurosci 2008;28(14): 3586–94.
450
I. Mürner-Lavanchy et al. / Early Human Development 90 (2014) 443–450
[15] Huttenlocher PR, Dabholkar AS. Regional differences in synaptogenesis in human cerebral cortex. J Comp Neurol 1997;387(2):167–78. [16] Shankle WR, Rafii MS, Landing BH, Fallon JH. Approximate doubling of numbers of neurons in postnatal human cerebral cortex and in 35 specific cytoarchitectural areas from birth to 72 months. Pediatr Dev Pathol 1999;2(3):244–59. [17] Gogtay N, Giedd JN, Lusk L, Hayashi KM, Greenstein D, Vaituzis AC, et al. Dynamic mapping of human cortical development during childhood through early adulthood. Proc Natl Acad Sci U S A 2004;101(21):8174–9. [18] Sowell ER, Thompson PM, Leonard CM, Welcome SE, Kan E, Toga AW. Longitudinal mapping of cortical thickness and brain growth in normal children. J Neurosci Off J Soc Neurosci 2004;24(38):8223–31. [19] Phillips JP, Montague EQ, Aragon M, Lowe JR, Schrader RM, Ohls RK, et al. Prematurity affects cortical maturation in early childhood. Pediatr Neurol 2011;45(4):213–9. [20] Ment LR, Kesler S, Vohr B, Katz KH, Baumgartner H, Schneider KC, et al. Longitudinal brain volume changes in preterm and term control subjects during late childhood and adolescence. Pediatrics 2009;123(2):503–11. [21] Crawford JR, Anderson V, Rankin PM, MacDonald J. An index-based short-form of the WISC-IV with accompanying analysis of the reliability and abnormality of differences. Br J Clin Psychol 2010;49(Pt 2):235–58. [22] Fischl B, Dale AM. Measuring the thickness of the human cerebral cortex from magnetic resonance images. Proc Natl Acad Sci U S A 2000;97(20):11050–5. [23] Fischl B, Sereno MI, Dale AM. Cortical surface-based analysis. II: inflation, flattening, and a surface-based coordinate system. NeuroImage 1999;9(2):195–207. [24] Dickerson BC, Fenstermacher E, Salat DH, Wolk DA, Maguire RP, Desikan R, et al. Detection of cortical thickness correlates of cognitive performance: reliability across MRI scan sessions, scanners, and field strengths. NeuroImage 2008;39(1):10–8. [25] Bjuland KJ, Lohaugen GC, Martinussen M, Skranes J. Cortical thickness and cognition in very-low-birth-weight late teenagers. Early Hum Dev 2013;89(6):371–80. [26] Petanjek Z, Judas M, Kostovic I, Uylings HB. Lifespan alterations of basal dendritic trees of pyramidal neurons in the human prefrontal cortex: a layer-specific pattern. Cereb Cortex 2008;18(4):915–29.
[27] Casey BJ, Tottenham N, Liston C, Durston S. Imaging the developing brain: what have we learned about cognitive development? Trends Cogn Sci 2005;9(3):104–10. [28] Lu LH, Dapretto M, O'Hare ED, Kan E, McCourt ST, Thompson PM, et al. Relationships between brain activation and brain structure in normally developing children. Cereb Cortex 2009;19(11):2595–604. [29] Ritter BC, Nelle M, Perrig W, Steinlin M, Everts R. Executive functions of children born very preterm—deficit or delay? Eur J Pediatr 2013;172(4):473–83. [30] Giedd JN, Rapoport JL. Structural MRI of pediatric brain development: what have we learned and where are we going? Neuron 2010;67(5):728–34. [31] Volpe JJ. The encephalopathy of prematurity—brain injury and impaired brain development inextricably intertwined. Semin Pediatr Neurol 2009;16(4): 167–78. [32] Baud O. Postnatal steroid treatment and brain development. Arch Dis Child Fetal Neonatal Ed 2004;89(2):F96–F100. [33] Weichsel Jr ME. The therapeutic use of glucocorticoid hormones in the perinatal period: potential neurological hazards. Ann Neurol 1977;2(5):364–6. [34] Flagel SB, Vazquez DM, Watson Jr SJ, Neal Jr CR. Effects of tapering neonatal dexamethasone on rat growth, neurodevelopment, and stress response. Am J Physiol Regul Integr Comp Physiol 2002;282(1):R55–63. [35] Schwab M, Antonow-Schlorke I, Kuhn B, Muller T, Schubert H, Walter B, et al. Effect of antenatal betamethasone treatment on microtubule-associated proteins MAP1B and MAP2 in fetal sheep. J Physiol 2001;530(Pt 3):497–506. [36] Martinussen M, Fischl B, Larsson HB, Skranes J, Kulseng S, Vangberg TR, et al. Cerebral cortex thickness in 15-year-old adolescents with low birth weight measured by an automated MRI-based method. Brain 2005;128(Pt 11):2588–96. [37] Nagy Z, Lagercrantz H, Hutton C. Effects of preterm birth on cortical thickness measured in adolescence. Cereb Cortex 2011;21(2):300–6.
Contents lists available at ScienceDirect
Early Human Development journal homepage: www.elsevier.com/locate/earlhumdev
Delay of cortical thinning in very preterm born children Ines Mürner-Lavanchy a,b,c,e, Maja Steinlin a,c, Mathias Nelle d, Christian Rummel b, Walter J. Perrig c,e, Gerhard Schroth b, Regula Everts a,b,c,⁎ a
Division of Neuropediatrics, Development and Rehabilitation, Children's University Hospital, Inselspital, Bern, Switzerland University Institute of Diagnostic and Interventional Neuroradiology, University Hospital, Inselspital, Bern, Switzerland Centre for Cognition, Learning and Memory, University of Bern, Bern, Switzerland d Division of Neonatology, Children's University Hospital, Inselspital, Bern, Switzerland e Institute of Psychology, University of Bern, Bern, Switzerland b c
a r t i c l e
i n f o
Article history: Received 11 December 2013 Received in revised form 2 May 2014 Accepted 29 May 2014 Keywords: Cortical thickness Very preterm Very low birth weight Children Development Delay
a b s t r a c t Background: Cortical gray matter thinning occurs during childhood due to pruning of inefficient synaptic connections and an increase in myelination. Preterms show alterations in brain structure, with prolonged maturation of the frontal lobes, smaller cortical volumes and reduced white matter volume. These findings give rise to the question if there is a differential influence of age on cortical thinning in preterms compared to controls. Aims: To investigate the relationship between age and cortical thinning in school-aged preterms compared to controls. Study design and outcome measures: The automated surface reconstruction software FreeSurfer was applied to obtain measurements of cortical thickness based on T1-weighted MRI images. Subjects: Forty-one preterms (b32 weeks gestational age and/or b 1500 g birth weight) and 30 controls were included in the study (7–12 years). Results: In preterms, age correlated negatively with cortical thickness in right frontal, parietal and inferior temporal regions. Furthermore, young preterms showed a thicker cortex compared to old preterms in bilateral frontal, parietal and temporal regions. In controls, age was not associated with cortical thickness. Conclusion: In preterms, cortical thinning still seems to occur between the age of 7 and 12 years, mainly in frontal and parietal areas whereas in controls, a substantial part of cortical thinning appears to be completed before they reach the age of 7 years. These data indicate slower cortical thinning in preterms than in controls. © 2014 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Important maturational events take place in the third trimester of pregnancy [1]. Since very preterm born children are born in this crucial phase, they are particularly vulnerable to primary injuries and secondary maturational difficulties. Various methods have been used to identify structural alterations in the preterm brain. Studies using magnetic resonance imaging (MRI) have reported lower cortical gray and white matter volumes, lower cerebellar volumes as well as smaller corpus callosum and hippocampus size in preterm born infants [2,3], children [4–7] and adolescents [8,9]. Furthermore, a smaller cortical surface area was observed in 7–10 year-old preterm children when compared to agematched term born controls [7]. A voxel-based morphometry study found atypical gray and white matter distribution in preterm adolescents aged 14–15 years [10]. Diffusion tensor imaging studies detected alterations in white matter fiber tract organization throughout the brain, ⁎ Corresponding author at: Division of Neuropediatrics, Development and Rehabilitation, Children's University Hospital, Inselspital, 3010 Bern, Switzerland. Tel.: +41 31 632 41 30. E-mail address: [email protected] (R. Everts).
http://dx.doi.org/10.1016/j.earlhumdev.2014.05.013 0378-3782/© 2014 Elsevier Ireland Ltd. All rights reserved.
suggesting differences in structural connectivity [11] as well as widespread microstructural white matter abnormalities in very preterm born children compared to same-aged term born controls [12]. The structural alterations in very preterm born children raise the question if the process of cortical thinning differs between preterms and controls. Cortical thickness is thought to be an indicator of the number of neurons per cortical column (groups of neurons which connect the six horizontal layers of the neocortex vertically) as well as glial support and dendritic arborization [13]. From early childhood to adolescence, decrease of cortical thickness co-occurs with the pruning of dispensable neurons and synapses [14]. This process leads to more efficient synaptic connections. The normal developmental cortical thinning does not occur simultaneously over the whole cortex. Findings concerning synaptic density, which is indirectly related to cortical thickness, suggest that during the course of development a synaptic loss occurs first in primary sensory and motor regions and later in multimodal association areas [15,16]. More recent studies confirmed this pattern using neuroimaging of cortical thinning in healthy children [14,17], others found contradictory results. Correspondingly, a longitudinal MRI study found not only regional specific thinning but also thickening in circumscribed perisylvian language relevant regions in 45 healthy
444
I. Mürner-Lavanchy et al. / Early Human Development 90 (2014) 443–450
children between 5 and 11 years of age [18]. Since the process of cortical thinning varies over different cerebral regions, it is important to examine cortical thickness over the whole cortex and on a regional level. A cross-sectional study investigated the development of cortical thinning in children with very low birth weight and term born controls at 18–22 months and 3–4 years of age [19]. At both age groups, children with very low birth weight displayed greater mean cortical thickness compared to control children. Although the difference did not reach statistical significance, the authors suggest that their results point to a delay in normal cortical thinning following prematurity, up to the age of four years. Correspondingly, a longitudinal study comparing 55 preterm children (born with 600 to 1250 g birth weight) with 20 term born control children at 8 and 12 years of age, identified different gray and white matter volume changes over time between the study groups: In preterms, gray matter volumes decreased and white matter volumes increased less than in controls, which suggests a different gray and white matter development in preterms than in controls [20]. To the authors' knowledge, no study has investigated the effect of age on global and regional cortical thinning in a sample of school-aged very preterm born children in a cross-sectional study design. Based on the existing findings, we hypothesize that age is differentially associated with cortical thinning in preterms and controls between 7 and 12 years of age. We further assume that the normal developmental cortical thinning is delayed in very preterm born children compared to their term born peers, even in our sample of relatively healthy very preterm born children. 2. Methods This study reports on a subset of data from the NEMO (NEuropsychology and meMOry) research project at the Children's University Hospital in Bern, Switzerland. The NEMO project examines cognitive development in very preterm born children including behavioral and neuroimaging data. The study protocol was approved by the local ethics committee. All children and caregivers provided informed written consent to the research and publication of the results prior to participation, consistent with the Code of Ethics of the World Medical Association (Declaration of Helsinki). 2.1. Very preterm group Medical reports of all very preterm born children (b 32 weeks of gestation and/or b 1500 g birth weight) born between 1998 and 2003 at the Children's University Hospital Bern, Switzerland were reviewed. Native German speakers aged 7 to 12 years with normal neonatal ultrasound, no or mild periventricular leukomalacia (grade I or II), no or mild neonatal cerebral lesions (hemorrhage grade I), no chronic illness, no pervasive developmental disorders, and Full-scale IQ N 85 in the neuropsychological follow-ups were included. A total of 247 children fulfilled the inclusion criteria and were contacted by letter out of which 75 children agreed to participate in the study. Fifty-five very preterm born children completed the MRI assessment. Fourteen preterms had to be excluded (technical problems n = 4, movement n = 10) resulting in 41 preterms for inclusion in the analysis. 2.2. Control group Term born controls from the same year cohorts were recruited by means of announcements in the hospital. Forty-two children completed the MRI examination, twelve children were excluded (technical problems n = 2 and movement, n = 10), resulting in 30 controls included in the analysis. Handedness was enquired by telephone interview prior to the first assessment. Socioeconomic status (SES) was defined as mother's and father's education level at the time of the neuropsychological assessment (no graduation = 1, college = 2, college of higher education =
3, university degree = 4). Full-scale IQ was assessed using the short form of the German version of the ‘Wechsler Intelligence Scale for Children, Fourth Edition’ [21]. 2.3. MR imaging Children underwent a one-hour MRI assessment at the Department of Diagnostic and Interventional Neuroradiology, University Hospital Bern. 2.3.1. Image acquisition MRI was performed on a Verio3-T whole body scanner (Siemens Erlangen, Germany) equipped with a 40 mT/m gradient system and a CP standard head coil (12 channels). The scanner was equipped with the Syngo MR 2002B (VA17) software. Anatomical imaging was obtained using a T1-weighted, 3D-MPRAGE sequence (TR 2300 ms, TE 2.98 ms, TI 900 ms, 0 mm gap, FoV 256, 1 mm voxel resolution, 160 contiguous sagittal slices) with an acquisition duration of 5.21 min, recommended by ADNI (http://www.adni-info.org/). 2.3.2. Image analysis The FreeSurfer software package (version 5.1.0, http://surfer.nmr. mgh.harvard.edu) was used for an automated cortical reconstruction of the T1-weighted images. The method used to create a threedimensional cortical surface model of cortical thickness using intensity and continuity information has been previously described in detail [22]. Briefly, the automated processing included removal of non-brain tissue, Talairach transformation, intensity normalization, tessellation of the gray matter/white matter boundary, topology correction and surface deformation to detect gray matter/white matter and gray matter/cerebrospinal fluid boundaries. The resulting representation of cortical thickness was measured as the shortest distance between tissue boundaries (gray matter/white matter and gray matter/cerebrospinal fluid). Thickness measures were mapped to the inflated surface of the reconstructed brain in order to allow for visualization of data across the entire cortical surface without being obscured by cortical folding. To ensure the accuracy of the automated segmentation, each scan was reviewed to check the delineation of gray and white matter differentiation. Where necessary, pial surface correction and/or white matter corrections were made according to the FreeSurfer guidelines. Small pial surface corrections (i.e. over 3–5 slices) were made in 13 preterms and 15 controls and a white matter correction was made in one control child. Morphologically homologous cortical locations were accurately matched across subjects by morphing each reconstructed brain to an average spherical surface representation which optimally aligned sulcal and gyral features across subjects while minimizing metric distortion [23]. To reduce noiseinduced variations and registration errors in measurements, a full width half maximum (FWHM) Gaussian blurring kernel of 15 mm was applied to smooth the thickness estimates. The applied method has been methodologically evaluated and has been applied in various settings, showing reliability even across different scanner platforms in terms of spatial localization and cortical thickness results [24]. 2.3.3. Statistics Pearson's chi square test (IBM SPSS Statistics 21.0) was used to examine group differences for categorical data. Unpaired two-sided t-tests were used to calculate group differences for continuous, normally distributed data. Cohen's delta coefficient d served as a measure of effect size, with r = 0.20 representing a small, r = 0.50 a medium and r = 0.80 a large effect. In order to compare mean global cortical thickness of both groups over the whole hemispheres, unpaired two-sided t-tests were used within SPSS. Associations of mean global cortical thickness with age were calculated for each group with Pearson correlations (data was distributed normally), p b .05 was considered statistically significant. Using the high-resolution surface-based averaging techniques of the FreeSurfer software, thickness maps were averaged within both groups
I. Mürner-Lavanchy et al. / Early Human Development 90 (2014) 443–450
and compared to each other [22]. Regional thickness difference maps were generated using t-tests between groups at each vertex using a general linear model approach as implemented by the FreeSurfer QDEC (query, design, estimate, contrast) application. The linear model was created with the DODS (different offset, different slope) method. A false discovery rate of p b .05 was applied to the difference maps to correct for multiple comparisons. The resulting difference maps show statistically significant differences of cortical thickness at each vertex. Simple linear regression was used in QDEC to analyze age effects on regional cortical thickness within each group. 3. Results 3.1. Sample characteristics Mean gestational age of the very preterm born children was 30.1 weeks (SD = 2.0, range = 25.7–32.0) and mean birth weight was 1294.0 gram (SD = 386.5, range = 570.0–2060.0). Groups were comparable with regard to age (Mpreterms = 9.9 years, SD = 1.6, range = 7.1–12.8, Mcontrols = 9.8 years, SD = 1.7, range = 7.6–12.9; t(69) = −.197, p = .844), sex (preterms 23 girls, 18 boys; controls 17 girls, 13 boys; χ2(1) = .002, p = .962) and handedness (preterms 34 right, 3 ambidexter, 4 left; controls 24 right, 2 ambidexter, 4 left; χ2(2) = .225, p = .893). Mothers' SES differed significantly between groups, with mothers of the control group holding higher educational degrees than mothers of the preterm group (SESmother χ2(2) = 11.357, p = .003; SESfather χ2(3) = 5.914, p = .116). Full-scale IQ differed significantly between groups, with preterm children showing a lower Full-scale IQ than controls (Mpreterms = 102.6, SD = 9.8, Mcontrols = 109.5, SD = 8.1; t(69) = 3.175, p = .002, d = 0.767). Full-scale IQ did not correlate with age in preterms and controls. Correspondingly, young and old preterms and young and old controls did not differ with regard to IQ. 3.2. Age effects on cortical thickness 3.2.1. Correlation of age and cortical thickness In preterms, mean global cortical thickness correlated negatively with age in the left hemisphere and right hemisphere (left r(41) = − .380, p = .014; right r(41) = − .394, p = .011). In controls, mean global cortical thickness did not correlate significantly with age in the left and
445
right hemispheres (left r(41) = −.191, p = .311; right r(41) = −.258, p = .168). On a regional level, the preterm group showed significant negative correlations between cortical thickness and age in right frontal and parietal regions, representing cortical thinning with age in these areas (Fig. 1 blue areas, the regions are listed in detail in Table 1). In controls, no correlations between regional cortical thickness and age were found. Cortical thickness did not correlate with SES and Full-scale IQ in very preterm born and control children. To rule out any possibility that sex, SES or Full-scale IQ influences the results, all analyses were additionally computed with correction for these potentially confounding variables. These additional analyses did not yield significantly different results. Therefore, all following analyses are presented without correction for sex, SES or Full-scale IQ. To further investigate the age effect on cortical thickness in preterms and controls, young and old age groups were built by separating the whole group by a median split (21 young preterms, age range = 7.1–9.9 years; 20 old preterms, age range = 10–12.8 years; 15 young controls, age range = 7.6–9.3 years; 15 old controls, age range = 9.4– 12.9 years). The young and old age groups did not differ significantly with regard to gestational age, birth weight, sex and SES.
3.2.2. Young vs. old preterms and young vs. old controls In preterms, mean global cortical thickness differed between young and old preterms in the left and right hemispheres with young preterms showing a thicker cortex than old preterms (left hemisphere: Myoung = 2.91 mm, SD = 0.09, Mold = 2.79 mm, SD = 0.12; t(39) = 3.485, p = .001, d = 1.131; right hemisphere: Myoung = 2.89 mm, SD = 0.09, Mold = 2.78 mm, SD = 0.12; t(39) = 3.363, p = .002, d = 1.037). In controls, mean global cortical thickness differed between young and old controls neither in the left hemisphere nor in the right hemisphere (left hemisphere: Myoung = 2.83 mm, SD = 0.09, Mold = 2.78 mm, SD = 0.09; t(28) = 1.468, p = .153, d = 0.556; right hemisphere: Myoung = 2.81 mm, SD = 0.09, Mold = 2.75 mm, SD = 0.08; t(28) = 1.899, p = .068, d = 0.705). On a regional level, there were significant cortical thickness differences between the young and old preterms in left parietal and frontal regions and right medial frontal and parietal areas, with old preterms showing a thinner cortex than young preterms (Fig. 2, Table 2). In controls, no differences in regional cortical thickness were found between young and old children.
Fig. 1. Association between cortical thickness and age in preterms. Dark to light blue areas represent negative correlations with age. The color scale represents the dynamic range of statistical change in p-values (FDR corrected; the statistical map is overlaid on the reconstructed surface with inflated brains, dark gray = sulci, light gray = gyri).
446
I. Mürner-Lavanchy et al. / Early Human Development 90 (2014) 443–450
Table 1 Anatomical areas with significant correlations (p b .05) between age and cortical thickness in preterms as determined by the cortical thickness maps. Anatomical areas
Hemisphere
Brain lobe
r
p value
Superior frontal gyrus Middle frontal gyrus Postcentral gyrus Inferior temporal sulcus Precuneus Anterior cingulate gyrus Lingual gyrus
R R R R R R R
Frontal Frontal Frontal Temporal Parietal Frontal Occipital
−0.488 −0.433 −0.396 −0.421 −0.428 −0.318 −0.382
.001 .002 .005 .003 .003 .021 .007
One-sided Pearson's correlations. L = left, R = right.
3.2.3. Young preterms vs. young controls and old preterms vs. old controls In the young age groups, global cortical thickness differed significantly between preterms and controls in the left hemisphere and in the right hemisphere (right hemisphere: Mpreterm = 2.91 mm, SD = 0.09, Mcontrol = 2.83 mm, SD = 0.09; t(34) = − 2.592, p = .014, d = 0.889; right hemisphere: Mpreterm = 2.89 mm, SD = 0.09, Mcontrol = 2.81 mm, SD = 0.09; t(34) = −2.510, p = .017, d = 0.889). In the old age groups, global cortical thickness did not differ in the left and right hemispheres (left: t(34) = − 0.274, p = .786, d = 0.094; right: t(34) = − 0.807, p = .425, d = 0.311). Neither young preterms and young controls nor old preterms and old controls differed significantly with regard to age. On a regional level young preterms had thicker cortex areas than young controls in small areas of the left frontal and right medial frontal and parietal regions (Fig. 3, Table 3). In the old age groups, no regional cortical thickness differences between preterms and controls were found.
3.3. Group differences in cortical thickness Mean global cortical thickness of the right hemisphere differed significantly between preterms and controls, with preterms having a thicker cortex than controls (Mpreterms = 2.84 mm, SD = 0.11, Mcontrols = 2.79 mm, SD = 0.09; t(69) = −2.123, p = .037, d = 0.498). Mean global cortical thickness in the left hemisphere did not differ significantly between preterms and controls (Mpreterms = 2.85 mm, SD = 0.12, Mcontrols = 2.80 mm, SD = 0.09; t(69) = −1.736, p = .087, d = 0.471). On a regional level, thicker cortex in preterms was found mainly in the left medial and superior frontal cortex (Fig. 4, red areas). The preterm group had a significantly thinner cortex than the control group in the left temporal region (Fig. 4, blue areas). In the right hemisphere, thicker cortex in preterms than in controls was found mainly in the parietal and medial frontal regions. However, introducing age as a covariate in this analysis, there remained no significant differences in cortical thickness between preterms and controls. 3.4. Sex differences On a global level, girls of the preterm group had a thicker cortex than preterm boys in the left hemisphere (t(39) = 2.203, p = .034, d = 0.690). No other sex-related differences in global or regional cortical thickness were found in preterms and controls. 4. Discussion To the authors' knowledge, no study has investigated the effect of age on regional cortical thinning in a sample of school-aged very
Fig. 2. Differences in cortical thickness between young and old preterms. Areas with significant differences are shown in color, red to yellow representing a thicker cortex in the young compared to the old children. The color scale represents the dynamic range of statistical difference in p-values (FDR corrected; the statistical map is overlaid on the reconstructed surface with inflated brains, dark gray = sulci, light gray = gyri).
I. Mürner-Lavanchy et al. / Early Human Development 90 (2014) 443–450
447
Table 2 Anatomical areas with significant differences (p b .05) in mean cortical thickness (in mm) between young and old preterms as determined by the cortical thickness maps. Anatomical areas
Hemisphere
Brain lobe
Young preterms (n = 21)
Old preterms (n = 20)
Difference in thickness (%)a
p value
Precuneus Superior frontal gyrus Middle frontal gyrus Pars opercularis Superior frontal gyrus Precuneus Pars orbitalis Lingual gyrus
L L L L R R R R
Parietal Frontal Frontal Frontal Frontal Parietal Frontal Occipital
2.88 3.13 2.69 2.92 2.97 2.85 3.14 2.57
2.75 2.95 2.58 2.82 2.78 2.71 2.88 2.47
−4.51 −5.75 −4.10 −3.43 −6.40 −4.91 −8.28 −3.89
.004 .001 .027 .032 b.001 .004 .001 .018
Two-sided t-test. L = left, R = right. a A negative value indicates that cortex was thicker in the young preterms than in old preterms.
preterm born and term born children so far. We assumed differential age influences on cortical thickness in very preterm and control children and hypothesized that developmental cortical thinning is delayed as a result of prematurity. In preterms, there was global and regional cortical thinning with age in the frontal and parietal areas, whereas in controls, cortical thickness was not associated with age. Up to date, it is unclear whether preterm children exhibit persistent atypical cortical development or whether they show a delay in development, and catch up with their peers with increasing age. In our preterm sample, age was significantly associated with cortical thickness. The older the children, the thinner their cortex mainly in the frontal and parietal regions, indicating that cortical thinning still occurred between the ages of 7 to 12 years in preterm children but not in controls. A study examining cortical thickness in 18- to 21-year-old adults found no
global differences between adults with very low birth weight (b1500 g) and controls but found specific regions of thicker and thinner cortex in the adults with very low birth weight [25]. The authors conclude that persistent cortical thickness deviations exist in preterm born adults. In the present study, however, we examined prematurely born children with a mean birth weight of approximately 1300 g (ranging from 570 g up to 2060 g) and therefore representing a sample with lower risk for neurodevelopmental problems. Additionally, the children examined in this study had no or minimal complications at birth, an IQ in the normal range and moderate to high socioeconomic status. Since not only perinatal variables but also environmental and social factors play a role in cortical development [26], we suggest that the preterm children in our sample are more likely to catch up with normal development across childhood than children with more severe neonatal problems.
Fig. 3. Differences in cortical thickness between young preterms and young controls. Areas with significant differences are shown in color, red to yellow representing a thicker cortex in the young preterm compared to the young controls. The color scale represents the dynamic range of statistical difference in p-values (FDR corrected; the statistical map is overlaid on the reconstructed surface with inflated brains, dark gray = sulci, light gray = gyri).
448
I. Mürner-Lavanchy et al. / Early Human Development 90 (2014) 443–450
Fig. 4. Differences in cortical thickness between preterms and controls. Areas with significant differences are shown in color, red to yellow representing a thicker cortex, dark to light blue representing a thinner cortex in the preterm than in the control group. The color scale represents the dynamic range of statistical difference in p-values (FDR corrected; the statistical map is overlaid on the reconstructed surface with inflated brains, dark gray = sulci, light gray = gyri).
Pruning – as reflected in cortical thinning – may underlie the emergence of functional specialization [27,28]. This functional specialization has been shown on the basis of cognitive data on executive functions previously published using the same study sample. These data indicate a developmental catch-up rather than an ongoing deficit in three of the core executive functions (shifting, working memory and inhibition) in very preterm born children between the ages of 7 to 12 years [29]. The present study is in line with this idea, suggesting that a delay in cortical thinning following prematurity is already present in early childhood, proceeds throughout middle childhood and is caught up on in the beginning of adolescence. It is unknown whether preterms follow the same developmental trajectories as term born children, only proceeding somewhat slower, or if development occurs through different developmental trajectories that serve to compensate structural brain alterations. In our sample of preterm children, the development of frontal and parietal areas is
found to be associated with age, with older preterms showing thinner cortex in these areas than young preterms. In the normal structural development of the brain, primary sensory and motor areas are known to mature first, followed by parietal and frontal higher-order areas [15,30]. We assume that in our sample, thinning has already taken place in early maturing brain regions. In our preterms, parietal and frontal regions still seemed to be in the process of thinning, up to the age of 12, when we no longer observed differences between preterms and controls. Consequently, our data led us to assume that the normal process of cortical thinning is somewhat delayed, but occurs under the same developmental trajectories as in term born controls. Two of the possible reasons for the delay of cortical thinning lie in the early premature development. A first possible reason for the delay of cortical thinning in premature children might be the early exposure to the extrauterine environment affecting brain structure through disturbance of the neuronal proliferation and migration taking place
Table 3 Anatomical areas with significant differences (p b .05) in mean cortical thickness (in mm) between young controls (n = 15) and young preterms (n = 21) as determined by the cortical thickness maps. Anatomical areas
Hemisphere
Brain lobe
Young controls (n = 15)
Young preterms (n = 21)
Difference in thickness (%)a
p value
Lateral orbitofrontal cortex Precentral gyrus Superior frontal gyrus Medial orbitofrontal gyrus Medial orbitofrontal cortex Superior parietal lobe Inferior temporal gyrus
L L L L R R R
Frontal Frontal Frontal Frontal Frontal Parietal Temporal
3.11 2.62 2.98 2.74 2.68 2.43 3.33
3.25 2.72 3.13 2.94 2.87 2.57 3.47
−4.50 −3.82 −5.03 −7.30 −7.09 −5.76 −4.20
.014 .010 .001 .008 .016 b.001 .015
Two-sided t-test. L = left, R = right. a A negative value indicates that cortex was thicker in the young preterms than in young controls.
I. Mürner-Lavanchy et al. / Early Human Development 90 (2014) 443–450
between the 24th and 40th weeks of gestation. During this time, axons from the thalamus, corpus callosum and other cortical sites synapse on subplate neurons and then enter the cortex, forming connections in the deep cortical plate. Because of its complexity and rapidity, this process is particularly liable to exogenous and endogenous events and hence the process is at risk in children born very preterm [31]. A second factor possibly associated with a delay of cortical thinning might be the preand postnatal administration of steroids to premature babies [32]. Steroid treatment during critical periods of brain development has been shown to impair myelination and brain cell division [33]. In animal studies, steroid injections during the first weeks of life enhanced the maturation of cholinergic neurons [34] and reduced the concentrations of microtubule-associated proteins, which are involved during late neurogenesis [35]. Thus, disturbances in neuronal development due to steroid treatment might be a reason for delay in cortical thinning in our preterm sample, given that 31 of the examined preterms received prenatal and four preterms additionally received postnatal steroid treatment. Taken together, exposure to physical and psychological stressors in early life might contribute to alterations in cortical thinning. The analyses showed that preterm and control children had global and regional differences in cortical thickness. However, when age was considered as a covariate, differences between the groups disappeared. Consequently, the differences in preterms and controls were mainly caused by the influence of age, indicating that differences in cortical thickness were not a result of the prematurity per se. Several studies [7,36,37] found differences in cortical thickness between preterms and controls. An MRI study found cortical thickness differences in preterm children without significant post-natal medical problems and control children in the age range of 8 to 10 years, without focusing on the particular influence of age [7]. According to our results, the catch-up of cortical thinning in children without major neonatal complications seems to take place around the ages of 10 to 12 years. Consequently, the children in the above-mentioned study might not yet have caught up with their term born peers [7]. In adolescents with very low birth weight (b1500 g) areas of thinner and thicker cortex in the preterm children compared to controls were found at the age of 15 [36,37]. It is possible that the lower birth weight of these children modifies cortical development to such an extent that persistent cortical deviations are the consequence. In the control children of the present study, there was no linear association between age and cortical thickness nor did cortical thickness differ between young and old control children on a global or regional level. A possible interpretation of these data is that the process of cortical thinning in the control children continues with such a minimal intensity that changes in cortical thickness cannot be detected with the MRI-based assessment used in the present study. There is reason to assume that even in our sample of control children, cortical thinning has not yet come to completion, since normal cortical thinning is thought to continue up to adolescence [30]. Indeed, gray matter thinning in the right frontal and bilateral parieto-occipital regions has been found in term born children between 5 and 11 years of age [18]. The voxel-based method used by these authors might have been more sensitive to subtle changes in cortical thickness than the method used in the present study. 4.1. Limitations As mentioned earlier, the sample examined in this study had no or minimal brain lesions at birth, no or minimal neurodevelopmental impairment and an IQ above 85. It is therefore not representative of the total of very preterm born children with more severe neonatal complications such as brain lesions and their more detrimental consequences for structural and functional outcome [4]. Moreover, as an indirect measure of cortical thickness, the surface based reconstruction does not yield the same morphological information as histology. Still, cortical thickness measurements based on surface reconstruction of
449
MRI images are useful markers of cortical morphology, because they allow in-vivo measurements of cortical changes. Finally, the results of this study are based on cross-sectional data, whereas longitudinal studies are required to give detailed insight into the process of cortical thinning across childhood. 5. Conclusion Our study reports differential influences of age on cortical thickness in preterms when compared to controls. In preterms, global and regional cortical thickness in the frontal and parietal areas decreased with age, whereas in controls, cortical thickness was not associated with age. These results point to a developmental delay of cortical thinning likely caused by prematurity. Our data suggest that this developmental delay is caught up on at the beginning of adolescence. Further studies are required to investigate the detailed trajectories of developmental cortical thinning and how they relate to functional outcome in very preterm born children. Conflict of interest None of the authors have any conflict of interest. Acknowledgments The authors gratefully acknowledge the contributions of Caroline Benninger for her help with scanning the children and Martin Zbinden for his ongoing technical support. This work was funded by project grants and fellowships from the Swiss National Science Foundation (PZ00P1_126309, IZK0Z1_137130/1 and PZ00P1_ 143173). References [1] Volpe JJ. Brain injury in premature infants: a complex amalgam of destructive and developmental disturbances. Lancet Neurol 2009;8(1):110–24. [2] Counsell SJ, Boardman JP. Differential brain growth in the infant born preterm: current knowledge and future developments from brain imaging. Semin Fetal Neonatal Med 2005;10(5):403–10. [3] Peterson BS, Vohr B, Staib LH, Cannistraci CJ, Dolberg A, Schneider KC, et al. Regional brain volume abnormalities and long-term cognitive outcome in preterm infants. JAMA 2000;284(15):1939–47. [4] Kesler SR, Ment LR, Vohr B, Pajot SK, Schneider KC, Katz KH, et al. Volumetric analysis of regional cerebral development in preterm children. Pediatr Neurol 2004;31(5):318–25. [5] Nosarti C, Rushe TM, Woodruff PW, Stewart AL, Rifkin L, Murray RM. Corpus callosum size and very preterm birth: relationship to neuropsychological outcome. Brain 2004;127(Pt 9):2080–9. [6] Isaacs EB, Edmonds CJ, Chong WK, Lucas A, Morley R, Gadian DG. Brain morphometry and IQ measurements in preterm children. Brain 2004;127(Pt 12):2595–607. [7] Lax ID, Duerden EG, Lin SY, Mallar Chakravarty M, Donner EJ, Lerch JP, et al. Neuroanatomical consequences of very preterm birth in middle childhood. Brain Struct Funct 2013;218(2):575–85. [8] Gimenez M, Junque C, Narberhaus A, Bargallo N, Botet F, Mercader JM. White matter volume and concentration reductions in adolescents with history of very preterm birth: a voxel-based morphometry study. NeuroImage 2006;32(4):1485–98. [9] Nosarti C, Al-Asady MH, Frangou S, Stewart AL, Rifkin L, Murray RM. Adolescents who were born very preterm have decreased brain volumes. Brain 2002;125(Pt 7):1616–23. [10] Nosarti C, Giouroukou E, Healy E, Rifkin L, Walshe M, Reichenberg A, et al. Grey and white matter distribution in very preterm adolescents mediates neurodevelopmental outcome. Brain 2008;131(Pt 1):205–17. [11] Constable RT, Ment LR, Vohr BR, Kesler SR, Fulbright RK, Lacadie C, et al. Prematurely born children demonstrate white matter microstructural differences at 12 years of age, relative to term control subjects: an investigation of group and gender effects. Pediatrics 2008;121(2):306–16. [12] Counsell SJ, Allsop JM, Harrison MC, Larkman DJ, Kennea NL, Kapellou O, et al. Diffusion-weighted imaging of the brain in preterm infants with focal and diffuse white matter abnormality. Pediatrics 2003;112(1 Pt 1):1–7. [13] la Fougere C, Grant S, Kostikov A, Schirrmacher R, Gravel P, Schipper HM, et al. Where in-vivo imaging meets cytoarchitectonics: the relationship between cortical thickness and neuronal density measured with high-resolution [18F]flumazenilPET. NeuroImage 2011;56(3):951–60. [14] Shaw P, Kabani NJ, Lerch JP, Eckstrand K, Lenroot R, Gogtay N, et al. Neurodevelopmental trajectories of the human cerebral cortex. J Neurosci Off J Soc Neurosci 2008;28(14): 3586–94.
450
I. Mürner-Lavanchy et al. / Early Human Development 90 (2014) 443–450
[15] Huttenlocher PR, Dabholkar AS. Regional differences in synaptogenesis in human cerebral cortex. J Comp Neurol 1997;387(2):167–78. [16] Shankle WR, Rafii MS, Landing BH, Fallon JH. Approximate doubling of numbers of neurons in postnatal human cerebral cortex and in 35 specific cytoarchitectural areas from birth to 72 months. Pediatr Dev Pathol 1999;2(3):244–59. [17] Gogtay N, Giedd JN, Lusk L, Hayashi KM, Greenstein D, Vaituzis AC, et al. Dynamic mapping of human cortical development during childhood through early adulthood. Proc Natl Acad Sci U S A 2004;101(21):8174–9. [18] Sowell ER, Thompson PM, Leonard CM, Welcome SE, Kan E, Toga AW. Longitudinal mapping of cortical thickness and brain growth in normal children. J Neurosci Off J Soc Neurosci 2004;24(38):8223–31. [19] Phillips JP, Montague EQ, Aragon M, Lowe JR, Schrader RM, Ohls RK, et al. Prematurity affects cortical maturation in early childhood. Pediatr Neurol 2011;45(4):213–9. [20] Ment LR, Kesler S, Vohr B, Katz KH, Baumgartner H, Schneider KC, et al. Longitudinal brain volume changes in preterm and term control subjects during late childhood and adolescence. Pediatrics 2009;123(2):503–11. [21] Crawford JR, Anderson V, Rankin PM, MacDonald J. An index-based short-form of the WISC-IV with accompanying analysis of the reliability and abnormality of differences. Br J Clin Psychol 2010;49(Pt 2):235–58. [22] Fischl B, Dale AM. Measuring the thickness of the human cerebral cortex from magnetic resonance images. Proc Natl Acad Sci U S A 2000;97(20):11050–5. [23] Fischl B, Sereno MI, Dale AM. Cortical surface-based analysis. II: inflation, flattening, and a surface-based coordinate system. NeuroImage 1999;9(2):195–207. [24] Dickerson BC, Fenstermacher E, Salat DH, Wolk DA, Maguire RP, Desikan R, et al. Detection of cortical thickness correlates of cognitive performance: reliability across MRI scan sessions, scanners, and field strengths. NeuroImage 2008;39(1):10–8. [25] Bjuland KJ, Lohaugen GC, Martinussen M, Skranes J. Cortical thickness and cognition in very-low-birth-weight late teenagers. Early Hum Dev 2013;89(6):371–80. [26] Petanjek Z, Judas M, Kostovic I, Uylings HB. Lifespan alterations of basal dendritic trees of pyramidal neurons in the human prefrontal cortex: a layer-specific pattern. Cereb Cortex 2008;18(4):915–29.
[27] Casey BJ, Tottenham N, Liston C, Durston S. Imaging the developing brain: what have we learned about cognitive development? Trends Cogn Sci 2005;9(3):104–10. [28] Lu LH, Dapretto M, O'Hare ED, Kan E, McCourt ST, Thompson PM, et al. Relationships between brain activation and brain structure in normally developing children. Cereb Cortex 2009;19(11):2595–604. [29] Ritter BC, Nelle M, Perrig W, Steinlin M, Everts R. Executive functions of children born very preterm—deficit or delay? Eur J Pediatr 2013;172(4):473–83. [30] Giedd JN, Rapoport JL. Structural MRI of pediatric brain development: what have we learned and where are we going? Neuron 2010;67(5):728–34. [31] Volpe JJ. The encephalopathy of prematurity—brain injury and impaired brain development inextricably intertwined. Semin Pediatr Neurol 2009;16(4): 167–78. [32] Baud O. Postnatal steroid treatment and brain development. Arch Dis Child Fetal Neonatal Ed 2004;89(2):F96–F100. [33] Weichsel Jr ME. The therapeutic use of glucocorticoid hormones in the perinatal period: potential neurological hazards. Ann Neurol 1977;2(5):364–6. [34] Flagel SB, Vazquez DM, Watson Jr SJ, Neal Jr CR. Effects of tapering neonatal dexamethasone on rat growth, neurodevelopment, and stress response. Am J Physiol Regul Integr Comp Physiol 2002;282(1):R55–63. [35] Schwab M, Antonow-Schlorke I, Kuhn B, Muller T, Schubert H, Walter B, et al. Effect of antenatal betamethasone treatment on microtubule-associated proteins MAP1B and MAP2 in fetal sheep. J Physiol 2001;530(Pt 3):497–506. [36] Martinussen M, Fischl B, Larsson HB, Skranes J, Kulseng S, Vangberg TR, et al. Cerebral cortex thickness in 15-year-old adolescents with low birth weight measured by an automated MRI-based method. Brain 2005;128(Pt 11):2588–96. [37] Nagy Z, Lagercrantz H, Hutton C. Effects of preterm birth on cortical thickness measured in adolescence. Cereb Cortex 2011;21(2):300–6.
