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
OBJECTIVE. The goal was to use diffusion tensor imaging to test the hypothesis that prematurely born children demonstrate long-term, white matter, microstructural differences, relative to term control subjects.
METHODS. Twenty-nine preterm subjects (birth weight: 600–1250 g) without neonatal brain injury and 22 matched, term, control subjects were evaluated at 12 years of age with MRI studies, including diffusion tensor imaging and volumetric imaging; voxel-based morphometric strategies were used to corroborate regional diffusion tensor imaging results. Subjects also underwent neurodevelopmental assessments.
RESULTS. Neurodevelopmental assessments showed significant differences in full-scale, verbal, and performance IQ and Developmental Test of Visual Motor Integration scores between the preterm and term control subjects. Diffusion tensor imaging studies demonstrated widespread decreases in fractional anisotropy (a measure of fiber tract organization) in the preterm children, compared with the control subjects. Regions included both intrahemispheric association fibers subserving language skills, namely, the right inferior frontooccipital fasciculus and anterior portions of the uncinate fasciculi bilaterally, and the deep white matter regions to which they project, as well as the splenium of the corpus callosum. These changes in fractional anisotropy occurred in subjects with significant differences in frontal, temporal, parietal, and deep white matter volumes. Fractional anisotropy values in the left anterior uncinate correlated with verbal IQ, full-scale IQ, and Peabody Picture Vocabulary Test-Revised scores for preterm male subjects. In addition, preterm male subjects were found to have the lowest values for fractional anisotropy in the right anterior uncinate fasciculus, and fractional anisotropy values in that region correlated with both verbal IQ and Peabody Picture Vocabulary Test-Revised scores for the preterm groups; these findings were supported by changes identified with voxel-based morphometric analyses.
CONCLUSIONS. Compared with term control subjects, prematurely born children with no neonatal ultrasound evidence of white matter injury manifest changes in neural connectivity at 12 years of age.
Preterm birth is associated with increased rates of disability, and many authors have attributed the neurodevelopmental handicaps that prematurely born children experience to white matter injury resulting from early birth.1–4 Over the years, the concept of injury to the developing white matter has changed. Early studies used cranial ultrasonography to detect lesions such as periventricular leukomalacia, which was reported to occur in 5% to 15% of very low birth weight, preterm infants.5 The advent of MRI led to the finding that white matter injury was far more common in prematurely born infants than previously suspected, and lesions were shown to occur in more than three fourths of very low birth weight, preterm infants.6–9 More recently, serial MRI studies of preterm neonates without evidence of gross neonatal brain injury demonstrated moderately decreased white matter volumes, compared with matched, term-equivalent, control subjects, which suggests that the injury of preterm birth must also be examined at the microstructural level.10
Prematurely born children exhibit both macrostructural11–14 and functional15–17 cerebral differences, compared with term control children. Moreover, these differences continue through many stages of development and are still evident at 12 years of age.18 In contrast, the long-term effects of preterm birth on neural connectivity remain largely unknown.
Diffusion tensor imaging (DTI) permits the investigation of microstructural alterations in corticogenesis19–21 and basic white matter structure.22–24 DTI data allow inferences to be made regarding the preferential directions of water diffusion, which are associated with the principal fiber directions. DTI can provide quantitative information by using measures such as fractional anisotropy (FA), which describes the degree to which water diffusion is restricted in one direction relative to all others and has been shown by several authors to increase with increasing gestational age among prematurely born infants.6,25 Furthermore, FA values have been reported to predict neurologic outcomes at 18 to 24 months of age for very low birth weight, preterm children with no evidence of brain injury in the newborn period.4
To examine the long-term impact of preterm birth on neural connectivity, we used DTI to compare prematurely born children with no ultrasonographic evidence of brain injury and matched, term, control children at 12 years of age. We hypothesized that FA values would differ significantly between the groups. Furthermore, because our previous studies of preterm children at 8 years of age suggested an influence of gender on white matter development,13 we explored the relationships of preterm birth, gender, and cognition in our subject groups.
This study was performed at the Yale University School of Medicine (New Haven, CT), Brown Medical School (Providence, RI), and Stanford University (Stanford, CA). The protocols were reviewed and approved by institutional review boards at each location. Children provided written assent and parents provided written consent for the study. All scans were obtained and analyzed at Yale University with the exception of voxel-based and volumetric morphometric analyses, which were performed at the Stanford Center for Interdisciplinary Brain Sciences Research.
The preterm cohort consisted of 29 children with no evidence of intraventricular hemorrhage, periventricular leukomalacia, and/or low-pressure ventriculomegaly. Subjects had normal neurologic findings and total ventricular cerebrospinal fluid volume (as measured with BrainImage 5 [Stanford University; Palo Alto, CA]) after segmentation within 2 SD of the mean ventricular volume of term control subjects at 12 years of age and no contraindications to MRI. Subjects with neonatal ultrasound evidence for intraventricular hemorrhage, white matter injury and/or ventriculomegaly were excluded from our study. All preterm subjects enrolled in the follow-up component of the Multicenter Randomized Indomethacin Intraventricular Hemorrhage Prevention Trial NCT0003391726,27 were recruited sequentially for the MRI study when they reached 12 years of age. These children were representative of the cohort of subjects with no evidence of neonatal brain injury from which they were selected, with respect to gender, handedness, full-scale IQ (FSIQ) scores, minority status, and maternal education. Twenty-two healthy term children, 12 years of age, were recruited from the local community and group-matched with the preterm group according to age, gender, and minority status. Control subjects were zip code-matched with preterm subjects for the purpose of neurocognitive testing.28 Control subjects were assessed with the same battery of cognitive, behavioral, and demographic measures as the preterm children at 12 years of age. Similarly, their MRI studies were evaluated for both clinical abnormalities and ventricular size. The assessments of neonatal health status and neurologic outcomes were described previously.29 Blinded assessment of intelligence was performed by using the Wechsler Intelligence Scale for Children-III.29 Children also were assessed with the Peabody Picture Vocabulary Test-Revised (PPVT-R)30 and the Developmental Test of Visual Motor Integration (VMI).31 The VMI is a culture-free, developmental sequence of geometric figures to be copied with pencil and paper. Visual perceptual and fine motor abilities are measured with the VMI. Visual impairment was not a factor in this study. Cerebral palsy was diagnosed if hypertonicity, hyperreflexia, and dystonia or spasticity were noted on neurologic examination.
MRI Protocols, Image Processing, and Measurements
MRI was performed with a GE-Signa 1.5-T scanner (General Electric, Milwaukee, WI). Sagittal brain images were acquired with a 3-dimensional, volumetric, radiofrequency spoiled gradient echo pulse sequence (repetition time: 24 milliseconds; echo time: 5 milliseconds; flip angle: 45°; number of excitations: 1; matrix size: 256 × 192 data points; field of view: 30 cm; slice thickness: 1.2 mm; number of contiguous slices: 124). DTI data were obtained by using a double-spin echo, echo planar imaging sequence with 6 directions, 2 b values (0 and 1000 seconds/mm2), and 3 averages (echo time: 81 milliseconds; repetition time: 5400 milliseconds; acquisition matrix: 128 × 128, interpolated to 256 × 256; field of view: 20 × 20 cm; number of slices: 40; slice thickness: 3 mm; skip: 0 mm).
Volumetric image processing was conducted by using BrainImage 5. Data processing included removal of non–brain tissues from the images, bias field correction, segmentation (gray matter, white matter, and cerebrospinal fluid), normalization of image position, and parcellation of the cerebral cortex into lobe and subcortical regions on the basis of a stereotaxic atlas template.32 This procedure, as described and validated in previous reports,33 results in reliable measurements for gray matter, white matter, and cerebrospinal fluid total cerebral, lobe, and deep cerebral volumes. Intrarater reliabilities for volumes described in this study were all ≥0.95, as determined with the intraclass correlation coefficient.
In the DTI processing, 6 separate acquisitions were averaged and the diffusion tensor was computed from these data. Mean diffusivity and FA were calculated from the tensor data and nonlinearly registered to a single-subject FA map selected from the control group of children. Both groups of subjects were registered to this single-subject template to form composite maps. The single subject who defined the common reference space to which the others were registered was a male control subject with good data (high signal/noise ratio and no motion artifacts). To take these data into a common reference space, a nonlinear registration between the reference FA map and the individual subject FA map within the Yale BioImage Suite software package34 was performed by using the intensity-only component of the method reported by Papademetris et al.35 We demonstrated previously that registration using the FA maps provides a better fit of the white matter than does registration using the raw DTI data.36
To assess white matter integrity, we confined our analysis to white matter by selecting FA maps only in the areas where all subjects included in the analysis had FA values of >0.15. A FA value of 0.15 was shown previously to provide a reliable threshold between gray matter and white matter.37
An average tensor across subjects was also computed after nonlinear registration of all subjects to a reference FA map, and the control group tensor was used to create a composite, tricolor, directionality map. This tricolor directionality map from the control group allowed fiber bundles to be delineated according to the direction of diffusion along the fibers, and it was used to define manually anatomic regions of interest (ROIs) on the basis of fiber bundle location. Each ROI contained only fibers oriented in a single direction (single color on the directional color map). Because all of the subjects were registered in the same composite space, these ROIs were applied directly to the single-subject and group FA maps to generate individual FA values for each ROI for each subject, for second-level statistical analyses.
In addition to the FA maps, the eigenvalue maps of the 3 eigenvectors of the diffusion tensor were obtained for each subject. The diffusion tensor eigenvalue maps were generated to assess how the eigenvalues changed in areas where there were significant changes in FA.
Because the preterm population exhibits verbal difficulties and the uncinate fasciculus comprises the main ventral pathway underlying semantic language systems in the developing brain,38–40 we targeted this region by using fiber tracking41 on the tensor data of the reference subject in the control group, to extract and to define the left uncinate and the right uncinate as separate ROIs. Fiber tracking was completed for each subject separately for the right uncinate and the left uncinate. A seed voxel was chosen in the tip of the temporal lobe, with a volume of 3 mm × 3 mm × 3 mm. The tracking parameters used were as follows: density of fiber tracking: 3; 4th order Runge-Kutta type: 4; step length: 0.25; primary eigenvector filters: FA between 0.2 and 1 and mean diffusivity between 0 and 10; minimal length: 10 mm; maximal length: 100 mm; maximal angle: 45°. Fibers were saved as binary images and were edited only to ensure that the uncinate fasciculus terminated in the inferior frontal gyrus. The edited ROIs were used to obtain a mean FA value for each individual fiber track. This fiber-tracking approach was registration-free and was performed entirely in the individual subject's brain space; therefore, each subject had a custom-tailored ROI. This secondary analysis was used to support the voxel-based results obtained in the direct FA comparison with anatomically defined ROIs.
To complement the DTI analyses, we used voxel-based morphometry (VBM)42 to examine regional volume differences between the groups. The optimized VBM process included (1) segmentation and extraction of the brain in native space, (2) normalization of the white matter to a standard space by using a custom template created from all subjects (preterm and term) in the study, (3) segmentation and extraction of the normalized brain, (4) modulation of the normalized white matter images to correct for tissue volume differences attributable to the normalization procedure, and (5) smoothing using a 12-mm full width at half-maximum kernel, to reduce the effects of noise.
Demographic and cognitive data were analyzed by using standard χ2 statistics for categorical data. Continuous-valued behavioral data were analyzed by using analysis of covariance including the terms group, gender, and group-gender interaction. Pearson correlations were used to evaluate associations in behavioral and MRI data. All P values reported were corrected for multiple comparisons.
For the DTI data, the ROI-based FA values were entered into an analysis of covariance model to examine main effects of group and gender and an interaction term. Similar analyses were performed examining the main and interaction effects of FA, cognitive measures, and gender.
Statistical analysis of the VBM data used SPM 2 software to obtain probability maps highlighting white matter differences. An analysis of variance model was used to test group-gender differences in regional white matter volumes. Absolute threshold masking (threshold: 0.15) was used to minimize gray matter/white matter boundary effects, and implicit masking was used to disregard voxels with 0 values. Analyses of variance were conducted with an explicit mask for the ROIs derived from DTI (group-gender interaction) analyses to constrain the VBM analysis to the inferior frontal gyrus (ie, anterior uncinate fasciculus). DTI ROIs were created as described above and normalized to the VBM white matter template by using BioImage Suite software. Statistics for the VBM analyses were normalized to standardized z scores, and significant clusters of activation were determined by using a height threshold of <0.001 (corrected). As used commonly, these stringent P value levels were used to account for the multiple comparisons performed during these analyses. All P values in this report are of the 2-sided type.
Neonatal characteristics of both groups are shown in Table 1. There were no marked differences in the numbers of male, nonwhite, or Hispanic children or years of maternal education.
Demographic and cognitive data are shown in Table 2 and demonstrate no significant differences in age at scan or number of right-handed subjects between the groups. Preterm subjects were significantly shorter than term control subjects (P = .032), but there was no significant effect of group on weight at 12 years of age (P = .14). Verbal IQ (VIQ) (P = .037), performance IQ (P < .001), and FSIQ (P = .002) scores were all significantly lower for the preterm subjects, compared with the term control children. Scores for the VMI were also significantly different between preterm and term subjects (P = .046), and a significant group-gender effect was noted for the VMI scores (P = .021). For this measure, scores for the preterm male subjects were significantly lower than those for preterm female subjects, whereas term male subjects had higher scores than did their female counterparts. None of the subjects had a VIQ score of <70.
Brain Volumes of the Study Subjects
Brain volume data for the study children are shown in Table 3 and demonstrate significant differences between the preterm and term study children for the temporal regions (left: P = .0392; right: P = .0157), deep gray matter regions (left: P = .0066; right: P < .001), frontal regions (left: P = .0042; right: P = .0135), temporal regions (left: P = .0274; right: P = .0062), parietal regions (left: P = .0409; right: P = .0226), and deep white matter regions (left: P = .0187; right: P = .0245). Furthermore, group-gender effects were found in both the left temporal white matter region (P = .037) and the left deep white matter region (P = .0403). For both regions, preterm female subjects were found to have values similar to those for term, control, female subjects but values for preterm male subjects were lower than those for term, control, male subjects.
Figure 1 shows a 3-orthogonal view of an example of the color maps used to define anatomically based ROIs for the primary DTI analysis. DTI data are shown in Table 4, and Talairach coordinates for each ROI are listed in Table 5. Significant group differences in FA values were seen for the following regions: left and right anterior uncinate fasciculi (left: P < .001; right: P = .003), left and right anteroinferior frontooccipital fasciculi (left: P = .040; right: P = .019), right posteroinferior frontooccipital fasciculus (P = .001), left superior frontooccipital fasciculus (P = .049), splenium of the corpus callosum (P = .021), and left and right external capsules (left: P = .024; right: P <.001). Significant differences were also noted in the subcortical white matter of the following regions: left and right precentral gyri (both: P < .001), right superior temporal gyrus (P = .002), right forceps major (P = .008), and left and right cinguli (both: P < .001). A trend for significance was noted in the left forceps major (P = .056).
In addition, a significant group-gender interaction was found in the FA values measured for the right anterior uncinate fasciculus ROI (P = .009), with preterm male subjects showing lower FA values than preterm female subjects, whereas term male subjects had higher values than preterm or term female subjects. The mean diffusivity measure demonstrated significant differences between the preterm and term study children only in the right posteroinferior frontooccipital fasciculus (P = .04).
Tractography ROIs Revealing Group-Gender Differences for Left Uncinate
Because both our previous studies and the current data demonstrated significant group-gender effects for left temporal white matter regions (with either volume or FA as the dependent variable), we tested the hypothesis that there would be a group-gender effect for FA values in the left uncinate fasciculus. A representative, single-subject, left uncinate ROI obtained through fiber tracking is shown in Fig 2. Using bilateral uncinate fasciculi ROIs, we observed significant group (P = .018) and group-gender (P = .043) effects for the left uncinate, with FA values being lower for preterm male subjects than for preterm female subjects and FA values for term male subjects being greater than those for term female subjects. FA values for the right uncinate also demonstrated a group effect (P = .043) and a trend for a group-gender effect (P = .094), suggesting that preterm male subjects also had the lowest FA values in this tract.
Analysis of Eigenvalue Maps for the Uncinate Fasciculi
Analysis of the eigenvalue maps for the contrast of preterm subjects minus control subjects revealed trends for significance for both the left and right anterior uncinate fasciculi. For both regions, we noted an increase in the lowest eigenvalues, compared with term control values (left: P = .095; right: P = .100).
VBM Data Demonstrating Group-Gender Effects for Uncinate Fasciculi
Preterm male subjects demonstrated significantly less white matter volume in the regions of the left and right anterior uncinate fasciculi (left: P = .003; right: P = .006), compared with term male subjects. Preterm female subjects showed no significant differences in these regions, compared with term female subjects. The reverse contrasts (preterm male/female minus term male/female) were not significant.
Exploratory Correlational Analyses
We tested the hypothesis that FA values in the left anterior uncinate fasciculus would correlate with language measures in the male and female preterm groups. These analyses demonstrated positive correlations between VIQ, FSIQ, and PPVT-R scores and left anterior uncinate FA values for preterm male subjects (VIQ: r = 0.513; P = .051; FSIQ: r = 0.535; P = .040; PPVT-R: r = 0.511; P = .052) and a trend for a negative correlation between VIQ and FA values for preterm female subjects (r = −0.539; P = .071). Figure 3 shows the correlation data for the left uncinate FA values, as a function of FSIQ, for the male and female preterm groups.
FA values in the right anterior uncinate fasciculus also showed positive correlations with VIQ (r = 0.635; P = .008) and PPVT-R scores (r = 0.61939; P = .011) for preterm male subjects. For female subjects, however, significant negative correlations were observed between both VIQ and PPVT-R scores and FA values in the right anterior uncinate fasciculus (VIQ: r = −0.74361; P = .004; PPVT-R: r = −0.75863; P = .003).
We report significant changes in FA, a measure of white matter organization, in the brains of prematurely born children, compared with term control subjects, at 12 years of age. These microstructural changes are particularly notable because our criteria for selection for the preterm group specifically excluded subjects with evidence of white matter changes on serial cranial ultrasound scans in the newborn period. Preterm subjects also were required to have normal ventricular size at 12 years of age. Regional volumetric analyses for the study subjects revealed significant white matter volume differences between the subject groups, which suggests both macrostructural and microstructural differences in white matter in the prematurely born subjects. Finally, our data suggested significant correlations between local FA values and VIQ, FSIQ, and PPVT-R scores, as well as group-gender effects.
Our data are notable for the localization of white matter abnormalities in several association tracts. Reductions in FA were found in the anteroinferior frontooccipital fasciculi bilaterally, as well as in the right posteroinferior frontooccipital fasciculus and the left superior frontooccipital fasciculus and inferior longitudinal fasciculus. In addition to these intrahemispheric association fibers, our results suggest abnormalities of fiber tract organization in the regions into which these fibers project, namely, the bilateral inferior frontal gyri and the precentral gyri.
In light of the previously reported temporal lobe volumetric findings and the language problems that prematurely born children experience,13,43,44 the abnormalities in the anteroinferior frontooccipital fasciculi and the uncinate fasciculi bilaterally are of particular interest. These tracks contribute to the temporal stem, the fibers responsible for connecting the polymodal association areas in each temporal lobe with the ipsilateral frontal lobe, thalamus, and contralateral temporal lobe.
The uncinate fasciculus connects the anterior temporal, orbitofrontal, and lateral-prefrontal cortices and is the main ventral pathway subserving semantic language systems in the developing brain.38–40 The fiber-tracking study reported here demonstrated significant group-gender effects for the left uncinate fasciculus, such that preterm male subjects were found to have the lowest FA values in this critical language pathway. Even larger effects were noted in the anterior uncinate ROIs, and an analysis of the eigenvalue maps for the anterior uncinate bilaterally showed that the main cause of lower FA values was a trend (P < .1) for increases in the lowest eigenvalues. These data suggest that apparent diffusion perpendicular to the principal direction of diffusion might have increased; a relative paucity of myelin was associated with increases in the lowest eigenvalues in preclinical studies.45 Furthermore, we noted decreased FA values for both the splenium of the corpus callosum and the forceps major, regions that connect the corpus callosum and parietooccipital white matter and are classically associated with reading difficulties in both injury and developmental studies.46,47 Considering the verbal difficulties that our subjects exhibited, these regions of significant changes in FA are of particular interest to clinicians and developmental neurobiologists.
DTI strategies were used previously in younger populations to investigate differences in white matter development in preterm subjects, compared with matched, term, control infants at term-equivalent age. Decreased FA was demonstrated in the white matter of preterm infants9 even when preterm infants with normal cranial ultrasound findings were compared with matched, term, control infants at term-equivalent age.6 Of note, Miller et al48 found an absence of the normal maturational increase in FA in the frontal regions of preterm infants with minimal white matter injury, which suggests regional vulnerability to injury in the developing preterm brain, and Vangberg et al49 reported decreased FA values in several white matter regions, including the corpus callosum, internal capsule, and superior fasciculus, in preterm subjects, compared with term control subjects, at 15 years of age. Finally, Arzoumian et al4 reported lower FA values in the posterior limb of the internal capsule in preterm infants who were subsequently found to have abnormal neurologic examination results at 20 months of age, and Nagy et al50 noted alterations in preterm children with attention-deficit/hyperactivity disorder at 11 years of age. Unlike other reports, our study evaluated only preterm subjects with normal ventricular size at 12 years of age. Furthermore, it is the first report, to our knowledge, that provides a combination of DTI, VBM, and volumetric data for adolescent subjects.
Significant developmental changes occur in white matter during childhood and adolescence, and several authors investigated both age-related changes and the relationship of white matter structure and cognition.51–53 Studying normally developing subjects 6 to 17 years of age, De Bellis et al54 found that male subjects had more-prominent, age-related, white matter volume increases, compared with female subjects. Blanton et al55 examined frontal lobe subregion volumes in 46 normal children and noted that the left inferior frontal gyrus was significantly larger in male subjects, compared with female subjects, at 6 to 17 years of age. Finally, left hemisphere FA values have been reported to correlate with both VIQ and FSIQ scores and academic measures, including reading, spelling, and rapid naming performance, in children 7 to 13 years of age.56,57 In the study presented here, correlations between FA values and cognitive measures were observed in our preterm group, although in opposite directions for preterm male subjects (positive in both left and right anterior uncinate) and female subjects (negative in right only). Different gender effects were reported previously across a number of other measures, including responses to indomethacin,58 and in terms of differences in language organization, as revealed with functional MRI.18 The results presented here are also consistent with an earlier publication that demonstrated gender-specific, white matter, microstructural changes with FA as the independent variable.59
Preterm infants were reported to have lower white matter volumes than matched control infants both in the newborn period and at school age,10,28 and previously we reported significant effects of preterm birth on white matter development in preterm male subjects at 8 years of age.13 In addition, using VBM strategies, Giménez et al60,61 detected widespread changes in white matter in preterm children at 15 years of age, and Caldú et al47 reported that decreases in the callosal area were correlated with performance IQ measures in preterm subjects at adolescence.
The question arises regarding the source of the selective vulnerability of specific cortical and white matter regions described here and in other reports. The relationship between this vulnerability and the stage of development of the brain at birth may play a role. As described in an article examining cortical gyrification,62 gyrification continues through the first postnatal year,63 and it has been shown to be a significant marker of neurodevelopment, with temporal areas being among the last to develop mature gyri.64–66 The temporal lobe also seems to represent an area of increased vulnerability in the preterm brain. In the temporal lobe, synaptogenesis and then gyrification begin during the third trimester of gestation,67,68 a time when many preterm children are born. Synapse density increases rapidly during this time and reaches a peak at 3 to 4 months after term in the temporal cortex.68 The alterations in corticogenesis reported by Kesler et al62 and the changes in white matter reported here are consistent with the presumptive time of injury. Most of our study participants were born near the onset of the third trimester, just as synaptogenesis was beginning in the temporal lobes. It seems that temporal lobe gyrification and white matter development in preterm children then proceeds in a suboptimal manner. The preterm cortex appears to be lissencephalic at birth64 and delayed in the development of gyrification, compared with control findings. Although gyrification in the frontal lobe also occurs later than in other cortical regions,65 cortical folding processes in the temporal lobe seem particularly affected by preterm birth.
The limitations of this study include the sample size and the paucity of information available concerning DTI changes in prematurely born children. Our subjects did not undergo MRI in the newborn period, and the changes in FA we reported might be influenced by neonatal white matter injury not readily detectable with cranial ultrasonography,1 by changes in axonal size, or by activity-dependent changes in myelination, rather than by changes in fiber organization. Although a single preliminary study of early intervention on FA in prematurely born subjects suggested the activity-dependent nature of FA in the developing brain,69 the impact of environmental factors such as maternal education and special school services on FA remains largely unknown. The links between cognitive function, gender, and white matter microstructural changes are poorly understood, and extensive research is needed to understand these relationships in the developing brain. The relationship between changes in white matter volumes and FA in typically developing adolescents is just beginning to be explored.70
The developing brain undergoes significant changes in functional organization with increasing age and skill.51,55 The neurobehavioral sequelae of preterm birth represent one of the major pediatric public health problems of our time,44,71 but the long-term microstructural effects of preterm birth on the developing brain remain largely unexplored. These studies extend our previous work and suggest that preterm birth results in significant, long-term, cerebral microstructural changes in children with no known evidence of intraventricular hemorrhage or cystic white matter injury in the newborn period.
This work was supported in part by National Institutes of Health grants NS27116, NS35476, M01-RR06022, M01-RR00125, K02-74677, DA017820, NS38467, and EB00473.
We thank Drs Deborah Hirtz and Walter Allan for scientific expertise; Marjorene Ainley for follow-up coordination; Jill Maller-Kesselman and Victoria Watson for neurodevelopmental testing; John Silbereis, Maolin Qiu, and Marcel Jackowski for technical help with the DTI acquisition and analysis; and Hedy Sarofin and Terry Hickey for technical assistance.
- Accepted July 20, 2007.
- Address correspondence to R. Todd Constable, PhD, Department of Diagnostic Imaging, Yale University School of Medicine, New Haven, CT 06520. E-mail:
The authors have indicated they have no financial relationships relevant to this article to disclose.
Drs Constable and Ment contributed equally to this work.
- ↵Dyet LE, Kennea NL, Counsell SJ, et al. Natural history of brain lesions in extremely preterm infants studied with serial magnetic resonance imaging from birth and neurodevelopmental assessment. Pediatrics.2006;118 (2):536– 548
- ↵Arzoumanian Y, Mirmiran M, Barnes PD, et al. Diffusion tensor brain imaging findings at term-equivalent age may predict neurologic abnormalities in low birth weight preterm infants. AJNR Am J Neuroradiol.2003;24 (8):1646– 1653
- ↵Counsell S, Allsop JM, Harrison MC, et al. Diffusion-weighted imaging of the brain in preterm infants with focal and diffuse white matter abnormality. Pediatrics.2003;112 (1):1– 7
- ↵Mewes AUJ, Hüppi PS, Als H, et al. Regional brain development in serial magnetic resonance imaging of low-risk preterm infants. Pediatrics.2006;118 (1):23– 33
- ↵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 (7):1616– 1623
- Nosarti C, Rushe TM, Woodruff PWR, Stewart AL, Rifkin L, Murray RM. Corpus callosum size and very preterm birth: relationship to neuropsychological outcome. Brain.2004;127 (9):2080– 2089
- ↵Peterson BS, Vohr BR, Kane M, et al. A functional MRI study of language processing and cognitive outcome in prematurely born children. Pediatrics.2002;110 (6):1153– 1162
- ↵Counsell S, Shen YJ, Boardman JP, et al. Axial and radial diffusivity in preterm infants who have diffuse white matter changes on magnetic resonance imaging at term-equivalent age. Pediatrics.2006;117 (2):376– 386
- ↵Dubois J, Hertz-Pannier L, Dehaene-Lambertz G, Cointepas Y, Le Bihan D. Assessment of the early organization and maturation of infants' cerebral white matter fiber bundles: a feasibility study using quantitative diffusion tensor imaging and tractography. NeuroImage.2006;30 (4):1121– 1132
- ↵Ment LR, Oh W, Ehrenkranz RA, Philip AGS. Low dose indomethacin and prevention of intraventicular hemorrhage: a multicenter randomized trial. Pediatrics.1994;93 (4):543– 450
- ↵Peterson BS, Vohr B, Cannistraci CJ, et al. Regional brain volume abnormalities and long-term cognitive outcome in preterm infants. JAMA.2000;284(15) :1939– 1947
- ↵Vohr BR, Allan WA, Westerveld M, et al. School age outcomes of very low birth weight infants in the Indomethacin Intraventricular Hemorrhage Prevention Trial. Pediatrics.2003;111 (4). Available at: www.pediatrics.org/cgi/content/full/111/4/e340
- ↵Dunn LM, Dunn LM. Peabody Picture Vocabulary Test-Revised. Circle Pines, NM: American Guidance Service; 1981
- ↵Beery KE. The VMI Developmental Test of Visual Motor Integration. Cleveland, OH: Modern Curriculum Press; 1989
- ↵Talairach J, Tournous P. Co-planar Stereotaxic Atlas of the Human Brain. New York, NY: Thieme Medical Publishers; 1988
- ↵Papademetris X, Jackowski M, Rajeevan N, Constable RT, Staib LH. BioImage Suite: An Integrated Medical Image Analysis Suite. New Haven, CT: Yale School of Medicine; 2007. Available at: www.bioimagesuite.org
- ↵Papademetris X, Jackowski AP, Schultz RT, Staib LH, Duncan JS. Integrated intensity and point-feature nonrigid registration. MICCAI.2004;3216 :763–770
- ↵Klang K, Jackowski M, Fulbright R, Constable RT. Anatomic versus fractional anisotropy images for registration of group DTI morphometry data. In: Proceedings of the International Society for Magnetic Resonance in Medicine. Berkeley, CA: International Society for Magnetic Resonance in Medicine; 2006:1064
- Duffau H, Gatignol P, Madonnet E, et al. New insights into the anatomo-functional connection of the semantic system: a study using cortico-subcortical electrostimulation. Brain.2005;128 (4):797– 810
- ↵Saigal S, den Ouden L, Wolke D, et al. School-age outcomes in children who were extremely low birth weight from four international population-based cohorts. Pediatrics.2003;112 (4):943– 950
- ↵Caldú X, Narberhaus A, Junqué C, et al. Corpus callosum size and neuropsychologic impairment in adolescents who were born preterm. J Child Neurol.2006;21 (5):406– 410
- Reiss AL, Abrams MT, Singer HS, Ross JL, Denckla MB. Brain development, gender and IQ in children: a volumetric imaging study. Brain.1996;119 (5):1763– 1774
- ↵Paus T, Zijdenbos A, Worsley K, et al. Structural maturation of neural pathways in children and adolescents: in vivo study. Science.1999;283(5409) :1908– 1911
- ↵De Bellis MD, Keshavan MS, Beers SR, Hall JWI, Frustaci K, Masalehdan A. Sex differences in brain maturation during childhood and adolescence. Cereb Cortex.2001;11 (6):552– 557
- ↵Ment LR, Peterson BS, Meltzer JA, et al. A functional MRI study of the long-term influences of early indomethacin exposure on language processing in the brains of prematurely born children. Pediatrics.2006;118 (3):961– 970
- ↵Giménez M, Junqué C, Vendrell P, et al. Abnormal orbitofrontal development due to prematurity. Neurology.2006;67(10) :1818– 1822
- ↵Battin MR, Maalouf EF, Counsell SJ, et al. Magnetic resonance imaging of the brain in very preterm infants: visualization of the germinal matrix, early myelination, and cortical folding. Pediatrics.1998;101 (6):957– 962
- ↵Armstrong E, Schleicher A, Omran H, Curtis M, Zilles K. The ontogeny of human gyrification. Cereb Cortex.1995;5 (1):56– 63
- ↵Als H, Duffy FH, McAnulty GB, et al. Early experience alters brain function and structure. Pediatrics.2004;113 (4):846– 857
- Copyright © 2008 by the American Academy of Pediatrics