Molecular Psychiatry (2014) 19, 380–384 OPEN

& 2014 Macmillan Publishers Limited All rights reserved 1359-4184/14

www.nature.com/mp

ORIGINAL ARTICLE
DNA evidence for strong genetic stability and increasing
heritability of intelligence from age 7 to 12
M Trzaskowski1, J Yang2, PM Visscher2,3 and R Plomin1

Two genetic findings from twin research have far-reaching implications for understanding individual differences in the
development of brain function as indexed by general cognitive ability (g, aka intelligence): (1) The same genes affect g throughout
development, even though (2) heritability increases. It is now possible to test these hypotheses using DNA alone. From 1.7 million
DNA markers and g scores at ages 7 and 12 on 2875 children, the DNA genetic correlation from age 7 to 12 was 0.73, highly similar
to the genetic correlation of 0.75 estimated from 6702 pairs of twins from the same sample. DNA-estimated heritabilities increased
from 0.26 at age 7 to 0.45 at age 12; twin-estimated heritabilities also increased from 0.35 to 0.48. These DNA results confirm the
results of twin studies indicating strong genetic stability but increasing heritability for g, despite mean changes in brain structure
and function from childhood to adolescence.

Molecular Psychiatry (2014) 19, 380–384; doi:10.1038/mp.2012.191; published online 29 January 2013
Keywords: cognitive development; Genome-wide complex trait analysis (GCTA); intelligence; twins

INTRODUCTION independent of heritability.11 The high genetic correlation

Although developmental research from childhood to adolescence implies that if a gene is found to be associated with g in
reveals species-general changes in brain structure and function,1,2 childhood, the gene is also highly likely to be associated with g in
much less is known about the development of individual adolescence. Later, we offer a hypothesis as to how heritability can
differences within our species, which has been called ‘one of increase when genetic effects are stable from age to age.
the preeminent challenges of neuroimaging’.3 It is important to These two genetic findings have not found much traction in the
understand the developmental etiology of individual differences, neurodevelopmental literature. This neglect might be due in part
because societal problems often involve individual differences— to a lack of attention to individual differences, but it might also be
for example, why some children are slow to speak, to learn or to due to skepticism about the twin method, which relies on some
read. The description and causes of species’ means are not major assumptions, most notably, equal environmental treatment
necessarily related to the description and causes of variances of monozygotic and DZ twins.11 Quantitative genetic designs such
within a species.4 Two well-replicated genetic findings from twin as the twin method would no longer be needed if it were possible
studies comparing monozygotic and dizygotic (DZ) twins suggest to identify all of the genes responsible for heritability.13 However,
hypotheses at the level of individual differences in cognitive it has proven more difficult than expected to identify genes for
ability that may be relevant to neuroscience, to the extent that complex traits,14 including g,15 which has led to the refrain of
brain structure and function underlie cognitive outcomes. These ‘missing heritability’.16,17 Nonetheless, it is now possible to use
twin-study findings involve general cognitive ability, which was DNA itself to estimate genetic variance and covariance in any
labeled g by Spearman more than a century ago,5 but is sample of unrelated individuals, not just samples consisting of
commonly known as intelligence.6 g is the most researched special family members such as twins or adoptees. The method,
cognitive trait in genetics7 and has important links with called genome-wide complex trait analysis (GCTA)18 correlates
neuroscience.8,9 genomic similarity across hundreds of thousands of single
First, the heritability of g increases during development, even nucleotide polymorphisms (SNPs) with phenotypic similarity in a
from childhood to adolescence.10 This finding is counterintuitive large sample of unrelated individuals.19 This population-based
to the extent that genetic effects are thought to be static, and DNA approach does not rely on the strong assumptions made in
environmental effects are expected to accumulate during classical twin studies. GCTA compares similarity across hundreds
development. The increasing heritability of g also seems at odds of thousands of SNPs with phenotypic similarity pair by pair in a
with the second genetic finding: The same genes largely affect g large sample of unrelated individuals. Although conventionally
throughout development.11 For example, in a longitudinal twin unrelated individuals only vary in their genetic similarity by a small
analysis from childhood to adolescence, the genetic correlation amount, GCTA accumulates all the genotype  phenotype
was estimated as 0.96, although the 95% confidence interval for association signals using the massive information available in a
this estimate was 0.74–1.0.12 The genetic correlation is literally the matrix of thousands of individuals, each compared pair by pair
correlation between genetic effects on g at the two ages with every other individual in the sample. GCTA has been used to

1
King’s College London, MRC Social, Genetic and Developmental Psychiatry Centre, Institute of Psychiatry, Denmark Hill, London, UK; 2University of Queensland Diamantina
Institute, The University of Queensland, Princess Alexandra Hospital, Brisbane, Queensland, Australia and 3Queensland Brain Institute, The University of Queensland, Brisbane,
Queensland, Australia. Correspondence: M Trzaskowski, MSc, King’s College London, MRC Social, Genetic and Developmental Psychiatry Centre, PO80, Institute of Psychiatry,
DeCrespigny Park, Denmark Hill, London SE5 8AF, UK.
E-mail: [email protected]
Received 16 October 2012; accepted 12 November 2012; published online 29 January 2013
 Genetic stability but increasing heritability
M Trzaskowski et al
381
estimate genetic influence for height,19 weight,20 psychiatric and elsewhere;36 at age 12, testing was conducted online.37 For each cognitive
medical disorders,21–23 personality24 and even economic and measure at each age, scores were regressed on sex and age and
political preferences.25 GCTA has also been applied to g in adults26 standardized residuals were derived, ranked with random values given to
and children.27 These GCTA estimates of genetic influence, tied data, and quantile normalized.38,39 Finally, total composites for g were
although substantial, have been lower than heritability typically created as unit-weighted means requiring complete data for at least three
of the four tests. All the procedures were executed using R (www.
found in twin studies of these traits. Using the 12-year data from r-project.org).40
the sample in the present report, GCTA and twin estimates of
heritability were compared explicitly for several cognitive
measures; the GCTA estimate of g was 35% and the twin Statistical analyses
estimate was 46%.28 Precision in comparing GCTA and twin Genome-wide complex trait analysis. The first step in GCTA is to calculate
estimates is important because, as explained later, this comparison pairwise genomic similarity between all pairs of individuals in the sample
reveals important information about a trait’s genetic architecture. using all genetic markers genotyped on the SNP array. Because GCTA is
designed to estimate genetic variance due to linkage disequilibrium
This previous GCTA research involves univariate analysis in that
between unknown causal variants and genotyped SNPs from a sample of
it decomposes the phenotypic variance of a single trait into unrelated individuals in the population, any close genetic relatedness is
genetic and non-genetic components of variance. Recently, GCTA eliminated; for this reason any individual whose genetic similarity is equal
has been extended to bivariate analysis, which decomposes the to or greater than a fourth cousin is removed (estimate of pairwise
phenotypic covariance between traits into components of relatedness 40.025). The essence of GCTA is to compare a matrix of
covariance. The first preliminary attempt to extend GCTA to pairwise genomic similarity to a matrix of pairwise phenotypic similarity
bivariate analysis reported a genetic correlation of 0.62 for g in using a random-effects mixed linear model.18 In univariate analysis, the
childhood (age 11) and old age.27 Here, we use a new bivariate variance of a trait can be partitioned using residual maximum likelihood
GCTA method18,29 to test the hypotheses of strong stability and into genetic and residual components. Detailed description of this method
can be found in Yang, Lee et al.18 and Yang, Benyamin et al.19 The bivariate
increasing heritability for g from age 7 to 12. We also compare
method extends the univariate model by relating the pairwise genetic
GCTA estimates with those from a twin analysis based on the similarity matrix to a phenotypic covariance matrix between traits 1 and 2,
same sample at the same ages using the same measures. allowing for correlated residuals.29 The eight principal components
described earlier were used as covariates in our GCTA analyses; as
mentioned, all phenotypes were age- and sex-regressed before analysis.
MATERIALS AND METHODS
Sample Twin modeling. The classical twin design and model-fitting is discussed
The sample was drawn from the Twins Early Development Study (TEDS), elsewhere.11 We fit a bivariate twin model using OpenMx,41 which
which is a multivariate longitudinal twin-study that recruited over 11 000 provided a direct comparison with the bivariate GCTA. The correlated
twin pairs born in England and Wales in 1994, 1995 and 1996.30,31 TEDS is factor solution is the least restricted model allowing variables to correlate
representative of the UK population.32 The project received approval from with one another via genetic, shared environment and non-shared
the Institute of Psychiatry ethics committee (05/Q0706/228), and parental environment. Because previous analyses of these data indicated
consent was obtained before data collection. Individuals were included if nonsignificant differences in model-fitting results between males and
their first language was English and they had no major medical or females,32,42 we combined same-sex and opposite DZ twin pairs in order to
psychiatric problems. GCTA was conducted on g at ages 7 and 12 for 2875 increase the power of the analyses. Twin analyses limited to same-sex
unrelated individuals in TEDS (only one member of a twin pair), of which twins yielded highly similar results (available from the first author).
1334 had g data at both ages. Twin model-fitting analyses of g at ages 7
and 12 were conducted for 6702 TEDS twin pairs, of which 2269 pairs had
g data at both ages. As expected for representative twin studies, the twins RESULTS AND DISCUSSION
included similar numbers of monozygotic twins, same-sex DZ twins and
Genetic stability
opposite-sex DZ twins.
As shown in Table 1, the GCTA genetic correlation between g at
ages 7 and 12 was 0.73 (0.29 standard error, s.e.). Table 2 shows
Genotyping
that the twin-study yielded a highly similar genetic correlation of
Although DNA is available for more than 12 000 TEDS participants, funds
0.75 (0.08 s.e.). The genetic correlation indexes the correlation
were available to genotype 3665 individuals (one member only per twin
pair) on Affymetrix GeneChip 6.0 (Affymetrix Inc., Santa Clara, CA, USA) SNP between genetic effects on g at the two ages independent of
genotyping arrays using standard experimental protocols as part of the heritability. That is, the genetic correlation can be high even if
WTCCC2 project. In addition to nearly 700 000 genotyped SNPs, more than heritability is low. It is also possible to weight the genetic
one million other SNPs were imputed using IMPUTE v.2 software (https:// correlation by heritability in order to estimate the genetic
mathgen.stats.ox.ac.uk/impute/impute.html).33 DNA for 3152 individuals contribution to the phenotypic correlation. The phenotypic
(1446 males and 1706 females) survived quality control criteria. Of these correlation for g between ages 7 and 12 was 0.46 (0.02) for
3152 individuals, 2875 had g scores at least at one age and 1344 had g 2408 children (one member randomly chosen from each twin pair)
scores at both ages. To control for ancestral stratification, we performed with g data at both ages. For GCTA, the genetic contribution to
principal component analyses on a subset of 100 000 quality-controlled
the phenotypic correlation was 0.25 (0.11), which is the GCTA
SNPs after removing SNPs in linkage disequilibrium (r240.2).34 Using the
Tracy  Widom test,35 we identified 8 axes with Po0.05, which were used genetic correlation weighted by heritability (that is, the product of
as covariates in GCTA analyses. the square roots of the GCTA heritabilities of g at the two ages).
Another way of expressing this is as bivariate heritability, which is
the proportion of the phenotypic correlation that can be
Measures
attributed to genetic covariance. GCTA bivariate heritability was
The measures and testing procedures have been described in detail for
0.60 (that is, 0.25C0.42), indicating that 60% of the phenotypic
age 736 and 12.37 At each age, a composite index of g was derived from
two verbal tests and two non-verbal tests. At age 7, the two verbal tests correlation could be accounted for by genetic factors. The
consisted of the Similarities subtest and the Vocabulary subtest from the comparable twin-study estimate of the genetic contribution to
WISC-III-UK, and the two non-verbal tests were the picture completion the phenotypic correlation was 0.31 (0.03), yielding a bivariate
subtest from the WISC-III-UK and the Conceptual Grouping subtest from heritability of 0.68.
the McCarthy Scales of Children’s Abilities. At age 12, the verbal tests
included the Information and Vocabulary subtests from the WISC-III-PI
Multiple Choice test, and the two non-verbal reasoning tests were WISC-III-
Increasing heritability
UK Picture Completion and Raven’s Standard and Advanced Progressive Despite the substantial genetic correlation of 0.73 from age 7–12,
Matrices. At age 7, testing was conducted by telephone as described GCTA estimates of genetic influence on g increased from 0.26

& 2014 Macmillan Publishers Limited Molecular Psychiatry (2014), 380 – 384
 Genetic stability but increasing heritability
M Trzaskowski et al
382
Table 1. Bivariate GCTA results (with standard errors) for general cognitive ability (g) from age 7 to 12a

(a) Genetic
Genetic variance at age 7 Genetic variance at age 12 Genetic covariance between age 7 and 12 Genetic correlation between age 7 and 12

0.26 (.17) 0.45 (0.14) 0.25 (0.11) 0.73 (0.29)

(b) Environmental
Environmental variance at Environmental variance at Environmental covariance between age 7 Environmental correlation between age 7
age 7 age 12 and 12 and 12

0.74 (0.17) 0.55 (0.14) 0.18 (0.11) 0.28 (0.15)

Abbreviation: GCTA, genome-wide complex trait analysis.
a
GCTA incorporates full-information maximum likelihood that uses the full sample of 2875 individuals with data at either 7 or 12. However, the variance
estimates at each age are based on individuals with data at that age (1908 at 7, 2329 at 12) and the covariance estimates are based on individuals with data at
both ages (1344).
b
The current version of GCTA does not report the environmental correlation or its standard error. The environmental correlation was derived here from the
GCTA estimates using the following algorithm: C(e)_tr12 / (OV(e)_tr1 *OV(e)_tr2), whereas the standard error was calculated using: Var(re) ¼ re * re * (VarVe1/
(4*Ve1*Ve1) þ VarVe2/(4*Ve2*Ve2) þ VarCe/(Ce*Ce) þ CovVe1Ve2/(2*Ve1*Ve2) - CovVe1Ce/(Ve1*Ce) - CovVe2Ce/(Ve2*Ce)); SE(re) ¼ O[Var(re)] where re is the
environmental correlation, Ve1 is the residual variance for trait 1, Ce is the residual covariance between two traits, VarVe1 is the sampling variance for Ve1
(residual variance for trait 1), VarCe is the sampling variance for Ce, CovVe1Ve2 is the sampling covariance between Ve1 and Ve2, and CovVe1Ce is the
sampling covariance between Ve1 and Ce.

Table 2. Bivariate twin model-fitting results (with standard errors) for general cognitive ability from age 7 to 12a

Genetic
Genetic variance at age 7 Genetic variance at age 12 Genetic covariance between age 7 and 12 Genetic correlation between age 7 and 12

0.36 (0.03) 0.49 (0.04) 0.31 (0.03) 0.75 (0.08)

Shared environment (C)

C variance at age 7 C variance at age 12 C covariance between age 7 and 12 C correlation between age 7 and 12

0.31 (0.03) 0.19 (0.03) 0.12 (0.03) 0.48 (0.11)

Non-shared environment (E)

E variance at age 7 E variance at age 12 E covariance between age 7 and 12 E correlation between age 7 and 12

0.33 (0.01) 0.32 (0.01) 0.03 (0.01) 0.09 (0.03)

a
OpenMx twin model-fitting incorporates full-information maximum likelihood that uses the full sample of 6702 pairs of twins with data at either 7 or 12.
However, the variance estimates at each age are based on twin pairs with data at that age (5320 at 7, 4061 at 12), and the covariance estimates are based on
twin pairs with data at both ages (2269).

(0.17 s.e.) at age 7 to 0.45 (0.14 s.e.) at age 12, although the large that genetic links between g and experience should strengthen
standard errors indicate that the increase did not reach statistical during development, but this has not yet been investigated. These
significance. Heritability increased significantly in the twin model- genetic links are expected especially for experiences in which
fitting analyses, from 0.36 (0.03) at age 7 to 0.49 (0.03) at age 12. children are able to select or modify their environments in line
Thus, GCTA estimates account for 74% of the twin-study with their genetic propensities, in contrast to environments
heritability estimate of g at age 7 and 94% at age 12. that are passively imposed on children. Supportive evidence to
date for this genotype  environment hypothesis relies on twin
Why genetic stability but increasing heritability? data, but GCTA can also be used to address these issues with DNA
In summary, GCTA confirms the twin-study hypotheses of strong alone.
genetic stability and increasing heritability. In other words, the
same genes are largely (about 75%) responsible for genetic Genetic architecture
influence on g at age 7 and age 12, yet the effect of these genes Our GCTA results clarify the genetic architecture of g in ways that
(heritability) increases substantially from age 7 to 12. How is this are relevant to solving the ‘missing heritability’ puzzle that has
possible? We hypothesize that the same genes affect g from age emerged from the limited success of genome-wide association
to age but heritability increases as children select their own studies to identify the genes responsible for heritability.46 Two of
environments that are correlated with their g-related genetic the major hypotheses to account for missing heritability are
propensities,10 a process called genotype  environment epistatic (nonadditive) genetic effects and rare variants, because
correlation.11 This hypothesis makes three predictions. The first genome-wide association research is limited to detecting additive
prediction is that g-related experiences will themselves show genetic effects and genetic effects that can be tagged by the
genetic influence, for which there is considerable evidence from common SNPs used to date on commercially available DNA
twin studies.43,44 Second, the links between these experiences and arrays.19 Because GCTA is also limited in these same two ways,
g are expected to be mediated genetically, evidence which is finding significant GCTA estimates of genetic influence provides
beginning to emerge from twin studies.45 The third prediction is strong evidence that current genome-wide association research

Molecular Psychiatry (2014), 380 – 384 & 2014 Macmillan Publishers Limited
 Genetic stability but increasing heritability
M Trzaskowski et al
383
strategies can detect the majority of the missing heritability if 7 Deary IJ, Johnson W, Houlihan LM. Genetic foundations of human intelligence.
samples are sufficiently large to provide power to detect Hum Genet 2009; 126: 215–232.
associations of small effect size. As noted above, our GCTA 8 Deary IJ, Penke L, Johnson W. The neuroscience of human intelligence differ-
estimates of genetic influence account for 74–94% of our twin- ences. Nat Rev Neurosci 2010; 11: 201–211.
study heritability estimates, which implies that most of the 9 Karama S, Colom R, Johnson W, Deary IJ, Haier R, Waber DP et al. Cortical thick-
missing heritability can be found with additive effects of common ness correlates of specific cognitive performance accounted for by the general
factor of intelligence in healthy children aged 6–18. Neuroimage 2011; 55:
SNPs. The heritability that remains missing might be due to
1443–1453.
epistatic effects and rare variants. 10 Haworth CMA, Wright MJ, Luciano M, Martin NG, de Geus EJC, van Beijsterveldt
In our longitudinal genetic analyses from age 7 to 12, the GCTA CEM et al. The heritability of general cognitive ability increases linearly from
estimate of genetic covariance is also somewhat lower than the childhood to young adulthood. Mol Psychiatry 2010; 15: 1112–1120.
twin-study estimate. As shown in Table 1, the genetic covariance 11 Plomin R, DeFries JC, Knopik VS, Neiderhiser JM. Behavioral Genetics. 6th edn.
for g between ages 7 and 12—that is, the genetic contribution to Worth: New York, 2013.
the phenotypic covariance—is 20% lower for GCTA than for twins 12 van Soelen ILC, Brouwer RM, van Leeuwen M, Kahn RS, Pol HEH, Boomsma DI.
(that is, 0.25 for GCTA and 0.31 for twins). However, the GCTA Heritability of verbal and performance intelligence in a pediatric longitudinal
genetic correlation of 0.73 is highly similar to the twin-study sample. Twin Res Hum Genet 2011; 14: 119–128.
13 Haworth CMA, Plomin R. Quantitative genetics in the era of molecular genetics:
genetic correlation of 0.76. The likely reason is that GCTA genetic
learning abilities and disabilities as an example. J Am Acad Child Adolesc Psychiatry
variance and covariance estimates are attenuated by imperfect 2010; 49: 783–793.
linkage disequilibrium between causal variants and genotyped 14 McCarthy MI, Abecasis GR, Cardon LR, Goldstein DB, Little J, Ioannidis JPA et al.
SNPs, but the GCTA estimate of the genetic correlation is Genome-wide association studies for complex traits: consensus, uncertainty and
unbiased, because the genetic correlation is derived from the challenges. Nat Rev Genet 2008; 9: 356–369.
ratio between genetic covariance and genetic variance. Because 15 Chabris CF, Hebert BM, Benjamin DJ, Beauchamp JP, Cesarini D, van der Loos M et
GCTA genetic variance and covariance estimates are biased to the al. Most reported genetic associations with general intelligence are probably false
same extent due to imperfect linkage disequilibrium, they cancel positives. Psychol Sci 2012; 23: 1314–1323.
each other out in the calculation of the genetic correlation, leaving 16 Maher B. Personal genomes: the case of the missing heritability. Nature 2008; 456:
an unbiased estimate of the genetic correlation. 18–21.
17 Plomin R. Child development and molecular genetics: 14 years later. Child Dev
2012; 30: 1467–8624.
Implications for brain structure and function 18 Yang JA, Lee SH, Goddard ME, Visscher PM. GCTA: a tool for genome-wide
complex trait analysis. Am J Hum Genet 2011; 88: 76–82.
To the extent that g indexes general brain function, the present 19 Yang J, Benyamin B, McEvoy BP, Gordon S, Henders AK, Nyholt DR et al. Common
results suggest hypotheses for the etiology of individual SNPs explain a large proportion of the heritability for human height. Nat Genet
differences in brain development. The same genes can be 2010; 42: 565–569.
expected to be responsible for individual differences throughout 20 Yang J, Manolio TA, Pasquale LR, Boerwinkle E, Caporaso N, Cunningham JM et al.
brain development despite the major mean changes that occur Genome partitioning of genetic variation for complex traits using common SNPs.
during development. The hypothesis of increasing heritability for Nat Genet 2011; 43: 519–525.
individual differences in brain development points to genotype  21 Lee SH, Wray NR, Goddard ME, Visscher PM. Estimating missing heritability for
environment correlation as the process by which genotypes disease from genome-wide association studies. Am J Hum Genet 2011; 88: 294–305.
become phenotypes. Importantly, the correspondence between 22 Lee SH, DeCandia TR, Ripke S, Yang J, Sullivan PF, Goddard ME et al. Estimating
the proportion of variation in susceptibility to schizophrenia captured by com-
GCTA and twin results indicates that special samples such as twins
mon SNPs. Nat Genet 2012; 44: 247–250.
are no longer needed to test such genetic hypotheses in 23 Lubke GH, Hottenga JJ, Walters R, Laurin C, de Geus EJ, Willemsen G et al. Esti-
neurodevelopment—GCTA makes it possible to test them in any mating the genetic variance of major depressive disorder due to all single
large sample of unrelated individuals. nucleotide polymorphisms. Biol Psychiatry 2012; 72: 707–709.
24 Vinkhuyzen AAE, Pedersen NL, Yang J, Lee SH, Magnusson PKE, Iacono WG et al.
Common SNPs explain some of the variation in the personality dimensions of
CONFLICT OF INTEREST neuroticism and extraversion. Transl Psychiatry 2012; 2: e102.
25 Benjamin DJ, Cesarini D, van der Loos MJHM, Dawes CT, Koellinger PD, Mag-
The authors declare no conflict of interest.
nusson PKE et al. The genetic architecture of economic and political preferences.
Proc Natl Acad Sci USA 2012; 109: 8026–8031.
26 Davies G, Tenesa A, Payton A, Yang J, Harris SE, Liewald D et al. Genome-wide
ACKNOWLEDGEMENTS association studies establish that human intelligence is highly heritable and
TEDS is supported by a program grant to RP from the UK Medical Research Council polygenic. Mol Psychiatry 2011; 16: 996–1005.
(G0901245, and previously G0500079), with additional support from the US National 27 Deary IJ, Yang J, Davies G, Harris SE, Tenesa A, Liewald D et al. Genetic con-
Institutes of Health (HD044454; HD046167). Genome-wide genotyping was made tributions to stability and change in intelligence from childhood to old age.
possible by a grant from the Wellcome Trust to the Wellcome Trust Case Control Nature 2012; 482: 212–215.
Consortium 2 project (085475/B/08/Z; 085475/Z/08/Z). RP is supported by a Medical 28 Plomin R, Haworth CMA, Meaburn EL, Price T. Wellcome Trust Case Control
Research Council Research Professorship award (G19/2) and a European Advanced Consortium 2, Davis OSP. Common DNA markers can account for more than half
Investigator award (295366); MT is supported by a Medical Research Council of the genetic influence on cognitive abilities. Psychol Sci 2001; 178: 41–48.
studentship. 29 Lee SH, Yang J, Goddard ME, Visscher PM, Wray NR. Estimation of pleiotropy
between complex diseases using SNP-derived genomic relationships and
restricted maximum likelihood. Bioinformatics 2012; 28: 2540–2542.
30 Haworth CMA, Davis OSP, Plomin R. Twins Early Development Study (TEDS): a
REFERENCES genetically sensitive investigation of cognitive and behavioural development
1 Blakemore SJ. Imaging brain development: the adolescent brain. Neuroimage from childhood to young adulthood. Twin Res Hum Genet 2012; FirstView: 1–9.
2012; 61: 397–406. 31 Oliver BR, Plomin R. Twins Early Development Study (TEDS): a multivariate, longi-
2 Power JD, Fair DA, Schlaggar BL, Petersen SE. The development of human func- tudinal genetic investigation of language, cognition and behaviour problems
tional brain networks. Neuron 2010; 67: 735–748. from childhood through adolescence. Twin Res Hum Genet 2007; 10: 96–105.
3 Giedd JN, Rapoport JL. Structural MRI of pediatric brain development: what have 32 Kovas Y, Haworth CMA, Dale PS, Plomin R. The genetic and environmental origins
we learned and where are we going? Neuron 2010; 67: 728–734. of learning abilities and disabilities in the early school years. Monogr Soc Res Child
4 Plomin R, Kosslyn S. Genes brain and cognition. Nat Neurosci 2001; 4: 1153–1155. Dev 2007; 72: 1–144.
5 Spearman C. ‘General intelligence’, objectively determined and measured. Am J 33 Howie BN, Donnelly P, Marchini J. A flexible and accurate genotype imputation
Psychol 1904; 15: 201–292. method for the next generation of genome-wide association studies. PLoS Genet
6 Deary IJ. Intelligence. Ann Rev Psychol 2012; 63: 453–482. 2009; 5: e1000529.

& 2014 Macmillan Publishers Limited Molecular Psychiatry (2014), 380 – 384
 Genetic stability but increasing heritability
M Trzaskowski et al
384
34 Fellay J, Shianna KV, Ge D, Colombo S, Ledergerber B, Weale M et al. A whole- 42 Davis O, Arden R, Plomin R. g in middle childhood: moderate genetic and
genome association study of major determinants for host control of HIV-1. Science shared environmental influence using diverse measures of general cognitive
2007; 317: 944–947. ability at 7, 9 and 10 years in a large population sample of twins. Intelligence 2008;
35 Patterson N, Price AL, Reich D. Population structure and eigenanalysis. PLoS Genet 36: 68–80.
2006; 2: 2074–2093. 43 Hanscombe KB, Haworth CMA, Davis OSP, Jaffee SR, Plomin R. The nature (and
36 Petrill SA, Rempell J, Oliver B, Plomin R. Testing cognitive abilities by telephone in nurture) of children’s perceptions of family chaos. Learn Individ Differ 2010; 20:
a sample of 6-to 8-year olds. Intelligence 2002; 30: 353–360. 549–553.
37 Haworth CMA, Harlaar N, Kovas Y, Davis OS, Oliver BR, Hayiou-Thomas ME et al. 44 Kendler KS, Baker JH. Genetic influences on measures of the environment: a
Internet cognitive testing of large samples needed in genetic research. Twin Res systematic review. Psychol Med 2007; 37: 615–626.
Hum Genet 2007; 10: 554–563. 45 Hanscombe KB, Haworth CMA, Davis OSP, Jaffee SR, Plomin R. Chaotic
38 van der Waerden BL. On the sources of my book "Moderne Algebra". Hist Math homes and school achievement: a twin study. J Child Psychol Psychiatry 2011; 52:
1975; 2: 31–40. 1212–1220.
39 Lehmann EL. Nonparametrics: Statistical Methods Based on Ranks. Holden-Day: San 46 Visscher PM, Brown MA, McCarthy MI, Yang J. Five years of GWAS discovery. Am J
Francisco, 1975. Hum Genet 2012; 90: 7–24.
40 Team RDC. R: A language and environment for statistical computing (Version
2.13.1). 2011, Available from http://www.R-project.org/.
41 Boker S, Neale M, Maes H, Wilde M, Spiegel M, Brick T et al. OpenMx: an open This work is licensed under a Creative Commons Attribution-
source extended structural equation modeling framework. Psychometrika 2011; NonCommercial-NoDerivs 3.0 Unported License. To view a copy of
76: 306–317. this license, visit http://creativecommons.org/licenses/by-nc-nd/3.0/

Molecular Psychiatry (2014), 380 – 384 & 2014 Macmillan Publishers Limited