Microstructural and functional connectivity in the developing preterm brain

Abstract

Prematurely born children are at increased risk for cognitive deficits, but the neurobiological basis of these findings remains poorly understood. Since variations in neural circuitry may influence performance on cognitive tasks, recent investigations have explored the impact of preterm birth on connectivity in the developing brain. Diffusion tensor imaging studies demonstrate widespread alterations in fractional anisotropy, a measure of axonal integrity and microstructural connectivity, throughout the developing preterm brain. Functional connectivity studies report that preterm neonates, children and adolescents exhibit alterations in both resting state and task-based connectivity when compared to term control subjects. Taken together, these data suggest that neurodevelopmental impairment following preterm birth may represent a disease of neural connectivity.

Introduction

Preterm subjects are at significant risk for cognitive impairment, and studies from the United States, England and Sweden have all revealed inferior educational outcomes in preterm adolescents and young adults.^{1, 2} A major focus of perinatal intensive care is the prevention of adverse outcomes in the prematurely born, yet the fundamental alterations in brain development responsible for these outcomes remain poorly understood. Magnetic resonance imaging (MRI) has facilitated the investigation of both structural and functional aspects of the developing brain and offers insight into the impact of preterm birth on brain development.

MRI has been used to characterize the macrostructural development of the brain by measuring gray and white matter volumes. Preterm subjects have globally and regionally smaller volumes of cortical gray and deep gray matter, smaller volumes of myelinated white matter and smaller corpus callosal areas when compared to term control subjects.^3–6 Nonetheless, these alterations do not completely explain the cognitive and behavioral difficulties of the prematurely born.

More recently, diffusion tensor imaging (DTI) has been used to characterize white matter microstructure in the developing brain. These investigations have demonstrated widespread decreases in fractional anisotropy (FA), a measure of axonal coherence, myelination and microstructural connectivity, in preterm children compared to term controls in studies ranging from infancy through young adulthood.^{7, 8} Correlations between microstructural connectivity data and cognitive and behavioral outcome measures in the prematurely born are of particular interest since recent data suggest that cognitive interventions may improve microstructural connectivity.^{9, 10}

In parallel with structural changes, functional MRI (fMRI) permits the investigation of the neural basis of language, memory and executive function in the developing brain. fMRI studies of language processing in preterm young adults who have thinning of the corpus callosum demonstrate increased blood oxygen level dependent (BOLD) signal in the right hemisphere of the preterm group compared to term controls, ^{11, 12} suggesting the engagement of alternative pathways for language. Similarly, fMRI data from several investigators suggest the engagement of auxillary circuits for memory, attention and executive function in the preterm brain,^13–16. Functional connectivity (fcMRI) analysis is a newer methodology that maps neural networks by identifying spatially distant brain regions that generate synchronized fluctuations in BOLD signal. Recent fcMRI reports suggest the engagement of alternative neural networks in the prematurely born.^{17, 18}

Emerging DTI data suggest that alterations of neural connectivity may underlie developmental disorders such as autism, Tourette Syndrome and schizophrenia. ^19–24 Preclinical studies have implicated both GABA-interneurons and oligodendroglia in the pathogenesis of these diseases^25–29. Since preterm birth influences both of these cellular targets in the developing brain,^30–32 we review published data and our own studies to suggest that preterm birth is also a disorder of connectivity.

Preterm birth results in significant disability

As described in the recent Institute of Medicine report, preterm birth is among the leading public health problems in the United States and Europe today.³³ There are over 4 million live births in the US each year, and more than 1.5% of these very low birth weight (VLBW) infants weigh less than 1500 grams. Survival for these medically fragile neonates now approaches 70 – 80%, but the annual perinatal care costs exceed $18 billion dollars per year in the United States alone.^33–37

Studies from the United States, Europe and Australia have all revealed poor educational outcomes in preterm subjects. At school entry, minor developmental impairment is diagnosed in 30 – 40% and major disabilities are found in almost 20% of preterm children.^{2, 34, 38–40} The intellectual deficits of preterm subjects may persist through adolescence and young adulthood.⁴¹ In the Victorian Infant Collaborative Study Group, at age 14 years 14% of extremely low birth weight preterm children (500 – 999 g) were classified as severely disabled and 15% as moderately so, compared to 2% of controls.⁴² Similarly, when 1097 Dutch preterm children of < 33 weeks GA and/or 1500 g BW were evaluated at the same age, more than half had disorders of learning, attention and social-emotional skills.⁴³ Finally, when compared to healthy controls, fewer preterm children pursue education after high school, fewer have full time employment and fewer live outside of the parental home.^{1, 44}

Injury is common in preterm birth: oligodendroglia and interneurons may play a role

During the late second and third trimesters of gestation, brain development is characterized by neurogenesis, gliogenesis, axonal ingrowth and the elaboration of synaptic connections. Injury is common in the preterm brain during this critical period.⁴⁵ Although intraventricular hemorrhage and focal periventricular leukomalacia (PVL) have historically been viewed as the major forms of brain injury accounting for handicap after preterm birth, more recent studies suggest that diffuse white matter injury is the major cause of neurodevelopmental sequelae.⁴⁶ Based on conventional MRI studies, up to 50% of VLBW neonates have evidence of white matter injury and accompanying neuronal/axonal disease.This condition has been variously called “white matter injury” or “diffuse periventricular leukomalacia” in the neonatal and neurologic literature.

First described by Banker,⁴⁷ PVL is characterized by both focal macroscopic and microscopic regions of necrosis, loss of premyelinating oligodendrocytes and marked astrogliosis and microgliosis of the white matter.⁴⁵ For many neonates, cerebral ventriculomegaly, detected by cerebral ultrasonography, has been the only biomarker used to detect this cascade of events.Ultrasound may be insufficient to detect the more commonly occurring diffuse white matter injury.⁴⁸ Histopathologic and MRI studies suggest that neuronal and axonal damage also occur in diffuse white matter injury, but these changes are less well understood.^49–52 MRI studies of preterm subjects demonstrate decreased cortical gray matter volumes as well as alterations in the thalamus, basal ganglia and cerebellum.^{5, 53–56}

Preclinical studies confirm oligodendroglial and neuro-axonal injury in periventricular white matter, which encompasses focal PVL and diffuse white matter injury.^{46, 57–59} Both hypoxic-ischemic insults and chronic hypoxia result in acute degeneration of late oligodendroglia precursors. The loss of these precursors is followed by a regenerative response of the surviving pre-oligodendroglia, but the surviving cells fail to produce myelin.³¹ Several authors have postulated that the vulnerable regions of the developing brain may harbor susceptible populations of late oligodendroglia precursors.^{58, 60}

Preclinical studies also suggest that PVL is associated with microstructural changes in connectivity. The late second and third trimesters are the time of both axonal in-growth and the elaboration of synaptic connections. In the chronic sublethal hypoxia rodent model of PVL, which mimics injury during that period, axonal guidance and synaptogenesis were disrupted resulting in aberrant axonal connectivity.⁶¹ Oligodendroglia produce axonal guidance molecules in addition to myelin, and oligodendrocyte production of the axon outgrowth inhibitor Nogo-A was decreased in the chronic sublethal hypoxia model. These results suggest at least one possible mechanism for disrupted connectivity.

Recent data suggest that neuronal vulnerability may also play a role in the alterations in connectivity found in the prematurely born. Preclinical studies demonstrate that gamma aminobutyric acid (GABA) interneurons are important for the assembly and maintenance of cerebral circuits,^{62, 63} and that GABAergic interneurons may serve as hub cells for circuit formation during brain development.⁶⁴ Multiple lines of evidence, including genetic, postmortem and imaging studies, suggest that the GABA system may be affected in disorders of connectivity.^{27, 65–67} Furthermore, although multiple cortical neuronal populations may be vulnerable to the injury of preterm birth, GABA is the predominant neurotransmitter in the developing telencephalon during the third trimester.⁶⁸ During this critical time period, GABAergic neurons migrate long distances from the ganglionic eminences in the ventral telencephalon through the ventricular zone, subventricular zone, intermediate zone and subplate to arrive at the cortex where they begin synaptogenesis.^{69, 70} This route of migration through the periventricular region puts the GABA interneurons at risk for injury.

Oxygen deprivation is a major cause of neurodevelopmental handicap in preterm subjects⁷¹ and has been shown to impact GABAergic neurons. Preclinical models of hypoxic injury are characterized by neuropathologic and behavioral changes. The chronic sublethal rodent model and that of repeated asphyxial injury to late fetal sheep both mimic the neuropathologic injury and behavioral outcomes of preterm birth, and these models result in loss of both oligodendroglia and GABAergic neurons.^{30, 72} Postmortem studies of preterm infants and children with white matter injury also demonstrate significant decreases in GABAergic interneurons and, perhaps in reciprocal response, their receptors in the subplate and cortex. ^{73, 74}

DTI studies document changes in inter-hemispheric connectivity

DTI is more sensitive than conventional MRI and assesses the integrity of white matter tracts at a microstructural level. In DTI analyses, fractional anisotropy (FA) values indicate the degree to which water diffusion is restricted along one axis relative to all others. Alterations in FA may result from changes in fiber organization, axonal size, or myelination,⁷⁵ and higher FA values serve as a marker for the coherence of white matter tracts.

FA values in white matter tracts increase over the course of fetal development and after birth. At term equivalent age FA values in multiple white matter tracts including the splenium and body of the corpus callosum and the left inferior longitudinal fasciculus demonstrate a linear correlation with gestational age at birth, indicating decreased white matter tract integrity in the most premature infants.^{76, 77}

These changes in FA in prematurely born subjects persist into adolescence and young adulthood. Lower FA values in the external capsules, longitudinal fasciculi and fornix and body of the corpus callosum have been reported in adolescents born preterm.⁷⁸ In a study of 12 year old subjects, multiple areas of decreased FA were found in preterm children compared to term children, including bilateral anterior portions of the uncinate fasciculi, the splenium of the corpus callosum, and the right inferior frontooccipital fasciculus.⁷⁹

FA values have been shown to correlate with performance on multiple measures of neurodevelopmental function.⁸⁰ Preterm children studied with DTI and the Griffiths Mental Development Scale at 2 years corrected age showed linear correlations between developmental quotient and FA values in the corpus callosum and right cingulum, correlations between performance sub-scores and FA in the corpus callosum and right cingulum, and correlations between eye-hand coordination scores and FA in the Icingulum, fornix, anterior commissure, corpus callosum, and right uncinate fasciculus.⁸¹ Similarly, in school age preterm children, whole-brain FA was an independent variable affecting full scale IQ after adjusting for birth weight, gestational age, and gender.⁸² In a group of 11 year old preterm children, reading performance scores positively correlated with FA values in the genu and body of the corpus callosum.⁸³ In 12 year old children born preterm, correlations were found between FA in the left anterior uncinate fasciculus and WISC verbal IQ, WISC full scale IQ and PPVT scores.⁷⁹ Finally, in prematurely born 15 year olds, Skranes reported correlations between low IQ and low FA in the external capsule and inferior and middle superior fasciculus, as well as between visual motor and visual perceptual deficits and low FA in the external capsule, posterior internal capsule, and inferior fasciculus.⁸⁴

Since functional imaging studies suggest the engagement of alternative pathways for language in the prematurely born, our group assessed language pathways in preterm subjects at adolescence. Forty-four PT subjects (600 – 1250 grams birth weight) without neonatal brain injury and 41 term controls were evaluated at age 16 years with DTI, the Peabody Picture Vocabulary Test - Revised (PPVT) and the Comprehensive Test of Phonological Processing (CTOPP). Preterm subjects scored comparably to term subjects on language measures but had lower FA values in multiple regions subserving language throughout the brain. Of note in the preterm group, FA values in both uncinate fasciculi correlated with PPVT scores, a semantic language task, while FA values in both arcuate fasciculi correlated with CTOPP Rapid Naming scores, a phonologic task.⁸⁵

These data support for the first time that the recently proposed concept of dual pathways underlying language function are present in reterm adolescents. These include a left-sided dorsal pathway associated with phonological and articulatory processing (arcuate fasciculus), and a bilateral ventral pathway for semantic, receptive language processing (uncinate fasciculus). The striking – and unexpected – bilateral dorsal correlations for the preterm group suggest that prematurely born subjects rely more heavily on the right hemisphere than typically developing adolescents for performance of phonological language tasks. Since recent findings in typically developing adults suggest that the language system becomes more confined during maturation,¹⁰ these findings may represent either a delay in maturation or the engagement of alternative neural pathways for language in the developing preterm brain.

Finally, to investigate long-term white matter microstructural differences in preterm children relative to term control subjects at 16 years of age and the impact of injury on the developing preterm brain, we evaluated 44 preterm subjects (birth weight 600 – 1250 g) without neonatal brain injury and 16 preterm subjects with Grades 1 and 2 intraventricular hemorrhage (IVH) with DTI and the testing measures described above. When we compared the 44 preterm subjects without neonatal brain injury and the 16 preterm subjects with IVH, subjects with IVH had significantly lower CTOPP Rapid Naming scores (p = 0.015) compared to subjects with no IVH, as shown in Table 1. We also observed significantly lower FA values in the preterm subjects with IVH compared to preterm subjects with no IVH in both the body (p = 0.018) and the splenium (p < 0.001) of the corpus callosum, as shown in Table 2 and Figures 1A and 2A. Although FA values in the corpus body significantly correlated with the CTOPP Rapid Naming score for all subjects as a group (R = 0.444, p = 0.0004), this correlation was only significant for the preterm subjects without IVH (R = 0.402, p = 0.007), as shown in Figure 1B. In contrast, while FA values in the splenium of the corpus callosum also correlated with the CTOPP Rapid Naming Score (R = 0.342 p = 0.008), this relationship was only significant for the preterm adolescents with a history of IVH (R = 0.593, p = 0.016), as shown in Figure 2B. Similar to the findings in preterm subjects compared to full-term controls, these data suggest that adolescents with a history of IVH are engaging alternative interhemispheric pathways for language or have a delay in maturation of language pathways compared to preterm subjects without IVH.

Table 1.

Demographic and Cognitive Data for Subjects in the DTI Study

	No IVH (n = 44)	IVH (n = 16)	P value
Age at scan (yrs)	16.3 ± 0.3	16.2 ± 0.2	0.12
Male gender	26 (59%)	9 (56%)	1.00
Right-handed	38 (86%)	11 (69%)	0.14
Minority status	15 (34%)	10 (63%)	0.075
Special services	8 (18%)	5 (31%)	0.30
Years of maternal education
< 12 years	5 (11%)	2 (13%)
12 years	16 (36%)	3 (20%)	0.86
13–15 years	11 (25%)	7 (47%)
≥ 16 years	12 (27%)	3 (20%)
Wechsler Intelligence Scale for Children-III
Verbal Comprehension Index	97.7 ± 15.3	90.3 ± 23.5	0.50
Verbal IQ	96.5 ± 15.3	88.8 ± 22.3	0.51
Performance IQ	93.1 ± 15.2	86.8 ± 18.0	0.40
Full Scale IQ	94.3 ± 14.2	86.5 ± 19.9	0.43
PPVT	101.8 ± 20.5	94.5 ± 26.8	0.46
CTOPP
Rapid Naming Comp.	101.7 ± 18.7	87.8 ± 17.4	0.015
Phonologic Awareness	82.4 ± 13.6	75.1 ± 16.5	0.28
Composite

Table 2.

Fractional Anisotropy Data

	FA value, mean ± SD		P value
Region	No IVH	IVH
Longitudinal fasciculi
L PIFOF	0.404 ± 0.030	0.400 ± 0.031	0.506
R PIFOF	0.413 ± 0.023	0.410 ± 0.021	0.300
L ILF	0.307 ± 0.024	0.299 ± 0.030	0.379
R ILF	0.304 ± 0.024	0.306 ± 0.021	0.888
L uncinate	0.290 ± 0.022	0.298 ± 0.028	0.168
R uncinate	0.277 ± 0.020	0.289 ± 0.023	0.054
Corpus callosum
Body	0.861 ± 0.092	0.810 ± 0.073	0.017
Genu	0.960 ± 0.092	0.935 ± 0.093	0.460
Splenium	0.668 ± 0.055	0.569 ± 0.056	< 0.0001
Internal capsule
L PLIC	0.502 ± 0.026	0.506 ± 0.023	0.583
R PLIC	0.496 ± 0.021	0.508 ± 0.023	0.158
External capsule
Left	0.326 ± 0.024	0.332 ± 0.029	0.375
Right	0.311 ± 0.019	0.311 ± 0.021	0.888
Other ROIs
L forceps major	0.341 ± 0.045	0.320 ± 0.040	0.050
R forceps major	0.356 ± 0.038	0.348 ± 0.041	0.293
L cingulum	0.310 ± 0.045	0.287 ± 0.046	0.075
R cingulum	0.303 ± 0.043	0.277 ± 0.037	0.057

Abbreviations: ILF, inferior longitudinal fasciculus; PIFOF, posteroinferior frontooccipital fasciculus; PLIC, posterior limb of the internal capsule; L, left; R, right.

Figure 1.

Significantly lower FA values were observed in the preterm subjects with IVH compared to PT subjects with no IVH in the body of the corpus callosum, shown in Panel A (p = 0.017). FA values in the corpus body significantly correlated with the CTOPP rapid naming score for the preterm subjects without IVH (R = 0.402, p = 0.007), as shown in Panel B.

Figure 2.

Significantly lower FA values were observed in the preterm subjects with IVH compared to preterm subjects with no IVH in the splenium of the corpus callosum, shown in Panel A (p < 0.0001). FA values in the splenium significantly correlated with the CTOPP rapid naming score for the preterm subjects with IVH (R=0.593, p=0.016), as shown in Panel B.

Taken together, data gathered from preterm subjects ranging from infancy through young adulthood suggest that preterm birth and early brain injury alter intra-and interhemispheric microstructural connectivity. These data correlate with both cognitive and specific language measures in the prematurely born, but the impact of maturation, early intervention and maternal education on these findings remains largely unknown.^{9, 10}

Functional images demonstrate alterations in neural connectivity

Functional MRI (fMRI) permits the investigation of the neural basis of language, memory and executive function in the developing brain. Depending on the subject population and the age at which they were studied, fMRI studies of language processing in preterm children compared to term controls demonstrate that blood oxygen dependent signals classically expected in the left temporal region are detected in the right hemisphere of the preterm group.^{11, 12} Further, even when gender, gestational age and early environmental interventions are taken into account, fMRI data reveal the engagement of auxillary circuits for memory, attention and executive function in the preterm brain,^13–16

In contrast, functional connectivity (fcMRI) studies assess “temporal correlations between spatially remote neurophysiological events.”^{86, 87} Spatially distant brain regions characterized by synchronized fluctuations in BOLD signal are mapped to visualize functionally connected neural networks. Recent reports suggest that developmental disorders such as autism,⁸⁸ Tourette Syndrome⁸⁹ and schizophrenia⁹⁰ may be secondary to alterations in neural connectivity. Studies using fcMRI suggest that premature birth may also alter connectivity.

Neural connectivity is currently assessed by either task-based strategies or seed-based resting state fcMRI examinations.⁹¹ Spontaneous low frequency BOLD signal fluctuations in functionally related gray matter regions show strong correlations at rest and are believed to relate to spontaneous neural activity.^{91, 92} Over time, the developing brain increases the strength of connections that exist in spatially remote regions in an anterior-posterior direction, weaving distal brain areas into highly cohesive and connected circuits, while reducing the strength of connections between anatomically proximal regions and contra-lateral homologues.⁸⁷ Resting state neural networks have been demonstrated in infancy^{93, 94}, as shown in Figure 3, and continue to mature through adolescence.^{86, 87, 95}

Figure 3.

Seed-based neural connectivity from left Wernicke’s region in 5 preterm neonates studied at term equivalent age. The left Wernicke’s seed is shown in yellow, and the regions to which it is significantly connected are shown in red (p < 0.05).

Data from task-based fcMRI studies demonstrate the engagement of alternative neural circuits for language in preterm subjects compared to term controls at both school age and adolescence.^{17, 18} Studies at school age demonstrated increased interhemispheric connectivity to both the right frontal and parietal language regions for preterm subjects compared to term controls during a semantic task. Of note, white matter volumes correlated with connectivity in the preterm subjects, showing that those with greater white matter volumes had greater connectivity. An important example of this is in the left hemisphere language system, where connectivity from left Wernicke’s area to the left angular gyrus was significantly correlated with left hemisphere deep white matter volumes for the preterm group (p = 0.015).⁹⁶

As shown in Figure 4, at adolescence, the preterm subjects continued to demonstrate increased interhemispheric connectivity to right-sided language regions compared to term controls, although less so than at age 8 years. In addition, those subjects with the greatest cognitive need were most likely to engage alternative, or perhaps “additional,” pathways for semantic language. Interhemispheric connectivity was inversely related to the level of maternal education in the preterm group, also suggesting auxiliary engagement or delayed pruning for those with the greatest need (Figure 5). It is of note, therefore, that a recent report of seed-based resting state connectivity data from preterm and term control neonates at term-equivalent age suggests that alterations in neural circuitry are present in the preterm group largely before environmental and educational strategies can intervene.⁹⁷

Figure 4.

Functional connectivity analyses demonstrated increased connectivity during a semantic language task from left Wernicke’s region (BA 22, red circle) to both right and left BA 40 regions (blue circles) in preterm subjects compared to term control subjects at age 16 years (p = 0.02).

Figure 5.

Connectivity from left Wernicke’s region (BA 22) to right BA 40 is inversely correlated to maternal education in preterm subjects at age 16 years (p = 0.02).

Since variations in neural circuitry may influence performance on cognitive tasks, we employed an fcMRI analysis of resting state intrinsic connectivity contrast (ICC) across the brain to identify behavioral correlations that differed between 20 preterm subjects (600 – 1250 g birth weight) and 23 matched controls at adolescence. Although there were no significant differences in IQ and PPVT-R scores between the preterm and term controls, preterm subjects performed significantly less well on the Delis Kaplan System Total Achievement Score (p < 0.01), a measurement of executive function in adolescents as shown in Table 3.

Table 3.

Demographic and Cognitive Data for the fcMRI Study Subjects

	Preterm	Term	P value
Number	20	23
Age at scan	17.8 ± 1.29	16.6 ± 1.21	< 0.001
Males	11 (55%)	8 (35%)	0.19
Right-handed	16 (80%)	22 (96%)	0.11
Minority status	6 (30%)	6 (26%)	0.62
WISC III Intelligence scores
Full Scale IQ	97.2 ± 10.0	105.0 ± 16.4	0.09
Verbal IQ	100.0 ± 11.05	103.8 ± 15.6	0.22
Performance IQ	95.1 ± 14.0	105.0 ± 17.2	0.06
Verbal Comprehension IQ	101.1 ± 11.55	104.8 ± 14.7	0.24
Language measures
Peabody Picture Vocabulary Test	101.5 ± 15.52	105.3 ± 19.6	0.45
CTOPP Rapid Naming Composite	93.4 ± 18.4	99.6 ± 11.2	0.31
Delis Kaplan Executive Function Scales
Total achievement scaled score	8.9 ± 1.8	10.7 ± 1.9	0.009
Maternal education < high school	3 (15%)	1 (4%)	0.24

Executive function is critical for academic success, and ICC correlational analyses identified multiple regions in which the correlation between intrinsic connectivity and Total Achievement Score significantly differed in preterm subjects compared to controls, as shown in Figure 6 (p < 0.05). These data suggest that preterm adolescents exhibit alterations in resting state connectivity in executive control networks when compared to term control subjects and support the hypothesis that there are long-lasting influences of preterm birth on connectivity in the developing brain.

Figure 6.

Resting state intrinsic connectivity correlations with an executive function task, the Delis Kaplan System Total Achievement Score, demonstrate significant differences between preterm subjects and term controls at age 16 years. For term controls (Panel A), increased intrinsic connectivity for this task (shown in yellow) is significantly correlated with better performance in the right temporal, right hippocampal and basal ganglia (p < 0.05). In contrast, increased intrinsic connectivity for this task is significantly correlated with better performance in the left temporal region for the preterm group and worse performance (shown in blue) in the frontal regions for the preterm subjects at young adulthood (p < 0.05).

Future directions

Advances in neuroimaging have provided exciting and important information about the impact of preterm birth on the developing brain. Diffusion tensor imaging studies demonstrate widespread alterations in fractional anisotropy throughout the developing brain, while functional connectivity studies suggest that preterm neonates, children and adolescents exhibit alterations in both resting state and task-based connectivity when compared to term control subjects.

These data support the hypothesis that preterm birth represents a disease of connectivity. Future preclinical investigations will explore the molecular markers of not only glial and neuronal survival in response to the stresses of preterm birth, but also those signaling mechanisms guiding axonal in-growth and synaptogenesis during the late second and third trimesters of gestation. Simultaneously, clinical studies will provide critical information about those early biomarkers and environmental events that impact connectivity in the developing brain.

ABBREVIATIONS

BA

Brodmann area

BOLD

Blood oxygen level dependent

CTOPP

Comprehensive Test of Phonological Processing

DTI

Diffusion tensor imaging

FA

Fractional anisotropy

fcMRI

Functional connectivity magnetic resonance imaging

fMRI

Functional magnetic resonance imaging

FSIQ

Full Scale Intelligence Quotient

GABA

Gamma aminobutyric acid

GLM

General linear model analysis

ICC

Intrinsic connectivity contrast

IVH

Intraventricular hemorrhage

PPVT -R

Peabody Picture Vocabulary Test-Revised

PT

Preterm

ROI

Region of interest

WISC-III

Wechsler Intelligence Scale for Children-III