Brain Injury, July 2010; 24(78): 9951002

ORIGINAL ARTICLE

Cognitive reserve in paediatric traumatic brain injury: Relationshipwith neuropsychological outcome

AMANDA FUENTES1, CHERISSE MCKAY2, & CHRISTINA HAY2

1York University, Toronto, Ontario, Canada and 2Bloorview Kids Rehab, Toronto, Ontario, Canada

(Received 2 September 2009; revised 7 April 2010; accepted 26 April 2010)

AbstractPrimary objective: The current study examined the relationship between neuropsychological performance and cognitivereserve (as measured by word reading and vocabulary tasks) in children with TBI.Research design: Retrospective records analysis of the neuropsychological test results of 52 participants with medicallydocumented traumatic brain injuries, ranging from 616 years of age.Main outcome and results: Indicators of cognitive reserve were not correlated with the majority of well-recognizedneuropsychological measures.Conclusions: Although past research has found that verbal ability is a valid indicator of CR in adult populations, the presentstudy found evidence against the validity of this traditional reserve proxy when applied to the paediatric population. Thesefindings suggest one of two conclusions: (1) measures used to indicate CR in adult populations (word reading, vocabulary)are not valid indicators of cognitive reserve in paediatric populations; and/or (2) the measures themselves are valid, yet thereis simply not a significant relationship between cognitive reserve and short-term (i.e. less than 6 months) neuropsychologicaloutcome in paediatric TBI.

Keywords: Traumatic brain injury, Pediatric, Cognitive reserve

Introduction

A rising amount of research has explored cognitivereserve (CR) in individuals with neurological condi-tions as it has proven to be a significant factor inestimating the degree of resulting cognitive impair-ment. Few studies have investigated cognitivereserve in paediatric populations, despite the highincidence of traumatic brain injuries (TBIs) inchildren and adolescents. Cognitive reserve hasbeen defined as the discrepancy between thedegree of pathology and the degree of functionalimpairment evidenced across individuals with thesame disorder ([1], p. 131). This concept is mea-sured indirectly by proxies such as pre-injury cogni-tive ability, post-injury intellectual and academicfunctioning, socioeconomic status, family functionand level of education (in adults) [2]. The hypothesis

that cognitive reserve is a moderator of the effects ofTBI is consistent with the recurring clinical obser-vation that there does not appear to be a direct linkbetween the degree of brain pathology and levelof resulting cognitive impairment across individ-uals [3]. Variability in neuropsychological outcomeappears to be related to CR differences stemmingfrom discrepancies in education, occupation, socialeconomic status and genetics.

Evidence of reserve can be observed by studyingindividuals that have sustained pathological insult,such as a traumatic brain injury (TBI). Bigler [4]points out that the functioning of the brain at the timeof injury is a crucial factor in determining the overallimpact of the insult. Essentially, the recovery processafter TBI relies heavily on the brains reserve capacityto endure the insult and then to repair itself [3]. Thisnotion is supported by the study conducted by

Correspondence: Amanda Fuentes, York University, Toronto, Ontario M3J 1P3, Canada. Tel: 647-293-4255. E-mail: amanda19@yorku.ca

ISSN 02699052 print/ISSN 1362301X online 2010 Informa Healthcare Ltd.DOI: 10.3109/02699052.2010.489791

Kesler et al. [5], which examined pre-injury and post-injury cognitive status of adult TBI survivors. In thisstudy, scores from standardized academic testingwere used as an index of pre-morbid ability and thencompared to post-injury IQ scores. In conjunctionwith the reserve hypothesis, the findings illustratedthat individuals with lower pre-morbid academicability showed greater vulnerability to the impact ofthe brain injury [5]. The concept of reserve alsocomes into play later in the lifespan after TBI, assurvivors have shown an increased risk of developingpsychiatric and neurodegenerative disorders [3, 5].In this way, brain injury is viewed as a risk factor thatdecreases reserve and, as such, increases vulnerabilityfor the onset of dementia.

Similarly, developmental research has shown thatCR is an important moderator of outcomes followingTBI in children [6, 7]. For example, Fay et al. [6]investigated post-insult cognitive ability as a proxy ofcognitive reserve to predict outcome after childhoodmild TBI. Fay et al. [6] hypothesized that cognitivereserve would moderate the occurrence of post-concussive symptoms attributable to mild TBI, suchthat outcome would be poorer among children oflower cognitive ability than higher cognitive ability(as assessed by a battery of standardized testsadministered within 3 weeks post-injury).Cognitive ability for each child was represented bya single overall composite score, which was deter-mined by computing the mean standard score acrossa variety of cognitive tests. Post-concussive symp-toms were assessed using the Post-concussiveSymptom Interview (PCS-I) and the Health andBehaviour Inventory (HBI). Hierarchical linearmodelling revealed that children with lower cognitivereserve capacity and complicated mild TBI showedgreater vulnerability to the development of post-concussive symptoms than children with highercognitive reserve capacity. Thus, the results sug-gested that outcome was moderated jointly by CRand injury severity. In a study that illustrated theinfluence of pre-injury abilities on outcomes follow-ing childhood brain insult, Farmer et al. [7] exam-ined memory functioning in children with varyingdegrees of brain insult who did or did not havepreviously identified learning problems. Childrenwith prior learning difficulties performed signifi-cantly worse on measures of memory functioningcompared to their peers with no prior learningproblems. These results were consistent with thehypothesis that pre-injury learning problems mod-erated the effects of TBI. Interestingly, children withprior learning problems demonstrated more cogni-tive weaknesses even when severity of injury and agefactors were held constant. Hence, the variability inneuropsychological outcomes of childhood TBI maybe explained in terms of differences in levels of CR.

Studies of neuropsychological outcome of child-hood TBI highlight developmental factors thatinteract with pre-injury cognitive ability to definelong-term recovery of brain insult. There is clearevidence that an earlier age at injury is moredetrimental to outcome than an injury suffered atan older age. To illustrate, greater vulnerability fordifficulties in language, intelligence and motor skillsare reported in children injured before 5 years ofage [8]. Additionally, younger children with TBIdemonstrate a longer recovery interval [8]. Denniset al. [2] suggest that injury to the immature brainadversely affects the development of cognitive skills.In this way, earlier age at injury diminishes CR bypreventing the child from acquiring efficientcognitive strategies that may have otherwise beenrecruited to maintain function after brain insult [2].Clearly, these empirical findings support the theorythat brain injury affects the adult brain differentlythan the child brain. The relationship between CRand neuropsychological outcome would also beexpected to differ since children have had less timeto develop such reserve. As such, it remains unclearwhether the measures used to assess cognitivereserve in adult populations would be reliable andvalid indicators in a paediatric population.

An accurate estimation of pre-morbid cognitivefunctioning is a crucial step in interpreting neuro-psychological assessments [9]. In an attempt toaccomplish this goal, many researchers have utilizedpast achievement information such as standardizedtests (e.g. IQ tests, aptitudes tests, the GraduateRecord Examination, etc.) [9]. However, thismethod of estimation is limited by the lack ofavailability of this information. Other variablesthought to represent reserve include indirect achieve-ment information, such as grade point average, levelof educational attainment and occupation [10]. Pre-morbid ability has also been inferred by examiningmeasurements of neural activity and brain size [10].However, Dennis et al. [2] point out that thismethodology is generally not employed with childrenbecause head and brain size change throughoutdevelopment. Demographic regression estimationmethodology has also been employed to predictscores, but has proven to be ineffective in providingaccurate estimates of individuals [11].

Finally, a useful alternative approach of estimationemploys what is referred to as hold tests. Thismethod relies on the assessment of abilities that areconsidered to be stable and resistant to braindamage [11]. It has been suggested that verbal abilityrepresents a hold measure because of its relativeresistance to cerebral insult [12]. Hence, verbalability constitutes a measure of crystallized intelli-gence. Manly et al. [13] support the significance ofliteracy level as a means of measuring reserve by

996 A. Fuentes et al.

affirming that literacy measures educational experi-ence more accurately than years of education, andthus is a superior assessment of the knowledge,strategy, and skills needed to perform well ontraditional neuropsychological tasks (p. 227). As aresult, various standardized reading tests have beenshown to be valid indicators of pre-injury functioningand cognitive reserve, including the Wide RangeAchievement Test-Third Edition (WRAT-3) [14]and the North American Adult Reading Test(NAART) [15]. The Wechsler Adult IntelligenceScale (WAIS) and its subsequent revisions (includingthe current WAIS-III) have also been viewed as ahold measure [11]. Vanderploeg et al. [9] reportedthat the Information and Vocabulary sub-tests werethe most accurate indicator of reserve. Ruff et al. [16]also investigated vocabulary as a predictor of func-tional outcome following head trauma in a severegroup between 1265 years and found that intactvocabulary was associated with return to work/school.

Past studies have documented mixed evidence forthe usefulness of verbal ability as an estimator ofcognitive reserve in children. Chapman andMcKinnon [8] examined 400 children with severeclosed head injuries and found that recovery of basiclanguage skills, vocabulary and syntactic abilitiesoccurred during the first 3 months post-injury,regardless of age at insult. Chapman andMcKinnon [8] suggested that the recovery ofhigher-order language functions (such as seman-tics, syntax, etc.) may be accomplished through thebrains ability to compensate for early injuries usingthe recruitment of new nerve cells. Conversely, otherresearchers have reported that literacy is particularlyvulnerable following childhood TBI [17].Specifically, children with early head injuries(before 6.5 years of age) perform more poorly onstandardized tests of reading than children withsimilar injuries suffered at an older age [17]. Barneset al. [17] suggested that a head injury sustainedbefore formal reading instruction constitutes a sig-nificant risk factor for difficulties in acquiring basicreading skills. In this way, age at injury was proposedto primarily affect the acquisition of new skills [17].Catroppa et al. [18] investigated the development ofreading skills post-TBI, from time of injury to7 years post-injury, in a mild-to-severe group injuredbetween 312 years. Their results showed thatreading ability was more compromised in childreninjured between 37 years of age than in childreninjured between 812 years. Thus, past research hasidentified reading difficulties following TBI in pre-school or early primary grades, however, little isknown regarding the resilience of these skills follow-ing TBI in older childhood.

In summary, past evidence indicates that thepicture of post-injury literacy skills as a proxy ofCR is not straightforward because recovery afterbrain injury relies on the developmental stage of theindividual at the time of injury. Within the context ofTBI, both literacy and hold tests have become wellrecognized as indicators of pre-morbid functioningwithin adult populations. Although past studies havedemonstrated recovery of language functions inolder children, children injured at a younger agehave been reported to have persisting literacy diffi-culties [8, 17]. At the same time, injury severity doesnot account for all variance in cognitive outcome, asresearchers have reported only modest contributionsof traditional indices of severity (such as GCSratings) to neuropsychological performance post-injury [2, 3, 8]. Given the importance of identifyingchildren who are at risk for negative outcomesfollowing TBI, research must address whether liter-acy tests are indeed valid estimates of cognitivereserve in children. Since previous studies havefound that literacy skills in children who havesustained an injury before 6.5 years are vulnerableto such injuries, only children 6 years old or olderwere used in this study. The overall purpose of thisstudy is to examine cognitive reserve in paediatricbrain injury and its relation to neuropsychologicaloutcome. Specifically, it will examine (a) if therelationship between cognitive and neuropsycholog-ical outcome observed in adult TBI translates topaediatric populations and (b) if well recognizedindicators of cognitive reserve, such as hold tests andliteracy tasks, are valid indicators in paediatricpopulations. To determine the possible efficacy ofthese hold tests, they will be compared to a well-recognized injury severity indicator, the GlasgowComa Scale (GCS) to see if cognitive reserve relatesto neuropsychological outcome as well, or betterthan, injury severity itself.

Method

Participants

The data used in this study consisted of 52 TBI cases36 months post-injury, meeting certain inclusion/exclusion criteria, referred to the neuropsychologyservice of a non-acute paediatric inpatient/daypatientneurorehabilitation hospital located in a large urbanarea in Canada. This hospital serves as the primaryrehabilitation hospital for the province of Ontarioand thus individuals were residents of a wide varietyof urban and rural areas across the province.Inclusion criteria required a medically documentedhistory of TBI, age between 616, current need forneuropsychological assessment and conventionalmedical signs of brain injury (e.g. loss of

Cognitive reserve in paediatric TBI 997

consciousness and post-traumatic amnesia) docu-mented in the medical record. Exclusion criteriaincluded history of pre-morbid developmental dis-ability, learning disability, autism spectrum disorder,attention deficit disorder or any other neurologicalimpairment or disorder (e.g. seizure disorder, previ-ous TBI, etc.). Data was also excluded for cases inwhich English was a second language, if the primarycognitive reserve measures (see Materials section)were not administered or when the results of theassessment were determined to be either invalid orunreliable by the treating neuropsychologist (asdocumented in the neuropsychological report) sec-ondary to inconsistent or questionable motivationand/or effort.

In terms of demographic variables, the group wascomprised of 34 males and 18 females, with anaverage age of 11.81 years (ranging again from616). The large majority of the children were righthanded (83%). A more detailed description ofdemographic characteristics of the group is pre-sented in Table I.

Overall, the group consisted of predominantlymoderate-to-severe brain injuries (see Table II).Specifically, admission Glasgow Coma Scale (GCS)[14] was indicative of moderate or severe degree ofbrain injury (513) in 70% of the cases. Fifty of 52cases revealed cerebral insult on neuroimaging, eventhose rated as mild severity according to the GCS. Infact, of the 15 cases of mild brain injury (as definedby a GCS score of 13 or greater), 14 of them hadpositive neuroimaging. Diffuse axonal injury wasreported in 27% of cases. In the 50 cases in whichmedical records included information regardingpost-injury seizure activity, 42 of them did nothave any documented seizure activity.

Materials

The dependent measures in this study were derivedfrom a variety of neuropsychological measures.Cognitive reserve was measured using the Word

Reading sub-test of the Wechsler IndividualAchievement Test-Second Edition (WIAT-II) [15]and the Vocabulary sub-test of the WechslerIntelligence Scale for Children-Fourth Edition(WISC-IV) [16]. Although there has not beenany direct research examining the WISC-IV andWIAT-II as indicators of pre-morbid functioning,they were chosen based on the aforementionedliterature that used comparable indices of literacyand vocabulary knowledge (e.g. WRAT-3,WAIS-III) as indicators of pre-injury functioning inadult TBI populations [8, 10]. The neuropsycholog-ical measures used in this study were those com-monly used as part of the neuropsychological batteryassessing the following cognitive domains: sustainedand divided attention (Test of Everyday Attention forChildren (TEA-Ch): Sky Search Attention ScaledScore (visual scanning and attention), ScoreScaled Score (sustained auditory attention), SkySearch Dual Task Decrement Scaled Score (multi-modal divided attention) and Score Dual Task Score

Table II. Injury characteristics of traumatic braininjury (TBI) participants (n52).

Count (%)

Cause of head traumaMVA (car only) 24 (46)MVA w/pedestrian or bike 15 (29)Bike/recreation 5 (10)Fall 2 (4)Boating 2 (4)Other 4 (8)

ER Admission GSCSevere (38) 25 (48)Moderate (912) 10 (19)Mild (1315) 15 (29)Unknown 2 (4)

NeuroimagingPositive 50 (96)Negative 2 (4)

LateralityLeft 13 (25)Right 11 (21)Bilateral 21 (40)Unknown 7 (14)

LobesFrontal 14 (27)Multiple 26 (50)Other 6 (12)Unknown 6 (12)

Diffuse Axonal InjuryPositive 14 (27)Negative 37 (71)Unknown 1 (2)

SeizureNegative 42 (81)Positive 8 (15)Unknown 2 (4)

MVAMotor vehicle accident; GCSGlasgowComa Scale.

Table I. Demographic characteristics of the TBIsample (n 52).

Count (%)

AgeMean (SD) 11.81 (3.13)Range 616

GenderMale 34 (65)Female 18 (35)

Handed dominanceRight 43 (83)Left 4 (8)Unknown 5 (9)

998 A. Fuentes et al.

(unimodal divided attention)) [17], visual-motorprocessing speed (WISC-IV Processing SpeedIndex score, Digit Symbol Coding Scaled Scoreand Symbol Search Scaled Score) [16], learning andmemory (California Verbal Learning Test-Childrens version (CVLT-C) List A Total T-score[18]; Rey Complex Figure Test (RCFT) ImmediateRecall T-score and Delayed Recall T-score) [19] andexecutive functioning (Wisconsin Card Sorting Test(WCST) Total Correct Raw Score [20]; ChildrensCategory Test (CCT) Total T-score) [21].

Results

To determine if there was a relationship betweenpre-morbid functioning (as measured by theWIAT-II Word Reading and WISC-IV Vocabularyscores) and neuropsychological outcome, Pearsoncorrelations were used. Neither the WIAT-II WordReading nor WISC-IV Vocabulary scaled scoreswere significantly correlated to the majority of theneuropsychological measures (see Table III).

In terms of WIAT-II Word Reading performance,it was only significantly correlated with the dividedattention measures (i.e. Dual Task scores) of theTEA-Ch. It was not significantly correlated with anyof the measures related to visual-motor processingspeed, memory, basic attention or executivefunctioning. Similarly, WISC-IV Vocabulary perfor-mance failed to relate to any of the neuropsycholog-ical measures, with the exception of verbal learningand memory (CVLT-C List A Total T-score). Injuryseverity, as measured by GCS score, was correlatedwith measures of divided attention (TEA-Ch SkySearch DT and Score DT) and processing speed

(WISC-IV Processing Speed Index, Coding sub-testand Symbol Search sub-test). Since neither of thecognitive reserve variables were significantly corre-lated to the neuropsychological outcome measures,regressional analyses were not conducted.

Discussion

The objective of the current study was to determinethe utility of reading and vocabulary tests asmeasures of pre-morbid ability for children withTBI. Specifically, two main research questions wereposed: (1) does the relationship between cognitiveand neuropsychological outcome observed in adultTBI translate to paediatric populations?; and (b) Dowell recognized indicators of cognitive reserve, suchas hold tests and literacy tasks, remain valid indica-tors in paediatric populations? Overall, the findingsshowed that neither reading ability (WIAT-II WordReading sub-test score) nor vocabulary skills(WISC-IV Vocabulary sub-test score) were relatedto levels of neurocognitive functioning followingTBI, with the exception of isolated divided attentiontasks. In comparison, moderate correlations werefound between Glasgow Coma Score (GCS) andneuropsychological measures of divided attentionand processing speed, indicating that severity ofinjury still serves as a modestly stronger, yet stillunreliable, predictor of neuropsychological outcome.Taken together, these findings suggest that themoderating effect of cognitive ability on neuropsy-chological outcome in adult TBI may not directlyapply to the paediatric population. Consequently,the results provide evidence against the validityof the traditional reserve proxy of reading

Table III. Pearson correlations (r2) between neuropsychological measures and WIAT-II word reading, WISC-IV vocabularyand GCS (respective sample sizes contained in brackets).

MeasureWIAT-II Word Reading WISC-IV Vocabulary Glasgow Coma Scale (GCS)

r2 (n) r2 (n) r2 (n)

TEA-ChSky search scaled score 0.138 (36) 0.025 (46) 0.124 (44)Score scaled score 0.108 (36) 0.050 (46) 0.062 (44)Sky search DT scaled score 0.388 (36)* 0.053 (46) 0.299 (44)*Score DT scaled score 0.449 (36)** 0.261 (45) 0.497 (43)**

CVLT-C List A Total T-score 0.225 (41) 0.295 (48)* 0.253 (46)RCFT

Immediate T-score 0.138 (33) 0.238 (37) 0.129 (36)Delay T-score 0.061 (33) 0.220 (37) 0.140 (36)

WISC-IVPSI 0.261 (42) 0.141 (52) 0.431 (50)**Coding scaled score 0.227 (42) 0.140 (52) 0.379 (50)**Symbol search scaled score 0.302 (42) 0.166 (52) 0.418 (50)**

WCST # correct raw score 0.093 (41) 0.054 (49) 0.141 (47)CCT Total T-score 0.270 (40) 0.229 (47) 0.313 (45)*

*p50.05; **p50.0, DTDual task/decrement score.

Cognitive reserve in paediatric TBI 999

ability/vocabulary when applied to the paediatricpopulation.

The current pattern of findings is inconsistentwith the threshold theory of cognitive reservedescribed by Stern [3]. According to the thresholdtheory, individuals with higher cognitive reserve aresomewhat protected from the effects of brain pathol-ogy, as they must sustain higher levels of pathologybefore performance is affected [3]. In this study,there was no relationship between neuropsycholog-ical outcome following TBI and cognitive reserve (asmeasured by verbal ability). The lack of associationbetween pre-morbid cognitive ability and outcomesfollowing childhood TBI is somewhat surprising,given that previous research has suggested that lowerCR is a significant risk factor for poorer neuropsy-chological outcome in both children and adults[57]. The failure of this study to find a relationbetween CR in children and neuropsychologicaloutcome following TBI may be due to the fact thatverbal ability was used as an indicator of CR. To theauthors knowledge, the current study is the first toassess the relationship between pre-morbid verbalability and subsequent neuropsychological outcomesfollowing TBI in children. It may be the case that achilds level of CR is not synonymous with theirreading abilities and vocabulary knowledge. In otherwords, vocabulary and reading may not be stable,crystallized measures in children as they are inadults. This would serve to undermine the effective-ness of reading and vocabulary abilities aspre-morbid predictors because they would not beresilient to the adverse effects of TBI. This conclu-sion is supported by past research describing thecomparability of a pure demographic approach withthe hold test prediction method in children [9]. Theactive model [3] of CR is grounded in the notionthat the brain actively attempts to compensate forthe disruption represented by TBI by utilizing pre-existing cognitive processing strategies. However, aspreviously suggested by others [4], these cognitiveabilities may not be fully developed in children at thetime of cerebral insult. Consequently, children maylack the cognitive strategies that are necessary tocounter against the damaging effects of TBI. Hence,developmental factors may exert a significant influ-ence on cognitive functioning after TBI by renderingthe child less capable of responding to the challengeof TBI. This implication is supported by a numberof studies demonstrating that early age at insult isassociated with more detrimental and persistingeffects in a number of cognitive domains [2830].In particular, previous research has documentedclear evidence that childhood TBI represents a riskfactor for difficulties in acquiring basic word decod-ing skills [17]. This observation is particularly

relevant to the present study, given that the meanage of participants was 11.8 years (range 616).

Alternatively, it may be the case that the levels ofperformance exhibited by children on verbal mea-sures do reflect crystallized abilities (and are thereforeaccurate markers of CR), but simply went undetectedin the current study. This explanation is supported byprevious research reporting the resilience of verbalabilities to the impact of brain pathology in childhoodTBI [8]. For example, the results from this studydiffer from the findings of Chapman and McKinnon[8], who observed good recovery of basic language,vocabulary and syntactic abilities in children withsevere head injuries 3 months post-injury. Thediffering results may be due to the fact that theChapman and McKinnon [8] study used a sample of400 children, whereas the current study examined asmall sample of children (n 52). The use of a smallsample size may have limited the power of the currentstudy to detect effects of CR. Finally, it remainspossible that the laterality of the damage from thehead injury may have affected reading and vocabularyskills (and may have therefore influenced CR). Forexample, Barnes et al. [17] observed that poorerreading was associated with contusions involving theleft side of the brain. Once again, there was aninsufficient number of participants to provide infor-mation on the relation between laterality and CR, asmeasured by reading and vocabulary.

Several caveats to the present study must be noted.First, an inherent weakness exists in the task ofdeveloping pre-morbid prediction strategies in clin-ical samples. Clearly, the best way to determine pre-injury functioning is to have completed baselineneuropsychological testing to which post-injury testscores can be directly compared. However, accuratemeasures of pre-morbid ability (such as standardizedtest data) are typically not present and were notavailable in the clinical sample of the present studyeither. Nevertheless, it remains suggested that onlyfluid skills are likely to be impacted by the negativeeffects of TBI, while crystallized abilities should beresistant to these effects [12]. Given that vocabularyand reading are considered to be crystallized abilities,they were expected to be minimally affected by brainpathology and were therefore utilized to estimateoverall levels of pre-morbid cognitive function.Although a number of past studies have validatedthis hold test approach in pre-morbid estimation foradults [911], brain damage may have affectedsimilar verbal hold measures in children, indicatingthe influence of developmental factors. A shortcom-ing of the present study is that it was not possible tocontrol for important differences in brain recoverythat may have been present among participants dueto the confounding variable of developmental plas-ticity. For example, past studies have also reported

1000 A. Fuentes et al.

that level of education is predictive of outcomefollowing TBI [5]. This raises a difficulty for thecurrent study, as younger participants may have beenmore vulnerable to the functional impact of TBI dueto their reduced number of years of education andless stabilized reserve. Further, it could be arguedthat a more accurate and possibly more clinicallymeaningful approach to assessing pre-morbid abilitywould be to examine the differences in neuropsycho-logical outcomes between high- and low-reservegroups. However, due to the small number ofparticipants (n52) there was no clear way ofdeveloping cut-off requirements that would serve toaccurately divide participants into clinically mean-ingful groups. Finally, given the moderating effect ofage at onset of TBI and developmental plasticity, itcould also be argued that a better understanding ofCR in children would have been achieved by dividingthe participants into age groups. Again, this was notpossible due to the small sample size.

In summary, the absence of any significant corre-lations between the Verbal IQ measures and neuro-psychological scores post-injury suggests that theseabilities are not valid indicators of cognitive reservein the paediatric TBI population. This findingemphasizes the importance of developing an inte-grated approach to studying how the neuropsycho-logical outcome of childhood TBI is mediated byCR. In turn, this will help to explain the indirect linkthat has been reported between severity of insult andclinical outcome. Developing a more comprehensiveinterpretation of the outcome of brain injuries inchildren requires teasing apart the components thatunderlie the recovery process of TBI in early life.Given that the current measures may not be validindicators of CR, then it is important to determinewhich measures are valid indicators in paediatricpopulations. Given what is known about the adultliterature, it is possible that socio-economic status(SES) would be related to recovery and outcome. Assuch, it is possible that parental education levels maybe predictive of childhood outcome more so than achilds educational abilities like reading and vocab-ulary [31, 32]. In order to define key areas of familyfunction that may serve as useful proxies of CR,future studies should investigate associationsbetween parental unemployment, parental conflictand parent mental health and neurocognitive out-comes. To this end, research examining the differ-ential outcomes of high- and low-reserve childrenand high- and low-reserve parents will prove to beuseful in determining whether or not children withsimilar levels of neuropathology experience similarlevels of impairment. Similarly, in order to gainfurther insight regarding the meaning of age at onsetof TBI and the effects for reading and vocabularyskills, it would be important to investigate

differential outcomes among younger and olderchildren. Advances in paediatric neuroimaging pro-vide the opportunity to investigate possible differ-ences in neural processing that may exist betweenhigh- and low-reserve children and between youngerand older children. Recent research of cognitivefunction in children survivors of TBI has demon-strated that neuropsychological outcome is alsomoderated by time since onset of CNS insult [33].For example, even when severity of head injury isheld constant, younger children have been observedto show a slower rate of recovery over time [33].Longitudinal studies of cognitive function of chil-dren survivors of TBI may therefore prove useful inproviding a more comprehensive view of develop-ment post-injury. The mental health of the child mayalso serve to buffer or compromise neuropsycholog-ical outcome post-insult. Hence, it would be usefulto explore the association between outcome and pre-injury or post-injury psychiatric disorders in chil-dren. Ultimately, the findings reported in the presentstudy support the current conception of CR as adynamic process that is modifiable at different stagesof the lifespan [10].

Declaration of interest: The authors report noconflict of interest. The authors alone are respon-sible for the content and writing of the paper.

References

1. Bieliaukas LA, Antonucci A. The impact of cognitive reserveon neuropsychological measures in clinical trials. In: Stern Y,editor. Cognitive Reserve: Theory and Applications.New York: Taylor & Francis; 2007. p 131142.

2. Dennis M, Yeates KO, Taylor HG, Fletcher JM. Brain reservecapacity, cognitive reserve capacity, and age-based functionalplasticity after congenital and acquired brain injury in children.In: Stern Y, editor. Cognitive Reserve: Theory andApplications. New York: Taylor & Francis; 2007. p 5384.

3. Stern Y. The concept of cognitive reserve: A catalyst forresearch. In: Stern Y, editor. Cognitive Reserve: Theory andApplications. New York: Taylor & Francis; 2007. p 14.

4. Bigler ED. Traumatic brain injury and cognitive reserve. In:Stern Y, editor. Cognitive Reserve: Theory and Applications.New York: Taylor & Francis; 2007. p 85116.

5. Kesler SB, Adams HF, Blasey CM, Bigler ED. Premorbidintellectual functioning, education, and brain size in traumaticbrain injury: An investigation of the cognitive reserve hypoth-esis. Applied Neuropsychology 2003;10:153162.

6. Fay TB, Yeates KO, Taylor HG, Bangert B, Dietrich A,Nuss KE, Rusin J, Wright M. Cognitive reserve as a moderatorof postcuncussive symptoms in children with complicated anduncomplicated mild traumatic brain injury. Journal of theInternational Neuropsychological Society 2010;16:94105.

7. Farmer JE, Kanne SM, Haut JS, Williams J, Johnstone B,Kirk K. Memory functioning following traumatic brain injuryin children with premorbid learning problems. DevelopmentalNeuropsychology 2002;22:455469.

Cognitive reserve in paediatric TBI 1001

8. Chapman SB, McKinnon L. Discussion of developmentalplasticity: Factors affecting cognitive outcome after pediatrictraumatic brain injury. Journal of Communication Disorders2000;33:333344.

9. Vanderploeg RD, Schinka JA, Baum KM, Tremont G,Mittenberg W. WISC-III premorbid prediction strategies:Demographic and best performance approaches.Psychological Assessment 1998;10:277284.

10. Richards M, Deary IJ. A life course approach to cognitivereserve: A model for cognitive aging and development?American Neurological Association 2005;58:617622.

11. Vanderploeg RD, Schinka JA, Axelrod BN. Estimation ofWAIS-R premorbid intelligence: Current ability and demo-graphic data used in a best-performance fashion.Psychological Assessment 1996;8:404411.

12. Richards M, Sacker A. Lifetime antecedents of cognitivereserve. Journal of Clinical and ExperimentalNeuropsychology 2003;25:614624.

13. Manly JJ, Touradji P, Tang M, Stern Y. Literacy andmemory decline among ethnically diverse elders. Journal ofClinical and Experimental Neuropsychology 2000;25:680690.

14. Orme DR, Johnstone B, Hanks R, Novack T. The WRAT-3reading subtest as a measure of premorbid intelligence amongpersons with brain injury. Rehabilitation Psychology 2004;49:250253.

15. Johnstone B, Callahan CD, Kapila CJ, Bouman DE. Thecomparability of the WRAT-R reading test and the NAARTas estimates of premorbid intelligence in neurologicallyimpaired patients. Archives of Clinical Neuropsychology1996;11:513520.

16. Ruff RM, Marshall LF, Crouch J, Klauber MR, Levin HS,Barth J, Kreutzer J, Blunt BA, Foulkes MA, Eisenberg HM,Jane JA, Marmarou A. Predictors of outcome following severehead trauma: Follow-up data from the Traumatic ComaData Bank. Brain Injury 1993;7:101111.

17. Barnes MA, Dennis M, Wilkinson M. Reading after closedhead injury in childhood: Effects on accuracy, fluency, andcomprehension. Developmental Neuropsychology 1999;15:124.

18. Catroppa C, Anderson VA, Muscara F, Morse SA,Haritou F, Rosenfeld JV, Heinrich LM. Educational skills:Long-term outcome and predictors following paediatrictraumatic brain injury. Neuropsychological Rehabilitation2009;19:716732.

19. Yeates KO, Taylor HG, Wade SL, Drotar D, Stancin T,Minich N. A prospective study on short- and long-term

neuropsychological outcomes after pediatric traumatic braininjury. Neuropsychology 2002;16:514523.

20. Teasdale G, Jennett B. Assessment of coma and impairedconsciousness: A practical scale. Lancet 1974;2:8184.

21. Wechsler D. Wechsler Individual Achievement Test(WIAT-II). 2nd ed. San Antonio, Texas: The PsychologicalCorporation; 2001.

22. Wechsler D. Wechsler Intelligence Scale for Children(WISC-IV). 4th Edition: American Manual. San Antonio,Texas: The Psychological Corporation; 2003.

23. Manly T, Robertson IH, Anderson V, Nimmo-Smith I. TheTest of Everyday Attention for Children. England: ThamesValley Test Company; 1999.

24. Fridlund AJ, Delis DC. California Verbal Learning Test:Childrens Version. Texas, USA: The PsychologicalCorporation; 1994.

25. Meyers JE, Meyers KR. Rey Complex Figure Test andRecognition Trial. Lutz, Florida: Psychological AssessmentResources, Inc; 1995.

26. Heaton RK, Chelune GJ, Talley JL, Kay GG, Curtiss G.Wisconsin Card Sorting Test. Lutz, Florida: PsychologicalAssessment Resources, Inc; 1981.

27. Boll T. Childrens Category Test. USA: The PsychologicalCorporation; 1993.

28. Anderson V, Moore C. Age at injury as a predictor ofoutcome following pediatric head injury: A longitudinalperspective. Child Neuropsychology 1995;1:187202.

29. Anderson VA, Morse SA, Klug G, Catroppa C, Haritou F,Rosenfield J, Pentland L. Predicting recovery from headinjury in young children: A prospective analysis. Journal ofInternational Neuropsycholgoical Society 1997;3:568580.

30. Dennis M, Barnes MA, Wilkinson M, Humphreys RP. Howchildren with head injury represent real and deceptiveemotion in short narrartives. Brain and Language 1998;61:450483.

31. Kaplan GA, Turrell G, Lynch JW, Everson SA, Helkala EL,Salonen J. Childhood socioeconomic position and cognitivefunction in adulthood. International Journal of Epidemiology2001;30:256263.

32. Capron D, Petrill SA, Pike A, Price T, Plomin R. Chaos inthe home and socioeconomic status are associated withcognitive development in early childhood: Environmentalmediators identified in a genetic design. Intelligence 2004;32:445460.

33. Dennis M. Developmental plasticity in children: The role ofbiological risk, development, time, and reserve. Journal ofCommunication Disorders 2000;33:321332.

1002 A. Fuentes et al.

Copyright of Brain Injury is the property of Taylor & Francis Ltd and its content may not be copied or emailed

to multiple sites or posted to a listserv without the copyright holder's express written permission. However,

users may print, download, or email articles for individual use.