Sensitivity and specificity of platform posturography for identifying patients with vestibular dysfunction.

Deficits in vestibular function typically cause dizziness, loss of balance, and diminished functional independence.[1-3] Adaptation of the central nervous system (CNS) may compensate for impairments following acute peripheral vestibular deficit (PVD) in humans.[4-7] Many vestibular impairments, however, do not appear to resolve spontaneously and may require physical therapy to improve or accelerate the resolution of dizziness or postural instability.[8] Various procedures have been described in the literature for the rehabilitation of patients with vestibular deficits.[8-22] Although there is no uniformly accepted definition of vestibular rehabilitation, these procedures commonly involve repeated head movements to habituate dizziness,[9,11,15,18] positioning maneuvers to prevent the onset of vertigo,[10,13,19] exercises to improve eye-head coordination and the fixation of gaze,[20-22] and balance retraining therapy.

The wide variation in patient symptoms and the lack of knowledge about the natural history of some vestibular impairments complicate the selection of patients who might be good candidates for vestibular rehabilitation. There is a need for sensitive and specific methods for identifying patients with vestibular impairment. Peripheral vestibular function in humans cannot be measured directly,[23-25] and clinicians must rely on the identification of abnormal vestibular reflexes and the patients' report of symptoms to infer the presence of a vestibular deficit. There is frequently a lack of association between symptoms and the actual detection of vestibular abnormalities.[24,25] The primary diagnostic criteria for Meniere's disease, for example, are the symptom triad of tinnitus, fluctuating hearing loss, and repeated attacks of vertigo.[24] This disease is thought to be associated with endolymphatic hydrops.[26] Black,[24] however, observed that the diagnosis of "Meneire's disease" is still applied when no cause for the triad of symptoms can be identified.

Several investigators[6,20,23,25,27] have characterized vestibular dysfunction in terms of deficits in the vestibuloocular reflex (VOR) system and the vestibulospinal reflex (VSR) system. Tests of the VOR system--bithermal caloric irrigation, vertical axis rotation of the whole body, and electronystagmographic (ENG) evaluations of spontaneous and positional nystagmus--provide information about the symmetry of a vestibular lesion affecting the horizontal semicircular canals.[20,23,25] These vestibular tests, however, do not characterize the vestibular deficit in terms of the patient's functional status--the ability to stand and walk--because the patient is evaluated in "passive" positions (ie, secured in a chair) in which balance is not required. The physiologic response induced by VOR testing is limited primarily to horizontal canal function. In addition, abnormal results for some tests (ie, caloric irrigation) tend to persist after the patient compensates and the symptoms resolve.[25,28] Vestibuloocular reflex testing, therefore, may guide therapy programs for gaze stabilization and habituation of dizziness,[20,29] but VOR tests have limited value for identifying balance deficits related to vestibular dysfunction.30

Posturography is an approach to the assessment of vestibular dysfunction that utilizes a force platform and provides various measures that reflect postural stability, such as the amount of body sway. There are essentially two types of clinical posturography: static platform posturography and dynamic platform posturography. Static platform posturography involves stance or tandem stance on a fixed platform with eyes open or closed.[23,31] This procedure uses the Romberg tests, and the outcome is quantified with respect to changes in center-of-force sway amplitude, distance, or velocity. Dynamic platform posturography also utilizes Romberg's tests with a fixed platform, but includes test conditions in which the platform and visual environment are moved to reduce the subject's ability to use visual and somatosensory information for balance.[32-35] In addition, dynamic platform posturography incorporates sudden displacements of the platform to test the subject's response to balance perturbations.[32-35] Dynamic posturography estimates body sway angles from the vertical projections of the center of force.

Platform posturography was developed to clarify the nature of motor coordination problems associated with vestibular impairment[6,23,34,35] and other sensorimotor deficits.[2,34] In order to provide a meaningful description of balance deficits related to vestibular dysfunction, however, posturography would have to have an appropriate level of sensitivity and specificity. The American Physical Therapy Association's Standards for Tests and Measurements in Physical Therapy Practice defined sensitivity as "how well a diagnostic test identifies people who should have a positive finding."[36] Sensitivity is expressed as a ratio of the number of subjects with a true-positive response divided by the number of subjects who should have had a positive test indicating pathology. Specificity is determined by "how well a diagnostic test identifies people who should have a negative finding."[36] Specificity is expressed as a ratio of the number of subjects with a true-negative normal) response on the test divided by the number of subjects who should have had a negative response.

Measures of sensitivity and specificity rely on the identification of patients with a "true-positive" or "true-negative" response. Clinical signs and symptoms, audiometry, and a neurological examination, coupled with tests of VOR function, often serve as the "criterion standard" for determining the probability of a vestibular deficit.[37,38] The inability to confirm vestibular histopathology in vivo means that sensitivity and specificity can only be estimated based on the strength of the clinical diagnosis. Given the potential error associated with the clinical examination, some studies have addressed sensitivity and specificity by focusing on patients with symptoms that had a high probability of association with vestibular dysfunction.[23,37] Even so, it is not clear whether the results of dynamic posturography provide enough specificity to discriminate among unimpaired subjects, patients with vestibular deficits, and those with neural lesions who do not have involvement of the vestibular system.[38-42] Reports of the level of sensitivity of posturography vary widely.[23,25,31,38-40] Asai et al,[25] for example, found that excessive sway, estimated during dynamic posturography, was associated with peripheral vestibular dysfunction in 40% of the patients with peripheral vestibular impairment. Black and Wall,[2-3] in contrast, reported that approximately 60% of the patients with PVDs could be identified with static posturography. The sensitivity of posturography also varies with the population studied. Voorhees, for example, reported a 45% level of sensitivity for identifying balance disorders related to peripheral vestibular dysfunction,[38] but a 90% level of sensitivity for detecting balance problems in patients with CNS disease.[39] Clarification of the factors influencing the sensitivity of posturography, therefore, is needed.

The lack of clear operational definitions for sensitivity and specificity in the literature related to posturography complicates the interpretation of results. Hamid et al,[40] for example, stated that the sensitivity of posturography was 95%, but this value actually corresponds to specificity as defined by the Standards.[36] stockwell[41] Stated that dynamic posturography was insensitive to peripheral vestibular disorders, but one of the tests in question was expected to be normal in this population. Some of Stockwell's data, therefore, would be more appropriately analyzed to describe specificity. Black and wall[23] noted that static posturography did not have adequate specificity to distinguish among patients with different types of PVDs, but the number of patients that should have had a negative response in each diagnostic category were not identified. The literature, therefore, may require a reinterpretation of results based on standard definitions of sensitivity and specificity.[36]

Discrepancies in the levels of sensitivity and specificity among studies may also be due to a variation in the definition of "abnormal balance." It is clear that there is no universally accepted description of "abnormal balance" that directly relates to vestibular dysfunction. In order to determine whether balance performance was impaired by vestibular deficits, Keim[42] utilized a composite posturography score--a weighted average of the test conditions--that was compared with the 95% confidence interval of a normative database. Vestibular dysfunction in other studies[25,38,39] was defined as abnormal posturography in any single test condition.

Some investigators have identified patterns of abnormal balance behavior that supposedly "detect" vestibular deficits.[23,40] The classification schemes for identifying vestibular deficits with dynamic posturography were not consistent, even when similar instrumentation(*) was used to evaluate equilibrium.[3,40,43,44] Shepard et al[3] and Hamid et al[40] described six categories for classifying patients with balance and vestibular dysfunctions, whereas others[43,44] described four categories. In studies in which the number of categories was the same, the criteria for describing the categories differed.[43,44] In some studies,[40,43,44] VOR testing was used in conjunction with posturography to define each category, whereas other authors[3,41] used only the posturography test results. In some cases, confirmation of a vestibular deficit could only occur when both the VOR and posturography showed abnormalities.[43,44] The rationale for this approach is ambiguous, because the pathologies measured by VOR tests (horizontal canal function) and posturography (vertical canal and otolith function) are thought to be different.[23,24,45] In addition, operational definitions of abnormal balance behavior were sometimes absent in the literature. Parker,[46] for example, found that posturography was "abnormal" in 25% of 536 patients selected for a retrospective review, but the definition of "abnormal posturography" was not stated. The purpose of this clinical perspective is to describe the sensitivity and specificity of dynamic and static platform posturography for detecting vestibular impairment. To place this topic in perspective, a review of the theoretical basis for evaluating vestibular dysfunction with dynamic platform posturography is presented. The vestibular deficit "patterns" identified with dynamic and static posturography are also reviewed.

Theoretical Basis for Evaluating Vestibular Dysfunction With Dynamic Platform Posturography

The rostral projections of the vestibular nerve eventually influence the ocular muscles via the medial longitudinal fasciculus, whereas the caudal projections of the vestibular nerve innervate the trunk and lower limb muscles via the vestibulospinal tract.[47] Early studies by Nashner[48,49] were designed to evaluate the "descending" influence of the vestibular system. Nashner examined the response of unimpaired subjects with eyes closed during horizontal perturbations of stance on a "sway-referenced" platform. Platform "sway-referencing" was achieved by a servo-controlled mechanism that rotated the platform in proportion to body sway motions. This technique essentially maintained a quasi-static ankle joint position during the response to perturbation so that the influence of ankle and foot proprioception would be minimized. The goal of these experiments was to evaluate the compensatory response to balance perturbation when the subjects were required to rely primarily on vestibular inputs. The threshold for detection of body sway was defined as the time from initiation of platform motion to the time that ankle torque changed from its baseline value. Nashner[48] found a consistent relationship between the induced sway rate and the latency of sway detection. He concluded that vestibular feedback was the primary sensory modality contributing to the detection of body sway, because visual cues were absent and somatosensory information from the feet and ankles was altered during the test conditions.

To provide an indication of the relative participation of linear and angular acceleration receptors in the control of body sway, Nashner[48,49] developed an analytical model based on the experimental data and known physiologic behavior of the semicircular canals and utricular otoliths. The function of the canals (detection of angular acceleration) and the function of the otoliths (detection of linear acceleration) were separated for analytical description. The model for the pitch-axis semicircular canals fit closely with observed detection of sway over the entire range of induced sway rates. The semicircular canals on the pitch axis, according to the model, had no static sensitivity to orientation with respect to gravity, but could sense angular acceleration at a lower threshold compared with the horizontal canals.[48] The model for linear motion detection showed that the otoliths contributed little to sway compensation because the induced sway was too rapid (ie, beyond the frequency response of these receptors). It should be noted that a more recent theory suggests that head orientation (tilt) information can be obtained from tonic otolith afferent signals.[50]

There were two hypotheses generated from Nashner's model[48,49] that are relevant to the current discussion: (1) the sensory organization hypothesis and (2) the vestibulo-motor-influence hypothesis.

Sensory Organization Hypothesis

One implication of the model proposed by Nashner[48,49] was that disruption of the gravitational reference (low-frequency otolith feedback) would create postural instability if visual and somatosensory inputs were not available as redundant systems to control balance. The redundancy of sensory inputs controlling balance were manipulated experimentally to test this hypothesis. A sway-referenced visual enclosure was used in conjunction with a sway-referenced platform to distort visual inputs and lower-extremity somatosensory inputs, respectively.[32,33] Black et al[32] hypothesized that disruption of vestibular function could be compensated by the use of vision and lower-extremity proprioception only in conditions in which the foot support was fixed with respect to earth horizontal and the surrounding visual surface was fixed with respect to earth vertical. Black et al[32] studied three categories of patients with vestibular disorders: a patient with complete loss of vestibular function (n=1), patients with unilateral vestibular deficit (n=5), and patients with Meniere's disease in remission (n=6). Subjects were classified into categories reflecting the severity of the vestibular deficit on the basis of VOR tests and the amount of sway on a fixed surface during Romberg and tandem Romberg tests, Two unimpaired subjects were used as "agematched" controls. The results were analyzed descriptively.

There were no apparent differences in the amount of body sway for unimpaired subjects and patients with vestibular deficits during stance on a fixed platform with eyes open, eyes closed, or with a sway-referenced visual surround.[32] The amount of body sway during distortion of visual and somatosensory inputs from the feet and ankles (sway-referenced platform and visual enclosure), in contrast, appeared to be linearly related to the severity of the vestibular deficit, thereby confirming the sensory organization hypothesis.

A battery of six test conditions have been referred to as the sensory organization test (SOT) for balance (Tab. 1).[6,34,35] The linear relationship between the severity of the vestibular deficit and sway amplitude during posturography provided support for concurrent validity of the SOT as a measure of vestibular dysfunction. Black et al[32] concluded that vestibular input provides a fixed gravitational reference that is normally used to evaluate discongruent visual and somatosensory information. When visual and somatosensory information are distorted, patients with vestibular deficits have greater difficulty maintaining stance compared with unimpaired subjects.

Table 1. Posturography sensory organization
Test (SOT) Conditions
Condition
(C)   Platform          Vision
C1    Fixed             Eyes open
C2    Fixed             Eyes closed or
                        blindfold
C3    Fixed             Sway-referenced
                        enclosure
C4    Sway-referenced   Eyes open
C5    Sway-referenced   Eyes closed or
                        blindfold
C6    Sway-referenced   Sway-referenced
                        enclosure

Vestibulo-motor-influence Hypothesis

A second implication of Nashner's model[48,49] was that the pitch-axis semicircular canals provided feedback that would modulate the VSR during induced body sway. This hypothesis was tested by measuring the electromyographic (EMG) response of lower-extremity muscles during unexpected displacements of the platform. The loss of balance in patients with vestibular deficits could be due to abnormal reflex responses in the lower extremities during conditions requiring the use of vestibular inputs to maintain balance (sway-referenced platform with eyes closed). Black et al[32] found that the EMG latencies of ankle muscles were within the normal range (about 200 milliseconds) for subjects with vestibular deficits even when perturbation conditions were expected to require the greatest use of vestibular inputs to maintain stance. The vestibulo-motor-influence hypothesis, therefore, was not confirmed. The results did not support a direct influence of the vestibular system on the reflex motor patterns in the lower extremity during externally imposed balance perturbations. other authors have also reported that externally imposed balance perturbations did not significantly involve vestibulospinal mechanisms. Horstmann and Dietz[51] and Dietz et al[52] attempted to isolate vestibular function by stimulating the vestibular end-organs with sudden head movements. The head perturbation device moved the upper trunk while maintaining a constant position of the center of force during standing.[51] The magnitude of the lower-extremity EMG response was small compared with the response of ankle muscles to platform displacements or perturbations of gait.[51,52] The contribution of VSRs to postural stabilization following an imposed balance disturbance, therefore, was considered to play an insignificant role in maintaining stance under these conditions.[51-53]

Dynamic Posturography. Tests of Balance (EquiTest)

Dynamic posturography (EquiTest) involves two tests of balance function: (1) the SOT (Tab. 1) and (2) a motor test that consists of sudden displacements of the platform to induce body sway.[34,35] The SOT evaluates postural stability with systematic changes in the type of visual and somatosensory information available to the patient to maintain stance (Tab. 1). The amount of sway is estimated from maximum peak-to-peak displacement of the center of force during the test interval.[54] An equilibrium score represents the estimated peak-to-peak sway normalized to a theoretical maximum body sway. Equilibrium scores beyond the 95% confidence interval of age-specific normative data are considered abnormal.[40.43,44,54,55]

A "motor" test of dynamic posturography was developed to evaluate the response to sudden force platform translation or rotation about the ankle joint. Measures of performance on this test include the latency and symmetry of sway-correcting torque (not EMG) applied to the platform.[38,39] Latency measurements are calculated as the time between an externally induced platform displacement and the patient's initiation of corrective floor reaction forces.[38,39,54] Mean latency measurements in unimpaired subjects under 60 years of age vary from 138 to 151 milliseconds,[54] and the detection of abnormality is also norm-based, with a confidence interval of approximately 95%. The asymmetry of torques during platform perturbation in unimpaired subjects can vary [+ or -]25%.39 Patients with PVDs typically have a response within normal limits for the latency and symmetry of lower-extremity activation during platform perturbation.[32,33,38,39] Central nervous system diseases that alter the long-loop reflex pathways, such as stroke[56] or multiple sclerosis,[57] however, may alter the symmetry or delay the response to balance perturbation.

Several authors[58,59] have described a test of postural control during pressure changes applied to the external auditory canal. Selected conditions of dynamic posturography SOT are used during the pressure application. The purpose of this "platform pressure test" is to identify patients with perilymphatic fistula. I did not include a discussion of the sensitivity and specificity of this test in this review because of the ambiguity associated with the diagnosis of this condition.[58]

Classification of Vestibular Dysfunction Based on Dynamic Platform Posturography

Hamid et al[40] conducted a retrospective review of posturography SOT from a database of approximately 3,000 patients and identified six balance deficit patterns that were based on abnormal equilibrium scores (Tab. 2). The interpretation of these patterns was dependent on supplemental information about the VOR gain obtained from vertical-axis rotation tests. The purpose of the VOR is to maintain the visual image on the retina when the head moves. As the head moves in one direction, the eyes rotate in the opposite direction to partially "nullify" the effects of head motion. Vertical-axis rotation stimulates the horizontal semicircular canals, and the characteristics of induced eye motion (nystagmus) can then be measured with respect to the velocity of rotation. The VOR gain is the ratio of angular eye velocity to angular head velocity. Theoretically, perfect ocular stabilization is reflected by a gain of - 1.0, but vestibular compensation of head motion is not complete in unimpaired subjects and normal gains are on the order of -0.8 (see Stockwell[41] for a detailed description of the VOR gain and other variables measured during rotation testing). A disruption of the VOR gain--either depressed or hyperactive--may indicate vestibular pathology,[23] and Hamid et al[40] recommended that this information be used to clarify the results of posturography (Tab. 2).

[TABULAR DATA 2 OMITTED]

The "vestibular loss pattern" described by Hamid et al,[40] for example, was characterized by abnormal body sway during stance on a sway-referenced platform with a blindfold (C5) and in a sway-referenced visual enclosure (C6). The "vestibular deficit pattern" was the same as the vestibular loss pattern with regard to posturography. The VOR gain, however, for the vestibular loss pattern approached zero, whereas the gain for the vestibular deficit pattern could be normal or hyperactive (Tab. 2).

Shumway-cook et al[43] and Horak et al[44] also classified patients with vestibular impairment according to the results of VOR tests and the posturography SOT (Tabs. 3 and 4). The classification system presented by Horak et al[44] defined abnormal VOR as a gain, phase, or asymmetry beyond two standard deviations of a normal sample for vertical-axis rotation at 0.5 Hz or 0.2 Hz. Shumway-Cook et al[43] used the combined deviations in the gain, phase, and symmetry of the VOR as the criteria for VOR abnormality. The vestibular loss category defined by Horak et al[44] was limited to abnormal SOT findings in C5 and C6 coupled with an abnormal VOR (Tab. 3). Shumway-cook et al[43] labeled this category as "reduced peripheral vestibular function" (Tab. 4). Both studies[43,44] described a sensory organization ("interaction") deficit category and a mixed category (sensory organization deficit combined with vestibular loss). The VOR was normal in the sensory organization deficit category and abnormal in the "mixed" deficit (Tabs. 3 and 4). The SOT patterns noted by Horak et al44 for these categories, however, were different compared with the patterns described by Shumway-Cook et al.[43] Horak et al[44] stated that an abnomal equilibrium score in any one of the six sensory conditions was sufficient to identify patients with sensory organization deficits or mixed deficits (Tab. 3). The analogous categories described by Shumway-Cook et al[43] required that the SOT be abnormal on all sway-referenced conditions rather than a minimum of just one condition (Tab. 4). There was no provision in either of these classification systems[43,44] for a vestibular deficit pattern abnormal C5 and C6 with a normal VOR).

In a retrospective review of 98 patients, Shepard et al[3] identified six posturography patterns (Tab. 5). This classification scheme was independent of VOR test results. The SOT patterns defining the "vestibular dysfunction" category presented by Shepard et al3 (Tab. 5) were the same as the vestibular loss categories[40,44] (Tabs. 2 and 3) and the "reduced peripheral vestibular function" category[43] (Tab. 4).

The classification of "visual and vestibular dysfunction" by Shepard et al[3] (Tab. 5) is the same as the "somatosensory dependence" for balance described by Hamid et al[40] (Tab. 2). Smith-Wheelock et al stated that "a patient who falls when standing on a compliant surface regardless of visual conditions is demonstrating a visual-vestibular dysfunction pattern, meaning they are not using visual or vestibular cues to maintain stance."[15(p221)] That is, the patient classified with visual-vestibular dysfunction has a somatosensory dependence for maintaining balance. Using this reasoning, the "somatosensory and vestibular dysfunction" category (Tab. 5) is the same as the "visual dependence" category (Tab. 2).

Shepard et al[3] did not identify a "sensory deficit" category, in contrast to Hamid et al,[40] Shumway-Cook et al,[43] and Horak et al.[44] As an alternative, Shepard et al[3] provided a general classification for "severe" dysfunction in sensory integration for balance (Tab. 5). Overall, the classification system presented by Shepard et al[3] in contrast to the other classifications (Tabs. 2-4), has been shown to have prognostic significance and face validity.[2,3] Shepard and colleagues[2,3] demonstrated that the outcome of vestibular rehabilitation was dependent on certain pretherapy SOT patterns. In the following discussion, I selected this classification system (Tab. 5) to provide operational definitions for the SOT patterns that potentially identified vestibular impairment.

Sensitivity and Specificity of Dynamic Platform Posturography

Sensory Organization Test Sensitivity

The operational definitions of sensitivity and specificity presented in the Standards for Tests and Measurements in Physical Therapy Practice[36] served as a guide to evaluate the literature. In order for posturography SOT to have maximum sensitivity, all patients with vestibular deficits would theoretically show some type of vestibular deficit pattern (Tab. 5). The categories of vestibular deficit by lesion site include peripheral end-organ, CNS disease affecting the vestibular system (eg, disruption of vestibulocerebellar pathways), and mixed peripheral and central pathology. Specificity cannot be measured in these populations because all patients with vestibular deficits would be expected to have a positive SOT for vestibular pathology. A summary of the studies that addressed sensitivity and specificity of dynamic posturography is presented in Table 6.

[TABULAR DATA 6 OMITTED]

Asai et al[25] evaluated the sensitivity of dynamic platform posturography. These authors studied patients with PVDs, including Meniere's disease, sudden deafness with vertigo, and vestibular neuronitis. The basis for each diagnosis was not clearly stated, but it was assumed that diagnostic classifications were justified by clinical signs and symptoms. Abnormal posturography was defined as an equilibrium score two standard deviations below normative values for any test condition (Tab. 1). The greatest percentages of abnormal conditions were found in C5 and C6, but posturography SOT was found to have only a 400/o level of sensitivity in detecting vestibular dysfunction. In contrast, the sensitivity of a static body sway test (area of sway during the Romberg test on a fixed force platform) for detecting vestibular impairment was 63% (Tab. 6). A three-test battery consisting of ENG evaluations of positional, spontaneous, and rotary-chair induced nystagmus (more than one test of three abnormal) had a 77% sensitivity. The authors concluded that ENG evaluations of pendular rotation and spontaneous and positional nystagmus should be coupled with dynamic platform posturography to obtain an optimal classification of the patient's disorder, because the sensitivity of detecting vestibular dysfunction with the SOT and ENG combined (89%) was greater than with either test alone.

Hamid et al[40] evaluated the sensitivity of posturography, retrospectively, by comparing vestibular function tests and clinical diagnosis with the results of posturography. The vestibular examination conflicted with posturography in only 5% of the cases. The conflicting cases were false positive (ie, posturography indicated an abnormality where no clinical abnormality was found). Hamid et al[40] concluded that the sensitivity of posturography, therefore, was 95%. Closer inspection of the data (see Tab. 1 in Hamid et al[40]), however, revealed that SOT patterns implicated vestibular pathology in 42% of the patients with peripheral vestibular dysfunction and in 50% of the patients with CNS involvement (Tab. 6). A hip sway pattern was described, but could not be meaningfully included in the reinterpretation of data because the hip sway pattern was derived from horizontal shear forces and the operational definition of "hip sway" was not provided.

In a retrospective chaft review, Voorhees[38] studied 53 patients with CNS dysfunction and 112 patients with peripheral vestibular disorders. Vestibular deficits were identified on the basis of brain imaging, neurological examination, ENG, dynamic posturography, and audiometry. All patients included in the study apparently had findings that suggested vestibular involvement. Abnormal posturography was defined as abnormal sway in any condition of the SOT. Nearly half of the patients with CNS disease (42%) also had abnormal ENGs, but the number of patients with abnormalities in both the SOT and ENG was not clear. Voorhees[38] reported that the sensitivity of SOT for detecting peripheral vestibular and CNS impairment was 45% and 49%, respectively. Whether the patients with CNS dysfunction in this study had central vestibular deficits is not clear. The detection of central vestibular deficits with posturography is a problem, because patients with sensorimotor impairments of central origin, including the ascending and descending pathways, may have balance deficits that are also detected with posturography.

In a subsequent study,[39] more than 90% of the patients with CNS deficits (n=28) were shown to have an abnormal SOT in at least one test condition (Tab. 6). This level of sensitivity may be slightly inflated, however, because a small number of patients (fewer than 4) had a visual preference pattern abnormal C3 and C6; Tab. 5) that did not indicate vestibular involvement.

Keim[42] also conducted a retrospective study and used the composite score of the six SOT conditions as the index of abnormal balance function. The data reported indicated that the SOT detected balance abnormalities in 63% of the patients with peripheral dysfunction and in 100% of the patients with CNS-vestibular dysfunction (Tab. 6). The author concluded that dynamic posturography provided a general measure of balance impairment, but could not be used to determine the lesion site.

Stockwell[41] used the results of ENG during tests of visual fixation, visual tracking, optokinetic stimuli, caloric stimulation, and Hallpike maneuvers as the "criterion standard" for determining the presence of a vestibular deficit. He reviewed 93 cases of PVD and 19 cases that were classified as CNS dysfunction "usually with evidence of vestibular involvement."[41](p5o) An additional 174 patients were considered to have a normal vestibular examination (see "Sensory Organization Test Specificity" section). For peripheral disorders, the SOT detected only 14 cases out of 93 (15%). The sensitivity for detection of CNS deficits was 21% (4 cases out of 19).

Parker[46] conducted a retrospective review and found that dynamic posturography detected peripheral vestibular impairment in 70 of 212 subjects (33%) and CNS vestibular deficits in 12 of 30 subjects (40%) (Tab. 6). The criteria for identifying abnormal balance were not stated, and the component of posturography used to detect abnormalities (ie, the SOT or the motor test) was not indicated.

Goebel and Paige[28] found that the mean equilibrium score was not different for subjects with recent vertigo compared with subjects with no history of vertigo. In their study, posturography was not based on a norm-referenced comparison, and only two SOT conditions (C5 and C6; Tab. 1) were used. Goebel and Paige[28] attempted to address the sensitivity of posturography, but the results of posturography were averaged to compare performance across groups. This approach was not compatible with an analysis of sensitivity, because the posturography outcomes for individual subjects were not described and the ratio of the number of patients with true-positive (abnormal) tests to the number of patients who should have had a true-positive test was not reported.

Sensory Organization Test Specificity

In order for posturography to have maximum specificity, all patients with nonvestibular disease or unimpaired subjects would be expected to have normal balance or show balance impairments that do not emulate vestibular deficit patterns (Tab. 5). Five percent of the patients with a normal vestibular examination (29 out of 581) reviewed by Hamid et al[40] had abnormal posturography. The specificity, therefore, would be 95% (Tab. 6).

Keirn[42] noted that the SOT composite score was within normal limits in 34% of 32 patients with nonvestibular CNS disease (Tab. 6). This finding means that the specificity was 34%. of the 174 patients with a normal ENG examination reviewed by Stockweff,[46] the SOT was normal for 161 patients (93%) (Tab. 6).

Motor Test Sensitivity

In order for the posturography motor test to have maximum sensitivity, all patients with CNS disease that involves the long-loop reflex pathways would theoretically show a delayed response latency. Prolonged latencies would be expected regardless of central vestibular system involvement, because a lesion at any point on the reflex pathway (ie, the posterior columns) could delay the response to balance perturbation.[57] The sensitivity of the motor test, therefore, is not restricted to the detection of central vestibular impairment.

Voorhees[38,39] evaluated the latency and symmetry of the torque response to perturbation in a retrospective study. Twenty-three percent (n=12) of the patients with CNS-vestibular involvement showed elevated latencies on the motor test.38 The sensitivity of the motor test, therefore, was 23% (Tab. 6). When the SOT (balance deficit in any condition) and motor test results were combined, the detection rate for balance dysfunction related to CNS-vestibular pathology was 72%. In a subsequent study,[39] the latency or symmetry of the response to perturbation was abnormal in 81% of the patients with CNS disorders. The marked improvement in sensitivity from the initial stud[38] to the follow-up study[39] was attributed to the development of software that provided greater accuracy in calculating latency.

Keim[42] reported that none of the 16 patients with CNS-vestibular deficits had abnormal latencies. Only 11 of 32 patients (34%) with nonvestibular CNS involvement had prolonged latencies on the motor test (Tab. 6).[12] The data presented by Stockwell[41] showed abnormalities in the motor test for 21% (4 out of 19) of the patients with CNS disease that usually involved the vestibular system.

Motor Test Specificity

In order for the posturography motor test to have maximum specificity, all unimpaired subjects and patients with PVDs would be expected to have a negative motor test. The latency of response to perturbation is expected to be within normal limits for these populations.[32,33] Voorhees found that only 3% (n = 4)[39] to 5% (n = 5)[38] of the patients diagnosed with peripheral vestibular impairments were found to have elevated latencies on the motor test. The specificity of the motor test in these studies, therefore, was 95% to 97%

Stockwell[41] found that 165 of 174 patients (95%) with a normal ENG examination also had a normal motor test. In addition, the data presented in that study[41] showed that 3 of 93 patients (97%) with a PVD also had a normal motor test. The specificity of the motor test, therefore, was 95% to 97%. Stockwell interpreted this finding as a measure of sensitivity and concluded that dynamic posturography "proved to be highly insensitive to peripheral vestibular disorders."[46(p52)] This interpretation may not be accurate, in my view, because patients with PVDs are not expected to show abnormalities in the motor test, as discussed earlier. Patients with PVDs, according to the operational definitions stated earlier, would be more appropriately considered as a population for the study of specificity.

Static Platform Posturography: Tests of Balance and Classffication of Vestibular Deficits

Black et al[60] studied patients with peripheral vestibular disease during stance with eyes opened and closed in the standard Romberg position (feet together) and with the feet in tandem stance. A PVD pattern was defined as normal performance with eyes open (below the 95th percentile of sway amplitude in unimpaired subjects), but as abnormal sway amplitude with eyes closed for either the standard or tandem Romberg position.[23]

Norre and Forrez[31] also studied patients with PVDs in the standard Romberg position, but included a condition with head extension and eyes closed. The area and distance traveled by the center of force were compared with the 95% confidence interval for unimpaired subjects. Modifications of this protocol included stance on a foam base placed on the force platform to disrupt balance and the addition of center-of-force velocity as a dependent variable.[61] Abnormalities of balance, in general, were found for patients with PVDs, but no specific pattern of abnormal sway could be attributed to vestibular impairment. The use of static platform posturography as a general measure of balance dysfunction, rather than a test to localize an anatomical lesion, was also confirmed in an investigation of nonvestibular dizziness by Gagey.[62] Other tests of balance involving static platform posturography included the use of galvanic current to stimulate the labyrinth.[63] These tests are not often used because it is impossible to predict the path of the current and, therefore, the structures tested by electrical stimulation. Recent improvements in the test protocol have been reported.64 A complete review of this topic, however, was beyond the scope of this article because of the infrequent use of this test to detect vestibular dysfunction.

Sensivity and Specificity of Static Platform Posturography

Black and Wall[23] found that the amplitude of body sway during the Romberg and tandem Romberg tests detected vestibular abnormalities in 60% of the patients with Meniere's disease (n=42) and in 57% (n=42) of the patients with benign paroxysmal positional nystagmus (Tab. 7).[23] Vestibuloocular screening tests (ie, vertical-axis rotation and caloric irrigation) detected abnormalities with even less precision (Tab. 7). The sensitivity of detection, however, increased to 83% for Meniere's disease when the Romberg tests were combined with rotation testing. With regard to specificity, Black and Wall[23] found that neither Romberg or VOR tests had "specificity" for a particular type of vestibular deficit because substantial numbers of patients were classified as "normal" with each test (Tab. 7). The number of subjects, however, who should have had a negative response was not determined. A measure meeting the operational definition of specificity stated in the Standards for Tests and Measurements in Physical Therapy Practice,[36] therefore, could not be derived from their data.

Wall and Black[37] classified patients as (1) likely to have a vestibular deficit, (2) vestibular deficit uncertain, and (3) normal vestibular function expected. Unimpaired subjects were used for the determination of specificity. Wall and Black[37] reported that static posturography and rotation tests each had a sensitivity on the order of 50% (Tab. 7). When posturography and rotation tests were combined (either test abnormal), the sensitivity increased to 78%. The specificity of static posturography was 95% in an unimpaired population. For patients with symptoms not related to vestibular impairment, the specificity was 71% (Tab. 7).

Norre et al[27] and Norre and Forrez[31] used static posturography during the Romberg tests, but added a third condition of head extension with eyes closed. The dependent measure was the length of the sway path during a 58-second stance test. An abnormal test was defined as a sway path that exceeded the upper limit of the sway path (mean + two standard deviations) demonstrated by 40 unimpaired subjects. Forty-seven of 82 patients (57%) with paroxysmal positional and 13 of 24 patients (54%) with Meniere's disease[31] had abnormal posturography (Tab. 7). Thirty-five percent of the patients with PVDs (29 of 82) had an asymmetrical VOR following vertical-axis rotation tests.[27]

Norre et al[65] also studied patients with unilateral vestibular hypofunction together with a group of patients with Meniere's disease. They reported that the sensitivity of static posturography was equivalent to rotation testing (Tab. 7) for the detection of both types of peripheral vestibular disorders. The level of sensitivity in each case was below 50% (Tab. 7).

Composite Levels of Sensitivity and Specificity

Data were pooled from the studies reviewed (Tabs. 6 and 7) so that a composite level of sensitivity and specificity could be determined for PVDs and central vestibular deficits. The sensitivity of the SOT was evaluated across five studies[23,38,40-42] (Tab. 6) involving a total of 836 patients with PVDs. Abnormalities in the SOT were detected in only 40% (n=338) of the patients (Fig. 1). The study by Parker[46] was not included in the composite because it was not clear when the SOT or the motor test was used to detect vestibular dysfunction. The sensitivity of static platform posturography was evaluated across six studies[23,25,27,31,37,65] (Tab. 7) involving a total of 571 patients with PVDs, and abnormalities were detected in 53% (n=302) of these patients. Tests of spontaneous and positional nystagmus and the horizontal component of the VOR, by comparison, detected PVD in 379 of 798 patients (48%) with suspected peripheral vestibular impairment.[23,25,27,37,38,41,42,65]

A total of 389 patients with central vestibular deficits were identified in five studies,[38-42] and SOT abnormalities were found in 540/o (n=209) of these patients (Fig. 2). The motor test was abnormal in 35% (n=41) of 119 patients with central vestibular disease.[38,39,41,42] The sensitivity for detecting CNS-vestibular deficits with dynamic posturography increased to 72% when the SOT was combined with the motor test (Fig. 2).[38]

When posturography was combined with other tests of vestibular function, an abnormality was defined by a positive result in either test. Combining static posturography with other tests of vestibular function increased the sensitivity of detecting PVD to 61% of 221 patients.[23,37] When the SOT was combined with other vestibular tests, the sensitivity in peripheral disorders increased to 89% of 52 patients (Fig. 1).[25] Although Asai et al[25] showed that combined testing increased overall sensitivity compared with the SOT alone, the contribution of dynamic posturography was small in relation to other vestibular tests (Fig. 1). Seventy-seven percent of the patients with PVDs were detected in this study (40 of 52 patients) with a three-test ENG battery. Only 6 additional patients with PVDs showed abnormal findings with the SOT, bringing the overall sensitivity to 89% (Fig. 1). In comparison, the combination of static posturography with other vestibular tests increased Sensitivity from 53% to 61% (Fig. 1), indicating that other vestibular tests did not add substantially to the sensitivity of static posturography. The discrepancy between the overall levels of sensitivity in combined tests involving static or dynamic posturography (shaded plus open portion of each bar in Fig. 1) might have been due to differences in the number and type of supplemental vestibular function tests across studies, the type of patients selected for analysis, and variations in the criteria for abnormal test results.

The specificity of the SOT was evaluated across three studies[40-42] involving a total of 787 patients with a normal vestibular examination and unimpaired subjects. The SOT was normal in 92% (n=725) of these subjects. Static posturography was normal in 130 of 139 (94%) unimpaired subjects and patients with nonvestibular disease (Tab. 7).[37] The motor test was normal in 500 of 522 (96%) unimpaired subjects and patients with PNDs.[8,39,41,42]

Discussion

Posturography is a clinical tool that measures balance and sensory interactions for equilibrium. The frequent complaint of postural instability in patients with vestibular dysfunction makes this tool a potentially useful component of patient assessment. There are many disorders, however, that can influence balance behavior that do not necessarily involve the vestibular system. This perspective, therefore, focused on disorders that were likely to involve a vestibular deficit.

The abnormal findings in posturography did not always identify the same patients as did results from other vestibular tests.[23,25,37,38,42,65] Asai et al,[25] for example, reported that both the SOT of dynamic posturography and the rotation test were abnormal in 8 out of 52 patients with PVDs. Posturography, however, detected abnormalities in 12 other patients who had normal rotation tests.[25] Discordance has also been observed between the results of static posturography and VOR tests.[23,25,65] These findings suggest that although the composite sensitivity of posturography is low, posturography provides information that is not always redundant with VOR testing. The criteria for classifying patients with vestibular dysfunction that is based primarily on VOR testing,[43,44] therefore, should be reexamined.

Sensitivity and Specificity of Posturography Sensory Organization Tests

A normal Romberg test does not necessarily rule out abnormal vestibular function, because the patient can use sensory inputs provided by the feet and ankles to compensate for vestibular impairment when the eyes are closed. Black et al[6] showed that the standard Romberg test was normal preoperatively and throughout a 4-year postoperative period for subjects with surgically induced unilateral loss of vestibular function. Black[6,66] suggested that dynamic posturography would increase the sensitivity of detecting PVDs compared with static posturography, because the SOT alters somatosensory information from the lower extreminities. This point of view, however, was not supported by the literature reviewed in this article. The sensitivity of dynamic posturography SOTs for detecting balance deficits related to peripheral vestibular dysfunction (40%) was slightly lower than that of static posturography (53%). The lack of sensitivity of the Romberg test in the study by Black et al[6] may have been due to the fact that only a standard Romberg position was used. Other studies addressing the sensitivity of Romberg's tests utilized both a standard position (feet together) and tandem stance (Tab. 7).[23,37]

[TABULAR DATA 7 OMITTED]

There are several factors that might alter the sensitivity of dynamic posturography. All of the studies that evaluated the sensitivity of dynamic platform posturography used a system (EquiTest) that estimated peak-to-peak body sway angles from changes in the center of force. Shepard et al[54] found that sway estimates on the EquiTest missed many subtle changes in sway that were detected with a video-kinematic analysis. In addition, dynamic platform posturography as originally described by Black et al[32] and the EquiTest posturography system used in all subsequent studies to evaluate sensitivity differed with respect to the test duration and sequence of conditions, method of measuring sway, and use of systematic variation in sway-referenced gains([dagger]) (Tab. 8).

[TABULAR DATA 8 OMITTED]

The position of the patient may be an additional factor that affects the sensitivity of dynamic posturography. Barin et al[67] reported that patients with unilateral vestibular deficits had more body sway when the head was tilted contralateral to the lesion compared with the head-neutral position. Lateral head tilt in unimpaired subjects did not alter postural stability.[67] This maneuver, therefore, might improve the sensitivity and specificity of posturography for detecting PVDs. Head extension has also been suggested as a maneuver for increasing the sensitivity of the posturography SOT.[55,67] The head extension maneuver, however, increased body sway and the number of falls in both unimpaired subjects and those with vestibular deficits.[55,67] Ghilardi et al[68] utilized a head-shaking maneuver to increase the sensitivity and specificity of static posturography,([double dagger]) but this procedure has not yet been studied extensively with dynamic posturography.

Various patterns of SOT abnormalities associated with vestibular dysfunction (Tabs. 2-5) have not been routinely used in studies of sensitivity. There is a need to define "abnormal balance" in order to improve the precision of sensitivity estimates. The use of a composite SOT score[42] or the criterion of abnormality in any single test condition[25,38,39] may be too general to describe balance behavior related to vestibular impairment. Keim,[42] for example, found that the composite SOT score was abnormal in 66% of 32 patients with nonvestibular CNS disease. This result was not surprising because an abnormal composite score may reflect a general state of balance dysfunction. Measurement of the sensitivity to detect vestibular dysfunction, however, should differ from general measures of balance dysfunction. Observations of abnormal balance can be expected in many patients with nonvestibular deficits,[42,62] but these patterns theoretically should not emulate the patterns "reserved" for describing vestibular deficits (Tabs. 2-5).

Motor Test Sensitivity and Specificity

Patients with peripheral vestibular impairment are expected to show normal sway-correcting response latencies following externally induced balance perturbations.[32,33,38] A positive motor test appeared to rule out peripheral vestibular impairment, because 95% to 100% of the patients with PVDs had normal responses to platform perturbation (Tab. 6).[38,42] The sensitivity of the motor test to identify vestibular deficits associated with CNS disease varied from 0% to 81%,[38,39,42] but was consistently above 70% when combined with the SOT.[38,39] It can be reasonably concluded, therefore, that patients who test positive on the motor test probably do not have a PVD unless it is coupled with a CNS lesion involving the long-loop reflex pathways.

Implications for Physical Therapy Practice

In spite of the low sensitivity of the SOT for detecting PVDs and central vestibular deficits, this test has been shown to correlate with the patient's symptoms. Asai[25] found that the SOT was normal in nearly all patients who did not report vertigo and abnormal in nearly all the patients who had vertiginous symptoms. The results of the SOT have also been correlated with scores on a self-assessment scale designed to quantify the functional, emotional, and physical effects of dizziness and unsteadiness.[30] Specific pretherapy SOT patterns have been found to be prognostic indicators of symptom and functional status improvement.[2] Patients with "visual-vestibular dysfunction" and those classified as having "severe dysfunction" (Tab. 5) had poorer outcomes with respect to posttreatment disability and symptom-intensity scores compared with other classifications.[2,69]

The difficulty in obtaining an objective measure of vestibular dysfunction has been further addressed by incorporating subjective assessments of perceived functional limitations. Several studies have dealt with this issue by including patient self-report scales for measuring changes in symptom frequency or intensity,[2,30,71] scales to rate the level of disability before and after therapy,[2,3] and measures of the patients' perceived sensitivity([sections]) to motion.[15,30]

Suggestions for Future Research

Nearly all of the studies that evaluated the sensitivity and specificity of dynamic posturography were retrospective analyses (Tab. 6). Retrospective designs have serious limitations with regard to experimenter bias and other threats to internal and external validity.71 In addition, when the goal is to compare the sensitivity of different vestibular function tests,[41,42,46] the outcomes may be influenced by preselection of patient cases. Future work, therefore, should involve prospective evaluations of sensitivity and specificity with clear operational definitions for "abnormal balance" and "vestibular dysfunction." In addition, the "criterion standards" for identifying patients with PVDs versus central vestibular deficits should be identified prior to initiating the study.

Conclusions

The literature addressing the sensitivity and specificity of platform posturography was evaluated with respect to the operational definitions presented in the American Physical Therapy Association's Standards for Tests and Measurements in Physical Therapy Practice.[36] Dynamic posturography SOT was found to have a slightly lower level of sensitivity compared with static posturography for the detection of peripheral vestibular dysfunction. Both types of posturography were highly specific for ruling out balance deficits in patients with a normal vestibular examination. The motor "perturbation" test of dynamic posturography also provided a specific measure of balance deficit. A positive motor test appeared to rule out peripheral vestibular impairment. The definition of "abnormal balance" related to vestibular dysfunction was not consistent across studies. Further research is needed to standardize the objective measurements used to identify vestibular dysfunction and to assess the sensitivity and specificity of those measures in an a priori fashion.

[dagger] Sway-referenced gain refers to the amount of instantaneous platform or visual enclosure displacement that is caused by a given degree of body sway. For example, a sway-referenced gain of 0.5 will rotate the force platform 0.5 degrees downward for every degree of forward body sway.

([sections]) Motion sensitivity is quantified by evaluating the duration and intensity of dizziness during rapid changes in specific body positions. Changes in position are initiated by the patient. A motion sensitivity quotient is calculated by multiplying the number of positions that provoked dizziness by the severity and duration scores.[15] On a scale of 0 to 100, higher scores indicate a greater sensitivity to motion.

References

[1] Kroenke K, Lucas CA, Rosenberg ML, et al. Causes of persistent dizziness: a prospective study of 100 patients in ambulatory care. Ann Intern Med. 1992; 1 17:898-904. [2] Shepard NT, Telian SA, Smith-Wheelock M, Raj A. Vestibular and balance rehabilitation therapy. Ann Otol Rhinol Laryngol. 1993; 102(3, part 1):198-205. [3] Shepard NT, Telian SA, Smith-Wheelock M. Habituation and balance retraining therapy: a retrospective review. Neurol Clin. 1990;8: 459-475. [4] Zee DS. Adaptation to vestibular disturbances; some clinical implications. Acta Neurol Belg, 1991:91:97-104. [5] Fetter M, Diener H-C, Dichgans J. Recovery of postural control after an acute unilateral vestibular lesion in humans. J Vestih Res. 1991;1:373-383. [6] 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. [7] Cass SP, Kartush JM, Graham MD. Clinical assessment of postural stability following vestibular nerve section. Laryngoscope. 1991; 101:1056-1059. [8] Konrad HR, Tomlinson D, Stockwell CW, et al. Rehabilitation therapy for patients with disequilibrium and balance disorders. Otolarynogol Head Neck Surg. 1992; 107:105-108. [9] Tangeman PT, Wheeler J. Inner ear concussion syndrome: vestibular implications and physical therapy treatment. Top Acute Care Trauma Rebabil. 1986;1:72-83. [10] Semont A, Freyss G, Vitte E. Curing the BPPV with a liberatory maneuver. Adv Otorbinolaryngol. 1988;42:290-293, [11] Shumway-Cook A, Horak FB. Vestibular rehabilitation: an exercise approach to managing symptoms of vestibular dysfunction. Semin Hearing. 1989;10:196-209. [12] Herdman SJ. Physical therapy management of vestibular disorders in older patients. Physical Therapy Practice. 1992;1(1):77-87, [13] Herdman SJ. Treatment of benign paroxysmal positional vertigo. Phjs Ther. 1990;70: 381-388. [14] Herdman SJ. Exercise strategies for vestibular disorders. Ear Nose Throat J 1989;68: 961-964. [15] Smith-Wheelock M, Shepard NT, Telian SA. Physical therapy program for vestibular rehabilitation. Am J Otol. 1991; 12:218-225. [16] Troost BT, Patton JM. Exercise therapy for positional vertigo. Neurology. 1992;42: 1441-1444. [17] Takemori S, Ida M, Umezu H. Vestibular training after sudden loss of vestibular functions. ORL J Otorbinolaryngol Relat Spec. 1985;47:76-83. [18] Dix MR. The rationale and technique of head exercises in the treatment of vertigo. Acta Otorbinolaryngol Beig. 1979;33:370-384. [19] Brandt T, Daroff RB. Physical therapy for benign paroxysmal positional veitigo. Arch Otolaryngol. 1980;106:484-485. [20] Herdman SJ. Assessment and treatment of balance disorders in the vestibular deficit patient. In: Balance: Proceedings of the APTA Forum; June 13-15, 1989; Nashville, Tenn. Alexandria, Va: American Physical Therapy Association; 1990:87-94. [21] Keim RJ, Cook M, Martini D. Balance rehabilitation therapy. Laryngoscope. 1992;102: 1302-1307. [22] Hecker HC, Haug CO, Herndon JW. Treatment of the vertiginous patient using Cawthorne's vestibular exercises. Laryngoscope. 1974;84:2065-2072. [23] Black FO, Wall C III. Comparison of vestibulo-ocular and vestibulospinal screening tests. Otolaryngol Head Neck Surg. 1981;89: 811-817. [24] Black FO. Vestibular function assessment in patients with Meniere's disease: the vestibulospinal system. Laryngoscope. 1982;92: 1419-1436. [25] Asai M, Watanabe Y, Ohashi N, Mizukoshi K. Evaluation of vestibular function by dynamic posturography and other equilibrium examinations. Acta Otolaryngol Suppl (Stockh). 1993;504:120-124. [26] Lindsay JR. Histopathology of Meniere's disease as observed by light microscopy. in: Pulec JL, ed. Meniere's Disease. Philadelphia, Pa: WB Saunders Co; 1968:319-337. [27] Norre ME, Forrez G, Beckers A. Vestibular habituation training and posturography in benign paroxysmal positioning vertigo. ORL J Otorbinolaryngol Relat Spec. 1987;49:22-25. [28] Goebel JA, Paige GD. Dynamic posturography and caloric test results in patients with and without vertigo. Otolaryngol Head Neck Surg. 1989;100:553-558. [29] Norre ME, De Weerdt W. Treatment of vertigo based upon habituation, II: technique and results of habituation training. J Laryngol Otol. 1980;94:971-977. [30] Jacobson GP, Newman CW, Hunter L, Balzer GK. Balance function test correlates of the Dizziness Handicap Inventory. J Am Acad Audiol. 1991;2:253-260. [31] Norre ME, Forrez G. Vestibulospinal function in otoneurology. ORL J Otorhinolaryngol Relat Spec. 1986;48:37-44. [32] Black FO, Wall C III, Nashner LM. Effects of visual and support surface orientation references upon postural control in vestibular deficient subjects. Acta Otolaryngol. 1983;95: 199-210. [33] Nashner LM, Black FO, Wall C III. Adaptation to altered support and visual conditions during stance: patients with vestibular deficits. J Neurosci, 1982;2:536-544. [34] Mirka A, Black FO. Clinical application of dynamic posturography for evaluating sensory integration and vestibular dysfunction. Neurol Clin. 1990;8:351-359. [35] Nashner LM, Peters JF. Dynamic posturography in the diagnosis and management of dizziness and balance disorders. Neurol Clin. 1990;8:331-349. [36] Rothstein JM, Echternach JL. Primer on Measurement: An Introductory Guide to Measurement Issues. Alexandria, Va: American Physical Therapy Association; 1993: Appendix, p 14. [37] Wall C III, Black FO. Posturat stability and rotational tests: their effectiveness for screening dizzy patients. Acta Otolaryngol (Stockh). 1983;95:235-246. [38] Voorhees RL. The role of dynamic posturography in neurotologic diagnosis. Laryngoscope. 1989;99:995-1001. [39] Voorhees RL. Dynamic posturography findings in central nervous system disorders. Otolaryngol Head Neck Surg. 1990; 103: 96-101. [40] Hamid MA, Hughes GB, Kinney SE. Specificity and sensitivity of dynamic posturography: a retrospective analysis. Acta Otolaryngol Suppl (Stockb). 1991;481:596-600. [41] Stockwell CW. Vestibular function testing: 4-year update. In: Cummings C, ed. Otolaryngology Head and Neck Surgery Update II. Chicago, Ill: Mosby-Yearbook; 1990:39-53. [42] Keim RJ. Clinical comparisons of posturography and electronystagmography. Laryngoscope. 1993;103:713-716. [43] Shumway-Cook A, Horak F, Black FO. A critical examination of vestibular function in motor-impaired learning-disabled children. Int J Pediatr Otorbinolaryngol. 1987; 14:21-30. [44] Horak FB, Shumway-Cook A, Crowe TK, Black FO. Vestibular function and motor proficiency of children with impaired hearing, or with learning disability and motor impairments. Dev Med Child Neurol. 1988;30:64-79. [45] Black FO, Nashner LM. Postural disturbance in patients with benign paroxysmal positional nystagmus. Ann Otol Rhinol Laryngol 1984;93:595-599. [46] Parker SW. Vestibular evaluation: electronystagmography, rotational testing, and posturography. Clin Electroencepbalogr. 1993; 24:151-159. [47] Truex RC, Carpenter MB. Human Neuroanatomy. 6th ed. Baltimore, Md: Williams & Wilkins; 1969:282, 354. [48] Nashner LM. A model describing vestibular detection of body sway motion. Acta Otolaryngol Stockb). 1971;72:429-436. [49] Nashner LM. Vestibular postural control model. Kybernetik. 1972;.10:106-110. [50] Hess BJM. Three-dimensional head angular velocity detection from otolith afferent signals. Biol Cybern. 1992;67:323-333. [51] Horstmann GA, Dietz V. The contribution of vestibular input to the stabilization of human posture: a new experimental approach. Neurosci Lett. 1988;95:179-184. [52] Deitz V, Horstmann GA, Berger W. Fast head tilt has only a minor effect on quick compensatory reactions during the regulation of stance and gait. Exp Brain Res. 1988;73: 470-476. [53] Dichgans J, Diener HC. The contribution of vestibulo-spinal mechanisms to the maintenance of human upright posture. Acta Otolaryngol (Stockb). 1989;107:338-345. [54] Shepard NT, Schultz A, Alexander NB, et al. Postural control in young and elderly adults when stance is challenged: clinical versus laboratory measurements. Ann Otol Rhinol Otolaryngol. 1993;102:508-517. [55] Jackson RT, Epstein CM, Boyette JE. Enhancement of posturography testing with head tilt and energy measurements. Am J Otol. 1991;12:420-425. [56] Di Fabio RP. Lower extremity antagonist muscle response following standing perturbation in subjects with cerebrovascular disease. Brain Res. 1987;406:43-51. [57] Diener HC, Dichgans J, Hulser PJ, et al. The significance of delayed long-loop responses to ankle displacement for the diagnosis of multiple sclerosis. Electroencepbalogr Clin Neurophysiol. 1984;57:336-342. [58] Shepard NT, Telian SA, Niparko JK, et al. Platform pressure test in identification of perdymphatic fistula. Am J Otol. 1992;13:49-54. [59] Black FO, Lilly DJ, Peterka RJ, et al. The dynamic posturographic pressure test for the presumptive diagnosis of perilymph fistulas. Neurol Clin. 1990;8:361-374. [60] Black FO, Wall C III, O'Leary D. Computerized screening of the human vestibulospinal system. Ann Otol. 1978; 87:853-860. [61] Norre ME. Sensory interaction testing in platform posturography. J Laryngol Otol. 1993:107:496-501. [62] Gagey PM. Non-vestibular dizziness and static posturography. Acta Otorbinolaryngol Belg. 1991;45:335-339. [63] Proud GO. Vestibular function testing. In: Paparella MM, Shumrick DA, eds. Otolaryngology:Basic Sciences and Related Disciplines, Volume 1. Philadelphia, Pa: WB Saunders Co; 1973. [64] Sekitani T. Differential diagnosis of vertigo-combined galvanic test and MRI. Acta Otolaryngol Suppl (Stockb). 1988:458:113-119. [65] Norre ME, Forrez G, Beckers A. Posturographic findings in two common peripheral vestibular disorders. J Otolaryngol. 1987;16: 340-344. [66] Black FO. Equilibrium (editorial). Otolaryngol Head Neck Surg. 1992;106:16-18. [67] Barin K, Seitz CM, Welling DB. Effect of head orientation on the diagnostic sensitivity of posturography in patients with compensated unilateral lesions. Otolaryngol Head Neck Surg. 1992;106:355-362. [68] Ghilardi PL, Fattori B, Casani A, Piragine F. Posturography in unilateral peripheral vestibular hypofunction. Acta Otorbinolaryngol Ital. 1990;10:347-356. [69] Telian SA, Shepard NT, Smith-Wheelock MS, Kemink JL. Habituation therapy for chronic vestibular dysfunction: preliminary results. Otolaryngol Head Neck Surg. 1990; 103:89-95. [70] Herdman SJ, Tusa RJ, Zee DS, et al. Single treatment approaches to benign paroxysmal position vertigo. Arch otolaryngol Head Neck Surg. 1993;119:450-454. [71] Bork CE. Research in Physical Therapy. Philadelphia, Pa: JB Lippincott Co; 1993:148-154.