Elsevier

NeuroImage

Volume 55, Issue 4, 15 April 2011, Pages 1443-1453
NeuroImage

Cortical thickness correlates of specific cognitive performance accounted for by the general factor of intelligence in healthy children aged 6 to 18

https://doi.org/10.1016/j.neuroimage.2011.01.016Get rights and content

Abstract

Prevailing psychometric theories of intelligence posit that individual differences in cognitive performance are attributable to three main sources of variance: the general factor of intelligence (g), cognitive ability domains, and specific test requirements and idiosyncrasies. Cortical thickness has been previously associated with g. In the present study, we systematically analyzed associations between cortical thickness and cognitive performance with and without adjusting for the effects of g in a representative sample of children and adolescents (N = 207, Mean age = 11.8; SD = 3.5; Range = 6 to 18.3 years). Seven cognitive tests were included in a measurement model that identified three first-order factors (representing cognitive ability domains) and one second-order factor representing g. Residuals of the cognitive ability domain scores were computed to represent g-independent variance for the three domains and seven tests. Cognitive domain and individual test scores as well as residualized scores were regressed against cortical thickness, adjusting for age, gender and a proxy measure of brain volume. g and cognitive domain scores were positively correlated with cortical thickness in very similar areas across the brain. Adjusting for the effects of g eliminated associations of domain and test scores with cortical thickness. Within a psychometric framework, cortical thickness correlates of cognitive performance on complex tasks are well captured by g in this demographically representative sample.

Research Highlights

► Abilities examined here had very similar patterns of cortical thickness associations. ► This was so despite important differences in the nature of the examined abilities. ► No cortical thickness correlates remained after controlling for g. ► This remained the case despite using very liberal thresholds.

Introduction

Human intelligence has been defined as “a very general mental capability that, among other things, involves the ability to reason, plan, solve problems, think abstractly, comprehend complex ideas, learn quickly and learn from experience. It is not merely book learning, a narrow academic skill, or test-taking smarts. Rather, it reflects a broader and deeper capability for comprehending our surroundings — ‘catching on’, ‘making sense’ of things, or ‘figuring out’ what to do” (Gottfredson, 1997). This definition, which stresses the general nature of the intelligence construct, is consistent with the phenomenon called ‘positive manifold’: scores on cognitive ability tasks of all kinds are positively correlated. In other words, people who achieve high scores on a test of any one aspect of intelligence are likely to score above average on others (Neisser et al., 1996). This empirical fact is nuanced by the finding that scores on tests that are more similar in content are more closely correlated with each other than with tests that have different content (Deary et al., 2010).

A widely accepted framework for representing these correlational patterns is a hierarchical arrangement with a general intelligence factor (g) at the apex that contributes to several more specialized cognitive ability domains arrayed below it, which in turn contribute to individual test scores (Carroll, 1993, Deary et al., 2010, Johnson et al., 2004, Johnson et al., 2008, Neisser et al., 1996). According to this view, individual differences in test performance can be accounted for by the combined influence of general intelligence (g), specific cognitive ability domains (sometimes called group factors of intelligence), and skills specific to each test. Of course, apart from this hierarchy of sources of cognitive variance, observed scores will reflect error variance and non-cognitive individual influences at the time of performing the test. That said, the general factor, g, is typically the major source of variance in test scores, accounting for 40% or more of total variance in performance on mental test batteries in representative samples (Carroll, 1993, Deary et al., 2010, Jensen, 1998).

Recently, efforts have been made to link components of this psychometric model of intellectual function to measures of brain structure and function revealed by brain imaging (Deary et al., 2010, Haier et al., 2009a). Indeed, it is reasonable to hypothesize that neural correlates of performance can be found for each of the components of the model. It has therefore been suggested that these sources of variance be partitioned in order to assess the respective contributions of brain structure and/or function to cognitive performance for the different levels of the model (Colom et al., 2009, Colom et al., 2007, Haier et al., 2009a, Jung and Haier, 2007).

An accumulating body of work has linked specific features of brain structure to general intelligence with a fair amount of consistency. A recent review concluded that a distributed network of multimodal association areas consisting of the dorsolateral prefrontal cortex (DLPF), the inferior and superior parietal lobule, the anterior cingulate cortex (ACC) and parts of the temporal and occipital lobes is highly correlated structurally, functionally and/or biochemically to general intellectual ability (Jung and Haier, 2007). These findings led to the proposal of a Parieto-Frontal Integration Theory (P-FIT) of intelligence; according to which, sensory information is first processed by temporal and occipital areas for subsequent integration and abstraction in parietal areas. Problem evaluation is then implemented by the prefrontal cortex and response selection mediated via the anterior cingulate. Consistent with this proposal, Colom et al. (2009) used voxel-based morphometry to demonstrate positive associations between scores representing g and gray matter concentration in several clusters including the dorsolateral prefrontal cortex, Broca's and Wernicke's areas, the somato-sensory association cortex, and the visual association cortex. Converging findings came from a study of 241 patients with focal brain damage, using voxel-based lesion-symptom mapping (Glascher et al., 2010), in which a measure of g was associated with damage within a distributed network in frontal and parietal brain regions as well as with lesions in white matter association tracts in frontopolar areas. These findings suggested that the general factor of intelligence reflects the efficacy of regions integrating verbal, visuospatial, working memory, and executive processes and their connections.

In a prior study, we demonstrated associations between a measure of g and measures of cortical thickness in a representative sample of 216 healthy children and adolescents between the ages of 6 and 18 from the NIH MRI Study of Normal Brain Development (Karama et al., 2009). Cortical thickness has significant advantages in terms of precision and interpretability over voxel-based morphometry, as detailed elsewhere (Karama et al., 2009). Significant associations were documented in most multimodal association areas of the cerebral cortex, somewhat more pronounced on the left side. Although the associations appeared somewhat stronger in adolescents than children, these differences did not achieve statistical significance. These results were generally consistent with the P-FIT model, although more associations were found with medial structures than that model would have predicted.

To date, neuroimaging studies have addressed structural brain correlates of crystallized and fluid intelligence (Gray and Thompson, 2004) as well as of verbal and performance IQ, in addition to full scale IQ (Andreasen et al., 1993, Luders et al., 2007, Luders et al., 2011, Witelson et al., 2006). However, brain imaging studies have yet to characterize the structural correlates of the different levels of the psychometric model including more specific cognitive ability domains and test performance. The objective of the current study was to extend this line of research further by examining associations between local measures of cortical thickness and (a) the general factor of intelligence; (b) cognitive domains/group factors; and (c) specific test scores in the same sample of children and adolescents from the NIH MRI Study of Normal Brain Development reported previously (Karama et al., 2009). In light of previous work (Colom et al., 2009, Deary et al., 2010, Flashman et al., 1997, Glascher et al., 2010, Jung and Haier, 2007, Luders et al., 2009), associations between cortical thickness and the various measures of cognitive performance were expected to be widely distributed and to cluster around association areas, but to differ depending upon task demands. Given the large proportion of test score variance explained by g, however, we hypothesized that adjusting statistically for the effects of g would significantly attenuate the degree and extent of the associations between cognitive domain and test scores, and local cortical thickness, thereby revealing more constrained and localized areas of association for the more specific abilities at lower levels of the hierarchy.

Section snippets

Sampling and recruitment

Data were obtained from the Pediatric MRI Data Repository (Release 2.0) created for the NIH MRI Study of Normal Brain Development (Evans and Brain Development Cooperative Group, 2006), a multi-site longitudinal project aimed at providing a normative database to characterize healthy brain maturation in relation to behavior. A listing of the participating sites and of the study investigators can be found at: http://www.bic.mni.mcgill.ca/nihpd/info/participating_centers.html.

This data base

Results

Cortical thickness was positively associated with the full g score (i.e., the g score extracted from the complete set of 7 tests) in a wide network of bilateral areas (Fig. 3). The magnitudes of these correlations for statistically significant foci were in the modest to moderate range (0.15 to 0.34). Analogous analyses for the three cognitive ability domain scores revealed a similar bilateral and distributed pattern of association (see left side of Fig. 4). Although not surprising given their

Discussion

The most important finding of the present work was that, after adjusting statistically for the effects of g, no significant associations between cortical thickness and performance remained for any of the cognitive domains or specific ability tests used here. This finding persisted, moreover, when a more lenient threshold of significance was applied to look for trends. Most of the association between the psychometrically distinct domains of cognitive performance measured here and local cortical

Conclusion

In conclusion, after adjusting for the effects of g, we did not find significant associations between cortical thickness and performance on more specific cognitive domains or ability tests for a large representative sample of healthy children and adolescents. It follows that cortical thickness correlates of cognitive performance on complex tasks are well captured by g. An important implication is that patterns of associations between cortical thickness (and perhaps other neural features) and

Disclaimer

The views herein do not necessarily represent the official views of the National Institute of Child Health and Human Development, the National Institute on Drug Abuse, the National Institute of Mental Health, the National Institute of Neurological Disorders and Stroke, the National Institutes of Health, the U.S. Department of Health and Human Services, or any other agency of the United States Government.

Acknowledgments

This project was conducted by the Brain Development Cooperative Group and supported by the National Institute of Child Health and Human Development, the National Institute on Drug Abuse, the National Institute of Mental Health, and the National Institute of Neurological Disorders and Stroke (Contract #s N01-HD02-3343, N01-MH9-0002, and N01-NS-9-2314, -2315, -2316, -2317, -2319 and -2320). Special thanks to the NIH contracting officers for their support. S.K. was supported by a Fellowship from

References (59)

  • W. Johnson et al.

    Just one g: consistent results from three test batteries

    Intelligence

    (2004)
  • W. Johnson et al.

    Still just 1 g: consistent results from five test batteries

    Intelligence

    (2008)
  • R.E. Jung et al.

    Sex differences in N-acetylaspartate correlates of general intelligence: an 1H-MRS study of normal human brain

    Neuroimage

    (2005)
  • S. Karama et al.

    Positive association between cognitive ability and cortical thickness in a representative US sample of healthy 6 to 18 year-olds

    Intelligence

    (2009)
  • J.S. Kim et al.

    Automated 3-D extraction and evaluation of the inner and outer cortical surfaces using a Laplacian map and partial volume effect classification

    Neuroimage

    (2005)
  • J.P. Lerch et al.

    Cortical thickness analysis examined through power analysis and a population simulation

    Neuroimage

    (2005)
  • E. Luders et al.

    Positive correlations between corpus callosum thickness and intelligence

    Neuroimage

    (2007)
  • E. Luders et al.

    Neuroanatomical correlates of intelligence

    Intelligence

    (2009)
  • E. Luders et al.

    The link between callosal thickness and intelligence in healthy children and adolescents

    Neuroimage

    (2011)
  • O. Lyttelton et al.

    An unbiased iterative group registration template for cortical surface analysis

    Neuroimage

    (2007)
  • D. MacDonald et al.

    Automated 3-D extraction of inner and outer surfaces of cerebral cortex from MRI

    Neuroimage

    (2000)
  • K.L. Narr et al.

    Asymmetries of cortical shape: effects of handedness, sex and schizophrenia

    Neuroimage

    (2007)
  • R.C. Oldfield

    The assessment and analysis of handedness: the Edinburgh inventory

    Neuropsychologia

    (1971)
  • M. Song et al.

    Brain spontaneous functional connectivity and intelligence

    Neuroimage

    (2008)
  • Y. Ad-Dab'bagh et al.

    The CIVET Image-Processing Environment: A Fully Automated Comprehensive Pipeline for Anatomical Neuroimaging Research

  • N.C. Andreasen et al.

    Intelligence and brain structure in normal individuals

    Am. J. Psychiatry

    (1993)
  • Y. Benjamini et al.

    Controlling the false discovery rate: a practical and powerful approach to multiple testing

    J. R. Stat. Soc. B Methodol.

    (1995)
  • K.A. Bollen

    Structure Equations with Latent Variables

    (1989)
  • J. Boyke et al.

    Training-induced brain structure changes in the elderly

    J. Neurosci.

    (2008)
  • Cited by (137)

    View all citing articles on Scopus
    1

    See Appendix B for author list and affiliations of the Brain Development Cooperative Group.

    View full text