Full Terms & Conditions of access and use can be found athttps://www.tandfonline.com/action/journalInformation?journalCode=nanc20

Aging, Neuropsychology, and CognitionA Journal on Normal and Dysfunctional Development

ISSN: 1382-5585 (Print) 1744-4128 (Online) Journal homepage: https://www.tandfonline.com/loi/nanc20

Normative data for the oldest old: Trail MakingTest A, Symbol Digit Modalities Test, VictoriaStroop Test and Parallel Serial Mental Operations

Katarina Fällman, Lina Lundgren, Ewa Wressle, Jan Marcusson & ElisabetClasson

To cite this article: Katarina Fällman, Lina Lundgren, Ewa Wressle, Jan Marcusson & ElisabetClasson (2019): Normative data for the oldest old: Trail Making Test A, Symbol Digit ModalitiesTest, Victoria Stroop Test and Parallel Serial Mental Operations, Aging, Neuropsychology, andCognition, DOI: 10.1080/13825585.2019.1648747

To link to this article: https://doi.org/10.1080/13825585.2019.1648747

© 2019 The Author(s). Published by InformaUK Limited, trading as Taylor & FrancisGroup.

Published online: 06 Aug 2019.

Submit your article to this journal

Article views: 355

View related articles

View Crossmark data

Normative data for the oldest old: Trail Making Test A,Symbol Digit Modalities Test, Victoria Stroop Test andParallel Serial Mental OperationsKatarina Fällman a, Lina Lundgrenb, Ewa Wresslec, Jan Marcussona

and Elisabet Classona

aDepartment of Acute Internal Medicine and Geriatrics, and Department of Clinical and ExperimentalMedicine, Linköping University, Linköping, Sweden; bDepartment of Geriatric Medicine, and Department ofSocial and Welfare Studies, Linköping University, Norrköping, Sweden; cDepartment of Acute InternalMedicine and Geriatrics, and Department of Social and Welfare Studies, Linköping University, Linköping,Sweden

ABSTRACTNormative data for evaluating cognitive function in the oldest old,aged 85 years and above, are currently sparse. The normative valuesused in clinical practice are often derived from younger old persons,from small sample sizes or from broad age spans (e.g. >75 years)resulting in a risk of misjudgment in assessments of cognitivedecline. This longitudinal study presents normative values for theTrail Making Test A (TMT-A), the Symbol Digit Modalities Test(SDMT), the Victoria Stroop Test (VST) and the Parallel SerialMental Operations (PaSMO) from cognitively intact Swedes aged85 years and above. 207 participants, born in 1922, were tested at85, 90 (n = 68) and 93 (n = 35) years of age with a cognitivescreening test battery. The participants were originally recruitedfor participation in the Elderly in Linköping Screening Assessment.Normative values are presented as mean values and standarddeviations, with and without adjustment for education. Therewere no clinically important differences between genders, but edu-cation had a significant effect on test results for the 85-year-olds.Age effects emerged in analyses of those participants who com-pleted the entire study and were evident for TMT-A, SDMT, VST1and PaSMO. When comparisons can be made, our results are inaccordance with previous data for TMT-A, SDMT and VST, and wepresent new normative values for PaSMO.

ARTICLE HISTORYReceived 29 April 2019Accepted 19 July 2019

KEYWORDSNeuropsychological tests;aged; 80 and over; executivefunction; attention;normative

Introduction

The process of normal aging is associated with changes in cognitive function, but highage is also the main risk factor for cognitive decline due to disease (Swedish Council onTechnology Assessment in Health Care, 2008). Neuropsychological tests are used toinvestigate whether a decline in a cognitive function is due to normal aging ora pathological process such as dementia, and the investigation relies on access toreliable normative data. Even though many of the tests used in clinical practice are

CONTACT Katarina Fällman katarina.fallman@liu.se

AGING, NEUROPSYCHOLOGY, AND COGNITIONhttps://doi.org/10.1080/13825585.2019.1648747

© 2019 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group.This is an Open Access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License(http://creativecommons.org/licenses/by-nc-nd/4.0/), which permits non-commercial re-use, distribution, and reproduction in anymedium, provided the original work is properly cited, and is not altered, transformed, or built upon in any way.

known to be affected by normal aging, and the very old population is the fastestgrowing age group in Sweden and in the world (Kinsella & Wan, 2008), normativedata for the oldest old (85 years of age and above) are sparse. Those data that areavailable are often from studies with few very old persons or from populations withlarge age spans, including younger elderly. To be able to give the correct diagnosis andavoid misjudgments, valid normative data for all ages are needed. This study focuses onfour widely used tests where the norms for the oldest population are of varying qualityand seldom include nonagenarians.

The neuropsychological tests in this study evaluate attention and executive function.Attention is the process whereby a person becomes receptive to stimuli and begins toprocess the information, while executive functions are the capacities that together allowfor independent and purposive actions (Harada, Natelson Love, & Triebel, 2013).Executive functions are crucial for us to manage our daily life, to be able to planahead, to work efficiently and to execute a task even in environments with disturbingmoments. It is the cognitive domain perhaps most closely associated with everydayfunction (Royall, Palmer, Chiodo, & Polk, 2005) and, together with attention, oftenaffected in neurocognitive as well as non-neurocognitive diseases. Difficulties withattention or executive function may for instance simulate memory problems due todecreased ability to stay focused on the task in individuals suffering from depression,one of the treatable differential diagnoses of dementia (Wright & Persad, 2007). Bothcapacities are also known to be affected by age (Harada et al., 2013) and a fair evaluationis dependent on reliable tools.

For a test to be reliable and valid in clinical practice without the help of a trainedneuropsychologist, it should be easy to administrate and not too time consuming, andthe effects of demographic factors such as gender, age and education should be known.The Trail Making Test A (TMT-A) (Reitan, 1955) measures attention and speed, andrequires visual scanning, graphomotor speed and visuomotor processing speed asmain cognitive functions. Gender is considered to have no or very little impact on theresults, in contrast to education and age where high age and low education affect theresults in a negative way (Llinas-Regla et al., 2017; Wei et al., 2018). The Symbol DigitModalities Test (SDMT) (Smith, 1991) evaluates complex attention, visual scanning andtracking, motor speed and memory (Strauss, Sherman, & Spreen, 2006) and has beenfound to predict functional decline in aging (Classon, Fallman, Wressle, & Marcusson,2016). SDMT is a fairly robust test, and even though some studies have found differencesbetween men and women (Jorm, Anstey, Christensen, & Rodgers, 2004; Kiely,Butterworth, Watson, & Wooden, 2014), it has often been considered as not clinicallyrelevant and different norms are not needed due to gender (Arango-Lasprilla et al., 2015;Peña-Casanova et al., 2009; Strauss et al., 2006). Many studies have however showndecreased test results with aging and higher points in well-educated subjects(Peña-Casanova et al., 2009; Richardson & Marottoli, 1996). The Victoria Stroop Test(VST) (Regard, 1981) is an executive test of selective attention and cognitive flexibility.The results from normative studies concerning impact of demographic variables such asgender and education are not consistent, but age is considered to have a negativeimpact on the results (Bayard, Erkes, & Moroni, 2011; Tremblay et al., 2016; Troyer, Leach,& Strauss, 2006). The Parallel Serial Mental Operations (PaSMO) (Nordlund et al., 2005) isan oral test of executive function and cognitive flexibility that does not require visual

2 K. FÄLLMAN ET AL.

scanning or psychomotor speed (Gothlin, Eckerstrom, Rolstad, Wallin, & Nordlund, 2017).To our knowledge, there are no published normative studies for PaSMO.

In the present study, community-dwelling elderly men and women were repeatedlytested at ages 85, 90 and 93 within the Elderly in Linköping Screening Assessment (ELSA85). The aim of the study is to present normative values for the very old population forthe four mentioned well-established tests and to investigate whether there are any ageeffects in the test results for the subgroup that participated throughout the study.

Material and method

Participants and procedure

Participants in this study were originally recruited for the longitudinal population studyElderly in Linköping Screening Assessment (ELSA 85). ELSA 85 initially invited all 85-year-olds (born in 1922) living in the municipality of Linköping in 2007 (n = 650 persons) toparticipate in the study. The study design is described in detail in Nagga, Dong,Marcusson, Skoglund, and Wressle (2012) and in Johansson, Marcusson, and Wressle(2019). 338 persons agreed to come to the Geriatric Clinic at Linköping UniversityHospital for an evaluation of cognitive function with neuropsychological assessments(described below) and for somatic check-up by a physician. Of the evaluated persons,336 completed the Mini Mental State Examination (MMSE) (Folstein, Folstein, & McHugh,1975) and were included in this study.

Follow-ups were carried out with home visits when the participants were 90 and 93 yearsold. Only participants who had undergone all the previous stages of the study and completedtheMMSE at each stagewere included, n = 107 at 90 years of age and n= 51 at 93 years of age(Figure 1).

The assessments were administered by an occupational therapist, a nurse ora physician from the Geriatric Department at Linköping University Hospital and all theevaluated tests were part of a screening test battery, the Cognitive Assessment Battery(CAB) (Nordlund, Pahlsson, Holmberg, Lind, & Wallin, 2011). During all phases of thestudy, the participants could cancel their participation or decline to participate in one ormore assessments and then continue with another test. Thus, not all participants havecarried out all the tests. They also had the possibility to end the examination andcontinue at another time.

Information about the participants’ health was obtained from the somatic check-ups,the subjects themselves and their medical records.

Exclusion criteria

The focus of this study was to find normative values for the cognitively healthy popula-tion aged 85 and above. Therefore, persons with known diseases that can causecognitive decline were excluded from the analysis. Exclusion criteria were as follows:severe head injury, traumatic cerebral bleeding, stroke, transient ischemic attack duringthe last six months, epilepsy, Parkinson’s disease, Parkinsonian disorders, Huntington’sdisease, multiple sclerosis, normal pressure hydrocephalus, dementia, objective mildcognitive impairment, chronic obstructive pulmonary disease with hypoxia (saturation

AGING, NEUROPSYCHOLOGY, AND COGNITION 3

less than 92% in air), insulin-dependent diabetes mellitus, generalized cancer, psychoticdisorders, ongoing depression and drug abuse (Lezak, Howieson, Bigler, & Tranel, 2012).When detailed somatic information was not available, the participant was excluded.

To further minimize the risk of including persons with cognitive decline, due tosomatic health problems not registered in their medical journal, persons who scoredfewer than 24 points in the MMSE were excluded (Creavin et al., 2016). A recent studyfrom Kvitting, Fallman, Wressle, and Marcusson (2019) suggests a cut-off of 26 points forpersons aged 85 and above. However, changing the cut-off did not give any clinicallyrelevant differences in results (results not shown).

Cognitive tests

The participants’ cognitive functions were evaluated with the Cognitive Assessment Battery(CAB) (Nordlund et al., 2011) and all the investigations were carried out in accordance with its

Figure 1. Flowchart of the participants in the Elderly in Linköping Screening Assessment.Excl. = number of persons who were excluded from the study for other reasons, e.g. moving toanother city during the study period. Survey = number of participants who answered a postal surveyincluding demographic information. Clinic = number of participants evaluated at the Geriatric Clinic,Linköping University Hospital. Home = number of participants evaluated during the follow-up homevisits. MMSE = number of participants who underwent the Mini Mental State Examination.Cognitively intact = number of participants fulfilling inclusion criteria.

4 K. FÄLLMAN ET AL.

manual. The CAB consists ofmultiple assessments, some of them in established short versions.It was developed to investigate the cognitive domains of speed and attention, learning andepisodicmemory, visuospatial abilities, language, and executive functions. The tests evaluatedin this study were executed in the following order: SDMT, VST, TMT-A and PaSMO.

The Trail Making Test (TMT) (Reitan, 1955) is one of the most widely used neuropsy-chological tests and consists of two subtests, TMT-A and TMT-B. In the CAB only TMT-A,the first subtest, is included. The participant is given the task of drawing a line betweenthe numbers 1–25, scattered across a sheet, in the correct ascending order as quickly aspossible. The score is the time in seconds to complete the task, with higher scoresrepresenting worse attentional capacity. No practice effects have been seen with longintervals (12 months) (Strauss et al., 2006).

The Symbol Digit Modalities Test (SDMT) (Smith, 1991) is a short, easily administeredcognitive test which exists in both written and oral forms, of which the written form isused in the CAB. The participant is presented with a coding key of nine geometricalsymbols, each representing a digit. After an initial 10 trial practice period, the subject isasked to recode as many symbols into digits as possible in 90 seconds using the key. Thescore is one point per correct digit, and the maximum result is 110 points. No significantpractice effects have been noted over a two-year interval (Uchiyama et al., 1994).

The Victoria Stroop Test (VST) (Regard, 1981) is a shorter version of the Stroop test(Stroop, 1935). It consist of three subtests, each presenting a stimulus card with six rowsof four items. The first test (VST1) shows 24 dots in the colors green, blue, red and yellowin a pseudorandom order. The second test (VST2) shows 24 common words, written inthe four different colors. The third test (VST3) shows 24 color words (green, yellow, blueand red) written in one of the four colors, but so that the print color never correspondsto the color name (e.g. “blue” is written in green ink). For every subtest, the participant isasked to name the color of the dot, word or color word as quickly as possible and theexaminer corrects any unnoticed errors. The score is the time in seconds, with higherscores indicating worse performance. There were no significant retest effects whenmiddle-aged to older persons carried out a similar Stroop test with three-year intervals(Adolfsdottir, Wollschlaeger, Wehling, & Lundervold, 2017).

The Parallel Serial Mental Operations (PaSMO) (Nordlund et al., 2005) is a verbal test that issimilar to the oral version of the TMT-B (Ricker, Axelrod, &Houtler, 1996) and theAlphanumericSequencing Test (Grigsby & Kaye, 1995), but includes all the letters in the alphabet witha reverse order between number and letter. It was originally developed at the Institute ofNeuroscience and Physiology at the University of Gothenburg, Sweden. The participant is firstgiven the training task of reciting the alphabet as quickly as possible (the Swedish versionwith28 letters). Then the participant is asked to recite the alphabet together with the number thatcorresponds to the letter (e.g. A-1, B-2). The score for the test is the time in seconds, and if theparticipant makes a mistake the examiner helps by saying the last correct letter and numberpair. Retest effects and practice effects on PaSMO or similar oral executive tests such as OralTMT-B have not yet been established (Strauss et al., 2006).

Ethical approval

The ELSA 85 study, including permission to obtain data from all registers held byÖstergötland County Council, has been approved by the Research Ethics Committee of

AGING, NEUROPSYCHOLOGY, AND COGNITION 5

Linköping University, Sweden (2006: 141–06, 2012: 332–31; 2014: 455–31) and complieswith the ethical rules for human experimentation stated by the Declaration of Helsinki(World Medical Association declaration of Helsinki, 1997). Written informed consent wascollected from all participants at each step of the study.

Statistical analysis

Statistical analysis was carried out using SPSS version 24. Mean values and standarddeviations were calculated for each age group and for each age divided by educationallevel: less than or equal to 9 years of schooling, or 10 years and above (initial analysesincluding three educational levels, 4–9 years, 10–12 years and 13 years and above,showed no difference between the two levels with higher years of schooling). Anydifferences between men and women and level of education were analyzed withindependent sample t-tests. Age effects were analyzed with repeated measures forthose individuals who had participated in all three test occasions (referred to below asCompleters) and post hoc analyses between the age groups were carried out with theBonferroni method. Sphericity was calculated using Mauchly’s test, assessing thehypothesis that the variance of the differences between conditions are equal, andwhen sphericity could not be assumed the significance was determined by correctionof Greenhouse-Geisser (Field, 2018). The level of significance was set at 0.05.

Mean values and standard deviations for the whole ELSA 85 population, as well asdemographic information about the participants excluded from this study, are presentedin the supplementary material. Analyses or test results regarding the whole ELSA 85population and the excluded participants will not be discussed further in this article.

Results

All the participants in this study, with known educational level, had at least four years ofschooling. Descriptive information about the participants is shown in Table 1.

When comparing the 172 persons who dropped out or were excluded at either of thefollow-ups for the 35 participants that were included at 93 years of age, it was notedthat there were lower proportions of men (42% vs 60%) and participants living withfamily member (43% vs 63%) at baseline. The proportions with a high educational level

Table 1. Descriptive information about the participants.85-year-olds 90-year-olds 93-year-olds

Count (%) Count (%) Count (%)

Gender Female 114 (55) 36 (53) 14 (40)Male 93 (45) 32 (47) 21 (60)

Living situation Living alone 111 (54) 38 (56) 25 (71)Living with family member 96 (46) 30 (44) 10 (29)

Residence Ordinary accommodation 203 (98) 62 (91) 30 (86)Sheltered housing 4 (2) 5 (7) 4 (11)Residential home 0 (0) 1 (2) 1 (3)

Education ≤ 9 144 (70) 48 (71) 24 (69)(years) ≥ 10 60 (29) 19 (28) 11(31)

Missing data 3 (1) 1 (1) 0 (0)Total 207 68 35

6 K. FÄLLMAN ET AL.

(28% vs 31%) and ordinary accommodation (98% vs 100%) were however almost thesame in the two groups.

There was no difference between men and women for any test at any age group in thisstudy except for the VST1 at age 85, where women had a lower mean (i.e. were faster) thanmen (Mwomen = 18, SD = 5.7, Mmen = 21, SD = 6.1, t(203) = −2,68, p = .008).

The normative data for TMT-A, SDMT, VST and PaSMO are presented in Table 2. Thenorms are divided into 85-, 90- and 93-year-olds, as a total and in two educational levels.As can be seen, there was a significant difference due to educational level for all tests,except VST1, in the 85-year-olds. However, only performance on SDMT differed due toeducational level at age 90, and at age 93 educational level had no significant impact ontest results. It is worth noting that the drop-out frequency for PaSMO was higher thanfor the other three tests – 16% at 85 years, 32% at 90 years and 28% at 93 years – andthat the drop-outs were mostly men or participants with low educational level.

As can be seen in Figure 2, TMT-A, SDMT, VST1 and PaSMO were all sensitive to agingwith successively lower performance. However, there was no difference due to age forVST2 and VST3. The actual change over time as a mean was an increase of 11 seconds(±21) for TMT-A, a decrease of 8 symbols (±6) for SDMT, and increases of 3 seconds (±3)for VST 1, 1 second (±7) for VST 2, 4 seconds (±9) for VST 3 and 19 seconds (±27) forPaSMO. The Completers had an overall somewhat better result than the whole cogni-tively intact population.

Discussion

We present normative values for four easily administered cognitive tests that evaluateattention and executive function, derived from a large population of cognitively intactvery old persons (i.e. over 85 years of age). The norms are presented as mean values andstandard deviations to facilitate applicability in clinical practice. To the authors’ knowl-edge, this is the first study to present normative values for very old persons for PaSMOand the study with the highest number of nonagenarians for the other three tests. Eventhough there are other studies, e.g. Royall, Palmer, Chiodo, and Polk (2004), investigatingattention and executive function in relation to high age, the present study is one of thelargest normative studies measuring attention and executive function in persons85 years of age and above.

As in various other normative studies, we found no significant gender differences inTMT-A or SDMT (Ashendorf et al., 2008; Peña-Casanova et al., 2009; Sheridan et al., 2006),or in PaSMO. In VST1 only, women were faster than men at the age of 85 years, but thedifference was only 3 seconds and the clinical relevance was questionable. Tremblayet al. (2016) found that women outperformed men in all parts of VST, but their normsare based on younger elderly. Most other studies on VST and similar variants of theStroop test do not recommend gender-specific norms (Strauss et al., 2006) that, in linewith the present results, are not required for the oldest old.

We present our results divided into two educational levels, ≤9 and ≥10 years. For theoctogenarians the results show an impact of educational level on performance, which isin accordance with previous studies (Ashendorf et al., 2008; Kiely et al., 2014; Tremblayet al., 2016). Surprisingly, there were no significant differences between high and lowlevels of education in the nonagenarians, except for SDMT at 90 years of age. This

AGING, NEUROPSYCHOLOGY, AND COGNITION 7

Table2.

Normativedata

fortheoldest

old:

TrailMakingTest-A

(TMT-A),Symbo

lDigitMod

alities

Test

(SDMT),Victoria

Stroop

Test

(VST)andParallelSerial

MentalO

peratio

ns(PaSMO)

TMT-A

SDMT

VST1

VST2

VST3

PaSM

O

nM*

SDp

nM**

SDp

nM*

SDp

nM*

SDp

nM*

SDp

nM*

SDp

Age

Edu

85To

tal***

198

6831

197

248

205

196

203

3111

203

4821

174

106

474-9

138

7233

137

227

142

206

140

3211

140

5023

115

115

51≥10

5757

23.002

5728

9<.001

6018

5n.s

6028

10.011

6042

16.012

5688

32<.001

90To

tal***

5566

2057

258

5820

958

3419

5750

2046

106

454-9

3766

1839

247

3919

1039

3622

3953

2130

113

48≥10

1761

22n.s

1728

9.049

1818

4n.s

1829

8n.s

1743

17n.s

1592

36n.s

93To

tal***

3367

2434

227

3320

532

299

3246

1725

115

504-9

2371

2723

227

2220

522

309

2248

1815

108

41≥10

1058

12n.s

1123

8n.s

1120

5n.s

1027

9n.s

1042

13n.s

10125

61n.s

*=second

s**=nu

mberscorrect

Edu=

yearsof

education

***=

allp

articipants,including

three85-year-olds

andon

e90-year-oldwith

unknow

neducationallevel

8 K. FÄLLMAN ET AL.

contradicts most other studies, which have found educational level to be important forattention and executive function as measured by these tests. The lack of statisticallysignificant differences could however be due to the sizes of the age groups in this study,with smaller groups at the very high ages. Mean performances show differences that areclinically significant (e.g. more than 5 seconds or 0.5 SD) for all ages in TMT-A, VST3 andPaSMO between the two educational levels. Bezdicek et al. (2017) found that ageseemed to have a more pronounced effect compared to education in old age, butthere are few studies with participants over 90 years of age with which to compare ourdata and more research needs to be done to be able to draw any conclusions about thepresent findings.

When scrutinizing the field for normative test results for the very old and TMT-A, theresults are conflicting. Ashendorf et al. (2008) and Bezdicek et al. (2017), two fairly largestudies (n = 117, 79–98 years of age and n = 104, 85–98 years of age, respectively) withmany similarities to our study in the choice of participants, have two diverse normativeresults where our study results end up in between, together with Zalonis et al. (2008).

Our results for the SDMT are in consensus with the results from a normative study fromAustralia, Kiely et al. (2014), and with Richardson and Marottoli (1996), a smaller study often

Figure 2. Age effects calculated with repeated measures for the Completers (participants with testresults for all three occasions). Error bars with 95% confidence interval. Lines above the bars indicatesignificance, p-value < .05. Number of Completers; Trail Making Test A n = 29, Symbol DigitModalities Test n = 28, Victoria Stroop Test 1 n = 30, Victoria Stroop Test 2 n = 29, VictoriaStroop Test 3 n = 29 and Parallel Serial Mental Operations n = 23.

AGING, NEUROPSYCHOLOGY, AND COGNITION 9

referred to in clinical practice. There are other studies with both higher and lower normvalues than those presented here, but they are from small samples or younger participants(Nielsen, Lolk, & Kragh-Sørensen, 1995; Vogel, Stokholm, & Jørgensen, 2013).

Only a few normative studies have been carried out for VST. In contrast to Bayard et al.(2011) and Tremblay et al. (2016), we chose not to focus on the interference score, but tocalculate the time score to facilitate use in clinical practice. Our results are consistent withthe results of Troyer et al. (2006) concerning VST3, but our participants have a higher score(i.e. take longer) for the VST1 (4–5 seconds longer mean) and VST2 (7–12 seconds longermean), and our results are more diverse. The participants in Troyer et al. (2006) arehowever younger and have a higher educational level than the participants in this study.

To our knowledge, there are no published normative studies concerning PaSMO. Thenorms presented in the CAB manual are only applicable up to the age of 79, and ourdata suggest a higher and more diverse range for normative values for persons aged85 years and above. With this in mind, and combined with the fact that there were manyparticipants – especially those with low educational level and men – who did notcomplete this test, it may be questioned whether this is a clinically useful test for thevery old population. Oral versions of alphanumerical tests such as the oral TMT-B andPaSMO are thus likely to be too challenging for the oldest old and not suitable alter-natives to pen-and-pencil formats in elderly people with disabilities, disproportionatelypenalizing low-educated men. It is however important to note that PaSMO is the last testin the CAB, which may be a contributing factor to the drop-out frequency. Morenormative research and validation for PaSMO and similar oral tasks need to be doneto investigate the clinical use of these executive tests.

When analyzing the results from the Completers, there were declines in results due toage in all the tests except for those that involve inhibition (VST2 and VST3). These resultsare in general accordance with previous studies showing age-related declines in cogni-tive speed, attention and executive functioning (Royall et al., 2004; Salthouse, 1996,2016). The lack of aging effects in inhibition conflicts with some (Adolfsdottir et al., 2017;Troyer et al., 2006; Van der Elst, Van Boxtel, Van Breukelen, & Jolles, 2006) but not all(Verhaeghen & De Meersman, 1998) previous research. It has been suggested thatinhibition per se is actually well preserved during aging and worsening Stroop perfor-mance, when seen, is more a reflection of cognitive slowing.

There are some limitations in the study design that needs to be highlighted. Thedefinition of “cognitively intact” is an estimation done after the data were collected. Thiscan lead both to the risk of including persons with preclinical cognitive disease, but also toexcluding cognitively healthy participants with e.g. somatic disease, leading to a biased testresult. The careful exclusion of participants with conditions known to impact on cognitionhelps to ensure that the norms presented reflect healthy aging, but as we did not have theinformation about e.g. biomarkers such as beta amyloid and tau proteins the definition is anestimation. Normative studies in the oldest old are scarce and for natural reasons thedecrease in number of participants in the highest age groups contribute to possiblepopulation differences between ages, like survival of the fittest. This could be an explanationto the somewhat unexpected result of similar mean test results for the evaluated teststhroughout the eight years of the study, where the healthier and more cognitively intactparticipants are more likely to continue to participate in the study, resulting in a morehomogenous group throughout the years of the study. Therefore, the age effect was studied

10 K. FÄLLMAN ET AL.

in the subgroup that participated in all time points, reducing the possibility of differences inpopulation between time points. The normative data for the nonagenarians in this study arebased on small sample sizes. This is not ideal for normative data because of the risk ofsampling errors and more normative research with larger samples needs to be done for theoldest old population to confirm our results. However, for a certain very old patient, thenormative data based on a small demographically fit sample may by more representativethan data based on a large age heterogeneous sample and therefore the norms presented inthis study are nonetheless useful in clinical practice. Where comparisons can be made, thepresent normative results are in accordance with those of previous studies.

Conclusion

The normative values presented in this study are valid and representative for very oldpersons in Sweden and other countries with similar social contexts and educationalsystems. The norms are useful in the clinical evaluation of attention and executive function,such as in the investigation of neurodegenerative disease, and the decline in test resultsdue to age highlights the need for updated normative values for this age group.

Acknowledgments

We would like to thank the participants for devoting their time and energy to this project. Wewould also like to thank our colleagues who have taken part in the data collection with enthu-siasm and endurance.

Disclosure statement

No potential conflict of interest was reported by the authors.

Funding

This study was performed by the authors while employed by Linköping University and/orÖstergötland County Council. Support staff (nurse, occupational therapist) were funded bya research grant from Linköping University (number LIO-696631, LIO 602761 and LIO 537591).The funders had no input into or influence on the study;Linköpings Universitet [LIO- 696631, LIO-602761, LIO- 537591].

Contributors

KF: drafting the manuscript, analysis and interpretation of the data, acquisition of data andstatistical analysis. LL: analysis and interpretation of data, revising the manuscript. EW: recruitmentof the study cohort, study concept and design, supervision of data collection and interpretation ofdata, and obtaining funding. JM: recruitment of the study cohort, study concept and design,supervision of data collection and interpretation of data, revising the manuscript and obtainingfunding. EC: drafting/revising the manuscript, study concept and design, analysis and interpreta-tion of the data, study supervision and obtaining funding. All authors have provided criticalcomments on the manuscript and approved the final version.

AGING, NEUROPSYCHOLOGY, AND COGNITION 11

Data sharing statement

Data cannot be made publicly available. Authorized researchers who have ethics approval and aninstitutionally approved study plan can apply for data sharing through the Department of Clinical andExperimentalMedicine, LinköpingUniversity. Formore details, please contact the corresponding author.

Impact statement

We certify that this work is recent novel clinical research. It presents useful clinical data on the useof Trail Making Test A, the Symbol Digit Modalities Test, the Victoria Stroop Test and Parallel SerialMental Operations, for persons aged 85 years and above.

ORCID

Katarina Fällman http://orcid.org/0000-0002-5289-8831

References

Adolfsdottir, S., Wollschlaeger, D., Wehling, E., & Lundervold, A. J. (2017). Inhibition and Switchingin Healthy Aging: A Longitudinal Study. Journal of the International Neuropsychological Society,23(1), 90–97.

Arango-Lasprilla, J. C., Rivera, D., Rodriguez, G., Garza, M. T., Galarza-Del-Angel, J., Rodriguez, W., . . .Perrin, P. B. (2015). Symbol Digit Modalities Test: Normative data for the Latin American Spanishspeaking adult population. NeuroRehabilitation, 37(4), 625–638.

Ashendorf, L., Jefferson, A. L., O’Connor, M. K., Chaisson, C., Green, R. C., & Stern, R. A. (2008). TrailMaking Test errors in normal aging, mild cognitive impairment, and dementia. Archives ofClinical Neuropsychology, 23(2), 129–137.

Bayard, S., Erkes, J., & Moroni, C. (2011). Victoria Stroop Test: Normative Data in a Sample Group ofOlder People and the Study of Their Clinical Applications in the Assessment of Inhibition inAlzheimer’s Disease. Archives of Clinical Neuropsychology, 26(7), 653–661.

Bezdicek, O., Stepankova, H., Axelrod, B. N., Nikolai, T., Sulc, Z., Jech, R., . . . Kopecek, M. (2017).Clinimetric validity of the Trail Making Test Czech version in Parkinson’s disease and normativedata for older adults. The Clinical Neuropsychologist, 31(1), 42–60.

Classon, E., Fallman, K., Wressle, E., & Marcusson, J. (2016). Relations between ConcurrentLongitudinal Changes in Cognition, Depressive Symptoms, Self-Rated Health and EverydayFunction in Normally Aging Octogenarians. PloS One, 11(8), e0160742.

Creavin, S. T., Wisniewski, S., Noel-Storr, A. H., Trevelyan, C. M., Hampton, T., Rayment, D., . . .Cullum, S. (2016). Mini-Mental State Examination (MMSE) for the detection of dementia inclinically unevaluated people aged 65 and over in community and primary care populations.Cochrane Database Syst Rev, 13(1), CD011145.

Field, A. (2018). Discovering Statistics Using IBM SPSS Statistics (5 ed.). California: Sage PublicationsInc.

Folstein, M. F., Folstein, S. E., & McHugh, P. R. (1975). “Mini-mental state”: A practical method for gradingthe cognitive state of patients for the clinician. Journal of Psychiatric Research, 12(3), 189–198.

Gothlin, M., Eckerstrom, M., Rolstad, S., Wallin, A., & Nordlund, A. (2017). Prognostic Accuracy ofMild Cognitive Impairment Subtypes at Different Cut-Off Levels. Dementia and GeriatricCognitive Disorders, 43(5–6), 330–341.

Grigsby, J., & Kaye, K. (1995). Alphanumeric sequencing and cognitive impairment among elderlypersons. Perceptual and Motor Skills, 80, 732–734.

Harada, C. N., Natelson Love, M. C., & Triebel, K. L. (2013). Normal cognitive aging. Clinics inGeriatric Medicine, 29(4), 737–752.

12 K. FÄLLMAN ET AL.

Johansson, M. M., Marcusson, J., & Wressle, E.(2019). Maintaining health-related quality of life from85 to 93 years of age despite decreasing functional ability. British Journal of OccupationalTherapy, 82(6), 348–356.

Jorm, A. F., Anstey, K. J., Christensen, H., & Rodgers, B. (2004). Gender differences in cognitiveabilities: The mediating role of health state and health habits. Intelligence, 32(1), 7–23.

Kiely, K. M., Butterworth, P., Watson, N., & Wooden, M. (2014). The Symbol Digit Modalities Test:Normative Data from a Large Nationally Representative Sample of Australians. Archives ofClinical Neuropsychology, 29(8), 767–775.

Kinsella, K., & Wan, H. (2008). An Aging World: 2008 (International Population Reports no. P95/09-1).Washington, DC: U.S. Census Bureau (US).

Kvitting, A. S., Fallman, K., Wressle, E., & Marcusson, J. (2019). Age-Normative MMSE Data for OlderPersons Aged 85 to 93 in a Longitudinal Swedish Cohort. Journal of American Geriatric Society,67(3), 534–538.

Lezak, M., Howieson, D. B., Bigler, E. D., & Tranel, D. (2012). Neuropsychological Assessment (5 ed.).New York: Oxford University Press.

Llinas-Regla, J., Vilalta-Franch, J., Lopez-Pousa, S., Calvo-Perxas, L., Torrents Rodas, D., & Garre-Olmo, J. (2017). The Trail Making Test. Assessment, 24(2), 183–196.

Nagga, K., Dong, H. J., Marcusson, J., Skoglund, S. O., & Wressle, E. (2012). Health-related factorsassociated with hospitalization for old people: Comparisons of elderly aged 85 in a populationcohort study. Archives of Gerontology and Geriatrics, 54(2), 391–397.

Nielsen, H., Lolk, A., & Kragh-Sørensen, P. (1995). Normative data for eight neuropsychologicaltests, gathered from a random sample of Danes aged 64 to 83 years. Nordisk Psykologi, 47(4),241–255.

Nordlund, A., Pahlsson, L., Holmberg, C., Lind, K., & Wallin, A. (2011). The Cognitive AssessmentBattery (CAB): A rapid test of cognitive domains. International Psychogeriatrics, 23(7), 1144–1151.

Nordlund, A., Rolstad, S., Hellström, P., Sjögren, M., Hansen, S., & Wallin, A. (2005). The GoteborgMCI study: Mild cognitive impairment is a heterogeneous condition. Journal of Neurology,Neurosurgery & Psychiatry, 76(11), 1485–1490.

Peña-Casanova, J., Quiñones-Úbeda, S., Quintana-Aparicio, M., Aguilar, M., Badenes, D.,Molinuevo, J. L., . . . Blesa, R. (2009). Spanish Multicenter Normative Studies (NEURONORMAProject): Norms for Verbal Span, Visuospatial Span, Letter and Number Sequencing, Trail MakingTest, and Symbol Digit Modalities Test. Archives of Clinical Neuropsychology, 24(4), 321–341.

Regard, M. (1981). Cognitive Rigidity and Flexibility: A Neuropsychological Study. University ofVictoria. unpublished doctoral dissertation.

Reitan, R. M. (1955). The relation of the trail making test to organic brain damage. Journal ofConsulting Psychology, 19(5), 393–394.

Richardson, E. D., & Marottoli, R. A. (1996). Education-Specific normative data on commonneuropsy-chological indices for individuals older than 75 Years. The Clinical Neuropsychologist,10(4), 375–381.

Ricker, J. H., Axelrod, B. N., & Houtler, B. D. (1996). Clinical Validation of the Oral Trail Making Test.Cognitive and Behavioral Neurology, 9(1), 50–53.

Royall, D. R., Palmer, R., Chiodo, L. K., & Polk, M. J. (2004). Declining executive control in normalaging predicts change in functional status: The Freedom House Study. Journal of AmericanGeriatrics Society, 52(3), 346–352.

Royall, D. R., Palmer, R., Chiodo, L. K., & Polk, M. J. (2005). Normal rates of cognitive change insuccessful aging: The freedom house study. Journal of the International NeuropsychologicalSociety, 11(7), 899–909.

Salthouse, T. A. (1996). The processing-speed theory of adult age differences in cognition.Psychological Review, 103(3), 403–428.

Salthouse, T. A. (2016). Little relation of adult age with cognition after controlling generalinfluences. Developmental Psychology, 52(10), 1545–1554.

Sheridan, L. K., Fitzgerald, H. E., Adams, K. M., Nigg, J. T., Martel, M. M., Puttler, L. I., . . . Zucker, R. A.(2006). Normative Symbol Digit Modalities Test performance in a community-based sample.Archives of Clinical Neuropsychology, 21(1), 23–28.

AGING, NEUROPSYCHOLOGY, AND COGNITION 13

Smith, A. (1991). Symbol Digit Modalities Test. Los Angeles, CA: Western Psychological Services.Strauss, E., Sherman, E. M. S., & Spreen, O. (2006). A compendium of neuropsychological tests:

Administration, norms, and commentary. Oxford: Oxford University Press.Stroop, J. R. (1935). Studies of interference in serial verbal reaction. Journal of Experimental

Psychology, 18, 643–662.Swedish Council on Technology Assessment in Health Care. (2008). Dementia- Etiology and

Epidemiology, a systematic review. Stockholm: SBU.Tremblay, M.-P., Potvin, O., Belleville, S., Bier, N., Gagnon, L., Blanchet, S., . . . Hudon, C. (2016). The

Victoria Stroop Test: Normative Data in Quebec-French Adults and Elderly. Archives of ClinicalNeuropsychology, 31(8), 926–933.

Troyer, A. K., Leach, L., & Strauss, E. (2006). Aging and Response Inhibition: Normative Data for theVictoria Stroop Test. Aging, Neuropsychology, and Cognition, 13(1), 20–35.

Uchiyama, C. L., D’Elia, L. F., Dellinger, A. M., Seines, O. A., Becker, J. T., Wesch, J. E., . . . Miller, E. N.(1994). Longitudinal comparison of alternate versions of the symbol digit modalities test: Issuesof form comparability and moderating demographic variables. Clinical Neuropsychologist, 8(2),209–218.

Van der Elst, W., Van Boxtel, M. P., Van Breukelen, G. J., & Jolles, J. (2006). The Stroop color-wordtest: Influence of age, sex, and education; and normative data for a large sample across theadult age range. Assessment, 13(1), 62–79.

Verhaeghen, P., & De Meersman, L. (1998). Aging and the Stroop effect: A meta-analysis.Psychology and Aging, 13(1), 120–126.

Vogel, A., Stokholm, J., & Jørgensen, K. (2013). Performances on Symbol Digit Modalities Test, ColorTrails Test, and modified Stroop test in a healthy, elderly Danish sample. Aging,Neuropsychology, and Cognition, 20(3), 370–382.

Wei, M., Shi, J., Li, T., Ni, J., Zhang, X., Li, Y., . . . Tian, J. (2018). Diagnostic Accuracy of the ChineseVersion of the Trail-Making Test for Screening Cognitive Impairment. Journal of AmericanGeriatrics Society, 66(1), 92–99.

World Medical Association declaration of Helsinki. (1997). Recommendations guiding physicians inbiomedical research involving human subjects. JAMA, 277(11), 925–926.

Wright, S. L., & Persad, C. (2007). Distinguishing between depression and dementia in olderpersons: Neuropsychological and neuropathological correlates. Journal of Geriatric Psychiatryand Neurology, 20(4), 189–198.

Zalonis, I., Kararizou, E., Triantafyllou, N. I., Kapaki, E., Papageorgiou, S., Sgouropoulos, P., &Vassilopoulos, D. (2008). A normative study of the trail making test A and B in Greek adults.Clinical Neuropsychologist, 22(5), 842–850.

14 K. FÄLLMAN ET AL.