J. Oates cognitive abilities in children and young peopleilsi.org/mexico/wp-content/uploads/sites/29/2016/09/Nutrition-and...J. Oates Nutrition and cognition: assessing cognitive abilities

Introduction

Our aim in this paper is to provide information andguidance about cognitive testing to individuals,especially those working in the field of nutrition,who do not have a background in psychology orcognitive neuroscience. It may be that they wish toconduct studies to determine the effects of somenutritional intervention or they may simply want tobe able to better interpret research reports or sci-entific articles. Our interest here is in the assess-ment of children; similar information for adultsmay be found in Schmitt et al. [34]. We havechosen to put certain constraints on the informa-tion we include in order to make the paper amanageable length. We consider here tests that aresuitable for use when assessing effects of nutrition

within a generally well-fed population consuming aWestern-style diet. The reader should be aware,however, that these effects may be more subtle thanthose that would be found in children in certaindeveloping countries. Studies of malnourishedgroups with marked deficiencies in energy, proteinand some specific micronutrients may require adifferent approach and are, therefore, not consid-ered here. Similarly, we discuss the assessment ofcognition specifically, although mood and psycho-motor function, requiring different forms ofassessment, are commonly affected by nutrition.

Although the main intention is to give practicaladvice, we start with a short theoretical discussionabout some aspects of cognition that are relevant to thenutritionist, followed by a few comments about thestructure of cognition. We then describe factors relat-ing to tests that need to be considered when designing a

E. IsaacsJ. Oates

Nutrition and cognition: assessingcognitive abilities in children andyoung people

j Abstract Although studying theeffects of diet on cognitive func-tion in children is of great interestto nutritionists, many have notreceived the formal training in theprinciples and practice of assess-ment necessary to properly evalu-ate research studies or to designand conduct research themselves.This paper is aimed at such anaudience and assumes little priorknowledge of the field. A shortdescription of neural developmentin childhood is followed by adiscussion of important principlesof assessment. A level of assess-ment approach is used to present a

selection of widely used cognitivetests organized by cognitive do-main and age group. Practicalinformation about the tests ispresented in tabular form and alist of useful websites is included.

j Key words cognition –children – psychological assess-ment – nutrition

SUPPLEMENTEur J Nutr (2008) 47[Suppl 3]:4–24DOI 10.1007/s00394-008-3002-y

EJN

3002

E. IsaacsMRC Childhood Nutrition ResearchCentre, Institute of Child HealthUniversity College London Institute ofChild Health30 Guilford StreetLondon WC1N 1EH, UK

J. OatesCentre for Childhood, Development andLearning. Faculty of Education andLanguage StudiesThe Open University, Walton HallMilton Keynes MK7 6AA, UK

ILSI Europe a.i.s.b.l. (&)Avenue E. Mounier 83, Box 61200 Brussels, BelgiumFax: +32-2/762-0044E-Mail: [email protected]

study, culminating in the choice of tests to be used. Ofthe many measures available, we have space to giveexamples only and the interested researcher isencouraged to look beyond the specific testsmentionedhere, guided by principles that we will outline.

Cognition and nutrition

The cognitive structure that we measure in a child hascome about through interaction between the brainand the environment over the course of development.In the initial stages of the development of the neuralsystem in the fetus, the physical characteristics of thebrain are determined by genetic factors but, imme-diately, the environment starts to influence thisstructure. Two fetuses with identical genetic make-upbut exposed to different nutritional regimes duringgestation, for example, may start to diverge in theirneural development. After birth, the variety of envi-ronmental factors increases but nutrition continues tobe important throughout life, particularly since it isnow recognized that brain development occurs over amuch longer portion of the life-span than originallythought [35]. Indeed, environmentally linked changesin brain architecture continue throughout the life-span: the brain is a very �plastic’ organ. Nutrition is aparticularly important environmental variable toconsider in relation to brain function because it canbe manipulated relatively easily.

It is a general principle that the central nervoussystem is most vulnerable when it is developing ra-pidly. The time when nutrition has the greatest effecton brain development, therefore, is during the peri-natal period known as the ‘‘growth spurt’’, consideredto include the third prenatal trimester and the firstfew months of postnatal life in human infants [12].Within this period of rapid growth, neural eventsoccur according to a well-established timetable [8] sothat the effects of nutrition will depend, to some ex-tent, on when they take place. For example, during thefirst months of human gestation, the brain cells thatare being produced are almost all neurons whilst,after 25 weeks, glial cells predominate [20]. Otherprocesses such as myelination and synaptogenesisfollow in order but it is important to note that theprecise timing of these may differ for different regionsof the brain. Determining the potential effects ofnutrition must thus take into account the processesthat are occurring at a particular time point in dif-ferent areas of the brain. It has been suggested [15]that further spurts in brain growth in humans occurbetween 2–4, 6–8, 10–12 and 14–16 years. Thesemay also prove to be times when nutritional inter-ventions are effective; during the adolescent spurt, for

example, myelination in the frontal lobe in particularis taking place [29] and nutrition might influence thisprocess. The above all describe possible effects on themacrostructure and microstructure of brain anatomybut nutrition may also have a role to play in brainphysiology by affecting both the level and operationof various neurotransmitters. To be clear, the influ-ence of nutrition does not stop when the initialgrowth spurt ends. Neural growth and developmentcontinue throughout childhood and adolescence andneed specific nutritional resources in sufficientquantities in order to occur and reach their maximumpotential.

As well as affecting the way the brain develops,nutrition plays a crucial role in maintaining brainfunction. Greenwood et al. [18] point out that thereare at least three important ways in which diet mayaffect neurochemistry. First, the ingestion of foodaffects the availability of the precursors required forthe synthesis of neurotransmitters. Second, foodserves as the source of the vitamins and minerals thatare essential co-factors for the enzymes that synthe-size neurotransmitters. Third, dietary fats alter thecomposition of the nerve cell membrane and myelinsheath, and that, in turn, influences neuronal func-tion. Another example is glucose. Since it is the mainmetabolic fuel of the brain, the rate of glucose deliveryfrom food to the blood stream and, thereby, the gly-cemic characteristics of the carbohydrate in the diet,could influence cognitive function as well [5]. Itshould be clear that not only severe malnutrition butalso variations in the normal diet can influence neu-ronal function and hence cognition.

We have emphasized the importance of knowingthe stage of neural development when trying todetermine the effects of nutrition on the brain. Tim-ing is important in another sense as well. Any changesto the basic neural architecture brought about bynutrition are likely to be long-term. Because neuraldevelopment is a cascade of events, even a smallchange early in the sequence may have a large anddiffuse impact on future development.

It is also important to recognize that these earlyemerging differences in cognitive capacity linked withdiet will, in turn, affect individuals’ behavior andhence their selective use of the environment. Humansdo not respond passively to the environments inwhich they find themselves; cognitive processes affectchoices of environments and cognition is clearlyimplicated in dietary choices. These complexinteractions among genetic, constitutional and envi-ronmental factors are usefully considered as �trans-actional’ processes [31].

Because nutritional interventions later in life aremore likely to be short-term and specific, we might

E. Isaacs et al. 5Nutrition and cognitive abilities in children and young people

expect that early interventions would have widespreadeffects on overall cognitive level, reflected in IQ scores,while later ones would have less influence on suchmeasures. Instead, they might affect performance inspecific cognitive domains, like attention andmemory.(Of course, early interventionmay affect specific as wellasmore general domains of cognition, so it is importantto test for these). The performance–competence dis-tinction is highly relevant here; cognitive performance,reflected in behavior, is open to dietary effects at alltimes, whereas competence, more akin to underlyingability, is subject to developmental variations in sus-ceptibility to dietary variations. The effects of eatingbreakfast on cognitive performance, for example,would be expected to be short-term (although possiblyrepeated daily) andmight affect a specific function suchas attention. Whenever they occur, nutritional effectsare likely to be subtle since they will not result in grossdistortions of brain structure. Rather than destructivelesions, we would expect to see ‘‘a quantitative andqualitative alteration of growth’’ [11]. As a result, wewould not predict such marked cognitive changesassociated with nutritional variance as we would in thecase of neurological damage. This is important from anassessment point of view since many neuropsycho-logical tests have been devised to detect brain damageof a sort we are not discussing here. These testsmay notbe sensitive enough to detect nutritional effects.

How a child performs on a cognitive test will alsobe affected by factors that are considered to be non-cognitive in themselves such as mood, motivation andeffort. Performance may also be affected by the con-text in which the testing takes place: are there peerspresent? Is the adult familiar or a stranger? Is thesetting home or school? In a study, one can usuallycontrol the context factors by keeping conditions thesame for all participants. Factors such as mood areharder to control but if they are thought to varywidely among the children in the group, it may beadvisable to obtain some objective measure that canbe used in the analyses of the data. It is well estab-lished that socio-economic factors have strong effectson the development of cognitive competence, througha variety of influences including parenting, availabil-ity of resources in the home and access to educationalopportunities. Given the multiplicity of transactingfactors influencing the development of cognitiveabilities, teasing out independent effects of nutritionpresents a daunting methodological challenge andrequires careful experimental design.

The structure of cognitive functions

We use the term ‘‘cognitive functions’’ to refer to theactivities of taking in information from the environ-

ment, processing this internally and then respondingin the form of behavior. We can think of cognitivefunctions as forming a hierarchy, moving from thegeneral to the specific.

The most general level is that of overall intellectualability, usually described as intelligence. There aretwo reasons why there has been so much debatewithin psychology about the existence of a generalintelligence factor, often referred to as ‘‘g’’. First, it isan empirical fact that performance differences be-tween individuals (both children and adults) in dif-ferent cognitive domains covary; people who areverbally fluent tend also to be good at mental arith-metic, for example. This suggests the presence ofoverall or global general intelligence differences be-tween people. But these covariations are not perfect;each individual has a unique �profile’ of strengths andweaknesses in different cognitive domains, doingbetter in some areas than others. So it is not possibleto simply rank people on a general intelligence scale.Some psychologists [23] think that performance isdetermined by a general factor and a series of morespecific factors relevant to certain, but not all, tasks. Itmay be that ‘‘g’’ will turn out to describe some aspectof the nervous system that affects all cognitive per-formance, such as speed of nerve conduction, but thisfactor has not yet been described (see [9] for dis-cussion of this whole topic).

The second reason for debate is that the idea of ‘‘g’’has often been linked with a notion of geneticallydetermined differences in intelligence between indi-viduals, and, more politically contentiously, withgenetically determined differences in intelligencelevels between groups. Without going deeply intothat debate, it is worth commenting that it is nowbecoming clear that virtually all human abilitiesand psychological characteristics are influenced to adegree by genetic factors, but that environments arealso powerful influences in the transactional pro-cesses referred to above. Increasingly, gene · nutrientinteractions are being discovered. Caspi et al. [7], forexample, have recently reported that genetic variationin fatty acid metabolism moderates the effects ofbreastfeeding on cognitive development, measured byIQ scores.

It is the concept of ‘‘g’’ that has led to the devel-opment of the well-known measure ‘‘IQ’’, standing for‘‘intelligence quotient’’ and based on a measure calledmental age. Originally, intellectual level was defined asthe ratio of a person’s mental age relative to theirchronological age; the formula to calculate it is MA/CA · 100. Thus, someone whose mental age is 12 witha chronological age of 10 will have an IQ of 120.Someone whose mental age is the same as theirchronological age will have an IQ of 100. Althoughmodern tests derive IQ in a slightly different way, they

6 European Journal of Nutrition, Vol. 47, Supplement 3 (2008)

are still based on the idea of comparing an individ-ual’s performance to that of a large group with thesame chronological age. Chronological age can bemeasured in a valid and reliable way as long as weknow accurately a person’s date of birth (except in thecase of premature birth, where corrections are usuallymade for the first 2 years). However, mental age is amuch more problematic measure, since it dependsalso on knowing what the mean performance levelsare for individuals of varying ages across the popu-lation from which the person comes. It also makesassumptions about the shape of the distribution ofsuch scores from the test being used to assess IQ. Forthese reasons and because of the many other sourcesof possible variation in an individual’s test perfor-mance, like all cognitive tests, IQ is never an exactmeasure, but will always have a degree of inbuilt errorvariance.

At the next level is a series of specific cognitivedomains, including memory, language, visual andexecutive functions; all except the last are familiarterms. Although there is less than total agreementamongst psychologists, the label of ‘‘executive func-tions’’ is generally used to refer to a set of abilities thatenable people to plan, initiate and carry through goal-directed behavior in organized and ‘‘thought out’’ways. Psychologists also draw a distinction betweenhigh-level cognitive functions such as integration,synthesis, planning, and organizing, and more basic,‘‘low-level’’ cognitive functions such as processinginformation from the senses. These cognitive domainsare not mutually exclusive (memories are often storedin linguistic forms, reasoning may involve visualstimuli) but they tend to be assessed using differenttests, so it is convenient to label them separately.There are some mental processes like attentionregulation and the maintenance of goal directionthat operate across a number of other cognitivedomains.

Each cognitive domain, in turn, can be decom-posed into a series of components or sub-processesthat are important in determining the level of functionwithin that domain. Some of these refer to differentprocesses, such as the understanding versus theexpression of language or recalling information ver-sus recognizing it. Others refer to the information onwhich the process acts, auditory versus visual stimuli,for example. In order to fully characterize perfor-mance within a cognitive domain it is necessary toadminister a series of tests designed to evaluate keycomponents. One should be suspicious of studies thatclaim to have assessed a complex cognitive domain,and then only report scores on a single test. If, how-ever, an hypothesis predicted a very selective effect onthe function this test assesses, this might be accept-able; even then, however, one might want to include

another type of test from this domain to determinewhether the hypothesized function was affectedselectively. We will describe both batteries of teststhat assess a variety of aspects of a cognitive domainas well as some commonly used single tests. Finally,there are very specific tests that are designed to assessbasic components of cognition, often across domains,and usually developed in a research context in linewith a specific hypothesis or model.

As a general point, it is important to note againthat when we assess cognition in a child we aremeasuring behavior, i.e., performance, rather thancompetence. Children are born with intellectual po-tential that needs a stimulating environment to befully realized. The quality of nutrition that the childreceives will be a contributing factor to the quality ofthe environment. Hebb [19] used the terms Intelli-gence A and Intelligence B to distinguish betweenpotential and the measured level of cognitive devel-opment. Intelligence A refers to the innate potentialfor intellectual development and can never be directlyobserved. What we observe, behavior on an intelli-gence test, for example, is Intelligence B, reflecting thelevel of development at the time of testing. A simpleexample: two individuals with the same Intelligence Aare tested on vocabulary, a component of mostintelligence tests. If a word has never been heard inthe environment of one but is common in the other,the two will differ in the behavior they display on thetest (Intelligence B), although their level of Intelli-gence A is the same. Intelligence A and B are notindependent of each other, of course, since Intelli-gence A will contribute to the development of Intel-ligence B but they are not the same and it is importantto bear this in mind. We have no way, at present, ofmeasuring potential, although one can see from theabove example that measuring nonverbal perfor-mance in children thought to have had an impover-ished cultural environment, in which languageexperience was sparse, may give a better estimate oftheir overall ability. Although we have been discuss-ing intelligence, the principle that we are assessingbehavior, and not potential, applies to all aspects ofcognition.

These cognitive domains exist across the agespectrum, and three papers published in 2005 [4, 22,33] relevant to nutritionists, contain an interestingdiscussion about these same issues in adults. Childrenmay need rather different methods of assessment,however, particularly if they are pre-verbal. One dif-ference between the cognitive structure of adults andchildren is that the neural processes that underliecognition are thought to be more focally representedin adults [14]. Although there is much disagreementabout the degree of such representation in infancy,there is general consensus that, as development con-


tinues, cognitive representation becomes more mod-ular. The practical result of this is that an interventionsuch as nutrition that affects a specific neural areamay have different consequences in children at dif-ferent ages and in adults. Cognitive domains, likebrain regions, follow their own trajectories of devel-opment so that adult levels of performance areachieved at different chronological ages.

Test choice in relation to study design

The topic of designing a study is a complex one and,while not the focus of the present paper, the choice oftests to be used in a study is not independent of itsdesign, so we will mention several relevant issueshere. Crucially, researchers must be clear whetherthey are interested in global nutritional effects acrossa range of cognitive functions, or whether theirinterest is focused on a specific set or subset withinthis wide spectrum. If the interest is in global effects,then it is unlikely that a global measure such as an IQresult will have sufficient power to determine any butthe most potent effects, unless very large sample sizesare feasible and controls for factors such as socio-economic background can be implemented.

If the interest is in specific cognitive functions, andin general such more tightly hypothesis-driven stud-ies are appropriate in this field, then the researchershould be careful to ensure that the psychometricparameters of the candidate test are adequate. Whatthis means is not just that the test shows adequatereliability and discrimination power, but that it alsohas been demonstrated to be a valid measure of thefunction that it purports to assess. Such informationis available from the publishers of tests or from testauthor journal publications. Tests without a robustevidential psychometric base should be viewed withcaution. We discuss this further below.

Historically, the most common cognitive outcomemeasure in nutrition studies, at least those conductedin a medical context, has been some global measure,like IQ. Studies are starting to include more specificmeasures, however, with the realization that a globalmeasure by itself can never adequately describe cog-nition and many more specific effects of nutritionmay be missed by reliance on the IQ score or equiv-alent. Since there is not a large literature to consult inorder to determine which cognitive domains to assess,a useful approach may be to include a battery thatcovers a variety of these (see below); if effects areshown in a particular domain, subsequent studiesmay attempt to further delineate the componentprocesses that are involved. Sometimes, we have someknowledge of which neural areas are affected by thenutritional intervention and this can guide us to the

choice of cognitive domains to investigate. If thefrontal lobes are known to be involved, for example,we would predict that executive functions would beworth assessing. Unfortunately, more often than not,we do not have this information.

Where there are likely to be wide ranges of verbalabilities and cultural/ethnic backgrounds in a studysample population, it would be appropriate to con-sider using nonverbal performance tests to reduceadditional variance arising from confounding sources,since language-based tests are much more vulnerableto such confounds.

Some studies are designed to measure performanceat one time point, but others repeat the measurement,for example, before and after some period of inter-vention. This type of test–retest design is often relevantto studyingnutritional effects. If the test is to be used onmore than one occasion, it is worth considering teststhat have two ormore equivalent forms, thus helping tominimize practice effects. If no such test is appropriatefor the study, then the initial score can be used as acovariate in the statistical analysis of the data. The effectof this is to look at the second test scores while statis-tically adjusting the first scores so that these are equalacross subjects. Testing performance onmore than oneoccasion allows the determination of the trajectory ofdevelopment of a function. An interesting researchquestion is the effect of nutritional intervention oncognitive developmental trajectories in terms of char-acteristics like their shape and slope.

There is one type of design that is particularlyappropriate to nutritional studies; it is mentionedhere because it is considered to be the ‘‘gold stan-dard’’ for research in this field. This design is thedouble-blind randomized, controlled trial, usuallyreferred to as an RCT [1]. It is rooted in the phar-macological model and is probably less widely used inother areas of psychological research. Samples ofindividuals who are eligible for inclusion in the studyare randomly assigned, using some coded method, toa treatment or a control group, i.e., they receive thesupplement or a placebo. Because of the randomnature of the assignment, the assumption is made thatmany extraneous factors that might affect outcomebut are not directly measured (for example, parentalIQ, regular diet) will be the same between the twogroups; any differences between the groups will be‘‘washed out’’ by the randomization. Double-blindrefers to the fact that neither the subjects in theexperiment nor the experimenters themselves areaware of which subjects have been assigned to whichgroup. The code revealing this is broken once thestudy is completed. In some studies, the treatmentgroup is compared to a group that has been given notreatment rather than a placebo. In such cases,assignment to groups may still be random, but the


subjects know which to which arm of the trial theyhave been assigned. It is important that this infor-mation is not known to the tester since even the mostprofessional test administrator can be subtly biased intheir assessment by knowing to which arm of a trialthe subject has been assigned.

Choosing tests

As noted, we have focused on tests that could be usedwithin the context of children living in developedcountries and consuming a typical Western diet. Thelanguage in which tests are available, overwhelminglyEnglish, is also a factor to consider in a Europeancontext. It is not possible for a researcher to simplytranslate a test from one language into another andassume that they are equivalent [17]. Even a simpletask like translating a list of single words for pre-sentation in a memory task may result in importantdifferences, because words vary in their frequency ofusage and their semantic associations across lan-guages. The effect is that the individuals given thetranslated task are, in practice, being given a differenttest. This may make comparisons with studies usingthe test in its ‘‘native language’’ difficult and, at worst,may lead to misleading results. A certain amount ofdiscretion must be used. Substituting a UK Englishword for an American English word—‘‘biscuit’’ for‘‘cookie’’ or ‘‘flat’’ for ‘‘apartment’’—will probably notresult in differences in performance. The more twolanguages and cultures diverge, however, the morecaution must be exercised. It is easier when the testsare nonverbal in nature but, even then, equivalencemay not be achieved. Visual material, for example,may picture objects that are unfamiliar in a differentcultural context. The experimenter should always bealert to the possible influence of test materials onoutcome measures. If, as is usual in nutrition studies,the goal of the study is to compare groups on out-come only, without relating their scores to the pop-ulation norms, then it is permissible to use translateditems, provided that this is made clear when the re-search is reported. The crucial thing is that all par-ticipants are given the same test. Nevertheless, theresearcher must be alive to the possibility thatchanging items by translation may affect the validityof the test. We include a table of some common testsand the European languages in which psychometri-cally adequate versions are available (Appendix 1).Many American tests have Spanish versions becauseof the size of the Hispanic population. Our choice oftests was guided by the two principles of (a) use in aWestern European context and (b) language avail-ability, in conjunction with an attempt to cover themain cognitive domains.

Another important consideration is whether a testis standardized or not. Standardized tests havenorms that reflect the scores that the population forwhich it is intended achieve on the test; the averageperformance for the group is usually, but not always,assigned a score of 10 or 100. The larger and morerepresentative of the population the better; tests withnorms derived from small groups may not be soreliable in reflecting the performance of an individ-ual relative to the group. The use of standardizedtests allows one to assess where an individual child’sperformance lies with respect to the population butif this is not a concern of the study, then nonstan-dardized tests may be suitable. This is usually thecase when two groups, such as one receiving anutritional supplement and another in a placebocondition, are compared and our interest is indetermining whether they differ significantly onsome cognitive outcome but not in relating this to abroader population. Providing that they are used inthe proper context, nonstandardized tests shouldcertainly be considered, although it will still becrucial to determine that the psychometric proper-ties are adequate. Another advantage of standardizedtests, relevant in some studies, is that they allowcomparison between scores on different tests. It isnot obvious that a child receiving raw scores of 38/62 on an arithmetic test and 21/40 on a test ofspelling, for example, is at the same level on the twotests of attainment. Standardized tests, however,produce standard scores and if we know that each ofthese raw scores has the same standard score, thenwe can legitimately draw this conclusion.

Researchers should check to see how recently a testhas been normed. Norms produced many years pre-vious may give less accurate estimates of currentpopulation performance levels. The Flynn effect [16]refers to a rise in average IQ scores in a populationthat occurs over time. The average rate of this riseseems to be around 3 IQ points every decade. Therehas been much discussion about this phenomenonand about the possible causes, one of which is sug-gested to be improved nutrition. A practical effect isthat norms become outdated and the ‘‘average’’ in-creases with time, i.e., 100 will no longer be the meanIQ score in a population. Test manufacturers, there-fore, produce new norms from time to time and theuse of outdated tests my give misleading measure-ments. Although only demonstrated for IQ tests, somepeople assume that other tests of cognition show in-creases over time as well.

To be of scientific value, psychological tests need tomeet a set of ‘‘psychometric’’ requirements [28]. Inorder to allow us to draw valid conclusions, measuresmust be both valid and reliable. A test’s validity re-flects the degree to which it actually measures what it


purports to do. There are various types of valid-ity—for example, content, construct, predictive—butall are concerned with determining how confident wemay be in using the test results to make inferences.Reliability refers to the consistency of test measure-ments—would a child taking the same test twice re-ceive the same score on both occasions? We can thinkof this in terms of the stability of the results, expressedin terms of a reliability coefficient. As a general rule,we would expect the reliability coefficient of a cogni-tive test to be at least 0.80. Standardized tests will haveinformation about these indices in the technicalmanuals that accompany the tests. The manual for theWISC-III, for example, reports reliability coefficientsof 0.95, 0.91 and 0.96 for VIQ, PIQ and FSIQ, respec-tively. It also shows that the reliability coefficients forthe subtests are generally lower than for the globalmeasures, illustrating that greater confidence may beplaced on the accuracy of the IQ score than for that ofan individual subtest. The standard error of mea-surement of a test and the confidence intervals sur-rounding a score are also useful indicators of a test’sreliability. Such information will not usually beavailable for nonstandardized tests and the researcherwill have to use more indirect sources: do the testscores correlate adequately with standardized mea-sures of the same function? Do the test results covarywith results from other tests in an expected way?

Tests must also be sensitive so that they are able tomeasure what we want to observe. Many of the testsdescribed below have been developed in a neuropsy-chological context, designed to detect changes in cog-nitive behavior in individuals with neurologicaldisorders or who have suffered brain trauma, but theymay not be so useful when the changes in cognitivebehavior are more subtle, as in many nutritionalstudies. There are no standardized tests that have beenspecifically designed for use in nutrition studies butseveral of the listed tests have been developed to beparticularly sensitive to small changes in performanceand have been used successfully in drug trials withnormal volunteers. This may make them particularlysuitable for nutrition research. A good general guide isto consult the literature to find tests that have provedsensitive in similar situations in the past. A useful ref-erence is Westenhoefer et al. [36] that contains adescription of tests that have been used in previousnutritional studies. If such information is not available,it is worth conducting a pilot study first to see if theproposed measures are sensitive enough for the study.

Tests for school-age children are usually in eitherpencil-and-paper form or presented by computer. Thevariety of the former is greater and their adminis-tration does not depend on costly equipment orproximity to a computer, so they are generally moreflexible. Both types of test vary in the degree of

experimenter involvement, ranging from tests that arecompletely automated to those in which each item ispresented by the researcher. The latter must beadministered individually but some tests can be givento groups. Computer presentation is generally moreobjective and finer measurements of variables likereaction time are possible to obtain, but it may notallow for detailed observation of test behavior. Pri-orities must be set by the researcher.

Whatever type of test is used, it is imperative thatstandard instructions are used with each child. In asense, each test administration is akin to a psycho-logical ‘‘experiment’’ in which a child is placed in astandard situation and fixed materials are used toelicit samples of behavior. The instructions given tothe child are a component of this standard situationand, unless the same instructions are used verbatimwith each child, the results between children will notbe comparable. Nonpsychologists do not alwaysrealize this and, for the best of motives, try to aidchildren to do their best by ‘‘helping them along’’ withadditional instructions or examples. Tests for childrenunder two years of age, before language acquisitionhas occurred, are rather different in form; by neces-sity, most such tests consist mainly of ‘‘performance’’items. Such items involve compliance with the tester’srequests and are, therefore, susceptible to tester bias.To avoid this and to acquire the skills necessary forsmooth administration of the test, extensive trainingis usually required.

The main factors in determining the choice of testswill be the age of the participants, the outcome to bemeasured and the language of administration. Inpractice, the choice of tests is also often constrainedby the time available for testing.

Cognitive tests

We present below a selection of tests, organized bylevels of assessment, from the wide variety available.Where possible, we have selected tests that are avail-able in European languages as well as in English (seeAppendix 1 for a list of common tests available in avariety of European languages). We have included alist of tests described in the text in Appendix 2, pro-viding information about their age range and time toadminister, as well as a list of websites for test pub-lishers. The publisher of a specific test is usually easilyobtained by using an internet search engine. Pub-lishers may be approached to find out whether a testof interest is available locally and/or in the locallanguage. Some tests will be available in differentlanguages but will have normative data only for theoriginal population; decisions about their suitabilitymust reflect the points we have described above.


Level I—overall ability

Tests of overall ability attempt to describe the generallevel of intellectual function that a child displays. It isassumed that this represents some sort of intellectualreserve that is deployed across more specific cognitivedomains. Tests that attempt to produce an overallassessment of cognitive ability reflect this as they areusually, although not always, composed of a variety ofsubtests that have been found to have shared variance.

Measuring overall ability forms an important partof most test batteries for several reasons. These arethe only measures that some nutrition studies report,making them necessary for comparison with the lite-rature. They also provide a standard against whichspecific tests can be evaluated. How we interpret ascore on a reading test of 80 in a child with an IQ of100 is likely to be different in a child with the samereading score but an IQ of 80. We also often need torule out overall intellectual differences betweengroups as an explanation for more specific differencesin cognition that we wish to attribute to nutritionalintervention. Overall tests, however, also haveimportant limitations. When they are composed ofvarious subtests, the same overall IQ score may reflectdifferent patterns of performance, sometimes quitesubstantial, as different combinations of subtestscores may add up to the same total score. It is alsowrong to assume that two groups of children with thesame overall score will necessarily perform at thesame level on specific tests—identical IQs may maskdifferent patterns of ability on other tests. There aresome cognitive domains, particularly executive func-tions, which are under-represented on most intelli-gence tests. As long as the limitations are recognized,however, a measure of overall ability can be a usefulbasis for a cognitive assessment. There is some evi-dence that early variations in diet may selectivelyinfluence VIQ scores [21, 36] while later interventionsmay have more effect on PIQ [3].

j Birth—2.5 years

In children of this age, tests of general ability areusually thought to assess the level of neurodevelop-ment rather than intelligence. Studies of nutritionaleffects in this young population are especially sig-nificant since the plasticity of neural development inthis period renders the young child particularlysensitive to nutritional factors. Probably the mostwidely used test in the English-speaking world is theBayley Scales of Infant and Toddler Development(Bayley-III: age range: 1–42 months; Test 1). The testcovers three major areas of development using ahigh degree of child–tester interaction: cognitive;

language (receptive and expressive) and motor (fineand gross). In addition, two scales use parentquestionnaires: social-emotional and adaptivebehavior. The Behavior rating scale is an optionalsixth subtest. As well as composite scores, the use ofsubtest scaled scores allows discrepancy analysis todetermine areas of strength and weakness. Testadministration is, of course, individual. Formerversions of the test provided two composite scores:the mental development index and the psychomotordevelopment index. Although the number of lan-guages in which it is available is limited, the per-formance nature of many of the items means thatmany could be used without translation, althoughthe level of a child’s familiarity with specific testmaterials may still influence performance. Learningto administer tests like the Bayley-III to infants andyoung children involves a high-degree of training,necessary if any confidence is to be placed in the testresults.

Scale 1 of The Griffiths Mental Development Scales(Test 2) covers the 0 to 2-years age range and providesmeasures in five areas of cognitive development:locomotor, personal-social, hearing and speech, eyeand hand coordination and performance. It is a UKbased test and is widely used by medical researchersand practitioners but is not available in other lan-guages. A new American test, the Mullen (Test 3) isstarting to be used widely by researchers. It can beused from birth to 68 months of age. It measures fineand gross motor function, visual reception, expressiveand receptive language, but is only available at pres-ent in English.

There is one more type of assessment that is oftenused as a measure of general ability in infants be-cause, although limited in scope to one type offunction, there is some evidence that it may serve asan indicator of intelligence scores later in childhood(although reported correlations vary and are mostlylow [6]). This assessment measures infants’ visualnovelty preference or visual recognition memory. Theresearcher exposes the infant to a target visual sti-mulus until the infant habituates to it and looks away.The target stimulus is presented again along with anovel stimulus and the relative proportion of timespent looking at the two stimuli is measured. It hasbeen argued that infants who tend to habituate morequickly and spend more time looking at the novelstimulus rather than the familiar stimulus also tend toperform well on tests of general intelligence at laterages. It is hypothesized that the processes underlyingthese early visual novelty preferences are basic tointellectual performance and are the same as thosethat will underlie intelligent behavior later in the life-span. One such formalized procedure is the FaganTest of Infant Intelligence (FTII, Test 4).


j Over 2.5 years

The choice of tests to assess intelligence in childrenolder than 2.5 years is much wider. Most are com-posed of a battery of subtests, sampling a broadspectrum of behavior, the scores of which are thencombined to give a single measurement, the full scaleIQ (FSIQ). These subtests can usually be grouped intothose that involve verbal material and those in whichthe stimuli are nonverbal—most tests provide sepa-rate measures of the child’s ability in these two broaddomains as well as FSIQ. By far the most widely usedtests worldwide, available in the greatest variety ofEuropean languages (see Appendix 1), are theWechsler scales. Two of these cover the age rangefrom 2 years 6 months to 16 years 11 months: theWechsler Preschool and Primary Scale of Intelligence(WPPSI-III, Test 5) and the Wechsler intelligenceScale for Children (WISC-IV, Test 6) for the oldergroup. An adult test (the Wechsler Adult IntelligenceScale—WAIS) allows assessment over the whole agerange, making the Wechsler tests suitable for longi-tudinal studies. All Wechsler tests provide a verbal IQ(VIQ) and a performance IQ (PIQ) in addition to theFSIQ. The Third edition of the WPPSI is somewhatdifferent in format to its predecessors, in that it hastwo different forms to be used with the 2 years6 months–3 years 11 months and the 4 years0 months–7 years 3 months age groups. In additionto IQ scores and a general language composite scorefor both groups, a processing speed quotient can alsobe obtained for the older group only. The WISC-IV, inaddition to IQ scores, also provides four index scoresallowing a finer-grained analysis of cognitive ability:verbal comprehension, perceptual reasoning, workingmemory and processing speed. The tests allow scorepatterns across subtests to be analyzed. If time con-straints exist, since they take around 90 min toadminister in full, the Wechsler Abbreviated Scale ofIntelligence (WASI) which provides VIQ, PIQ andFSIQ scores based on four subtests may prove useful(Test 7). Other widely used intelligence tests in Eng-lish are: the Kaufman Assessment Battery for Children(KABC-II, Test 8), the British Ability Scales (BAS2,Test 9) and the Differential Ability Scales (DAS, Test10), an American test based on the BAS; all are astime-intensive as the Wechsler scales.

All the above tests share a similar structure, butthere are other options worth noting. The CognitiveAbilities Test (CAT3, Test 11) is designed for groupadministration in school-age children. It measuresreasoning in three domains: verbal, nonverbal andnumerical, and could prove useful in studies wheresome measure of overall ability is needed but indi-vidual face-to-face testing is not feasible. The Leiter

International Performance Scale (Leiter-R, Test 12) isentirely nonverbal and performance is, therefore,independent of the language of administration. It usesvisualization and reasoning scales to measure IQ andalso has scales that measure attention and memory.Finally, Raven’s Progressive Matrices (RPM, Test 13)is a widely used nonverbal test of intelligence whichwas designed as a measure of ‘‘g’’. A colored form foruse with children and a standard form for olderparticipants are available; in both, the oral instruc-tions are very simple. The test does not measureverbal ability although vocabulary scales are availableto use with it, but only in English. The correlationsbetween scores on the RPM and the Wechsler teststend to be in the 0.70–0.80 range. The test can beadministered to a group and it is available withmanuals in a wide range of European languages (seeAppendix 1).

Another useful nonverbal intelligence test is theSnijders–Oomen (SON-R, Test 14). Seven subtestsmeasure four areas of cognition: abstract reasoning(categories and analogies), concrete reasoning (situ-ations and stories), spatial abilities (mosaics andpatterns) and perceptual competence (hidden pic-tures). Although the time needed to administer theSON-R is, on average, 1.5 h, there is a shortenedversion consisting of four subtests: categories,mosaics, situations and analogies, taking around45 min.

The most recently published nonverbal test is theWechsler Nonverbal Scale of Ability (WNV, Test 15),which gives an overall IQ. It has two forms, dependingon age at assessment. At each age-level, there are two-subtest (administration time 10–20 min) and four-subtest (30–45 min) batteries. It has been designedfor use in different countries and it is claimed that thenative language of the child is irrelevant to perfor-mance. All directions are given in pictorial form sothere is no need to adapt the administration of the testor to pantomime instructions. Initially, norms areavailable for USA and Canada only, but it could beused for group comparisons in many countries.

Level II—batteries measuring multiple cognitivedomains

We are concerned here with obtaining measures ofperformance on a variety of cognitive domains, usu-ally with a series of sub-tests within each domain. Themain advantage to this approach is that norms for allthe sub-tests have been obtained from the samepopulation and they can thus be compared easily andreliably with each other. In nutrition studies, wemight hypothesize that one particular domain would


be affected more than others and the use of such abattery would allow us to confirm the selectivity of theintervention. There are no such batteries for use withchildren under 2.5 years of age since the functionsthat these tests measure are not considered directlytestable in infants and toddlers. Even in older chil-dren, the choice is much narrower than for tests ofoverall ability.

j Over 3 years

The best known is the NEPSY-II (from NEuroPSY-chology), originally designed to measure the neuro-psychological status of children aged 3–12 years (Test16). It covers five domains: attention/executive func-tion, language, memory and learning, sensorimotorfunction and visuospatial processing. A series of coresubtests for each domain measures key componentprocesses, as well as supplementary tests if more de-tailed assessment is required. For example, the mem-ory and learning subtests are: immediate memory forsentences, immediate and delayed recall of names andfaces, list learning and narrative memory under freerecall and cued conditions. Five domain scores can becalculated, provided that the core subtests have beenadministered, but there is no overall score as there isin the case of IQ, since it is the pattern of performancethat is of interest. A new sixth domain, social per-ception, has been added in the second edition and theupper age limit has been extended to 16 years. The testis published in the USA, although it was developedoriginally in Finland and it is available in someEuropean languages with local norms.

Another option is the CANTAB (CAmbridge Neu-ropsychological Test Automated Battery), a computer-based battery of tests originally developed for adultsbut now with norms for children (Test 17). Until re-cently, all test stimuli were nonverbal and described asindependent of language and culturally nonbiased;several verbal sub-tests have been added recently.These 22 subtests are grouped into batteries to providemeasures of domains like executive function; it ispossible to create customized batteries for any partic-ular study. A unique feature of the CANTAB is thatperformance on the various subtests is known to cor-relate with activity in particular regions of the brain,one reason why the CANTAB is very widely used inacademic research settings. All responses are auto-matically recorded during performance, with precisetimingwhere relevant, and awide variety ofmeasures isprovided for each subtest. The CANTAB has proved tobe sensitive to small changes in behavior, such as thoseassociated with medication in normal subjects and thisfeature has also made it a popular choice for researchstudies and a good candidate for use in nutrition

studies. While the new verbal sub-tests may not besuitable for use with non English participants, thenonverbal tests should be.

Another UK-based computer battery is CognitiveDrug Research, known as CDR (Test 18). As the nameimplies, this battery was developed for use in drugtrials, which may make it suitable for nutrition studies.It has been used very little in academic research, sincethe company is usually contracted to conduct thestudies itself. The company is flexible in its approach,however, and attempts to meet researchers’ needs. Thebattery was developed for use with adults but mostsubtests would probably be suitable for older childrenand possibly with younger ones after a pilot study.Similar in structure to CANTAB, the core subtestsmeasure motor control, attention, working and sec-ondary memory with a series of supplementary tasksto provide further information. Although the tests aremainly nonverbal, the few that use verbal stimuli areavailable in most European languages.

Level III—individual cognitive domains

Wemay have an hypothesis predicting that function ina particular cognitive domain will be affected by ournutritional intervention or even that some aspectwithin a domain will be selectively affected, leavingother aspects intact. In cases such as this, we will wantto administer a test that measures a variety of cognitivecomponents within a domain. Many studies, however,use just one or two tests, often combining individualtests from different domains to form an ad hoc battery.Thismeans, of course, that the researcher has chosen toevaluate only one or two of the many component pro-cesses of the cognitive domain. This may be appro-priate if there is an hypothesis predicting that certainvery specific processes will be affected selectively by thetreatment. The tests will not have been normed on thesame population, however, making comparison be-tween test scores more difficult than if a multiple-component test is used. It is usually best, under thesecircumstances, to also include a test of another functionin the same domain not hypothesized to be affected, inorder to establish the selectivity of the effect. Examplesof both types of test, multiple and single component,are described in this section. Many more may be foundby consulting test catalogues and textbooks. Few suchtests are available for the younger age group.

It is worth noting that some Level II Batteries,notably the NEPSY, can be used to assess single do-mains in isolation and single tests from these domainscan also be used. Some of the tests are suitable for 3 to4-year-olds and these are often the only standardizedtests available for this age group. Almost all the


subtests can be used between 5 and 12 years, so it isworth considering such batteries as a source of usefultests. They must, however, allow performance onindividual tests and domains to be quantified and nothave a single score as the only outcome measure.

j Memory

Memory refers to the registration, storage, retentionand later retrieval of information to which the personhas been exposed and is a complex system made up ofcomponents that may be dissociated from one an-other [2]. Thus, a comprehensive assessment ofmemory will need to have a series of sub-tests. Wemay, for example, measure memory for verbal versusnonverbal material, vary the presentation so thatsome stimuli are visual and some auditory, test dif-ferent modes of retrieval such as recall or recognition,and vary the interval between first presentation of thematerial and subsequent recall.

Multiple component tests

No such comprehensive memory batteries are avail-able for children under five.

Age 5 and over There are three widely used batteriesto measure memory function: childrens’ MemoryScale (CMS, Test 19), test Of Memory And Learning(TOMAL, Test 20) and the wide Ranging Assessmentof Memory and Learning (WRAML-2, Test 21). Al-though there are minor differences between them, wewill describe the CMS to give some idea of how theyare all structured. The CMS uses visual and verbalmaterial in short and long-delay conditions. Recall,recognition and working memory are assessed andlearning and attentional characteristics are described.A core of six subtests can be administered with addi-tional subtests for more complete evaluation. A seriesof composite index scores (for example, verbalimmediate, verbal delayed, attention/concentration) isproduced, as well as a general memory index. Thereare tables that allow the determination of whetherdissociations between different aspects of memory arestatistically meaningful, for example, attentional fac-tors versus memory per se. A useful feature of the CMSis that it allows memory scores to be linked to per-formance on the Wechsler scales so that we can see ifmemory function is in line with expectations from IQ.

Different in structure and purpose to the above isThe Rivermead Behavioral Memory Test (RBMT-II;RBMT-III due in 2008, Test 22), designed to assessmemory skills related to everyday situations. It claimsto be more ecologically valid and includes subtests

measuring the recognition of faces and objects,‘‘prospective’’ memory for future activities, recall of aroute and a story, both immediately and after a delayand items measuring orientation in time and place.Quick to administer, this test provides two measures:a screening score that offers a simple way to estimatethe extent of everyday memory problems, and aprofile score which gives a more sensitive measure.Four parallel forms of the test are included, allowingaccurate measures of change over time in test–retestdesigns. The adult version has been used with ado-lescents from 11 to 15 years of age. A children’s ver-sion (RBMT-C; Test 22) is available for ages 5–10.

Single component tests

Under 3 years Memory in children under 3 years ofage is rarely assessed in the same way that it is in olderchildren and the measures are often primarily de-signed for other purposes. The Bus Story (Test 23), forexample, involves telling the child a story while lookingat a series of accompanying pictures and then askingthe child to tell the story back using the pictures asprompts; this was devised as a language test, althoughit obviously has a large memory component. Tests ofvisual recognition memory such as that of Fagan werementioned above as measures of overall cognitiveability. Tests like the BSID and the WPPSI containsome items that measure memory. Uncertainty aboutthe processes involved in memory development inyoung children have hampered the development ofappropriate standardized tests although research par-adigms may be of interest (see below).

Over 5 years The Children’s Auditory VerbalLearning Test (CAVLT-2, Test 24) measures auditoryverbal learning and memory abilities in children andadolescents aged 7–18 years, using a word-list recallparadigm. Although it is a single test, it uses onerecognition and two free-recall memory word listsand yields measures of immediate memory span, levelof learning, immediate recall, delayed recall, recog-nition accuracy, and total intrusions.

The Visual Aural Digit Span test (VADS, Test 25) issuitable for school-aged children and consists of foursubtests in which numeric sequences involving thedigits 1–9 are used as stimuli. The child orally repeatsor writes the sequences frommemory. The first subtestrequires oral repetition of orally presented digits. Inthe second subtest, the child must orally repeat digitsthat are presented visually. The third subtest requiresthe child to write digits that are orally presented whilethe fourth subtest involves writing digits that arevisually presented. This provides a great deal ofinformation (11 measures) compared to the usual digit


span test involving oral presentation and recall only,although the latter has the advantage of recall both inorder of presentation of the stimuli and backwardsand can result in some interesting dissociations be-tween performance in the two conditions. The VADS iseasily adapted for use in different languages.

The Rey–Osterrieth (RCFT, Test 26) test usescomplex visual material to measure visuospatialability and visuospatial memory. It is suitable forchildren from 6 years as well as for adults. Althoughthere are various research versions of the test and avariety of scoring systems, it is now available as acommercial test with appropriate norms. The figure iscopied when first presented and the child is then re-quired to reproduce it from memory both immedi-ately and after a 30 min delay. There is also arecognition trial in which the child has to select whichelements of the figure have been seen before bychoosing between correct items and incorrect foils.The RCFT measures visuospatial construction ability(copying trial) and visuospatial memory (immediateand delayed memory, recognition). Also available isthe developmental scoring system for the Rey–Ost-errieth Complex Figure (DSS-ROCF) which is suitablefor children aged 5–14 years. It is specifically con-cerned with evaluating performance within a devel-opmental context and emphasizes qualitative aspectsof the child’s drawing such as organization and style.Determining the maturity of the child’s performancemay be of interest in some studies.

j Language

Language and verbal processing play a key role, ofcourse, in the assessment of VIQ. Such tests, however,do not assess language per se but rather how it reflectsverbal knowledge and verbal reasoning. A child askedto define a word on an IQ vocabulary sub-test, forexample, may use ungrammatical language but iscredited for the item if the reply contains the correctinformation. Language assessment, on the other hand,concentrates more on the formal and pragmatic as-pects of verbal usage. Given the obvious impact ofpast learning experiences, ethnic background, andother factors on language abilities, language-basedtests provide highly confounded measures of cogni-tive abilities. Because of the importance of language ineducational achievement and its use in verbal rea-soning and verbal problem-solving, however, it isimportant to determine whether nutrition affects thisdomain of cognitive function. Any assessment incor-porating verbal measures must ensure that adequatecontrols are implemented to minimize or account forconfounds; details of maternal and paternal educationand achievement, socio-economic status and infor-

mation about the level of stimulation in the homeenvironment (preferably using a measure such as theHOME inventory [13] should be collected and used inanalyzes of the data). Because of the sensitivity tocultural differences, it is best to use a test devised foruse in a particular country. Speech and languageprofessionals should be able to provide advice forsuitable materials available locally.


At the very least, a comprehensive language assess-ment should provide measures of both receptive andexpressive language. It may also measure componentswithin each of these two domains.

Under 3 years The MacArthur–Bates Communica-tive Development Inventories (Test 27) are checklistsreflecting language comprehension and expression thatare completed by parents; scores obtained from themhave proved to correlate highly with assessments bylanguage professionals. The CDI: words and gestures(infant form) and CDI: words and sentences (toddlerform) are used from8 to 30 months. There is anupwardextension, the CDI-III, suitable for assessing languageskills in children aged 30–37 months. Typically,vocabulary checklists are presented and the parentindicates which items the child comprehends andproduces. For older children, there are questions aboutthe use of gestures, combining words, the degree ofgrammatical complexity in the child’s speech, and yes/no questions concerning semantics, pragmatics, andcomprehension. The tests have been very widely usedand there is a large literature available. Relevant to thisarticle is the test availability in a wide variety of Euro-pean languages. As well, the website contains usefulinformation about test development in different lan-guages which has relevance beyond this particular test.

The NEPSY language domain includes four sub-tests assessing receptive and expressive language for 3and 4-year-olds.

Over 3 years The Clinical Evaluation of LanguageFundamentals (CELF 4th edition, Test 28) is anexample of a test designed to measure a wide varietyof language components, containing both core andsupplementary subtests. The four core subtests pro-vide a total language score as well as composite scoresfor language structure, language content, languagecontent and memory and working memory. Addi-tional tests measure the skills and behaviors thatunderlie the child’s language: phonological awareness,rapid automatic naming, digit span, sequences, wordassociations and memory. These are useful in trying


to determine whether there is a selective effect of, forexample, nutrition on certain aspects of overall lan-guage function. There is also a preschool version(CELF Preschool, Test 28). It too measures a widerange of language components.


Over 3 years Probably the single most widely usedtest to assess language is an American test, the Pea-body Picture Vocabulary Test; its incarnation in theUK is called the British Picture Vocabulary Scale(BPVS-II, Test 29). This is a quick measure ofreceptive language that estimates the size of thechild’s lexicon. Pictures are presented, four to a page,and the child must point to the one that matches aspoken word; no reading, writing or speaking arerequired of the child. The stimulus items range fromconcrete to abstract. It does not measure expressivelanguage but is often used to estimate overall verbalability (the highest correlation between a subtest andoverall IQ is that for vocabulary).

The Test for Reception Of Grammar (TROG-2, Test30) tests the understanding of 20 English grammaticalconcepts, four times each, using a picture multiple-choice format. The child points to the picture thatmatches a spoken phrase or sentence by the tester so,like the BPVS-II, this is a measure of receptive lan-guage but at a more complex level than single-wordknowledge.

The most usual way to measure expressive lan-guage, using a single test, is to administer a namingtest. The child is presented with a series of pictures ofobjects, usually sequenced in order of difficulty, andhas to produce the correct label for the object. There isno one test that is dominant in this field but the WordFinding Vocabulary Test (Test 31) is typical. Morecomplex expressive language in children aged 3–8 years is often assessed by the Bus Story Test (Test23). As described above, the child is told a story whilelooking at accompanying pictures and then re-tells thestory. The age level of consecutive speech used in thisre-telling can be assessed from the information con-tent, sentence length and grammatical usage. Manylanguage tests can be found by consulting educationalsuppliers such as Taskmaster (see websites list).

j Attention

The term ‘‘attention’’ refers to our ability to select andfocus on specific aspects of the stimuli, both internaland external, to which we are exposed. It is closelyallied to working memory, the system that allows us tooperate cognitively on information held in con-sciousness. Attention, like other cognitive domains, is

complex and current research indicates that differentfunctions within this domain are underpinned bydifferent neural systems. Commonly identified com-ponents of attention are focusing, sustaining andshifting [26] or, referring to similar concepts but usingdifferent names, detecting, alerting and orienting [30].Tests should assess attention in both visual andauditory sensory modalities. There is increasing re-search interest in attention processes, in part becausethey are implicated in the performance of many higherlevel cognitive functions, but also because deficits anddifficulties in attention regulation are prevalent inchildren and young people, with serious effects onlearning and associated with behavioral disorders.

Disturbed attention is frequently associated withnutritional disorders.


There are no attention batteries for use with childrenunder 6 years of age at present.

Over 6 years The Test of Everyday Attention forChildren (TEA-Ch, Test 32) assesses a range ofimportant attentional capacities and comprises ninesubtests which measure children’s abilities to attendselectively, to sustain attention, to divide attentionbetween tasks, to switch attention and to inhibit verbaland motor responses. Both auditory and visual stimuliare used. The TEA-Ch is suitable for children aged 6–16 years, with separate norms for boys and girls.


Three to four years There are two NEPSY subteststhat assess attention/executive function in this agegroup.

Over 4 years Conners’ Continuous PerformanceTest (CPT II, Version 5, Test 33) is an attention testthat is widely used in ADHD research and clinicalassessments for people aged 6 years or older, and isquick to administer. Responses to target letters on acomputer screen provide information about the spe-cific type of deficits that may be present. For example,some response patterns suggest inattentiveness orimpulsivity, while others may indicate activation/arousal problems or difficulties maintaining vigilance.This test can only be presented by computer. TheKiddie version (K-CPT, Test 33), for 4 to 5-year-olds,uses the same basic paradigm but with pictures ofobjects rather than words. It takes 7 min to admin-ister. A similar test, also presented by computer, is theTest of Variables of Attention, for 4 to 80-year-olds,


which has extensive norms for general, as well asclinical, populations (Test 34). One advantage of thistest is that it has two versions, one visual (TOVA) andthe other auditory (TOVA-A); it takes longer toadminister (25–30 min) but provides more informa-tion. NonVerbal tests of attention can usually be usedin any country.

The CANTAB, described above, includes severaltests of attention that can be used alone and severalsubtests on the Wechsler IQ tests are also said tomeasure attention. The CMS, also described above,gives scores on an attention/concentration dimension.

j Executive functions

We have earlier noted the significance of executivefunctions in cognitive performance. They may bethought of as higher-order cognitive skills that areassociated with the child’s ability to engage in inde-pendent, goal-directed behavior. A child with goodexecutive function will demonstrate mental flexibility,will be able to form and maintain sets, to plan and self-monitor behavior, and will be able to inhibit impulsiveresponses, as well as demonstrating abstract reason-ing, concept formation and rule learning. Deficits inexecutive function may show themselves as the lack ofbehavioral control (disinhibition), the inability tochange behavior despite corrective feedback (persev-eration), or inefficiency in problem-solving, difficultyin understanding cause–effect relationships and a lackof motivation. Although there are some experimentalprocedures for measuring problem-solving in youngchildren, the full range of executive functions isthought to only develop in later childhood [10] andmay continue well into adolescence as myelination ofthe frontal lobes proceeds. Although not tests, as such,two rating scales are available to evaluate a variety ofexecutive functions in school and home settings inchildren aged 2–18 years. The Behavioral Rating ofExecutive Function-Preschool (BRIEF-P, Test 35) isfor children aged 2–5 years 11 months and the BRIEF(Test 35) is for 5 to 18-year-olds. Scales are completedby parents, teachers or day care providers dependingon the age of the child. They are quick to administerbut provide a different type of information to thatobtained from test administration.


Seven years and over The Behavioral Assessmentof the Dysexecutive Syndrome in Children (BADS-C,Test 36) is a test battery that has recently becomeavailable to measure a range of executive functions inchildren aged 7–16 years. Its subtests assess the fol-lowing functions:

• inflexibility and perseveration;• novel problem solving;• impulsivity;• planning;• the ability to utilize feedback and moderate one’sbehavior accordingly.

The subtests and materials are novel and of interest tomost children. Each of the six subtests yields an age-scaled score, combined to form an overall scaled scorewhich is classified into six categories, ranging fromimpaired to superior. There is also a 20 item ques-tionnaire, the Dysexecutive Questionnaire for Children(DEX-C), to be completed by a parent and/or teacherwho has frequent contact with the child. The subtestsmay be used separately but the real strength of the testlies in the assessment and profiling of different func-tions.


One of the most common ways to measure (one aspectof) executive function is to use a verbal fluency test. Inthis kind of test, the requirement is to produce as manydifferent words as possible in a short time (usually 30–60 s), while following specified rules. Typically, thechild is asked to produce words that begin with a cer-tain letter, like M, (phonological condition) or belongto a certain category, like flowers (semantic condition),without repetition. This is considered to be a test ofexecutive function because it involves a number ofdifferent cognitive processes that are disrupted in thedysexecutive syndrome. For example, the child mustuse a strategy to extract items from its own lexiconwhile inhibiting nonrelevant responses, and must notbreak the rules by using words from outside the cate-gory or repeating a work (perseveration). There aremany versions of verbal fluency tests–a useful one, withnorms for children, is one of the sub-tests of the NE-PSY: the semantic condition is given to 3 to 12-year-olds, but the phonological condition is only used withchildren aged 7 years and above.

The Children’s Color Trails Test (CCTT, Test 37) issuitable for 8 to 16-year-olds and assesses sustainedattention, sequencing, and other executive functions.The child must connect stimuli scattered on a page bydrawing lines between them in a specified sequence. Itis suitable for use within any linguistic context as thestimuli are nonverbal and visual instructions may beused. It is quick to administer (5–7 min).

Another widely used single test is the Stroop colorand Word Test (Test 38). The children’s version is for5 to 14-year-olds and takes 5 min to administer. It tapsinto the executive function processes of cognitiveflexibility and resistance to interference from outsidestimuli. The test consists of three basic parts: a word


page consisting of the names of colors (red, green,blue) printed in black ink; a color page of semanticallymeaningless symbols (X) printed in different coloredinks; a color–word page comprising the words fromthe first page, printed in the colors from the secondpage, with the restriction that the word and color donot match. The task is to look at each sheet, and movedown the columns, reading words or naming the inkcolors, as quickly as possible, within a given time limit.The test yields three scores, and Interference, based onthe number of items completed.

j Visual functions

This refers not to the assessment of vision itself, butrather to cognitive activities that involve the use of vi-sion. For example, visuo-perceptual tasksmeasure howan individual perceives visual information, i.e., how theincoming visual stimuli information are interpreted bythe brain. Visuo-motor tasks look at abilities that in-volve vision to guide and monitor motor behavior,while visuo-spatial tasks assess the ability to makecorrect judgments about spatial relations. There arerelatively few tests to measure these abilities comparedto memory and language, particularly in children.Some individual tests are widely used but there are fewbatteries. It is worth pointing out here, particularly ifvisual functions are the focus of research interest, that itcan be quite difficult to isolate these functions fromverbal processing. Even when presented with visualstimuli, humans tend to label them verbally and willoften use verbal strategies when processing them.


Three years and over The Wide Range Assessmentof Visual Motor Abilities (WRAVMA, Test 39) is de-signed for use between the ages of 3 and 17 years. Itmeasures performance in three domains: visuo-motor(drawing), visual spatial (matching) and fine motor(peg moving). The latter is important since any motordifficulties may influence the score on visuo-motortasks. It is quick to administer, taking about 10 min.


Two years and above The Beery–Buktenica Devel-opmental Test of Visual–Motor Integration (BeeryVMI, 5th edition, Test 40) assesses the extent to whichindividuals can integrate their visual and motorabilities. The short format and full format testspresent drawings of geometric forms arranged in or-der of increasing difficulty that the individual is askedto copy. The age range is 2–18 years; the short formatis often used with children aged 2–8 years. It contains

two supplementary tests that help determine whetherany deficit noted on the test is primarily due to dif-ficulty with the visual or motor components. Theinitial test takes around 15 min, while the visualperception and motor coordination supplementarytests take 5 min each. It is linguistically neutral.

The Benton Judgment of Line Orientation Test(BJLO, Test 41) is a widely used neuropsychologicaltest and is a ‘‘pure’’ measure of visuospatial judgmentthat is language-free. It assesses the accuracy ofjudgments of the slope of visually presented lines byrequiring the testee to identify target stimuli con-sisting of pairs of lines of different slopes on a mul-tiple choice display. Originally designed for adults, itcan be used with children from 7 years upwards.

The NEPSY includes two measures of visuospatialfunction for use in 3 to 4-year-olds.

Level IV—research tests

These are tests that have been designed by researchersto fulfil the requirements of a particular study and areoften theory-based. Referring to the scientific literaturecan prove very fruitful as a source of tests; it is usuallypossible to contact the test authors to gain additionalinformation regarding psychometric properties,administration protocols and subsequent revisions, aswell as permission to use the test material. Wherecomparisons between treatment groups are concerned,and reference to population norms is not necessary,such research tests may provide adequately sensitiveindicators for very specific cognitive functions. Nev-ertheless, the researcher must not ignore the psycho-metric properties of such tests. Such tests are oftenparticularly suitable for use with young persons sincethis age group is under-represented in the body ofstandardized tests. Many researchers, for example, useresearch paradigms to study function in domains likememory [24] and problem-solving [37] in toddlers.

The number of such tests is legion but to illustratetheir character we will describe here just one—theinspection time paradigm, developed as a test of visualprocessing speed [27]. It is not concerned with howrapidly a response is made to the visual stimulus(reaction time) but rather with the amount of time thatan individual needs to be exposed to the stimulus tomake a correct response (inspection time). The stim-ulus is presented on a computer screen (Fig. 1) and thetask is to press a key to indicate which of the two legsof the figure is longer, left or right. Given unlimitedtime, this is a cognitively trivial task but the durationof presentation ranges from 6 to 200 ms; at the lowestexposure times, performance is usually random. Thereare many repetitions of the figure at different stimulus


exposure times and the number of correct decisions ateach duration is calculated. This sort of test is notstandardized and the exact details of the presentationprocedure may differ between studies, but it has led toa great deal of productive research.

Measuring indirect factors affecting cognition

As noted earlier, a multitude of factors can potentiallyaffect children’s cognitive performance when they arebeing assessed. This is especially true for youngerchildren. The testing situation itself can be perceivedin many different ways by children, and it should notbe assumed that children understand that the aim ofthe tester is to evoke behavior that is as close aspossible to the child’s best achievable performance,that is, as close an approximation as possible to thechild’s underlying competence [30].

While most test manuals stress the importance ofestablishing �rapport’ with a child testee, the situationis inevitably one that places a child under stress, notleast the stress of dealing with what have to be, bytheir very nature, difficult challenges, some of whichthe child will fail. The child is also face with a chal-lenging social situation, where the motives of astrange and more powerful adult need to be assessedand related to in a way that the child thinks appro-priate. Naturally, the child brings to such situationsexpectations of the most appropriate ways to behave,based on past experience. For some children, this maybe to withdraw and be highly compliant; for others itmay be to ingratiate themselves with the adult; ‘‘doing

their best’’ is not a universal response. It will alwaysbe important to ensure that staff who are testing theparticipants in a study have adequate training in testadministration with the relevant populations.

Motivation, temperament, illness, emotional dis-turbances, weather (windy conditions increase dis-tractibility in young children), noise levels, lighting,time of day, time of year and many other factors canall impact on a child’s performance. Apart from thefirst two, these factors are generally negative in theireffects, degrading performance and hence leading toinaccuracy in measures of competence or potential.Because of these nonrandom effects, the noise thatthey introduce into the data does not simply reducepower, but may also produce erroneous and mis-leading apparent effects.

This brief discussion of such contextual effects isgiven here primarily to stress the importance ofimplementing controls or measuring potential con-founds to minimize such sources of error, and toemphasize the need to use manualized procedures toensure that protocols are consistently administered.More about such issues are to be found in booksdevoted to testing considerations [25, 33, 34].

Concluding remarks

The focus of this paper has been on offering practicaladvice with regard to selecting tests that can be usedin nutrition studies, rather than on theoretical con-cepts. We have presented a variety of tests, using alevels-of-assessment approach to cognitive measure-ment, which may aid the researcher in selectingappropriate tests for their study.

Wemust emphasize that the more thought that goesinto planning the study before data collection starts, thebetter the outcome will almost certainly be. This meansarticulating very specificallywhat the aimof the study isand choosing a design that will allow the question to beanswered. It is a good idea to plan the way that the datawill be analyzed ahead of time to ensure that theappropriate data are collected to allow the statisticalanalysis. Particular attention needs to be paid to tryingto isolate factors that may confound the results. If thereare socio-economic differences associated with theoutcome of interest, for example, then that factor needsto be controlled, either by demonstrating that thegroups do not differ or, if they do, by using it as acovariate in analysis. Either way, it is necessary tomeasure socio-economic status at the time of datacollection and, as mentioned above, it is also useful toinclude some sort of measure like HOME [13].

If analyses of the data show that statistically signif-icant differences exist between treatment and control

a

c

b

L R

*

Hand

Fixation point appears

Figure appears for 6–200 ms

Which leg is longer?

Fig. 1 In the inspection time task, the child fixates the star. The experimentalfigure then appears for a period of time ranging from 6 to 200 ms. The childpresses either the L or R computer key to indicate which of the two legs of thefigure is longer. A large number of trials with randomized duration of stimuli ispresented


groups, it is important to realize that the underlyingeffect sizes are of great importance in interpretingfindings and publishing them. It can be misleading toreport statistically significant but very small effect sizesin popular media or advertising since the public maynot understand that other factors with much greatereffect sizes may in practice overwhelm the reportedeffect. Although it may be found that a difference isstatistically significant, it may nevertheless be so smallin absolute size that it in practical, physiological anddevelopmental terms it is of no significance whatever.In part this is because of the inherent instability ofcognitive testmeasures over time.Ameandifference of,say, 5 IQ points between two groups, for example, atreatment and a control group, is unlikely to be sus-tained as a group difference over time because of thenormal fluctuations in individual IQ scores over timethat can be expected. Longitudinal designs, where thepersistence of effects over time can be assessed by re-peated measures taken from the same individuals atsuccessive time points, offer one solution to this issue.

For reasons discussed earlier, tests of cognitivefunctions can never be 100% reliable. The variancethat even relatively reliable tests introduce into mea-surements always needs to be taken account of, andthis is a consideration that will always push for largerand hence more expensive, sample sizes. At the sametime, the issue of effect size is also relevant to deci-sions that have to be taken regarding the plannedsample sizes for research studies. While it is nowcommon practice to carry out statistical power cal-culations to estimate the minimum sample sizes re-quired to show statistically significant results, thismay lead to the use of excessively large and hencecostly samples if reaching statistical significance is theonly criterion. Smaller samples may be adequate todetect practically significant differences. Thoroughexamination of previously published results, ormaking use of a two-stage design where a pilot studyseeks to indicate potential effect sizes, are recom-mended so that meaningful effect sizes can be plan-ned for in the research design, rather than merelystatistical significance.

The present is an exciting time for research intonutrition and cognition since brain imaging proce-dures allow us to examine whether cognitive effectsmay be found to have measurable neural substrates.Whether nutritional intervention can alter brain anat-omy and/or physiology as well as behavior opensmanynew avenues of research. In tandem with this, newdiscoveries about the effects of gene · nutrient inter-actions on cognition also herald a new era of investi-gations. Such studies require collaboration betweenscientists trained in different disciplines to be fullyeffective. In all such studies, careful and appropriatemeasures of cognitive function are of prime impor-

tance.We hope some readers will go on to contribute tothis body of literature.

Notes on References

In the reference section, we have included three booksthat are excellent sources of general information abouttests [25–34]. All contain far more detail than spaceallows here. References to specific tests mentioned inthe text are not presented in this section, but rather inAppendix 2 which summarizes some of the maincharacteristics of each. The tests are grouped by levels(this should help to identify areas where there are test‘‘gaps’’) and appear in the order in which they do so inthe text.

j Acknowledgments The authors would like to thank the reviewersMaureen M. Black and Bonnie Kaplan for their useful commentsand discussions. This work was commissioned by the Nutrition andMental Performance Task Force of the European branch of theInternational Life Sciences Institute (ILSI Europe). Industry mem-bers of this task force are Barilla G. & R. Fratelli, Coca-Cola Euro-pean Union Group, DSM, Groupe Danone, Kraft Foods, Nestle,Sudzucker/BENEO Group, Unilever and Wild Flavors. For furtherinformation about ILSI Europe, please call +32-2-771.00.14 or email:[email protected]. The opinions expressed herein are those of theauthors and do not necessarily represent the views of ILSI Europe.

j Conflict of interests The authors have no financial or otherinterests that might conflict with the views expressed.

Appendix 1

Appendix 1 presents a list of widely used tests that areavailable in different European languages (all avail-able in English) (Table 1).

Appendix 2

Appendix 2 lists all tests described in the text inthe order in which they appear in the text, i.e.,grouped by levels. Summary information is given foreach (Table 2).

Table 1 Table of test availability in different European languages for somewidely used tests (all available in English)

Language Tests available

Dutch BSID-III, WPPSI-R, WISC-R, NEPSYFrench WISC-R, WPPSI-IIIGerman Raven’s, WPPSI-III, WISC-IVDanish Raven’s, BSID-III, WISC-III, WPPSI-RGreek WISC-IIISpanish NEPSY, BSID-I, WISC-IV, Raven’sHungarian Raven’sItalian Raven’s, WPPSI-III, WISC-III


Table

2Tableof

testsreferred

tointhetext,intheorderinwhich

they

appear

Levelof

assessment

Testname

Agerange

Timeto

administer

Notes

I—Overallability

1.Bayley-III

1–42

months

30–90

min

Threeadministeredscales:cognitive,language

andmotor;twoquestionnaires

2.Griffiths

MentalDevel-

opmentScale

0–2years

N/K

Measuresforfivedomains

offunction:

locomotor,personal-social,learning

and

speech,eye-handcoordination,performance

3.MullenScales

ofEarly

Learning

Birth,68

months

15minat

1year

Five

scales:fineandgrossmotor

function,

visualreception,expressive

andreceptive

language

4.FTII

Itemsfortestingat

27,29,39

and52

postnatalweeks

Measuresvisualnovelty

preference/

recognition

mem

ory

5.WPPSI-III

2years6months–7years

3months:(a)2years

6months–3years11

months

(b)4years0months–7years

3months

45minforcore

sixsubtests

(a)IQ

scores

+generallanguage

composite

available(b)Processing

speedquotient

inadditionto

above

6.WISC-IV

7years4months–16

years

11months

Variesby

numberof

subtests.

Approximately90

mininfull

VIQ,PIQandFSIQ

scores

+four

indexscores

7.WASI

6–89

years

30minforfour

subtests

VIQ,PIQandFSIQ

8.KABC-II

3–18

years

25–70

min

Scores

canbe

calculated

toexcludeverbal

ability

ifappropriate.Scales:simultaneous

processing,sequential,planning.Learning,

andknow

ledge(if

appropriate)

9.BAS2

Early

years:2years6months–

7years11

months

Schoolage:6years–

17years11

months

Variesby

form

andnumberof

subtests

Generalconceptualability

score(GCA)and

threeem

bedded

clusterscores

from

six

subtests.Furtherscales

availableformore

specificinformation

10.

DAS-II

2years6months–17

years

11months

Core

45–60

months

Diagnostics30

min

Often

used

toprovidebasisforintervention

inlearning

problems

11.

CAT3

7years6months–17+years

45min/section

Group

administration.Threesections:verbal,

nonverbalandnumericalreasoning.Often

used

inschoolsettings

inUK

12.

Leiter-R

2–21

years

25–40

min

Completelynonverbal.Usedto

determine

iflowachievem

entisdueto

lowIQ

ormorespecificcauses

13.

RPM

ColoredPM

—childrenStandard

PM—

generalpopulation

Advanced

PM—

top20%

ofpopulation

30minapproximately

Group

andindividualadministration.

Nonverbalstimuliand

simpleoralinstructions

14.

SON-R

5years6months–17

years

Averageof

90min(45–60

for

shortform)

Nonverbal.Measuresfour

areasof

cognition.

Short-form

available

15.

WNV

(a)4–7years

(b)8–21

years11

months

30–45

minforfour

subtests

Nativelanguage

irrelevantto

performance

II—Multicognitive

domains

16.

NEPSY-II

3–4years

5–16

years

Core

45minpreschool,1h

schoolage.Diagnostics

available

Measuressixcognitive

domains.Sub-tests

canbe

used

separately


Table

2Continued

Levelof

assessment

Testname

Agerange

Timeto

administer

Notes

17.

CANTAB

Adultandchild

norms

Tailoredto

research

needsso

variesconsiderably

Computerpresentation.Widelyused

inresearch.

Sensitive

tosubtlecognitive

changes

18.

CDR

Adulttestbutcouldbe

adapted

forusewith

children

Computerpresentation.Usuallyused

ina

commercialsetting—

little

academ

icresearch.

III—

Individualcognitive

domains

Mem

ory

19.

CMS

5–16

years

30min

Sixcore

sub-testsgive

indexscores,including

generalmem

ory

20.

TOMAL

5years–19

years11

months

45min

Tencore,four

supplementarysubtests

21.

WRAML2

5–90

years

45–60

min

Sixcore

sub-testsgive

threeindexscores

that

give

generalmem

oryquotient

22.

(a)RBMT-II

(b)RBMT-C

Use

over

ageof

11years

5years–10

years11

months

Approximately30

min

Four

parallelversions

available.Em

phasisison

mem

oryineveryday

situations

notthe

laboratory

23.

BusStoryTest

3–8years

15min

Designedas

alanguage

testbutcanbe

used

tomeasure

narrativemem

ory

24.

CAVLT-2

6years6months–17

years

11months

25min

Wordlist-learning

format

from

which

awide

variety

ofscores

isobtained

25.

VADS

5years6months–12

years

10min

Variedmeasuresof

short-term

mem

ory

26.

RCFT

6–89

years

45min,including

30mindelay

interval

Measuresvisuospatialability

andmem

ory.

Developmentalscoringsystem

savailable

Language

27.

MacArthur-Bates

Com-

municativeDevelop-

mentInventories

(a)CDIwords

and

gestures

(b)CDIwords

and

sentences(c)CDI-III

8–16

months

16–30

months

30–37

months

Checklistformat,parentalreportwell-validated

againstotherdata.Manylanguage

versions

available

28.

CELF-preschool2

CELF-4

3–6years

5years–16

years11

months

30–45

min30–60

min

Primarymeasuresof

receptiveandexpressive

language

plus

subtests

29.

BPVS-II

3years–15

years8months

5–8min

Measuressize

ofthechild’sreceptivelexicon

30.

TROG-2

4-year

adult

10–20

min

Measureshowwellgram

maticalconstructsare

understood

usingafour

picture

multiple-choiceformat

31.

Wordfinding

vocabularytest

3–9years

15min

50picturecardsfornaming

Attention

32.

TEA-Ch

6–16

years

55–60

min

Five

aspectsof

attention,no

overallquotient.

Separate

normsforboys

andgirls

33.

(a)CPT-II,Version5

(b)K-CPT

6-year

adult

4–5years

14min7min

Computerpresentationforboth

34.

TOVA

4–80

years

25–30

min

Twoversions:visual(TOVA)andAuditory

(TOVA-A)


Appendix 3: Notes on Websites

The British Psychological Society operates a websitethat includes details of a wide range of tests and givesaccess to reviews of them, along with discussion of arange of issues associated with psychological testing.We have also included a list of some major test sup-pliers’ websites; you can search the websites for spe-cific tests by name. Some suppliers requireprofessional eligibility and registration for buyingcertain tests, so that they may only be available forpsychologists, speech therapists, etc., but this infor-mation will be available on the website. There aremany smaller suppliers so we hope this list is useful,but it cannot be exhaustive and an internet searchengine will prove invaluable.

(a) Weblink to British Psychological Society websitefor psychological testing:

http://www.psychtesting.org.uk/

(b) Weblinks to test suppliers:

Association of Test Publishershttp://www.testpublishers.org/memserv.htm

Cambridge Cognition (CANTAB)http://www.cantabeclipse.com

Cognitive Drug Research (CDR)http://www.cognitivedrugresearch.com

Educational and Industrial Testing Servicehttp://www.edits.net

ERIC Test Publisher Directoryhttp://buros.unl.edu/buros/jsp/search.jsp

Harcourt Assessmenthttp://harcourtassessment.com

nferNELSONhttp://www.nfer-nelson.co.uk

Pearson Assessmentshttp://ags.pearsonassessments.com/

Pro-Ed, Inc.http://www.proedinc.com

Psychological Assessment Resources Inc.http://www.parinc.com

Taskmasterhttp://www.taskmasteronline.co.uk

Western Psychological Serviceshttp://www.wpspublish.com

Table

2Continued

Levelof

assessment

Testname

Agerange

Timeto

administer

Notes

Executivefunctions

35.

BRIEF-PBRIEF

2years–5years11

months

5–18

years

10–15

min

Ratingscales.BRIEFhasformsforparents

andteachers

36.

BADS-C

7–16

years

35–45

min

37.

CCTT

8–16

years

5–7min

Threeadditionalformsavailableforusein

research

studies

38.

Stroop

5–14

years

10–15

minutes

Visualfunctions

39.

WRAVM

A3–17

years

5–10

min

40.

BeeryVM

IFullform;Shortform;Tw

osupplementarytests

15min;10

min;5mineach

41.

BJLO

Adulttestbuthassomechild

norms

15min

Partof

theBenton

Laboratoryof

Neuropsy-

chology


References

1. Armitage P, Berry G, Matthews JNS(2002) Statistical methods in medicalresearch, 4th edn. Blackwell Sciences,Oxford

2. Baddeley A (1990) Human memory:theory and practice. Lawrence ErlbaumAssociates, Hove and London

3. Benton D (2001) Micro-nutrient sup-plementation and the intelligence ofchildren. Neurosci Behav Rev 25:298–309

4. Benton D, Kallus KW, Schmitt JAJ(2005) How should we measure nutri-tion-induced improvements in mem-ory? Eur J Nutr 44:485–498

5. Benton D, Ruffin HP, Lassel T, Nabb S,Messaoudi M, Vinoy S, Desor D, LangV (2003) The delivery rate of dietarycarbohydrates affects cognitive perfor-mance in both rats and humans. Psy-chopharm (Berlin) 166:86–90

6. Bornstein MH, Sigman MD (1986)Continuity in mental developmentfrom infancy. Child Dev 57:251–274

7. Caspi A et al (2007) Moderation ofbreastfeeding effects on the IQ by ge-netic variation in fatty acid metabo-lism. PNAS Early edition, 0704292104

8. de Graff Peters VB, Hadders-Algra M(2006) Ontogeny of the human centralnervous system: what is happeningwhen? Early Hum Dev 82:257–266

9. Deary IJ (2000) Looking down on hu-man intelligence. Oxford UniversityPress, Oxford

10. Diamond A (2002) Normal develop-ment of prefrontal cortex from birth toyoung adulthood: cognitive functions,anatomy, and biochemistry. In: StussDT, Knight RT (eds) Principles offrontal lobe function. Oxford Univer-sity Press, New York, pp 466–503

11. Dobbing J (1981) The later develop-ment of the brain and its vulnerabilityIn: Davis JA, Dobbing J (eds) Scientificfoundations of paediatrics, 2nd edn.Heinemann, London

12. Dobbing J, Sands J (1973) Quantitativegrowth and development of humanbrain. Arch Dis Child 48:757–767

13. Elardo R, Bradley R (1981) Homeobservation for measurement of theenvironment: a review of research. DevRev 1:113–145

14. Elman JL, Bates EA, Johnson MH,Karmiloff-Smith A, Parisi D, Plunkett K(1998) Rethinking innateness. MITPress, Cambridge/London

15. Epstein HT (1978) Growth spurts dur-ing brain development: implicationsfor educational policy and practice. In:Chard JS, Mirsky AF (eds) Educationand the brain (NSSE Yearbook, pt. 2).University of Chicago Press, Chicago,pp 343–371

16. Flynn JR (1984) The mean IQ ofAmericans: massive gains 1932 to 1978.Psychol Bull 95:29–51

17. Geisinger K (1994) Cross-cultural nor-mative assessment: translation andadaptation issues influencing the nor-mative interpretation of assessmentinstruments. Psychol Assess 6:304–312

18. Greenwood CE, Craig REA (1987) Die-tary influences on brain function:implications during periods of neuro-nal maturation. Curr Topics Nutr Dis16:159–216

19. Hebb DO (1966) A textbook of psy-chology, 2nd edn. WB Saunders Co.,Philadelphia/London

20. Herschkowitz N (1988) Brain develop-ment in the fetus, neonate and infant.Biol Neonate 54:1–19

21. Horwood LJ, Darlow BA, Mogridge N(2001) Breast milk feeding and cogni-tive ability at 7–8 years. Arch Dis Child84:F23–F27

22. Kallus KW, Schmitt JAJ, Benton D(2005) Attention, psychomotor func-tions and age. Eur J Nutr 44:465–484

23. Kline P (1996) Intelligence: the psy-chometric view. Routledge, London/New York

24. Kressley RA, Knopf M (2006) A com-parison of between- and within-sub-jects imitation designs. Infant BehavRes 29:566–573

25. Lezak MD, Howieson DB, Loring DW,Hannay HJ, Fischer JS (2004) Neuros-psychological assessment, 4th edn.OUP, New York/Oxford

26. Mirsky AF, Anthony BJ, Duncan CC,Ahearn MB, Kellam SG (1991) Analysisof the elements of attention: a neuro-psychological approach. NeuropsycholRev 2:109–145

27. Nettlebeck T (2001) Correlation be-tween inspection time and psychomet-ric abilities: a personal interpretation.Intelligence 29:459–474

28. Nunnally J, Bernstein I (1994) Psycho-metric theory, 3rd edn. McGraw Hill,New York

29. Paus T, Collins DL, Evans AC, LeonardG, Pike B, Zijdenbos A (2001) Matura-tion of white matter in the humanbrain: a review of magnetic resonancestudies. Brain Res Bull 54:255–266

30. Perret-Clermont A-N, Carugati F, OatesJ (2004) A socio-cognitive perspectiveon learning and cognitive development.In: Oates J, Grayson A (eds) Cognitiveand language development in children.Blackwell, Oxford, pp 303–332

31. Posner MI, Petersen SE (1990) Theattention system of the human brain.Annu Rev Neurosci 13:25–42

32. Sameroff AJ, Mackenzie MJ (2003) Re-search strategies for capturing trans-actional models of development: thelimits of the possible. Dev Psychopa-thol 15:613–640

33. Sattler JM (1994) Assessment of chil-dren: cognitive applications, 4th edn.Jerome M. Sattler, Publisher, Inc., SanDiego

34. Schmitt JAJ, Benton D, Kallus KW(2005) General methodological consid-erations for the assessment of nutri-tional influences on human cognitivefunctions. Eur J Nutr 44:459–464

35. Toga AW, Thompson PM, Sowell ER(2006) Mapping brain maturation.Trends Neurosci 29:148–159

36. Westenhoefer J, Bellisle F, Blundell JE,de Vries J, Edwards D, Kallus W, MilonH, Pannemans D, Tuijtelaars S, TuorilaH (2004) PASSCLAIM mental state andperformance. Eur J Nutr 43(Suppl 2):II/85–II/117

37. Willatts P, Forsyth JS (2000) The role oflong-chain polyunsaturated fatty acidsin infant cognitive development. Pro-stoglandins Leukot Essent Fatty Acids63:95–100


Documents

J. Oates cognitive abilities in children and young peopleilsi.org/mexico/wp-content/uploads/sites/29/2016/09/Nutrition-and...J. Oates Nutrition and cognition: assessing cognitive abilities