Gait & Posture 41 (2015) 13–18
Contents lists available at ScienceDirect
Gait & Posture journal homepage: www.elsevier.com/locate/gaitpost
Normative data for arm swing asymmetry: How (a)symmetrical are we? A. Plate a, D. Sedunko a, O. Pelykh b, C. Schlick b, J.R. Ilmberger b, K. Bo¨tzel a,* a b
Department of Neurology, Ludwig-Maximilians-University Munich, Marchioninistr. 15, 81377 Munich, Germany Department of Orthopedics, Physical Medicine and Rehabilitation, Ludwig-Maximilians-University Munich Marchioninistr. 15, 81377 Munich, Germany
A R T I C L E I N F O
A B S T R A C T
Article history: Received 18 February 2014 Received in revised form 10 July 2014 Accepted 14 July 2014
Arm swing asymmetry during gait may be a sensitive sign for early Parkinson’s disease. There is only very limited information about how much asymmetry can be considered to be physiological. To assess the normal range of arm swing asymmetry, we investigated 60 healthy subjects. The influence of age, gender, and additional mental tasks (dual-tasking) on arm swing asymmetry was assessed. Limb kinematics of 60 healthy persons in three age groups (between 40 and 75 years) were measured with an ultrasound motion capture system while subjects walked on a treadmill. Treadmill velocity was varied (3 steps) and mental loads (2 different tasks) were applied in different trials. Additionally, a group of 7 patients with early Parkinson’s disease was investigated. Arm swing amplitude as well as arm swing asymmetry varied considerably in the healthy subjects. Elderly subjects swung their arms more than younger participants. Only the more demanding mental load caused a significant asymmetry, i.e., arm swing was reduced on the right side. In the patient group, asymmetry was considerably higher and even more enhanced by mental loads. Our data indicate that an asymmetry index above 50 (i.e., one side has twice the amplitude of the other) may be considered abnormal. Evaluation of arm swing asymmetry may be used as part of a test battery for early Parkinson’s disease. Such testing may become even more important when disease-modifying drugs become available for Parkinson’s disease. ß 2014 Elsevier B.V. All rights reserved.
1. Introduction In Parkinson’s disease (PD) 60% of the dopaminergic neurons of the S. nigra are lost even before motor symptoms appear [1]. The reduced arm swing during gait, which initially appears one-sided, might be one of the first motor symptoms because subconscious movement generation relies heavily on intact dopaminergic innervation [2]. Even though the exact point in time at which to start treatment in PD remains a controversial topic, there is some evidence that early medical treatment has a positive impact on progression [3,4]. As a consequence, recording of arm swing and gait parameters may be an adequate method to detect PD at an early stage. Several studies on this topic were recently presented.
Lewek et al. (2010) state that one-sided reduced arm swing is a clear and early indication of PD [5]. Huang et al. (2012) furthermore describe reduced coordination and increased arm swing asymmetry in PD patients [6]. The study by Roggendorf et al. (2012) confirmed these results and also reported reduced active arm retroversion on the more affected side [7]. The aim of the present study was to establish a normative database for gait-related arm swing with respect to different ages, left–right asymmetry, male–female differences, and the influence of cognitive tasks because additional cognitive tasks can have a significant effect on gait speed of elderly subjects [8] and an even greater effect in PD patients [9]. 2. Material and methods
* Corresponding author. Tel.: +498970953673. E-mail addresses: [email protected] (A. Plate), [email protected] (D. Sedunko), [email protected] (O. Pelykh), [email protected] (C. Schlick), [email protected] (J.R. Ilmberger), [email protected] (K. Bo¨tzel). http://dx.doi.org/10.1016/j.gaitpost.2014.07.011 0966-6362/ß 2014 Elsevier B.V. All rights reserved.
2.1. Participants Sixty healthy persons between the age of 40 and 75 years were examined (mean 55.3). The subjects were assigned to three subgroups: aged 40–49 years, 50–59 years, and 60–75 years. In
A. Plate et al. / Gait & Posture 41 (2015) 13–18
14
values indicate a larger amplitude on the right side. An ASI of 33.33 indicates that the amplitude on the left is 1.5 times the value of that on the right (ASI of 50: ampleft = 2 ampright). For further analysis, ASI and absolute ASI were used.
each group, there were 10 women and 10 men. All subjects gave written informed consent for their participation in the study that was approved by the local ethics committee. All but one subject fulfilled the criteria for right-handedness according to the Edinburgh Handedness Inventory [10]. All subjects answered standardized questionnaires and underwent a neurological assessment to exclude persons with Parkinson’s disease, stroke, dementia, polyneuropathy, arthrosis or heart failure that might influence walking or arm swing. In addition, seven PD patients in an early disease state were analyzed (Hoehn and Yahr 1–1.5, disease duration less than 3 years, age 40–80 years, mean 57.3). Four patients were treated with dopamine-agonists and the remainder was untreated. Four patients had a left-sided and 3 a right-sided preponderance of PD symptoms. All but one patient fulfilled the criteria for righthandedness according to the Edinburgh Handedness Inventory [10].
2.3. Gait protocol Walking at velocities of 3 and 4 km/h and at the speed preferred by the individual was recorded for a minimum of 30 s. This allowed analysis of at least 20 gait cycles per person and type of walking. Also two additional cognitive tasks were given: counting backwards and a Stroop test, in which the subjects had to name the color of the words ‘red,’ ‘blue,’ and ‘green’ which were displayed in differing colors on a large screen in front of the subject. In these tests, the subject walked at his/her preferred gait speed. 2.4. Statistical analysis
2.2. Recording methods
Data was analyzed with SPSS 20 statistical software (IBM SPSS Statistics). Normal distribution of the data was confirmed with the Kolmogorov–Smirnov test. The influence of velocity, gender, age, and additional cognitive tasks on arm swing amplitude and arm swing asymmetry was evaluated with ANOVAs for repeated measures and post-hoc t-tests, if indicated.
Gait-related arm swing measurements were recorded on a treadmill (‘‘quasar’’ by h/p/cosmos sports & medical gmbh, Nussdorf-Traunstein, Germany). The position of the shoulders, elbows, and wrists were recorded with six ultrasound emitters (Zebris CMS20, sampling frequency 50 Hz, resolution 0.5 mm). Analysis and graphic presentation of arm swing data were performed with a Matlab algorithm (Matlab 7.0, MathWorks, Natick, USA). The angular position of the upper arm (against earth vertical) was computed using the markers over the deltoid muscle and the elbow. After visual inspection of the curves, a 30-s epoch with regular arm movements was selected. Then, peaks and troughs were automatically detected and the angular distance of adjacent peaks and troughs of the upper arm (arm swing angle) was computed (Fig. 1). An asymmetry index (ASI) was calculated using the following formula [11]:
ASI ¼
3. Results 3.1. Healthy subjects: differences of arm swing amplitude related to gender, side and gait velocity, and age Group mean arm swing amplitudes were in the range of 20–26 degrees in healthy subjects as well as in the patients (Table 1). However, the individual variability was much higher (Fig. 2). To evaluate differences in arm swing amplitude related to gait speed, gender, and side, an ANOVA for repeated measures was computed. There was a clear significant positive influence of ‘gait velocity’ (F2,58 = 31, p < 0.001) but ‘gender’ and ‘side’ were not significant although there was a tendency for women to have larger amplitudes than men and for the left arm to swing more than the right (Fig. 2). The influence of age on arm swing amplitude (right arm, left arm; Fig. 3) was investigated with Pearson’s product-moment correlation. The correlation was positive, low, but significant (right arm: r = 0.18, p = 0.011; left arm: r = 0.15, p = 0.039).
LR 100; maxðL; RÞ
3.2. Evaluation of asymmetry
where L and R denote the mean left and right arm swing angle during a 30-s recording epoch. 0 indicates full symmetry, positive values indicate a larger amplitude on the left side, and negative
The mean ASI during 3 km/h gait velocity was 16.4 (i.e., a 19% larger amplitude on the left), reflecting the non-significant left-sided preponderance. The absolute ASI, which reflects overall asymmetry without regard to side, was 32.7 (49% larger
20
[ deg ]
15 10 5
right arm
-10 0
5
10
15
0
5
10
15
20
25
30
35
20
25
30
35
15
[ deg ]
10
left arm
Arm Swing Amplitude
0 -5
5 0 -5 -10
Time [ s ] Fig. 1. Graphs of the amplitudes of the arm swing of the right and left arm. The upper row shows the arm swing amplitude of the right arm, the lower graph shows the amplitude of the left arm over a period of 35 s of gait. Arm swing amplitude (dotted line) was defined as the difference between peaks and troughs of the time-varying arm position (average from approx. 20 cycles).
A. Plate et al. / Gait & Posture 41 (2015) 13–18
15
Table 1 Average arm swing parameters of healthy subjects and patients in four different conditions. L arm, R arm: arm swing amplitude of upper arm in degrees. ASI: asymmetry index; s.p.: self-preferred). Velocity 3 km/h
Velocity 4 km/h
Stroop test (s.p. velocity)
Counting backwards (s.p. velocity)
L arm
R arm
ASI
ASI abs
L arm
R arm
ASI
ASI abs
L arm
R arm
ASI
ASI abs
L arm
R arm
ASI
ASI abs
Healthy
Mean Std dev Confid interv of mean (95%)
22.6 7.5
19.8 11.4
16.4 36.0
32.7 22.0 27.6–39.1
26.2 7.9
25.0 12.3
9.3 33.7
27.9 20.8 22.9–33.8
26.1 9.4
20.0 11.2
24.2 36.2
36.6 23.4 30.5–42.7
28.2 7.3
26.0 12.4
11.5 35.7
30.1 22.0 24.8–36.3
Patients
Mean Std dev
20.9 16.6
20.0 9.8
3.2 69.7
57.4 32.1
24.6 16.4
24.7 11.2
0.7 64.4
53.5 28.5
15.4 12.1
14.1 14.0
9.7 84.1
76.5 19.1
21.0 14.4
19.4 14.0
4.5 75.4
64.7 28.8
amplitude, Table 1, Fig. 3). An ANOVA for repeated measures with the factors velocity,’ ‘gender,’ and ‘age group’ was computed for ASIabs. There was a tendency for reduced ASIabs at higher velocities, but the factor ‘velocity’ failed to reach statistical significance (F2,54 = 2.6, p = 0.082). None of the other factors (gender, age group) were significant. 3.3. Influence of cognitive tasks on arm swing asymmetry Cognitive tasks were performed at preferred gait speed (mean 3.4 km/h, range 2.1–5.1 km/h). Comparing ASI at ‘preferred speed’ with ASI in ‘counting backwards’ and the ‘Stroop Test’ (ANOVA for repeated measures) resulted in a highly significant factor ‘Test’ (F2,57 = 11.2, p < 0.01) which was clearly due to an ASI of 24.2 during the Stroop test (Table 1) as shown by post-hoc t-tests. 3.4. Patients: amplitude and asymmetry index Mean arm swing amplitude apparently did not differ much between patients and control subjects (Table 1). However, amplitude differences between healthy subjects and patients became obvious during dual tasks. Absolute asymmetry of arm swing was considerably higher in the patient group compared to the control group (Table 1, Fig. 4). Without additional cognitive tasks, ASIabs was higher than 50 in 5/7 patients (Fig. 4). With additional tasks, none of the patients had an ASIabs of less than 50. Because of the small size of the patient group, a formal ROC analysis was not attempted. However, with an ASIabs of 50, the specificity of the test is 0.79 (based on data of healthy subjects) and the sensitivity is 0.666 (based on patient data).
60
The functional role of arm swing during gait and whether it is an active or a passive phenomenon are still the subject of much discussion [12]. Energy consumption during gait has been shown to be reduced by arm swing, which reduces the ground reaction moment during the impact of the foot on the ground [13]. There is evidence that the arms act as passive mass dampers during human walking and running although the evidence is clearest for running [14]. However, other authors conclude that active and passive mechanisms are involved in the generation of arm swing [15]. Passive dynamic walking robots move their arms in a humanoid fashion that supports the contribution of passive phenomena [13]. Arm swing may be a leftover from quadrupedal limb coordination realized by specialized thoraco-lumbar spinal fibers and midbrain centers which seem to be intact in Parkinson’s
40 35 30 25 20 left arm
15
right arm
10 5
40
0 40-49
50-59
60-75
Age group
30
20
10
0 M R
3 km/h M F L R
F L
M R
s.p.speed M F L R
F L
M R
4 km/h M F L R
F L
Fig. 2. Arm swing amplitude related to gait velocity (3 km/h; self-preferred (s.p.) speed; 4 km/h), gender (M/F), and side (R/L). The box denotes the 50% percentile, whiskers denote the 95% percentile. Horizontal line: median. There is a considerable variability of arm swing amplitude in healthy subjects. Arm swing amplitude increases with gait velocity, women tend to have larger arm swing amplitude (not significant (n.s.)), and the left arm is generally swung more (n.s.).
Asymmetry - Index [ absolute ]
Arm Swing Amplitude [ deg ]
50
4. Discussion
Amplitude [ deg ]
Group
70 60 50 40 40-49
30
50-59
20
60-75
10 0 3 km/h V1
s.p. spee V2 d
4 km/h V3
Velocity Fig. 3. Upper panel: age-related changes of arm swing amplitude during preferred gait speed in healthy subjects. Standard deviation is indicated. Lower panel: agerelated changes of absolute asymmetry index for healthy subjects in 3 different age groups and standard deviation (3 km/h; self-preferred (s.p.) speed; 4 km/h).
A. Plate et al. / Gait & Posture 41 (2015) 13–18
16
100
90
80
Asymmetry Index absolute
70
60
50
40
30
20
10
0 3km/h CONT
PAT
s.p.speed CONT
PAT
4km/h CONT
PAT
100back CONT
PAT
Stroop CONT
PAT
Fig. 4. Absolute ASI of control group (CONT) and patients (PAT) in different conditions (3 km/h; self-preferred (s.p.) speed; 4 km/h; counting 100 backwards at s.p. speed, Stroop test at s.p. speed). The box denotes the 50% percentile, whiskers denote the 95% percentile. Horizontal line: median. Data of individual patients (n = 7) shown as circles. Note that dual-tasking (counting and Stroop test) increases arm swing asymmetry of 3 patients who had ‘normal’ asymmetry in the other conditions.
disease [16]. The clinical observation that arm swing asymmetry is increased in early PD has been confirmed in several studies [5– 7,17]. The reason for this is less clear. Rigidity of the more affected side may contribute as a passive phenomenon [18] and attenuation of automated movements may play a role as well [2]. The intention of the present study is to establish a database for ‘normal’ arm swing and to define the impact of the factors ‘gait speed,’ ‘age,’ ‘gender,’ and ‘additional mental tasks.’ As previous studies have shown [5], we found that amplitude of arm swing correlates with gait speed, resulting in higher amplitudes when walking faster. Furthermore, in our data the left arm swung more, especially in females and females also tended to have higher amplitudes in general. Age seems to be an important factor since our data showed a significant positive correlation between arm swing amplitude and age, meaning that the elderly swing their arms more. Since the passive factors contributing to arm swing should be comparable, it is most likely an active process that seems to be increased in the elderly due to higher demands for gait and postural stabilization. This assumption is supported by the observation that, even in PD, trunk rotation magnitude does not differ from that in control subjects [5] and arm reactions are used more in the elderly when compensating for sudden postural disturbances [19]. Furthermore there have been discussions about whether the elderly have increased energy consumption during gait, but arm swing reduces energy consumption during gait in young and elderly subjects [20]. This might highlight the need for regular arm swing in the elderly as a possible mechanism to save energy. There have been examinations of sex-specific patterns of trunk muscle coordination [21] that mainly detected differences in mean amplitude levels and their proportions as well as activations during stride. In our study the factor ‘gender’ does not influence arm swing significantly. However, previous studies
have not specifically looked at gender-specific differences in arm swing. Asymmetry of arm swing amplitude was considerable in our set of 60 healthy persons, which indicates that strictly symmetrical controlled arm swing movements may not be necessary for normal gait. This asymmetry was not significantly influenced by age or velocity. The asymmetry index found in our study is in the range of previous findings [11]. These authors reported an ASI of 12.5 that is close to our 10.1 (self-preferred velocity, not in Table 1), not taking the age into account. These values reflect a left-sided preponderance of arm swing amplitude which was seen in both right- and left-handed healthy subjects [11] and thus is not related to handedness. In general amplitude was smaller in the PD group, as reported before [7], whereas arm swing asymmetry, especially in early stages seems to detect differences more reliably [5]. As a threshold for maximum ‘normal’ asymmetry we suggest an ASIabs of 50, which indicates that one side has twice the amplitude of the other. This value is only 1.5 times the mean of the ASIabs of our healthy subjects, thus providing a good sensitivity. An ASIabs of 50 signifies a specificity of 0.79 (based on our data from 60 healthy subjects), which seems to be acceptable for a screening test. With our small group of patients, it is not possible to obtain reliable data on sensitivity. Lewek at al. [5], who used another definition of ASI, suggested a threshold of 7.4% which corresponds to 2.4 times the mean ASI of their healthy subjects (n = 8). Asymmetry was greatly enhanced by putting a heavy, but not a light, mental load on our healthy subjects. Only the attentiondemanding Stroop test, but not counting backwards, increased the ASI of the healthy subjects. In the patients, the effect of the Stroop test seemed to be a lot stronger; both arms reduced their swing amplitude (Table 1) and the ASI increased enormously. Counting backwards resulted in a higher ASI in the PD group, but
A. Plate et al. / Gait & Posture 41 (2015) 13–18
it stayed fairly stable in the HC group. Number counting causes a bihemispheric activation in comparison to the predominantly left-hemisphere activation of object-naming [22]. Even though there have been differing results, the Stroop test does seem to activate the left hemisphere more due to word processing and language [23]. This might be the reason for less arm swing asymmetry during counting backwards than during the Stroop test and less motor activation of the right arm during the Stroop test. Interestingly, increases of asymmetry were much more obvious in PD patients than in HC in both dual tasks, indicating that the remaining automaticity breaks down in this situation. This might be due to the fact that the subconscious influence of the basal ganglia on motor tasks is reduced in PD. PD patients are therefore forced to use a more conscious pattern, resulting in limitations during dual-tasking, especially in more complex tasks like the Stroop test. It has been shown that gait variability in PD is increased by dual-tasking [9]. The effects of dual-tasking in PD are not improved by medication or deep brain stimulation [24]. In another study with non-PD subjects, this influence was prominent in elderly persons with a history of falls and reduced executive function, whereas there was no increase in gait variability in young and healthy elderly during cognitive load [25]. Effects of dual-tasking on arm swing have rarely been examined. Remarkably, our study revealed that arm swing asymmetry in healthy ‘‘non-faller’’ subjects was clearly influenced by dual-tasking. Second, in our small sample of patients with early PD, challenging dual-tasking massively enhanced arm swing asymmetry and a reduced arm swing amplitude and therefore the effect of a complex second task on arm swing asymmetry should be exploited in further studies dealing with early motor signs in PD. Our findings are supported by the ‘‘posture-second’’ hypothesis as suggested by Bloem et al. [26] who assume that PD patients treat all elements of a task with equal priority, which might lead to balancing difficulties and possible falls. A possible shortcoming of our study is the use of a treadmill device that may have an impact on the gait pattern. Treadmill walking can increase step length and was thus advocated as a therapeutic procedure in PD [27]. However, the influence of the treadmill on arm swing seems to be negligible [28]. Looking at the literature, studies used quite different methods to vary velocity. In some studies velocities were indicated by the examiner verbally, e.g., walk ‘‘as fast as you can’’ [5,28], some chose, for example, 20% less or more of the self-preferred speed [29] and others used fixed velocities [7,11,21]. In our study subjects were asked to walk at their self-preferred speed as well as at fixed velocities. Using fixed velocities might have potential implications due to operating as an additional cognitive task and therefore having an influence on gait parameters. In order to align our data with the above-mentioned studies we chose to use two fixed velocities and preferred speed in addition to the dual tasks. Our results did not indicate that velocity has a significant influence on the ASI of the healthy subjects after all, but there was a correlation of velocity and arm swing amplitude, as reported before. A comparison of different ways of selecting velocities would be very interesting for further studies. However, it remains to be determined whether ‘preferred speed on a treadmill is a reliable individual marker. Our results confirm the potential role for arm swing asymmetry as a marker of early PD. They also highlight the fact that, in healthy subjects, arm swing amplitude can be considerably asymmetric. In addition to other assessments, such as smell or ultrasound of the basal ganglia, arm swing measurements should also be included in a battery for testing early signs of PD, as has been suggested [17,30].
17
Acknowledgments This project was supported financially by Lu¨neburg heritage (no involvement in the study design, the collection, analysis and interpretation of data, in the writing of the manuscript or in the decision to submit the manuscript for publication). Conflict of interest statement Annika Plate: reports no disclosures. Daniela Sedunko: reports no disclosures. Olena Pelykh: reports no disclosures. Cornelia Schlick: reports no disclosures. Josef R. Ilmberger: reports no disclosures. Kai Bo¨tzel: reports no disclosures. References [1] Becker G, Muller A, Braune S, Buttner T, Benecke R, Greulich W, et al. Early diagnosis of Parkinson’s disease. J Neurol 2002;249(Suppl 3):III/40–/48. [2] Redgrave P, Rodriguez M, Smith Y, Rodriguez-Oroz MC, Lehericy S, Bergman H, et al. Goal-directed and habitual control in the basal ganglia: implications for Parkinson’s disease. Nat Rev Neurosci 2010;11(11):760–72. [3] Fahn S, Oakes D, Shoulson I, Kieburtz K, Rudolph A, Lang A, et al. Levodopa and the progression of Parkinson’s disease. N Engl J Med 2004;351(24):2498– 508. [4] Olanow CW, Rascol O, Hauser R, Feigin PD, Jankovic J, Lang A, et al. A doubleblind, delayed-start trial of rasagiline in Parkinson’s disease. N Engl J Med 2009;361(13):1268–78. [5] Lewek MD, Poole R, Johnson J, Halawa O, Huang X. Arm swing magnitude and asymmetry during gait in the early stages of Parkinson’s disease. Gait Posture 2010;31(2):256–60. [6] Huang X, Mahoney JM, Lewis MM, Guangwei D, Piazza SJ, Cusumano JP. Both coordination and symmetry of arm swing are reduced in Parkinson’s disease. Gait Posture 2012;35(3):373–7. [7] Roggendorf J, Chen S, Baudrexel S, van de Loo S, Seifried C, Hilker R. Arm swing asymmetry in Parkinson’s disease measured with ultrasound based motion analysis during treadmill gait. Gait Posture 2012;35(1):116–20. [8] Autenrieth CS, Karrasch S, Heier M, Gorzelniak L, Ladwig KH, Peters A, et al. Decline in gait performance detected by an electronic walkway system in 907 older adults of the population-based KORA-Age study. Gerontology 2013;59(2):165–73. [9] Hausdorff JM, Balash J, Giladi N. Effects of cognitive challenge on gait variability in patients with Parkinson’s disease. J Geriatr Psychiatry Neurol 2003;16(1): 53–8. [10] Oldfield RC. The assessment and analysis of handedness: the Edinburgh inventory. Neuropsychologia 1971;9(1):97–113. [11] Kuhtz-Buschbeck JP, Brockmann K, Gilster R, Koch A, Stolze H. Asymmetry of arm swing not related to handedness. Gait Posture 2008;27(3):447–54. [12] Meyns P, Bruijn SM, Duysens J. The how and why of arm swing during human walking. Gait Posture 2013;38(4):555–62. [13] Collins SH, Adamczyk PG, Kuo AD. Dynamic arm swinging in human walking. Proc Biol Sci 2009;276(1673):3679–88. [14] Pontzer H, Holloway 4th JH, Raichlen DA, Lieberman DE. Control and function of arm swing in human walking and running. J Exp Biol 2009;212(Pt 4):523– 34. [15] Kuhtz-Buschbeck JP, Jing B. Activity of upper limb muscles during human walking. J Electromyogr Kinesiol 2012;22(2):199–206. [16] Dietz V. Quadrupedal coordination of bipedal gait: implications for movement disorders. J Neurol 2011;258(8):1406–12. [17] Schneider SA, Drude L, Kasten M, Klein C, Hagenah J. A study of subtle motor signs in early Parkinson’s disease. Mov Disord 2012;27(12):1563–6. [18] Winogrodzka A, Wagenaar RC, Booij J, Wolters EC. Rigidity and bradykinesia reduce interlimb coordination in Parkinsonian gait. Arch Phys Med Rehabil 2005;86(2):183–9. [19] Maki BE, Edmondstone MA, McIlroy WE. Age-related differences in laterally directed compensatory stepping behavior. J Gerontol A Biol Sci Med Sci 2000;55(5):M270–7. [20] Ortega JD, Fehlman LA, Farley CT. Effects of aging and arm swing on the metabolic cost of stability in human walking. J Biomech 2008;41(16):3303–8. [21] Anders C, Wagner H, Puta C, Grassme R, Scholle HC. Healthy humans use sexspecific coordination patterns of trunk muscles during gait. Eur J Appl Physiol 2009;105(4):585–94. [22] Petrovich Brennan NM, Whalen S, de Morales Branco D, O’Shea JP, Norton IH, Golby AJ. Object-naming is a more sensitive measure of speech localization than number counting: converging evidence from direct cortical stimulation and fMRI. Neuroimage 2007;37(Suppl 1):S100–8. [23] Brown TL, Gore CL, Pearson T. Visual half-field Stroop effects with spatial separation of words and color targets. Brain Lang 1998;63(1):122–42. [24] Seri-Fainshtat E, Israel Z, Weiss A, Hausdorff JM. Impact of sub-thalamic nucleus deep brain stimulation on dual-tasking gait in Parkinson’s disease. J Neuroeng Rehabil 2013;10(1):38.
18
A. Plate et al. / Gait & Posture 41 (2015) 13–18
[25] Springer S, Giladi N, Peretz C, Yogev G, Simon ES, Hausdorff JM. Dual-tasking effects on gait variability: the role of aging, falls, and executive function. Mov Disord 2006;21(7):950–7. [26] Bloem BR, Grimbergen YA, van Dijk JG, Munneke M. The ‘‘posture second’’ strategy: a review of wrong priorities in Parkinson’s disease. J Neurol Sci 2006;248(1–2):196–204. [27] Bello O, Sanchez JA, Fernandez-del-Olmo M. Treadmill walking in Parkinson’s disease patients: adaptation and generalization effect. Mov Disord 2008;23(9): 1243–9.
[28] Carpinella I, Crenna P, Rabuffetti M, Ferrarin M. Coordination between upperand lower-limb movements is different during overground and treadmill walking. Eur J Appl Physiol 2010;108(1):71–82. [29] Krasovsky T, Lamontagne A, Feldman AG, Levin MF. Effects of walking speed on gait stability and interlimb coordination in younger and older adults. Gait Posture 2014;39(1):378–85. [30] Maetzler W, Hausdorff JM. Motor signs in the prodromal phase of Parkinson’s disease. Mov Disord 2012;27(5):627–33.
Contents lists available at ScienceDirect
Gait & Posture journal homepage: www.elsevier.com/locate/gaitpost
Normative data for arm swing asymmetry: How (a)symmetrical are we? A. Plate a, D. Sedunko a, O. Pelykh b, C. Schlick b, J.R. Ilmberger b, K. Bo¨tzel a,* a b
Department of Neurology, Ludwig-Maximilians-University Munich, Marchioninistr. 15, 81377 Munich, Germany Department of Orthopedics, Physical Medicine and Rehabilitation, Ludwig-Maximilians-University Munich Marchioninistr. 15, 81377 Munich, Germany
A R T I C L E I N F O
A B S T R A C T
Article history: Received 18 February 2014 Received in revised form 10 July 2014 Accepted 14 July 2014
Arm swing asymmetry during gait may be a sensitive sign for early Parkinson’s disease. There is only very limited information about how much asymmetry can be considered to be physiological. To assess the normal range of arm swing asymmetry, we investigated 60 healthy subjects. The influence of age, gender, and additional mental tasks (dual-tasking) on arm swing asymmetry was assessed. Limb kinematics of 60 healthy persons in three age groups (between 40 and 75 years) were measured with an ultrasound motion capture system while subjects walked on a treadmill. Treadmill velocity was varied (3 steps) and mental loads (2 different tasks) were applied in different trials. Additionally, a group of 7 patients with early Parkinson’s disease was investigated. Arm swing amplitude as well as arm swing asymmetry varied considerably in the healthy subjects. Elderly subjects swung their arms more than younger participants. Only the more demanding mental load caused a significant asymmetry, i.e., arm swing was reduced on the right side. In the patient group, asymmetry was considerably higher and even more enhanced by mental loads. Our data indicate that an asymmetry index above 50 (i.e., one side has twice the amplitude of the other) may be considered abnormal. Evaluation of arm swing asymmetry may be used as part of a test battery for early Parkinson’s disease. Such testing may become even more important when disease-modifying drugs become available for Parkinson’s disease. ß 2014 Elsevier B.V. All rights reserved.
1. Introduction In Parkinson’s disease (PD) 60% of the dopaminergic neurons of the S. nigra are lost even before motor symptoms appear [1]. The reduced arm swing during gait, which initially appears one-sided, might be one of the first motor symptoms because subconscious movement generation relies heavily on intact dopaminergic innervation [2]. Even though the exact point in time at which to start treatment in PD remains a controversial topic, there is some evidence that early medical treatment has a positive impact on progression [3,4]. As a consequence, recording of arm swing and gait parameters may be an adequate method to detect PD at an early stage. Several studies on this topic were recently presented.
Lewek et al. (2010) state that one-sided reduced arm swing is a clear and early indication of PD [5]. Huang et al. (2012) furthermore describe reduced coordination and increased arm swing asymmetry in PD patients [6]. The study by Roggendorf et al. (2012) confirmed these results and also reported reduced active arm retroversion on the more affected side [7]. The aim of the present study was to establish a normative database for gait-related arm swing with respect to different ages, left–right asymmetry, male–female differences, and the influence of cognitive tasks because additional cognitive tasks can have a significant effect on gait speed of elderly subjects [8] and an even greater effect in PD patients [9]. 2. Material and methods
* Corresponding author. Tel.: +498970953673. E-mail addresses: [email protected] (A. Plate), [email protected] (D. Sedunko), [email protected] (O. Pelykh), [email protected] (C. Schlick), [email protected] (J.R. Ilmberger), [email protected] (K. Bo¨tzel). http://dx.doi.org/10.1016/j.gaitpost.2014.07.011 0966-6362/ß 2014 Elsevier B.V. All rights reserved.
2.1. Participants Sixty healthy persons between the age of 40 and 75 years were examined (mean 55.3). The subjects were assigned to three subgroups: aged 40–49 years, 50–59 years, and 60–75 years. In
A. Plate et al. / Gait & Posture 41 (2015) 13–18
14
values indicate a larger amplitude on the right side. An ASI of 33.33 indicates that the amplitude on the left is 1.5 times the value of that on the right (ASI of 50: ampleft = 2 ampright). For further analysis, ASI and absolute ASI were used.
each group, there were 10 women and 10 men. All subjects gave written informed consent for their participation in the study that was approved by the local ethics committee. All but one subject fulfilled the criteria for right-handedness according to the Edinburgh Handedness Inventory [10]. All subjects answered standardized questionnaires and underwent a neurological assessment to exclude persons with Parkinson’s disease, stroke, dementia, polyneuropathy, arthrosis or heart failure that might influence walking or arm swing. In addition, seven PD patients in an early disease state were analyzed (Hoehn and Yahr 1–1.5, disease duration less than 3 years, age 40–80 years, mean 57.3). Four patients were treated with dopamine-agonists and the remainder was untreated. Four patients had a left-sided and 3 a right-sided preponderance of PD symptoms. All but one patient fulfilled the criteria for righthandedness according to the Edinburgh Handedness Inventory [10].
2.3. Gait protocol Walking at velocities of 3 and 4 km/h and at the speed preferred by the individual was recorded for a minimum of 30 s. This allowed analysis of at least 20 gait cycles per person and type of walking. Also two additional cognitive tasks were given: counting backwards and a Stroop test, in which the subjects had to name the color of the words ‘red,’ ‘blue,’ and ‘green’ which were displayed in differing colors on a large screen in front of the subject. In these tests, the subject walked at his/her preferred gait speed. 2.4. Statistical analysis
2.2. Recording methods
Data was analyzed with SPSS 20 statistical software (IBM SPSS Statistics). Normal distribution of the data was confirmed with the Kolmogorov–Smirnov test. The influence of velocity, gender, age, and additional cognitive tasks on arm swing amplitude and arm swing asymmetry was evaluated with ANOVAs for repeated measures and post-hoc t-tests, if indicated.
Gait-related arm swing measurements were recorded on a treadmill (‘‘quasar’’ by h/p/cosmos sports & medical gmbh, Nussdorf-Traunstein, Germany). The position of the shoulders, elbows, and wrists were recorded with six ultrasound emitters (Zebris CMS20, sampling frequency 50 Hz, resolution 0.5 mm). Analysis and graphic presentation of arm swing data were performed with a Matlab algorithm (Matlab 7.0, MathWorks, Natick, USA). The angular position of the upper arm (against earth vertical) was computed using the markers over the deltoid muscle and the elbow. After visual inspection of the curves, a 30-s epoch with regular arm movements was selected. Then, peaks and troughs were automatically detected and the angular distance of adjacent peaks and troughs of the upper arm (arm swing angle) was computed (Fig. 1). An asymmetry index (ASI) was calculated using the following formula [11]:
ASI ¼
3. Results 3.1. Healthy subjects: differences of arm swing amplitude related to gender, side and gait velocity, and age Group mean arm swing amplitudes were in the range of 20–26 degrees in healthy subjects as well as in the patients (Table 1). However, the individual variability was much higher (Fig. 2). To evaluate differences in arm swing amplitude related to gait speed, gender, and side, an ANOVA for repeated measures was computed. There was a clear significant positive influence of ‘gait velocity’ (F2,58 = 31, p < 0.001) but ‘gender’ and ‘side’ were not significant although there was a tendency for women to have larger amplitudes than men and for the left arm to swing more than the right (Fig. 2). The influence of age on arm swing amplitude (right arm, left arm; Fig. 3) was investigated with Pearson’s product-moment correlation. The correlation was positive, low, but significant (right arm: r = 0.18, p = 0.011; left arm: r = 0.15, p = 0.039).
LR 100; maxðL; RÞ
3.2. Evaluation of asymmetry
where L and R denote the mean left and right arm swing angle during a 30-s recording epoch. 0 indicates full symmetry, positive values indicate a larger amplitude on the left side, and negative
The mean ASI during 3 km/h gait velocity was 16.4 (i.e., a 19% larger amplitude on the left), reflecting the non-significant left-sided preponderance. The absolute ASI, which reflects overall asymmetry without regard to side, was 32.7 (49% larger
20
[ deg ]
15 10 5
right arm
-10 0
5
10
15
0
5
10
15
20
25
30
35
20
25
30
35
15
[ deg ]
10
left arm
Arm Swing Amplitude
0 -5
5 0 -5 -10
Time [ s ] Fig. 1. Graphs of the amplitudes of the arm swing of the right and left arm. The upper row shows the arm swing amplitude of the right arm, the lower graph shows the amplitude of the left arm over a period of 35 s of gait. Arm swing amplitude (dotted line) was defined as the difference between peaks and troughs of the time-varying arm position (average from approx. 20 cycles).
A. Plate et al. / Gait & Posture 41 (2015) 13–18
15
Table 1 Average arm swing parameters of healthy subjects and patients in four different conditions. L arm, R arm: arm swing amplitude of upper arm in degrees. ASI: asymmetry index; s.p.: self-preferred). Velocity 3 km/h
Velocity 4 km/h
Stroop test (s.p. velocity)
Counting backwards (s.p. velocity)
L arm
R arm
ASI
ASI abs
L arm
R arm
ASI
ASI abs
L arm
R arm
ASI
ASI abs
L arm
R arm
ASI
ASI abs
Healthy
Mean Std dev Confid interv of mean (95%)
22.6 7.5
19.8 11.4
16.4 36.0
32.7 22.0 27.6–39.1
26.2 7.9
25.0 12.3
9.3 33.7
27.9 20.8 22.9–33.8
26.1 9.4
20.0 11.2
24.2 36.2
36.6 23.4 30.5–42.7
28.2 7.3
26.0 12.4
11.5 35.7
30.1 22.0 24.8–36.3
Patients
Mean Std dev
20.9 16.6
20.0 9.8
3.2 69.7
57.4 32.1
24.6 16.4
24.7 11.2
0.7 64.4
53.5 28.5
15.4 12.1
14.1 14.0
9.7 84.1
76.5 19.1
21.0 14.4
19.4 14.0
4.5 75.4
64.7 28.8
amplitude, Table 1, Fig. 3). An ANOVA for repeated measures with the factors velocity,’ ‘gender,’ and ‘age group’ was computed for ASIabs. There was a tendency for reduced ASIabs at higher velocities, but the factor ‘velocity’ failed to reach statistical significance (F2,54 = 2.6, p = 0.082). None of the other factors (gender, age group) were significant. 3.3. Influence of cognitive tasks on arm swing asymmetry Cognitive tasks were performed at preferred gait speed (mean 3.4 km/h, range 2.1–5.1 km/h). Comparing ASI at ‘preferred speed’ with ASI in ‘counting backwards’ and the ‘Stroop Test’ (ANOVA for repeated measures) resulted in a highly significant factor ‘Test’ (F2,57 = 11.2, p < 0.01) which was clearly due to an ASI of 24.2 during the Stroop test (Table 1) as shown by post-hoc t-tests. 3.4. Patients: amplitude and asymmetry index Mean arm swing amplitude apparently did not differ much between patients and control subjects (Table 1). However, amplitude differences between healthy subjects and patients became obvious during dual tasks. Absolute asymmetry of arm swing was considerably higher in the patient group compared to the control group (Table 1, Fig. 4). Without additional cognitive tasks, ASIabs was higher than 50 in 5/7 patients (Fig. 4). With additional tasks, none of the patients had an ASIabs of less than 50. Because of the small size of the patient group, a formal ROC analysis was not attempted. However, with an ASIabs of 50, the specificity of the test is 0.79 (based on data of healthy subjects) and the sensitivity is 0.666 (based on patient data).
60
The functional role of arm swing during gait and whether it is an active or a passive phenomenon are still the subject of much discussion [12]. Energy consumption during gait has been shown to be reduced by arm swing, which reduces the ground reaction moment during the impact of the foot on the ground [13]. There is evidence that the arms act as passive mass dampers during human walking and running although the evidence is clearest for running [14]. However, other authors conclude that active and passive mechanisms are involved in the generation of arm swing [15]. Passive dynamic walking robots move their arms in a humanoid fashion that supports the contribution of passive phenomena [13]. Arm swing may be a leftover from quadrupedal limb coordination realized by specialized thoraco-lumbar spinal fibers and midbrain centers which seem to be intact in Parkinson’s
40 35 30 25 20 left arm
15
right arm
10 5
40
0 40-49
50-59
60-75
Age group
30
20
10
0 M R
3 km/h M F L R
F L
M R
s.p.speed M F L R
F L
M R
4 km/h M F L R
F L
Fig. 2. Arm swing amplitude related to gait velocity (3 km/h; self-preferred (s.p.) speed; 4 km/h), gender (M/F), and side (R/L). The box denotes the 50% percentile, whiskers denote the 95% percentile. Horizontal line: median. There is a considerable variability of arm swing amplitude in healthy subjects. Arm swing amplitude increases with gait velocity, women tend to have larger arm swing amplitude (not significant (n.s.)), and the left arm is generally swung more (n.s.).
Asymmetry - Index [ absolute ]
Arm Swing Amplitude [ deg ]
50
4. Discussion
Amplitude [ deg ]
Group
70 60 50 40 40-49
30
50-59
20
60-75
10 0 3 km/h V1
s.p. spee V2 d
4 km/h V3
Velocity Fig. 3. Upper panel: age-related changes of arm swing amplitude during preferred gait speed in healthy subjects. Standard deviation is indicated. Lower panel: agerelated changes of absolute asymmetry index for healthy subjects in 3 different age groups and standard deviation (3 km/h; self-preferred (s.p.) speed; 4 km/h).
A. Plate et al. / Gait & Posture 41 (2015) 13–18
16
100
90
80
Asymmetry Index absolute
70
60
50
40
30
20
10
0 3km/h CONT
PAT
s.p.speed CONT
PAT
4km/h CONT
PAT
100back CONT
PAT
Stroop CONT
PAT
Fig. 4. Absolute ASI of control group (CONT) and patients (PAT) in different conditions (3 km/h; self-preferred (s.p.) speed; 4 km/h; counting 100 backwards at s.p. speed, Stroop test at s.p. speed). The box denotes the 50% percentile, whiskers denote the 95% percentile. Horizontal line: median. Data of individual patients (n = 7) shown as circles. Note that dual-tasking (counting and Stroop test) increases arm swing asymmetry of 3 patients who had ‘normal’ asymmetry in the other conditions.
disease [16]. The clinical observation that arm swing asymmetry is increased in early PD has been confirmed in several studies [5– 7,17]. The reason for this is less clear. Rigidity of the more affected side may contribute as a passive phenomenon [18] and attenuation of automated movements may play a role as well [2]. The intention of the present study is to establish a database for ‘normal’ arm swing and to define the impact of the factors ‘gait speed,’ ‘age,’ ‘gender,’ and ‘additional mental tasks.’ As previous studies have shown [5], we found that amplitude of arm swing correlates with gait speed, resulting in higher amplitudes when walking faster. Furthermore, in our data the left arm swung more, especially in females and females also tended to have higher amplitudes in general. Age seems to be an important factor since our data showed a significant positive correlation between arm swing amplitude and age, meaning that the elderly swing their arms more. Since the passive factors contributing to arm swing should be comparable, it is most likely an active process that seems to be increased in the elderly due to higher demands for gait and postural stabilization. This assumption is supported by the observation that, even in PD, trunk rotation magnitude does not differ from that in control subjects [5] and arm reactions are used more in the elderly when compensating for sudden postural disturbances [19]. Furthermore there have been discussions about whether the elderly have increased energy consumption during gait, but arm swing reduces energy consumption during gait in young and elderly subjects [20]. This might highlight the need for regular arm swing in the elderly as a possible mechanism to save energy. There have been examinations of sex-specific patterns of trunk muscle coordination [21] that mainly detected differences in mean amplitude levels and their proportions as well as activations during stride. In our study the factor ‘gender’ does not influence arm swing significantly. However, previous studies
have not specifically looked at gender-specific differences in arm swing. Asymmetry of arm swing amplitude was considerable in our set of 60 healthy persons, which indicates that strictly symmetrical controlled arm swing movements may not be necessary for normal gait. This asymmetry was not significantly influenced by age or velocity. The asymmetry index found in our study is in the range of previous findings [11]. These authors reported an ASI of 12.5 that is close to our 10.1 (self-preferred velocity, not in Table 1), not taking the age into account. These values reflect a left-sided preponderance of arm swing amplitude which was seen in both right- and left-handed healthy subjects [11] and thus is not related to handedness. In general amplitude was smaller in the PD group, as reported before [7], whereas arm swing asymmetry, especially in early stages seems to detect differences more reliably [5]. As a threshold for maximum ‘normal’ asymmetry we suggest an ASIabs of 50, which indicates that one side has twice the amplitude of the other. This value is only 1.5 times the mean of the ASIabs of our healthy subjects, thus providing a good sensitivity. An ASIabs of 50 signifies a specificity of 0.79 (based on our data from 60 healthy subjects), which seems to be acceptable for a screening test. With our small group of patients, it is not possible to obtain reliable data on sensitivity. Lewek at al. [5], who used another definition of ASI, suggested a threshold of 7.4% which corresponds to 2.4 times the mean ASI of their healthy subjects (n = 8). Asymmetry was greatly enhanced by putting a heavy, but not a light, mental load on our healthy subjects. Only the attentiondemanding Stroop test, but not counting backwards, increased the ASI of the healthy subjects. In the patients, the effect of the Stroop test seemed to be a lot stronger; both arms reduced their swing amplitude (Table 1) and the ASI increased enormously. Counting backwards resulted in a higher ASI in the PD group, but
A. Plate et al. / Gait & Posture 41 (2015) 13–18
it stayed fairly stable in the HC group. Number counting causes a bihemispheric activation in comparison to the predominantly left-hemisphere activation of object-naming [22]. Even though there have been differing results, the Stroop test does seem to activate the left hemisphere more due to word processing and language [23]. This might be the reason for less arm swing asymmetry during counting backwards than during the Stroop test and less motor activation of the right arm during the Stroop test. Interestingly, increases of asymmetry were much more obvious in PD patients than in HC in both dual tasks, indicating that the remaining automaticity breaks down in this situation. This might be due to the fact that the subconscious influence of the basal ganglia on motor tasks is reduced in PD. PD patients are therefore forced to use a more conscious pattern, resulting in limitations during dual-tasking, especially in more complex tasks like the Stroop test. It has been shown that gait variability in PD is increased by dual-tasking [9]. The effects of dual-tasking in PD are not improved by medication or deep brain stimulation [24]. In another study with non-PD subjects, this influence was prominent in elderly persons with a history of falls and reduced executive function, whereas there was no increase in gait variability in young and healthy elderly during cognitive load [25]. Effects of dual-tasking on arm swing have rarely been examined. Remarkably, our study revealed that arm swing asymmetry in healthy ‘‘non-faller’’ subjects was clearly influenced by dual-tasking. Second, in our small sample of patients with early PD, challenging dual-tasking massively enhanced arm swing asymmetry and a reduced arm swing amplitude and therefore the effect of a complex second task on arm swing asymmetry should be exploited in further studies dealing with early motor signs in PD. Our findings are supported by the ‘‘posture-second’’ hypothesis as suggested by Bloem et al. [26] who assume that PD patients treat all elements of a task with equal priority, which might lead to balancing difficulties and possible falls. A possible shortcoming of our study is the use of a treadmill device that may have an impact on the gait pattern. Treadmill walking can increase step length and was thus advocated as a therapeutic procedure in PD [27]. However, the influence of the treadmill on arm swing seems to be negligible [28]. Looking at the literature, studies used quite different methods to vary velocity. In some studies velocities were indicated by the examiner verbally, e.g., walk ‘‘as fast as you can’’ [5,28], some chose, for example, 20% less or more of the self-preferred speed [29] and others used fixed velocities [7,11,21]. In our study subjects were asked to walk at their self-preferred speed as well as at fixed velocities. Using fixed velocities might have potential implications due to operating as an additional cognitive task and therefore having an influence on gait parameters. In order to align our data with the above-mentioned studies we chose to use two fixed velocities and preferred speed in addition to the dual tasks. Our results did not indicate that velocity has a significant influence on the ASI of the healthy subjects after all, but there was a correlation of velocity and arm swing amplitude, as reported before. A comparison of different ways of selecting velocities would be very interesting for further studies. However, it remains to be determined whether ‘preferred speed on a treadmill is a reliable individual marker. Our results confirm the potential role for arm swing asymmetry as a marker of early PD. They also highlight the fact that, in healthy subjects, arm swing amplitude can be considerably asymmetric. In addition to other assessments, such as smell or ultrasound of the basal ganglia, arm swing measurements should also be included in a battery for testing early signs of PD, as has been suggested [17,30].
17
Acknowledgments This project was supported financially by Lu¨neburg heritage (no involvement in the study design, the collection, analysis and interpretation of data, in the writing of the manuscript or in the decision to submit the manuscript for publication). Conflict of interest statement Annika Plate: reports no disclosures. Daniela Sedunko: reports no disclosures. Olena Pelykh: reports no disclosures. Cornelia Schlick: reports no disclosures. Josef R. Ilmberger: reports no disclosures. Kai Bo¨tzel: reports no disclosures. References [1] Becker G, Muller A, Braune S, Buttner T, Benecke R, Greulich W, et al. Early diagnosis of Parkinson’s disease. J Neurol 2002;249(Suppl 3):III/40–/48. [2] Redgrave P, Rodriguez M, Smith Y, Rodriguez-Oroz MC, Lehericy S, Bergman H, et al. Goal-directed and habitual control in the basal ganglia: implications for Parkinson’s disease. Nat Rev Neurosci 2010;11(11):760–72. [3] Fahn S, Oakes D, Shoulson I, Kieburtz K, Rudolph A, Lang A, et al. Levodopa and the progression of Parkinson’s disease. N Engl J Med 2004;351(24):2498– 508. [4] Olanow CW, Rascol O, Hauser R, Feigin PD, Jankovic J, Lang A, et al. A doubleblind, delayed-start trial of rasagiline in Parkinson’s disease. N Engl J Med 2009;361(13):1268–78. [5] Lewek MD, Poole R, Johnson J, Halawa O, Huang X. Arm swing magnitude and asymmetry during gait in the early stages of Parkinson’s disease. Gait Posture 2010;31(2):256–60. [6] Huang X, Mahoney JM, Lewis MM, Guangwei D, Piazza SJ, Cusumano JP. Both coordination and symmetry of arm swing are reduced in Parkinson’s disease. Gait Posture 2012;35(3):373–7. [7] Roggendorf J, Chen S, Baudrexel S, van de Loo S, Seifried C, Hilker R. Arm swing asymmetry in Parkinson’s disease measured with ultrasound based motion analysis during treadmill gait. Gait Posture 2012;35(1):116–20. [8] Autenrieth CS, Karrasch S, Heier M, Gorzelniak L, Ladwig KH, Peters A, et al. Decline in gait performance detected by an electronic walkway system in 907 older adults of the population-based KORA-Age study. Gerontology 2013;59(2):165–73. [9] Hausdorff JM, Balash J, Giladi N. Effects of cognitive challenge on gait variability in patients with Parkinson’s disease. J Geriatr Psychiatry Neurol 2003;16(1): 53–8. [10] Oldfield RC. The assessment and analysis of handedness: the Edinburgh inventory. Neuropsychologia 1971;9(1):97–113. [11] Kuhtz-Buschbeck JP, Brockmann K, Gilster R, Koch A, Stolze H. Asymmetry of arm swing not related to handedness. Gait Posture 2008;27(3):447–54. [12] Meyns P, Bruijn SM, Duysens J. The how and why of arm swing during human walking. Gait Posture 2013;38(4):555–62. [13] Collins SH, Adamczyk PG, Kuo AD. Dynamic arm swinging in human walking. Proc Biol Sci 2009;276(1673):3679–88. [14] Pontzer H, Holloway 4th JH, Raichlen DA, Lieberman DE. Control and function of arm swing in human walking and running. J Exp Biol 2009;212(Pt 4):523– 34. [15] Kuhtz-Buschbeck JP, Jing B. Activity of upper limb muscles during human walking. J Electromyogr Kinesiol 2012;22(2):199–206. [16] Dietz V. Quadrupedal coordination of bipedal gait: implications for movement disorders. J Neurol 2011;258(8):1406–12. [17] Schneider SA, Drude L, Kasten M, Klein C, Hagenah J. A study of subtle motor signs in early Parkinson’s disease. Mov Disord 2012;27(12):1563–6. [18] Winogrodzka A, Wagenaar RC, Booij J, Wolters EC. Rigidity and bradykinesia reduce interlimb coordination in Parkinsonian gait. Arch Phys Med Rehabil 2005;86(2):183–9. [19] Maki BE, Edmondstone MA, McIlroy WE. Age-related differences in laterally directed compensatory stepping behavior. J Gerontol A Biol Sci Med Sci 2000;55(5):M270–7. [20] Ortega JD, Fehlman LA, Farley CT. Effects of aging and arm swing on the metabolic cost of stability in human walking. J Biomech 2008;41(16):3303–8. [21] Anders C, Wagner H, Puta C, Grassme R, Scholle HC. Healthy humans use sexspecific coordination patterns of trunk muscles during gait. Eur J Appl Physiol 2009;105(4):585–94. [22] Petrovich Brennan NM, Whalen S, de Morales Branco D, O’Shea JP, Norton IH, Golby AJ. Object-naming is a more sensitive measure of speech localization than number counting: converging evidence from direct cortical stimulation and fMRI. Neuroimage 2007;37(Suppl 1):S100–8. [23] Brown TL, Gore CL, Pearson T. Visual half-field Stroop effects with spatial separation of words and color targets. Brain Lang 1998;63(1):122–42. [24] Seri-Fainshtat E, Israel Z, Weiss A, Hausdorff JM. Impact of sub-thalamic nucleus deep brain stimulation on dual-tasking gait in Parkinson’s disease. J Neuroeng Rehabil 2013;10(1):38.
18
A. Plate et al. / Gait & Posture 41 (2015) 13–18
[25] Springer S, Giladi N, Peretz C, Yogev G, Simon ES, Hausdorff JM. Dual-tasking effects on gait variability: the role of aging, falls, and executive function. Mov Disord 2006;21(7):950–7. [26] Bloem BR, Grimbergen YA, van Dijk JG, Munneke M. The ‘‘posture second’’ strategy: a review of wrong priorities in Parkinson’s disease. J Neurol Sci 2006;248(1–2):196–204. [27] Bello O, Sanchez JA, Fernandez-del-Olmo M. Treadmill walking in Parkinson’s disease patients: adaptation and generalization effect. Mov Disord 2008;23(9): 1243–9.
[28] Carpinella I, Crenna P, Rabuffetti M, Ferrarin M. Coordination between upperand lower-limb movements is different during overground and treadmill walking. Eur J Appl Physiol 2010;108(1):71–82. [29] Krasovsky T, Lamontagne A, Feldman AG, Levin MF. Effects of walking speed on gait stability and interlimb coordination in younger and older adults. Gait Posture 2014;39(1):378–85. [30] Maetzler W, Hausdorff JM. Motor signs in the prodromal phase of Parkinson’s disease. Mov Disord 2012;27(5):627–33.
