Elsevier

NeuroImage

Volume 191, 1 May 2019, Pages 36-48
NeuroImage

The emergence of long-range language network structural covariance and language abilities

https://doi.org/10.1016/j.neuroimage.2019.02.014Get rights and content

Abstract

Language skills increase as the brain matures. Language processing, especially the comprehension of syntactically complex sentences, is supported by a brain network involving functional interactions between left inferior frontal and left temporal regions in the adult brain, with reduced functional interactions in children. Here, we examined the gray matter covariance of the cortical thickness network relevant for syntactic processing in relation to language abilities in preschool children (i.e., 5-year-olds) and analyzed the developmental changes of the cortical thickness covariance cross-sectionally by comparing preschool children, school age children, and adults. Further, to demonstrate the agreement of cortical thickness covariance and white matter connectivity, tractography analyses were performed. Covariance of language-relevant seeds in preschoolers was strongest in contralateral homologous regions. A more adult-like, significant cortical thickness covariance between left frontal and left temporal regions, however, was observed in preschoolers with advanced syntactic language abilities. By comparing the three age groups, we could show that the cortical thickness covariance pattern from the language-associated seeds develops progressively from restricted in preschoolers to widely-distributed brain regions in adults. In addition, our results suggest that the cortical thickness covariance difference of the left inferior frontal gyrus to superior temporal gyrus/sulcus between preschoolers and adults is accompanied by distinctions in the white matter tracts linking these two areas, with more mature white matter in the arcuate fasciculus in adults compared to preschoolers. These findings provide anatomical evidence for developmental changes in language both from the perspective of gray matter structure co-variation within the language network and white matter maturity as the anatomical substrate for the structural covariance.

Introduction

The brain's structure undergoes remarkable changes during human development and scales up with individual differences in behavior and cognitive functions, such as intelligence and memory (Giedd et al., 1999; Kanai and Rees, 2011; Shaw et al., 2008; Sowell, 2004; Sowell et al., 2004). The interest in structural connectivity of the cerebral cortex has recently been emphasized in human brain research, stressing that cognitive functions are supported by connections between distributed, rather than isolated cortical regions (Fedorenko and Thompson-Schill, 2014). One approach for aspects of cortical networks is structural covariance, describing the phenomenon that gray matter properties of one brain region may co-vary with those of other broadly distributed cortical regions (Alexander-Bloch et al., 2013a). It has been assumed that structural covariance is related to axonal connectivity and that it underlies functional connectivity between regions as a result of mutually trophic influences (Ferrer et al., 1995), shared experience-related plasticity (Draganski et al., 2004; Mechelli, 2005; Montembeault et al., 2012), or a combination of these factors (Seeley et al., 2009). Like for the brain's gray and white matter, structural covariance has also been observed to change with age, as well as to interact with behavioral and cognitive functions and its disorders (Alexander-Bloch et al., 2013a; Bernhardt et al., 2014b; Montembeault et al., 2016; Qi et al., 2016; Seeley et al., 2009). For instance, for children up to the age of 2 years, structural covariance has been observed to be restricted by showing mostly correlations with the surrounding regions of the seed for both higher-order (e.g., default-mode, dorsal attention, speech, semantic, and executive control networks) and lower-order networks (e.g., primary visual and sensorimotor networks) (Geng et al., 2016). Between the ages 5 and 18 years, diverging developmental trajectories for lower- and higher-order networks were shown. The lower-order networks demonstrate well-established structural covariance already in early childhood, peaking in early adolescence, before pruning induces further reorganization towards adulthood. In contrast, structural covariance of higher-order networks is not yet fully developed in early childhood, which is mainly characterized by co-variation between contralateral homologous regions. The structural covariance of higher-order networks then develops progressively to more widely distributed regions across the teenage years (Zielinski et al., 2010). A similar developmental pattern has been demonstrated for intrinsic functional connectivity with increased short-range and sparse long-range functional connectivity in children, but a coherent long-range architecture in young adults across a range of cognitive networks (i.e., control and default mode networks) (Fair et al., 2009, 2007; Kelly et al., 2009; Supekar et al., 2009). Thus, developmental findings for both functional and structural covariance networks suggest that our brain develops by pruning short-range connections and strengthening long-range connections (Hagmann et al., 2010).

Structural covariance has also been suggested to be associated with behavioral and cognitive abilities, such as working memory, intellectual abilities, and vocabulary abilities, in developing children and adults (Lee et al., 2014; Lerch et al., 2006; Oh et al., 2011). For instance, children with higher intelligence were shown to have stronger thickness covariance between the left inferior frontal region and other frontal and parietal regions (Lerch et al., 2006). Furthermore, typically developing children with higher vocabulary skills were shown to have greater thickness covariance in language-related regions, including left inferior parietal, inferior temporal, and middle frontal regions (Lee et al., 2014). Similar to other higher cognitive functions, language processes require a highly specialized network within which the distributed areas subserving the relevant processes are integrated. In contrast to vocabulary skills, however, little is known about the cortico-cortical structural covariance network of cortical thickness underlying syntax - a crucial component of the language processing system.

In previous studies using a variety of different tasks, a left temporo-frontal network (perisylvian cortex) has been identified as language-specific (Fedorenko et al., 2011; Lohmann et al., 2010). In particular, the left inferior frontal and posterior superior temporal lobe have been suggested to be involved in the comprehension of syntactically complex sentences in adults (Friederici, 2011, 2009; Grewe et al., 2007; Kinno et al., 2008; Makuuchi et al., 2009; Meltzer et al., 2010; Santi and Grodzinsky, 2010). From the perspective of development, the processing of syntactically complex sentences develops slowly and gradually in children (Skeide and Friederici, 2016). Three-year-old children are able to detect grammatical case-marking cues, but are not able to use this information for sentence comprehension until the age of six years (Schipke et al., 2012). The reason for this slow and gradual development of the processing of syntactically complex sentences can be found in studies examining the functional and structural developmental trajectories of the language-specific brain network. From the functional perspective, for example, preschoolers show reduced activation in the left inferior frontal and superior temporal regions during sentence processing compared to adults (Brauer and Friederici, 2007; Friederici et al., 2011; Wu et al., 2016). Interestingly, this long-range functional connectivity of language relevant brain regions was positively correlated with the preschoolers' competences to process syntactically complex sentences, suggesting individual abilities to be interrelated with the brain's functional development (Xiao et al., 2016). Further, although children start to show functional selectivity in the left posterior superior temporal gyrus (pSTG) between the ages of 9–10 years, they still do not show any adult-like functional selectivity in the left BA44 - a core region for syntactic processes - during syntactic processing (Skeide et al., 2016). Moreover, from the anatomical perspective, a positive correlation was found between sentence comprehension abilities and the gray matter volume of the left inferior frontal gyrus (IFG), inferior temporal gyrus (ITG), and the left parieto-temporal regions in children between the ages of 5–8 years (Fengler et al., 2015). In addition, the fractional anisotropy (FA) of the arcuate fasciculus, referred to as the dorsal language pathway connecting the left IFG and the pSTG (Frey et al., 2008; Wilson et al., 2011), was positively correlated with the sentence comprehension abilities in children between the ages of 3–10 years (Skeide et al., 2016). The dorsal language pathway, which has been suggested to be crucial for syntactic processing, does not fully mature until late childhood (Brauer et al., 2011; Perani et al., 2011). Furthermore, concerning cortical thickness covariance in adolescents, it was found that the cortical thickness of BA44 was co-varied with the thickness in the temporal regions, and this covariance was further suggested to change with age (Lerch et al., 2006). Together, functional and white matter structural connectivity of language-relevant brain regions has been shown to develop across childhood and to be influenced by individual language abilities. However, the development of the cortical thickness covariance of language-relevant brain regions and its relation to syntactic language abilities in children remains unknown.

The aim of the present study was to first investigate how cortical thickness covariance of language-related brain regions in preschoolers is associated with language comprehension abilities in the syntactic domain. In order to do so, we acquired structural magnetic resonance imaging (sMRI) data in 5-year-olds and applied a sentence picture-matching test with two syntactic conditions: 1) simple canonical subject-initial sentences and 2) syntactically complex non-canonical object-initial sentences (Schipke et al., 2012). Structural covariance mappings of cortical thickness were obtained by correlating the cortical thickness of the pre-defined language relevant seeds, that are, the left IFG and left posterior superior temporal gyrus/sulcus (pSTG/STS) (Friederici et al., 2011), to the rest of the cortex across individuals. Further, the relation of cortical thickness covariance mappings with sentence comprehension abilities of syntactically complex sentences in preschoolers was analyzed. Second, to be able to identify the developmental trajectory of cortical thickness covariance of language regions relevant for syntactic processing, we aimed to cross-sectionally compare cortical thickness covariance mappings of preschoolers, school age children, and adults. Thus, we additionally acquired sMRI data of school age children (9–13 years old) and adults (19–33 years old). Third, to demonstrate the agreement of cortical thickness covariance and white matter connectivity, we reconstructed white matter tracts between the left IFG and pSTG/STS using diffusion weighted imaging (DWI) data and compared the white matter tracts across development in relation to cortical thickness covariance mappings. Since DWI scanning parameters of school age children differed from preschool children and adults, we only compared white matter tracts in relation to cortical thickness covariance between preschool children and adults.

First, we expected that young preschool children display immature patterns of cortical thickness structural covariance of seeds that are specifically relevant for syntactic language processing (i.e., left IFG, left pSTG/STS), such that the seeds show covariance only with their contralateral homologous regions. Second, we expected that preschooler's cortical thickness covariance of language-relevant seeds was positively related to their ability to process syntactically complex non-canonical object-initial sentences (i.e., more mature left IFG to left pSTG/STS covariance in preschool children with enhanced language abilities). Third, in comparison to school age children and adults, we expected reduced structural coupling between the left IFG and STG in preschoolers. Fourth, we expected cortical thickness covariance mappings of the language-relevant seeds to develop in a similar way as the white matter tracts connecting the left IFG and left pSTG/STS.

Section snippets

Participants

A total number of 65 five-year-old preschool children (30 girls, mean age = 5.49 years, standard deviation (SD) = 0.28), 52 school age children in the age range of 9–13 years (23 girls, mean age = 10.37 years, SD = 0.59), and 53 adults in the age range of 19–35 years (22 females, mean age = 26.51 years, SD = 3.69) were included in the current study. The three groups were matched concerning gender (χ2̣ = 0.26, p = 0.880). Handedness was assessed using the modified version of Edinburgh Handedness

Seed-based cortical thickness covariance in preschoolers

The seed-based structural covariance mapping of cortical thickness was obtained first for preschoolers (Fig. 2 and Supplementary Table S1). We found that the left IFG only showed significant cortical thickness covariance with its surrounding area and the right IFG in the preschool children. Further, the left pSTG/STS was observed to structurally co-vary with proximal regions of the left pSTG/STS as well as with the bilateral middle temporal gyrus (MTG), bilateral inferior parietal lobule (IPL),

Discussion

In the present study, we investigated how cortico-cortical thickness covariance is associated with the syntactic processing abilities of preschool children, as well as how the syntax-related gray matter structural covariance pattern of cortical thickness develops from preschool children to adults. By computing cortical thickness covariance of two language-associated seeds (i.e., left IFG and left STG/STS), we found restricted covariance with the anatomically proximal and contralateral

Conclusion

In the current study, we demonstrate that the cortical thickness covariance pattern of brain regions relevant for syntactic processes develops from a less mature (i.e., co-varying with the contralateral homologous regions) in preschoolers to a more mature one (i.e., increasing long-range intra-hemispheric covariance) in school age children and adults. The structural covariance of cortical thickness between the left frontal and left temporal regions was shown to be positively related to the

Acknowledgements

This work was funded by the German Research Foundation (project FR 519/20-1 (FOR 2253) awarded to A.D.F.) and by the Max Planck Society and the Fraunhofer Society (Grant number M.FE.A.NEPF0001).

T.Q. was finically supported by the China Scholarship Council (201506040035). We would like to thank Indra Kraft and Kodjo Vissiennon for collecting the data; Thomas C. Gunter and Clara Ekerdt (Max Planck Institute for Human Cognitive and Brain Sciences, Leipzig, Germany) for providing suggestions on our

References (117)

  • G. Gong et al.

    Convergence and divergence of thickness correlations with diffusion connections across the human cerebral cortex

    Neuroimage

    (2012)
  • D.N. Greve et al.

    False positive rates in surface-based anatomical analysis

    Neuroimage

    (2018)
  • T. Grewe et al.

    The role of the posterior superior temporal sulcus in the processing of unmarked transitivity

    Neuroimage

    (2007)
  • X. Han et al.

    Reliability of MRI-derived measurements of human cerebral cortical thickness: the effects of field strength, scanner upgrade and manufacturer

    Neuroimage

    (2006)
  • S.K. Holland et al.

    Normal fMRI brain activation patterns in children performing a verb generation task

    Neuroimage

    (2001)
  • M. Jenkinson et al.

    FSL. Neuroimage

    (2012)
  • L.J. Knoll et al.

    Left prefrontal cortex activation during sentence comprehension covaries with grammatical knowledge in children

    Neuroimage

    (2012)
  • S.K. Krogsrud et al.

    Changes in white matter microstructure in the developing brain—a longitudinal diffusion tensor imaging study of children from 4 to 11 years of age

    Neuroimage

    (2016)
  • C. Lebel et al.

    Microstructural maturation of the human brain from childhood to adulthood

    Neuroimage

    (2008)
  • J.P. Lerch et al.

    Mapping anatomical correlations across cerebral cortex (MACACC) using cortical thickness from MRI

    Neuroimage

    (2006)
  • M. Montembeault et al.

    The impact of aging on gray matter structural covariance networks

    Neuroimage

    (2012)
  • S.C. Nunez et al.

    fMRI of syntactic processing in typically developing children: structural correlates in the inferior frontal gyrus

    Dev. Cogn. Neurosci.

    (2011)
  • H. Oh et al.

    ??-Amyloid affects frontal and posterior brain networks in normal aging

    Neuroimage

    (2011)
  • R.C. Oldfield

    The assessment and analysis of handedness: the Edinburgh inventory

    Neuropsychologia

    (1971)
  • D.K. Patterson et al.

    Bidirectional iterative parcellation of diffusion weighted imaging data: separating cortical regions connected by the arcuate fasciculus and extreme capsule

    Neuroimage

    (2014)
  • T. Qi et al.

    More bilateral, more anterior: alterations of brain organization in the large-scale structural network in Chinese dyslexia

    Neuroimage

    (2016)
  • A. Santi et al.

    fMRI adaptation dissociates syntactic complexity dimensions

    Neuroimage

    (2010)
  • W.W. Seeley et al.

    Neurodegenerative diseases target large-scale human brain networks

    Neuron

    (2009)
  • M.A. Skeide et al.

    Syntax gradually segregates from semantics in the developing brain

    Neuroimage

    (2014)
  • A. Alexander-Bloch et al.

    Imaging structural co-variance between human brain regions

    Nat. Rev. Neurosci.

    (2013)
  • A. Alexander-Bloch et al.

    The convergence of maturational change and structural covariance in human cortical networks

    J. Neurosci.

    (2013)
  • A.L. Alexander et al.

    Characterization of cerebral white matter properties using quantitative magnetic resonance imaging stains

    Brain Connect.

    (2011)
  • M. Ashtari et al.

    White matter development during late adolescence in healthy males: a cross-sectional diffusion tensor imaging study

    Neuroimage

    (2007)
  • T.R. Barrick et al.

    White matter structural decline in normal ageing: a prospective longitudinal study using tract-based spatial statistics

    Neuroimage

    (2010)
  • S. Bava et al.

    Longitudinal characterization of white matter maturation during adolescence

    Brain Res.

    (2010)
  • C. Beaulieu

    The basis of anisotropic water diffusion in the nervous system - a technical review

    NMR Biomed.

    (2002)
  • T.E.J. Behrens et al.

    Characterization and propagation of uncertainty in diffusion-weighted MR imaging

    Magn. Reson. Med.

    (2003)
  • B.C. Bernhardt et al.

    Structural covariance networks of the dorsal anterior insula predict females' individual differences in empathic responding

    Cerebr. Cortex

    (2014)
  • B.C. Bernhardt et al.

    Selective disruption of sociocognitive structural brain networks in autism and alexithymia

    Cerebr. Cortex

    (2014)
  • M.J. Blanca et al.

    Effect of variance ratio on ANOVA robustness: might 1.5 be the limit?

    Behav. Res. Methods

    (2017)
  • J. Brauer et al.

    Neuroanatomical prerequisites for language functions in the maturing brain

    Cerebr. Cortex

    (2011)
  • J. Brauer et al.

    Functional neural networks of semantic and syntactic processes in the developing brain

    J. Cognit. Neurosci.

    (2007)
  • R. Cafiero et al.

    The concurrence of cortical surface area expansion and white matter myelination in human brain development

    Cerebr. Cortex

    (2018)
  • E.L. Dennis et al.

    Development of brain structural connectivity between ages 12 and 30: a 4-Tesla diffusion imaging study in 439 adolescents and adults

    Neuroimage

    (2013)
  • B. Draganski et al.

    Neuroplasticity: changes in grey matter induced by training

    Nature

    (2004)
  • T.J. Eluvathingal et al.

    Quantitative diffusion tensor tractography of association and projection fibers in normally developing children and adolescents

    Cerebr. Cortex

    (2007)
  • D.A. Fair et al.

    Functional brain networks develop from a “local to distributed” organization

    PLoS Comput. Biol.

    (2009)
  • D.A. Fair et al.

    Development of distinct control networks through segregation and integration

    Proc. Natl. Acad. Sci. Unit. States Am.

    (2007)
  • E. Fedorenko et al.

    Functional specificity for high-level linguistic processing in the human brain

    Proc. Natl. Acad. Sci. U. S. A

    (2011)
  • E. Fedorenko et al.

    Reworking the language network

    Trends Cognit. Sci.

    (2014)
  • Cited by (17)

    • Associated functional network development and language abilities in children

      2021, NeuroImage
      Citation Excerpt :

      For instance, a long-range white matter connection between the left IFG and left posterior STG and superior temporal sulcus (STS) has been reported to emerge gradually with age and to relate to language abilities (Friederici, 2011; Skeide and Friederici, 2016). Moreover, brain regions of the gray matter linked to this fiber tract, e.g., the left IFG, left inferior parietal lobule (IPL), and left supramarginal gyrus, have also been observed to be associated with language skills in children and to change with development (Fengler et al., 2015; Lu et al., 2007; Qi et al., 2019a, 2019b; Richardson et al., 2010; Sowell et al., 2004). In general, myelination development parallels cortical growth, such as cortical surface expansion and thinning (Cafiero et al., 2018; Natu et al., 2019), gradually developing across frontal, temporal, and parietal regions (Brown et al., 2012; Cafiero et al., 2018; Raznahan et al., 2011; Tamnes et al., 2017), which are proposed to be associated with language comprehension (Roehrich-Gascon et al., 2015).

    • Warped rhythms: Epileptic activity during critical periods disrupts the development of neural networks for human communication

      2021, Behavioural Brain Research
      Citation Excerpt :

      Here the authors compared covariance of cortical thickness in syntactic network nodes in preschoolers (5-year olds), school-aged children (9−13 year olds), and young adults (18−33 year olds). This study found that covariance patterns developed from being relatively localized in preschoolers to more widespread in adults, and this change was linked to maturation in white matter tracts [48]. Thus, language networks show a consistent pattern of lateralization over early childhood as well as overall similar maturation patterns of increased within network connectivity that develop at a pace commensurate with language abilities.

    • Structural covariance network and applications in brain imaging

      2023, Chinese Journal of Radiology (China)
    View all citing articles on Scopus
    View full text