Usefulness of galvanically induced nystagmus as a diagnostic test of peripheral vestibular hypofunction

Abstract

BACKGROUND:

Galvanic Vestibular Stimulation (GVS) has long been used as a vestibular stimulus. A major issue observed in GVS research was the high variability of the responses, which has led some researchers to question its diagnostic utility.

OBJECTIVES:

Determine the diagnostic accuracy of galvanically induced nystagmus for the diagnosis of subjects with unilateral peripheral vestibular hypofunction.

METHODS:

A total of 195 dizzy patients were prospectively enrolled, forming a consecutive series as they were received. Individuals with spontaneous nystagmus or using vestibular sedatives were excluded. Nystagmus induced by a 4 mA galvanic stimulus was compared with the caloric test as a reference standard.

RESULTS:

Of the 195 subjects tested with GVS, 115 were subjects with a unilateral peripheral vestibular hypofunction. The presence of nystagmus showed an AUC of only 0.529 (0.478 to 0.580, p = 0.125), and the maximum slow phase velocity of nystagmus showed an AUC of only 0.523 (0.472 to 0.573, p = 0.439).

CONCLUSIONS:

Since neither AUC is better than random discrimination, this study concludes that GVS-induced nystagmus is not useful as a diagnostic tool for unilateral peripheral vestibular hypofunction. These findings discourage the use of GVS-induced nystagmus in the clinical setting.

Keywords

Vestibular diagnostic tests routine galvanic skin response

1. Introduction

Although Galvanic Vestibular Stimulation (GVS) has been a subject of research for more than a century, it has had little, if any, use in the clinical setting. GVS was first described in 1820 by physiologist Johann Evangelist Purkinje [22]. He noticed patients falling down whilst applying a galvanic current in their heads. In 1871, neurologist Eduard Hitzig [11] observed nystagmus when applying galvanic current to the brains of dogs. These two researchers were the ones who led the way in GVS description, by defining its two main effects: vestibulo-spinal reflexes and vestibulo-ocular reflexes, respectively.

Investigating the neurophysiological effects of GVS is a continuing concern within neurotology. It has previously been observed than GVS applied to the mastoid process of healthy subjects causes a mainly horizontal nystagmus [14] and a subjective sensation of tilt [8]. Studies in guinea pigs have shown that GVS modulates the firing rate of vestibular afferents neurons [12]; cathodic currents increase the discharge firing rate but anodic currents decrease the discharge firing rate [9]. It has been reported that GVS-induced nystagmus displays a slow phase velocity (SPV) of approximately 5°/sec toward the anode (in darkness, in response to 5 mA stimuli), and a linearrelationship exists between the SPV and the intensity of GVS current of approximately 1°/sec/mA [16]. Thus, increasing GVS intensity also increases the SPV of the induced nystagmus.

Whether a surface-delivered GVS acts primarily as an otolithic or a canalicular stimulus has been a matter of considerable debate [5, 6], however, it is well documented that it activates the vestibular system [12]. For this reason, some attempts were made to determine its usefulness in a clinical diagnostic setting [20]. In contrast, highly variable nystagmus resulting from the use of GVS has resulted in researchers questioning whether GVS could ever be used as a diagnostic tool for vestibular pathology [2].

The last two decades have seen a growing body of literature describing particular effects of GVS, and its use as an experimental tool. However, a lack of studies with a focus on its clinical use is evident. To date, the usefulness of GVS as a diagnostic test has received scant attention in the research literature.

This study intends to determine the extent to which GVS can be useful as a diagnostic tool to diagnose patients with unilateral peripheral vestibular hypofunction.

2. Methods

2.1. Study design

This study has been performed following the Standards for Reporting of Diagnostic Accuracy (STARD) guidelines [3], in order to report all the information required in studies of diagnosticaccuracy.

The ultimate goal of this study was to determine whether GVS can identify those who are having any form of unilateral peripheral vestibular hypofunction. Data collection was planned before the index test and reference standard were performed. It was performed in a tertiary referral hospital and was approved by the institutional ethics committee (approval code FPNT-CEIB-04A).

2.2. Participants

Inclusion criteria were designed that capture patients with dizziness referred for assessment in a tertiary referral center. Enrollment was consecutive in “real-time”, including all new patients who were seen in a 1-year period. This consecutive enrollment was chosen because it represents a higher methodologic quality than prospectively enrollment of nonconsecutive patients [24]. All potentially eligible participants were asked for written informed consent, and both the index test and reference standard were performed on the same day in which they were recruited.

The exclusion criteria were: spontaneous nystagmus (because this sign can guide the physician, GVS is intended to be used in subjects with uncertain diagnosis), and use of vestibular sedatives in the previous 3 days.

2.3. Test methods

The index test being studied is GVS. It was delivered via surface electrodes of 660 mm2 (Natus JellyTab, Natus Medical Incorporated, Mundelein, IL, USA) placed over each mastoid process. A battery-powered galvanic stimulator (IDEE, Maastricht University, Maastricht, The Netherlands) was used for all subjects. A 4 mA constant DC current was delivered to the electrodes. The output current was reversed in polarity by means of a selection switch. A 20-second recording during GVS delivery was performed. There were two conditions: Cathode Left Anode Right (CLAR) and Anode Left Cathode Right (ALCR).

Subjects were evaluated in a dark room, seated in a padded chair, using a 50 Hz sampling infrared videonystagmography (VNG Ulmer Synapsys, Marseille, France); the complete absence of possible sources of light was confirmed before the start of the recordings to ensure the absence of visual fixation. The subjects Reid’s plane (imaginary line between the inferior margin of the orbit and the upper margin of external auditory meatus) was positioned in an approximately earth-horizontal plane, allowing for comparable orientation across subjects.

For each recording, the horizontal slow phase velocity (SPV) was measured automatically and registered for every eye-movement. A positive result in the index test was defined as an eye-movement composed of more than five fast and slow phase eye movements that occurred consecutively at a slow phase eye velocity faster than 1 deg/sec in either CLAR or ALCR conditions. Fig. 1 illustrates examples of records of one subject with a positive response and another subject with no response.

Fig. 1.

A “pathological” response was defined as a positive response to GVS, as can be seen in the upper record, if the response was absent or the nystagmus reached velocities lower than 1 deg / sec, it was considered negative.

The reference standard chosen for this study was the caloric test. It was chosen as the reference standard because it is the most widely used test to define peripheral vestibular hypofunction, and because the alternatives (rotating chair, video head impulse test) are not available in most vestibular examinationcenters.

Caloric testing was done using irrigations of 200 cc of water at 30°C sequentially in each ear, followed by irrigations of 200 cc of water at 44°C. Again, subjects were evaluated in a dark room, seated in a padded chair, using a 50 Hz sampling infrared videonystagmography (VNG Ulmer Synapsys, Marseille, France). The individuals were reclined until the head was 30º from the horizontalplane.

Unilateral peripheral vestibular hypofunction was defined as a labyrinth asymmetry of more than 24% in the caloric test, as defined by Proctor et al. [21].

Neither the clinical information nor the results of the reference standard were available to the index test reader. Likewise, the reader of the reference standard did not have access to the index test.

2.4. Analysis

Sensitivity, specificity, positive and negative predictive values as well as likelihood ratios were calculated for the presence of galvanically-induced nystagmus. Sensitivity, specificity, as well as positive and negative predictive values were expressed as percentages. Confidence intervals for sensitivity and specificity have been calculated as Clopper-Pearson confidence intervals [4]. Confidence intervals for the likelihood ratios were calculated using the Log method as described by Altman et al. [1]. Confidence intervals for the predictive values were calculated as the standard logit confidence intervals as described by Mercaldo et al. [17]. Positive and Negative predicted values cannot be extrapolated to be used in the general population. Conversely, sensitivity, specificity and likelihood ratios are not influenced by the prevalence of the target condition, and these results can be used in other populations.

The diagnostic performance of this test (i.e.: the accuracy of the test to discriminate subjects with and without the target condition) was evaluated using Receiver Operating Characteristic (ROC) curve analysis [18]. ROC curves were also used to compare the diagnostic performance of the test either analyzed as presence / absence of nystagmus, or as maximum SPV achieved by nystagmus. The method of Delong et al. [7] was used for the calculation of the Standard Error of the Area Under the Curve (AUC) and of the difference between two AUCs.

Indeterminate results did not occur in this study, because nystagmus was recorded by videonystagmography, and cut-off points were clearly established prior to the start of the study. Therefore, no statistical adjustment was necessary for handling indeterminate results either in the index test or in the reference standard. The same goes for missing data, since ocular movements were recorded and stored electronically in the videonistagmograph, and all records were accessible at the end of the study. Therefore, no statistical adjustment was necessary for handling missing data.

A Youden’s index analysis was expected to be made from the results of the ROC curves, if the diagnostic accuracy of the index test would have been different than random discrimination. However, as it will be seen in the Results section, this analysis was not necessary, and only the pre-specified cutoff point was used (i.e.: nystagmus was considered present if it has SPV≥1 deg/sec).

Sample size calculation was based on a pre-study survey that indicated that the prevalence of unilateral peripheral vestibular hypofunction was expected to be 50%, and the expected sensitivity of GVS to be 90%, then the minimum number of subjects required is 141, according to the Hajian-Tilaki equations [10].

The methodological approach taken in this diagnostic accuracy study is a comparison between galvanically induced nystagmus and caloric deficit. Although GVS stimulates more than just the horizontal canal, as has been confirmed by neurophysiological studies such as that of Kim & Curthoys [12], this study was conducted with the focus solely on the clinical utility of GVS, that is, if GVS was able to provide information so similar to the caloric stimulation that it could replace it. Therefore, GVS was studied in its ability to predict caloric deficit, and not in its ability to differentiate between healthy subjects and subjects with vestibulopathies, since there are currently no vestibular tests that can make this distinction. Comparison of the galvanically-induced nystagmus between healthy subjects and subjects with vestibulopathies has already been made [15], and there is no need to repeat these comparisons. What is missing is a study that assesses the clinical utility in a real scenario. That is the reason why this study was conducted in subjects with dizziness, which is the target population to which vestibular tests are performed in real life.

3. Results

3.1. Participants

Two hundred sixty-two individuals were received in our center due to dizziness (potentially eligible participants). Twenty-two subjects were excluded because they had spontaneous nystagmus, and another thirty-one subjects were excluded because they had used vestibular sedatives in the previous three days. Of these eligible participants, 14 refused to participate in the study. Therefore, the index test (i.e.: GVS) was performed on 195 individuals. The diagram in Fig. 2 shows the flow of participants.

Fig. 2.

STARD diagram reporting flow of participants through the study.

The mean age (±SD) of the participating subjects was 52.7±14.2 years; 108 (55%) were women. All subjects were referred to the tertiary referral center by the secondary care specialist with the diagnosis of dizziness for further study. Those with the target condition (unilateral peripheral vestibular hypofunction) had a caloric deficit of 59.4±30.7 %. The distribution of diagnoses among the subjects with the target condition was: 45% of subjects with unilateral meniere’s disease (MD), 23% with vestibular neuritis, 11% with bilateral MD, 9% with vestibular schwannoma (VS), 7% with uncertain diagnosis / non-vestibular dizziness, and 4% with benign paroxysmal positional vertigo. Those without the target condition had a diagnosis distribution as follows: 56% with uncertain diagnosis / non-vestibular dizziness, 18% with vestibular migraine, 15% with benign paroxysmal positional vertigo, 4% with Charcot-Marie-Tooth disease, 2% with bilateral MD, 2% withVS, and 2% with unilateral MD.

No clinical interventions were performed on the subjects between the index test and the reference standard, as both tests were performed in the same session with a time interval between them of less than 5 minutes.

3.2. Test results

Of the 195 subjects tested with GVS, 115 were subjects with a unilateral peripheral vestibular hypofunction. Of these, 104 were correctly identified with a positive result in the test (true positives) and 11 obtained a negative test result (false negatives). With regard to these 11 false negatives, 5 were obtained in subjects with unilateral MD, 2 in bilateral MD, 2 in vestibular neuritis, and 2 in VS. Of the 80 subjects without hypofunction, 7 were correctly identified with a negative result in the test (true negatives) and 73 subjects obtained a positive result (false positive). The cross tabulation of the index test results by the results of the reference standard is shown in Table 1.

Table 1
Unilateral peripheral vestibular hypofunction detection by GVS in studied subjects

Reference Standard

Index test Present n Absent n Total

Positive True Positive 104 False Positive 73 177

Negative False negative 11 True Negative 7 18

Total 115 80 195

	Reference Standard
Positive	True Positive	104	False Positive	73	177
Negative	False negative	11	True Negative	7	18
Total		115		80	195

From these data, sensitivity, specificity, positive and negative predictive values, and positive and negative likelihood ratios were calculated, as well as 95% Confidence Intervals for every result. Table 2 shows the results obtained.

Table 2

Results for sensitivity, specificity, positive and negative predictive values, and positive and negative likelihood ratios in subjects tested with GVS.

	Presence of nystagmus
Statistic	Value	95% CI
Sensitivity	90.43%	83.53% to 95.13%
Specificity	8.75%	3.59% to 17.20%
Positive Likelihood Ratio	0.99	0.91 to 1.08
Negative Likelihood Ratio	1.09	0.44 to 2.70
Positive predictive value *	58.76%	56.56% to 60.92%
Negative predictive value *	38.89%	20.49% to 61.10%

^*Predictive values observed in one study do not apply universally. The predictive values of a test depend upon the prevalence of the target condition in the patients being tested.

In order to determine the discriminatory ability of GVS in each modality (i.e.: maximum SPV or Presence of nystagmus) a ROC curves analysis was performed. Both modalities were statistically non-different in discriminating ability compared with random discrimination. That is, the area under the ROC curve of each modality was close to 0.5 with non-statistically significant p values. This analysis can be seen in Table 3.

Table 3

Area Under the Curve (AUC) obtained by each modality in the ROC curves analysis

	Maximum SPV	Presence of nystagmus
Area under the ROC curve (AUC)	0.523	0.529
Standard Error	0.0294	0.0189
95% Confidence Interval	0.472 to 0.573	0.478 to 0.580
Significance level P (Area = 0.5)	0.439	0.125

Both curves are represented in Fig. 3. Although the Presence of nystagmus had an AUC greater than Maximum SPV, no statistically significant differences were found between them. The result of this pairwise comparison is shown in Table 4.

Fig. 3.

Receiver Operating Characteristic (ROC) curve analysis showing the performance of the two modalities used. Maximum SPV has an Area Under the Curve (AUC) of 0.523, and the Presence of nystagmus has an AUC of 0.529. None of these AUCs proved to be better than random selection (both p-values were >0.05).

Table 4

Pairwise comparison of ROC curves

Maximum SPV ∼ Presence of nystagmus
Difference between areas	0.00635
Standard Error	0.0226
95% Confidence Interval	–0.0379 to 0.0506
Significance level P	0.779

When tested with the index test, 2 subjects reported headache of a few minutes duration as the only adverse effect. No adverse effects were reported with the use of the reference standard.

4. Discussion

From the data in Table 1, it is evident that despite a high sensitivity of the index test, the very low specificity takes away much of the utility of GVS as a diagnostic test. This is a direct consequence of the large number of false positives that were obtained.

The most obvious findings to emerge from the analysis are the very poor likelihood ratios, both positive (0.99) and negative (1.09). Interestingly, a positive result in the index test reduces the probability of having the target condition, and vice versa, a negative result increases the probability of having it. These results are somewhat counterintuitive. This occurs because the index test has been positive in only 90.4% of subjects with the target condition, but has been positive in 91.2% of subjects without it. With these likelihood ratios, obtaining a positive result in our sample changes the probability of having the target condition from 58.9% (prevalence in our sample) to 58.4%. And conversely, obtaining a negative result changes the probability to 64.2%. This makes GVS clinically useless as a diagnostic test.

A Fagan Nomogram was made to display this lack of utility when it comes to using GVS in decision making (see Fig. 4).

Fig. 4.

Fagan Nomogram displaying how the probability of having the target condition changes according to the likelihood ratios in subjects tested with GVS.

Neither of the two modalities (i.e.: rely on SPV vs. rely on the presence of nystagmus) proved to have greater diagnostic ability than the other. But more important than this, neither modality proved to be better than random discrimination, which means that if a diagnosis were randomly assigned to each subject, a diagnostic accuracy similar to GVS would be observed (as can be seen in the two non-significant p values of both AUCs in Table 3).

The results of the present study should not be generalized to all possible uses of GVS. This study set out with the aim of assessing the diagnostic accuracy of GVS in detecting unilateral peripheral vestibular hypofunction. But GVS has been used successfully in a wide variety of experimental procedures. Ocular response to GVS has been used to assess remnant brain activity in comatose patients, and a correlation between null response to GVS and brain death has been established [23]. It has been observed that GVS simulates many of the characteristics of post-flight locomotive dysfunction observed in astronauts following short- or long-duration missions, so GVS has been suggested as part of the pre-astronaut training program [19]. And GVS has been successfully used to stimulate central neurological processes in experimental models of the velocity-storage mechanism [13]. As can be seen, GVS has been used in multiple areas, and more uses may be expected in the future for this stimulation tool. However, it has not proved useful as a diagnostic tool in subjects with vestibular pathologies.

4.1. Limitations

The population characteristics found in this sample may not be representative of the population at other centers where the case mix of dizzy patients may differ. Furthermore, only dizzy patients who presented without nystagmus were included, since GVS as a diagnostic test should be used in subjects in whom the physician has a diagnostic uncertainty. The videonystagmographic-recorded GVS was used, and cannot comment on the performance of the bedside examination. No adjustment was made to identify subjects with bilateral caloric hypofunction, because they represented only 2% of the sample. The absence of asymmetry in this group could have slightly changed the results.

In addition, a positive response was arbitrarily considered to be pathological. Formally, neither response nor absence of response should be considered pathological, since galvanic stimulation may produce responses that are more related to the intensity of the stimulus than to the patient’s disease. And this was exactly the objective of the study, to show that the information obtained from the presence of intrastimulus nystagmus is not useful to predict caloric deficit. A ‘positive’ response was considered pathological due to strictly statistical reasons: considering the positive response as pathological produces an AUC of 0.529, whereas if a ‘negative’ response were to be considered pathological, it would produce an AUC of 0.471 (even worse than random discrimination). Note that, in either of the two scenarios, the diagnostic capacity is negligible.

Finally, this study does not include a correlation between the degree of caloric deficit and the degree of galvanic response, however, it is unlikely that such a relationship would have been found, given that 59% of the patients had a caloric deficit, but only 12% presented a galvanic response in only one ear. Since the laterality analysis has already been done by MacDougall et al. [15], it was considered unnecessary to repeat it.

4.2. Conclusion

The present study confirms previous findings and provides evidence that GVS has no diagnostic ability to diagnose subjects with unilateral peripheral vestibular hypofunction. This is the first study to adequately investigate the diagnostic accuracy of GVS, and has confirmed a widely believed opinion among experts, its lack of usefulness. Overall, this study strengthens the idea that GVS should not be used for the diagnosis of subjects with vestibular hypofunction in the clinical setting.

Ethics committee approval

Ethics committee approval was received for this study from the ethics committee, reference number FPNT-CEIB-04A.

Informed consent

Written informed consent was obtained from subjects who participated in this study.

Conflict of interest

The author declares that he has no conflict of interest.

Financial disclosure

The author declares that this study has received no financial support.

Footnotes

Acknowledgments

None.

References

D.G.

Altman ,

Machin ,

T.N.

Bryant and

M.J.

Gardner , editors. Statistics with confidence. 2nd ed. Bristol: BMJ Books; 2000.

E.J.

Blonder and

Davis , The galvanic falling reaction in patients with verified intracranial neoplasms, JAMA 107 (1936), 411–412.

P.M.

Bossuyt ,

J.B.

Reitsma ,

D.E.

Bruns , et al., STARD 2015: An updated list of essential items for reporting diagnostic accuracy studies, BMJ 351 (2015), h5527.

Clopper and

E.S.

Pearson , The use of confidence or fiducial limits illustrated in the case of the binomial, Biometrika 26 (1934), 404–413.

Cohen ,

S.B.

Yakushin and

G.R.

Holstein , What Does Galvanic Vestibular Stimulation Actually Activate? Front Neurol 2 (2011), 90.

I.S.

Curthoys and

H.G.

MacDougall , What Galvanic Vestibular Stimulation Actually Activates, Front Neurol 3 (2012), 117.

E.R.

DeLong ,

D.M.

DeLong and

D.L.

Clarke-Pearson , Comparing the areas under two or more correlated receiver operating characteristic curves: A nonparametric approach, Biometrics 44(3) (1988), 837–845.

Fitzpatrick and

Day , Probing the human vestibular system with galvanic stimulation, J Appl Physiol 96(6) (2004), 2301–2316.

Goldberg ,

Fernandez and

Smith , Responses of vestibular nerve afferents in the squirrel monkey to externally applied galvanic currents, Brain Res 252(1) (1982), 156–160.

10.

Hajian-Tilaki , Sample size estimation in diagnostic test studies of biomedical informatics, J Biomed Inform 48 (2014), 193–204.

11.

Hitzig , Über die beim Galvanisieren des Kopfes entstehenden Störungen der Muskelinnervation und der Vorstellungen vom Verhalten im Raume[About perturbations of muscle innervation and imaginations of behavior in space during head galvanization], Reichert und du Bois-ReymondsArch (1871), 716–770.

12.

Kim and

I.S.

Curthoys , Responses of primary vestibular neurons to galvanic vestibular stimulation (GVS) in the anaesthetised guinea pig, Brain Res Bull 64(3) (2004), 265–271.

13.

Krstulovic ,

Atrache Al Attrache ,

Perez Garrigues ,

Argente-Escrig ,

Bataller Alberola and

Morera Perez , Increased Velocity Storage in Subjects with Meniere’s Disease. JIntAdv Otol 12(1) (2016), 87–91.

14.

MacDougall ,

Brizuela ,

Burgess and

Curthoys , Between-subject variability and within-subject reliability of the human eye-movement response to bilateral galvanic (DC) vestibular stimulation, Exp Brain Res 144 (2002), 69–78.

15.

H.G.

MacDougall ,

A.E.

Brizuela ,

A.M.

Burgess ,

I.S.

Curthoys and

G.M.

Halmagyi , Patient and normal three-dimensional eye-movement responses to maintained (DC) surface galvanic vestibular stimulation, Otol Neurotol 26(3) (2005), 500–511.

16.

MacDougall ,

Brizuela and

Curthoys , Linearity, symmetry and additivity of the human eye-movement response to maintained unilateral and bilateral surface galvanic (DC) vestibular stimulation, Exp Brain Res 148(2) (2003), 166–175.

17.

N.D.

Mercaldo ,

K.F.

Lau and

X.H.

Zhou , Confidence intervals for predictive values with an emphasis to case control studies, Statistics in Medicine 26 (2007), 2170–2183.

18.

C.E.

Metz , Basic principles of ROC analysis, Semin Nucl Med 8(4) (1978), 283–298.

19.

Moore ,

MacDougall ,

Peters ,

Bloomberg ,

Curthoys and

Cohen , Modeling locomotor dysfunction following spaceflight with Galvanic vestibular stimulation, Exp Brain Res 174(4) (2006), 647–659.

20.

C.R.

Pfaltz , The diagnostic importance of the galvanic test in otoneurology, Pract Otorhinolaryngol (Basel) 31(4) (1969), 193–203.

21.

Proctor ,

Glackin ,

Shimizu ,

Smith and

Liet-man , Reference values for serial vestibular testing, Ann Otol Rhinol Laryngol 95(1 Pt 1) (1986), 83–90.

22.

Purkyne , Commentatio de examine physiologico organi visus et systematis cutanei [Commentary on the Physiological Examination of the Visual Organs and the Cutaneous System]. 1st ed. Pragae: Vratislaviae; 1819, pp. 88.

23.

Schlosser ,

Ünterberg and

Clarke , Üsing video-oculography for galvanic evoked vestibulo-ocular monitoring in comatose patients, JNeurosciMethods 145(1–2) (2005), 127–131.

24.

D.L.

Simel ,

Rennie and

P.M.

Bossuyt , The STARD statement for reporting diagnostic accuracy studies: Application to the history and physical examination, J Gen Intern Med 23(6) (2008), 768–774.

	Reference Standard
Index test	Present	n	Absent	n	Total
Positive	True Positive	104	False Positive	73	177
Negative	False negative	11	True Negative	7	18
Total		115		80	195

Usefulness of galvanically induced nystagmus as a diagnostic test of peripheral vestibular hypofunction

Abstract

BACKGROUND:

OBJECTIVES:

METHODS:

RESULTS:

CONCLUSIONS:

Keywords

1. Introduction

2. Methods

2.1. Study design

2.2. Participants

2.3. Test methods

3. Results

3.1. Participants

Table 1 Unilateral peripheral vestibular hypofunction detection by GVS in studied subjects Reference Standard Index test Present n Absent n Total Positive True Positive 104 False Positive 73 177 Negative False negative 11 True Negative 7 18 Total 115 80 195

4.2. Conclusion

Ethics committee approval

Informed consent

Conflict of interest

Financial disclosure

Footnotes

Acknowledgments

References

Table 1
Unilateral peripheral vestibular hypofunction detection by GVS in studied subjects

Reference Standard

Index test Present n Absent n Total

Positive True Positive 104 False Positive 73 177

Negative False negative 11 True Negative 7 18

Total 115 80 195