Journal of Gerontology: BIOLOGICAL SCIENCES 1991. Vol.46, No. 6, B238-244
Copyright 1991 by The Gerontological Society of America
Postural Sway Characteristics of the Elderly Under Normal and Altered Visual and Support Surface Conditions Normand Teasdale,1 George E. Stelmach,2 and Ann Breunig3 'Laval University, Quebec, Canada. Exercise and Sport Research Institute, Arizona State University-Tempe. 3 Motor Behavior Laboratory, University of Wisconsin-Madison.
2
One of the most pervasive findings in the literature on the aged is the general slowing of cognitive-motor responses with advancing age. Hence, an increased slowness in the processing of information from vestibular, visual, and somatosensory systems could contribute greatly to a decline in postural stability. To examine this question, in a crosssectional investigation, postural sway behavior of elderly (n = 18) and young (n = 10) adults was examined under conditions that stressed the slower integrative mechanisms rather than the reflexive mechanisms ofpostural control. The postural sway behavior ofyoung and elderly subjects was examined for a prolonged duration (80 s), under altered visual and/or support surface (5 cm thickfoam surface) conditions, and contrasted with normal stance. Results showed that the exclusion or disruption of one of the sensory inputs, alone, was not consistently sufficient to differentiate between elderly and young adults, because of compensation by the remaining sensory sources. Both alterations together (i.e., visual and surface), however, had a substantially greater effect upon the elderly than the young.
N
UMEROUS experimenters have demonstrated that postural sway increases with age (e.g., Era and Heikkinen, 1985; Murray et al., 1975; Sheldon, 1963). Very little is known, however, about how the different sensory systems responsible for control of posture are affected by the aging process. Both the monosynaptic (Appenzeller et al., 1966; Carel et al., 1979; Clarkson, 1978) and the long-latency reflexive systems (Woollacott et al., 1986) have been shown to exhibit small latency increases with age. In a 1989 study, Stelmach, Teasdale, DiFabio, and Phillips found that elders' sway and reflexive responses, during stance, to large-fast rotational ankle perturbations (7 deg at 20 deg/s) were similar to those of young subjects, suggesting that the elders' reflexive mechanisms were relatively intact. When smallslow rotational perturbations (2 deg at 4 deg/s) were given, however, the elders swayed more than the young subjects and could not adapt to repetitive presentations of perturbations. These results were interpreted as evidence that the aged may be prone to sway when posture is controlled by slower, higher-level sensory integrative mechanisms, inasmuch as the small-slow perturbation corrections took place mainly through the slower and higher level sensory systems (Stelmach etal., 1989b). A possible cause for the decreased stability observed with increasing age is the deterioration of primary peripheral sensory processes. Indeed, elderly individuals have been shown to have higher proprioceptive thresholds to passive movement than the young (Birren, 1947; Skinner et al., 1984; Whanger and Wang, 1974) and are less accurate at reproducing and matching joint angles (Kokmen et al., 1978; Stelmach and Sirica, 1986). In fact, from a clinical standpoint, a small and progressive degeneration of sensory input from the lower extremities is often the first, and most B238
common, manifestation of aging (Calne, 1985). Similarly, diminution of the axon population in the optic nerves (Johnson et al., 1987), loss of spatial sensitivity (Sekuler et al., 1980), reduction of hair cells in the semicircular canals (Rosenhall and Rubin, 1975), and reduction of the macula of the uticulus and sacculus (Johnsson and Hawkins, 1972) have all been reported with aging. Yet, little evidence has been found to support the suggestion that a decrease in peripheral sensory acuity is associated with increased postural sway (Brocklehurst et al., 1982; Era and Heikkinen, 1985). The lack of an obvious causal relationship between specific peripheral sensory impairments and postural sway implies that posture control is dependent upon complex, integrative processing from a variety of inputs. Stelmach and Worringham (1985) suggested that deterioration of the central integrative processes may further explain the decreased postural stability observed with aging. When reflexive mechanisms are not elicited, the central nervous system must undertake certain processing stages (i.e., peripheral input, response selection, and execution) to prevent a fall. Increased slowness in processing information from the vestibular, visual, and somatosensory systems could be a major source of disruption in postural stability (Allum and Pfaltz, 1985; Diener etal., 1984; Skinner etal., 1984). Indeed, one of the most convincing findings in the literature on aging is the general slowing of cognitive-motor responses (e.g., Salthouse, 1985). For this reason we examined postural control in a sample aging population under conditions that stressed the slower integrative mechanisms rather than the reflexive mechanisms of posture control. This experiment was designed to determine whether a reduction of visual, somatosensory, or both visual and somatosensory inputs were differentially disruptive to posture
POSTURAL SWAY
control across young and older adults. Sway behavior was examined for prolonged periods of upright posture (80 s) and quantified, using several dependent measures to maximize the ability to localize age changes. METHODS
Subjects. — Ten young subjects (5 males and 5 females, mean age 21.5 yrs, range 21—22) and 18 elderly subjects (9 males and 9 females, mean age 74 yrs, range 70-80) participated in the experiment. The elderly subjects were volunteers from local aging coalitions and support groups, and the young subjects were students from the University of Wisconsin-Madison. Elderly subjects, chosen from a large pool of subjects, had to meet the following criteria: no musculoskeletal defects; not taking any medication; not undergoing treatment for any neurological disease; no history of falling. This conservative selection procedure serves to minimize any age-related differences. All subjects gave informed consent for the procedures used. Apparatus. — A force platform with 4 load cells (AMTI model OR6-5-1) measured force and moment components along the x, y, and z axes. The signals from the force platform were amplified (AMTI SGA6-4) before being processed by a computer (PDP 11-73). All signals were sampled at 40 Hz and then filtered with a dual-pass Butterworth second-order filter (10 Hz cut-off) to remove any highfrequency artifacts. Sway behavior in the sagittal and frontal planes was then computed. Procedures. — The experiment comprised four different conditions: (a) control vision and normal surface, referred to as the normal vision-surface condition; (b) normal vision and altered surface, referred to as the altered surface condition; (c) normal surface and no vision, referred to as the alteredvision condition; and (d) no vision and altered surface, referred to as the altered vision-surface condition. Subjects stood barefoot on the force platform with feet together. First, the support surface was normal for the control (eyes open) and altered-vision (eyes closed) conditions. Then the support surface was altered by adding an open-cell, polyurethane foam surface (5 cm thick) upon which the subjects stood, for the altered-surface (eyes open) and altered vision-surface (eyes closed) conditions. The addition of a foam support surface alters the reliability of the somatosensory contributions to posture control and, therefore, is believed to force greater reliance on vision and/or vestibular information (Straube et al., 1988). A trial consisted of four sampling periods of 20 s each, spaced at 3-5 s intervals. Two consecutive trials (i.e., 80 s of total sampling/trial) were run for each condition. Data analysis. — Numerous sway measures and methods of displaying sway characteristics have been presented in the literature. No attempt has been made, however, to determine whether these measures reflect different characteristics of sway behavior. In the present experiment, sway was evalu-
B239
ated by using four different sway measures to determine whether they reflected differential characteristics of sway behavior and whether certain of these measures were agesensitive. Specifically used were range and standard deviations of sway behavior in both planes, average sway velocity, and sway density histograms. The range is thought to give an indication of how far center-of-foot pressure (COP) deviates from its original baseline, whereas standard deviation gives the variability of balance control over the sampled period. The range of sway behavior may be misleading because it does not provide information regarding the distribution of the COP within that range. Nevertheless, range was employed, as it is often reported in the literature and represents a standard of comparison with other published studies. Average sway velocity is the sum of the displacement vectors divided by total sampling time. In this experiment, the sampling period was constant for all conditions and subjects. Hence, the average sway velocity is by definition proportional to the summation of all the postural displacements of a given subject. Thus, for comparison purposes, the greater the value, the greater the displacement, and, therefore, the greater the average velocity. Finally, a measure of sway derived from density histograms of the sway path (Harris et al., 1982) was applied. Specifically, density histograms give a qualitative description of the distribution of sway behavior. Distributions were quantified by calculating an average COP and the percentage of time the subjects spent in eight arbitrarily defined concentric circles centered around the average COP in radial increments of 5 mm. Each of these areas was labeled from area 1 (closest to mean COP) to area 8 (farthest from mean COP). Thus, a higher percentage of time spent away from the average COP could indicate a postural behavior at risk. On average, the young and elderly spent more than 99% of the time within a radius of 40 mm of their mean center of foot pressure (COP). Data obtained for the sway range and standard deviation of sway range were submitted to a Group x Sway Plane X Vision x Surface x Sampling Period Analysis of Variance (ANOVA), with repeated measures on the last four factors. Data obtained for the sway velocity, after a square root transformation was performed, were submitted to a Group x Vision x Surface x Sampling Period ANOVA, with repeated measures on the last three factors. The raw data for sway velocity are presented in the text. Finally, to highlight group differences in each vision/surface experimental condition, the dispersion of sway data were submitted to Group x Area (8 radii of 5 mm) ANOVAs, with repeated measures on the last factor. RESULTS
In this experiment, the normal vision-surface condition represented a control condition, as none of the primary sensory inputs responsible for postural control (i.e., visual, vestibular, and proprioceptive) was altered. On the other hand, manipulation of a specific sensory input gave an estimation of the importance of that information for posture control in addition to indicating how the central nervous system adapted and reorganized information provided by the remain-
B240
TEASDALE ET AL.
ing sources of sensory information. A stability effect was derived from each of the different sensory input alterations. Sway range. — Overall, for group comparisons, the elderly had a greater sway range than did the young subjects (27.3 mm vs 21.8 mm; F (1,26) = 6.67, p < .01). Unlike the young subjects, who had a similar range of sway in both planes (21.6 mm and 22.0 mm), the elderly subjects exhibited greater sway along the medio-lateral plane than along the anterior-posterior plane (29.2 mm vs 25.4 mm; F (1,26) = 4.97, p < .05 for the Sway Plane x Group interaction. However, as the sway plane factor was significant only in this case, data obtained for both planes were pooled. Both groups were affected similarly by visual conditions (F (1,26) = 1.65, p > .05 for the Vision x Group interaction. As may be seen in Table 1, withdrawing vision produced an 11.2 mm increase in sway (cost) for the young (from 14.8 mm to 26.0 mm), and a 10.8 mm increase in sway (cost) for the elderly (from 17.4 mm to 28.2 mm); (F (1,26) = 97.08, p < .001 for the main effect of vision). Young and elderly showed different results for the alteredsurface condition; the sway range of the young subjects only increased by 2.1 mm, whereas it increased by 4.2 mm for the elderly (F (1,26) = 12.97, p < .005 for the Group x Surface interaction. However, the combined-perturbations condition (altered vision-surface) produced the most dramatic cost of all, especially in the elderly subjects (F (1,26) = 14.05,/? < .001 for the Surface x Vision x Group triple interaction. Indeed, with vision withdrawn, surface alteration created a 15.0 mm sway-range increase in the young, and a 24.7 mm sway-range increase in the elderly. No indication of any increased sway range over time appeared for either elderly or young subjects. In fact, for both groups, during the altered-vision condition there was a gradual decrease in sway range from the first 20 s of sampling to the last 20 s of sampling (7.0 mm for the elderly and 4.1 mm for the young subjects). No such decrease was observed for the altered-surface condition (F (3,78) = 2.67, p < .05). With vision, sway range remained relatively constant (a decrease of 0.9 mm for both groups; F (3,78) = 8.07, p < .001 for the Vision x Sampling Period interaction. Variability of the sway range. — A pattern of results similar to that of the sway range was obtained for variability. Overall, the elderly subjects exhibited more variability than did the young subjects (5.6 mm vs 4.6 mm; F( 1,26) = 5.08, p < .05). The young subjects exhibited similar variability in both planes (4.7 mm and 4.5 mm), whereas the elderly subjects exhibited greater sway along the medio-lateral plane than along the anterior-posterior plane (5.8 mm vs 5.3 mm; F ( 1,26) = 5.98, p< .05 for the Sway Plane x Group interaction. Again, as the sway plane factor yielded no other significant effect, the data obtained for both planes were pooled. Variability (Table 1) for both groups was similar under the altered-vision condition (2.0 mm increase, or cost, for the elderly vs 2.2 mm for the young; F (1,26) = 0.96, p > .05 for the Vision x Group interaction). Elderly subjects exhibited greater variability than young subjects under the alteredsurface condition (0.8 mm increased variability for the el-
Table 1. Sway Range, Variability, and Velocity for Visual, Surface, and Combined Visual-Surface Perturbations Condition Control Cost of altered vision Cost of altered surface Cost of altered vision and surface
Sway range (mm)
Variability (mm)
Velocity (mm/s)
Young
Young
Young
Elders
Elders
Elders
14.8
17.4
3.2
3.7
10.0
14.3
11.2
10.8"
2.2
2.0"
7.2
13.8*
0.4
0.8»"
2.1
3.2-
3.0
4.7'
10.9
25.2
b
24.7C
Note. Values in row Control represent sway range, variability, and velocity, exhibited by both young and elderly subjects, under the normal surface-vision condition. Values, or costs, in rows Altered vision, Altered surface, and Altered vision and surface represent the sway range, variability, and velocity in excess of that exhibited in the control condition. "Main Effect x Group two-way interaction. b Main Effect x Group two-way interaction. 'Vision x Surface x Group three-way interaction.
derly vs 0.4 mm for the young; F( 1,26) = 40.82,/? < .001). Just as the sway range for the elderly proved to be greatest under the altered vision-surface condition, so did its variability (F (1,26) = 11.01,/? < .001 for the Surface X Vision x Group triple interaction). Indeed, when vision was withdrawn and the surface altered, the variability of the elderly subjects increased by 4.7 mm, while it only increased by 3.0 mm for the young subjects. For both groups, there was a small, but significant, decrease in the variability of sway under the altered-vision condition over time. Between the first and last 20 s of sampling, there was a 1.1 mm decrease, without vision, and only a 0.1 mm decrease, with vision (F (3,78) = 7.53, p < .001 for the Sampling Period x Vision interaction. No such decrease was observed with altered surface, 0.6 mm with altered surface, and 0.4 mm with normal surface (F (3,78) = 2.19, /? < .05 for the Sampling Period X Surface interaction. Sway velocity. — Overall, the elderly had a greater average velocity than did the young subjects, and thus swayed more (24.9 mm/s vs 15.7 mm/s; F (1,26) = 9.49,/? < .005). As seen in Table 1, both groups had greater velocities when vision was withdrawn (increase of 7.2 mm/s for the young, and 13.8 mm/s for the elderly; (F (1,26) = 46.97, p < .001 for the main effect of vision). Despite a larger cost for the elderly than for the young subjects, the Vision x Group interaction did not reach the significance level (F (1,26) = 2.81, p > .05). Under the altered-surface condition, the young subjects showed a 2.1 mm/s increase, and the elderly, a 3.2 mm/s increase, in sway velocity (F (1,26) = 4.20,/? < .05 for the Surface x Group interaction. Again, the combination of both perturbations was most detrimental to the elderly's posture (F (1,26) = 4.23, p < .05 for the Surface x Vision X Group triple interaction. When both surface and vision were altered, the young
POSTURAL SWAY
subjects exhibited a 10.9 mm/s increase, whereas the elderly subjects exhibited a 25.2 mm/s increase, in sway velocity. Table 1 shows that the postural cost of both alterations was nearly additive for the young subjects, but multiplicative for the elders. As in the cases of sway range and variability, the sway velocity decreased over time, implying some occurrence of adaptation. For the young subjects, sway velocity decreased from 17.4 mm/s to 14.8 mm/s (from the first sampling period to the last sampling period), whereas for the elderly subjects, it decreased from 27.5 mm/s to 23.2 mm/s. A decomposition of the main effect of the sampling period into its orthogonal components showed that the linear and quadratic components were statistically significant (F (1,26) = 58.64 and 25.26, p < .001), suggesting that there was a more rapid decrease in velocity at the beginning, than at the end, of the sampling periods. Moreover, the decrease was more rapid for the young than for the elderly (100% of the decrease in velocity occurred within the first sampling period for the young, whereas only 49% of the decrease occurred within that time period for the elderly; F (1,26) = 4.88, p < .05 for the quadratic component of the Sampling Period x Group interaction).
B241
elderly subjects spent 4.1, 4.9, 1.3, 0.5, and 0.2% more time in areas 2, 3, 4, 5, and 6, respectively (F (7,182) = 3.43, p < .005 for the Group x Area interaction). For the altered-vision condition, both groups showed a similar dispersion of sway. Indeed, young subjects spent 2.1,0.6, and 0.4% more time in areas 2, 3, and 4, respectively, and
18r
2 15 -
(mm) Medio-Lateral
Dispersion of sway. — Representative density histograms of the sway path for an elderly and a young subject appear in Figures 1 and 2, respectively. Table 2 shows the percentage of time spent by elderly and young subjects in the eight areas centered around COP, under the different visual and surface experimental conditions. In Figure 3, group differences in the dispersion of sway were obtained by subtracting the percentages of time spent by young subjects in a given area from those of the elderly in that same area. Positive values indicate that young subjects spent more time in an area than elderly subjects, whereas negative values indicate that elderly subjects spent more time in an area than young subjects. As seen in Figures 1 and 2, elderly sway behavior spread over a wider area than that of the young subjects for the normal vision-surface, or control, condition. Young subjects spent 10.9% more time in area 1 than elderly subjects, while the elderly subjects spent 3.8, 5.0, 1.7, and 0.3% more time in areas 2, 3, 4, and 5, respectively, than the young subjects (F (7,182) = 3.64, p < .005 for the Group x Area interaction). The altered-surface condition yielded similar results; the young spent 11.2% more time in area 1, and the
RIGHT
Figure 1. Density histogram of the sway path of an elderly subject.
18r
-40 (mm) Medio-Lateral
Figure 2. Density histogram of the sway path of a young subject.
Table 2. Percentage of Time Spent by Young and Elderly Subjects in the Eight Areas Centered Around COP, Under Different Experimental Conditions Vision/normal surface
Vision/altered surface
No-vision/ normal surface
No-vision/ altered surface
Area
Young
Elders
Young
Elders
Young
Elders
Young
Elders
1 2 3 4 5 6 7 8
54.3 38.3
43.5 42.1 11.9
48.1 38.2 10.2
36.9 42.3 15.1
28.8 39.6 20.0
31.5 37.5 19.4
2.1 0.3 0.0 0.0 0.0
2.6 0.4 0.2 0.1 0.0
3.9 0.9 0.4 0.1 0.0
8.2 2.5 0.6 0.1 0.0
7.8 2.7 0.8 0.2 0.0
25.3 36.6 20.9 10.3
17.8 30.4 23.9 13.9
4.7 1.3 0.5 0.2
7.3 3.7 1.7 0.7
6.9 0.4 0.0 0.0 0.0 0.0
TEASDALE ET AL.
B242
DISCUSSION
not to suggest that vision is more important to balance control than proprioception; the degree to which proprioceptive information was altered in this experiment cannot be precisely documented). Dispersion of sway data showed that both groups spent a higher percentage of time in an "at risk'' posture-control area, centered around COP, for the alteredsurface condition than for the altered-vision condition. For the altered-surface condition, the elderly spent 3.8% of the time more than 20 mm from their mean COP, compared with 1.8%, for the altered-vision condition. Likewise, for the altered-surface condition, the young spent 3.4% of the time more than 20 mm from their COP, compared with 0.9%, for the altered-vision condition. The importance of proprioception to balance control is well documented. Woollacott et al. (1986) have shown that balance control is affected similarly by minimized swayrelated inputs (by stabilizing the force platform with respect to body sway) and by removal of visual information, or by rendering it to be noninformative (by stabilizing vision with respect to body sway). Findings from Diener et al. (1984) further reinforce the importance of proprioception to balance control. The investigators reported that their subjects had to be prevented from falling when proprioceptive information from the lower limbs was eliminated through ischemic blocking and upon presentation of small sinusoidal platform movement. In the present experiment, withdrawal of vision, alone, did not affect the elders' sway significantly more than that of young subjects. In fact, in absolute terms, young subjects were more affected by the visual perturbation than the elders. This lack of expected detrimental effect on the elders' posture could be interpreted as being different from previous results (Sheldon, 1963; Straube et al., 1988). However, we believe the data obtained may simply reflect the conservative nature of our subjects' selection procedures; namely, the assessed posture control of our subjects may have been better due to their overall fitness. In particular, in this study, the altered vision-surface condition highlighted marked age-impairments in balance control. A small degradation of the information provided by the somatosensory system became very detrimental to posture when vision was withdrawn as well. To explain the less stable posture of the elderly, Woollacott et al. (1986) and Straube et al. (1988) have suggested that either a slowing of, or a defect in, the central integrative mechanisms is responsible for reorganizing the postural system. For situations that did not elicit the reflexive mechanisms, our results indicated that the less stable posture of the elders was the result of a defect, rather than a slowing, in the central integrative mechanisms. Indeed, when vision was withdrawn and/or the surface altered, both groups showed similar postural adaptation (i.e., decreased sway across time) to the different perturbations, thus not supporting the hypothesis that a slower reorganizing of the postural control system takes place in the elderly.
Overall, we found that range, variability, velocity, and dispersion of sway behavior were greater in the elderly than in the young subjects. In both groups, range, variability, and velocity of sway increased more under the altered-vision condition than under the altered-surface condition. (This is
The decreased-sway results yielded over time for range, variability, and velocity in this experiment may be of clinical value, regarding timed balance tests (Bohannon et al., 1984; Potvin and Tourtelotte, 1975) often used to measure postural control. Our findings suggest that a small alteration in an
V I S I O N / N O R M A L SURFACE
e
1
1
-J -e 12 VISION/ALTERED SURFACE
t
-3
NO-VISION/NORMAL SURFACE
NO-VISION/ALTERED SURFACE
LJ LJ •—' — 1
1
1
—1—
—i
1
1
1
1
Figure 3. Group differences in the dispersion of sway. Positive values indicate that young subjects spent more time in an area than elderly subjects; negative values indicate that elderly subjects spent more time in an area than young subjects.
elderly subjects spent 2.7, 0.2, 0.2, and 0.1% more time in areasl,5,6,and7,respectively(F(7,182) = 0.22,/?> .05 for the Group x Area interaction. Group differences were most striking for the altered vision-surface condition, as young subjects spent 7.5 and 6.2% more time in areas 1 and 2, respectively, and elderly subjects spent 3.0,3.6,2.6,2.4,1.2, and 0.5% more time in areas 3 , 4 , 5 , 6,7, and 8, respectively (F (7,182) = 121.46, p < .05 for the Group x Area interaction). Hence, in response to both perturbations, the elderly exhibited far more spread in their sway behavior than did the young. More importantly, however, the young subjects were able to compensate for these perturbations, containing their sway within a much smaller area than the elderly subjects. The young spent 93.1% of the time within 20 mm of their COP and only 2.0% outside of 25 mm, whereas the elderly spent 86.0% of the time within 20 mm of their COP and 14.0% outside of 20 mm (6.7% outside 25 mm).
POSTURAL SWAY
individual does not degrade balance control over time but rather improves it, implying some adaptation to change. Hence, the question should be asked whether timed balance tests, which concentrate the body's center of gravity over one leg, reflect the same mechanisms observed in normal standing posture. While such tests have been significantly correlated with age (Bohannon et al., 1984; Era and Heikkinen, 1985; Potvin and Tourtelotte, 1975), they do not necessarily reflect the operation of the different systems responsible for postural control in normal standing and may, instead, be testing more peripheral factors such as muscle strength. In the present investigation, elderly subjects spent 0.4% of the time 20 mm away from their mean COP in the control condition; 1.8% with altered vision; 3.8% with altered surface; and 14.0% with altered vision-surface. Similarly, the young subjects showed 0.1, 0.9, 0.4, and 6.9%, respectively, for the same conditions. The percentage of time spent away from a "central" position may not seem crucial. However, it has been suggested that with age, the size of the standing support-base tends to decrease (Lee and Deming, 1987; Rothstein et al., 1987). In addition, there is need for an increased displacement of sway, before detection of instability occurs (Stelmach and Worringham, 1985). Thus, even a small percentage of time spent away from a central position would be more critical for the elderly than the young in postural control for prevention of a fall. Our results also showed that the exclusion or disruption of only one of the sensory inputs was not always sufficient to differentiate elderly from young persons, which does not support the suggestion that elders are using more "primitive control strategies" based on a greater reliance upon visual information (Pykko et al., 1988; Straube et al., 1988). (It is possible, however, that our elderly subjects were not old enough to exhibit a balance-control strategy based mainly on the reliance of visual information, and that the remaining sources of sensory information were sufficient to allow compensation through reorganization of the postural control system.) When both vision was withdrawn, and proprioceptive information altered, however, the elders' balance control was seriously degraded. Age-related deterioration of balance control may be the result of a degeneration of the different sensory systems responsible for posture control (e.g., Calne, 1985; Stelmach and Worringham, 1985). Specifically, it has been suggested that, with increasing age, the elderly rely less on vestibular input (Straube et al., 1988). Our results indirectly support this idea, as the elders' balance control was seriously impaired when they had to rely on vestibular information (altered vision-surface condition). Alternatively, it is possible that the deterioration of posture is not simply due to a greater reliance on visual information, and lesser reliance on vestibular information, but rather to the inability of the postural control system to reorganize itself when two senses are altered (Diener et al., 1984). ACKNOWLEDGMENTS
Address correspondence to Dr. George E. Stelmach, Exercise and Sport Research Institute, P.E. East, Arizona State University, Tempe, AZ 852870404.
B243
Dr. Normand Teasdale is now at Laval University, Laboratoire de Performance Motrice Humain, Quebec, Canada G1K 7L4. REFERENCES
Allum, J. H.; Pfaltz, C. R. Visual and vestibular contributions to pitch sway stabilization in the ankle muscles of normals and patients with bilateral peripheral vestibular deficits. Exp. Brain Res. 58:82-94; 1985. Appenzeller, O.; Imarisio, J. J.; Gilbert, J. J. The effect of age and neurological disease on the ankle jerk. Arch. Neurol. 15:147-150; 1966. Birren, J. E. Vibratory sensitivity in the aged. J. Gerontol. 2:267-268; 1947. Bohannon, R. W.; Larkin, P.; Cook, A.; Gear, J.; Singer, J. Decrease in timed balance test scores with aging. Phys. Ther. 64:1067-1070; 1984. Brocklehurst, J. C ; Robertson, D.; James-Groom, P. Clinical correlates of sway in old age-sensory modalities. Age Ageing 11:1-10; 1982. Calne, D. B. Normal aging of the nervous systems. In: Andres, R.; Bierman, L.; Hazard, W. R. (eds.). Principles of geriatric medicine. New York: McGraw-Hill, 1985:231-235. Carel, R. S.; Korczyn, A. D.; Hochberg, Y. Age and sex dependency of the achilles tendon reflex. Am. J. Med. Sci., 278:57-63; 1979. Clarkson, P. The relationship of age and level of physical activity with the fractionated components of patellar reflex time. J. Gerontol. 33:650656; 1978. Diener, H. C ; Dichgans, J.; Guschlbauer, B.; Mau, H. The significance of proprioception on postural stabilization as assessed by ischemia. Brain Res. 296:103-109; 1984. Era, P.; Heikkinen, E. Postural sway during standing and unexpected disturbance of balance in random samples of men of different ages. J. Gerontol. 40:287-295; 1985. Harris, G. F.; Knox, T. A.; Larson, S. J.; Sances, A.; Millar, E. A. A method for the display of balance platform center of pressure data. J. Biomechanics 15:741-745; 1982. Johnson, B. M.; Miao, M.; Sadum, A. A. Age-related decline of human optic nerve. Age 10:5-9; 1987. Johnsson, L. G.; Hawkins, J. E. Sensory and neural degeneration with aging, as seen in microdissections of the inner ear. Ann. Otol. Rhinol. Laryngol. 81:179-193; 1972. Kokmen, E.; Bossemeyer, R. W.; Williams, W. J. Quantitative evaluation of joint motion sensation in an aging population. J. Gerontol. 33:62-67; 1978. Lee, W. A.; Deming, L. Correlation between age and the size of the normalized static support base while standing. In: Goodman, D; Reid, G.; McAuley, E. (eds.). Psychology of motor behaviour and sport. Vancouver: University of British Columbia, 1987:41. Murray, M. D.; Seirig, A. A.; Sepic, S. B. Normal postural stability and steadiness: Quantitative assessment. J. Bone Joint Surg. [Am] 57:510— 516;1975. Potvin, A. R.; Tourtelotte, W. W. The neurological examination: Advancements in its quantification. Arch. Phys. Med. Rehabil. 56:425-437; 1975. Pykko, I.; Aalto, H.; Hytonen, M.; Starck, J.; Jantti, P.; Ramsay, H. Effect of age on postural control. In Amblard, B.; Berthoz, A.; Clarac, F. (eds.). Posture and gait: Development, adaptation and modulation. Amsterdam: Elsevier, 1988:105-114. Rosenhall, U.; Rubin, W. Degenerative changes in the human vestibular sensory epithelia. Acta Otolaryngol. 79:67-81; 1975. Rothstein, D. E.; Larish, D. D.; Mungiole, M. Age-related differences in dynamic postural control. The Gerontologist 27 (Suppl.): 160A; 1987. Salthouse, T. A. A theory of cognitive aging. Amsterdam: Elsevier; 1985. Sekuler, R.; Hutman, L. P.; Owsley, C. J. Human aging and spatial vision. Science 209:1255-1256; 1980. Sheldon, J. H. The effect of age on the control of sway. Gerontologia Clinica 5:129-138; 1963. Skinner, H. B.; Barrack, R. L.; Cook, S. D. Age-related decline in proprioception. Clin. Orthop. Related Res. 184:208-211; 1984. Stelmach, G. E.; Sirica, A. Aging and proprioception. Age 9:99-103; 1986. Stelmach, G. E.; Worringham, C. J. Sensorimotor deficits related to postural stability: Implications for falling in the elderly. Clin. Geriatr. Med. 1:679-694; 1985. Stelmach, G. E.; Teasdale, N.; DiFabio, R. P.; Phillips, J. Age-related
B244
TEASDALE ET AL.
decline in postural control mechanisms. Int. J. Aging Human Dev. 29:205-223; 1989a. Stelmach, G. E.; Phillips, J.; DiFabio, R. P.; Teasdale, N. Age, functional postural reflexes, and voluntary sway. J. Gerontol. Biol. Sci. 44:B100B106; 1989b. Straube, A.; Botzel, K.; Hawken, M.; Paulus, W.; Brandt, T. Postural control in the elderly: Differential effects of visual, vestibular, and somatosensory input. In: Amblard, B.; Berthoz, A.; Clarac, F. (eds.). Posture and gait: Development, adaptation and modulation. Amster-
dam: Elsevier, 1988:105-114. Whanger, A. D.; Wang, H. S. Clinical correlates of the vibratory sense in elderly psychiatric patients. J. Gerontol. 29:39-45; 1974. Woollacott, M. H.; Shumway-Cook, A.; Nashner, L. M. Aging and posture control: Changes in sensory organization and muscular coordination. Int. J. Aging Human Dev. 23:97-114; 1986. Received October 2, 1989 Accepted June 5, 1990
Copyright 1991 by The Gerontological Society of America
Postural Sway Characteristics of the Elderly Under Normal and Altered Visual and Support Surface Conditions Normand Teasdale,1 George E. Stelmach,2 and Ann Breunig3 'Laval University, Quebec, Canada. Exercise and Sport Research Institute, Arizona State University-Tempe. 3 Motor Behavior Laboratory, University of Wisconsin-Madison.
2
One of the most pervasive findings in the literature on the aged is the general slowing of cognitive-motor responses with advancing age. Hence, an increased slowness in the processing of information from vestibular, visual, and somatosensory systems could contribute greatly to a decline in postural stability. To examine this question, in a crosssectional investigation, postural sway behavior of elderly (n = 18) and young (n = 10) adults was examined under conditions that stressed the slower integrative mechanisms rather than the reflexive mechanisms ofpostural control. The postural sway behavior ofyoung and elderly subjects was examined for a prolonged duration (80 s), under altered visual and/or support surface (5 cm thickfoam surface) conditions, and contrasted with normal stance. Results showed that the exclusion or disruption of one of the sensory inputs, alone, was not consistently sufficient to differentiate between elderly and young adults, because of compensation by the remaining sensory sources. Both alterations together (i.e., visual and surface), however, had a substantially greater effect upon the elderly than the young.
N
UMEROUS experimenters have demonstrated that postural sway increases with age (e.g., Era and Heikkinen, 1985; Murray et al., 1975; Sheldon, 1963). Very little is known, however, about how the different sensory systems responsible for control of posture are affected by the aging process. Both the monosynaptic (Appenzeller et al., 1966; Carel et al., 1979; Clarkson, 1978) and the long-latency reflexive systems (Woollacott et al., 1986) have been shown to exhibit small latency increases with age. In a 1989 study, Stelmach, Teasdale, DiFabio, and Phillips found that elders' sway and reflexive responses, during stance, to large-fast rotational ankle perturbations (7 deg at 20 deg/s) were similar to those of young subjects, suggesting that the elders' reflexive mechanisms were relatively intact. When smallslow rotational perturbations (2 deg at 4 deg/s) were given, however, the elders swayed more than the young subjects and could not adapt to repetitive presentations of perturbations. These results were interpreted as evidence that the aged may be prone to sway when posture is controlled by slower, higher-level sensory integrative mechanisms, inasmuch as the small-slow perturbation corrections took place mainly through the slower and higher level sensory systems (Stelmach etal., 1989b). A possible cause for the decreased stability observed with increasing age is the deterioration of primary peripheral sensory processes. Indeed, elderly individuals have been shown to have higher proprioceptive thresholds to passive movement than the young (Birren, 1947; Skinner et al., 1984; Whanger and Wang, 1974) and are less accurate at reproducing and matching joint angles (Kokmen et al., 1978; Stelmach and Sirica, 1986). In fact, from a clinical standpoint, a small and progressive degeneration of sensory input from the lower extremities is often the first, and most B238
common, manifestation of aging (Calne, 1985). Similarly, diminution of the axon population in the optic nerves (Johnson et al., 1987), loss of spatial sensitivity (Sekuler et al., 1980), reduction of hair cells in the semicircular canals (Rosenhall and Rubin, 1975), and reduction of the macula of the uticulus and sacculus (Johnsson and Hawkins, 1972) have all been reported with aging. Yet, little evidence has been found to support the suggestion that a decrease in peripheral sensory acuity is associated with increased postural sway (Brocklehurst et al., 1982; Era and Heikkinen, 1985). The lack of an obvious causal relationship between specific peripheral sensory impairments and postural sway implies that posture control is dependent upon complex, integrative processing from a variety of inputs. Stelmach and Worringham (1985) suggested that deterioration of the central integrative processes may further explain the decreased postural stability observed with aging. When reflexive mechanisms are not elicited, the central nervous system must undertake certain processing stages (i.e., peripheral input, response selection, and execution) to prevent a fall. Increased slowness in processing information from the vestibular, visual, and somatosensory systems could be a major source of disruption in postural stability (Allum and Pfaltz, 1985; Diener etal., 1984; Skinner etal., 1984). Indeed, one of the most convincing findings in the literature on aging is the general slowing of cognitive-motor responses (e.g., Salthouse, 1985). For this reason we examined postural control in a sample aging population under conditions that stressed the slower integrative mechanisms rather than the reflexive mechanisms of posture control. This experiment was designed to determine whether a reduction of visual, somatosensory, or both visual and somatosensory inputs were differentially disruptive to posture
POSTURAL SWAY
control across young and older adults. Sway behavior was examined for prolonged periods of upright posture (80 s) and quantified, using several dependent measures to maximize the ability to localize age changes. METHODS
Subjects. — Ten young subjects (5 males and 5 females, mean age 21.5 yrs, range 21—22) and 18 elderly subjects (9 males and 9 females, mean age 74 yrs, range 70-80) participated in the experiment. The elderly subjects were volunteers from local aging coalitions and support groups, and the young subjects were students from the University of Wisconsin-Madison. Elderly subjects, chosen from a large pool of subjects, had to meet the following criteria: no musculoskeletal defects; not taking any medication; not undergoing treatment for any neurological disease; no history of falling. This conservative selection procedure serves to minimize any age-related differences. All subjects gave informed consent for the procedures used. Apparatus. — A force platform with 4 load cells (AMTI model OR6-5-1) measured force and moment components along the x, y, and z axes. The signals from the force platform were amplified (AMTI SGA6-4) before being processed by a computer (PDP 11-73). All signals were sampled at 40 Hz and then filtered with a dual-pass Butterworth second-order filter (10 Hz cut-off) to remove any highfrequency artifacts. Sway behavior in the sagittal and frontal planes was then computed. Procedures. — The experiment comprised four different conditions: (a) control vision and normal surface, referred to as the normal vision-surface condition; (b) normal vision and altered surface, referred to as the altered surface condition; (c) normal surface and no vision, referred to as the alteredvision condition; and (d) no vision and altered surface, referred to as the altered vision-surface condition. Subjects stood barefoot on the force platform with feet together. First, the support surface was normal for the control (eyes open) and altered-vision (eyes closed) conditions. Then the support surface was altered by adding an open-cell, polyurethane foam surface (5 cm thick) upon which the subjects stood, for the altered-surface (eyes open) and altered vision-surface (eyes closed) conditions. The addition of a foam support surface alters the reliability of the somatosensory contributions to posture control and, therefore, is believed to force greater reliance on vision and/or vestibular information (Straube et al., 1988). A trial consisted of four sampling periods of 20 s each, spaced at 3-5 s intervals. Two consecutive trials (i.e., 80 s of total sampling/trial) were run for each condition. Data analysis. — Numerous sway measures and methods of displaying sway characteristics have been presented in the literature. No attempt has been made, however, to determine whether these measures reflect different characteristics of sway behavior. In the present experiment, sway was evalu-
B239
ated by using four different sway measures to determine whether they reflected differential characteristics of sway behavior and whether certain of these measures were agesensitive. Specifically used were range and standard deviations of sway behavior in both planes, average sway velocity, and sway density histograms. The range is thought to give an indication of how far center-of-foot pressure (COP) deviates from its original baseline, whereas standard deviation gives the variability of balance control over the sampled period. The range of sway behavior may be misleading because it does not provide information regarding the distribution of the COP within that range. Nevertheless, range was employed, as it is often reported in the literature and represents a standard of comparison with other published studies. Average sway velocity is the sum of the displacement vectors divided by total sampling time. In this experiment, the sampling period was constant for all conditions and subjects. Hence, the average sway velocity is by definition proportional to the summation of all the postural displacements of a given subject. Thus, for comparison purposes, the greater the value, the greater the displacement, and, therefore, the greater the average velocity. Finally, a measure of sway derived from density histograms of the sway path (Harris et al., 1982) was applied. Specifically, density histograms give a qualitative description of the distribution of sway behavior. Distributions were quantified by calculating an average COP and the percentage of time the subjects spent in eight arbitrarily defined concentric circles centered around the average COP in radial increments of 5 mm. Each of these areas was labeled from area 1 (closest to mean COP) to area 8 (farthest from mean COP). Thus, a higher percentage of time spent away from the average COP could indicate a postural behavior at risk. On average, the young and elderly spent more than 99% of the time within a radius of 40 mm of their mean center of foot pressure (COP). Data obtained for the sway range and standard deviation of sway range were submitted to a Group x Sway Plane X Vision x Surface x Sampling Period Analysis of Variance (ANOVA), with repeated measures on the last four factors. Data obtained for the sway velocity, after a square root transformation was performed, were submitted to a Group x Vision x Surface x Sampling Period ANOVA, with repeated measures on the last three factors. The raw data for sway velocity are presented in the text. Finally, to highlight group differences in each vision/surface experimental condition, the dispersion of sway data were submitted to Group x Area (8 radii of 5 mm) ANOVAs, with repeated measures on the last factor. RESULTS
In this experiment, the normal vision-surface condition represented a control condition, as none of the primary sensory inputs responsible for postural control (i.e., visual, vestibular, and proprioceptive) was altered. On the other hand, manipulation of a specific sensory input gave an estimation of the importance of that information for posture control in addition to indicating how the central nervous system adapted and reorganized information provided by the remain-
B240
TEASDALE ET AL.
ing sources of sensory information. A stability effect was derived from each of the different sensory input alterations. Sway range. — Overall, for group comparisons, the elderly had a greater sway range than did the young subjects (27.3 mm vs 21.8 mm; F (1,26) = 6.67, p < .01). Unlike the young subjects, who had a similar range of sway in both planes (21.6 mm and 22.0 mm), the elderly subjects exhibited greater sway along the medio-lateral plane than along the anterior-posterior plane (29.2 mm vs 25.4 mm; F (1,26) = 4.97, p < .05 for the Sway Plane x Group interaction. However, as the sway plane factor was significant only in this case, data obtained for both planes were pooled. Both groups were affected similarly by visual conditions (F (1,26) = 1.65, p > .05 for the Vision x Group interaction. As may be seen in Table 1, withdrawing vision produced an 11.2 mm increase in sway (cost) for the young (from 14.8 mm to 26.0 mm), and a 10.8 mm increase in sway (cost) for the elderly (from 17.4 mm to 28.2 mm); (F (1,26) = 97.08, p < .001 for the main effect of vision). Young and elderly showed different results for the alteredsurface condition; the sway range of the young subjects only increased by 2.1 mm, whereas it increased by 4.2 mm for the elderly (F (1,26) = 12.97, p < .005 for the Group x Surface interaction. However, the combined-perturbations condition (altered vision-surface) produced the most dramatic cost of all, especially in the elderly subjects (F (1,26) = 14.05,/? < .001 for the Surface x Vision x Group triple interaction. Indeed, with vision withdrawn, surface alteration created a 15.0 mm sway-range increase in the young, and a 24.7 mm sway-range increase in the elderly. No indication of any increased sway range over time appeared for either elderly or young subjects. In fact, for both groups, during the altered-vision condition there was a gradual decrease in sway range from the first 20 s of sampling to the last 20 s of sampling (7.0 mm for the elderly and 4.1 mm for the young subjects). No such decrease was observed for the altered-surface condition (F (3,78) = 2.67, p < .05). With vision, sway range remained relatively constant (a decrease of 0.9 mm for both groups; F (3,78) = 8.07, p < .001 for the Vision x Sampling Period interaction. Variability of the sway range. — A pattern of results similar to that of the sway range was obtained for variability. Overall, the elderly subjects exhibited more variability than did the young subjects (5.6 mm vs 4.6 mm; F( 1,26) = 5.08, p < .05). The young subjects exhibited similar variability in both planes (4.7 mm and 4.5 mm), whereas the elderly subjects exhibited greater sway along the medio-lateral plane than along the anterior-posterior plane (5.8 mm vs 5.3 mm; F ( 1,26) = 5.98, p< .05 for the Sway Plane x Group interaction. Again, as the sway plane factor yielded no other significant effect, the data obtained for both planes were pooled. Variability (Table 1) for both groups was similar under the altered-vision condition (2.0 mm increase, or cost, for the elderly vs 2.2 mm for the young; F (1,26) = 0.96, p > .05 for the Vision x Group interaction). Elderly subjects exhibited greater variability than young subjects under the alteredsurface condition (0.8 mm increased variability for the el-
Table 1. Sway Range, Variability, and Velocity for Visual, Surface, and Combined Visual-Surface Perturbations Condition Control Cost of altered vision Cost of altered surface Cost of altered vision and surface
Sway range (mm)
Variability (mm)
Velocity (mm/s)
Young
Young
Young
Elders
Elders
Elders
14.8
17.4
3.2
3.7
10.0
14.3
11.2
10.8"
2.2
2.0"
7.2
13.8*
0.4
0.8»"
2.1
3.2-
3.0
4.7'
10.9
25.2
b
24.7C
Note. Values in row Control represent sway range, variability, and velocity, exhibited by both young and elderly subjects, under the normal surface-vision condition. Values, or costs, in rows Altered vision, Altered surface, and Altered vision and surface represent the sway range, variability, and velocity in excess of that exhibited in the control condition. "Main Effect x Group two-way interaction. b Main Effect x Group two-way interaction. 'Vision x Surface x Group three-way interaction.
derly vs 0.4 mm for the young; F( 1,26) = 40.82,/? < .001). Just as the sway range for the elderly proved to be greatest under the altered vision-surface condition, so did its variability (F (1,26) = 11.01,/? < .001 for the Surface X Vision x Group triple interaction). Indeed, when vision was withdrawn and the surface altered, the variability of the elderly subjects increased by 4.7 mm, while it only increased by 3.0 mm for the young subjects. For both groups, there was a small, but significant, decrease in the variability of sway under the altered-vision condition over time. Between the first and last 20 s of sampling, there was a 1.1 mm decrease, without vision, and only a 0.1 mm decrease, with vision (F (3,78) = 7.53, p < .001 for the Sampling Period x Vision interaction. No such decrease was observed with altered surface, 0.6 mm with altered surface, and 0.4 mm with normal surface (F (3,78) = 2.19, /? < .05 for the Sampling Period X Surface interaction. Sway velocity. — Overall, the elderly had a greater average velocity than did the young subjects, and thus swayed more (24.9 mm/s vs 15.7 mm/s; F (1,26) = 9.49,/? < .005). As seen in Table 1, both groups had greater velocities when vision was withdrawn (increase of 7.2 mm/s for the young, and 13.8 mm/s for the elderly; (F (1,26) = 46.97, p < .001 for the main effect of vision). Despite a larger cost for the elderly than for the young subjects, the Vision x Group interaction did not reach the significance level (F (1,26) = 2.81, p > .05). Under the altered-surface condition, the young subjects showed a 2.1 mm/s increase, and the elderly, a 3.2 mm/s increase, in sway velocity (F (1,26) = 4.20,/? < .05 for the Surface x Group interaction. Again, the combination of both perturbations was most detrimental to the elderly's posture (F (1,26) = 4.23, p < .05 for the Surface x Vision X Group triple interaction. When both surface and vision were altered, the young
POSTURAL SWAY
subjects exhibited a 10.9 mm/s increase, whereas the elderly subjects exhibited a 25.2 mm/s increase, in sway velocity. Table 1 shows that the postural cost of both alterations was nearly additive for the young subjects, but multiplicative for the elders. As in the cases of sway range and variability, the sway velocity decreased over time, implying some occurrence of adaptation. For the young subjects, sway velocity decreased from 17.4 mm/s to 14.8 mm/s (from the first sampling period to the last sampling period), whereas for the elderly subjects, it decreased from 27.5 mm/s to 23.2 mm/s. A decomposition of the main effect of the sampling period into its orthogonal components showed that the linear and quadratic components were statistically significant (F (1,26) = 58.64 and 25.26, p < .001), suggesting that there was a more rapid decrease in velocity at the beginning, than at the end, of the sampling periods. Moreover, the decrease was more rapid for the young than for the elderly (100% of the decrease in velocity occurred within the first sampling period for the young, whereas only 49% of the decrease occurred within that time period for the elderly; F (1,26) = 4.88, p < .05 for the quadratic component of the Sampling Period x Group interaction).
B241
elderly subjects spent 4.1, 4.9, 1.3, 0.5, and 0.2% more time in areas 2, 3, 4, 5, and 6, respectively (F (7,182) = 3.43, p < .005 for the Group x Area interaction). For the altered-vision condition, both groups showed a similar dispersion of sway. Indeed, young subjects spent 2.1,0.6, and 0.4% more time in areas 2, 3, and 4, respectively, and
18r
2 15 -
(mm) Medio-Lateral
Dispersion of sway. — Representative density histograms of the sway path for an elderly and a young subject appear in Figures 1 and 2, respectively. Table 2 shows the percentage of time spent by elderly and young subjects in the eight areas centered around COP, under the different visual and surface experimental conditions. In Figure 3, group differences in the dispersion of sway were obtained by subtracting the percentages of time spent by young subjects in a given area from those of the elderly in that same area. Positive values indicate that young subjects spent more time in an area than elderly subjects, whereas negative values indicate that elderly subjects spent more time in an area than young subjects. As seen in Figures 1 and 2, elderly sway behavior spread over a wider area than that of the young subjects for the normal vision-surface, or control, condition. Young subjects spent 10.9% more time in area 1 than elderly subjects, while the elderly subjects spent 3.8, 5.0, 1.7, and 0.3% more time in areas 2, 3, 4, and 5, respectively, than the young subjects (F (7,182) = 3.64, p < .005 for the Group x Area interaction). The altered-surface condition yielded similar results; the young spent 11.2% more time in area 1, and the
RIGHT
Figure 1. Density histogram of the sway path of an elderly subject.
18r
-40 (mm) Medio-Lateral
Figure 2. Density histogram of the sway path of a young subject.
Table 2. Percentage of Time Spent by Young and Elderly Subjects in the Eight Areas Centered Around COP, Under Different Experimental Conditions Vision/normal surface
Vision/altered surface
No-vision/ normal surface
No-vision/ altered surface
Area
Young
Elders
Young
Elders
Young
Elders
Young
Elders
1 2 3 4 5 6 7 8
54.3 38.3
43.5 42.1 11.9
48.1 38.2 10.2
36.9 42.3 15.1
28.8 39.6 20.0
31.5 37.5 19.4
2.1 0.3 0.0 0.0 0.0
2.6 0.4 0.2 0.1 0.0
3.9 0.9 0.4 0.1 0.0
8.2 2.5 0.6 0.1 0.0
7.8 2.7 0.8 0.2 0.0
25.3 36.6 20.9 10.3
17.8 30.4 23.9 13.9
4.7 1.3 0.5 0.2
7.3 3.7 1.7 0.7
6.9 0.4 0.0 0.0 0.0 0.0
TEASDALE ET AL.
B242
DISCUSSION
not to suggest that vision is more important to balance control than proprioception; the degree to which proprioceptive information was altered in this experiment cannot be precisely documented). Dispersion of sway data showed that both groups spent a higher percentage of time in an "at risk'' posture-control area, centered around COP, for the alteredsurface condition than for the altered-vision condition. For the altered-surface condition, the elderly spent 3.8% of the time more than 20 mm from their mean COP, compared with 1.8%, for the altered-vision condition. Likewise, for the altered-surface condition, the young spent 3.4% of the time more than 20 mm from their COP, compared with 0.9%, for the altered-vision condition. The importance of proprioception to balance control is well documented. Woollacott et al. (1986) have shown that balance control is affected similarly by minimized swayrelated inputs (by stabilizing the force platform with respect to body sway) and by removal of visual information, or by rendering it to be noninformative (by stabilizing vision with respect to body sway). Findings from Diener et al. (1984) further reinforce the importance of proprioception to balance control. The investigators reported that their subjects had to be prevented from falling when proprioceptive information from the lower limbs was eliminated through ischemic blocking and upon presentation of small sinusoidal platform movement. In the present experiment, withdrawal of vision, alone, did not affect the elders' sway significantly more than that of young subjects. In fact, in absolute terms, young subjects were more affected by the visual perturbation than the elders. This lack of expected detrimental effect on the elders' posture could be interpreted as being different from previous results (Sheldon, 1963; Straube et al., 1988). However, we believe the data obtained may simply reflect the conservative nature of our subjects' selection procedures; namely, the assessed posture control of our subjects may have been better due to their overall fitness. In particular, in this study, the altered vision-surface condition highlighted marked age-impairments in balance control. A small degradation of the information provided by the somatosensory system became very detrimental to posture when vision was withdrawn as well. To explain the less stable posture of the elderly, Woollacott et al. (1986) and Straube et al. (1988) have suggested that either a slowing of, or a defect in, the central integrative mechanisms is responsible for reorganizing the postural system. For situations that did not elicit the reflexive mechanisms, our results indicated that the less stable posture of the elders was the result of a defect, rather than a slowing, in the central integrative mechanisms. Indeed, when vision was withdrawn and/or the surface altered, both groups showed similar postural adaptation (i.e., decreased sway across time) to the different perturbations, thus not supporting the hypothesis that a slower reorganizing of the postural control system takes place in the elderly.
Overall, we found that range, variability, velocity, and dispersion of sway behavior were greater in the elderly than in the young subjects. In both groups, range, variability, and velocity of sway increased more under the altered-vision condition than under the altered-surface condition. (This is
The decreased-sway results yielded over time for range, variability, and velocity in this experiment may be of clinical value, regarding timed balance tests (Bohannon et al., 1984; Potvin and Tourtelotte, 1975) often used to measure postural control. Our findings suggest that a small alteration in an
V I S I O N / N O R M A L SURFACE
e
1
1
-J -e 12 VISION/ALTERED SURFACE
t
-3
NO-VISION/NORMAL SURFACE
NO-VISION/ALTERED SURFACE
LJ LJ •—' — 1
1
1
—1—
—i
1
1
1
1
Figure 3. Group differences in the dispersion of sway. Positive values indicate that young subjects spent more time in an area than elderly subjects; negative values indicate that elderly subjects spent more time in an area than young subjects.
elderly subjects spent 2.7, 0.2, 0.2, and 0.1% more time in areasl,5,6,and7,respectively(F(7,182) = 0.22,/?> .05 for the Group x Area interaction. Group differences were most striking for the altered vision-surface condition, as young subjects spent 7.5 and 6.2% more time in areas 1 and 2, respectively, and elderly subjects spent 3.0,3.6,2.6,2.4,1.2, and 0.5% more time in areas 3 , 4 , 5 , 6,7, and 8, respectively (F (7,182) = 121.46, p < .05 for the Group x Area interaction). Hence, in response to both perturbations, the elderly exhibited far more spread in their sway behavior than did the young. More importantly, however, the young subjects were able to compensate for these perturbations, containing their sway within a much smaller area than the elderly subjects. The young spent 93.1% of the time within 20 mm of their COP and only 2.0% outside of 25 mm, whereas the elderly spent 86.0% of the time within 20 mm of their COP and 14.0% outside of 20 mm (6.7% outside 25 mm).
POSTURAL SWAY
individual does not degrade balance control over time but rather improves it, implying some adaptation to change. Hence, the question should be asked whether timed balance tests, which concentrate the body's center of gravity over one leg, reflect the same mechanisms observed in normal standing posture. While such tests have been significantly correlated with age (Bohannon et al., 1984; Era and Heikkinen, 1985; Potvin and Tourtelotte, 1975), they do not necessarily reflect the operation of the different systems responsible for postural control in normal standing and may, instead, be testing more peripheral factors such as muscle strength. In the present investigation, elderly subjects spent 0.4% of the time 20 mm away from their mean COP in the control condition; 1.8% with altered vision; 3.8% with altered surface; and 14.0% with altered vision-surface. Similarly, the young subjects showed 0.1, 0.9, 0.4, and 6.9%, respectively, for the same conditions. The percentage of time spent away from a "central" position may not seem crucial. However, it has been suggested that with age, the size of the standing support-base tends to decrease (Lee and Deming, 1987; Rothstein et al., 1987). In addition, there is need for an increased displacement of sway, before detection of instability occurs (Stelmach and Worringham, 1985). Thus, even a small percentage of time spent away from a central position would be more critical for the elderly than the young in postural control for prevention of a fall. Our results also showed that the exclusion or disruption of only one of the sensory inputs was not always sufficient to differentiate elderly from young persons, which does not support the suggestion that elders are using more "primitive control strategies" based on a greater reliance upon visual information (Pykko et al., 1988; Straube et al., 1988). (It is possible, however, that our elderly subjects were not old enough to exhibit a balance-control strategy based mainly on the reliance of visual information, and that the remaining sources of sensory information were sufficient to allow compensation through reorganization of the postural control system.) When both vision was withdrawn, and proprioceptive information altered, however, the elders' balance control was seriously degraded. Age-related deterioration of balance control may be the result of a degeneration of the different sensory systems responsible for posture control (e.g., Calne, 1985; Stelmach and Worringham, 1985). Specifically, it has been suggested that, with increasing age, the elderly rely less on vestibular input (Straube et al., 1988). Our results indirectly support this idea, as the elders' balance control was seriously impaired when they had to rely on vestibular information (altered vision-surface condition). Alternatively, it is possible that the deterioration of posture is not simply due to a greater reliance on visual information, and lesser reliance on vestibular information, but rather to the inability of the postural control system to reorganize itself when two senses are altered (Diener et al., 1984). ACKNOWLEDGMENTS
Address correspondence to Dr. George E. Stelmach, Exercise and Sport Research Institute, P.E. East, Arizona State University, Tempe, AZ 852870404.
B243
Dr. Normand Teasdale is now at Laval University, Laboratoire de Performance Motrice Humain, Quebec, Canada G1K 7L4. REFERENCES
Allum, J. H.; Pfaltz, C. R. Visual and vestibular contributions to pitch sway stabilization in the ankle muscles of normals and patients with bilateral peripheral vestibular deficits. Exp. Brain Res. 58:82-94; 1985. Appenzeller, O.; Imarisio, J. J.; Gilbert, J. J. The effect of age and neurological disease on the ankle jerk. Arch. Neurol. 15:147-150; 1966. Birren, J. E. Vibratory sensitivity in the aged. J. Gerontol. 2:267-268; 1947. Bohannon, R. W.; Larkin, P.; Cook, A.; Gear, J.; Singer, J. Decrease in timed balance test scores with aging. Phys. Ther. 64:1067-1070; 1984. Brocklehurst, J. C ; Robertson, D.; James-Groom, P. Clinical correlates of sway in old age-sensory modalities. Age Ageing 11:1-10; 1982. Calne, D. B. Normal aging of the nervous systems. In: Andres, R.; Bierman, L.; Hazard, W. R. (eds.). Principles of geriatric medicine. New York: McGraw-Hill, 1985:231-235. Carel, R. S.; Korczyn, A. D.; Hochberg, Y. Age and sex dependency of the achilles tendon reflex. Am. J. Med. Sci., 278:57-63; 1979. Clarkson, P. The relationship of age and level of physical activity with the fractionated components of patellar reflex time. J. Gerontol. 33:650656; 1978. Diener, H. C ; Dichgans, J.; Guschlbauer, B.; Mau, H. The significance of proprioception on postural stabilization as assessed by ischemia. Brain Res. 296:103-109; 1984. Era, P.; Heikkinen, E. Postural sway during standing and unexpected disturbance of balance in random samples of men of different ages. J. Gerontol. 40:287-295; 1985. Harris, G. F.; Knox, T. A.; Larson, S. J.; Sances, A.; Millar, E. A. A method for the display of balance platform center of pressure data. J. Biomechanics 15:741-745; 1982. Johnson, B. M.; Miao, M.; Sadum, A. A. Age-related decline of human optic nerve. Age 10:5-9; 1987. Johnsson, L. G.; Hawkins, J. E. Sensory and neural degeneration with aging, as seen in microdissections of the inner ear. Ann. Otol. Rhinol. Laryngol. 81:179-193; 1972. Kokmen, E.; Bossemeyer, R. W.; Williams, W. J. Quantitative evaluation of joint motion sensation in an aging population. J. Gerontol. 33:62-67; 1978. Lee, W. A.; Deming, L. Correlation between age and the size of the normalized static support base while standing. In: Goodman, D; Reid, G.; McAuley, E. (eds.). Psychology of motor behaviour and sport. Vancouver: University of British Columbia, 1987:41. Murray, M. D.; Seirig, A. A.; Sepic, S. B. Normal postural stability and steadiness: Quantitative assessment. J. Bone Joint Surg. [Am] 57:510— 516;1975. Potvin, A. R.; Tourtelotte, W. W. The neurological examination: Advancements in its quantification. Arch. Phys. Med. Rehabil. 56:425-437; 1975. Pykko, I.; Aalto, H.; Hytonen, M.; Starck, J.; Jantti, P.; Ramsay, H. Effect of age on postural control. In Amblard, B.; Berthoz, A.; Clarac, F. (eds.). Posture and gait: Development, adaptation and modulation. Amsterdam: Elsevier, 1988:105-114. Rosenhall, U.; Rubin, W. Degenerative changes in the human vestibular sensory epithelia. Acta Otolaryngol. 79:67-81; 1975. Rothstein, D. E.; Larish, D. D.; Mungiole, M. Age-related differences in dynamic postural control. The Gerontologist 27 (Suppl.): 160A; 1987. Salthouse, T. A. A theory of cognitive aging. Amsterdam: Elsevier; 1985. Sekuler, R.; Hutman, L. P.; Owsley, C. J. Human aging and spatial vision. Science 209:1255-1256; 1980. Sheldon, J. H. The effect of age on the control of sway. Gerontologia Clinica 5:129-138; 1963. Skinner, H. B.; Barrack, R. L.; Cook, S. D. Age-related decline in proprioception. Clin. Orthop. Related Res. 184:208-211; 1984. Stelmach, G. E.; Sirica, A. Aging and proprioception. Age 9:99-103; 1986. Stelmach, G. E.; Worringham, C. J. Sensorimotor deficits related to postural stability: Implications for falling in the elderly. Clin. Geriatr. Med. 1:679-694; 1985. Stelmach, G. E.; Teasdale, N.; DiFabio, R. P.; Phillips, J. Age-related
B244
TEASDALE ET AL.
decline in postural control mechanisms. Int. J. Aging Human Dev. 29:205-223; 1989a. Stelmach, G. E.; Phillips, J.; DiFabio, R. P.; Teasdale, N. Age, functional postural reflexes, and voluntary sway. J. Gerontol. Biol. Sci. 44:B100B106; 1989b. Straube, A.; Botzel, K.; Hawken, M.; Paulus, W.; Brandt, T. Postural control in the elderly: Differential effects of visual, vestibular, and somatosensory input. In: Amblard, B.; Berthoz, A.; Clarac, F. (eds.). Posture and gait: Development, adaptation and modulation. Amster-
dam: Elsevier, 1988:105-114. Whanger, A. D.; Wang, H. S. Clinical correlates of the vibratory sense in elderly psychiatric patients. J. Gerontol. 29:39-45; 1974. Woollacott, M. H.; Shumway-Cook, A.; Nashner, L. M. Aging and posture control: Changes in sensory organization and muscular coordination. Int. J. Aging Human Dev. 23:97-114; 1986. Received October 2, 1989 Accepted June 5, 1990
