- Email: [email protected]
Brainstem representation of vestibular evoked myogenic potentials
Clinical Neurophysiology 121 (2010) 1102–1108
Contents lists available at ScienceDirect
Clinical Neurophysiology journal homepage: www.elsevier.com/locate/clinph
Brainstem representation of vestibular evoked myogenic potentials Günther Heide a,b, Bettina Luft a, Jens Franke a, Peter Schmidt c, Otto W. Witte a, Hubertus Axer a,* a
Hans Berger Clinic for Neurology, University Hospital Jena, Friedrich-Schiller-University, Erlanger Allee 101, D-07747 Jena, Germany Klinik für Neurologie, Klinikum Meinigen GmbH, Bergstraße 3, D-98617 Meiningen, Germany c Institute of Diagnostic and Interventional Radiology, Department of Neuroradiology, Friedrich-Schiller-University Jena, Erlanger Allee 101, D-07747 Jena, Germany b
a r t i c l e
i n f o
Article history: Accepted 7 February 2010 Available online 3 March 2010 Keywords: Vestibular evoked myogenic potentials VEMPs Brainstem Vestibulocollic reflex
a b s t r a c t Objective: Vestibular evoked myogenic potentials (VEMPs) are caused by a short-latency reflex recorded from averaged electromyography from the sternocleidomastoid muscle evoked by intense auditory clicks. Besides peripheral vestibulopathy, abnormal VEMPs can be caused by lesions of the brainstem. The aim of this study was to analyze the topology of ischemic brain lesions generating pathological VEMPs. Methods: Twenty-nine patients with brainstem infarcts were prospectively studied using VEMPs and MR imaging to evaluate the brainstem representation of the VEMP reflex. Individual brainstem lesions were projected to a standard MR-dataset for normalization. Probabilistic lesion maps were calculated. A digital brainstem atlas was fitted to the lesion maps. Results: Twelve patients showed unilaterally abnormal VEMPs, 10 patients had normal VEMPs. Seven patients with bilaterally absent VEMPs were not analyzed. Most lesions were located in the lateral medulla oblongata involving the spinal accessory nerve. Most lesions in the pons were associated to anterolateral parts of pyramidal tract fibers. In a few cases, lesions were located in the tegmental area of the pons, including the vestibular nuclei. Conclusions: Abnormal VEMPs may be produced not only by peripheral vestibulopathy but also by brainstem lesions. VEMPs may be influenced by effects caused by lesions located above the level of the vestibular nuclei. Significance: This study adds to the knowledge of anatomical brainstem representation of VEMP. Ó 2010 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights reserved.
1. Introduction In the recent past, vestibular evoked myogenic potentials (VEMPs) have been used to detect peripheral vestibular damage (Welgampola and Colebatch, 2005). The method was initially described by Colebatch and Halmagyi in 1992 (Colebatch and Halmagyi, 1992; Colebatch et al., 1994) as click evoked myogenic potentials. An auditory click stimulus leads to a stimulation of the sacculus activating vestibular neurons, which in turn cause a modulation of tonic muscle activity of the sternocleidomastoid muscle. The measurement of the VEMPs requires tonic contraction of the muscle and can be observed after averaged EMG (electromyography) of the ipsilateral sternocleidomastoid muscle. The VEMP starts with a positive peak (p13) corresponding to an inhibition of the underlying motor unit firing (Colebatch and Rothwell, 2004) and then shows a secondary negative peak (n23). VEMPs are particularly relevant in the differential diagnosis of peripheral vestib* Corresponding author. Address: Department of Neurology, Friedrich-SchillerUniversity Jena, Erlanger Allee 101, D-07747 Jena, Germany. Tel.: +49 3641 9323454; fax: +49 3641 9323402. E-mail address: [email protected] (H. Axer).
ular lesions (Heide et al., 1999), e.g. Ménière disease (Kim-Lee et al., 2009), vestibular neuritis (Hong et al., 2008), bilateral vestibulopathy, or vestibular schwannomas (Welgampola and Colebatch, 2005). It has been shown recently, that cortical representation of saccular vestibular stimulation using VEMPs involves the multisensory cortical network within both hemispheres including the posterior insular cortex, the middle and superior temporal gyri, and the inferior parietal cortex (Schlindwein et al., 2008). The inferior parts of the temporo-parietal network seem to be mainly influenced by the ipsilaterality of the pathways, while the upper parts mainly reflect the predominance of vestibular processing of the non-dominant hemisphere (Janzen et al., 2008). Although it is known that lesions of the brainstem also lead to VEMP alterations (Deftereos et al., 2008), the brainstem pathway of VEMPs is still not fully understood. The latency of the earliest responses implies a rapidly conducting disynaptic (Colebatch et al., 1994) or oligosynaptic pathway with the vestibular nuclei as afferent part and the accessory nucleus as efferent part. It has been suggested that the reflex is mediated by ipsilateral descending projections via the lateral vestibulospinal tract (Colebatch et al., 1994). Another hypothesis favours a connection between vestibular
1388-2457/$36.00 Ó 2010 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.clinph.2010.02.007
G. Heide et al. / Clinical Neurophysiology 121 (2010) 1102–1108
nuclei and accessory nucleus via the medial vestibulospinal tract (Watson and Colebatch, 1998), which has been experimentally demonstrated to contain major otolith–sternocleidomastoid motoneuron pathways in the cat (Kushiro et al., 1999). To elucidate the brainstem structures involved in the human, we analyzed MR lesion patterns of brainstem strokes (Axer et al., 2007) generating alterations of VEMPs with regard to their topological distribution. 2. Methods 2.1. Subjects Patients with acute brainstem infarction admitted to our stroke unit from 2006 to 2008 were included in this study. The study was performed with the approval of the local ethics committee at our institution (0942-09/02, 2193-01/08), and written informed consent was obtained from each patient. Inclusion criteria were: first neurological deficit in the history, acute to subacute single brainstem infarction identified in diffusion weighted imaging (DWI), no other concomitant brain lesion; and no history of peripheral vestibulopathy and hearing loss. Exclusion criteria were: contraindication for MR (e.g. pacemaker), other brain lesions, history of peripheral vestibulopathy and/or hearing loss, the use of hearing aids, or unilateral abnormal wave I in AEP (acoustic evoked potentials). All patients underwent MR imaging at the admission to the emergency room and/or in the course of the first week after symptom onset. VEMPs and AEP were registered in the first week after onset of symptoms. 2.2. MR protocol MRI data were acquired on a 1.5 T scanner (Symphony, Siemens, Erlangen, Germany). For diffusion weighted imaging (DWI) a singleshot, multislice spin-echo, echo planar imaging (EPI) sequence was used. Diffusion gradients were applied in three orthogonal directions to generate DWI using the following parameters: TR/TE 4600/108 ms, Flip angle 90°, field of view (FoV) 230 230, matrix 128 128, b-values 0 and 1000 s/mm2, thickness 5 mm, gap 1.5 mm. The voxel size of the DWI sequence (1.8 1.8 5.0 mm3) meets the requirements for fast lesion detection in the acute stroke routine at our hospital. DWI was used to define a lesion as acute. Twenty-five DWI slices were acquired spanning the whole brain. Sections of the brainstem were visible on 12 slices. In addition, a high-resolution T2-weighted sequence of the brainstem was acquired using the following parameters: TR/TE 5060/99 ms, Flip angle 180°, FoV 167 210, matrix 408 512, thickness 3 mm, gap 0.3 mm, voxel size 0.41 0.41 3 mm3. This sequence comprised 25 slices specifically spanning the brainstem and was therefore used for registration of the lesion profiles to the ‘reference’ brain 2.3. Measurement of VEMPs Vestibular evoked myogenic potentials were acquired using Synergy 12.2 (VIASYS Healthcare UK Ltd.). The vestibulocollic reflex was evoked by rarefaction clicks with a loudness of 110 dB NHL (normal hearing level), duration of 0.1 ms, and frequency of 5 Hz. The loudness of the clicks was adjusted to 110 dB NHL in this study, because most studies used a loudness of 95 (Lee et al., 2008; Wang et al., 2008; Maes et al., 2009), 100 (Welgampola and Colebatch, 2001), or 110 dB NHL (Debatisse et al., 2005; Heide et al., 1999). Others used 140 dB SPL (sound pressure level) (Eleftheriadou et al., 2008). Since a click intensity of 95 dB NHL is equivalent to 140 dB SPL, the loudness of 110 dB NHL is slightly under the
1103
pain threshold and a higher loudness could possibly induce noise induced hearing loss. The click was delivered to one ear by a pair of headphones. A continuous white noise of 40 dB was applied to the contralateral ear. There was no adjustment to individual hearing loss. Surface EMG signals were recorded from the sternocleidomastoid muscles symmetrically on both sides with a reference placed on the clavicles. Patients were lying supine and had to lift their head during recording to reach tonic muscle contraction. Time of analysis was 100 ms. The signal was bandpass filtered from 20 Hz to 2 kHz. A total of 256 recordings was averaged, and the measurements were reproduced in a second turn. VEMP amplitudes can be influenced by EMG level (Akin et al., 2004). Thus, EMG target levels ranging from 300 to 500 lV have been suggested for clinical application of the VEMP, which has been used as a guidance level of EMG activity in our study. It has been paid attention to the symmetrical contraction of the sternocleidomastoid muscles in each patient using simultaneous bilateral EMG-measurements, so that asymmetric tonic muscle contraction could be prevented as amplitudes of VEMP of both sides were compared to each other. VEMPs were classified as abnormal when they were absent or if an asymmetry of amplitudes larger than 50% was found. 2.4. Calculation of probabilistic lesion maps 3D Slicer 3.2 (www.slicer.org) was used for image processing. It is an open-source, cross-platform application for visualizing and analyzing medical image data. 3D Slicer was developed by the Artificial Laboratory of the Massachusetts Institute of Technology and the Brigham and Womens Hospital, a teaching affiliate of Harvard Medical School (Gering et al., 2001). DWI and high-resolved T2w brainstem sequences of each patient were imported into the 3D Slicer. A high-resolved T2w dataset of a healthy brainstem was used as a ‘standard’ brain, where all lesion profiles were projected upon. Linear affine co-registration was used which is implemented in the Slicer software. This way, each lesion was transferred to a ‘reference’ coordinate system allowing for performance of group analyses. Each single lesion profile was projected onto the ‘standard’ brainstem and thereby normalized. The ischemic lesion was segmented manually and was saved for each patient. According to the results of VEMP measurements three groups were built: group 1 consisted of patients with unilaterally abnormal VEMPs (on the lesion side). Group 2 consisted of patients with normal VEMPs. Group 3 was constituted from patients with bilaterally absent VEMPs, so that a lesion dependency could not reliably be attributed in this group. Algorithms developed on the Matlab R2008b environment (Mathworks Inc., Natick, MA) were used for group analysis of the lesion data. Therefore, lesion data from patients of group 1 and group 2 were used to calculate ‘probabilistic’ lesion maps with the number of lesions in each voxel relative to the number of patients in each group in percentage. It can be assumed that lesion profiles of group 1 may have an impact to produce VEMP pathology, while lesion profiles of group 2 should not be associated to pathological VEMPs. Therefore, a difference-map between probabilistic lesion maps of group 1 and group 2 was calculated. Negative values in the difference-map correspond to lesions which are not associated to VEMP pathology while positive values may be capable to produce pathological VEMPs. The lesion maps were imported into 3D Slicer for visualization and analysis. 2.5. Correlation of lesion maps and anatomy For correlation of lesion data to anatomic structures a digitized atlas based on the ‘Atlas of the Human Brainstem’ (Paxinos and
1104
G. Heide et al. / Clinical Neurophysiology 121 (2010) 1102–1108
Huang, 1995) was used. The 64 templates of the Paxinos atlas were digitized. Afterwards the coordinate system existent on each template of the atlas was used to define reference landmarks. Matlab routines were used to perform a landmark registration of each template according to the coordinate system of the atlas using scaling, rotation, and translations. After this registration process, the image data was imported into the 3D Slicer, where the anatomical structures of the atlas were segmented manually, so that a 3D digitized atlas of the brainstem was created consisting of the registered Paxinos templates and segmented 3D structures of the brainstem. This brainstem atlas was projected onto the ‘standard’ brainstem with the lesion maps using 3D Slicer. For this, linear affine co-registration was used as implemented in the Slicer Software. The atlas data containing 64 slices was fitted to MR brainstem dataset with the associated lesion maps. These MR-data spanned the whole human brainstem with 26 slices of 408 512 pixel. Anatomical landmarks for registration were surface structures of the brainstem (e.g. ventricular floor line, pontomesencephalic junction, pontomedullary junction, etc.) and anatomical structures in the brainstem visible on T2w MRI (e.g. red nucleus, pyramis of the medulla oblongata, etc.). This way, the lesion maps could be correlated to anatomical data, so that all histological structures contained in the lesions could be identified. 3. Results Twenty-nine patients with acute brainstem infarcts were enrolled into the study (Table 1), while about 80 patients with brainstem ischemia were screened for inclusion into the study. However, most of these patients were relatively old of age and had other concomitant brain lesions (e.g. former stroke, subcortical arteriosclerotic encephalopathy, and others). Many patients complained peripheral vestibulopathy and/or hearing loss in the past, many patients used hearing aids and could therefore not be enrolled into the study. Twelve patients (group 1) showed unilaterally abnormal VEMPs on the lesion side of the brainstem infarction. Differences in p13 latencies between right and left side only occurred in 1 case (case #6) in addition to amplitude reduction. A separate behavior of positive (p13) and negative waves (n23) could not be found. Wave I of
the acoustic evoked potentials (AEP) was measured as a parameter of cochlear function. All 12 patients of group 1 showed bilaterally symmetrical latencies of wave I (delayed in 4 and normal in 8 patients, Table 1). Fig. 1 gives an example of a patient with abnormal VEMPs on the left hand side, who had an ischemic lesion in the left medulla oblongata. Ten patients (group 2) had normal VEMPs. Group 3 consisted of 7 patients with bilaterally absent VEMPs, which could not definitely be attributed to the brainstem infarction. The lesion profiles of these patients did not show a systematic lesion anatomy. Therefore, these patients were not further analyzed regarding lesion anatomy. The lesion maps of group 1 and group 2 are shown in Fig. 2. Unilaterally abnormal VEMPs were produced by lesions reaching from the mesencephalon down to the medulla oblongata. However, after subtraction of the lesion maps of group 1 and group 2 a more specific lesion pattern could be found. One region capable to produce pathological VEMPs is the lateral medulla oblongata in contrast to the classical Wallenberg infarct (Fitzek et al., 2006) located more dorsally in the lateral medulla oblongata. Another lesion site is the anterolateral part of the pyramidal tract fibers mainly located in the upper pons. In fewer cases lesions were found in the tegmentum of the pons, in one case only the vestibular nuclear group was lesioned. The correlation of the lesions to anatomical structures was done by the projection of the brainstem atlas to the lesion maps. Figs. 3 and 4 summarize the anatomical structures identified to be contained in the lesion maps which may be attributed to pathological VEMPs. In the medulla oblongata the spinal accessory nerve (n11) can be attributed to a disturbance of the efferent part of the VEMP reflex. In addition, ventral and dorsal spinocerebellar tracts, spinothalamic tract, and most parts of the medullar trigeminal system were included into the medullary lesion profile. It has to be noted that neither the ventral nor the lateral vestibulospinal tract was included in the lesion area in the medulla oblongata. In the pons most lesions were associated to anterolateral parts of the pontine pyramidal tract fibers. In fewer cases lesions of the tegmental area of the pons, but also the trapezoid body and medial lemniscus, as well as the vestibular nuclei were found. No significant weakness in the sternocleidomastoid muscles was noted in group 1 and group 2. To rule out the possibility that
Fig. 1. Example of a patient with unilaterally abnormal VEMPs. (A) Normal VEMP at the right side. (B) Missing VEMP at the left side. (C) Hyperintense DWI-signal in the left lateral medulla oblongata as a sign of acute brainstem ischemia. (D) Lesion in T2w MRI. (E) Ischemic lesion as 3D volume projected onto the right side of the reference brainstem.
G. Heide et al. / Clinical Neurophysiology 121 (2010) 1102–1108
1105
Fig. 2. Probabilistic lesion maps. The left column shows lesion profiles of patients with unilaterally pathological VEMPs (group 1). The middle column shows lesion profiles of patients with normal VEMPs (group 2). The right column shows a difference-map between group 1 and group 2. Note that lesions associated to normal VEMPs are visualized as cold colors and lesions associated to pathological VEMPs are visualized as warm colors. Color coding is shown in the right lower corner.
a different activation strategy due to muscle weakness could result in different VEMPs, the NIHSS scores for facial and arm motor function were compared between group 1 and 2 using the Mann–Whitney U-test. No statistically significant difference between these groups could be found in the facial (p = 0.974) and arm (p = 0.497) motor scores of the NIHSS. 4. Discussion The analysis of VEMPs in diseases of the brainstem has been suggested several times (Chen and Young, 2003). It has been reported that VEMPs may be abnormal in patients with multiple sclerosis (Patkó et al., 2007; Versino et al., 2007) and basilar artery migraine (Liao and Young, 2004). However, these are diseases which cause disseminated and not well circumscribed brainstem dysfunctions. Thus, it seems not to be reliable to study these patients for anatomical localization. Besides single case studies (Deftereos et al., 2006; Rosengren et al., 2007; Shin et al., 2009) describing VEMP abnormalities due to circumscribed brainstem lesions, only few group studies have been performed so far (Itoh et al., 2001; Deftereos et al., 2008). They concluded that VEMPs
may be a useful diagnostic tool to identify lower brainstem lesions especially in the lateral lower pons and the upper medulla oblongata. In contrast, Pollak et al. (2006) demonstrated that cerebellar strokes did not influence VEMPs (19 patients), but they also could not find statistically significant differences between 15 patients with lower brainstem ischemic stroke and 53 normal controls. However, they found abnormal VEMPs in single patients with brainstem lesions while other patients had normal VEMPs. Thus, for localization purposes the lesion profiles of patients with normal VEMPs have to be compared to patients with abnormal VEMPs, because lesion profiles of patients with normal VEMPs should provide information about anatomical structures which seem not to be involved in the VEMP reflex. Moreover, the use of an anatomical atlas as done in this study helps to identify anatomical structures in higher detail than using MR-data only. However, it has to be kept in mind that structures identified in lesion studies not necessarily may be correlated to the functional disturbance under study, because these may be included in the lesion volume caused by random inclusion into vascular territories. Therefore, anatomical plausibility has to be
1106
G. Heide et al. / Clinical Neurophysiology 121 (2010) 1102–1108
Fig. 3. Anatomical structures located in the lesion profiles of the medulla oblongata as derived from the atlas.
Fig. 4. Anatomical structures located in the lesion profiles of the pons as derived from the atlas.
1107
G. Heide et al. / Clinical Neurophysiology 121 (2010) 1102–1108
proven regarding a possible correlation between lesion anatomy and clinical disturbance. Bilaterally missing VEMPs were found in 7 patients. It is known that amplitudes and latency of VEMPs are dependent on age (Welgampola and Colebatch, 2001; Basta et al., 2007) and bilaterally absent VEMPs may be found in normal subjects over age 60 (Welgampola and Colebatch, 2005). Such changes may be due to age-related hair cell loss or other age-related changes of the vestibular system. Therefore, patients with bilaterally absent VEMPs had to be excluded from lesion analysis. The lesion profiles of these 7 patients did not show a systematic lesion anatomy making it very unlikely that a circumscribed central lesion could cause bilaterally absent VEMPs. Mostly, peripheral vestibular dysfunction is the major source for abnormal VEMPs. Although the history of peripheral vestibulopathy or hearing loss was an exclusion criterion in this study, it may be a limitation of the study that these functions have not been specifically tested (e.g. using audiometric measurements and calo-
ric tests). Bilateral symmetrical latencies of the AEP in the patients with asymmetrical VEMP pathology at least did not reveal an asymmetric dysfunction of the peripheral auditory system. Moreover, the lesions shown in MRI may include functional penumbra since all tests were done in the first week of stroke. This may be another limitation of the study. We demonstrated that VEMPs may be pathological in brainstem lesions not only being confined to the lower brainstem but reaching from medulla oblongata to the mesencephalon. The afferent part of the reflex is conducted by the vestibular nuclei at the level of the lateral lower pons and the efferent part by the accessory nucleus at the level of the medulla oblongata (Colebatch et al., 1994). Consequently, we found patients with lesions of the vestibular nuclei at the level of the lateral lower pons or lesions of the spinal accessory nerve in the lateral medulla oblongata. There is little information about the anatomy of the lateral vestibulospinal tract in man (Nathan et al., 1996; Brodal, 1981). However, the effects of the lateral vestibulospinal tract projecting from
Table 1 Patients enrolled in the study, clinical data, lesion location of brainstem infarcts, and VEMP pathology. VEMPs
Patients
Age
Gender
NIHSS
Lesion location
VEMP pathology
AEP wave I
Unilaterally pathological VEMPs
1
55
M
1
Absent
2
71
M
1
Right paramedian pons, tegmentum Left paramedian pons
3
66
M
3
Left dorsolateral medulla
Absent
4
50
M
5
Left paramedian pons
Amplitude reduction > 70%
5
59
M
2
Right dorsolateral medulla
Absent
6
65
M
4
7
61
M
1
Amplitude reduction > 50%, latency 20.9 ms Absent
8
68
M
4
Right lower paramedian pons, tegmentum Right paramedian pons, tegmentum Right dorsolateral medulla
Amplitude reduction > 80%
9
68
M
3
Left dorsolateral medulla
Absent
10
80
F
2
Right paramedian pons
Absent
11
63
M
1
Absent
12
54
M
1
Left paramedian pons, tegmentum Left paramedian pons
Wave I bilaterally delayed Wave I bilaterally delayed Bilaterally normal wave I Bilaterally normal wave I Bilaterally normal wave I Bilaterally normal wave I Wave I bilaterally delayed Bilaterally normal wave I Bilaterally normal wave I Bilaterally normal wave I Wave I bilaterally delayed Bilaterally normal wave I
n = 12 (41.4%)
Mean 63.3 ± 8.2
11 M, 1F
Mean 2.3 ± 1.4
13
65
M
1
14 15 16 17 18
66 68 71 63 68
M F M F M
3 1 8 1 1
19 20 21 22 n = 10 (34.5%)
62 68 67 70 Mean 66.8 ± 2.9
M M F M 7 M, 3 F
2 3 5 7 Mean 3.2 ± 2.6
23
55
M
3
Right dorsolateral medulla
Absent
24 25 26 27 28 29 n=7 (24.1%)
75 77 55 70 55 45 Mean 57.2 ± 12.2
M F F F F M 3 M, 4 F
6 4 3 1 4 2 Mean 3.6 ± 1.6
Left paramedian pons Left paramedian pons Right paramedian pons Right basal mesencephalon Right paramedian pons Right paramedian pons
Absent Absent Absent Absent Absent Absent
Normal VEMPs
Bilaterally absent VEMPs
Absent
Amplitude reduction > 75%
Left paramedian pons, tegmentum Right dorsolateral medulla Left dorsolateral medulla Right paramedian pons Left paramedian pons Left paramedian mesencephalon Left dorsolateral medulla Right paramedian pons Left paramedian pons Right paramedian pons
1108
G. Heide et al. / Clinical Neurophysiology 121 (2010) 1102–1108
the lateral vestibular nucleus of Deiters to the spinal cord are mainly excitatory. Consequently, it has been hypothesized recently that the medial vestibulospinal tract contains the pathway responsible for the inhibitory effects of the VEMP reflex (Shin et al., 2009; Kushiro et al., 1999). The medial vestibulospinal tract originates from the other three vestibular nuclei and runs within the medial longitudinal fascicle (Nieuwenhuys and Voogd, 1988). However, in our patient group no lesion of the medial longitudinal fascicle could be demonstrated in the lower brainstem – only in the midpons above the level of the vestibular nuclei. Although it has been suggested on purely anatomical considerations that only lesions of the lower pons and upper medulla oblongata may be capable to produce disturbances of VEMPs (Welgampola et al., 2005), we also demonstrated that lesions of the upper pons with pyramidal tract lesion in its upper ventrolateral part is capable to modify the VEMP reflex. This has not been described in the literature so far. The localization of the lesion in the ventrolateral (and not dorsomedial) part of the basis of the pons corresponds well to the localization of corticobulbar fibers (Kataoka et al., 2003) and corticofacial fibers (Urban et al., 2001). There could be a potential interaction in significant lesion location in the sense that patients with lesions in lower pons/vestibular nuclei level also could have lesions stretched up to the upper pons pyramidal tract level. However, 4 patients had upper paramedian pons infarction without additional infarction of the lower pons and vestibular nuclei (patients #2, 4, 10, 12, Table 1), which excludes the possibility that this area could randomly be included into the lesion profiles. In addition, none of the patients had multiple lesions, they all had single lacunar brainstem infarction. Thus, it may be hypothesized that abnormal VEMPs can also be caused by lesioned projections of the pyramidal tract. The hypothesis that a different activation strategy due to muscle weakness could result in VEMP alteration was contradicted by the observation that no significant differences between NIHSS scores of facial and arm motor function could be found between patients with pathological and normal VEMPs. A similar ‘supravestibular’ pathomechanism may be attributed to an influence from the pontine tegmentum above the level of the vestibular nuclei also including the medial longitudinal fascicle. In conclusion, abnormal VEMPs can be produced by lesions of the brainstem reaching from the medulla oblongata to the mesencephalon. Main foci of lesions capable to produce abnormal VEMPs are the lateral medulla oblongata (including the 11th nerve) and the ventrolateral basis of the upper pons (including parts of the pyramidal tract), but also the tegmentum of the lower pons (including the vestibular nuclei). It has to be considered from a clinical point of view that abnormal VEMPs may be produced not only by peripheral vestibulopathy but also by brainstem lesions. Moreover, VEMPs may be influenced by effects caused by lesions located above the level of the vestibular nuclei. References Akin FW, Murnane OD, Panus PC, Caruthers SK, Wilkinson AE, Proffitt TM. The influence of voluntary tonic EMG level on the vestibular-evoked myogenic potential. J Rehab Res Dev 2004;41:473–80. Axer H, Grässel D, Brämer D, Fitzek S, Kaiser WA, Witte OW, et al. Time course of diffusion imaging in acute brainstem infarcts. J Magn Reson Imaging 2007;26:905–12. Basta D, Todt I, Ernst A. Characterization of age-related changes in vestibular evoked myogenic potentials. J Vestib Res 2007;17:93–8. Brodal A. Neurological anatomy in relation to clinical medicine. New York: Oxford University Press; 1981. Chen C, Young Y. Vestibular evoked myogenic potentials in brainstem stroke. Laryngoscope 2003;113:990–3. Colebatch JG, Halmagyi GM. Vestibular evoked potentials in human neck muscles before and after unilateral vestibular deafferentation. Neurology 1992;42:1635–6.
Colebatch JG, Halmagyi GM, Skuse NF. Myogenic potentials generated by a clickevoked vestibulocollic reflex. J Neurol Neurosurg Psychiatr 1994;57:190–7. Colebatch JG, Rothwell JC. Motor unit excitability changes mediating vestibulocollic reflexes in the sternocleidomastoid muscle. Clin Neurophysiol 2004;115:2567–73. Debatisse D, Pralong E, Guerit JM, Bisdorff A. Recording click-evoked myogenic potentials (CEMPs) with a setup for brainstem auditory evoked potentials (BAEPs). Neurophysiol Clin 2005;35:109–17. Deftereos SN, Panagopoulos G, Eleftheriadou A, Korres S, Georgonikou D, Kandiloros D, et al. Using vestibular evoked myogenic potentials to localise brainstem lesions. A preliminary report. B-ENT 2008;4:215–9. Deftereos SN, Panagopoulos G, Gryllia M, Georgonikou D, Polyzoi M, Kechagias E, et al. Neurophysiological monitoring of brainstem function in a patient with Wallenberg syndrome, using Vestibular Evoked Myogenic Potentials. Neurol Neurophysiol Neurosci 2006:3. Eleftheriadou A, Deftereos SN, Zarikas V, Panagopoulos G, Korres S, Sfetsos S, et al. Vestibular evoked myogenic potential eliciting in normal subjects: comparison of four different methods. J Otolaryngol Head Neck Surg 2008;37:704–11. Fitzek S, Baumgärtner U, Marx J, Joachimski F, Axer H, Witte OW, et al. Pain and itch in Wallenberg’s syndrome: anatomical–functional correlations. Suppl Clin Neurophysiol 2006;58:187–94. Gering DT, Nabavi A, Kikinis R, Hata N, O’Donnell LJ, Grimson WE, et al. An integrated visualization system for surgical planning and guidance using image fusion and an open MR. J Magn Reson Imaging 2001;13:967–75. Heide G, Freitag S, Wollenberg I, Iro H, Schimrigk K, Dillmann U. Click evoked myogenic potentials in the differential diagnosis of acute vertigo. J Neurol Neurosurg Psychiatry 1999;66:787–90. Hong SM, Yeo SG, Kim SW, Cha CI. The results of vestibular evoked myogenic potentials, with consideration of age-related changes, in vestibular neuritis, benign paroxysmal positional vertigo, and Meniere’s disease. Acta Otolaryngol 2008;128:861–5. Itoh A, Kim YS, Yoshioka K, Kanaya M, Enomoto H, Hiraiwa F, et al. Clinical study of vestibular-evoked myogenic potentials and auditory brainstem responses in patients with brainstem lesions. Acta Otolaryngol Suppl 2001;545:116–9. Janzen J, Schlindwein P, Bense S, Bauermann T, Vucurevic G, Stoeter P, et al. Neural correlates of hemispheric dominance and ipsilaterality within the vestibular system. NeuroImage 2008;42:1508–18. Kataoka S, Miaki M, Saiki M, Saiki S, Yamaya Y, Hori A, et al. Rostral lateral pontine infarction: neurological/topographical correlations. Neurology 2003;61:114–7. Kim-Lee Y, Ahn JH, Kim YK, Yoon TH. Tone burst vestibular evoked myogenic potentials: diagnostic criteria in patients with Meniere’s disease. Acta Otolaryngol 2009;129:924–8. Kushiro K, Zakir M, Ogawa Y, Sato H, Uchino Y. Saccular and utricular inputs to sternocleidomastoid motoneurons of decerebrate cats. Exp Brain Res 1999;126:410–6. Lee SK, Cha CI, Jung TS, Park DC, Yeo SG. Age-related differences in parameters of vestibular evoked myogenic potentials. Acta Otolaryngol 2008;128:66–72. Liao L, Young Y. Vestibular evoked myogenic potentials in basilar artery migraine. Laryngoscope 2004;114:1305–9. Maes L, Vinck BM, De Vel E, D’haenens W, Bockstael A, Keppler H, et al. The vestibular evoked myogenic potential: a test-retest reliability study. Clin Neurophysiol 2009;120:594–600. Nathan PW, Smith M, Deacon P. Vestibulospinal, reticulospinal and descending propriospinal nerve fibres in man. Brain 1996;119:1809–33. Nieuwenhuys R, Voogd J, Van Huijzen C. The human central nervous system. Berlin: Springer; 1988. Patkó T, Simó M, Arányi Z. Vestibular click-evoked myogenic potentials: sensitivity and factors determining abnormality in patients with multiple sclerosis. Mult Scler 2007;13:193–8. Paxinos G, Huang X. Atlas of the human brainstem. San Diego: Academic Press; 1995. Pollak L, Kushnir M, Stryjer R. Diagnostic value of vestibular evoked myogenic potentials in cerebellar and lower-brainstem strokes. Neurophysiol Clin 2006;36:227–33. Rosengren SM, Nogajski JH, Cremer PD, Colebatch JG. Delayed vestibular evoked responses to the eyes and neck in a patient with an isolated brainstem lesion. Clin Neurophysiol 2007;118:2112–6. Schlindwein P, Mueller M, Bauermann T, Brandt T, Stoeter P, Dieterich M. Cortical representation of saccular vestibular stimulation: VEMPs in fMRI. NeuroImage 2008;39:19–31. Shin J, Song H, Koo J, Lee H, Kim JS. Abnormal vestibular evoked myogenic potentials in medial medullary infarction. J Clin Neurol 2009;5:101–3. Urban PP, Wicht S, Vucorevic G, Fitzek S, Marx J, Thömke F, et al. The course of corticofacial projections in the human brainstem. Brain 2001;124:1866–76. Versino M, Ranza L, Colnaghi S, Alloni R, Callieco R, Romani A, et al. The N3 potential compared to sound and galvanic vestibular evoked myogenic potential in healthy subjects and in multiple sclerosis patients. J Vestib Res 2007;17: 39–46. Wang S, Yeh T, Chang C, Young Y. Consistent latencies of vestibular evoked myogenic potentials. Ear Hear 2008;29:923–9. Watson SR, Colebatch JG. Vestibulocollic reflexes evoked by short-duration galvanic stimulation in man. J Physiol (Lond) 1998;513:587–97. Welgampola MS, Colebatch JG. Vestibulocollic reflexes: normal values and the effect of age. Clin Neurophysiol 2001;112:1971–9. Welgampola MS, Colebatch JG. Characteristics and clinical applications of vestibular-evoked myogenic potentials. Neurology 2005;64:1682–8.
Contents lists available at ScienceDirect
Clinical Neurophysiology journal homepage: www.elsevier.com/locate/clinph
Brainstem representation of vestibular evoked myogenic potentials Günther Heide a,b, Bettina Luft a, Jens Franke a, Peter Schmidt c, Otto W. Witte a, Hubertus Axer a,* a
Hans Berger Clinic for Neurology, University Hospital Jena, Friedrich-Schiller-University, Erlanger Allee 101, D-07747 Jena, Germany Klinik für Neurologie, Klinikum Meinigen GmbH, Bergstraße 3, D-98617 Meiningen, Germany c Institute of Diagnostic and Interventional Radiology, Department of Neuroradiology, Friedrich-Schiller-University Jena, Erlanger Allee 101, D-07747 Jena, Germany b
a r t i c l e
i n f o
Article history: Accepted 7 February 2010 Available online 3 March 2010 Keywords: Vestibular evoked myogenic potentials VEMPs Brainstem Vestibulocollic reflex
a b s t r a c t Objective: Vestibular evoked myogenic potentials (VEMPs) are caused by a short-latency reflex recorded from averaged electromyography from the sternocleidomastoid muscle evoked by intense auditory clicks. Besides peripheral vestibulopathy, abnormal VEMPs can be caused by lesions of the brainstem. The aim of this study was to analyze the topology of ischemic brain lesions generating pathological VEMPs. Methods: Twenty-nine patients with brainstem infarcts were prospectively studied using VEMPs and MR imaging to evaluate the brainstem representation of the VEMP reflex. Individual brainstem lesions were projected to a standard MR-dataset for normalization. Probabilistic lesion maps were calculated. A digital brainstem atlas was fitted to the lesion maps. Results: Twelve patients showed unilaterally abnormal VEMPs, 10 patients had normal VEMPs. Seven patients with bilaterally absent VEMPs were not analyzed. Most lesions were located in the lateral medulla oblongata involving the spinal accessory nerve. Most lesions in the pons were associated to anterolateral parts of pyramidal tract fibers. In a few cases, lesions were located in the tegmental area of the pons, including the vestibular nuclei. Conclusions: Abnormal VEMPs may be produced not only by peripheral vestibulopathy but also by brainstem lesions. VEMPs may be influenced by effects caused by lesions located above the level of the vestibular nuclei. Significance: This study adds to the knowledge of anatomical brainstem representation of VEMP. Ó 2010 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights reserved.
1. Introduction In the recent past, vestibular evoked myogenic potentials (VEMPs) have been used to detect peripheral vestibular damage (Welgampola and Colebatch, 2005). The method was initially described by Colebatch and Halmagyi in 1992 (Colebatch and Halmagyi, 1992; Colebatch et al., 1994) as click evoked myogenic potentials. An auditory click stimulus leads to a stimulation of the sacculus activating vestibular neurons, which in turn cause a modulation of tonic muscle activity of the sternocleidomastoid muscle. The measurement of the VEMPs requires tonic contraction of the muscle and can be observed after averaged EMG (electromyography) of the ipsilateral sternocleidomastoid muscle. The VEMP starts with a positive peak (p13) corresponding to an inhibition of the underlying motor unit firing (Colebatch and Rothwell, 2004) and then shows a secondary negative peak (n23). VEMPs are particularly relevant in the differential diagnosis of peripheral vestib* Corresponding author. Address: Department of Neurology, Friedrich-SchillerUniversity Jena, Erlanger Allee 101, D-07747 Jena, Germany. Tel.: +49 3641 9323454; fax: +49 3641 9323402. E-mail address: [email protected] (H. Axer).
ular lesions (Heide et al., 1999), e.g. Ménière disease (Kim-Lee et al., 2009), vestibular neuritis (Hong et al., 2008), bilateral vestibulopathy, or vestibular schwannomas (Welgampola and Colebatch, 2005). It has been shown recently, that cortical representation of saccular vestibular stimulation using VEMPs involves the multisensory cortical network within both hemispheres including the posterior insular cortex, the middle and superior temporal gyri, and the inferior parietal cortex (Schlindwein et al., 2008). The inferior parts of the temporo-parietal network seem to be mainly influenced by the ipsilaterality of the pathways, while the upper parts mainly reflect the predominance of vestibular processing of the non-dominant hemisphere (Janzen et al., 2008). Although it is known that lesions of the brainstem also lead to VEMP alterations (Deftereos et al., 2008), the brainstem pathway of VEMPs is still not fully understood. The latency of the earliest responses implies a rapidly conducting disynaptic (Colebatch et al., 1994) or oligosynaptic pathway with the vestibular nuclei as afferent part and the accessory nucleus as efferent part. It has been suggested that the reflex is mediated by ipsilateral descending projections via the lateral vestibulospinal tract (Colebatch et al., 1994). Another hypothesis favours a connection between vestibular
1388-2457/$36.00 Ó 2010 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.clinph.2010.02.007
G. Heide et al. / Clinical Neurophysiology 121 (2010) 1102–1108
nuclei and accessory nucleus via the medial vestibulospinal tract (Watson and Colebatch, 1998), which has been experimentally demonstrated to contain major otolith–sternocleidomastoid motoneuron pathways in the cat (Kushiro et al., 1999). To elucidate the brainstem structures involved in the human, we analyzed MR lesion patterns of brainstem strokes (Axer et al., 2007) generating alterations of VEMPs with regard to their topological distribution. 2. Methods 2.1. Subjects Patients with acute brainstem infarction admitted to our stroke unit from 2006 to 2008 were included in this study. The study was performed with the approval of the local ethics committee at our institution (0942-09/02, 2193-01/08), and written informed consent was obtained from each patient. Inclusion criteria were: first neurological deficit in the history, acute to subacute single brainstem infarction identified in diffusion weighted imaging (DWI), no other concomitant brain lesion; and no history of peripheral vestibulopathy and hearing loss. Exclusion criteria were: contraindication for MR (e.g. pacemaker), other brain lesions, history of peripheral vestibulopathy and/or hearing loss, the use of hearing aids, or unilateral abnormal wave I in AEP (acoustic evoked potentials). All patients underwent MR imaging at the admission to the emergency room and/or in the course of the first week after symptom onset. VEMPs and AEP were registered in the first week after onset of symptoms. 2.2. MR protocol MRI data were acquired on a 1.5 T scanner (Symphony, Siemens, Erlangen, Germany). For diffusion weighted imaging (DWI) a singleshot, multislice spin-echo, echo planar imaging (EPI) sequence was used. Diffusion gradients were applied in three orthogonal directions to generate DWI using the following parameters: TR/TE 4600/108 ms, Flip angle 90°, field of view (FoV) 230 230, matrix 128 128, b-values 0 and 1000 s/mm2, thickness 5 mm, gap 1.5 mm. The voxel size of the DWI sequence (1.8 1.8 5.0 mm3) meets the requirements for fast lesion detection in the acute stroke routine at our hospital. DWI was used to define a lesion as acute. Twenty-five DWI slices were acquired spanning the whole brain. Sections of the brainstem were visible on 12 slices. In addition, a high-resolution T2-weighted sequence of the brainstem was acquired using the following parameters: TR/TE 5060/99 ms, Flip angle 180°, FoV 167 210, matrix 408 512, thickness 3 mm, gap 0.3 mm, voxel size 0.41 0.41 3 mm3. This sequence comprised 25 slices specifically spanning the brainstem and was therefore used for registration of the lesion profiles to the ‘reference’ brain 2.3. Measurement of VEMPs Vestibular evoked myogenic potentials were acquired using Synergy 12.2 (VIASYS Healthcare UK Ltd.). The vestibulocollic reflex was evoked by rarefaction clicks with a loudness of 110 dB NHL (normal hearing level), duration of 0.1 ms, and frequency of 5 Hz. The loudness of the clicks was adjusted to 110 dB NHL in this study, because most studies used a loudness of 95 (Lee et al., 2008; Wang et al., 2008; Maes et al., 2009), 100 (Welgampola and Colebatch, 2001), or 110 dB NHL (Debatisse et al., 2005; Heide et al., 1999). Others used 140 dB SPL (sound pressure level) (Eleftheriadou et al., 2008). Since a click intensity of 95 dB NHL is equivalent to 140 dB SPL, the loudness of 110 dB NHL is slightly under the
1103
pain threshold and a higher loudness could possibly induce noise induced hearing loss. The click was delivered to one ear by a pair of headphones. A continuous white noise of 40 dB was applied to the contralateral ear. There was no adjustment to individual hearing loss. Surface EMG signals were recorded from the sternocleidomastoid muscles symmetrically on both sides with a reference placed on the clavicles. Patients were lying supine and had to lift their head during recording to reach tonic muscle contraction. Time of analysis was 100 ms. The signal was bandpass filtered from 20 Hz to 2 kHz. A total of 256 recordings was averaged, and the measurements were reproduced in a second turn. VEMP amplitudes can be influenced by EMG level (Akin et al., 2004). Thus, EMG target levels ranging from 300 to 500 lV have been suggested for clinical application of the VEMP, which has been used as a guidance level of EMG activity in our study. It has been paid attention to the symmetrical contraction of the sternocleidomastoid muscles in each patient using simultaneous bilateral EMG-measurements, so that asymmetric tonic muscle contraction could be prevented as amplitudes of VEMP of both sides were compared to each other. VEMPs were classified as abnormal when they were absent or if an asymmetry of amplitudes larger than 50% was found. 2.4. Calculation of probabilistic lesion maps 3D Slicer 3.2 (www.slicer.org) was used for image processing. It is an open-source, cross-platform application for visualizing and analyzing medical image data. 3D Slicer was developed by the Artificial Laboratory of the Massachusetts Institute of Technology and the Brigham and Womens Hospital, a teaching affiliate of Harvard Medical School (Gering et al., 2001). DWI and high-resolved T2w brainstem sequences of each patient were imported into the 3D Slicer. A high-resolved T2w dataset of a healthy brainstem was used as a ‘standard’ brain, where all lesion profiles were projected upon. Linear affine co-registration was used which is implemented in the Slicer software. This way, each lesion was transferred to a ‘reference’ coordinate system allowing for performance of group analyses. Each single lesion profile was projected onto the ‘standard’ brainstem and thereby normalized. The ischemic lesion was segmented manually and was saved for each patient. According to the results of VEMP measurements three groups were built: group 1 consisted of patients with unilaterally abnormal VEMPs (on the lesion side). Group 2 consisted of patients with normal VEMPs. Group 3 was constituted from patients with bilaterally absent VEMPs, so that a lesion dependency could not reliably be attributed in this group. Algorithms developed on the Matlab R2008b environment (Mathworks Inc., Natick, MA) were used for group analysis of the lesion data. Therefore, lesion data from patients of group 1 and group 2 were used to calculate ‘probabilistic’ lesion maps with the number of lesions in each voxel relative to the number of patients in each group in percentage. It can be assumed that lesion profiles of group 1 may have an impact to produce VEMP pathology, while lesion profiles of group 2 should not be associated to pathological VEMPs. Therefore, a difference-map between probabilistic lesion maps of group 1 and group 2 was calculated. Negative values in the difference-map correspond to lesions which are not associated to VEMP pathology while positive values may be capable to produce pathological VEMPs. The lesion maps were imported into 3D Slicer for visualization and analysis. 2.5. Correlation of lesion maps and anatomy For correlation of lesion data to anatomic structures a digitized atlas based on the ‘Atlas of the Human Brainstem’ (Paxinos and
1104
G. Heide et al. / Clinical Neurophysiology 121 (2010) 1102–1108
Huang, 1995) was used. The 64 templates of the Paxinos atlas were digitized. Afterwards the coordinate system existent on each template of the atlas was used to define reference landmarks. Matlab routines were used to perform a landmark registration of each template according to the coordinate system of the atlas using scaling, rotation, and translations. After this registration process, the image data was imported into the 3D Slicer, where the anatomical structures of the atlas were segmented manually, so that a 3D digitized atlas of the brainstem was created consisting of the registered Paxinos templates and segmented 3D structures of the brainstem. This brainstem atlas was projected onto the ‘standard’ brainstem with the lesion maps using 3D Slicer. For this, linear affine co-registration was used as implemented in the Slicer Software. The atlas data containing 64 slices was fitted to MR brainstem dataset with the associated lesion maps. These MR-data spanned the whole human brainstem with 26 slices of 408 512 pixel. Anatomical landmarks for registration were surface structures of the brainstem (e.g. ventricular floor line, pontomesencephalic junction, pontomedullary junction, etc.) and anatomical structures in the brainstem visible on T2w MRI (e.g. red nucleus, pyramis of the medulla oblongata, etc.). This way, the lesion maps could be correlated to anatomical data, so that all histological structures contained in the lesions could be identified. 3. Results Twenty-nine patients with acute brainstem infarcts were enrolled into the study (Table 1), while about 80 patients with brainstem ischemia were screened for inclusion into the study. However, most of these patients were relatively old of age and had other concomitant brain lesions (e.g. former stroke, subcortical arteriosclerotic encephalopathy, and others). Many patients complained peripheral vestibulopathy and/or hearing loss in the past, many patients used hearing aids and could therefore not be enrolled into the study. Twelve patients (group 1) showed unilaterally abnormal VEMPs on the lesion side of the brainstem infarction. Differences in p13 latencies between right and left side only occurred in 1 case (case #6) in addition to amplitude reduction. A separate behavior of positive (p13) and negative waves (n23) could not be found. Wave I of
the acoustic evoked potentials (AEP) was measured as a parameter of cochlear function. All 12 patients of group 1 showed bilaterally symmetrical latencies of wave I (delayed in 4 and normal in 8 patients, Table 1). Fig. 1 gives an example of a patient with abnormal VEMPs on the left hand side, who had an ischemic lesion in the left medulla oblongata. Ten patients (group 2) had normal VEMPs. Group 3 consisted of 7 patients with bilaterally absent VEMPs, which could not definitely be attributed to the brainstem infarction. The lesion profiles of these patients did not show a systematic lesion anatomy. Therefore, these patients were not further analyzed regarding lesion anatomy. The lesion maps of group 1 and group 2 are shown in Fig. 2. Unilaterally abnormal VEMPs were produced by lesions reaching from the mesencephalon down to the medulla oblongata. However, after subtraction of the lesion maps of group 1 and group 2 a more specific lesion pattern could be found. One region capable to produce pathological VEMPs is the lateral medulla oblongata in contrast to the classical Wallenberg infarct (Fitzek et al., 2006) located more dorsally in the lateral medulla oblongata. Another lesion site is the anterolateral part of the pyramidal tract fibers mainly located in the upper pons. In fewer cases lesions were found in the tegmentum of the pons, in one case only the vestibular nuclear group was lesioned. The correlation of the lesions to anatomical structures was done by the projection of the brainstem atlas to the lesion maps. Figs. 3 and 4 summarize the anatomical structures identified to be contained in the lesion maps which may be attributed to pathological VEMPs. In the medulla oblongata the spinal accessory nerve (n11) can be attributed to a disturbance of the efferent part of the VEMP reflex. In addition, ventral and dorsal spinocerebellar tracts, spinothalamic tract, and most parts of the medullar trigeminal system were included into the medullary lesion profile. It has to be noted that neither the ventral nor the lateral vestibulospinal tract was included in the lesion area in the medulla oblongata. In the pons most lesions were associated to anterolateral parts of the pontine pyramidal tract fibers. In fewer cases lesions of the tegmental area of the pons, but also the trapezoid body and medial lemniscus, as well as the vestibular nuclei were found. No significant weakness in the sternocleidomastoid muscles was noted in group 1 and group 2. To rule out the possibility that
Fig. 1. Example of a patient with unilaterally abnormal VEMPs. (A) Normal VEMP at the right side. (B) Missing VEMP at the left side. (C) Hyperintense DWI-signal in the left lateral medulla oblongata as a sign of acute brainstem ischemia. (D) Lesion in T2w MRI. (E) Ischemic lesion as 3D volume projected onto the right side of the reference brainstem.
G. Heide et al. / Clinical Neurophysiology 121 (2010) 1102–1108
1105
Fig. 2. Probabilistic lesion maps. The left column shows lesion profiles of patients with unilaterally pathological VEMPs (group 1). The middle column shows lesion profiles of patients with normal VEMPs (group 2). The right column shows a difference-map between group 1 and group 2. Note that lesions associated to normal VEMPs are visualized as cold colors and lesions associated to pathological VEMPs are visualized as warm colors. Color coding is shown in the right lower corner.
a different activation strategy due to muscle weakness could result in different VEMPs, the NIHSS scores for facial and arm motor function were compared between group 1 and 2 using the Mann–Whitney U-test. No statistically significant difference between these groups could be found in the facial (p = 0.974) and arm (p = 0.497) motor scores of the NIHSS. 4. Discussion The analysis of VEMPs in diseases of the brainstem has been suggested several times (Chen and Young, 2003). It has been reported that VEMPs may be abnormal in patients with multiple sclerosis (Patkó et al., 2007; Versino et al., 2007) and basilar artery migraine (Liao and Young, 2004). However, these are diseases which cause disseminated and not well circumscribed brainstem dysfunctions. Thus, it seems not to be reliable to study these patients for anatomical localization. Besides single case studies (Deftereos et al., 2006; Rosengren et al., 2007; Shin et al., 2009) describing VEMP abnormalities due to circumscribed brainstem lesions, only few group studies have been performed so far (Itoh et al., 2001; Deftereos et al., 2008). They concluded that VEMPs
may be a useful diagnostic tool to identify lower brainstem lesions especially in the lateral lower pons and the upper medulla oblongata. In contrast, Pollak et al. (2006) demonstrated that cerebellar strokes did not influence VEMPs (19 patients), but they also could not find statistically significant differences between 15 patients with lower brainstem ischemic stroke and 53 normal controls. However, they found abnormal VEMPs in single patients with brainstem lesions while other patients had normal VEMPs. Thus, for localization purposes the lesion profiles of patients with normal VEMPs have to be compared to patients with abnormal VEMPs, because lesion profiles of patients with normal VEMPs should provide information about anatomical structures which seem not to be involved in the VEMP reflex. Moreover, the use of an anatomical atlas as done in this study helps to identify anatomical structures in higher detail than using MR-data only. However, it has to be kept in mind that structures identified in lesion studies not necessarily may be correlated to the functional disturbance under study, because these may be included in the lesion volume caused by random inclusion into vascular territories. Therefore, anatomical plausibility has to be
1106
G. Heide et al. / Clinical Neurophysiology 121 (2010) 1102–1108
Fig. 3. Anatomical structures located in the lesion profiles of the medulla oblongata as derived from the atlas.
Fig. 4. Anatomical structures located in the lesion profiles of the pons as derived from the atlas.
1107
G. Heide et al. / Clinical Neurophysiology 121 (2010) 1102–1108
proven regarding a possible correlation between lesion anatomy and clinical disturbance. Bilaterally missing VEMPs were found in 7 patients. It is known that amplitudes and latency of VEMPs are dependent on age (Welgampola and Colebatch, 2001; Basta et al., 2007) and bilaterally absent VEMPs may be found in normal subjects over age 60 (Welgampola and Colebatch, 2005). Such changes may be due to age-related hair cell loss or other age-related changes of the vestibular system. Therefore, patients with bilaterally absent VEMPs had to be excluded from lesion analysis. The lesion profiles of these 7 patients did not show a systematic lesion anatomy making it very unlikely that a circumscribed central lesion could cause bilaterally absent VEMPs. Mostly, peripheral vestibular dysfunction is the major source for abnormal VEMPs. Although the history of peripheral vestibulopathy or hearing loss was an exclusion criterion in this study, it may be a limitation of the study that these functions have not been specifically tested (e.g. using audiometric measurements and calo-
ric tests). Bilateral symmetrical latencies of the AEP in the patients with asymmetrical VEMP pathology at least did not reveal an asymmetric dysfunction of the peripheral auditory system. Moreover, the lesions shown in MRI may include functional penumbra since all tests were done in the first week of stroke. This may be another limitation of the study. We demonstrated that VEMPs may be pathological in brainstem lesions not only being confined to the lower brainstem but reaching from medulla oblongata to the mesencephalon. The afferent part of the reflex is conducted by the vestibular nuclei at the level of the lateral lower pons and the efferent part by the accessory nucleus at the level of the medulla oblongata (Colebatch et al., 1994). Consequently, we found patients with lesions of the vestibular nuclei at the level of the lateral lower pons or lesions of the spinal accessory nerve in the lateral medulla oblongata. There is little information about the anatomy of the lateral vestibulospinal tract in man (Nathan et al., 1996; Brodal, 1981). However, the effects of the lateral vestibulospinal tract projecting from
Table 1 Patients enrolled in the study, clinical data, lesion location of brainstem infarcts, and VEMP pathology. VEMPs
Patients
Age
Gender
NIHSS
Lesion location
VEMP pathology
AEP wave I
Unilaterally pathological VEMPs
1
55
M
1
Absent
2
71
M
1
Right paramedian pons, tegmentum Left paramedian pons
3
66
M
3
Left dorsolateral medulla
Absent
4
50
M
5
Left paramedian pons
Amplitude reduction > 70%
5
59
M
2
Right dorsolateral medulla
Absent
6
65
M
4
7
61
M
1
Amplitude reduction > 50%, latency 20.9 ms Absent
8
68
M
4
Right lower paramedian pons, tegmentum Right paramedian pons, tegmentum Right dorsolateral medulla
Amplitude reduction > 80%
9
68
M
3
Left dorsolateral medulla
Absent
10
80
F
2
Right paramedian pons
Absent
11
63
M
1
Absent
12
54
M
1
Left paramedian pons, tegmentum Left paramedian pons
Wave I bilaterally delayed Wave I bilaterally delayed Bilaterally normal wave I Bilaterally normal wave I Bilaterally normal wave I Bilaterally normal wave I Wave I bilaterally delayed Bilaterally normal wave I Bilaterally normal wave I Bilaterally normal wave I Wave I bilaterally delayed Bilaterally normal wave I
n = 12 (41.4%)
Mean 63.3 ± 8.2
11 M, 1F
Mean 2.3 ± 1.4
13
65
M
1
14 15 16 17 18
66 68 71 63 68
M F M F M
3 1 8 1 1
19 20 21 22 n = 10 (34.5%)
62 68 67 70 Mean 66.8 ± 2.9
M M F M 7 M, 3 F
2 3 5 7 Mean 3.2 ± 2.6
23
55
M
3
Right dorsolateral medulla
Absent
24 25 26 27 28 29 n=7 (24.1%)
75 77 55 70 55 45 Mean 57.2 ± 12.2
M F F F F M 3 M, 4 F
6 4 3 1 4 2 Mean 3.6 ± 1.6
Left paramedian pons Left paramedian pons Right paramedian pons Right basal mesencephalon Right paramedian pons Right paramedian pons
Absent Absent Absent Absent Absent Absent
Normal VEMPs
Bilaterally absent VEMPs
Absent
Amplitude reduction > 75%
Left paramedian pons, tegmentum Right dorsolateral medulla Left dorsolateral medulla Right paramedian pons Left paramedian pons Left paramedian mesencephalon Left dorsolateral medulla Right paramedian pons Left paramedian pons Right paramedian pons
1108
G. Heide et al. / Clinical Neurophysiology 121 (2010) 1102–1108
the lateral vestibular nucleus of Deiters to the spinal cord are mainly excitatory. Consequently, it has been hypothesized recently that the medial vestibulospinal tract contains the pathway responsible for the inhibitory effects of the VEMP reflex (Shin et al., 2009; Kushiro et al., 1999). The medial vestibulospinal tract originates from the other three vestibular nuclei and runs within the medial longitudinal fascicle (Nieuwenhuys and Voogd, 1988). However, in our patient group no lesion of the medial longitudinal fascicle could be demonstrated in the lower brainstem – only in the midpons above the level of the vestibular nuclei. Although it has been suggested on purely anatomical considerations that only lesions of the lower pons and upper medulla oblongata may be capable to produce disturbances of VEMPs (Welgampola et al., 2005), we also demonstrated that lesions of the upper pons with pyramidal tract lesion in its upper ventrolateral part is capable to modify the VEMP reflex. This has not been described in the literature so far. The localization of the lesion in the ventrolateral (and not dorsomedial) part of the basis of the pons corresponds well to the localization of corticobulbar fibers (Kataoka et al., 2003) and corticofacial fibers (Urban et al., 2001). There could be a potential interaction in significant lesion location in the sense that patients with lesions in lower pons/vestibular nuclei level also could have lesions stretched up to the upper pons pyramidal tract level. However, 4 patients had upper paramedian pons infarction without additional infarction of the lower pons and vestibular nuclei (patients #2, 4, 10, 12, Table 1), which excludes the possibility that this area could randomly be included into the lesion profiles. In addition, none of the patients had multiple lesions, they all had single lacunar brainstem infarction. Thus, it may be hypothesized that abnormal VEMPs can also be caused by lesioned projections of the pyramidal tract. The hypothesis that a different activation strategy due to muscle weakness could result in VEMP alteration was contradicted by the observation that no significant differences between NIHSS scores of facial and arm motor function could be found between patients with pathological and normal VEMPs. A similar ‘supravestibular’ pathomechanism may be attributed to an influence from the pontine tegmentum above the level of the vestibular nuclei also including the medial longitudinal fascicle. In conclusion, abnormal VEMPs can be produced by lesions of the brainstem reaching from the medulla oblongata to the mesencephalon. Main foci of lesions capable to produce abnormal VEMPs are the lateral medulla oblongata (including the 11th nerve) and the ventrolateral basis of the upper pons (including parts of the pyramidal tract), but also the tegmentum of the lower pons (including the vestibular nuclei). It has to be considered from a clinical point of view that abnormal VEMPs may be produced not only by peripheral vestibulopathy but also by brainstem lesions. Moreover, VEMPs may be influenced by effects caused by lesions located above the level of the vestibular nuclei. References Akin FW, Murnane OD, Panus PC, Caruthers SK, Wilkinson AE, Proffitt TM. The influence of voluntary tonic EMG level on the vestibular-evoked myogenic potential. J Rehab Res Dev 2004;41:473–80. Axer H, Grässel D, Brämer D, Fitzek S, Kaiser WA, Witte OW, et al. Time course of diffusion imaging in acute brainstem infarcts. J Magn Reson Imaging 2007;26:905–12. Basta D, Todt I, Ernst A. Characterization of age-related changes in vestibular evoked myogenic potentials. J Vestib Res 2007;17:93–8. Brodal A. Neurological anatomy in relation to clinical medicine. New York: Oxford University Press; 1981. Chen C, Young Y. Vestibular evoked myogenic potentials in brainstem stroke. Laryngoscope 2003;113:990–3. Colebatch JG, Halmagyi GM. Vestibular evoked potentials in human neck muscles before and after unilateral vestibular deafferentation. Neurology 1992;42:1635–6.
Colebatch JG, Halmagyi GM, Skuse NF. Myogenic potentials generated by a clickevoked vestibulocollic reflex. J Neurol Neurosurg Psychiatr 1994;57:190–7. Colebatch JG, Rothwell JC. Motor unit excitability changes mediating vestibulocollic reflexes in the sternocleidomastoid muscle. Clin Neurophysiol 2004;115:2567–73. Debatisse D, Pralong E, Guerit JM, Bisdorff A. Recording click-evoked myogenic potentials (CEMPs) with a setup for brainstem auditory evoked potentials (BAEPs). Neurophysiol Clin 2005;35:109–17. Deftereos SN, Panagopoulos G, Eleftheriadou A, Korres S, Georgonikou D, Kandiloros D, et al. Using vestibular evoked myogenic potentials to localise brainstem lesions. A preliminary report. B-ENT 2008;4:215–9. Deftereos SN, Panagopoulos G, Gryllia M, Georgonikou D, Polyzoi M, Kechagias E, et al. Neurophysiological monitoring of brainstem function in a patient with Wallenberg syndrome, using Vestibular Evoked Myogenic Potentials. Neurol Neurophysiol Neurosci 2006:3. Eleftheriadou A, Deftereos SN, Zarikas V, Panagopoulos G, Korres S, Sfetsos S, et al. Vestibular evoked myogenic potential eliciting in normal subjects: comparison of four different methods. J Otolaryngol Head Neck Surg 2008;37:704–11. Fitzek S, Baumgärtner U, Marx J, Joachimski F, Axer H, Witte OW, et al. Pain and itch in Wallenberg’s syndrome: anatomical–functional correlations. Suppl Clin Neurophysiol 2006;58:187–94. Gering DT, Nabavi A, Kikinis R, Hata N, O’Donnell LJ, Grimson WE, et al. An integrated visualization system for surgical planning and guidance using image fusion and an open MR. J Magn Reson Imaging 2001;13:967–75. Heide G, Freitag S, Wollenberg I, Iro H, Schimrigk K, Dillmann U. Click evoked myogenic potentials in the differential diagnosis of acute vertigo. J Neurol Neurosurg Psychiatry 1999;66:787–90. Hong SM, Yeo SG, Kim SW, Cha CI. The results of vestibular evoked myogenic potentials, with consideration of age-related changes, in vestibular neuritis, benign paroxysmal positional vertigo, and Meniere’s disease. Acta Otolaryngol 2008;128:861–5. Itoh A, Kim YS, Yoshioka K, Kanaya M, Enomoto H, Hiraiwa F, et al. Clinical study of vestibular-evoked myogenic potentials and auditory brainstem responses in patients with brainstem lesions. Acta Otolaryngol Suppl 2001;545:116–9. Janzen J, Schlindwein P, Bense S, Bauermann T, Vucurevic G, Stoeter P, et al. Neural correlates of hemispheric dominance and ipsilaterality within the vestibular system. NeuroImage 2008;42:1508–18. Kataoka S, Miaki M, Saiki M, Saiki S, Yamaya Y, Hori A, et al. Rostral lateral pontine infarction: neurological/topographical correlations. Neurology 2003;61:114–7. Kim-Lee Y, Ahn JH, Kim YK, Yoon TH. Tone burst vestibular evoked myogenic potentials: diagnostic criteria in patients with Meniere’s disease. Acta Otolaryngol 2009;129:924–8. Kushiro K, Zakir M, Ogawa Y, Sato H, Uchino Y. Saccular and utricular inputs to sternocleidomastoid motoneurons of decerebrate cats. Exp Brain Res 1999;126:410–6. Lee SK, Cha CI, Jung TS, Park DC, Yeo SG. Age-related differences in parameters of vestibular evoked myogenic potentials. Acta Otolaryngol 2008;128:66–72. Liao L, Young Y. Vestibular evoked myogenic potentials in basilar artery migraine. Laryngoscope 2004;114:1305–9. Maes L, Vinck BM, De Vel E, D’haenens W, Bockstael A, Keppler H, et al. The vestibular evoked myogenic potential: a test-retest reliability study. Clin Neurophysiol 2009;120:594–600. Nathan PW, Smith M, Deacon P. Vestibulospinal, reticulospinal and descending propriospinal nerve fibres in man. Brain 1996;119:1809–33. Nieuwenhuys R, Voogd J, Van Huijzen C. The human central nervous system. Berlin: Springer; 1988. Patkó T, Simó M, Arányi Z. Vestibular click-evoked myogenic potentials: sensitivity and factors determining abnormality in patients with multiple sclerosis. Mult Scler 2007;13:193–8. Paxinos G, Huang X. Atlas of the human brainstem. San Diego: Academic Press; 1995. Pollak L, Kushnir M, Stryjer R. Diagnostic value of vestibular evoked myogenic potentials in cerebellar and lower-brainstem strokes. Neurophysiol Clin 2006;36:227–33. Rosengren SM, Nogajski JH, Cremer PD, Colebatch JG. Delayed vestibular evoked responses to the eyes and neck in a patient with an isolated brainstem lesion. Clin Neurophysiol 2007;118:2112–6. Schlindwein P, Mueller M, Bauermann T, Brandt T, Stoeter P, Dieterich M. Cortical representation of saccular vestibular stimulation: VEMPs in fMRI. NeuroImage 2008;39:19–31. Shin J, Song H, Koo J, Lee H, Kim JS. Abnormal vestibular evoked myogenic potentials in medial medullary infarction. J Clin Neurol 2009;5:101–3. Urban PP, Wicht S, Vucorevic G, Fitzek S, Marx J, Thömke F, et al. The course of corticofacial projections in the human brainstem. Brain 2001;124:1866–76. Versino M, Ranza L, Colnaghi S, Alloni R, Callieco R, Romani A, et al. The N3 potential compared to sound and galvanic vestibular evoked myogenic potential in healthy subjects and in multiple sclerosis patients. J Vestib Res 2007;17: 39–46. Wang S, Yeh T, Chang C, Young Y. Consistent latencies of vestibular evoked myogenic potentials. Ear Hear 2008;29:923–9. Watson SR, Colebatch JG. Vestibulocollic reflexes evoked by short-duration galvanic stimulation in man. J Physiol (Lond) 1998;513:587–97. Welgampola MS, Colebatch JG. Vestibulocollic reflexes: normal values and the effect of age. Clin Neurophysiol 2001;112:1971–9. Welgampola MS, Colebatch JG. Characteristics and clinical applications of vestibular-evoked myogenic potentials. Neurology 2005;64:1682–8.