Relationship between standing posture and stability.

Key Words: Center of gravity, Measurement, Posture, Stability, Vestibular hypofunction.

Quantification of the relationship between standing posture and stability has received limited attention.[1-3] Posture and stability require the integration of mechanical, sensory, and motor processing strategies that permit upright standing. Posture can be defined as the rotational and translational positions of adjoining body segments and their orientation relative to gravity.[2,3] We define stability as the ability to control center-of-gravity (COG) amplitude and velocity of displacement while remaining standing. Generally, smaller amplitudes and velocities of displacement of the COG yield greater stability. Unpublished research[2] has shown that individuals without vestibular impairment efficiently maintain equilibrium in quiet standing with minimal muscular activation, metabolic cost, and joint loading. Posture and stability are coupled mechanically.[2,3] Body-segment alignment or postural changes affect COG location, which may alter stability.[2] There appears to be no standard evaluation and treatment method used for patients with impaired stability. For example, increasing trunk flexion would alter whole-body COG position, which may in turn impair standing stability. Whether treatment for postural alignment impairments improves standing stability, however, is unknown.

Reports of posture analysis using joint kinematics have been limited. The total body mass can be assumed to be concentrated at the COG without an alteration of the body's translational inertia properties.[3] In quiet standing, the whole-body COG is in constant motion.[4,5] The locations of the joint center and the COG determine the rotatory moment of the body and its extremities. Gravity produces rotatory movements via external mechanical force imposition.[6] Braune and Fischer[7] concluded that the knee, hip, shoulder, and ear normally align perfectly in the sagittal plane. Other authors[8-11] have reported that the knee, hip, and shoulder joints and the ear are aligned anterior to the ankle joint. Some authors[8,9,12,13] have suggested that the COG is anterior to the ankle, knee, and shoulder joints but posterior to the hip joint and ear in individuals without impairment during standing with feet apart. Woodhull et al[8] reported that in typical standing with the feet 20 to 30 cm apart, a slight gravitational force extends the hip and knee joints and posteriorly tilts the pelvis. Levine et al[14] reported that female subjects without impairment had an 11.3-degree anterior pelvic tilt during quiet standing, as measured using a VICON three-dimensional kinematic system.(*)[1] Five infrared-sensing television cameras detected reflective targets that were affixed to anatomical landmarks.[14]

Persons with acute bilateral vestibular hypofunction (BVH) or unilateral vestibular hypofunction (UVH) have been reported to display postural aberrations.[15] Horak and Shupert[15] reported that 4 subjects with BVH, aged 49 to 69 years, had a more forward head position than did subjects without impairment. They theorized that individuals with acute BVH align themselves in quiet standing near the posterior limit of stability (by extending the trunk) or near the anterior limit of stability.[15] Takemori et al[16] reported that 12 patients with unilateral labyrinthectomies, 3 patients with unilateral VIII nerve section, and 22 patients with unilateral streptomycin sulfate infusion into the middle ear cavity had a shift in the COG toward the involved side during quiet standing in the acute phase. Quantitative measures of variables that reflect posture in persons with chronic vestibular hypofunction have not been reported.

Individuals with vestibular hypofunction are reported to be less stable than individuals without impairment.[5,17] In one study,[5] subjects with BVH standing with feet together and eyes closed had higher phase-plane stability than did subjects without impairment. Phase-plane stability can be defined as the combined variances in whole-body COG displacement and velocity in the anteroposterior and mediolateral planes. Center-of-gravity displacement variances alone do not discriminate between persons with BVH and persons without impairments, thus establishing the importance of measuring momentum (or velocity) control for analyzing stability.[5] Phase-plane stability has not been analyzed in individuals with UVH. The loss of unilateral vestibular function, however, has been reported to cause an increase in body sway when there are coexisting impairments in both the visual and somatosensory systems.[17]

There are no published reports that describe, using quantitative methods, whole-body COG displacement changes when an individual stands with feet together and eyes closed as compared with standing with feet apart and eyes open. Clinicians frequently evaluate and treat balance impairments in subjects with visual, proprioceptive, and vestibular system impairments. All three systems are critical sources of afferent information that influence the control of stability.[18] By eliminating visual input and narrowing the base of support, it is possible to investigate stability in a more challenging situation. Information regarding whole-body COG displacement under various conditions would be useful in clinical decision making because control of whole-body COG is the goal with individuals who have stability impairments.

The primary purpose of our study was to determine whether persons with stability impairments have postural aberrations. Individuals with vestibular hypofunction have been reported to have stability impairments.[5,17] We therefore wanted to determine whether these individuals also display postural deviations. We hypothesized that any subject whose neck joint (approximated at the atlanto-occipital joint) was anterior to the whole-body COG (forward head) and whose hip joint was posterior to the COG would display an impairment in stability, as reflected by a larger variance in COG amplitude and velocity of displacement in the anteroposterior and mediolateral planes than in individuals without these postural deviations.

A second purpose of our study was to describe sagittal joint center alignments and their relationship to the whole-body COG in subjects with vestibular hypofunction and in subjects without impairments (controls) during quiet standing. Based on the literature, we hypothesized that subjects with vestibular hypofunction will have a more posteriorly tilted pelvis and extended trunk compared with control subjects. We also investigated postural changes that occur during standing with feet together and eyes closed, a more challenging equilibrium condition than standing with feet apart and eyes open. Our third hypothesis was that subjects with vestibular hypofunction will demonstrate an anterior COG shift when standing with feet together and eyes closed as compared with standing with feet apart and eyes open. This hypothesis was based on our second hypothesis (subjects with vestibular hypofunction will have an extended trunk). If these individuals are already positioned near the posterior limit of stability, then it seems logical that they would have to move anteriorly. When joints such as the knee and trunk are extended, there is only one direction in which it is possible to sway (eg, anteriorly). The body cannot sway posteriorly unless the feet come off the ground. If the knee and trunk joints are in more flexion (which shifts the COG anteriorly), then it is possible to move in more degrees of freedom (eg, anteriorly and posteriorly). An anterior COG shift, therefore, would enable individuals to adjust their body sway by moving with more degrees of freedom in this more challenging standing position.

Method

Subjects

The sample consisted of 27 subjects with vestibular hypofunction and 26 volunteers who served as a control group. Both groups of subjects were of similar age and sex. The control subjects reported no orthopedic or neurological impairments that affect standing stability or posture and had no impairments based on a brief physical examination by the tester (eg, gait analysis, balancing on one foot, analysis of the transfer from a sitting position to a standing position). Testers were physicians or physical therapists. All control subjects were community ambulators without assistive devices. Subjects with UVH (n = 17) and BVH (n = 10) had no other neuromusculoskeletal impairments that could affect standing stability. The mean time since establishment of the diagnosis was 48 months, with a range of 2 months to 56 years. Of the 17 subjects with UVH, 9 subjects had right UVH and 8 subjects had left UVH. All subjects with vestibular hypofunction had abnormal vestibular function tests and clinical examinations, as evaluated by an otoneurologist. They had reduced vestibular function, not distorted vestibular function as would be seen in subjects with benign paroxysmal positional vertigo or Meniere disease. All subjects with vestibular hypofunction were capable of ambulating without an assistive device during the testing procedure.

Subjects with BVH had vestibulo-ocular reflex gains on computerized sinusoidal vertical-axis rotation (SVAR) of at least 2.5 standard deviations below normal and bilaterally decreased caloric responses.[19,20] During the SVAR test, the subject sits in the dark on a chair connected to a motorized platform that rotates left to right at frequencies of 0.01 to 1 Hz.[19,21] The subjects' open-eye-induced movements are recorded using direct current electro-oculography.[19,22] The gain is the ratio of slow-phase eye movement velocity to stimulus (chair) velocity.[19] Caloric testing involves infusing the ear with warm and cold water with a closed-loop irrigation, inducing a burst of nystagmus lasting 1 to 3 minutes. Ice water is infused into the external auditory canal if the warm and cold stimulus does not induce a response.[19] Asymmetries in the peak slow-phase velocity of the nystagmus suggest the possibility of a unilateral lesion.[21] Subjects with UVH had unilateral damage and at least one of the following: 30% unilaterally reduced caloric response, confirmatory abnormalities on rotational testing (mildly decreased low-frequency gains, asymmetrical rotation-induced nystagmus, or increased phase leads), or positional nystagmus while lying with the involved ear down.[20] Phase is the time-relationship between head and eye velocity.[19] Table 1 presents the subjects' demographic data.

Table 1. Descriptive Statistics for Subjects Without Impairment (n = 26) and Subjects With Vestibular Hypofunction (n = 27)

Subjects Without Impairment

Sex(a)    Age (y)   Height(m)   Weight(kg)

M         22.2      1.80          56.82
M         25.2      1.80          72.73
M         27.7      1.78          76.36
M         29.1      1.73          65.91
M         30.1      1.71          78.18
M         31.5      1.80          88.64
M         47.5      1.80         102.27
M         56.2      1.80          97.73
M         66.3      1.80          79.55
M         68.8      1.68          71.36
M         71.1      1.68          75.00
M         71.6      1.79          78.18
M         72.9      1.80          84.09
M         77.8      1.75          78.18
F         20.2      1.68          61.36
F         26.4      1.62          51.36
F         28.2      1.63          52.27
F         28.5      1.52          50.91
F         29.7      1.55          53.64
F         31.1      1.67          55.91
F         71.2      1.75          87.73
F         74.6      1.70          73.64
F         76.2      1.64          60.45
F         81.2      1.55          58.64
F         84.4      1.52          50.00
F         88.0      1.55          43.18

14=M
12=F
[bar] X   49.9      1.69          68.9
SD        23.9      0.10          16.3
Minimum   20.2      1.52          43.2
Maximum   88.0      1.80         225.0

Subjects With Vestibular Hypofunction

Group(b)   Sex       Age (y)   Height(m)   Weight(kg)

1          M         46.7      1.78         72.73
3          M         48.0      1.79        147.73
3          M         48.5      1.72         84.09
3          M         50.3      1.72         95.45
1          M         54.1      1.55         60.91
2          M         56.9      1.78         99.55
3          M         66.4      1.66         69.09
3          M         66.8      1.72         72.73
3          M         67.8      1.75         81.82
1          M         68.4      1.74         90.91
3          M         69.7      1.75         75.00
3          M         78.2      1.80         79.55
3          M         81.7      1.78         77.27
3          F         25.5      1.57         68.18
2          F         27.7      1.55         47.73
1          F         33.5      1.65         61.36
3          F         43.9      1.65         58.18
3          F         44.3      1.63         81.82
3          F         47.4      1.68        125.00
1          F         60.2      1.57         72.73
3          F         63.3      1.55         52.73
3          F         65.1      1.61         75.91
3          F         76.5      1.52         56.82
1          F         77.2      1.63         88.18
2          F         77.4      1.62         59.09
1          F         79.5      1.63         50.91
3          F         80.9      1.57         70.45
           13=M
           14=F
           [bar] X   59.5      1.7           76.9
           SD        16.4      0.1           22.0
           Minimum   25.5      1.5           47.7
           Maximum   81.7      1.8          147.7

(a) M = male, F = female.

(b) 1 = bilateral vestibular hypofunction with sinusoidal vertical axis rotation at 0.05 Hz [is less than or equal to] 0.1 (eg, nearly zero vestibular function), 2 = bilateral vestibular hypofunction with sinusoidal vertical axis rotation at 0.05 Hz [is greater than or equal to] 0.1, 3 = unilateral vestibular hypofunction.

Instrumentation

We collected the postural and stability data at the Massachusetts General Hospital Biomotion Laboratory (Boston, Mass). A 4-optoelectric camera Selspot II/TRACK (Telemetered Rapid Automatic Computerized Kinematic software) full-body kinematic data acquisition system([dagger]) was used to collect the kinematic postural data. Plastic rigid arrays secured to 11 rigid body segments (head, trunk, pelvis, thighs, shank, feet, and upper arms) held the infrared light-emitting diodes (LEDs) (Fig. 1). An infrared detector within each camera tracked the 60 LEDs. Conversion from camera to array position and orientation and then to body-segment position (segment linear data) and orientation (joint angular data) were implemented using methods previously described in the literature.[23,24] The body-segment linear data consisted of the sagittal room-referenced joint center positions of the ankle, knee, hip, shoulder, back, and neck. The joint angular data consisted of the room-referenced sagittal and coronal positions of the trunk and pelvis (eg, pelvic anteroposterior tilt) and the sagittal positions of the head and knee. Thus, if the trunk or pelvis were perpendicular to the floor and parallel to the vertical line of gravity, it would be in zero or neutral flexion or extension. Our system resolution is [is less than] 1 mm for linear displacement and [is less than] 1 degree for angular displacement.[23] Using this kinematic data acquisition system to measure upper-body angular kinematics, Krebs et al[25] found high between-trial reliability for these measures within subjects without vestibular impairments (trial-to-trial Pearson r [is greater than or equal to] 88). Each body segment was modeled with 6 degrees of freedom (3 rotations and 3 translations). Ten unconstrained joints, therefore, connected the 11 body segments.

[Figure 1 ILLUSTRATION OMITTED]

Subject Procedures

Barefoot subjects stood in two positions: (1) eyes open, feet apart but parallel and with the midheels 30 cm apart, and (2) eyes closed, feet together (feet parallel and [is less than]1 cm apart). The subjects stood on two Kistler force plates,([double dagger]) which measured vertical ground reaction forces (expressed as a percentage of body weight) from each lower extremity to assess the amount of weight bearing on each lower extremity in quiet standing. Body forces were normalized by expressing them as a percentage of total body mass calculated from a multisegment model. The subjects maintained each standing posture for 17 seconds. They stood about 10 seconds before 7 seconds of data were collected. Subjects stood with their arms folded across their chest, grasping the elbows, with their feet pointed forward. The tester instructed each subject to "Look straight ahead; stand as still as possible." Two trials of each standing position were collected, and data from the first (most naive) trial were analyzed. One chair rise trial was also performed. These data were used in conjunction with the standing data for the determination of the knee and hip joint center locations.[23] During this trial, the subject transferred from a sitting position to a standing position. The chair height was equal to the subject's knee height (distance from floor to medial tibial plateau). The subject's feet were 10 cm apart, and the greater trochanters were 4 cm behind the edge of the chair during chair rise.

Procedure for hint Center Locations

Two methods were used to determine joint center locations. In the first method, two hand-held TRACK pointers containing an array of LEDs were carefully aligned by the tester, reflecting the segment orientation in all 3 planes while the tip of the pointer contacted specific anatomical landmarks (Fig. 1). The pointer data were then used to calculate the segment's origin and transformation matrix relative to the segment's fixed array. These quiet-standing data alone determined the back, shoulder, ankle, and neck joints. The anatomical landmarks were the peaks of the iliac crests, acromial processes, ankle joint mediolateral centers and second metatarsal heads, and mastoid process, respectively. The midpoint of the line connecting the two iliac crests was the back joint that approximated L4-5. The neck joint was one neck radius medial to the pointer point location, as defined by the pointer array Z-axis. The neck joint approximated the atlanto-occipital joint.

The second method used average axes of rotation defined by kinematic and standing data for the knees and hips. Riley et al[26] found this method to have more accuracy for locating these lower-limb joint centers than the pointing trial alone. The pointer data and the rotations during the chair rise trial estimated the hip and knee axes of rotation. For the hip, the pointing plane intersected the midpoint of a line extending between the pubic symphysis and the anterior superior iliac spines. The pointer plane passed through the center of the knee joint using the tibial tubercle to define the midline plane. Each axis was an average screw axis, including rotation and translation.[24] The joint centers were the intersection of the axes of rotation and the joint midsagittal plane.[24]

Procedure for Whole-Body COG Location

The mass and center of mass of each body segment were computed on a subject-specific basis using anatomical measurements (length and diameter) and regression equations.[27] The summation of body-segment masses and their positions and orientations in space defined the location of the COG. Our system has been shown to be accurate within 1 cm for the estimation of COG.[24]

Data Collection

Three-dimensional computer graphics assisted in organizing, viewing, and analyzing the data using Superplot software, created at the Massachusetts General Hospital Biomotion Laboratory (Boston, Mass) in PV-WAVE([sections]) (Fig. 2). The Superplot software displays an 11-segment android kinematic model (Fig. 3). In the 3-dimensional global coordinate frame, we obtained the sagittal joint center positions for the ankle, knee, hip, shoulder, back, and neck; the sagittal and coronal positions for the trunk and pelvis; and sagittal positions for the head and knee. The system calculated a mean for each 7-second standing position. Kinematic and kinetic data were rejected for subjects who took a step, as determined by viewing the android and force-plate data. These data were excluded for one subject with vestibular hypofunction who took a step during standing with feet together. Data for another subject with vestibular hypofunction were excluded from the analysis of vertical ground reaction force data because of a problem with force-plate data collection during standing with feet together and eyes closed. One subject with vestibular hypofunction was excluded from postural data analysis involving the shoulder joint sagittal position secondary to a problem with data collection with the right arm.

[Figure 2-3 ILLUSTRATION OMITTED]

Data Analysis

We used the ankle joint center as the reference point for construction of the "virtual" computer-generated plumb-line anteroposterior position and subtracted each sagittal joint center position from the sagittal ankle joint position (eg, ankle-knee) (Tab. 2). We also subtracted all sagittal joint center positions from the sagittal COG position (eg, COG-hip) (Tab. 2). A combined phase-plane stability variable, [Sigma.sub.r] (root mean square variance of the COG anteroposterior and mediolateral displacement and COG anteroposterior and mediolateral velocity), obtained from phase-plane analysis measured standing stability (Fig. 4).[5] The SAS statistical software package([parallel]) generated descriptive and correlational (Pearson product-moment correlation coefficient) statistics for the phase-plane stability variable and all postural variables listed in Table 2. The correlation between the phase-plane stability variable and the two postural variables (COG to hip and COG to neck) was used to determine whether any subjects whose neck joint was interior to the whole-body COG and whose hip joint was posterior to the COG would display a stability impairment.

[TABULAR DATA 2 NOT REPRODUCIBLE IN ASCII]

A repeated-measures multivariate analysis of variance (MANOVA) was used to determine whether postural and stability differences existed across subjects with vestibular hypofunction and subjects without impairment in the two standing positions (eyes open and eyes closed). The dependent variables used in this analysis were the following variables in the two standing positions: (1) sagittal angles of the pelvis, trunk, head, and knee, (2) coronal angles of the trunk and pelvis, (3) sagittal distances from two variables (ankle to neck, ankle to shoulder, ankle to back, ankle to hip, ankle to knee, ankle to COG,COG to neck, COG to shoulder, COG to back, COG to hip, and COG to knee), (4) phase-plane stability, and (5) left and right vertical ground reaction forces. The independent variable was group.

Pair-wise post hoc simple contrasts were used to compare the sagittal and coronal angles and the sagittal alignments shown in Table 2 between the two standing positions (feet apart and eyes open versus feet together and eyes closed). Pair-wise post hoc simple contrasts compared specific postural differences (the sagittal and coronal angles and the sagittal alignments in Tab. 2) across the subjects without impairment and the subjects with vestibular hypofunction to determine whether the subjects with vestibular hypofunction had a more posteriorly tilted pelvis and extended trunk than did the subjects without impairment. The pair-wise contrast of the variable "ankle to COG" in the two standing positions was used to determine whether subjects with vestibular hypofunction would demonstrate an anterior COG shift when standing with feet together and eyes closed as compared with standing with feet apart and eyes open.

A pair-wise post hoc simple contrast compared the phase-plane stability variable between the two groups and between the two standing positions. Three subjects were excluded from the contrasts analyzing the stability variable. One subject without impairment was excluded from this analysis because he was a researcher in this field and, therefore, may have demonstrated bias in the postural adjustments he made. One subject with vestibular hypofunction took a step during standing with feet together and, therefore, the phase-plane data were not valid throughout this position. With one subject with vestibular hypofunction, the head array was not visible for the entire 7 seconds of data collection, which prevented the processing of the whole-body COG phase-plane data.

We also used pair-wise post hoc simple contrasts to determine whether a difference existed in the mean vertical ground reaction force distribution between the left and right lower extremities and between the subjects with vestibular hypofunction and the subjects without impairment. This contrast was performed twice using the data from subjects standing with feet apart as well as standing with feet together. Furthermore, a pair-wise post hoc simple contrast was used to determine whether the left vertical ground reaction force distribution was different from the right vertical ground reaction force distribution. This contrast comparing left and right vertical ground reaction force distributions was analyzed for the subjects with vestibular hypofunction, for the subjects without impairment, and for all subjects in both standing positions.

A repeated-measures MANOVA was used to determine whether differences in posture and stability existed between the two standing positions (feet together and eyes closed versus feet apart and eyes open). The dependent variables were the sagittal angles of the pelvis, trunk, head, and knee; the coronal angles of the pelvis and trunk; the sagittal alignments stated in Table 2; phase-plane stability; and the left and right vertical ground reaction forces. The independent variable was feet position. We used alpha = .05 as the level of significance. We chose r [is greater than] .60 as the level of correlation that was clinically meaningful.

Results

Correlation of Posture and Stability

There was a low correlation between posture (the sagittal and coronal angles and sagittal alignments shown in Tab. 2) and standing stability (the phase-plane variable) in both groups (r [is less than] .3). When standing stability was correlated with all postural variables in the 6 subjects with the highest phase-plane measurements (the 6 subjects who were least stable) during standing with feet together, we found one strong correlation: right trunk lateral flexion increased as the standing stability increased (r = .99).

Standing Stability Differences Between the Two Groups

Standing stability was less in the subjects with vestibular dysfunction (feet apart, [[Sigma].sub.r] = 0.98; feet together, [[Sigma].sub.r] = 2.36) than in the subjects without impairment (feet apart, [[Sigma].sub.r] = 0.73; feet together, [[Sigma].sub.r] = 1.14) (P [is less than] .05) (Tab. 2).

Postural Differences Between the Two Groups

The repeated-measures MANOVA showed no differences in the postural variables, the phase-plane stability variable, and the vertical ground reaction forces between the subjects with vestibular hypofunction and the subjects without impairment (P [is greater than] .5). No pair-wise contrasts of postural variables between the subjects with vestibular hypofunction and the subjects without impairment were significant (P [is greater than] .05). Even in the more challenging standing condition (feet together, eyes closed), posture did not differ between the two groups (Tab. 2). Table 3 presents the mean vertical ground reaction forces for the left and right lower extremities for the two groups in the two standing positions. A difference in the vertical ground reaction force distribution between the left and right lower extremities existed between the two groups during standing with feet apart (P [is less than] .05) (Tab. 3). The subjects with vestibular hypofunction had more weight on the left lower extremity during standing with feet apart (P [is less than] .05) (Tab. 3).

Table 3. Vertical Ground Reaction Force Data (Percentage of Body Weight) During Standing With Feet Apart and Standing With Feet Together in Subjects Without Impairment and Subjects With Vestibular Hypofunction

Vertical                                 Subjects With Vestibular
Ground             Total Sample                 Hypofunction
Reaction
Forces(a)   Feet Apart   Feet Together   Feet Apart   Feet Together

Left          50.05         49.22         52.41(b)      50.76
Right         49.49         49.17         47.46(b)      47.97

Vertical
Ground      Subjects Without Impairment
Reaction
Forces(a)   Feet Apart   Feet Together

Left          47.61         47.74
Right         51.60         50.33

(a) Mean percentage of body weight for a 7-second trial.

(b) Left significantly different from right (pair-wise post hoc simple contrast, P [is less than] .05). The small discrepancy in total body weight was due to the fact that body forces were normalized to total body mass calculated from a multisegment model.

Table 2 summarizes the means for the sagittal and coronal positions of the pelvis and trunk, the sagittal positions of the head and knee, and the combined anteroposterior and mediolateral stability variable ([[Sigma].sub.r]) for the two standing trials (feet together and feet apart) for the subjects with vestibular hypofunction, the subjects without impairment, and the total sample. During standing with feet apart, the total sample stood with an anterior pelvic tilt (1.62 [degrees]), extended trunk (1.87 [degrees]) and head (6.17 [degrees]), and flexed knees (5.04 [degrees]). The average subject in both groups also had a slight right laterally flexed trunk and pelvis of [is less than] 2 degrees. Table 2 also presents the mean distances in the sagittal plane from the ankle joint to the COG, back, neck, shoulder, hip, and knee joints and the mean distances from the COG to the neck, shoulder, knee, back, and hip joints. In all subjects, the knee, hip, shoulder, neck, and back joints were anterior to the ankle joint in the sagittal plane during standing with feet together and with feet apart (Tab. 2). The COG was anterior to the ankle, knee, back, and shoulder joints and posterior to the hip and neck joints in all subjects (Tab. 2).

Differences in Posture and Stability Between the Two Standing Positions

The repeated-measures MANOVA did not show differences in the postural variables, the phase-plane stability variable, and the vertical ground reaction forces between the two standing positions (feet together and eyes closed versus feet apart and eyes open) (P [is greater than] .2) across all subjects. The total sample of subjects increased their anterior pelvic tilt, trunk and head flexion, and knee flexion in standing with feet apart compared with standing with feet together (pair-wise contrast, P [is less than] .05) (Tab. 2). Coronal postures did not change between the standing positions (Tab. 2). Table 2 specifies pair-wise contrasts between the standing positions within the total sample, the subjects with vestibular hypofunction, and the subjects without impairment. Standing with feet together and eyes closed was a more challenging equilibrium condition than standing with feet apart and eyes open, as demonstrated by increases in the phase-plane stability in both groups (Tab. 2). In this more challenging position, the average subject in the total sample had an anterior shift in the COG from the ankle joint (3.96 cm) (Tab. 2). In all subjects standing with feet together, the COG also moved further anterior to the knee, hip, back, and shoulder (Tab. 2). In the group with vestibular hypofunction, standing with feet together, the COG moved farther anterior to the neck (Tab. 2).

Correlations Between the Postural Variables

Table 4 displays strong correlations among the postural variables. In all subjects, the majority of strong correlations existed during standing with feet together and eyes closed as compared with standing with feet apart and eyes open. One pair of postural variables (hip-COG distance and ankle-hip distance) was strongly correlated only during standing with feet apart (Tab. 4). Six pairs of postural variables were highly correlated only during standing with feet together (Tab. 4). The subjects with vestibular hypofunction demonstrated a greater number of correlations among postural variables than did the subjects without impairment.

Table 4. Correlations Among Postural Variables (r)(a) in Subjects Without Impairment and Subjects With Vestibular Hypofunction in the Two Standing Positions

Variable                                        Total Sample

Standing with feet together/sagittal plane
  Ankle-shoulder distance/trunk angle                .60
  Ankle-knee distance/knee angle                     .74
  Ankle-shoulder distance/ankle-neck distance        .69
  Trunk angle/neck-COG(b) distance
  Trunk angle/ankle-neck distance                    .65
  Neck-COG distance/ankle-neck distance             -.67
  Neck-COG distance/ankle-shoulder distance
  Ankle-COG distance/neck-COG distance               .62
  Ankle-COG distance/hip-COG distance               -.81
  Ankle-COG distance/trunk angle

Standing with feet apart/sagittal plane
  Hip-COG distance/ankle-hip distance
  Hip-COG distance/ankle-COG distance               -.65
  Neck-COG distance/ankle-neck distance             -.69
  Ankle-shoulder distance/ankle-neck distance        .67
  Trunk angle/ankle-shoulder distance

                                                Subjects With
                                                Vestibular
Variable                                        Hypofunction

Standing with feet together/sagittal plane
  Ankle-shoulder distance/trunk angle              .72
  Ankle-knee distance/knee angle                   .61
  Ankle-shoulder distance/ankle-neck distance      .74
  Trunk angle/neck-COG(b) distance                -.73
  Trunk angle/ankle-neck distance                  .65
  Neck-COG distance/ankle-neck distance           -.67
  Neck-COG distance/ankle-shoulder distance       -.66
  Ankle-COG distance/neck-COG distance             .76
  Ankle-COG distance/hip-COG distance             -.83
  Ankle-COG distance/trunk angle

Standing with feet apart/sagittal plane
  Hip-COG distance/ankle-hip distance
  Hip-COG distance/ankle-COG distance
  Neck-COG distance/ankle-neck distance           -.70
  Ankle-shoulder distance/ankle-neck distance      .83
  Trunk angle/ankle-shoulder distance              .62

                                                Subjects Without
Variable                                        Impairment

Standing with feet together/sagittal plane
  Ankle-shoulder distance/trunk angle
  Ankle-knee distance/knee angle                    .83
  Ankle-shoulder distance/ankle-neck distance       .65
  Trunk angle/neck-COG(b) distance
  Trunk angle/ankle-neck distance
  Neck-COG distance/ankle-neck distance            -.69
  Neck-COG distance/ankle-shoulder distance
  Ankle-COG distance/neck-COG distance
  Ankle-COG distance/hip-COG distance              -.77
  Ankle-COG distance/trunk angle                    .65

Standing with feet apart/sagittal plane
  Hip-COG distance/ankle-hip distance               .61
  Hip-COG distance/ankle-COG distance              -.79
  Neck-COG distance/ankle-neck distance            -.68
  Ankle-shoulder distance/ankle-neck distance
  Trunk angle/ankle-shoulder distance

(a) Clinically significant results only (r [is less than or equal to] 6).

(b) COG = center of gravity.

Discussion

Correlation of Posture and Stability

Posture and stability were not strongly correlated in either group of subjects. Body-segment alignment changes alter the whole-body COG location.[2] A stable individual is able to control the whole-body COG amplitude and velocity of displacement. Despite using a very sophisticated kinematic analysis system, however, we found a very low relationship between posture and stability in our subjects. The finding of one highly significant correlation, which was between the phase-plane stability variable and trunk lateral flexion in the 6 most unsteady subjects, demonstrates that further research is warranted. Treating minute postural deviations in patients who do not have musculoskeletal dysfunction is not warranted as a means of improving stability. Individuals with larger postural deviations may benefit from "postural correction" to improve stability. Larger postural differences may have a higher correlation to stability.

Posture in Subjects With Vestibular Hypofunction

The unsteady subjects with vestibular hypofunction did not have postural aberrations, which differs from observations by Horak and Shupert[15] that subjects with BVH may have a more forward head position compared with subjects without impairment. Our data also do not support Horak and Shupert's theory that subjects with BVH may align themselves in quiet standing near the posterior limit of stability.[15] In our study, both the subjects with vestibular hypofunction and the subjects without impairment demonstrated a slightly extended trunk and head during standing with feet apart and eyes open (Tab. 2). Although the average posture was not different between the two groups in our study, the subjects with vestibular hypofunction had higher ranges for the following joint angles in both standing positions: pelvic flexion and extension, head flexion and extension, and pelvic and trunk lateral side bending. One reason for the differences in the findings of the two studies may be related to the diagnoses of the subjects. We included subjects with BVH and UVH; Horak and Shupert[15] studied only subjects with BVH. The acuity of the hypofunction could also account for the differences in results. In our study, the subjects were not in the acute phase; their posture may have compensated over time, although their stability was still impaired. Nonetheless, our study included the largest number of subjects with vestibular hypofunction ever reported; thus, this group may be a more representative group than those in other studies.

The subjects in our study did not have acute vestibular hypofunction, yet they stood with more body weight shifted toward the left lower extremity during standing with feet apart. We did not find a difference in weight distribution between subjects with BVH and subjects with UVH. Each subject with vestibular dysfunction may compensate differently, and we cannot draw firm conclusions from this small sample. During standing with feet apart, however, the subjects with vestibular hypofunction had an unequal weight distribution compared with the subjects without impairment. These findings suggest that subjects without impairment do not have an asymmetric weight distribution, in contrast to the report of Nichols et al[28] that young adults without impairment shift their weight toward their left lower extremity. Our findings also differ from the report by Kirby and colleagues[29] that subjects without impairment shift their weight toward their right lower extremity while standing with their feet together.

Comparison of Postural Findings With Those of Previous Studies

Our data are similar to the data reported by other researchers, with the exception of Braune and Fischer,[7] who concluded that joint centers aligned perfectly in the sagittal plane among subjects standing with feet apart (Tab. 5). Our sample of postural data is the largest ever reported, using the most advanced data acquisition system. Many past studies used photographs to measure posture. It is difficult, however, to identify bony landmarks on photographs (Fig. 5). According to Kendall and McCreary's "ideal" alignment, the plumb line referenced from anterior to the ankle joint passes anterior to the knee, posterior to the hip, through the bodies of the cervical and lumbar vertebrae, and through the shoulder joint.[30] This statement was made without supporting data or published research. In our study, the plumb line drawn from the ankle joint center fell posterior to all joint centers (Tab. 2). We verified that the location of the whole-body COG while our subjects stood with feet apart agreed with the relevant body of literature (Tab. 6). The COG location in relationship to joint centers in our subjects was similar to that described by Kendall and McCreary.[30] They contended that when the whole-body COG is slightly anterior to the knee and posterior to the hip, these joints are stable.[30] It can be argued, therefore, that minimal muscular effort is necessary to maintain neutral alignment, secondary to ligamentous and muscular constraints,[30] but data to support this argument have not been collected.

[Figure 5 ILLUSTRATION OMITTED]

Table 5. Comparison of the Total Sample's Sagittal Joint Center Alignment Relative to the Ankle Joint With That of Individuals Without Impairment in Prior Studies

                    Sagittal Coordinates(a)

Investigators     Knee                     Hip

Braune and
  Fischer[7]      0                        0
Akerblom[9],(b)   0.8 (23)                 4.4 110)
Woodhull et
  al[8],(c)       3.8 (15)                 6.2 (15)
Barry-Greb[10]
Harrison et
  al[11],(c)
Present
  study(d)        4.24[+ or -] 2.14 (53)   5.42[+ or -] 2.86 (53)

                         Sagittal Coordinates(a)

Investigators     Shoulder                        Neck

Braune and
  Fischer[7]      0
Akerblom[9],(b)
Woodhull et
  al[8],(c)       3.8 (15)
Barry-Greb[10]    4.2 [+ or -] 2.0 (23)
Harrison et
  al[11],(c)      5.5 [+ or -] 2.4 (41)
Present
  study(d)        1.89 [+ or -] 3.01 (52)   4.84 [+ or -] 4.03 (53)

                  Sagittal Coordinates(a)

Investigators     Back

Braune and
  Fischer[7]
Akerblom[9],(b)
Woodhull et
  al[8],(c)
Barry-Greb[10]
Harrison et
  al[11],(c)
Present
  study(d)        3.28 [+ or -] 2.42 (53)

(a) Numbers are anteroposterior distances (in centimeters) relative to the ankle. Positive numbers indicate that the joint centers are anterior to the ankle. When available, standard deviations are given. N is in parentheses.

(b) Ankle joint center was "weight-bearing trapezium" found by radiography.

(c) Ankle joint center was 1 cm anterior to posterior border of lateral malleolus.

(d) Ankle joint center determined by pointing trial with feet apart, eyes open.

Table 6. Comparison of the Total Sample's Sagittal Joint Center Alignment in Relationship to the Whole-Body Center of Gravity (COG) Alignment With That of Individuals Without Impairment in Prior Studies

                            Whole-Body COG(a) in the
                           Sagittal Plane Relative to

Investigators               Ankle             Knee

Hellebrandt[12],(b)     5.0
Akerblom[9]             2.24 [+ or -]   1.6 [+ or -] 0.2(23)
                          0.26 (11)
Fox and Young[13],(c)   5.36 [+ or -]   Slightly anterior
                          1.46 (66)       to joint
                                          center (66)
Woodhull et al(8)       4.9 [+ or -]    1.1 [+ or -] 2.4(15)
                          1.3 (15)
Present study           4.34 [+ or -]   0.10 [+ or -] 3.59(53)
                          3.30 (53)

                            Whole-Body COG(a) in the
                           Sagittal Plane Relative to

Investigators               Hip                Shoulder

Hellebrandt[12],(b)
Akerblom[9]             -1.8 110)
Fox and Young[13],(c)                   In anterior portion
                                           of acromial
                                           process (66)
Woodhull et al(8)       -1.3 [+ or -]    1.2 [+ or -] 1.6 (15)
                           1.2 (15)
Present study           -1.08 [+ or -]   2.40 [+ or -] 3.73 (52)
                           3.56 (53)

                            Whole-Body COG(a) in the
                           Sagittal Plane Relative to

Investigators              Neck                 Back

Hellebrandt[12],(b)
Akerblom[9]
Fox and Young[13],(c)

Woodhull et al(8)
Present study           -0.49 [+ or -]   1.07 [+ or -] 3.61 (53)
                           4.35 (53)

(a) Numbers are anteroposterior distances (in centimeters) relative to the whole-body COG. A negative number denotes that the COG is posterior to the joint center. When available, standard deviations are given. N is in parentheses.

(b) Did not specify which part of lateral malleolus for ankle reference point.

(c) Posterior border of lateral malleolus was ankle reference point.

Our research contradicts Woodhull and colleagues' conclusion that the pelvis typically tilts posteriorly and the knees extend.[8] Our results support Levine and Whittle's finding that female subjects without impairment have an anterior pelvic tilt in quiet standing.[14] In our system, neutral pelvic tilt is described as zero degrees of pelvic flexion and extension in relationship to the floor. Although difficult to detect by looking at Figure 5, the subject without impairment shown in the figure is standing with 2 degrees of anterior pelvic tilt. We believe our instrumented postural data provide greater detail about joint angles during standing than is possible by visual observation. Thus, positions described as neutral[30] actually consist of a few degrees of flexion or extension, which are angles that are probably imperceptible by visual observation. Our postural data only slightly deviate from the "ideal" posture described by Kendall and McCreary.[30] Clinicians should be cautious not to automatically link postural impairments (or observed deviations from the "ideal" posture) with functional limitations because perfect alignment has never been reported, even for individuals without impairment.[7-11]

Standing With Feet Together and Eyes Closed: A More Challenging Equilibrium Condition

The large anterior COG shift that occurred when subjects changed from standing with their feet apart and eyes open to standing with their feet together and eyes closed may have allowed them to obtain more degrees of freedom of movement in the anterior and posterior directions. This greater freedom of movement thus allows individuals to adjust their balance in two directions while standing in the more challenging equilibrium condition, as compared with individuals who stand with their knees extended, allowing only one direction of movement (eg, flexion). For example, all subjects' knees were more flexed during standing with feet together, allowing flexion and extension movements of the knee to control balance. In subjects with vestibular hypofunction, the anterior COG shift came primarily from anterior rotations of the pelvis, trunk, and head (Tab. 2) and not from anterior translations of the neck, shoulder, and knee in relationship to the ankle. In these subjects, the hip and back actually shifted posteriorly (Tab. 2), closer to the ankle joint, which is what would be expected with a flexion moment of the upper torso. The subjects without impairment, however, had an anterior shift in the shoulder and neck in relationship to the ankle joint during standing with feet together (Tab. 2) that could have contributed to the anterior COG shift.

Postural Patterns Controlling Stability When Subjects Change From Standing With Feet Apart to Standing With Feet Together

Our results support Krebs and colleagues' contention that whole-body movement patterns control posture in persons with vestibular hypofunction and in persons without impairment during standing with feet together and eyes closed.[31] In our study, changes in posture at multiple body segments were used to control stability when subjects changed from standing with feet apart and eyes open to standing with feet together and eyes closed. These findings differ from Nashner and colleagues' report that subjects without impairment used an ankle strategy when standing with eyes closed.[32]

One reason that weight shift became more symmetrical in the subjects with vestibular hypofunction during standing with feet together may have been their attempt to deal with a more challenging position (Tab. 3). By narrowing the base of support, individuals may be more likely to sway and to take a step. Distributing body weight more evenly in this position may allow upright stance to be maintained without stepping. With a larger base of support (eg, during standing with feet apart), the medial and lateral limits of stability extend from the outer edge of one foot to the outer edge of the other foot. When the base of support is narrowed, the limits of stability decrease. If body weight is borne primarily on one lower extremity, the base of support becomes even smaller. Remaining stable during standing with feet together, therefore, may be easier with a more symmetrical distribution of body weight.

[Figure 4 ILLUSTRATION OMITTED]

Head/Trunk Movement as Subjects Change From Standing With Feet Apart to Standing With Feet Together

Changes in postural alignment at the trunk and head to stabilize the head against gravity did not exist in either group as subjects changed from standing with feet apart to standing with feet together. For example, the head did not extend when the trunk moved into more flexion (Tab. 2). In both groups, the pelvis, trunk, and head moved toward a flexed posture during standing with feet together and eyes closed (Tab. 2). Thus, the head was stabilized with respect to the trunk (head stabilization on trunk).[33] Because visual input was impaired during standing with eyes closed, visual feedback may be more useful for head stabilization to the trunk than has been reported.[33] Our findings are similar to those reported by Shupert et al[34] but differ from a later report by Shupert et at[35] that individuals whose trunk flexes compensate by extending the head to stabilize the head in relationship to gravity (head stabilization in space[33]). Shupert et al[35] contend that patients with vestibular loss do not coordinate the head and trunk movement because such patients predominantly use an ankle strategy.

Postural Patterns in the Two Groups

Certain postural patterns occurred in both groups. As we expected, when the sagittal distance between the ankle and shoulder increased, the neck and COG distance decreased (Tab. 4). These data support the conclusion that whole-body movement patterns occur in quiet standing to control posture.[31] As the sagittal distance between the ankle and shoulder increased, the distance between the ankle and neck increased in both groups during standing with feet together (Tab. 4). Furthermore, as the sagittal distance between the ankle and COG increased, the sagittal distance between the COG and hip decreased (Tab. 4). In this example, the COG moved anteriorly, closer to the hip. Taken collectively, these postural patterns demonstrated that the subjects adapted a "forward posture" pattern during standing with feet together and eyes closed.

Subjects with vestibular hypofunction demonstrated unique postural patterns as compared with the subjects without impairment. In both standing positions, they increased their trunk flexion as the sagittal distance between the shoulder and ankle and the neck and ankle increased (Tab. 4). This finding suggests that whole-body movement patterns are used to maintain stability in quiet standing. In addition, as trunk flexion and the distance between the ankle and shoulder increased, the distance between the neck and COG decreased (Tab. 4). This finding is likely due to the COG moving anteriorly with trunk flexion due to the relatively large trunk mass. The subjects with vestibular hypofunction, during standing with feet apart, also had positive cross-correlations between the distance from ankle to shoulder and the distance from ankle to neck and between the trunk angle and the distance from ankle to shoulder (Tab. 4). Because the subjects with vestibular hypofunction had greater phase-plane stability, their shoulders, neck, and trunk could have been swaying more, as compared with the subjects without impairment. In quiet standing with feet apart, the subjects with vestibular hypofunction stood with more weight on the left lower extremity. Compensation for this asymmetry may have caused increased body sway.

Our study supplied data to support the concept that quiet standing is not completely static.[4,5] As shown in Figures 2 and 6, body segments moved over a course of 7 seconds. The 7-second mean could possibly obscure a correlation between posture and balance. Within a given subject, some movements were directly correlated to COG displacement, whereas other movements were compensatory and inversely correlated to COG displacement (Fig. 6). Such stability strategies, therefore, may be obscured by pooling the data across subjects. A larger sample size is needed to increase the statistical power necessary to reveal whether a larger posture-stability relationship exists.

[Figure 6 ILLUSTRATION OMITTED]

Conclusions

Posture was minimally correlated with stability, and posture did not differ between the subjects with vestibular hypofunction and the subjects without impairment. An anterior COG shift occurred as subjects moved from standing with feet apart and eyes open to standing with feet together and eyes closed. The subjects with vestibular hypofunction had greater phase-plane stability than did the subjects without impairments in both standing positions, demonstrating that they had less stability in a standing position compared with the subjects without impairment.

We report the postural data from the largest sample of both subjects with vestibular hypofunction and subjects without impairment ever reported. The average individual in the total sample stood in quiet standing with a slight anterior pelvic tilt (1.62 [degrees]), minimally extended trunk (1.87 [degrees]) and head (6.17 [degrees]), and flexed knees (5.04 [degrees]). In the more challenging standing condition (feet together, eyes closed), all subjects increased their anterior pelvic tilt, trunk and head flexion, knee flexion, and anterior COG position. These data support the concept that segmentally coupled whole-body movement patterns control posture in both groups. Changes in postural alignment at the trunk and head (head stabilization in space) to stabilize the head against gravity were not demonstrated in either group as subjects changed from standing with feet apart and eyes open to standing with feet together and eyes closed.

Acknowledgments

We thank Rita Popat, PT, NCS, and the Massachusetts General Hospital Biomotion Laboratory staff for their computer assistance.

(*) Oxford Metrics Ltd, Oxford, England.

([dagger]) The Selspot component of the optoelectronic system is manufactured by Selective Electronics, Partille, Sweden. TRACK is the copyrighted trademark of a computer program developed at the Massachusetts Institute of Technology, Cambridge, Mass.

([double dagger]) Kistler Instruments, Winterthur, Switzerland.

([sections]) Visual Numerics Inc, 5775 Flatiron Pkwy, Suite 220, Boulder, CO 80301.

([parallel]) SAS Institute Inc, SAS Campus Dr, Cary, NC 27513.

References

[1] Lee WA. A control systems framework for understanding normal and abnormal posture. Am J Occup Ther. 1989;43:291-301.

[2] Riley PO. Modeling of the Biomechanics of Posture. Boston, Mass: Massachusetts Institute of Technology; 1987. Doctoral dissertation.

[3] Benda BJ, Riley PO, Krebs DE. Biomechanical relationship between center of gravity and center of pressure during standing. IEEE Trans Rehabil Eng. 1994;2(1):3-10.

[4] Shumway-Cook A, Woollacott MH. Motor Control: Theory and Practical Applications. Baltimore, Md: Williams & Wilkins; 1995.

[5] Riley PO, Benda BJ, Gill-Body KM, Krebs DE. Phase plane analysis of stability in quiet standing. J Rehabil Res Dev. 1995;32:227-235.

[6] Norkin CC, Levangie PK. Joint Structure and Function: A Comprehensive Analysis. Philadelphia, Pa: FA Davis Co; 1983.

[7] Braune W, Fischer O; Maquet PGJ, Furlong R, trans. On the Centre of Gravity of the Human Body: As Related to the Equipment of the German Infantry Soldier. New York, NY: Springer-Verlag; 1985.

[8] Woodhull AM, Maltrud K, Mello BL. Alignment of the human body in standing. Eur J Appl Physiol. 1985;54:109-115.

[9] Akerblom B. Standing and Sitting Posture, With Special Reference to the Construction of Chairs. Stockholm, Sweden: Nordiska Bokhandeln; 1948.

[10] Barry-Greb T. Measurement of Normal and Forward Head Posture. Lexington, Ky: University of Kentucky; 1990. Master's thesis.

[11] Harrison AL, Barry-Greb T, Wojtowicz G. Clinical measurement of head and shoulder posture variables. J Orthop Sports Phys Ther. 1996;23:353-361.

[12] Hellebrandt FA. Standing as a geotropic reflex. Am J Physiol (Lond). 1938;121:471-474.

[13] Fox MG, Young DG. Placement of the gravity line in anteroposterior standing posture. Res Q. 1954;25:277-285.

[14] Levine D, Whittle MW. The effects of pelvic movement on lumbar lordosis in the standing position. J Orthop Sports Phys Ther. 1996;24:130-135.

[15] Horak FB, Shupert C. Role of the vestibular system in postural control. In: Herdman S, ed. Vestibular Rehabilitation. Philadelphia, Pa: FA Davis Co; 1994:22-46.

[16] Takemori S, Ida M, Umezu H. Vestibular training after sudden loss of vestibular functions. ORL J Otorhinolaryngol Relat Spec. 1985;47:76-83.

[17] Black FO, Shupert CL, Peterka RJ, Nashner LM. Effects of unilateral loss of vestibular function on the vestibulo-ocular reflex and postural control. Ann Otol Rhinol Laryngol. 1989;98:884-889.

[18] Shumway-Cook A, Horak FB. Assessing the influence of sensory interaction on balance: suggestion from the field. Phys Ther. 1986;66:1548-1550.

[19] Krebs DE, Gill-Body KM, Riley PO, Parker SW. Double-blind, placebo-controlled trial of rehabilitation for bilateral vestibular hypofunction: preliminary report. Otolaryngol Head Neck Surg. 1993;109:735-741.

[20.] Evans MK, Krebs DE. Relationship of vestibulospinal and vestibulocular tests. Otolaryngol Head Neck Surg. In press.

[21] Parker SW. Vestibular evaluation: electronystagmography, rotational testing, and posturography. Clin Electroencephalogr. 1993;24:151-159.

[22] Herdman SJ, Sandusky AL, Hain TC, et al. Characteristics of postural stability in patients with aminoglycoside toxicity. J Vestib Res. 1994;4:71-80.

[23] Antonsson EK, Mann RW. Automatic 6-DOF kinematic trajectory acquisition and analysis. Journal of Dynamic Systems, Measurement, and Control. 1989;111:31-39.

[24] Riley PO, Mann RW, Hodge WA. Modeling of the biomechanics of 5 posture and balance. J Biomech. 1990;23:503-506.

[25] Krebs DE, Wong D, Jevsevar DS, et al. Trunk kinematics during locomotor activities. Phys Ther. 1992;72:505-514.

[26] Riley PO, Fijan RS, Hodge WA, Mann RW. Determination of joint centers for posture studies. In: Stein JL, ed. The Biomechanics of Normal and Prosthetic Gait. Fairfield, NJ: American Society of Mechanical Engineering; 1987;4:131-136.

[27] McConville JT, Churchill TD, Kaleps I, et al. Anthropometric Relationships of Body Segment Moments of Inertia. Dayton, Ohio: Wright-Patterson Air Force Base; 1980. Report No. AFAMRL-TR-80-119.

[28] Nichols DS, Glenn TM, Hutchinson KJ. Changes in the mean center of balance during balance testing in young adults. Phys Ther. 1995;75:699-706.

[29] Kirby RL, Price NA, MacLeod DA. The influence of foot position on standing balance. J Biomech. 1987;20:423-427.

[30] Kendall FP, McCreary EK. Muscles: Testing and Function. 3rd ed. Baltimore, Md: Williams & Wilkins; 1983.

[31] Krebs DE, Riley PO, Gill-Body KM. Quantitative assessment of posture and stability: Does hip strategy exist? In: Proceedings of the 12th International Congress of the World Confederation for Physical Therapy, June 25-30, 1995, Washington, DC. Alexandria, Va: American Physical Therapy Association; 1995:640. Abstract.

[32] Nashner LM, Black FO, Wall C. Adaptation to altered support and visual conditions during stance: patients with vestibular deficits. J Neurosci. 1982;2:536-544.

[33] Di Fabio RP, Emasithi A. Aging and the mechanisms underlying head and postural control during voluntary motion. Phys Ther. 1997;77:458-475.

[34] Shupert CL, Black FO, Horak FB, Nashner LM. Coordination of the head and body in response to support surface translations in normals and patients with bilaterally reduced vestibular function. In: Amblard B, Berthoz A, Clarac F, eds. Posture and Gait: Development, Adaptation, and Modulation--Proceedings of the 9th International Symposium on Postural and Gait Research, Marseilles, France, 29 May-1 June, 1998. Amsterdam, the Netherlands: Elsevier Science Publishers BV; 1988:281-289.

[35] Shupert CL, Horak FB, Black FO. The effect of peripheral disorders on head-trunk coordination during postural sway in humans. In: Berthoz A, Graf W, Vidal P, eds. The Head-Neck Sensory Motor System. New York, NY: Oxford University Press Inc; 1992:1.

CG Danis, PT, is a recent graduate of the postprofessional physical therapy, orthopedic-sports specialty, master's degree program at MGH Institute of Health Professions, Boston, Mass.

DE Krebs, PhD, PT, is Professor, MGH Institute of Health Professions, and Director, Massachusetts General Hospital Biomotion Laboratory, 101 Merrimac St, Boston, MA 02114-4719 (USA) ([e-mail]). Address all correspondence to Dr Krebs.

KM Gill-Body, PT, is Assistant Professor, MGH Institute of Health Professions.

S Sahrmann, PhD, PT, FAPTA, is Associate Professor of Physical Therapy and Neurology, Physical Therapy Program, Washington University School of Medicine, St Louis, Mo.

This study was approved by the MGH Institutional Review Board.

This study was supported in part by NIH grants RO1AG11255 and NIDRR H133G60045.