CYBERPSYCHOLOGY & BEHAVIOR

Volume 12, Number 2, 2009© Mary Ann Liebert, Inc.DOI: 10.1089/cpb.2008.0204

A Virtual Reality Test Identifies the Visuospatial Strengths of Adolescents with Dyslexia

Elizabeth A. Attree, BSc., C.Psychol.,1 Mark. J. Turner, M.Sc.,1 and Naina Cowell, M.A.2

Abstract

Research suggests that the deficits characterizing dyslexia may also be associated with superior visuospatialabilities. Other research suggests that superior visuospatial abilities of people with dyslexia may not have beenso far identified because of the lack of appropriate tests of real-life spatial ability. A recent small-scale studyfound that visuospatial superiority was evident in men with dyslexia. This study assessed the visuospatial abil-ity of adolescents with dyslexia in order to determine whether these adolescents performed better on a pseudoreal-life visuospatial test than did their nondyslexic peers. Forty-two adolescents took part in the study. Therewas an equal numerical split between the experimental and control groups. The experimental group all had adiagnosis of dyslexia by an educational psychologist or specialist teacher. Visuospatial ability was assessed us-ing the Recall of Designs and the Pattern Construction subtests from the British Ability Scales (2nd edition;BAS-11) together with a computer-generated virtual environment test. The assessments were administered ina counterbalanced order. Adolescents with dyslexia tended to perform less well than their nondyslexic peerson the BAS-11 tests; however, this difference was not statistically significant. For the computer-generated vir-tual environment test (pseudo real-life measure), statistically significant higher scores were achieved by thedyslexic group. These findings suggest that adolescents with dyslexia may exhibit superior visuospatialstrengths on certain pseudo real-life tests of spatial ability. The usefulness of these findings is discussed in re-lation to possible implications for assessment and educational intervention programs for adolescents withdyslexia.

163

Introduction

DYSLEXIA IS A SPECIFIC LEARNING DISABILITY that is “evidentwhen accurate and fluent word reading or spelling de-

velops very incompletely or with great difficulty.” This def-inition focuses on literacy learning at the word level, imply-ing a severe and persistent problem despite appropriatelearning opportunities.1 If taken as descriptive rather thanoperational, this definition provides a neutral starting pointfrom which different theoretical rationales and research ini-tiatives can be considered. It enables the examination of as-sociated difficulties, whether cognitive, emotional, or phys-ical, in order to specify individual educational needs.2

The etiology of dyslexia is not fully understood. However,neurological explanations such as Nicolson’s3 hypothesisthat congenital cerebellar dysfunction provides a coherentexplanation of all the symptoms of dyslexia, including the“core” phonological difficulties, ties in with similar theoriesdeveloped by Frith.4

Atypical findings in the brains of people with dyslexia haveled to speculation that they might not just lead to the condi-tion itself but to the enhancement of other cognitive functions.5

If this speculation can be substantiated, then increased knowl-edge and understanding of enhanced cognitive functions inindividuals with dyslexia could have important implicationsfor teaching and learning.6 The specific cognitive domainwithin which these purportedly enhanced functions existshould be relatively straightforward to identify, because theauditory-verbal and the visuospatial information processingsystems are neuropsychologically partially separate and dis-tinct networks within the brain.7 For example, one cognitivefeature of dyslexia that has long been thought to be univer-sally superior is visuospatial ability. Case studies of talentedindividuals with dyslexia8 and research studies9,10 suggestthat dyslexia may be positively associated with visuospatialabilities, artistic talents, or creativity.11 Research also indicatesthat the regions of the brain associated with visuospatial abil-ities are larger in individuals with dyslexia.12

1School of Psychology, University of East London, London, United Kingdom.2Multi-Agency and Psychology Services, Children, Schools and Families, Hertfordshire County Council, United Kingdom.

ATTREE ET AL.164

Nevertheless, in terms of relative strengths in relation tospatial versus verbal abilities, strengths in the spatial areasfor individuals with dyslexia exist, but evidence for absolutestrengths compared to the general population is not consis-tent. Some suggest relative strengths in visuospatial ar-eas,13,14 while others indicate spatial skills that are neitherinferior nor superior to the rest of the population.15–17 Ex-planations for such talents among people with dyslexia havebeen proposed using a compensatory argument. These dis-tributions of talents may represent compensation contribut-ing to the evolutionary resistance of dyslexic genes.11 Analternative explanation is the default and channeling hy-pothesis,18 which suggests that individuals with dyslexiawho have enhanced spatial talents are more likely to choosespatial as opposed to verbal occupations because the choiceof the latter avenue may not be open to them. This does notnecessarily suggest any difference in spatial abilities betweenthose with or without dyslexia.

Moreover, Winner et al.18 found that individuals withdyslexia performed either equivalently to or worse than con-trol peers on a wide range of visuospatial tasks. Further re-search was suggested using more real-life tests of spatial abil-ity to determine whether dyslexia is associated with spatialability. Visuospatial ability has been given only token atten-tion as an important dimension of cognitive functioning andoften has been ignored by the psychological and educationalcommunity.8 The paucity of research in this area may be due,in part, to a widely agreed definition of “visuospatial abili-ties.” Indeed, they are within a somewhat complex cognitivedomain generally related to the broader field of “intelli-gence.” For example, Carroll’s19 description of intelligenceand associated cognitive skills derive from a multifactorialtheory of intelligence. He proposed multiple dimensions ofhuman abilities, which are related in complex ways to learn-ing, achievement, and problem-solving.

Carroll’s19 factor analytical research identified five majortypes of visuospatial abilities: (a) spatial orientation (the abil-ity to imagine how an image will appear from another per-spective), (b) spatial visualization (the ability to apprehend,encode, and mentally manipulate spatial forms), (c) figuralflexibility (the ability to come up with a variety of ways toinvolve a spatial problem), (d) closure speed (the ability toinitiate an apparently disparate perceptual field into a uni-fied concept), and (e) reference memory (which assesses spa-tial memory).

In order to investigate visuospatial abilities in individualswith dyslexia, it is important to assess a range of abilities, asidentified by Carroll.19 In summarizing their inconclusive re-search, Winner et al.18 asked whether real-world spatial tasksmight indicate more readily observable spatial talents in in-dividuals with dyslexia. Evidence has emerged from a small-scale study carried out by Brunswick and Martin20 thatfound that men with dyslexia had a greater level of accuracyon a real-life task. Hence, before dismissing an associationbetween dyslexia and visuospatial talents, more research isneeded using real-life tests of spatial ability. These ideas in-spired this research study.

This study used a virtual real-world task alongside stan-dardized psychometric tests measuring a range of visu-ospatial abilities to explore whether the performance of ado-lescents with dyslexia was in any way superior to theirnondyslexic counterparts. If superior visuospatial abilities

are identified, this will not only have implications for as-sessment and intervention for people with dyslexia but willalso help to bridge the gaps between pure psychological re-search with outcomes relevant to the pragmatic concerns ofdyslexic individuals themselves and those who supportthem.

Based on the proposal that individuals with dyslexia mayexhibit enhanced visuospatial abilities on assessments thatreflect real-world tasks, we hypothesized that the dyslexicgroup would score significantly higher than controls on thecomputer-generated pseudo real-life visuospatial measure.Performance on the more traditional assessments of visu-ospatial ability is unlikely to yield differences between thedyslexic and control group, which would be consistent withprevious research.

Materials and Methods

Design

This was a quasi-experimental mixed 2 � 3 design. Thebetween-participants factor had two levels (dyslexic andnondyslexic), and the within-participants factor had threelevels (the three tests). The dependent variable was the per-formance score on each test. Participants were presentedwith a counterbalanced within-participants design with thethree tests: (A) virtual environment (VE) test, (B) PatternConstruction test, and (C) Recall of Designs test. Tests B andC were taken from the British Ability Scales (BAS) II.21 Thetests were administered individually in sessions lasting 30to 40 minutes. All participants were seen within a 4-weekperiod.

Participants

A total of 21 dyslexic and 21 control participants volun-teered to take part in the study. Participants attended a sec-ondary school in either Cambridgeshire or Essex.

The dyslexic group consisted of 18 males and 3 femaleswhose ages ranged from 12 years 9 months to 14 years 6months. Ten (9 male, 1 female) participants attended adyslexia unit attached to an Essex secondary school. The stu-dents attending the dyslexia unit had statements for specialeducational needs22 identifying dyslexia as the primary spe-cial educational need. The remaining participants (9 males,2 females) were on School Action Plus (DfES SEN Code ofPractice 2001), receiving support from the school’s learningsupport unit. They had a diagnosis of dyslexia that had beenconfirmed either by specialist staff at the dyslexia unit or byeducational psychologist assessment. These criteria provideexternal validation of the dyslexia diagnosis.

The control group of 15 males and 6 females had no re-ported history of literacy difficulties, which was confirmedby teacher assessment. Literacy attainments and the meanchronological age (SD in parentheses) of each group were asfollows: spelling age, dyslexic 8.1 (0.992), controls 12.6(1.877); reading age, dyslexic 9.0 (1.578), controls not avail-able; chronological age, dyslexic13.4 (0.717), controls 13.9(0.702).

All participants used English as their first language andhad no known physical or sensory difficulties. Two pupilsfrom the control group used glasses (for reading) and onepupil from the dyslexic group used blue-tinted glasses. The

two schools had similar catchment areas in terms of socioe-conomic status.

Measures and procedure

The tests chosen to investigate this hypothesis assessed themajor types of visuospatial abilities identified by Carroll:19

The VE test measured spatial orientation and visualizationand reference memory. The Pattern Construction test mea-sured spatial visualization, orientation, matching, and anal-ysis; visual holistic information processing; and speed ofprocessing. The Recall of Designs test measured spatial vi-sualization and orientation, visual discrimination of figuresor designs, spatial memory, and visual holistic informationprocessing.

VE test. The VE task was constructed using SuperscapeVRT software. It ran on a Toshiba laptop computer with a17-inch screen, and exploration was controlled by the key-board arrow keys. The environment depicted four intercon-nected rooms in a bungalow: a bedroom, a music room, alounge, and a kitchen. For example, Figure 1 is a black andwhite illustration of the VE bedroom (the actual VE was incolor). Throughout these rooms, 25 household objects weresituated. This VE task was used in several previous stud-ies,eg,23 and has been shown to be a reliable method of iden-tifying spatial recognition skills.

Each participant was seated individually in front of thecomputer. They were shown how to use the arrow keys andgiven the following instructions: “You are taking part in avirtual reality test program. Your task is to use the arrowkeys to find a route through the rooms you are about to seeon the screen. At the same time, I would like you to studythe objects you will see and try to find a toy car.” As par-ticipants entered each room on commencement of the test,the experimenter named the room. If participants did not ap-pear to be searching the rooms properly, they were remindedof their task to find a toy car. This was in the kitchen, thelast room of the VE.

On completion of the VE test, all participants were askedhow many rooms there were in the virtual bungalow, andtheir responses were noted. They then sat in front of a largeboard to perform the spatial layout memory task. This con-sisted of constructing a plan of the virtual bungalow on theboard using shapes from 2D cardboard shapes: squares, rec-tangles presented vertically and horizontally, and circlesplus long rectangular shapes representing corridors. Whenthe participant had constructed the plan of the virtual bun-galow on the board, the experimenter copied it onto the testsheet. The actual VE layout is shown in Figure 2.

Pattern construction test (from the BAS II). Participantswere presented with 2D patterns in a booklet and asked touse yellow and black plastic cubes to construct the designpresented. A stopwatch was used, and timing began the mo-ment the experimenter finished the instruction. A record andnote of the time taken to complete each pattern correctly wasnoted on the record sheet. The timing of responses was ex-plained to participants by saying that they would be timedbut that they needed to work carefully and get the patternsright. The test was administered and scored on correctnessof design and response time for each item using the stan-dardized instructions from the Administration and ScoringManual for the BAS II.24

Recall of designs test (from BAS II). This test was also ad-ministered and scored using the standardized instructionsfrom the Administration and Scoring Manual for the BAS II.24

Participants were presented with an abstract line drawingfor 5 seconds and then asked to draw it from memory. Par-ticipants were presented with Example A initially. After 5seconds, it was removed and participants were presentedwith a sheet of blank paper (with a 10 � 15 cm wide faintgrid) and asked to draw what they saw. Feedback was pro-vided using the example in the booklet. Similar procedureswere adopted for Examples B and C. Participants were thenadministered items 6 to 16 from the booklet using the deci-sion points suggested in the manual. The scoring was based

VR TEST OF VISUOSPATIAL ABILITIES IN DYSLEXIA 165

FIG. 1. Black and white illustration of the VE bedroom.

on whether participants remembered the essential featuresof the design and the item-by-item criteria.

Results

The participants’ spatial layout floor plans from the VEtest were scored according to the scoring system developedby Brooks et al.23 (i.e., participants’ plans of the spatial lay-out were compared to the actual VE layout), which allocatedfour marks each to the correct number and shapes of rooms,entry doorway positions, exit walls, and exit passageway po-sitions that participants correctly identified; this gave a max-imum score of 20.

Each participant’s performance on the Pattern Constructiontest and the Recall of Designs test was scored according to theassessment protocol outlined in the Administration and ScoringManual of the BAS II.24 Two judges scored the data from theRecall of Designs test. An interjudge reliability of 95% wasreached. Raw scores on these two tests were converted to nor-malized, standard, age-adjusted scores (t scores).

Summary statistics for both groups (with and withoutdyslexia) are shown in Table 1. For the spatial recognition

memory test, it can be seen that the children with dyslexiagained the higher scores. Children without dyslexia scoredmarginally higher on the Pattern Construction test and Re-call of Designs test.

Because the scoring of the two types of assessment wasundertaken using different parameters, these were analyzedseparately. First, the spatial memory recognition scores wereconsidered. VE test scores were correlated with chronologi-cal age (p � �0.014), so a univariate ANCOVA, with age asthe covariate, was performed on this measure. The dyslexicgroup scored significantly higher on this test, F(2, 39) �6.077, p � 0.018, partial �2 � 0.14.

The BAS measures were considered next. Data were en-tered into a split-plot two-way ANOVA with BAS spatial testT score (Pattern Construction and Recall of Designs tests) asthe repeated measures factor and groups (dyslexic and con-trols) as the independent factor. There was a significant maineffect of BAS spatial test score, F(1, 40) � 5.027, p � 0.031,partial �2 � 0.112, and condition, F(1, 40) � 3.999, p � 0.052,partial �2 � 0.091. There was no significant interaction be-tween condition and BAS spatial test score, F(1, 40) � 1.744,p � 0.194, partial �2 � 0.042.

ATTREE ET AL.166

TABLE 1. MEANS AND STANDARD ERRORS ON ALL DEPENDENT VARIABLES FOR THE DYSLEXIC AND NONDYSLEXIC GROUPS

Dyslexic Control

Mean SE Mean SE

Spatial recognition memory* 14.61 0.56 12.67 0.45British Ability Scales (t score) Recall of Designs 51.05 2.46 57.76 1.82

Pattern Construction 56.57 2.06 59.19 1.44

*Means for spatial recognition memory test adjusted for chronological age.

FIG. 2. Floor plan of the virtual bungalow.

Post hoc tests were carried out to further examine the dif-ferences. Two sets of paired samples t tests were carried outon the BAS spatial test T scores (Pattern Construction andRecall of Designs tests). A Bonferroni correction adjusted thealpha to a criterion probability value of 0.025. The pairedsamples t test revealed statistically nonsignificant effects forboth the dyslexic group (t20 � 2.168, p � 0.042, d � 0.585)and the control group (t20 � 0.809, p � .428, d � 0.216).

Two sets of independent samples t tests were carried outon the group BAS T scores (dyslexic vs. control). A Bonfer-roni correction adjusted the alpha to a criterion probabilityvalue of 0.025. The independent samples t test revealed sta-tistically nonsignificant effects for both the Pattern Con-struction test T scores (t40 � 1.041, p � 0.304, d � 0.321) andthe Recall of Designs test T scores (t40 � 2.194, p � 0.035, d �0.677).

Discussion

Results showed that there were overall statistical differ-ences between and within the groups in terms of the BASspatial test scores. Looking at the means in isolation, it ap-pears that children with dyslexia performed less well thanchildren without dyslexia. However, on closer examination,no statistical significant differences were found on the re-peated measures factor (Pattern Construction vs. Recall ofDesigns) for either group. Furthermore, children withdyslexia did not score significantly lower than children with-out dyslexia on either the Pattern Construction test or theRecall of Designs test. These findings are commensurate withprevious research10 that performance of individuals withdyslexia on standard paper-and-pencil tests of spatial abil-ity are equivalent to or slightly below the performance of in-dividual without dyslexia.

Interestingly, when the children’s spatial ability was as-sessed on a real-life spatial test, the performance of childrenwith dyslexia was statistically significantly higher than thatof children without dyslexia. Although the findings need tobe replicated, they do lend support to the proposal10 thatdyslexia could possibly be associated with superior perfor-mance on assessments that reflect real-world tasks. The testused in this study was one that assessed incidental spatialmemory, which is more reflective of everyday activity be-cause children (and adults) rarely pay attention to every de-tail of their journey from one place to another. Spatial recog-nition memory is just one aspect of spatial ability, and futureresearch needs to extend real-life assessment by investigat-ing, for example, whether dyslexia is associated with greateruse of environmental cues (e.g., geometric properties or land-marks). It should also be noted that VE exposure time mayinfluence recognition scores; this was not measured in thepresent study. However, all participants took an averagetime of 10 minutes to navigate the VE with no noticeablevariations.

The present findings also need to be considered with re-gard to motivational factors. All the children, irrespective ofgroup, found the computer-generated task more engagingthan the paper-and-pencil task and the puzzle-type task. Itis possible that if interesting and exciting standard spatialassessments could be devised, then children with dyslexiamay show superior performance on standard psychologicaltests.

The study was exploratory in nature, and results must beconsidered with caution. A more robust theoretical frame-work and a larger sample representative of the dyslexic pop-ulation are needed to come to any significant conclusion re-garding whether there might be a subgroup of individualswith dyslexia with spatial talents. Common with manyquasi-experimental designs, it was difficult to control for in-ternal validity. In spite of the limitations of the study, interms of a small sample, results are encouraging, particularlybecause the assessment of spatial abilities has frequentlybeen treated as tangential to the more conventional measuresof academic abilities. Researchers8 are increasingly advocat-ing the use of spatial tests to identify academic ability inpupils with learning difficulties through the use of computergraphics–based assessment tools.

Because the VE assessment revealed differences betweenthe two groups, which the standard assessments did not, itmay be a sensible suggestion to include real-life assessmentswhen measuring children’s abilities, especially since stan-dardized psychological assessments may not be ecologicallyvalid.25 Spatial tests added to the program of standardizedassessment that psychologists undertake may identify abili-ties or strengths that currently either go undetected or areminimized. VE testing could play a role in the assessment ofspatial skills as well as assist in planning educational ap-proaches that utilize a spatial orientation for those who pos-sess strong spatial skills and have difficulties learning inother modalities.

Many individuals with dyslexia tend to be best at holis-tic, 3D thinking and problem solving, keeping the big pic-ture in mind,26 and often do well on tasks with spatial com-ponents. For such individuals, using a gestalt approach tolearning and using techniques that help them learn throughtheir strengths can enable successful learning.27 For exam-ple, educators need to align teaching and learning strategieswith pupils’ learning styles.28 A visuospatial learning style(i.e., ensuring students see the whole picture before they un-derstand the parts) is often not addressed in schools, andstudents’ self-esteem suffers accordingly.29 Visuospatialtools such as virtual reality technology can offer ways to en-gage the learner through exercising visuospatial intelli-gence.30 Such tools can help transform what might be mean-ingless, intangible abstractions for many individuals withdyslexia into understandable, visible reality.

From an applied perspective, the findings provide someinsight into how educators might tailor assessment and ed-ucational intervention programs to suit the particular needsof the dyslexic learner. Future research needs to substanti-ate these findings. Examining the long-term effects of edu-cational approaches with individuals with dyslexia that uti-lize visuospatial modes of thought could prove an interestingarea for future research.

Disclosure Statement

The authors have no conflict of interest.

References

1. British Psychological Society. (1999) Dyslexia, literacy and psy-chological assessment. Report of a working party of the divisionof educational and child psychology. Leicester: British Psycho-logical Society.

VR TEST OF VISUOSPATIAL ABILITIES IN DYSLEXIA 167

2. Reason R. Letters to the editor re: critical response todyslexia, literacy and psychological assessment. Dyslexia2001; 7:174–7.

3. Nicolson RI. The dyslexia ecosystem. Dyslexia 2002; 8:55–66.4. Frith, U. Paradoxes in the definition of dyslexia. Dyslexia

1999; 5:192–214.5. Boehm G, Sherman G, Rosen G, et al. Neocortical ectopias

in BXSB mice: effects upon reference & working memorysystems. Cerebral Cortex 1996; 6:696–700.

6. Rose D. Universal design for learning associate editor col-umn. Journal of Special Education Technology 2001; 16:4.

7. Reichle ED Carpenter PA, Just MA. The neural basis of strat-egy and skill in sentence-picture verification. Cognitive Psy-chology 2000; 40:261–95.

8. West T. (1999) The abilities of those with reading disabili-ties: focusing on the talents of people with dyslexia. In Du-ane DD, ed. Reading and attention disorders—neurobiologicalcorrelates. Baltimore: York Press.

9. Sherman G. Dyslexia and talent: a connection? InternationalDyslexia Association, Perspectives, Spring 2002.

10. Geschwind N, Galaburda AM. (1987) Cerebral lateralization:biological mechanisms, associations and pathology. Cambridge,MA: MIT Press.

11. Wolf U, Lundberg I. The prevalence of dyslexia among artstudents. Dyslexia 2002; 8:34–42.

12. Galaburda A, Sherman G, Rosen G, et al. (1985) Develop-mental dyslexia: four consecutive cases with cortical anom-alies. Annals of Neurology 1985; 18:222–33.

13. Palmer S. Phonological recoding deficit in working memoryof dyslexic teenagers. Journal of Research in Reading 2000;23:28–40.

14. von Károlyi C, Winner E, Gray W, et al. Dyslexia linked totalent: global visual spatial ability. Brain & Language 2003;85:427–31.

15. Everatt J, Steffert B, Smythe I. An eye for the unusual: cre-ative thinking in dyslexics. Dyslexia 1999; 5:28–46.

16. Koenig O, Kosslyn S, Wolf P. Mental imagery and dyslexia:a deficit in processing multipart visual objects? Brain & Lan-guage 1991; 41:381–94.

17. La France. The gifted/dyslexic child: characterising and ad-dressing strengths and weaknesses. Annals of Dyslexia 1997;47:163–82.

18. Winner E, von Károlyi C, Malinsky D, et al. Dyslexia andvisual-spatial talents compensation vs deficit model. Brain& Language 2001; 76:81–110.

19. Carroll JB. (1993) Human cognitive abilities. Cambridge: Cam-bridge University Press.

20. Brunswick N, Martin GN. Developmental dyslexia, visu-ospatial ability and learning style: do they interact? BritishPsychological Society Annual Conference, University ofYork, 2007.

21. Elliott CD. Smith P, McCulloch K (1997) British Ability ScalesII technical manual. London: NferNelson.

22. Great Britain Department for Education and Science. (2001)Special educational needs code of practice. Annesley: DfES.

23. Brooks BM, Attree EA, Rose FD, et al. The specificity ofmemory enhancement during interaction with a virtual en-vironment. Memory 1999; 7:65–78.

24. Elliott CD, Smith P, McCulloch K (1996) British AbilityScales(BAS II) administration and scoring manual, 2nd ed. Lon-don: NferNelson.

25. Banaji M R, Crowder R C. The bankruptcy of everyday mem-ory. American Psychologist 1989; 44:1185–93.

26. West T. (1997) In the mind’s eye: visual thinkers, gifted peoplewith dyslexia and other learning difficulties, computer images andironies of creativity, 3rd ed. Amherst, NY: Promethesus.

27. Silverman LK. The visual–spatial learner: preventing schoolfailure. Gifted Child Today Magazine 1989; 34:15–20.

28. Riding R, Rayner S. (1998) Cognitive styles and learning strate-gies. London: Fulton.

29. Silverman LK, Freed JN. Strategies for the visual-spatiallearner. The Dyslexic Reader 1996; 4.

30. Winn WD. (1997) The impact of three-dimensional immersivevirtual environments on modern pedagogy. (Report). Seattle,WA: Human Interface Technology Laboratory, p.15.

Address reprint requests to:Elizabeth A. Attree

School of Psychology, University of East LondonRomford Road

London, E15 4LZUnited Kingdom

E-mail: e.a.attree@uel.ac.uk

ATTREE ET AL.168