Robert Plomin - The Genetics of Cognitive Abilities

People differ greatly in all aspectsof what is casually known as in-telligence. The differences are

apparent not only in school, from kin-dergarten to college, but also in the mostordinary circumstances: in the wordspeople use and comprehend, in theirdiffering abilities to read a map or fol-low directions, or in their capacities forremembering telephone numbers orfiguring change. The variations in thesespecific skills are so common that theyare often taken for granted. Yet whatmakes people so different?

It would be reasonable to think thatthe environment is the source of differ-

ences in cognitive skills—that we arewhat we learn. It is clear, for example,that human beings are not born with afull vocabulary; they have to learnwords. Hence, learning must be themechanism by which differences in vo-cabulary arise among individuals. Anddifferences in experience—say, in theextent to which parents model and en-courage vocabulary skills or in the qual-ity of language training provided byschools—must be responsible for indi-vidual differences in learning.

Earlier in this century psychology wasin fact dominated by environmental ex-planations for variance in cognitive abil-

ities. More recently, however, most psy-chologists have begun to embrace amore balanced view: one in which na-ture and nurture interact in cognitivedevelopment. During the past fewdecades, studies in genetics have point-ed to a substantial role for heredity inmolding the components of intellect,and researchers have even begun totrack down the genes involved in cogni-tive function. These findings do not re-fute the notion that environmental fac-tors shape the learning process. Insteadthey suggest that differences in people’sgenes affect how easily they learn.

Just how much do genes and envi-

The Genetics of Cognitive Abilities and Disabilities62 Scientific American May 1998

The Genetics of CognitiveAbilities and DisabilitiesInvestigations of specific cognitive skills can help clarify

how genes shape the components of intellect

by Robert Plomin and John C. DeFries

ronment matter for specific cognitiveabilities such as vocabulary? That is thequestion we have set out to answer.Our tool of study is quantitative genet-ics, a statistical approach that exploresthe causes of variations in traits amongindividuals. Studies comparing the per-formance of twins and adopted chil-dren on certain tests of cognitive skills,for example, can assess the relative con-tributions of nature and nurture.

In reviewing several decades of suchstudies and conducting our own, wehave begun to clarify the relations amongspecialized aspects of intellect, such asverbal and spatial reasoning, as well asthe relations between normal cognitivefunction and disabilities, such as dys-lexia. With the help of molecular genet-ics, we and other investigators have alsobegun to identify the genes that affectthese specific abilities and disabilities.Eventually, we believe, knowledge ofthese genes will help reveal the biochem-ical mechanisms involved in human in-telligence. And with the insight gainedfrom genetics, researchers may somedaydevelop environmental interventionsthat will lessen or prevent the effects ofcognitive disorders.

Some people find the idea of a geneticrole in intelligence alarming or, at thevery least, confusing. It is important tounderstand from the outset, then, what

exactly geneticists mean when they talkabout genetic influence. The term typi-cally used is “heritability”: a statisticalmeasure of the genetic contribution todifferences among individuals.

Verbal and Spatial Abilities

Heritability tells us what proportionof individual differences in a pop-

ulation—known as variance—can be as-cribed to genes. If we say, for example,that a trait is 50 percent heritable, weare in effect saying that half of the vari-ance in that trait is linked to heredity.Heritability, then, is a way of explainingwhat makes people different, not whatconstitutes a given individual’s intelli-gence. In general, however, if heritabili-ty for a trait is high, the influence ofgenes on the trait in individuals wouldbe strong as well.

Attempts to estimate the heritabilityof specific cognitive abilities began withfamily studies. Analyses of similaritiesbetween parents and their children and

between siblings have shown that cog-nitive abilities run in families. Results ofthe largest family study done on specificcognitive abilities, which was conduct-ed in Hawaii in the 1970s, helped toquantify this resemblance.

The Hawaii Family Study of Cogni-tion was a collaborative project betweenresearchers at the University of Colora-do at Boulder and the University of Ha-waii and involved more than 1,000 fam-ilies and sibling pairs. The study deter-mined correlations (a statistical measureof resemblance) between relatives ontests of verbal and spatial ability. A cor-relation of 1.0 would mean that thescores of family members were identi-cal; a correlation of zero would indicatethat the scores were no more similarthan those of two people picked at ran-dom. Because children on average sharehalf their genes with each parent andwith siblings, the highest correlation intest scores that could be expected ongenetic grounds alone would be 0.5.

The Hawaii study showed that fami-

The Genetics of Cognitive Abilities and Disabilities Scientific American May 1998 63

TWINS ARE COMMON RESEARCH SUBJECTS in studies of specific cognitive abil-ities. The identical (opposite page) and fraternal (below) pairs depicted here are partic-ipants in the authors’ research. They are performing a task in a test of spatial ability,trying to reconstruct a block model with their own toy building blocks. On such tests,which are given to each child individually, the scores of identical twins (who have allthe same genes) are more similar than the scores of fraternal twins (who share abouthalf their genes)—a sign that genetic inheritance exerts an influence on spatial ability. D

AVID

KA

MPF

NER

Gam

ma

Liai

son

ly members are in factmore alike than unrelat-ed individuals on mea-sures of specific cognitiveskills. The actual correla-tions for both verbal andspatial tests were, on aver-age, about 0.25. These cor-relations alone, however, donot disclose whether cogni-tive abilities run in familiesbecause of genetics or becauseof environmental effects. Toexplore this distinction, ge-neticists rely on two “experi-ments”: twinning (an experi-ment of nature) and adoption(a social experiment).

Twin studies are the work-horse of behavioral genetics.They compare the resemblanceof identical twins, who have thesame genetic makeup, with the resem-blance of fraternal twins, who shareonly about half their genes. If cognitiveabilities are influenced by genes, identi-cal twins ought to be more alike thanfraternal twins on tests of cognitiveskills. From correlations found in thesekinds of studies, investigators can esti-mate the extent to which genes accountfor variances in the general population.Indeed, a rough estimate of heritability

can be made by doubling the differencebetween identical-twin and fraternal-twin correlations.

Adoption provides the most directway to disentangle nature and nurturein family resemblance, by creating pairsof genetically related individuals whodo not share a common family environ-ment. Correlations among these pairsenable investigators to estimate the con-tribution of genetics to family resem-blance. Adoption also produces pairs of

genetically unrelated individuals whoshare a family environment, and theircorrelations make it possible to esti-mate the contribution of shared envi-ronment to resemblance.

Twin studies of specific cognitive abil-ities over three decades and in four coun-tries have yielded remarkably consistentresults [see illustration on page 66].Correlations for identical twins greatlyexceed those for fraternal twins on testsof both verbal and spatial abilities inchildren, adolescents and adults. Re-sults of the first twin study in the elder-


TESTS OF VERBAL ABILITY

1. VOCABULARY: In each row, circle the word that means the same or

nearly the same as the underlined word. There is only one correct choice

in each line.

a. aridcoarse clever modest dry

b. piquantfruity pungent harmful upright

2. VERBAL FLUENCY: For the next three minutes, write as many words as

you can that start with F and end with M.

3. CATEGORIES: For the next three minutes, list all the things you can

think of that are FLAT.

Since the dawn of psychology, experts have disagreed aboutthe fundamental nature of intelligence. Some have claimed

that intelligence is an inherent faculty prescribed by heredity,whereas others have emphasized the effects of education andupbringing. Some have portrayed intelligence as a global qualitythat permeates all facets of cognition; others believe the intellectconsists of discrete, specialized abilities—such as artistic talentor a flair for mathematics—that share no common principle.

In the past few decades, genetic studies have convinced mostpsychologists that heredity exerts considerable influence on in-telligence. In fact, research suggests that as much as half of thevariation in intelligence among individuals may be attributed togenetic factors.

And most psychologists have also come to accept a globalconceptualization of intelligence. Termed general cognitive abil-ity, or “g,” this global quality is reflected in the apparent overlapamong specific cognitive skills. As Robert Plomin and John C. De-Fries point out, people who do well on tests of one type of cog-nitive skill also tend to do well on tests of other cognitive abil-ities. Indeed, this intercorrelation has provided the rationale forIQ (intelligence quotient) tests, which yield a single score fromcombined assessments of specific cognitive skills.

Because specific and general cognitive abilities are related inthis manner, it is not surprising that many of the findings regard-ing specific abilities echo what is already known about generalability. The heritabilities found in studies of specific cognitiveabilities, for example, are comparable with the heritability deter-mined for g. The developmental trend described by the au-thors—in which genetic influence on specific cognitive abilitiesseems to increase throughout childhood, reaching adult levelsby the mid-teens—is also familiar to researchers of general cog-nitive ability.

And because measures of g are derived from intercorrelationsof verbal and spatial abilities, a gene that is linked with boththose traits is almost guaranteed to have some role in generalcognitive ability as well—and vice versa. This month in the jour-nal Psychological Science, Plomin and various collaborators re-port the discovery of the first gene associated with general cog-nitive ability. Although the finding should further understandingof the nature of cognition, it is also likely to reignite debate. In-deed, intelligence research may be one realm where under-standing does little to quell disagreement.

KAREN WRIGHT is a freelance writer living in New Hampshire.

How Do Cognitive Abilities Relate to General Intelligence?

by Karen Wright

JEN

NIF

ER C

. CH

RIST

IAN

SEN

ly—reported last year by Ger-ald E. McClearn and his col-leagues at Pennsylvania StateUniversity and by Stig Bergand his associates at the In-stitute for Gerontology inJönköping, Sweden—showthat the resemblances be-tween identical and frater-nal twins persist even intoold age. Although geron-tologists have assumedthat genetic differencesbecome less important asexperiences accumulateover a lifetime, researchon cognitive abilitieshas so far demonstrat-ed otherwise. Calcula-tions based on thecombined findings inthese studies implythat in the generalpopulation, geneticsaccounts for about60 percent of thevariance in verbalability and about 50 percentof the variance in spatial ability.

Investigations involving adoptees haveyielded similar results. Two recent stud-ies of twins reared apart—one by Thom-as J. Bouchard, Jr., Matthew McGueand their colleagues at the University ofMinnesota, the other an internationalcollaboration headed by Nancy L. Ped-ersen at the Karolinska Institute inStockholm—have implied heritabilitiesof about 50 percent for both verbal andspatial abilities.

In our own Colorado Adoption Proj-ect, which we launched in 1975, wehave used the power of adoption stud-ies to further characterize the roles ofgenes and environment, to assess devel-opmental trends in cognitive abilities andto explore the extent to which specificcognitive skills are related to one anoth-er. The ongoing project compares thecorrelations between more than 200adopted children and their birth andadoptive parents with the correlationsfor a control group of children raised by

their biological parents [see illustrationon page 67].

These data provide some surprisinginsights. By middle childhood, for ex-ample, birth mothers and their childrenwho were adopted by others are just assimilar as control parents and their chil-dren on measures of both verbal andspatial ability. In contrast, the scores ofadopted children do not resemble thoseof their adoptive parents at all. Theseresults join a growing body of evidencesuggesting that the common family en-vironment generally does not contrib-ute to similarities in family members.Rather family resemblance on suchmeasures seems to be controlled almostentirely by genetics, and environmentalfactors often end up making familymembers different, not the same.

The Colorado data also reveal an in-teresting developmental trend. It appearsthat genetic influence increases during

childhood, so that by the mid-teens,heritability reaches a level comparablewith that seen in adults. In correlationsof verbal ability, for example, resem-blance between birth parents and theirchildren who were adopted by othersincreases from about 0.1 at age three toabout 0.3 at age 16. A similar pattern isevident in tests of spatial ability. Somegenetically driven transformation incognitive function seems to take placein the early school years, around ageseven. The results indicate that by thetime people reach age 16, genetic fac-tors account for 50 percent of the vari-ance for verbal ability and 40 percentfor spatial ability—numbers not unlikethose derived from twin studies ofspecific cognitive abilities.

The Colorado Adoption Project andother investigations have also helped


TESTS OF SPECIFIC ABILITIESadministered to adolescents andadults include tasks resembling theones listed here. The tests gaugeeach cognitive ability in severalways, and multiple tests are com-bined to provide a reliable mea-sure of each skill. (Answers ap-pear on page 69.)

TESTS OF SPATIAL ABILITY1. IMAGINARY CUTTING: Draw a line or lines showing where the figure on the

left should be cut to form the pieces on the right. There may be more than

one way to draw the lines correctly.

2. MENTAL ROTATIONS: Circle the two objects on the right that are the same

as the object on the left.

3. CARD ROTATIONS: Circle the figures on the right that can be rotated

(without being lifted off the page) to exactly match the one on the left.

4. HIDDEN PATTERNS: Circle each pattern below in which the figure

appears. The figure must always be in this position, not upside down or on

its side.

ab

cd

ab

cd

e

ab

cd

ef

g

JEN

NIF

ER C

. CH

RIST

IAN

SEN

clarify the differences and similaritiesamong cognitive abilities. Current cog-nitive neuroscience assumes a modularmodel of intelligence, in which differentcognitive processes are isolated anatom-ically in discrete modules in the brain.The modular model implies that specificcognitive abilities are also genetically dis-tinct—that genetic effects on verbal abil-ity, say, should not overlap substantial-ly with genetic effects on spatial ability.

Psychologists, however, have long rec-ognized that most specialized cognitiveskills, including verbal and spatial abili-ties, intercorrelate moderately. That is,people who perform well on one type oftest also tend to do well on other types.Correlations between verbal and spatialabilities, for example, are usually about0.5. Such intercorrelation implies a po-tential genetic link.

From Abilities to Achievement

Genetic studies of specific cognitiveabilities also fail to support the

modular model. Instead it seems thatgenes are responsible for most of theoverlap between cognitive skills. Analy-sis of the Colorado project data, for ex-ample, indicates that genetics governs70 percent of the correlation betweenverbal and spatial ability. Similar resultshave been found in twin studies in child-

hood, young adulthood andmiddle age. Thus, there is agood chance that when genesassociated with a particularcognitive ability are identi-fied, the same genes will beassociated with other cogni-tive abilities.

Research into schoolachievement has hinted thatthe genes associated with cog-nitive abilities may also berelevant to academic perfor-mance. Studies of more than2,000 pairs of high school– age twins were done in the1970s by John C. Loehlin ofthe University of Texas atAustin and Robert C. Nich-ols, then at the NationalMerit Scholarship Corporation inEvanston, Ill. In these studies the scoresof identical twins were consistently andsubstantially more similar than those offraternal twins on all four domains ofthe National Merit Scholarship Quali-fying Test: English usage, mathematics,social studies and natural sciences.These results suggest that genetic fac-tors account for about 40 percent ofthe variation on such achievement tests.

Genetic influence on school achieve-ment has also been found in twin stud-ies of elementary school–age children as

well as in our work with the ColoradoAdoption Project. It appears that genesmay have almost as much effect onschool achievement as they do on cog-nitive abilities. These results are surpris-ing in and of themselves, as educatorshave long believed that achievement ismore a product of effort than of ability.Even more interesting, then, is the find-ing from twin studies and our adoptionproject that genetic effects overlap be-tween different categories of achieve-ment and that these overlapping genesare probably the very same genetic fac-


0.0

0.2

0.4

0.6

0.8

1.0

0.0

0.2

0.4

0.6

0.8

1.0

CHILDHOOD ADOLESCENCE YOUNG ADULTHOOD MIDDLE AGE OLD AGE

CHILDHOOD ADOLESCENCE YOUNG ADULTHOOD MIDDLE AGE OLD AGE

VERBAL ABILITY

SPATIAL ABILITY

TWIN

CO

RREL

ATIO

N(IN

CRE

ASI

NG

SIM

ILA

RITY

)TW

IN C

ORR

ELAT

ION

(INC

REA

SIN

G S

IMIL

ARI

TY)

IDENTICAL TWINS FRATERNAL TWINS

TWIN STUDIES have examined correlations in verbal (top) andin spatial (bottom) skills of identical twins and of fraternaltwins. When the results of the separate studies are put side byside, they demonstrate a substantial genetic influence on specific

cognitive abilities from childhood to old age; for all age groups,the scores of identical twins are more alike than those of frater-nal twins. These data seem to counter the long-standing notionthat the influence of genes wanes with time.

The implications of heritability data are commonlymisunderstood. As the main text indicates, heri-

tability is a statistical measure, expressed as a percent-age, describing the extent to which genetic factors con-tribute to variations on a given trait among the membersof a population.

The fact that genes influence a trait does not mean,however, that “biology is destiny.” Indeed, genetics re-search has helped confirm the significance of environ-mental factors, which generally account for as much vari-ance in human behavior as genes do. If intelligence is 50percent heritable, then environmental factors must bejust as important as genes in generating differencesamong people.

What Heritability Means

tors that can influence cognitive abilities. This evidence supports a decidedly

nonmodular view of intelligence as apervasive or global quality of the mindand underscores the relevance of cogni-tive abilities in real-world performance.It also implies that genes for cognitiveabilities are likely to be genes involvedin school achievement, and vice versa.

Given the evidence for genetic influ-ence on cognitive abilities and achieve-ment, one might suppose that cognitivedisabilities and poor academic achieve-ment must also show genetic influence.

But even if genes are in-volved in cognitive disor-ders, they may not be thesame genes that influencenormal cognitive function.The example of mental re-tardation illustrates thispoint. Mild mental retarda-tion runs in families, but se-vere retardation does not. In-stead severe mental retarda-tion is caused by genetic andenvironmental factors—nov-el mutations, birth complica-tions and head injuries, toname a few—that do notcome into play in the normalrange of intelligence.

Researchers need to assess,rather than assume, genetic

links between the normal and the ab-normal, between the traits that are partof a continuum and true disorders ofhuman cognition. Yet genetic studies ofverbal and spatial disabilities have beenfew and far between.

Genetics and Disability

Most such research has focused onreading disability, which afflicts

80 percent of children diagnosed with alearning disorder. Children with read-ing disability, also known as dyslexia,

read slowly, show poor comprehensionand have trouble reading aloud [see“Dyslexia,” by Sally E. Shaywitz, Sci-entific American, November 1996].Studies by one of us (DeFries) haveshown that reading disability runs infamilies and that genetic factors do in-deed contribute to the resemblanceamong family members. The identicaltwin of a person diagnosed with read-ing disability, for example, has a 68percent risk of being similarly diag-nosed, whereas a fraternal twin hasonly a 38 percent chance.

Is this genetic effect related in anyway to the genes associated with nor-mal variation in reading ability? Thatquestion presents some methodologicalchallenges. The concept of a cognitivedisorder is inherently problematic, be-cause it treats disability qualitatively—

you either have it or you don’t—ratherthan describing the degree of disabilityin a quantitative fashion. This focuscreates an analytical gap between disor-ders and traits that are dimensional(varying along a continuum), which areby definition quantitative.

During the past decade, a new genet-ic technique has been developed thatbridges the gap between dimensionsand disorders by collecting quantitativeinformation about the relatives of sub-jects diagnosed qualitatively with a dis-


–0.05

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

–0.05

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

CHILDRENAND THEIR BIRTH PARENTS

ADOPTED CHILDRENAND THEIR BIRTH PARENTS

ADOPTED CHILDREN AND THEIR ADOPTIVE PARENTS

3 YEARS 4 YEARS 7 YEARS 9 YEARS 10 YEARS 12 YEARS 14 YEARS 16 YEARS

3 YEARS 4 YEARS 7 YEARS 9 YEARS 10 YEARS 12 YEARS 14 YEARS 16 YEARS

CHILD’S AGE

VERBAL ABILITY

SPATIAL ABILITY

CH

ILD

-PA

REN

T C

ORR

ELAT

ION

(INC

REA

SIN

G S

IMIL

ARI

TY)

CH

ILD

-PA

REN

T C

ORR

ELAT

ION

(INC

REA

SIN

G S

IMIL

ARI

TY)

COLORADO ADOPTION PROJECT, which followed subjectsover time, finds that for both verbal (top) and spatial (bottom)abilities, adopted children come to resemble their birth parents(white bars) as much as children raised by their birth parents do

(gray bars). In contrast, adopted children do not end up resem-bling their adoptive parents (black bars). The results imply thatmost of the family resemblance in cognitive skills is caused bygenetic factors, not environment.

Moreover, even when genetic factors have an especial-ly powerful effect, as in some kinds of mental retardation,environmental interventions can often fully or partlyovercome the genetic “determinants.” For example, thedevastating effects of phenylketonuria, a genetic diseasethat can cause mental retardation, can often be nullifiedby dietary intervention.

Finally, the degree of heritability for a given trait is notset in stone. The relative influence of genes and environ-ment can change. If, for instance, environmental factorswere made almost identical for all the members of a hy-pothetical population, any differences in cognitive abilityin that population would then have to be attributed togenetics, and heritability would be closer to 100 percentthan to 50 percent. Heritability describes what is, ratherthan what can (or should) be. —R.P. and J.C.D.

JEN

NIF

ER C

. CH

RIST

IAN

SEN

ability. The method is calledDF extremes analysis, after itscreators, DeFries and David W.Fulker, a colleague at the Uni-versity of Colorado’s Institutefor Behavioral Genetics.

For reading disability, theanalysis works by testing theidentical and fraternal twins ofreading-disabled subjects onquantitative measures of read-ing, rather than looking for ashared diagnosis of dyslexia[see illustration at right]. Ifreading disability is influencedby genes that also affect varia-tion within the normal rangeof reading performance, thenthe reading scores of the identi-cal twins of dyslexic childrenshould be closer to those of thereading-disabled group than thescores of fraternal twins are. (Asingle gene can exert differenteffects if it occurs in more thanone form in a population, sothat two people may inheritsomewhat different versions.The genes controlling eye colorand height are examples ofsuch variable genes.)

It turns out that, as a group,identical twins of reading-dis-abled subjects do perform al-most as poorly as dyslexic sub-jects on these quantitative tests, where-as fraternal twins do much better thanthe reading-disabled group (though stillsignificantly worse than the rest of thepopulation). Hence, the genes involvedin reading disability may in fact be thesame as those that contribute to thequantitative dimension of reading abili-ty measured in this study. DF extremesanalysis of these data further suggeststhat about half the difference in readingscores between dyslexics and the generalpopulation can be explained by genetics.

For reading disability, then, therecould well be a genetic link between thenormal and the abnormal, even thoughsuch links may not be found universallyfor other disabilities. It is possible thatreading disability represents the extremeend of a continuum of reading ability,rather than a distinct disorder—that dys-lexia might be quantitatively rather thanqualitatively different from the normalrange of reading ability. All this sug-gests that if a gene is found for readingdisability, the same gene is likely to beassociated with the normal range ofvariation in reading ability. The defini-

tive test will come when a specific geneis identified that is associated with ei-ther reading ability or disability. In fact,we and other investigators are alreadyvery close to finding such a gene.

The Hunt for Genes

Until now, we have confined our dis-cussion to quantitative genetics, a

discipline that measures the heritabilityof traits without regard to the kind andnumber of genes involved. For informa-tion about the genes themselves, re-searchers must turn to molecular genet-ics—and increasingly, they do. If scien-tists can identify the genes involved inbehavior and characterize the proteinsthat the genes code for, new interven-tions for disabilities become possible.

Research in mice and fruit flies hassucceeded in identifying single genes re-lated to learning and spatial perception,and investigations of naturally occur-ring variations in human populationshave found mutations in single genesthat result in general mental retardation.These include the genes for phenylketo-

nuria and fragile X syndrome,both causes of mental retarda-tion. Single-gene defects thatare associated with Duchenne’smuscular dystrophy, Lesch-Ny-han syndrome, neurofibroma-tosis type 1 and Williams syn-drome may also be linked tothe specific cognitive disabili-ties seen in these disorders [see“Williams Syndrome and theBrain,” by Howard M. Len-hoff, Paul P. Wang, FrankGreenberg and Ursula Bellugi;Scientific American, Decem-ber 1997].

In fact, more than 100 sin-gle-gene mutations are knownto impair cognitive develop-ment. Normal cognitive func-tioning, on the other hand, isalmost certainly orchestrated bymany subtly acting genes work-ing together, rather than by sin-gle genes operating in isolation.These collaborative genes arethought to affect cognition in aprobabilistic rather than a de-terministic manner and arecalled quantitative trait loci, orQTLs. The name, which appliesto genes involved in a complexdimension such as cognition,emphasizes the quantitativenature of certain physical and

behavioral traits. QTLs have alreadybeen identified for diseases such as dia-betes, obesity and hypertension as wellas for behavioral problems involvingdrug sensitivity and dependence.

But finding QTLs is much more diffi-cult than identifying the single-gene mu-tations responsible for some cognitivedisorders. Fulker addressed this prob-lem by developing a method, similar toDF extremes analysis, in which certainknown variations in DNA are correlat-ed with sibling differences in quantita-tive traits. Because genetic effects are eas-ier to detect at the extremes of a dimen-sion, the method works best when atleast one member of each sibling pair isknown to be extreme for a trait. Inves-tigators affiliated with the ColoradoLearning Disabilities Research Centerat the University of Colorado first usedthis technique, called QTL linkage, totry to locate a QTL for reading disabili-ty—and succeeded. The discovery wasreported in 1994 by collaborators atBoulder, the University of Denver andBoys Town National Research Hospitalin Omaha.


NU

MB

ERO

F IN

DIV

IDU

ALS

HIGHLOW

READINGDISABLED

MEAN

MEAN

READING SCORES

RANDOMLYSELECTED

TWINS

IDENTICAL TWINSOF READING-DISABLED

INDIVIDUALS

FRATERNAL TWINSOF READING-DISABLED

INDIVIDUALS

NU

MB

ERO

F IN

DIV

IDU

ALS

NU

MB

ERO

F IN

DIV

IDU

ALS

MEAN

READING SCORES of twins suggest a possible genetic linkbetween normal and abnormal reading skills. In a group ofrandomly selected members of twin pairs (top), a smallfraction of children were reading disabled (blue). Identical(middle) and fraternal (bottom) twins of the reading-dis-abled children scored lower than the randomly selectedgroup, with the identical twins performing worse than thefraternal ones. Genetic factors, then, are involved in readingdisability. The same genes that influence reading disabilitymay underlie differences in normal reading ability.

JEN

NIF

ER C

. CH

RIST

IAN

SEN

Like many techniques in moleculargenetics, QTL linkage works by iden-tifying differences in DNA markers:stretches of DNA that are known tooccupy particular sites on chromo-somes and that can vary somewhatfrom person to person. The differentversions of a marker, like the differentversions of a gene, are called alleles.Because people have two copies of allchromosomes (except for the gender-determining X and Y chromosomes inmales), they have two alleles for anygiven DNA marker. Hence, siblingscan share one, two or no alleles of amarker. In other words, for eachmarker, siblings can either be likeidentical twins (sharing both alleles),like fraternal twins (sharing half theiralleles) or like adoptive siblings (shar-ing no alleles).

The investigators who found theQTL for reading disability identified areading-disabled member of a twinpair and then obtained reading scoresfor the other twin—the “co-twin.” Ifthe reading scores of the co-twins wereworse when they shared alleles of aparticular marker with their reading-disabled twins, then that marker waslikely to lie near a QTL for readingdisability in the same chromosomalregion. The researchers found such amarker on the short arm of chromo-some 6 in two independent samples,one of fraternal twins and one of non-twin siblings. The findings have sincebeen replicated by others.

It is important to note that whereasthese studies have helped point to thelocation of a gene (or genes) implicatedin reading disability, the gene (or genes)has not yet been characterized. This dis-tinction gives a sense of where the ge-netics of cognition stand today: poised

on the brink of a new level of discovery.The identification of genes that influ-ence specific cognitive abilities will rev-olutionize researchers’ understandingof the mind. Indeed, molecular geneticswill have far-ranging consequences forthe study of all human behavior. Re-searchers will soon be able to investi-

gate the genetic connections betweendifferent traits and between behaviorsand biological mechanisms. They willbe able to better track the develop-mental course of genetic effects and todefine more precisely the interactionsbetween genes and the environment.

The discovery of genes for disordersand disabilities will also help clini-cians design more effective therapiesand to identify people at risk long be-fore the appearance of symptoms. Infact, this scenario is already being en-acted with an allele called Apo-E4,which is associated with dementiaand cognitive decline in the elderly. Ofcourse, new knowledge of specificgenes could turn up new problems aswell: among them, prejudicial labelingand discrimination. And genetics re-search always raises fears that DNAmarkers will be used by parents pre-natally to select “designer babies.”

We cannot emphasize too much thatgenetic effects do not imply genetic de-terminism, nor do they constrain envi-ronmental interventions. Although somereaders may find our views to be contro-versial, we believe the benefits of iden-tifying genes for cognitive dimensionsand disorders will far outweigh the po-tential abuses.


Further Reading

Nature, Nurture and Psychology. Edited by Robert Plomin andGerald E. McClearn. American Psychological Association, Washing-ton, D.C., 1993.

Genetics of Specific Reading Disability. J. C. DeFries and Mari-cela Alarcón in Mental Retardation and Developmental DisabilitiesResearch Reviews, Vol. 2, pages 39–47; 1996.

Behavioral Genetics. Third edition. Robert Plomin, John C. De-Fries, Gerald E. McClearn and Michael Rutter. W. H. Freeman, 1997.

Susceptibility Loci for Distinct Components of DevelopmentalDyslexia on Chromosomes 6 and 15. E. L. Grigorenko, F. B. Wood,M. S. Meyer, L. A. Hart, W. C. Speed, A. Schuster and D. L. Pauls inAmerican Journal of Human Genetics, Vol. 60, pages 27–39; 1997.

NU

MBE

R O

F IN

DIV

IDU

ALS

HIGHLOW READING SCORES

SINGLE-GENE MODEL

READINGDISABLED

HIGHLOW READING SCORES

NU

MBE

R O

F IN

DIV

IDU

ALS

Indi

vidu

als w

ithou

t

perfo

rman

ce-

low

erin

g al

lele

s

QUANTITATIVE-TRAIT MODEL

Single-gene variant causes disability

Variants of multiple reading-related genes combine to lower scores

Indi

vidu

als w

ith o

nly

som

e pe

rform

ance

-

low

erin

g al

lele

s

Indi

vidu

als w

ith

man

y pe

rform

ance

-

low

erin

g al

lele

s

Individuals withdisability-related allele

Individuals withoutdisability-related allele

READINGDISABLED

TWO MODELS illustrate how geneticsmay affect reading disability. In the clas-sic view (top), a single variant, or allele,of a gene is able to cause the disorder;everyone who has that allele becomesreading disabled (graph). But evidencepoints to a different model (bottom), inwhich a single allele cannot produce thedisability on its own. Instead variants ofmultiple genes each act subtly but cancombine to lower scores and increasethe risk of disability.

The Authors

ROBERT PLOMIN and JOHN C. DEFRIES have collaboratedfor more than 20 years. Plomin, who worked with DeFries at theUniversity of Colorado at Boulder from 1974 to 1986, is now atthe Institute of Psychiatry in London. There he is research profes-sor of behavioral genetics and deputy director of the Social, Ge-netic and Developmental Psychiatry Research Center. DeFries di-rects the University of Colorado’s Institute for Behavioral Genet-ics and the university’s Colorado Learning Disabilities ResearchCenter. The ongoing Colorado Adoption Project, launched by theauthors in 1975, has so far produced three books and more than100 research papers. Plomin and DeFries are also the lead authorsof the textbook Behavioral Genetics, now in its third edition.

TEST ANSWERS VERBAL: 1a. dry; 1b. pungent SPATIAL: 1. 2. b, c; 3. a, c, d; 4. a, b, f

SA

JEN

NIF

ER C

. CH

RIST

IAN

SEN

Robert Plomin - The Genetics of Cognitive Abilities

Documents

Transcript of Robert Plomin - The Genetics of Cognitive Abilities