Velocity discrimination: Reliability and construct validity in older adults

Velocity discrimination: Reliability and construct validity in older adults

Human Movement Science 26 (2007) 443–456 Velocity discrimination: Reliability and construct validity in older adults Ke...

363KB Sizes 0 Downloads 10 Views

Human Movement Science 26 (2007) 443–456

Velocity discrimination: Reliability and construct validity in older adults Kelly P. Westlake ¤, Yuhsiao Wu, Elsie G. Culham School of Rehabilitation Therapy, Queen’s University, Kingston, ON, Canada K7L 3N6 Available online 7 February 2007

Abstract The aim of this study was to determine whether a test of velocity discrimination is a reliable and valid measure of proprioception in healthy older adults. Results revealed excellent test–retest reliability over a 2-week period. Velocity discrimination also indicated good construct validity with modest correlations with center of pressure sway outcomes in eyes open and closed conditions as well as stair climbing time. Good construct validity was identiWed by velocity discrimination sensitivity to age with a higher mean value for the older participants than for the younger participants. These Wndings suggest velocity discrimination is a valid and reliable measure of velocity sense, which may be included with measures of position and movement sense to enhance the proprioceptive testing repertoire among researchers. Implications of these results are discussed in terms of evaluation of proprioceptive training programs aimed to enhance postural control. © 2006 Elsevier B.V. All rights reserved. PsycINFO classiWcation: 2221; 2320 Keywords: Proprioception; Kinesthesis; Validity; Elderly; Postural control

1. Introduction Proprioception is deWned as the aVerent information arising from receptors within the muscles, joints, and skin and provides orientation information about movement and position of the joints, velocity of muscular contractions, and the force, eVort, and heaviness * Corresponding author. Address: Rehabilitation Research and Development Center, VA Palo Alto HCC, 3801 Miranda Avenue, Palo Alto, CA 94304, United States. E-mail address: [email protected] (K.P. Westlake).

0167-9457/$ - see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.humov.2006.12.002


K.P. Westlake et al. / Human Movement Science 26 (2007) 443–456

associated with muscular contractions (Gandevia, McCloskey, & Burke, 1992; Lephart, Riemann, & Fu, 2000). This sensation, in addition to tactile signals, is crucial to the maintenance of postural stability (Lord, Clark, & Webster, 1991; Maurer, Mergner, & Peterka, 2006). Traditionally, proprioception has been measured using tests of threshold to perception of passive movement (TPM) and joint position sense (JPS). TPM represents either the degree of passive joint displacement prior to the detection of movement and direction or the angle at which 70% of displacements are detected. JPS is identiWed as the error between a predetermined test position of the joint and the active or passive replicated position of the same or contralateral joint. Often, results of only one proprioceptive test are reported (Dracoglu, Aydin, Baskent, & Celik, 2005; Fong & Ng, 2006; Tiedemann, Sherrington, & Lord, 2005). However, a distinct diVerence in the neural mechanisms controlling position and movement sense (Brown, Rosenbaum, & Sainburg, 2003) suggests the need to conduct more than one independent assessment to gain a true perspective of proprioceptive function. The lack of correlation between measures of these submodalities (De Jong, Kilbreath, Refshauge, & Adams, 2005; Grob, Kuster, Higgins, Lloyd, & Yata, 2002) and between measures of position and velocity sense (Djupsjobacka & Domkin, 2005) further supports this notion. Nevertheless, unlike the sense of position and movement, the sense of velocity is one component of proprioception that has received relatively little attention. Given that velocity information is essential and more accurate than position and acceleration information for the small postural corrections required during quiet stance (Jeka, Kiemel, Creath, Horak, & Peterka, 2004; Masani, Popovic, Nakazawa, Kouzaki, & Nozaki, 2003) and the association between postural sway and fall risk (Maki, Holliday, & Topper, 1994; Melzer, Benjuya, & Kaplanski, 2004), particular attention should be paid to the eVects of aging on the use of velocity information at the ankle. In the few studies that have assessed velocity sense, testing was completed by replicating a target velocity, discriminating between the faster of two movement velocities, or by replicating position during movement at varying velocities (dynamic position sense) (Deshpande, Connelly, Culham, & Costigan, 2003; Lonn, Djupsjobacka, & Johansson, 2001; Verschueren, Brumagne, Swinnen, & Cordo, 2002). Only two studies were found to have examined age eVects. Using velocity replication and dynamic position sense, both investigations failed to identify the age-related changes known to occur using tests of TPM and JPS (Deshpande et al., 2003; Verschueren et al., 2002). These results may have been due to methodological issues relating to the complexity of task requirements. Deshpande et al. (2003) assessed active replication of velocity while participants met the postural demands of standing throughout the task. Participants in the study by Verschueren et al. (2002) focused on quickly opening the hand when the ankle rotated through a target angle at varying velocities. The sensory, motor, and cognitive requirements of these tasks may have masked the presence of independent age changes in velocity sense. Thus, a measurement of threshold to velocity discrimination at the ankle was developed, which aimed to increase test validity by simplifying task performance. Pilot work using this measure demonstrated a signiWcant relationship with activity level in older adults, unlike measures of TPM and JPS (Westlake & Culham, 2005). However, before the eVects of activity level and exercise on velocity sense with respect to postural control may be further explored, an evaluation of the psychometric properties of velocity discrimination was necessary. The aim of this study was to determine whether velocity discrimination is a reliable and valid measure of proprioception.

K.P. Westlake et al. / Human Movement Science 26 (2007) 443–456


It was hypothesized that velocity discrimination would (i) demonstrate excellent test– retest reliability in older adults; (ii) correlate with measures of postural sway and functional tasks (construct validity); and (iii) be sensitive to age diVerences (construct validity). 2. Methods 2.1. Participants Forty-six healthy older volunteers were recruited through local newspaper advertisements and Xyers and 24 healthy young adults were recruited by an e-mail sent to the students in the academic department of the investigators. Eight of the older participants were randomly selected to undergo velocity discrimination test–retest reliability assessments within a 2-week period. Exclusion criteria included the following: (1) major lower extremity joint or muscle pathology (e.g., chronic ankle instability or severe osteoarthritis); (2) neurological disorders and/or balance diYculties such as prior stroke or vertigo; and (3) use of an assistive device for mobility. A brief clinical examination conducted by the principle investigator tested for symptoms of peripheral neuropathy including the presence, diminution, or absence of Achilles tendon reXex, and position sense of the big toe (Richardson, 2002), and light touch sensation to the dorsal and plantar aspect of the foot. Participants demonstrating the absence or diminution on two or three of these tests were excluded from participation. The study was approved by the University and AYliated Teaching Hospitals Health Sciences Human Research Ethics Board and informed consent was obtained from all participants prior to participation. Table 1 presents summary statistics for age, gender, height, and weight of participants in the younger and older groups. Group equivalency was established for weight (p > .05), while age, gender distribution, and height diVerences were identiWed (p < .05). 2.2. Procedures Proprioceptive testing procedures were conducted using a torque motor (Compumotor, model 605) with an attached footplate and potentiometer measuring the angular displacement of the footplate (Fig. 1). Participants sat on a plinth with the right knee slightly Xexed Table 1 Participant characteristics

Age (yrs) Male:Female Weight (kg) Height (cm) TVD (°/s) TPM (°) JPS (°)

Older adults (n D 46)

Younger adults (n D 24)


76.28 § 4.75 27:19 73.55 § 12.66 165.60 § 9.60 1.69 § 1.24 1.64 § 0.80 2.99 § 1.27

26.17 § 2.79 6:18 67.92 § 11.13 170.23 § 6.97 0.46 § 0.18 0.82 § 0.43 3.27 § 1.16

<.001¤ .029 .07 .04¤ <.001¤ <.001¤ .39

Note: values are mean § standard deviation. Abbreviations: TVD D threshold to velocity discrimination; TPM D threshold to perception of passive movement; JPS D joint position sense. ¤ SigniWcant diVerence between groups at p < .05 (independent t-tests). 9 SigniWcant diVerence between groups at p < .05 (Mann–Whitney U-test).


K.P. Westlake et al. / Human Movement Science 26 (2007) 443–456

Fig. 1. Experimental set-up for proprioceptive tests. Participants were seated on a plinth with the right foot resting in a foot plate attached to a torque motor. A potentiometer located on the left of the footplate recorded angular movement.

to 5° and the ankle in a neutral position. The right foot rested in the footplate, with the dorsal and lateral aspects of the foot free from contact to avoid cutaneous cues. This position was selected for comfort since previous work in younger and older populations revealed no diVerences in ankle joint threshold to perception of passive movement (TPM) or passive joint position sense (JPS) tested in the weight-bearing compared with the nonweight bearing position used in the present study (Refshauge & Fitzpatrick, 1995; Westlake & Culham, 2006). Participants were blindfolded and wore headphones with white noise to reduce visual and auditory cues. All tests were randomized and one practice trial was allowed for each measure. TPM was determined by passively rotating the ankle joint at a velocity of 0.25°/s (Deshpande et al., 2003). Participants pressed a button to stop joint rotation at the moment movement was perceived and the direction of displacement could be stated with accuracy. The test was performed 6 times with a random order of dorsiXexion (df) and plantarXexion (pf) directions. The ankle joint was returned to a neutral starting position after each trial. Occasional sham trials and constant reminders of the instructions aimed to reduce the incidence of guessed responses. Trials in which participants verbally indicated that they were not absolutely positive of the movement direction once the stop button was pressed were discarded. A mean was taken of the remaining trials. Absolute diVerences between the start and stop positions (in °) were calculated as the TPM. This measure demonstrated excellent test–retest reliability (ICC D .95) at the ankle within a 2-week period (Deshpande et al., 2003). Passive JPS was tested by passively plantarXexing the ankle from a neutral position at a velocity of 5°/s (Deshpande et al., 2003; Gross, 1987), to one of the following three test positions: 10°pf, 12°pf, and 15°pf (Westlake & Culham, 2005). Each position was tested twice in a random order for a total of six trials. Relatively plantarXexed positions were selected to avoid excess cutaneous input from the plantar aspect of the foot as the footplate moved upward. Following a 5 s hold, the ankle was passively moved to a new start position of 5°df and then passively moved towards the test angle once again. Participants pressed a

K.P. Westlake et al. / Human Movement Science 26 (2007) 443–456


stop switch once they perceived the test angle was reproduced. Absolute diVerences in degrees between the test and reproduced position were calculated and the mean of 6 trials was determined. Test–retest reliability of this measure at the subtalar joint yielded a Pearson correlation coeYcient of .99 (Gross, 1987). Tests for velocity discrimination of ankle movement followed a forced choice methodology, whereby participants were required to choose the faster of two passive movements. Displacement was from 0 to 20°pf for the Wrst velocity and then 20°pf to 5°df for the second velocity (Fig. 2). The next trial began at 5°df and moved to 20°pf before returning to 0°. Subsequent velocity pairs alternated between the 20° and 25° ranges of displacement by beginning at the end angle (0° or 5°df) from the previous trial and always returning from 20°pf. The variable range of movement was used to reduce memory eVects. The variable range was deemed acceptable since previous work demonstrated that velocity discrimination was not dependent on displacement cues at velocities lower than 45°/s (Kerr & Worringham, 2002; Lonn et al., 2001). The order of the reference and test velocity was randomized for each presented pair. Since a response bias in which participants tended to overestimate lower velocities and underestimate higher velocities was evident in a velocity discrimination test of the upper extremity (Lonn et al., 2001), the reference velocity of the current study was held constant at 5°/s. The initial test velocity was 10°/s and was reduced by 1°/s until an incorrect response was reported. Test velocity was then increased in increments of 0.25°/s. The threshold value was the smallest diVerence between the reference and test velocities in which a correct response was obtained and conWrmed over three trials. Reasons for selecting the reference and test velocity values used for tests of velocity discrimination in this study were threefold. First, these values were well above the threshold of 2°/min (0.03°/s), below which point the detection of movement is abolished (Clark, Burgess, Cahpin, & Lipscomb, 1985). Second, since the velocity-dependent nature of movement detection saturates at speeds above 5°/s in both younger and older populations

Fig. 2. Schematic representation of velocity discrimination test depicting angular displacements. The Wrst test begins at 0°, moves to 20°pf at one velocity, and then to 5°df at another velocity. Each subsequent test begins at the end angle of the previous test. 䊏 First test; 䊐 subsequent test.


K.P. Westlake et al. / Human Movement Science 26 (2007) 443–456

(Gilsing et al., 1995; Thelen, Brockmiller, Ashton-Miller, Schultz, & Alexander, 1998), it could be assumed that age-related diVerences in TPM would not interfere with velocity discrimination results. Third, the chosen velocities were deemed low enough to overcome agerelated diVerences in central processing time of proprioceptive signals (Chaput & Proteau, 1996), yet high enough to reduce time constraints and limit fatigue while moving through the 20–25° displacements. Center of pressure (COP) mean velocity of total path length and standard deviation of COP amplitude in the mediolateral and anterior–posterior direction were measured while participants stood quietly on a force platform (ATMI, Model OR6-7). Participants were barefoot with heels positioned according to individual height and forefeet were splayed to a comfortable stance position. Tests were performed in a random order for 30 s with eyes open and closed with three trials for each condition. During the trials with eyes open, participants were encouraged to stare at a visual target placed at eye level a distance of 3 m in front of the force platform. A rest period of 1 min was allowed between trials. Gait speed and a timed stair test were also performed in a random order. Gait speed was measured over the middle 10 m of a 14 m walkway to allow for acceleration and deceleration. Participants walked at a “quick, but safe, speed”. Timing began once the Wrst foot crossed the 10 m-start line and ended once the Wrst foot crossed the Wnish line. One practice trial was allowed and an average gait speed of three test trials was recorded. The timed stair test required participants to ascend and descend 13 steps of standard tread depth and rise height using one rail at a “quick, but safe speed”. A separate time was recorded for ascent and descent and a 1-min rest period was allowed at the top and bottom landing. An average of two test trials was recorded. 2.3. Data analysis Baseline variables were compared between younger (n D 24) and older (n D 46) groups. Independent t-tests were used for continuous data (age, height, weight, velocity discrimination, TPM, JPS), and the Mann–Whitney U-test was used for nominal data (gender). 2.3.1. Test–retest reliability Test–retest reliability, measuring the ability of velocity discrimination to produce consistent scores over a 2-week period, was assessed using an intraclass correlation coeYcient (ICC2,1) with the 95% conWdence intervals calculated according to procedures described by Shrout and Fleiss (Shrout PE, 1979). ICC values greater than .75 were considered to demonstrate excellent reliability, while ICC values less than .40 were considered to have poor reliability. The selection of the 2-week time interval between tests was believed to be short enough so that neither true change nor the confounding eVects of practice would have occurred. ICC gives a relative index of the ratio of subject variability to subject variability plus the random error variability. The ICC2,1 equation also takes the variance between test–retest scores into account and was applicable because each participant was measured on each occasion (Rankin & Stokes, 1998). Bland Altman plots (Bland & Altman, 1986) were used to visualize systematic variations in the mean diVerence between repeated measures and to identify any data points that were outliers. The diVerence between the test and retest trials for each participant was plotted against the mean of these two measurements as a representation of the relationship between measurement error and the best estimate of a true velocity discrimination value.

K.P. Westlake et al. / Human Movement Science 26 (2007) 443–456


Ninety-Wve percent limits of agreement (i.e., 1.96 £ standard deviation (SD), above and below the mean of the two trials) were then calculated and plotted. The preciseness of these limits was estimated using calculations of the standard error (square root of 3SD2/n) and 95% conWdence intervals (CIs) (§1.96 £ standard error). 2.4. Construct validity Without a gold standard to measure velocity sense, construct validity was determined by testing the relationship between velocity discrimination and the two other proprioceptive measures, TPM and JPS, using Pearson correlation coeYcients (Finch, Brooks, Stratford, & Mayo, 2002). Given the importance of proprioception to postural control, velocity discrimination was also assessed for relationships with measures of COP outcomes (COP velocity, standard deviation of mediolateral and anterior–posterior COP amplitude), in addition to functional measures of stair climbing time, and gait speed. Correlation coeYcients were interpreted in terms of the proportion of variance associated with changes in velocity discrimination scores using the coeYcient of determination (r2 £ 100) (Evans, 1998). Construct validity was also analyzed using an independent t-test with group (older vs. younger) as the independent factor. Outliers were identiWed as having an individual mean of 3SDs outside of the group mean with a signiWcant eVect on the results (Stevens, 1999). Statistical signiWcance was established at p < .05 and all statistical procedures were conducted using SPSS, version 11.5 (SPSS Inc.). 3. Results 3.1. Test–retest reliability Table 2 presents the results for test–retest velocity discrimination reliability using the intraclass correlation coeYcient (ICC). ICC2,1 was excellent (.86) indicating that older adults had highly consistent performances. Fig. 3 illustrates the Bland Altman plot of the diVerences between the two test–retest trials against the mean of these two trials for each participant. The mean velocity discrimination diVerence was 0.03°/s with SD of 0.42°/s, giving 95% limits of agreements of ¡0.83 to 0.86°/s. All participants were within these limits, indicating good repeatability of the velocity discrimination measure. Standard error and conWdence intervals (CIs) were calculated for the limits of agreement as an estimate of sampling error (Altman & Bland, 2003). Standard error, the square root (sqrt) of 3SD2/n, was (sqrt)3(0.42)2/8 D 0.007. Therefore, the 95% CI for the limits of agreement (§1.96 £ standard error) was .013. For the lower limit,

Table 2 TVD test–retest reliability (n D 8) Test week 1


Retest week 3


Mean § SD


Mean § SD



95% CI

1.44 § 0.80


1.41 § 0.92




Abbreviations: TVD D threshold to velocity discrimination; ICC D intraclass correlation coeYcient; CI D conWdence interval; SD D standard deviation.


K.P. Westlake et al. / Human Movement Science 26 (2007) 443–456

Difference TVD test-retest (degrees/s)


M e an+ 1.96 SD

1.0 0.5

M e an=0.031 0.0 -0.5

M e an- 1.96 SD

-1.0 -1.5 0.0









Mean TVD test retest values (degrees/s) Fig. 3. Bland Altman plot of diVerence against mean for TVD test–retest TVD values, with mean diVerence and 95% limits of agreement indicated.

CI was ¡.84 to ¡.82 and for the upper limit, CI was .85–.87. The very small CIs reXect little variation between test and retest scores. 3.2. Construct validity Construct validity of velocity discrimination was determined using the relationships of velocity discrimination in the older group (n D 46) with other proprioceptive measures, COP sway outcomes, and functional tasks. Two outliers were identiWed as having a TPM value and stair descent time 3SD outside of the group mean, demonstrating a signiWcant eVect on the results. These data points were removed from subsequent analysis. Table 3 delineates the associations, expressed as Pearson correlation coeYcients, between proprioceptive measures and COP sway measures with eyes open and closed. Relationships between proprioceptive and functional measures are found in Table 4. SigniWcant correlations, although modest, revealed that participants with low velocity discrimination values demonstrated lower COP sway outcomes with eyes open and closed as well as shorter time to ascend and descend stairs. Stated diVerently, 10.9% of the COP velocity and 16.0% of the standard deviation of mediolateral COP amplitude variance with Table 3 Correlations between TVD, TPM, and P-JPS, and COP outcomes (n D 46) Eyes open Velocity r TVD TPM JPS

SD-x p


.33 .13 .16

Eyes closed

.03 .39 .29


SD-y p


.40 .15 .24

.01 .33 .10

r .29 .33¤ ¡.04

Velocity p .05 .03 .81

r .32 .12 .12

SD-x p


.03 .42 .45


SD-y p


.30 .18 ¡.03

.05 .23 .83


p ¤

.35 .41¤ .08

.03 .01 .58

Abbreviations: TVD D threshold to velocity discrimination; TPM D threshold to perception of passive movement; JPS D joint position sense; SD-x D standard deviation of COP amplitude in the mediolateral direction; SDy D standard deviation of COP amplitude in the anterior–posterior direction; r D Pearson correlation coeYcient. ¤ Correlation is signiWcant at p < .05.

K.P. Westlake et al. / Human Movement Science 26 (2007) 443–456


Table 4 Correlations between TVD, TPM, P-JPS, and functional tasks Proprioception

Timed stair climb

TPM (n D 46)

P-JPS (n D 46)

Ascent (n D 46)












.27 .28

.07 .07

.40¤ .45¤ .42¤

.01 .002 .004

.31¤ .24 .26

.04 .12 .09

¡.16 ¡.20 ¡.11

.28 .20 .49

TVD (n D 46) .29 TPM (n D 45) JPS (n D 46)

Descent (n D 45)

Gait speed (n D 46)

Abbreviations: TVD D threshold to velocity discrimination; TPM D threshold to perception of passive movement; JPS D joint position sense; r D Pearson correlation coeYcient. ¤ Correlation is signiWcant at p < .05.

eyes open, 10.2% of COP velocity, 9% of standard deviation of mediolateral COP amplitude, and 12.3% of standard deviation of anterior–posterior COP amplitude variance with eyes closed, and 16.0% stair ascent and 9.6% descent time variance may be explained by velocity discrimination scores (r2 £ 100)(Evans, 1998). Positive associations between velocity discrimination and the other two proprioceptive measures, TPM and JPS, revealed only a trend towards signiWcance, as did the relationship between TPM and JPS. Correlations between both TPM and JPS and the ascending timed stair task were signiWcant. As TPM and JPS increased, so did the time taken to ascend stairs. In other words, velocity discrimination, TPM, and JPS each contribute approximately 16–20% to the variance of stair ascent time. In contrast, other than associations between TPM and standard deviation of anterior–posterior COP amplitude with eyes open and closed, no signiWcant correlations were identiWed between TPM and JPS and other COP outcomes. Neither velocity discrimination nor the other proprioceptive measures demonstrated a signiWcant relationship with gait speed. Velocity discrimination was found to be sensitive to age related changes, t D ¡4.8, p < .001, with a higher mean value for the older participants than for the younger participants (Table 1), indicating good construct validity. 3.3. Participants To assess whether gender diVerences aVected velocity discrimination values, an independent t-test was applied to male (mean velocity discrimination value § SD, 1.48 § 1.32°) and female (1.07 § 1.0°) data of the combined older and younger participants (n D 70). The results (t D 1.46, p D .15) suggest that gender did not inXuence velocity discrimination outcomes. A Pearson correlation analysis was also applied to the combined group of younger and older participants to determine whether relationships existed between height and velocity discrimination. Results indicated no relationship (r D ¡.12, p D .34), suggesting that height diVerences between groups were not likely to confound velocity discrimination results. Mean § SD for tests of velocity discrimination, threshold to perception of passive movements (TPM), and joint position sense (JPS) for each cohort are also reported in Table 1. Age-related changes in TPM are consistent with pervious Wndings (Deshpande et al., 2003; Westlake & Culham, 2006), while the lack of change in JPS is generally not supported (Westlake & Culham, 2006; You, 2005). Nevertheless, performance was similar to that found previously for healthy older participants in terms of TPM and JPS (Westlake & Culham, 2006), COP outcomes under eyes open (velocity 1.10 § 0.37 cm/s; standard


K.P. Westlake et al. / Human Movement Science 26 (2007) 443–456

deviation of COP amplitude in the mediolateral direction 0.19 § 0.07 cm; and anterior–posterior direction 0.44 § 0.13 cm) and eyes closed (COP velocity 1.63 § 0.71 cm/s; standard deviation of COP amplitude in the mediolateral direction 0.22 § 0.09 cm; and anterior–posterior direction 0.51 § 0.15 cm) conditions (Benjuya, Melzer, & Kaplanski, 2004; Laughton et al., 2003), stair ascent (3.95 § 1.14 s) and descent (3.30 § 0.88 s) time (Vincent et al., 2002), and fast gait speed (1.94 § 0.31 m/s) (Lopopolo, Greco, Sullivan, Craik, & Mangione, 2006). 4. Discussion To examine the test–retest reliability of velocity discrimination or its performance consistency across time, eight participants were asked to return to the laboratory for a second measurement 2 weeks following the initial testing. The very high intraclass correlation coeYcient (ICC) value of .86 and the very small Bland Altman 95% limits of agreement (¡.83 to .86°/s), suggest that velocity discrimination is quite stable across time. Since the present study is the Wrst attempt to measure velocity discrimination at the ankle in older adults, no comparative data are available. One study measuring active replication of velocity at the ankle in a standing position demonstrated a slightly lower ICC value of .79 (Deshpande et al., 2003). Other ankle proprioceptive measures of threshold to perception of passive movement (TPM) and an active measure of joint position sense (JPS) revealed similar high ICC values of .95 and .83, respectively (Deshpande et al., 2003). With neither an accepted criterion standard nor published comparison data for velocity discrimination, construct validity of this test was Wrst assessed using other proprioceptive measures and functional outcomes. The lack of relationship between velocity discrimination, TPM, and JPS was not surprising given the previously demonstrated lack of correlation between tests of position and movement sense (De Jong et al., 2005; Grob et al., 2002) and between tests of position and velocity discrimination (Djupsjobacka & Domkin, 2005). However, similar relationships between all three proprioceptive measures (r D .27–.29, p D .07) and similar relationships between each measure and stair ascent time (r D .40–.45, p D .004–.01), suggests an equal contribution to the test variance of each of these measures. The trend towards a signiWcant correlation between proprioceptive measures implies that a greater number of participants may be required to identify a signiWcant relationship. Where velocity discrimination diVers from TPM and JPS is in the relationship with stair descent time, which was also found to be modest, although signiWcant. One possibility for this Wnding is that velocity sense may be more sensitive and accurate than position or movement cues arising from the eccentrically contracting muscles that control stair descent. Reduced JPS following eccentric compared to concentric exercise provides some evidence in support of this theory (Brockett, Warren, Gregory, Morgan, & Proske, 1997). The lack of relationship between velocity discrimination and gait speed may be explained by the fact that a fast walking task does not involve the postural challenges and awareness of body orientation cues to the same extent as the stair-climbing task. Tiedemann et al. (2005) reported that rather than one independent factor, gait speed is best predicted using a range of physiological and psychological factors such as lower limb strength, proprioception, balance, reaction time, vision, pain and emotional well-being. Construct validity was also assessed using COP stability measures. SigniWcant, but modest, relationships were found between velocity discrimination and nearly all COP outcomes in both eyes open and eyes closed conditions. In contrast, only standard deviation of anterior–posterior COP amplitude in the eyes open and closed conditions demonstrated rela-

K.P. Westlake et al. / Human Movement Science 26 (2007) 443–456


tionships with TPM and no outcomes were found to correlate with JPS. The distinct diVerence between velocity discrimination and the other two proprioceptive measures in terms of the relationship to COP outcomes, particularly COP velocity, is important for two reasons. First, COP sway velocity is known to be the most consistent stability parameter demonstrating age-related changes, whereas standard deviation of mediolateral and anterior–posterior COP amplitude are less discriminatory (Prieto, Myklebust, HoVmann, Lovett, & Myklebust, 1996; Raymakers, Samson, & Verhaar, 2005). Second, center of mass (COM) velocity information is critical for anticipating changes in COM position and determining which compensatory measures need to be taken to regain control of quiet standing (Jeka et al., 2004; Masani et al., 2003). Assuming the body moves as an inverted pendulum, COP is used as an estimate of the angular deviations of COM during quiet standing (Winter, 1995), thereby making the previous point especially relevant. Thus, the modest correlations between velocity discrimination and COP velocity in addition to standard deviation of mediolateral and anterior–posterior COP amplitude may suggest a clinical advantage of using velocity discrimination over measures of TPM and JPS when evaluating the inXuence of proprioceptive declines on postural control. Future longitudinal studies are needed to assess whether declines in velocity discrimination may predict reduced stability and impaired functional balance in older adults. Construct validity of velocity discrimination was also assessed using diVerences between young and older participants. Age-related changes in proprioception are fairly well established in the literature (Bullock-Saxton, Wong, & Hogan, 2001; Thelen et al., 1998; Westlake & Culham, 2006) and were reXected in older and younger group diVerences in velocity discrimination in the current study. These results are in contrast to previous studies that have attempted to identify age eVects in velocity sense (Deshpande et al., 2003; Verschueren et al., 2002). Apart from distinctions in task complexity, these discrepancies are best explained by methodological diVerences. Verschueren et al. (2002) investigated the eVects of diVerent ankle movement velocities on dynamic position sense, representing a departure from the experimental procedures used in the current study. The test velocities of 15, 20, 25, 30°/s were considerably higher than the speed of testing (5–10°/s) used in the current study. Perhaps age-related changes become evident only at lower test velocities. Indeed, Kokmen, Bossemeyr, and Williams (1978) considered the eVect of velocity on movement detection thresholds and discovered signiWcant diVerences between younger and older groups at low velocities, but not at high velocities. Such low movement velocities are considered to be comparable to the low COP velocities evident during quiet standing (Fernie, Gryfe, Holliday, & Llewellyn, 1982), thereby supporting the relationship between the COP outcomes and velocity discrimination found in the current study. In the case of Deshpande et al. (2003), participants were asked to actively replicate an active self-selected velocity at the ankle with movement occurring through a constant range of motion. Lower velocity replication errors have been reported during active as opposed to passive velocity replication (Lonn et al., 2001), suggesting that older participants may have taken advantage of the additional corollary discharge associated with eVerent activity. In addition, velocity replication accuracy is known to improve at lower compared to higher velocities (Lonn et al., 2001). Thus, the range of self-selected reference velocities may have induced enough within group variability to mask diVerences between older and younger participants. Moreover, several inherent problems have been identiWed with the velocity replication methodology, such as diVerences between peak and average velocities (Lonn et al., 2001). For these reasons, tests of replication were not deemed as


K.P. Westlake et al. / Human Movement Science 26 (2007) 443–456

valid as tests of discrimination that were not inXuenced by an active contraction or cues provided by the extent of movement or time. Even with evidence of age-related changes in velocity discrimination, the clinical relevance of low velocity discrimination values in older adults is questionable. The majority of falls occur when the base of support is perturbed while walking, such as during trips and slips (Maki & McIlroy, 1996), requiring lower extremity joint velocities upwards of 106°/s to regain postural stability (Madigan & Lloyd, 2005). While it may be true that, in isolation, a small reduction in velocity sense may not lead to instability or increased fall risk, the integration of this age-related change with muscle weakness and decreased torque generation in older adults (for review see Simpson, 2004) may be important. Calvin-Figuiere, Romaiguere, and Roll (2000) eVectively demonstrated the close relationship between the integration of proprioception information and the activation of motor commands in the antagonistic muscle. It follows that the reduced peak velocities at the hip, knee, and ankle joints during fall recovery in older adults (Madigan & Lloyd, 2005; Thelen et al., 2000) may, at least partially, be due to a reduction in velocity cues leading to slowed muscle activation patterns. Therefore, future research exploring the possibility of enhancing the sense of velocity along with other proprioceptive submodalities through exercise in older adults may be warranted. In conclusion, evidence from this study indicates that this new method of assessing velocity sense at the ankle in healthy older adults is both valid and reliable. Results support the addition of velocity discrimination along with tests of position and movement sense to enhance the proprioceptive testing repertoire by representing several submodalities. Preliminary evidence of the relationships of velocity discrimination with measures of COP sway and stair climbing time may prompt future research investigating the central integration of velocity sense during postural control and whether age-related changes may be reversed with exercise. As a Wnal note, since the proprioceptive testing for this study was completed using equipment that is typically inaccessible to clinicians, the described tests are presently targeted towards researchers in the area of proprioception. Future research is recommended to establish comparable tests for clinical use. Acknowledgements This research was supported by a Physical Therapy Foundation of Canada Research Grant Kelly Westlake is supported by a Canadian Institutes of Health Research (CIHR), Institute of Aging Fellowship. References Altman, D. G., & Bland, J. M. (2003). Interaction revisited: The diVerence between two estimates. British Medical Journal, 326, 219. Benjuya, N., Melzer, I., & Kaplanski, J. (2004). Aging-induced shifts from a reliance on sensory input to muscle cocontraction during balanced standing. The Journals of Gerontology. Series A, Biological Sciences and Medical Sciences, 59, 166–171. Bland, J. M., & Altman, D. G. (1986). Statistical methods for assessing agreement between two methods of clinical measurement. Lancet, 1, 307–310. Brockett, C., Warren, N., Gregory, J. E., Morgan, D. L., & Proske, U. (1997). A comparison of the eVects of concentric versus eccentric exercise on force and position sense at the human elbow joint. Brain Research, 771, 251–258.

K.P. Westlake et al. / Human Movement Science 26 (2007) 443–456


Brown, L. E., Rosenbaum, D. A., & Sainburg, R. L. (2003). Limb position drift: Implications for control of posture and movement. Journal of Neurophysiology, 90, 3105–3118. Bullock-Saxton, J. E., Wong, W. J., & Hogan, N. (2001). The inXuence of age on weight-bearing joint reposition sense of the knee. Experimental Brain Research, 136, 400–406. Calvin-Figuiere, S., Romaiguere, P., & Roll, J. P. (2000). Relations between the directions of vibration-induced kinesthetic illusions and the pattern of activation of antagonist muscles. Brain Research, 881, 128–138. Chaput, S., & Proteau, L. (1996). ModiWcations with aging in the role played by vision and proprioception for movement control. Experimental Aging Research, 22, 1–21. Clark, F. J., Burgess, R. C., Cahpin, J. W., & Lipscomb, W. T. (1985). The role of intramuscular receptors in the awareness of limb position. Journal of Neurophysiology, 54, 1529–1540. De Jong, A., Kilbreath, S. L., Refshauge, K. M., & Adams, R. (2005). Performance in diVerent proprioceptive tests does not correlate in ankles with recurrent sprain. Archives of Physical Medicine and Rehabilitation, 86, 2101– 2105. Deshpande, N., Connelly, D. M., Culham, E. G., & Costigan, P. A. (2003). Reliability and validity of ankle proprioceptive measures. Archives of Physical Medicine and Rehabilitation, 84, 883–889. Djupsjobacka, M., & Domkin, D. (2005). Correlation analysis of proprioceptive acuity in ipsilateral positionmatching and velocity-discrimination. Somatosensory Motor Research, 22, 85–93. Dracoglu, D., Aydin, R., Baskent, A., & Celik, A. (2005). EVects of kinesthesia and balance exercises in knee osteoarthritis. Journal of Clinical Rheumatology, 11, 303–310. Evans, A. N. (1998). Using basic statistics in the social sciences (Third ed.). New Jersey: Prentice-Hall, Inc. Fernie, G. R., Gryfe, C. I., Holliday, P. J., & Llewellyn, A. (1982). The relationship of postural sway in standing to the incidence of falls in geriatric subjects. Age and Ageing, 11, 11–16. Finch, E., Brooks, D., Stratford, P. W., & Mayo, N. E. (2002). Physical rehabilitation outcome measures (Second ed.). Hamilton, ON: Canadian Physiotherapy Association. Fong, S. M., & Ng, G. Y. (2006). The eVects on sensorimotor performance and balance with Tai Chi training. Archives of Physical Medicine and Rehabilitation, 87, 82–87. Gandevia, S. C., McCloskey, D. I., & Burke, D. (1992). Kinaesthetic signals and muscle contraction. Trends in Neurosciences, 15, 62–65. Gilsing, M. G., Van den Bosch, C. G., Lee, S. G., Ashton-Miller, J. A., Alexander, N. B., Schultz, A. B., et al. (1995). Association of age with the threshold for detecting ankle inversion and eversion in upright stance. Age and Ageing, 24, 58–66. Grob, K. R., Kuster, M. S., Higgins, S. A., Lloyd, D. G., & Yata, H. (2002). Lack of correlation between diVerent measurements of proprioception in the knee. The Journal of Bone and Joint Surgery [Br], 614–618. Gross, M. T. (1987). EVects of recurrent lateral ankle sprains on active and passive judgements of joint position. Physical Therapy, 67, 1505–1509. Jeka, J., Kiemel, T., Creath, R., Horak, F., & Peterka, R. (2004). Controlling human upright posture: Velocity information is more accurate than position or acceleration. Journal of Neurophysiology, 92, 2368–2379. Kerr, G. K., & Worringham, C. J. (2002). Velocity perception and proprioception. Advances in Experimental Medicine and Biology, 508, 79–86. Kokmen, E., Bossemeyr, R. W., & Williams, W. J. (1978). Qualitative evaluation of joint motion sensation in an aging population. Journal of Gerontology, 33, 62–67. Laughton, C. A., Slavin, M., Katdare, K., Nolan, L., Bean, J. F., Kerrigan, D. C., et al. (2003). Aging, muscle activity, and balance control: Physiologic changes associated with balance impairment. Gait and Posture, 18, 101– 108. Lephart, S. M., Riemann, B. L. & Fu, F. H. (2000). Introduction to the sensorimotor system. In Lephart, S.M., Fu, F. H. (Eds.), Propriocpetion and neuromuscular control in joint stability (pp. xvii–xxiv). Windsor, ON: Human Kinetics. Lonn, J., Djupsjobacka, M., & Johansson, H. (2001). Replication and discrimination of limb movement velocity. Somatosensory Motor Research, 18, 76–82. Lopopolo, R. B., Greco, M., Sullivan, D., Craik, R. L., & Mangione, K. K. (2006). EVect of therapeutic exercise on gait speed in community-dwelling elderly people: A meta-analysis. Physical Therapy, 86, 520–540. Lord, S. R., Clark, R. D., & Webster, I. W. (1991). Postural stability and associated physiological factors in a population of aged persons. Journal of Gerontology, 46, M69–M76. Madigan, M. L., & Lloyd, E. M. (2005). Age and stepping limb performance diVerences during a single-step recovery from a forward fall. The Journals of Gerontology. Series A, Biological Sciences and Medical Sciences, 60, 481–485.


K.P. Westlake et al. / Human Movement Science 26 (2007) 443–456

Maki, B. E., Holliday, P. J., & Topper, A. K. (1994). A prospective study of postural balance and risk of falling in an ambulatory and independent elderly population. Journal of Gerontology, 49, M72–M84. Maki, B. E., & McIlroy, W. E. (1996). Postural control in the older adult. Clinics in Geriatric Medicine, 12, 635– 658. Masani, K., Popovic, M. R., Nakazawa, K., Kouzaki, M., & Nozaki, D. (2003). Importance of body sway velocity information in controlling ankle extensor activities during quiet stance. Journal of Neurophysiology, 90, 3774– 3782. Maurer, C., Mergner, T., & Peterka, R. J. (2006). Multisensory control of human upright stance. Experimental Brain Research, 171, 231–250. Melzer, I., Benjuya, N., & Kaplanski, J. (2004). Postural stability in the elderly: A comparison between fallers and non-fallers. Age and Ageing, 33, 602–607. Prieto, T. E., Myklebust, J. B., HoVmann, R. G., Lovett, E. G., & Myklebust, B. M. (1996). Measures of postural steadiness: DiVerences between healthy young and elderly adults. IEEE Transactions on Bio-Medical Engineering, 43, 956–966. Rankin, G., & Stokes, M. (1998). Reliability of assessment tools in rehabilitation: An illustration of appropriate statistical analyses. Clinical Rehabilitation, 12, 187–199. Raymakers, J. A., Samson, M. M., & Verhaar, H. J. (2005). The assessment of body sway and the choice of the stability parameter(s). Gait and Posture, 21, 48–58. Refshauge, K. M., & Fitzpatrick, R. (1995). Perception of movement at the human ankle: EVects of leg position. Journal of Physiology, 488, 243–248. Richardson, J. K. (2002). The clinical identiWcation of peripheral neuropathy among older persons. Archives of Physical Medicine and Rehabilitation, 83, 1553–1558. Shrout PE, F. J. (1979). Intraclass correlations: Uses in assessing rater reliability. Psychological Bulletin, 86, 420– 428. Simpson, J. M. (2004). Elderly people at risk of falling: The role of muscle weakness. Physiotherapy Canada, 79, 831–834. Stevens, J. (1999). Intermediate statistics: A modern approach (Second ed.). Mahwah, NJ: Lawrence Erlbaum Associates, Inc. Thelen, D. G., Brockmiller, C., Ashton-Miller, J. A., Schultz, A. B., & Alexander, N. B. (1998). Thresholds for sensing foot dorsi- and plantarXexion during upright stance: EVects of age and velocity. The Journals of Gerontology. Series A, Biological Sciences and Medical Sciences, 53, M33–M38. Thelen, D. G., Muriuki, M., James, J., Schultz, A. B., Ashton-Miller, J. A., & Alexander, N. B. (2000). Muscle activities used by young and old adults when stepping to regain balance during a forward fall. Journal of Electromyography and Kinesiology, 10, 93–101. Tiedemann, A., Sherrington, C., & Lord, S. R. (2005). Physiological and psychological predictors of walking speed in older community-dwelling people. Gerontology, 51, 390–395. Verschueren, S. M., Brumagne, S., Swinnen, S. P., & Cordo, L. (2002). The eVect of aging on dynamic position sense at the ankle. Behavioural Brain Research, 136, 593–603. Vincent, K. R., Braith, R. W., Feldman, R. A., Magyari, P. M., Cutler, R. B., Persin, S. A., et al. (2002). Resistance exercise and physical performance in adults aged 60 to 83. Journal of the American Geriatrics Society, 50, 1100–1107. Westlake, K. P., & Culham, E. G. (2005). Relationship between physical activity level, proprioception, and postural sway among older adults. In G. R. Jones, A. W. Taylor, & N. A. Ecclestone (Eds.), ScientiWc proceedings from the 6th World Congress on Aging and Physical Activity, August 3–7, 2004 (pp. 43–47). London, ON: Canadian Centre for Activity and Aging. Westlake, K. P., & Culham, E. G. (2006). InXuence of testing position and age on measures of ankle proprioception. Advances in Physiotherapy, 8, 41–48. Winter, D. A. (1995). Human balance and posture control during standing and walking. Gait and Posture, 3, 193– 214. You, S. H. (2005). Joint position sense in elderly fallers: A preliminary investigation of the validity and reliability of the SENSERite measure. Archives of Physical Medicine and Rehabilitation, 86, 346–352.