Anda di halaman 1dari 5

Age and Ageing 2014; 43: 285–289 © The Author 2013.

thor 2013. Published by Oxford University Press on behalf of the British Geriatrics Society.
doi: 10.1093/ageing/aft157 All rights reserved. For Permissions, please email:
Published electronically 14 October 2013

A Stroop Stepping Test (SST) using low-cost

computer game technology discriminates
between older fallers and non-fallers

Neuroscience Research Australia, Falls and Balance Research Group, Sydney, NSW, Australia
School of Public Health and Community Medicine, University of New South Wales, Sydney, NSW, Australia
Address correspondence to: Stephen R. Lord. Email:


Background: impaired stepping and reduced cognitive functioning have both been identified as fall-risk factors in older
people. We developed a Stroop Stepping Test (SST) that combines stepping and response inhibition using low-cost computer
game technology to provide a functional measure that reflects real-life behaviour and determined whether this test discrimi-
nates between older fallers and non-fallers.
Methods: a cross-sectional study, including 103-independent living cognitively intact older people (70–93 years), was con-
ducted. Participants were assessed on the SST and other outcome measures associated with fall-risk. The SST presented
arrows on a computer screen with words written within them. Participants were asked to step in the direction indicated by the
word and ignore the arrow orientation. Participants also reported whether they had fallen or not in the past 12 months.
Results: twenty-eight percent of participants reported falling in the past year. SST mean time per trial [OR: 1.72 (95% confi-
dence interval 1.02–2.91) and SST errors (OR: 1.53 (1.14–2.07)] were associated with falls. After adjusting for other fall-risk
factors in a multivariate logistic regression analysis, each error made during SST increased the odds of falling by a factor 1.7
[OR: 1.65 (1.17–2.34)].
Conclusions: this study shows the SST—a low-cost video game device—is feasible for older people to undertake. The SST
was able to distinguish fallers from non-fallers, providing a novel way to explore cognitive mechanisms for fall-risk in
older people.

Keywords: accidental falls, assessment, stepping, executive function, inhibition, older people

Introduction We developed a low-cost Stroop Stepping Test (SST)

using dance pads adapted from exercise-based video games.
Poor reactive and volitional stepping have been identified SST involves both stepping and inhibition to provide a
as risk factors for falls in older people [1–5] showing that measure that reflects complex real-life behaviour. We
reflexive responses and executive function (EF) are im- assessed (i) whether the SST is feasible for older people to
portant to maintain balance [6]. An important aspect of EF undertake, (ii) whether SST performance is associated with
involves attending selectively to one stimulus while inhibit- validated fall-risk measures and (iii) whether the SST can
ing other sensory inputs and motor outputs. Studies distinguish fallers from non-fallers.
showing that impaired inhibition is associated with falls
have used neuropsychological assessments that are not spe-
cific to balance control and therefore might not identify
early or small functional impairments resulting in slowed Methods
or erroneous motor responses as a consequence of Participants were recruited from retirement villages in
impaired inhibition [4, 7, 8]. Other studies have used Sydney, Australia. Inclusion criteria were age over 70 years,
sophisticated laboratory equipment making their use in independent-living, ambulant with or without a walking aid,
clinical practice questionable [9, 10]. able to step unassisted on a step pad. People were excluded if

D. Schoene et al.

Time was used to measure the ability to perform accurate

rapid steps [12]. Functional mobility was assessed with the
Timed-Up and Go test [13] and the Alternate Step Test
(AST) [14]. Attention and processing speed was assessed
using the Digit Symbol Substitution Test [15] and Trail
Making Test (TMT) Part A. EF was assessed using the TMT
Part B [16] and Stroop test [16]. Delayed recall was taken
from the MMSE. The 10-item Iconographical Falls Efficacy
Scale was used to measure fear of falling [17].
Participants who reported ≥1 falls in the past 12 months
were classified as fallers. A fall was defined as ‘an unexpected
event in which the participants come to rest on the ground,
floor, or lower level’ [18].
Pearson and Spearman-Rank correlations were used to
assess associations between outcome variables. Univariate
and multivariate logistic regression analyses were used to cal-
culate odds ratios for the associations between the outcome
measures and falls while adjusting for other fall-risk variables
(if P < 0.15 in univariate analysis). Age, gender, education,
delayed recall (dichotomised), lower limb pain and use of
walking aids were selected as potential confounders for the
relationship between SST performance and falls. Data were
Figure 1. SST: participant stepping according to the word and analysed with SPSS version 20 for Windows.
not the arrow orientation.

they were cognitively impaired (MMSE <24), were colour- Results

blind or had visual impairments that could not be corrected, One hundred and fourteen individuals showed interest in
had neurodegenerative disorders or suffered from limiting participating in this study, of which 11 were excluded
lower limb pain. Written informed consent was obtained because they did not fulfil the inclusion criteria. All remain-
from all participants prior to the study which was approved ing 103 participants (mean age = 79.5 ± 4.8 years, mean
by the Human Research Ethics Committee at the University MMSE = 28.9 ± 1.1) were able to complete SST in under
of New South Wales. three minutes. No adverse events were recorded.
Taking longer to complete each SST trial was associated
with poorer scores in nearly all other measures
(Supplementary data are available in Age and Ageing online,
SST was administered using a custom-made dance pad Appendix 1, Table S1). Making more errors on the SST was
(150 × 90 cm) which was connected to a computer and associated with poorer EF and proprioception but not with
display screen (1280 × 768 pixels; 60 Hz; 58 cm) (Figure 1). any other sensorimotor measures.
An arrow was presented in the centre of the screen pointing Twenty-nine participants (28%) reported falling in the
in one of four directions (up, down, left and right) that past year. Characteristics of fallers and non-fallers are
matched the four possible step directions (forward, back- reported in Table 1. Fallers took longer to complete each
ward, left and right). A word indicating a different direction SST trial (3.0 ± 1.2 versus 2.6 ± 0.6 s) and made more step-
was written inside the arrow. Participants were instructed to ping errors (1.7 ± 2.2 versus 0.6 ± 1.1) than non-fallers.
‘Step by the word’ and therefore had to inhibit the response They also scored worse on postural sway, knee extension
indicated by the arrow’s orientation. Four practice trials, strength, AST and fear of falling. No difference was observed
followed by a random sequence of 20 trials in which the in any of the cognitive measures except for delayed recall,
directions of word and orientation never matched were admi- where participants with better memory scores were more
nistered. The required step size was small (25–30 cm). The likely to have reported a fall.
average time (ms) to complete a trial (excluding trials with After adjusting for sensorimotor and cognitive explana-
error) and number of errors were recorded. tory variables related to falls in older people, each error
made on the SST task increased the odds of having had a fall
1.7 times [OR: 1.65 (1.17–2.34)]. Fear of falling, sway and
Fall-risk assessment delayed recall also independently discriminated fallers from
Sensorimotor function was assessed with visual contrast sen- non-fallers in the final model when adjusting for potential
sitivity, proprioception, knee extension strength, simple reac- confounding factors [Supplementary data are available in Age
tion time and postural sway [11]. Choice Stepping Reaction and Ageing online, Appendix 2, Table S2)].

Stroop stepping test and fall risk

Table 1. Stepping, sensorimotor, cognitive, mobility and psychological characteristics of fallers and non-fallers (means ± SD)
Variable Fallers (n = 29) Non-fallers (n = 74) ORa (95% CI) P-value
SST time/trial (s)b 3.0 ± 1.2 2.6 ± 0.6 1.72 (1.02–2.91) 0.044
SST errorb 1.7 ± 2.2 0.6 ± 1.1 1.53 (1.14–2.07) 0.005
Age (years) 80.2 ± 5.8 79.3 ± 4.4 1.04 (0.95–1.14) 0.394
Gender % (female) 72.4% 69.0% 1.18 (0.46–3.07) 0.728
MMSE (0–30) 28.9 ± 1.0 28.9 ± 1.2 0.99 (0.67–1.46) 0.970
Recall (0/1) % (correct) 79.3% 58.1% 2.771 (1.008–7.614) 0.048
Pain lower extremities (0/1) % (yes) 75.9% 59.5% 2.14 (0.81–5.65) 0.123
Use of walking aids (0/1) % (yes) 17.2% 6.8% 2.88 (0.77–10.80) 0.118
CSRT RT (ms)b 835 ± 130 790 ± 95 1.004 (1.004–1.008) 0.066
CSRT MT (ms)b 286 ± 91 277 ± 63 1.002 (0.996–1.008) 0.535
DSST (number of correct symbols)c 39.8 ± 11.0 38.7 ± 9.1 1.004 (0.959–1.050) 0.879
TMT A (s)b,c 42.8 ± 12.9 42.7 ± 12.8 1.02 (0.99–1.05) 0.332
TMT B (s)b,c 107.3 ± 40.7 105.2 ± 41.3 1.005 (0.995–1.014) 0.361
TMT B-A (s)b,c 64.9 ± 31.2 62.5 ± 35.3 1.006 (0.994–1.018) 0.337
Stroop symbolsc 0.69 ± 0.19 0.68 ± 0.22 1.39 (0.16–11.99) 0.763
Stroop errorsb,c 1.07 ± 1.7 1.05 ± 3.1 1.003 (0.855–1.177) 0.970
Falls efficacy (10–40)b 20.2 ± 5.2 16.9 ± 4.4 1.16 (1.05–1.28) 0.002
Sway velocity (mm)b 362 ± 152 315 ± 127 1.003 (0.999–1.006) 0.114
SwayAP (mm)b 42 ± 16 34 ± 13 1.035 (1.005–1.065) 0.021
SwayML (mm)b 47 ± 20 38 ± 17 1.025 (1.001–1.049) 0.040
Contrast sensitivity (dB) 20.9 ± 2.5 21.4 ± 1.7 0.88 (0.70–1.10) 0.250
Knee extension (kg) 26.0 ± 6.6 30.0 ± 9.1 0.94 (0.89–0.97) 0.037
Proprioception (degrees)b 2.5 ± 1.8 2.3 ± 1.7 1.07 (0.83–1.37) 0.598
Hand reaction time (ms)b 240 ± 55 245 ± 43 0.998 (0.988–1.007) 0.615
TUG (s)b 11.7 ± 3.1 10.6 ± 2.9 1.12 (0.97–1.29) 0.119
AST (s)b 13.1 ± 5.6 10.8 ± 4.1 1.11 (1.01–1.21) 0.028
Per unit change.
Lower values indicate better performance.
Adjusted for education.
OR, odds ratio; CI, confidence interval; SST, Stroop Stepping Test; MMSE, Mini-Mental State Examination; CSRT, Choice Stepping Reaction Time; RT, reaction
time; MT, movement time; DSST, Digit Symbol Substitution Test; TMT, Trailmaking Test; Stroop CW symbols, number of correct named colour-symbols in
incongruent colour-word task per second; Stroop CW errors, number of errors made in incongruent colour-word task; AP, antero-posterior; ML, medio-lateral; TUG,
Timed-up and go test; AST, Alternate Step Test; variables entered in multivariate logistic regression in bold (P < 0.15).

Discussion Fear of falling is associated with self-reported and

performance-based measures of psychological, cognitive and
The time to complete a SST trial correlated significantly with physical function [20]. The inclusion of this measure
measures of EF, processing speed, balance and step coordin- suggests that it measures an aspect of fall-risk that is not
ation. This suggests SST time shares characteristics with a completely encompassed by balance and stepping measures.
range of physical and cognitive functioning measures that are Several studies have found that standard neuropsycho-
associated with fall-risk. The SST may therefore be consid- logical measures of EF predict falls in older people [4, 7, 8].
ered a composite measure and may provide a more robust However, in line with others, we could not confirm this [21,
estimate of fall-risk compared with clinical tests that assess 22]. These discrepancies may be due to differing sample
individual components. The number of errors made during characteristics, faller group classifications and neuropsycho-
the SST was an independent and significant risk factor for logical measures used. In contrast, the time and error mea-
falls in the final multivariate model that only weakly corre- sures of the SST, which combine EF within a stepping task,
lated with EF but not with most other measures. This indi- showed differences between fallers and non-fallers. This sug-
cates that SST errors might provide important additional gests that the composite measurement of inhibition and step-
information regarding fall-risk. Fallers required 0.4 s longer ping in a functional context increases discriminative ability
to complete each SST trial and made 1.1 more errors during for fall-risk. This is in agreement with recent findings
20 trials, suggesting that the difference between faller groups showing that the incorporation of Stroop tasks into real-
is not due to a speed-accuracy trade-off but rather to world scenarios improved the ability to detect functional dif-
impaired functional performance. Our findings are in agree- ferences in healthy and cognitively impaired people [23, 24].
ment with previous results showing that inhibition is an Finally, we found that older adults with better recall ability
important factor in step initiation and that a deficit leads to were more likely to report falls, most likely reflecting better
increased error rates and slowed reaction times [9, 10]. fall events recall [25].
The multivariate model included SST, postural sway and The SST is portable, easy to set-up and quick to adminis-
fear of falling. Sway is a well-known risk factor for falls [19]. ter. All participants could complete the task including those

D. Schoene et al.

who normally used walking aids. The lack of a floor effect is data collection and analysis, decision to publish, or prepar-
important as often those people at higher risk are unable to ation of the manuscript.
complete assessments due to difficulty of the involved tasks.
The low-cost of the equipment also makes it more feasible
for translation into clinical practice than highly specialised la- Supplementary data
boratory equipment.
We acknowledge that this study has certain limitations. Supplementary data mentioned in the text is available to
The retrospective recording of falls may have underestimated subscribers in Age and Ageing online.
the true prevalence of fall events. However, a history of falls
has been shown to be a good predictor of future falls [26]
and by including a delayed recall item in the analysis, we tried References
to control for recall bias. Second, older people with physical
frailty or cognitive impairment might not be able to under- 1. Maki BE, McIlroy WE. Control of rapid limb movements for
take the SST possibly limiting its external validity. Finally, the balance recovery: age-related changes and implications for fall
SST in its current form did not include any cross-over steps, prevention. Age Ageing 2006; 35: ii12–8.
a possible movement strategy for avoiding falls. 2. Maki BE, Zecevic A, Bateni H, Kirshenbaum N, McIlroy WE.
Cognitive demands of executing postural reactions: does aging
impede attention switching? NeuroReport 2001; 12: 3583–7.
3. Chapman GJ, Hollands MA. Evidence that older adult fallers
Conclusions prioritise the planning of future stepping actions over the ac-
curate execution of ongoing steps during complex locomotor
The SST, a screening test that combines stepping with
tasks. Gait Posture 2007; 26: 59–67.
response inhibition, was able to distinguish between older 4. Lord SR, Fitzpatrick RC. Choice stepping reaction time: a
fallers and non-fallers. The current study shows the feasibility composite measure of fall-risk in older people. J Gerontol A
of using a low-cost video game device for the assessment of Biol Sci Med Sci. 2001; 56: M627–32.
older people. Further studies are needed to prospectively val- 5. Melzer I, Kurz I, Shahar D, Levi M, Oddsson L. Application
idate this test, assess its ability to detect clinically meaningful of the voluntary step execution test to identify elderly fallers.
changes and determine its reliability. Age Ageing 2007; 36: 532–7.
6. Hsu CL, Nagamatsu LS, Davis JC, Liu-Ambrose T.
Examining the relationship between specific cognitive pro-
Key points cesses and fall-risk in older adults: a systematic review.
Osteoporos Int 2012; 23: 2409–24.
• A SST that combines executive functioning and rapid step- 7. Anstey KJ, Wood J, Kerr G, Caldwell H, Lord SR. Different
ping using computer game technology was developed. cognitive profiles for single compared with recurrent fallers
• In contrast to standard neuropsychological tests, the SST without dementia. Neuropsychology 2009; 23: 500–8.
was able to discriminate between fallers and non-fallers. 8. Mirelman A, Herman T, Brozgol M et al. Executive function
and falls in older adults: new findings from a five-year pro-
• By combining both cognitive and physical factors, the SST
spective study link fall-risk to cognition. PLoS ONE 2012; 7:
is a functionally relevant means of assessing fall-risk. e40297.
9. Sparto PJ, Fuhrman SI, Redfern MS et al. Postural adjustment
errors reveal deficits in inhibition during lateral step initiation
in older adults. J Neurophysiol 2012; 109: 415–28.
Acknowledgements 10. Uemura K, Oya T, Uchiyama Y. Effects of visual interference
on initial motor program errors and execution times in the
The authors would like to thank the staff of the ARV retire- choice step reaction. Gait Posture 2013; 38: 68–72.
ment villages for their help in this project. 11. Lord SR, Menz HB, Tiedemann A. A physiological profile
approach to fall-risk assessment and prevention. Phys Ther
2003; 83: 237–52.
Conflicts of interest 12. Schoene D, Lord SR, Verhoef P, Smith ST. A novel video
game–based device for measuring stepping performance and
None declared. fall-risk in older people. Arch Phys Med Rehabil 2011; 92:
13. Podsiadlo D, Richardson S. The Timed ‘Up & Go’: a test of
Funding basic functional mobility for frail elderly persons. J Am Geriatr
Soc 1991; 39: 142–8.
This work was supported by a National Health and Medical 14. Tiedemann A, Shimada H, Sherrington C, Murray S, Lord S.
Research Council (NHMRC) Partnership Grant (grant The comparative ability of eight functional mobility tests for
number 568975), an International Postgraduate Research predicting falls in community-dwelling older people. Age
Scholarship (DS), a NHMRC Career Development Ageing 2008; 37: 430–5.
Fellowship (KD) and a NHMRC Senior Principal Research 15. Salthouse TA. What do adult age differences in the digit
Fellowship (SRL). The funders had no role in study design, symbol substitution test reflect? J Gerontol 1992; 47: P121–8.

C-reactive protein, APOE genotype

16. Lezak MD, Howieson DB, Lohring DW. Neuropsychological older adults: a prospective study linking thinking, walking, and
assessment, 4th edition. New York: Oxford University Press, falling. J Gerontol A Biol Sci Med Sci 2010; 65: 1086–92.
2004. 22. Berry SD, Quach L, Procter-Gray E et al. Poor adherence to
17. Delbaere K, Smith ST, Lord SR. Development and initial valid- medications may be associated with falls. J Gerontol A Biol Sci
ation of the iconographical falls efficacy scale. J Gerontol A Med Sci 2010; 65: 553–8.
Biol Sci Med Sci. 2011; 66: 674–80. 23. Parsons TD, Courtney CG, Arizmendi B, Dawson M. Virtual
18. Lamb SE, Jørstad-Stein EC, Hauer K, Becker C, Prevention of Reality Stroop Task for neurocognitive assessment. Stud
Falls Network Europe and Outcomes Consensus Group. Health Technol Inform 2011; 163: 433–9.
Development of a common outcome data set for fall injury 24. Perrochon A, Kemoun G, Watelain E, Berthoz A. Walking
prevention trials: the prevention of falls network europe con- Stroop carpet: an innovative dual-task concept for detecting
sensus. J Am Geriatr Soc 2005; 53: 1618–22. cognitive impairment. Clin Interv Aging 2013; 8: 317–28.
19. Piirtola M, Era P. Force platform measurements as predictors 25. Delbaere K, Close JCT, Kochan NA, Sachdev PS, Lord SR.
of falls among older people: a review. Gerontology 2006; 52: Good memory as a predictor of falls: fact or artefact. Age
1–16. Ageing 2012; 41: 404–8.
20. Deshpande N, Metter EJ, Bandinelli S, Lauretani F, Windham 26. Deandrea S, Lucenteforte E, Bravi F, Foschi R, La Vecchia C,
BGFerrucci L. Psychological, physical, and sensory correlates Negri E. Risk factors for falls in community-dwelling older
of fear of falling and consequent activity restriction in the people: a systematic review and meta-analysis. Epidemiology
elderly: the InCHIANTI study. Am J Phys Med Rehabil 2008; 2010; 21: 658–68.
87: 354–62.
21. Herman T, Mirelman A, Giladi N, Schweiger A, Hausdorff Received 23 April 2013; accepted in revised form 26 June
JM. Executive control deficits as a prodrome to falls in healthy 2013

Age and Ageing 2014; 43: 289–292 © The Author 2013. Published by Oxford University Press on behalf of the British Geriatrics
doi: 10.1093/ageing/aft193 Society. This is an Open Access article distributed under the terms of the Creative Commons
Published electronically 3 December 2013 Attribution License (, which permits unrestricted
reuse, distribution, and reproduction in any medium, provided the original work is properly cited.

C-reactive protein, APOE genotype and

longitudinal cognitive change in an older
Hospital Militar de Área de Brasília, Brasília, Brazil
Wolfson Diabetes and Endocrine Clinic, Institute of Metabolic Sciences, Addenbrooke’s Hospital, Cambridge CB2 2QQ, UK
Department of Public Health and Primary Care, University of Cambridge, Cambridge CB2 0SR, UK
MRC Biostatistics Unit, Cambridge CB2 0SR, UK
Department of Psychology, University of Edinburgh, 7 George Square, Edinburgh, EH8 9JZ, UK
Centre for Cognitive Ageing and Cognitive Epidemiology, University of Edinburgh, 7 George Square, Edinburgh, EH8 9JZ, UK
Address correspondence to: R. E. Marioni. Tel: 0044 131 650 3422. Email:

Background: circulating measures of inflammatory markers, such as C-reactive protein (CRP) have been associated with an
increased risk of future cognitive decline. However, the nature of the relationship among the very old (>75 years) is unclear.
Cross-sectional evidence suggests that elevated CRP may even be protective in this age group. This study examines these asso-
ciations longitudinally.