- Email: [email protected]
Galvanic and acoustic vestibular stimulation activate different populations of vestibular afferents
Clinical Neurophysiology 114 (2003) 359–365 www.elsevier.com/locate/clinph
Galvanic and acoustic vestibular stimulation activate different populations of vestibular afferents Ann M. Bacsi, Shaun R.D. Watson, James G. Colebatch* Institute of Neurological Sciences and UNSW Clinical School, Prince of Wales Hospital, Randwick, Sydney, NSW 2031, Australia Accepted 31 October 2002
Abstract Objective: To deduce whether similar or distinct populations of vestibular afferents are activated by acoustic and galvanic vestibular stimulation by comparing the effectiveness of ‘matched’ stimuli in eliciting vestibulospinal reflexes. Methods: Twelve subjects (5 men, 7 women) underwent individual ‘matching’ of 2 ms tone burst and galvanic stimuli, using vestibulocollic reflexes so that corrected reflex amplitudes to tone burst and galvanic stimuli were within 10% of each other. These same intensities were then administered using 20 ms durations to determine whether they were equally effective in evoking vestibulospinal responses. Results: Corrected reflex amplitudes for vestibulocollic responses to tone burst and galvanic stimulation were not significantly different for the right (P ¼ 0:45) or left (P ¼ 0:68) sides. All subjects had vestibulospinal responses to galvanic stimulation (average intensity 4.0 mA for both sides). The short latency (SL) and medium latency (ML) components of the vestibulospinal reflexes were larger after galvanic compared to tone burst stimulation in 11 of 12 subjects (P , 0:01). Conclusions: Despite evoking equal-sized vestibulocollic reflexes, there was a clear dissociation between the magnitude of tone burst and galvanic-induced vestibulospinal reflexes. Galvanic stimulation evoked SL and ML reflexes in all subjects. Tone burst stimuli evoked only small SL reflexes and, in most cases, no ML reflexes. Acoustically-evoked vestibulocollic reflexes are likely to be due to saccular excitation. The limited effectiveness of longer tone burst stimuli to evoke ML vestibulospinal reflexes suggests that saccular afferents have, at most, only a minor role in the production of these reflexes. We conclude that galvanic stimulation is more effective in eliciting vestibulospinal reflexes than tone burst stimulation, and that the two methods activate different populations of vestibular afferents. q 2002 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Galvanic; Vestibulocollic; Vestibulospinal; Human
1. Introduction Both acoustic (Colebatch et al., 1994; Cheng and Murofushi, 2001) and galvanic stimuli (Watson and Colebatch, 1998b) are capable of evoking short latency vestibulocollic reflexes in tonically contracting sternocleidomastoid (SCM) muscles. The waveform and latencies of both methods of stimulation are very similar; the responses show an initial positive deflection (p13) followed by a negative wave (n23) ipsilateral to the side of stimulation (Colebatch et al., 1994). Galvanic stimulation evokes a positive-negative response slightly earlier than that for tone bursts but is otherwise similar (Watson and Colebatch, 1998b; Welgampola and Colebatch, 2001a; Murofushi et al., 2002). The size of these responses scales with the level of tonic activation (Watson and Colebatch, 1998b). In the case of acousti* Corresponding author. Tel.: 161-612-9382-2407; fax: 161-612-93822428. E-mail address: [email protected] (J.G. Colebatch).
cally-evoked vestibular reflexes, the evidence strongly suggests an action on the saccule (Young et al., 1977; Cazals et al., 1983; Didier and Cazals, 1989; McCue and Guinan, 1995, 1997; Murofushi and Curthoys, 1997). Galvanic stimulation excites distal vestibular nerve afferents (Spiegel and Scala, 1943; Goldberg et al., 1984), particularly those with an irregular discharge (Goldberg et al., 1984), but any selectivity of this action on afferents arising from the different vestibular end organs is uncertain. Longer duration galvanic stimulation evokes a characteristic electromyographic (EMG) response in the soleus muscles of standing subjects (Britton et al., 1993; Fitzpatrick et al., 1994; Watson and Colebatch, 1997). This response consists of two parts- the short (SL) and medium latency (ML) components. It differs from the galvanicevoked vestibulocollic response in that it is strongly influenced by both postural task and head position, such that changing direction of head rotation leads to a reversal in polarity of both components. Although it is not certain
1388-2457/02/$ - see front matter q 2002 Elsevier Science Ireland Ltd. All rights reserved. doi:10.1016/S1 388-2457(02)00 376-0
CLINPH 2002118
360
A.M. Bacsi et al. / Clinical Neurophysiology 114 (2003) 359–365
that both components travel via the vestibulospinal tract (Britton et al., 1993), for simplicity we will use the term vestibulospinal for both reflexes. Watson and Colebatch (1998a) found that click vestibular activation could evoke vestibulospinal reflexes in some subjects which resembled those following galvanic stimulation. The effects following clicks had a similar appearance and latency to those following cathodal vestibular stimulation, i.e. the effects of clicks were consistent with an excitatory effect on vestibular afferents. However, these authors were unable to compare the relative effectiveness of the two methods given the very different durations of each type of stimulus. Our study was designed to deduce whether similar or distinct populations of vestibular afferents are activated by acoustic and galvanic stimulation by comparing the effectiveness of ‘matched’ stimuli in eliciting vestibulospinal reflexes. Tone bursts excite saccular afferents (McCue and Guinan, 1994, 1997) and evoke short latency reflexes which resemble those to clicks (Welgampola and Colebatch, 2001b). Tone bursts have the advantage that their duration can be safely varied to match that commonly used with galvanic stimulation in evoking vestibulospinal reflexes. Initially, the intensities of the acoustic and galvanic stimuli were adjusted so that the sizes of the corrected reflex amplitudes recorded from the SCM muscles were within 10% of one another. We then tested the relative effectiveness of these ‘matched’ stimulus intensities using longer (but equal) durations in evoking vestibulospinal reflexes. If both forms of vestibular stimulation were to activate a similar population of fibres and, in particular, if saccular fibres provide an important contribution to vestibulospinal reflexes, then the effect of both methods of stimulation should be similar for reflexes recorded in both the neck and the leg. Failure to see similar sized effects in the legs would allow us to conclude that each technique activates different populations of vestibular afferents. A further consequence of such an observation would be that the use of both techniques is likely to provide information about vestibular function additional to either one alone.
2. Methods 2.1. Subjects and study design Twelve healthy adults aged 25–46 (7 females, 5 males) were studied after obtaining informed consent and local ethics committee approval. The study consisted of two parts. Firstly, the initial ‘matching’ was performed by measuring vestibulocollic reflexes to short (2 ms) acoustic and galvanic (direct current) stimuli, with the latter adjusted to evoke the same size of short latency vestibulocollic reflex as the former. Secondly, recordings were made of vestibulospinal reflexes from over the soleus muscles, in standing subjects, using longer (20 ms) duration acoustic and galvanic stimuli. The current intensity was that previously deter-
mined as ‘matching’ the acoustic stimulus for vestibulocollic reflexes. For all acoustic stimulation rise time and fall time ¼ 0 ms. 2.2. Vestibulocollic reflexes Tone bursts (124 dB SPL at 500 Hz, 2 V peak to peak) were generated using a CED 1401 interface (Cambridge Electronic Design, Cambridge, UK) and phase-locked to the trigger pulse. The tone burst was delivered pseudorandomly to either the right or left ear through calibrated headphones (TDH 49, Telephonics Corp., USA). A total of 512 stimuli (256 per ear) were presented at 5/s. Subjects were tested in the recumbent position, on a couch with an adjustable backrest, while their heads were elevated against gravity to activate both SCM muscles. Electromyographic activity was recorded from Ag/AgCl surface electrodes (3 M, St Paul, MN, USA) over symmetrical sites on both sides of the neck, amplified and bandpass filtered (32 Hz–1.6 kHz). Active recording electrodes were placed on the upper third of the muscle belly and reference electrodes on the medial ends of the clavicles. An earth electrode was placed on either the upper end of the sternum or the medial forearm. EMG was sampled at 5000 Hz, from 20 ms before to 100 ms after stimulus onset using a 1401 plus analogue to digital converter and Sigavg software (Cambridge Electronic Design, Cambridge, UK) on a PC. Averages were made of the unrectified and rectified EMG signals. The reflex responses were measured from the unrectified average, while the rectified average was used to quantify the level of tonic activation. The ipsilateral p13n23 response was used for matching but the contralateral n12p20 response (Watson and Colebatch, 1998b) was also measured, if present. ‘Corrected reflex amplitudes’ were calculated by determining the peak to peak amplitude and then dividing by the mean rectified EMG activity level for the 20 ms preceding the stimulus onset to correct for background activation (Colebatch et al., 1994; Welgampola and Colebatch, 2001b). Corrected reflex amplitude or reflex gain is dimensionless and is largely independent of the level of tonic activity during the stimulus presentation (Colebatch et al., 1994). Short duration galvanic stimulation (model DS2 A; Digitimer, UK) was via electrodes of 27–30 cm 2, individually cut from commercially available electrosurgical plating, coated with electrode gel and placed over the mastoids. These were then secured with tape and an elastic bandage was wrapped around the head. An indifferent electrode was placed either over the spine at C7 or on the shoulder opposite to the side of stimulation. By convention we commenced this part of the experiment with the cathode over the left mastoid, followed by the cathode over the right mastoid. For galvanic-evoked vestibulocollic reflexes, the stimulating and recording sites are close together, and an artefact eliminator was used (Watson and Colebatch, 1998b). Two trials were performed, one at rest and the
A.M. Bacsi et al. / Clinical Neurophysiology 114 (2003) 359–365
other during contraction with the subject raising the head against gravity, with a brief rest period (,1 min) between recordings. In each, stimuli 2 ms long were delivered at 3/s to a total of 256 stimuli per trial. The resting trace (artefact only) was subtracted from the active trace (artefact plus response) to eliminate the stimulus artefact (Watson and Colebatch, 1998b). Vestibulocollic reflexes were then recorded beginning with a 4 mA current which was adjusted to produce a corrected p13n23 reflex amplitude within 10% of that for the tone burst-evoked vestibulocollic reflex. Early responses in the contralateral SCM were also measured, but not used for matching. 2.3. Vestibulospinal reflexes Nine of 12 subjects underwent vestibulospinal testing on a subsequent day. Three subjects underwent all testing on the same day. The vestibulospinal part of the experiment was performed with subjects standing, eyes closed, feet together, head rotated 90 degrees to the left and leaning slightly forwards. All subjects received a standard tone burst stimulus of 124 dB SPL at 500 Hz (2 V peak to peak), duration 20 ms and 256 tones per trial (128 for each ear; presented pseudorandomly as for the vestibulocollic part of the experiment). This stimulus is sufficient to activate the stapedial reflex but there is little attenuation after 400 ms (Welgampola and Colebatch, 2001b), thus, we chose a repetition interval of 500 ms. This tone burst is also sufficient to elicit a startle reflex, and these occur earlier and more frequently in standing subjects (Brown et al., 1991). Nevertheless we consider any effect of startle would have been minimal, both because the effects would have rapidly habituated with the rapid repetition of the stimuli and also because the reflex excitability would not have inverted for the two ears, a characteristic feature of vestibulospinal reflexes (Watson and Colebatch, 1997). Only reciprocal effects were measured in our estimation of the SL and ML reflexes. Unipolar cathodal stimulation was given at the intensity previously determined as ‘matching’ the corrected reflex amplitudes from the neck recordings for each subject. These stimuli were also of 20 ms duration, administered at 2 Hz with 128 stimuli per trial. By convention, we also commenced recordings with cathode left. Surface electrodes were placed over the soleus, 4.5–5.5 cm apart. EMG recordings were made from the soleus contralateral to the direction of head rotation (in this case, the right soleus), as this has previously been shown to be the larger response (Britton et al., 1993; Watson and Colebatch, 1997). EMG was amplified and filtered (32 Hz–1.6 kHz) and then sampled at 5000 Hz, from 50 ms before to 250 ms after stimulus onset. Each trial was rectified before averaging. The artefact eliminator was not required for this part of the experiment. Traces were analysed by superimposing the cathode-left and cathode-right recordings after digital filtering and
361
subtraction of prestimulus rectified EMG (Watson and Colebatch, 1997; Welgampola and Colebatch, 2001a). The SL reflex was defined as the response beginning between 40 and 70 ms and the onset taken as the point at which the traces of the two stimulus montages first diverged. The response immediately following the SL and with opposite polarity, commencing at about 90 ms after the stimulus onset, was taken as the end of the short latency response and the beginning of the ML response. The point at which the traces crossed over a second time was taken as the end of the ML response (Britton et al., 1993; Fitzpatrick et al., 1994). The SL and ML reflex amplitudes were expressed as a percentage of mean rectified EMG, after subtraction of the level of prestimulus EMG. This method allowed for differences in prestimulus EMG activity. The average of the two values obtained for the two stimulus polarities was calculated to give a percentage estimate of the overall size of the SL and ML reflexes (Welgampola and Colebatch, 2001a). Theoretically, because our vestibulospinal stimulus was relatively short, a response to its cessation (an ‘off’ response) might have affected the size of the ML reflex. Any such effect should have been constant for a constant SL reflex so should not have affected the hypothesis being tested. In the case of tone bursts, responses were often small and 10 of 12 subjects had repeat trials to verify the presence or absence of a response. If a (small) response was present for only one trial and appeared to be absent in the other, the values for SL and ML amplitudes for the positive trial were halved. 2.4. Statistical methods Student’s paired t tests and Pearson Correlations were used to compare the sizes of vestibulocollic and vestibulospinal reflexes. Chi squared and Fisher’s exact tests were used to analyse contralateral responses to tone burst and galvanic stimuli in the neck. Unless stated otherwise, results are given as mean ^ standard deviation. 3. Results 3.1. Vestibulocollic reflexes All subjects had an ipsilateral positive (p13) followed by a negative (n23) response to tone bursts of 124 dB SPL at 500 Hz (2 V peak to peak). The average latencies for the p13 and n23 peaks were 13.6 ^ 1.75 and 22 ^ 2.28 ms, respectively, for the right side, and 12.6 ^ 0.72 and 21.5 ^ 1.95 ms, respectively, for the left side. The results for right and left corrected reflex amplitudes were, respectively, 0.87 ^ 0.31 (range 0.28–1.31) and 0.86 ^ 0.32 (range 0.39–1.47). The average galvanic current intensity was 4.0 mA (range 2–6 mA; Table 1) and all stimuli were well tolerated. Overall, the galvanic-evoked reflex amplitudes were very closely matched, with corrected values for the right and left sides of 0.85 ^ 0.33 and 0.85 ^ 0.31 (Fig.
362
A.M. Bacsi et al. / Clinical Neurophysiology 114 (2003) 359–365
Table 1 Intensities of galvanic stimulation used to match vestibulocollic reflexes Subject
Right (mA)
Left (mA)
1 2 3 4 5 6 7 8 9 10 11 12
5 2 4.5 3.5 4 4.5 6 5 3 6 5 4.5
4.75 4 5 3 4 4.5 2.5 4 3 6 4 4
uniform – subjects with the largest galvanic-evoked SL reflexes had the greatest reduction in their responses to acoustic stimulation. There was a significant negative correlation between the relative size of the acoustic-evoked SL reflex and the SL reflex amplitude evoked by galvanic stimulation (r ¼ 20:81, P , 0:01). ML reflexes showed a different pattern. ML reflexes were nearly always absent in response to acoustic stimulation, being detectable in only 2 of 24 trials (with amplitudes of 3 and 4%). Overall, both SL and ML amplitudes were significantly smaller for tone burst
1). Matching the reflex evoked by galvanic stimulation to within 10% of the tone burst-evoked reflex was successful in 20 of the 24 ears. In four ears (three subjects) this was not possible; the galvanic stimulus amplitude used evoked smaller responses (by 26, 28 and 29%), or, in one case evoked a larger reflex (by 29%) than the target amplitude. Nevertheless, there was no significant difference between the corrected amplitudes to acoustic and galvanic stimulation on either the right (P ¼ 0:45, paired t test) or left sides (P ¼ 0:68, paired t test). Contralateral responses (i.e. from the left SCM with stimulation on the right and vice-versa) were more common with galvanic stimulation (Fig. 2). This difference was significant for right sided stimulation, (3 crossed responses to acoustic stimulation versus 8 crossed responses to galvanic stimulation; P , 0:05) but not for left sided stimulation (2 versus 6 crossed responses; P ¼ 0:08). Overall, contralateral responses were significantly smaller than ipsilateral responses for both methods of stimulation (tð11Þ ¼ 4:03–8:5; P , 0:002 in all cases). 3.2. Vestibulospinal reflexes Galvanic stimulation with an intensity ‘matching’ that for tone bursts, (i.e. an average of 4.0 ^ 1.0 mA) with a duration of 20 ms produced SL and ML reflexes in all subjects. The mean onset latencies were 53.8 ms (range 47.6–63.5) and 88.5 ms (range 84.6–93.4), respectively. Average SL amplitude (expressed as a percentage of baseline EMG activity) was 18 ^ 11% (range 7–42%) and, for ML, was 18 ^ 9% (range 8–32%). Tone bursts of 20 ms duration (124 dB SPL) produced an SL response in 16 of 24 trials with an average onset at 51.8 ^ 6.0 ms. The mean SL amplitude to the tone burst (as a percentage of baseline EMG activity) was 7 ^ 3% (Fig. 3). The tone burst-evoked SL reflexes had smaller amplitudes than the galvanic-evoked reflexes in 11 of 12 subjects (Fig. 4). In one subject with a small response to galvanic stimulation, there was a small increase (from 8.5 to 10%) in the SL response to tone bursts. The difference between the size of the galvanic-evoked SL reflex and that evoked by acoustic stimulation was not
Fig. 1. Matching of tone burst and galvanic vestibulocollic reflexes in a single subject, consisting of averaged EMG for 256 stimuli per ear for both acoustic and galvanic stimulation. The top pair of tracings was obtained after tone burst stimulation, and the corrected reflex amplitudes are 1.17 for both the right and left sides. The bottom pair of tracings shows the response to galvanic stimulation, where corrected reflex amplitudes are 1.08 for the right side and 1.14 for the left side. The dotted lines indicate stimulus onset.
A.M. Bacsi et al. / Clinical Neurophysiology 114 (2003) 359–365
363
Fig. 3. Vestibulospinal reflexes recorded over the right soleus muscle to galvanic and tone burst stimulation: rectified, averaged EMG traces are shown for 128 acoustic or galvanic stimuli. The DC component has been removed and traces have been digitally filtered. Galvanic stimuli were given in pulses of 20 ms. Intensities were 5 mA for the right side and 4.75 mA for the left side. Tone bursts were administered bilaterally at 124 dB SPL; each tone lasted 20 ms. The galvanic intensities were those which ‘matched’ the vestibulocollic reflex amplitudes. SL size for galvanic stimulation was 29% whereas no response was measured after tone burst stimulation. The dotted lines indicate stimulus onset. For galvanic stimulation, the dark line indicates results for cathode left; lighter line indicates results for cathode right. In the case of tone bursts, the dark line represents the EMG trace for left-sided acoustic stimulation; the lighter line represents the EMG trace for right-sided acoustic stimulation.
activation using acoustic stimuli appears to be selective for the saccule (Townsend and Cody, 1971; Cazals et al., 1983; Didier and Cazals, 1989; Murofushi and Curthoys, 1997). Murofushi and Curthoys (1997) found that click sensitive neurons were predominantly found in the inferior division of the vestibular nerve which supplies the saccule, although some were also present in the posterior part of the superior
Fig. 2. Ipsilateral and contralateral responses to right-sided galvanic and tone burst stimulation consisting of averaged EMG for 256 stimuli per ear for both methods of stimulation. The upper pair of traces shows the responses to galvanic stimulation, with corrected reflex amplitudes of 1.08 on the ipsilateral side and 0.36 for the contralateral n12p20 response. The lower pair of traces shows the responses to tone burst stimulation, where the ipsilateral side had a corrected reflex amplitude of 1.17, and the contralateral n12p20 was absent. The dotted lines indicate stimulus onset.
stimulation than for galvanic stimulation (tð11Þ ¼ 3:37, P , 0:01; tð11Þ ¼ 6:8, P , 0:001 for SL and ML, respectively). There was no significant correlation between the sizes of tone burst or galvanic-evoked vestibulocollic reflex amplitudes and corresponding SL reflexes for either right or left sides (P range 0.148–0.979).
4. Discussion Previous studies have indicated that the site of vestibular
Fig. 4. Comparison of sizes for galvanic and tone burst-evoked vestibulospinal reflexes: SL component. The left column of points corresponds to the size of the SL reflex evoked by galvanic stimulation. The right column shows the sizes of the SL reflexes evoked using tone burst stimulation. Individual subject results are paired by a joining line. Eleven of 12 subjects had a greater response to galvanic stimulation, a difference which was statistically significant (t ¼ 3:37, P , 0:01). Similar findings applied to the ML amplitudes. The dotted lines indicate subjects whose galvanicevoked vestibulocollic responses were not matched to within 10% of tone burst-evoked responses.
364
A.M. Bacsi et al. / Clinical Neurophysiology 114 (2003) 359–365
division of the vestibular nerve. In the same study, biocytin tracing of click sensitive neurons revealed nerve endings throughout the saccular maculae, whereas the utricular maculae were free of labelled nerve endings. Galvanic stimulation, in contrast, appears to work at the level of the vestibular nerve terminals. Spiegel and Scala (1943) found that destruction of the end-organ did not suppress the response to galvanic stimulation, whereas the response was abolished by sectioning the vestibular nerve. Goldberg et al. (1984) identified the site of action as the spike trigger zone. Cathodal stimulation increased, and anodal stimulation decreased, the discharge rate of vestibular afferents. Irregularly firing afferents were six times more sensitive than regularly firing afferents to applied current. Day et al. (1997) proposed that the central nervous system interprets galvanic stimulation as a tilt of the support surface (presumably an otolith effect), with resultant movement of body and head towards the anodal ear. Other studies (Zink et al., 1998; Watson et al., 1998) reported that low intensity (3 mA) galvanic stimulation applied to the mastoid processes may be relatively selective for the otolith, resulting in perceived tilt of the visual field and ocular torsion towards the anode but provoking minimal or no nystagmus. This finding is in contrast to earlier studies, such as Breson et al. (1971), who recorded horizontal nystagmus in 61 normal subjects following galvanic stimulation at 1.6 mA. Zink et al. (1998) recorded 5 s trials, which may have been too short to record sustained nystagmus which can require a ‘buildup’ time of up to 6 s (Karlberg et al., 2000). In addition, the study by Watson et al. (1998) involved a visual fixation point, known to suppress nystagmus, as well as a low sampling frequency of 2 Hz. Inter-individual differences may also influence the pattern of responses. Kleine et al. (1999) found that, after sinusoidal galvanic stimulation, some subjects had a large nystagmic response with little tonic response (indicating predominant semicircular canal stimulation) whereas other subjects had a large tonic response with little nystagmus (indicating predominant otolith stimulation). The evidence overall is compatible with galvanic stimulation having an action on both otolith and semicircular canal afferents, with significant inter-individual variation (MacDougall et al., 2002). Our experiment showed that despite evoking equal-sized vestibulocollic reflexes, there was a clear dissociation between tone burst and galvanic-induced vestibulospinal reflexes. There was also a trend for galvanic stimulation to evoke more crossed vestibulocollic responses, but this was only significant for right-sided stimulation. Future studies with larger sample sizes may clarify this issue. Vestibulospinal responses to galvanic stimuli were significantly larger than those for tone bursts in nearly all subjects. As noted above, acoustically-evoked vestibulocollic reflexes are likely to be the result of saccular excitation. The very limited effectiveness of longer tone burst stimuli in evoking ML reflexes suggests that saccular afferents have only a minor role at most in the production of this reflex.
The findings for the SL component of the vestibulospinal reflex were more complex. Our findings suggest that saccular afferents may be sufficient to evoke small SL reflexes but that additional vestibular afferents contribute in those subjects with large SL reflexes. In contrast to tone bursts, galvanic stimulation evoked prominent SL and ML vestibulospinal responses in all subjects. Previous studies have provided evidence that galvanic stimulation causes activation of otolith and semicircular canal afferents (see above). The SL reflex is thought to be mediated by vestibulospinal pathways (Britton et al., 1993). The main descending vestibular projection to the legs is the lateral vestibulospinal tract, originating from Deiter’s nucleus (Brodal, 1981). This, in turn, receives a major part of its input from otolith receptors. It is likely that galvanicevoked vestibulospinal reflexes are, at least in part, the result of otolith activation. If so, our results suggest that the saccular contribution is less important than afferents arising from the utricle, particularly for the ML component of vestibulospinal reflexes in soleus. The greater frequency of crossed responses to galvanic-evoked vestibulocollic reflexes would also be consistent with an effect on utricular afferents. Utricular afferents, unlike saccular afferents, have been shown to evoke crossed excitation in SCM motoneurons (Kushiro et al., 1999). Our observations indicate that different populations of vestibular afferents are activated by galvanic and tone burst stimulation. This, in turn, implies that tone burst and galvanic-evoked reflexes may probe different aspects of vestibular function. Acknowledgements This work was supported by the National Health and Medical Research Council of Australia. We thank Ms Sally Rosengren for assistance with data collection and editorial comments. References Breson K, Elberling C, Fangel J. Galvanic nystagmography. Acta Otolaryngol 1971;71(6):449–455. Britton TC, Day BL, Brown P, Rothwell JC, Thompson PD, Marsden CD, Postural EMG. responses in the arm and leg following galvanic vestibular stimulation in man. Exp Brain Res 1993;94:143–151. Brodal A. Neurological anatomy in relation to clinical medicine. 3rd ed. Oxford: Oxford University Press, 1981. pp. 202–205. Brown P, Day BL, Rothwell JC, Thompson PD, Marsden CD. The effect of posture on the normal and pathological auditory startle reflex. J Neurol Neurosurg Psychiatry 1991;54:892–897. Cazals Y, Aran J, Erre J, Guilhaume A, Aurousseau C. Vestibular acoustic reception in the guinea pig: a saccular function? Acta Otolaryngol 1983;95:211–217. Cheng P, Murofushi T. The effects of plateau time on vestibular-evoked myogenic potentials triggered by tone bursts. Acta Otolaryngol 2001;121:935–938. Colebatch JG, Halmagyi GM, Skuse NF. Myogenic potentials generated by a click-evoked vestibulocollic reflex. J Neurol Neurosurg Psychiatry 1994;57:190–197.
A.M. Bacsi et al. / Clinical Neurophysiology 114 (2003) 359–365 Day BL, Severac-Cauquil A, Bartolomei L, Pastor MA, Lyon IN. Human body-segment tilts induced by galvanic stimulation: a vestibularly driven balance protection mechanism. J Physiol 1997;500(3):661– 672. Didier A, Cazals Y. Acoustic responses recorded from the saccular bundle on the eighth nerve of the guinea pig. Hearing Res 1989;37:123–128. Fitzpatrick R, Burke D, Gandevia SG. Task-dependent reflex responses and movement illusions evoked by galvanic vestibular stimulation in standing humans. J Physiol 1994;478.2:363–372. Goldberg JM, Smith CE, Ferna´ ndez C. Relation between discharge regularity and responses to externally applied galvanic currents in vestibular nerve afferents of the squirrel monkey. J Neurophysiol 1984;51:1236– 1256. Karlberg M, McGarvie L, Magnusson M, Aw ST, Halmagyi GM. The effects of galvanic stimulation on the human vestibulo-ocular reflex. Neuroreport 2000;11:3897–3901. Kleine JF, Guldin WO, Clarke AH. Variable otolith contribution to the galvanically induced vestibulo-ocular reflex. Neuroreport 1999;10:1143–1148. Kushiro K, Zakir M, Ogawa Y, Sato H, Uchino Y. Saccular and utricular inputs to SCM motoneurons of decerebrate cats. Exp Brain Res 1999;126(3):410–416. MacDougall HG, Brizuela AE, Burgess AM, Curthoys IS. Between-subject variability and within-subject reliability of the human eye-movement response to bilateral galvanic (DC) vestibular stimulation. Exp Brain Res 2002;144:69–78. McCue MP, Guinan JJ. Acoustically responsive fibres in the vestibular nerve of the cat. J Neurosci 1994;14:6058–6070. McCue MP, Guinan JJ. Spontaneous activity and frequency selectivity of acoustically responsive vestibular afferents in the cat. J Neurophysiol 1995;74.4:1563–1572. McCue MP, Guinan JJ. Sound-evoked activity in primary afferent neurons of a mammalian vestibular system. Am J Otol 1997;18:355– 360. Murofushi T, Curthoys IS. Physiological and anatomical study of click-
365
sensitive primary vestibular afferents in the guinea pig. Acta Otolaryngol 1997;117(1):66–72. Murofushi T, Takegoshi H, Ohki M, Ozeki H. Galvanic-evoked myogenic responses in patients with an absence of click-evoked vestibulocollic reflexes. Clin Neurophysiol 2002;113(2):305–309. Spiegel SA, Scala NP. Response of the labyrinthine apparatus to electrical stimulation. Arch Otol 1943;38:131–138. Townsend GL, Cody DTR. The averaged inion response evoked by acoustic stimulation: its relation to the saccule. Ann Otorhinolaryngol 1971;80:121–131. Watson SRD, Colebatch JG. EMG responses in the soleus muscles evoked by unipolar galvanic vestibular activation. Electroenceph clin Neurophysiol 1997;105:476–483. Watson SRD, Colebatch JG. Vestibular-evoked EMG responses in soleus: a comparison between click and galvanic stimulation. Exp Brain Res 1998a;119:504–510. Watson SRD, Colebatch JG. Vestibulocollic reflexes evoked by short duration galvanic stimulation in man. J Physiol 1998b;513(2):587–597. Watson SRD, Brizuela AE, Curthoys IS, Colebatch JG, MacDougall HG, Halmagyi GM. Maintained ocular torsion produced by bilateral and unilateral galvanic (DC) vestibular stimulation in humans. Exp Brain Res 1998;122:453–458. Welgampola MS, Colebatch JG. Vestibulospinal reflexes: quantitative effects of sensory feedback and postural task. Exp Brain Res 2001a;139:345–353. Welgampola MS, Colebatch JG. Characteristics of tone burst-evoked myogenic potentials in SCM muscles. Otol Neurotol 2001b;22:796– 802. Young ED, Ferna´ ndez C, Goldberg JM. Responses of squirrel monkey vestibular neurons to audio frequency sound and head vibration. Acta Otolaryngol 1977;84:352–360. Zink R, Bucher SF, Weiss A, Brandt T, Dieterich M. Effects of galvanic vestibular stimulation on otolithic and semicircular canal eye movements and perceived vertical. Electroenceph clin Neurophysiol 1998;107:200–205.
Galvanic and acoustic vestibular stimulation activate different populations of vestibular afferents Ann M. Bacsi, Shaun R.D. Watson, James G. Colebatch* Institute of Neurological Sciences and UNSW Clinical School, Prince of Wales Hospital, Randwick, Sydney, NSW 2031, Australia Accepted 31 October 2002
Abstract Objective: To deduce whether similar or distinct populations of vestibular afferents are activated by acoustic and galvanic vestibular stimulation by comparing the effectiveness of ‘matched’ stimuli in eliciting vestibulospinal reflexes. Methods: Twelve subjects (5 men, 7 women) underwent individual ‘matching’ of 2 ms tone burst and galvanic stimuli, using vestibulocollic reflexes so that corrected reflex amplitudes to tone burst and galvanic stimuli were within 10% of each other. These same intensities were then administered using 20 ms durations to determine whether they were equally effective in evoking vestibulospinal responses. Results: Corrected reflex amplitudes for vestibulocollic responses to tone burst and galvanic stimulation were not significantly different for the right (P ¼ 0:45) or left (P ¼ 0:68) sides. All subjects had vestibulospinal responses to galvanic stimulation (average intensity 4.0 mA for both sides). The short latency (SL) and medium latency (ML) components of the vestibulospinal reflexes were larger after galvanic compared to tone burst stimulation in 11 of 12 subjects (P , 0:01). Conclusions: Despite evoking equal-sized vestibulocollic reflexes, there was a clear dissociation between the magnitude of tone burst and galvanic-induced vestibulospinal reflexes. Galvanic stimulation evoked SL and ML reflexes in all subjects. Tone burst stimuli evoked only small SL reflexes and, in most cases, no ML reflexes. Acoustically-evoked vestibulocollic reflexes are likely to be due to saccular excitation. The limited effectiveness of longer tone burst stimuli to evoke ML vestibulospinal reflexes suggests that saccular afferents have, at most, only a minor role in the production of these reflexes. We conclude that galvanic stimulation is more effective in eliciting vestibulospinal reflexes than tone burst stimulation, and that the two methods activate different populations of vestibular afferents. q 2002 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Galvanic; Vestibulocollic; Vestibulospinal; Human
1. Introduction Both acoustic (Colebatch et al., 1994; Cheng and Murofushi, 2001) and galvanic stimuli (Watson and Colebatch, 1998b) are capable of evoking short latency vestibulocollic reflexes in tonically contracting sternocleidomastoid (SCM) muscles. The waveform and latencies of both methods of stimulation are very similar; the responses show an initial positive deflection (p13) followed by a negative wave (n23) ipsilateral to the side of stimulation (Colebatch et al., 1994). Galvanic stimulation evokes a positive-negative response slightly earlier than that for tone bursts but is otherwise similar (Watson and Colebatch, 1998b; Welgampola and Colebatch, 2001a; Murofushi et al., 2002). The size of these responses scales with the level of tonic activation (Watson and Colebatch, 1998b). In the case of acousti* Corresponding author. Tel.: 161-612-9382-2407; fax: 161-612-93822428. E-mail address: [email protected] (J.G. Colebatch).
cally-evoked vestibular reflexes, the evidence strongly suggests an action on the saccule (Young et al., 1977; Cazals et al., 1983; Didier and Cazals, 1989; McCue and Guinan, 1995, 1997; Murofushi and Curthoys, 1997). Galvanic stimulation excites distal vestibular nerve afferents (Spiegel and Scala, 1943; Goldberg et al., 1984), particularly those with an irregular discharge (Goldberg et al., 1984), but any selectivity of this action on afferents arising from the different vestibular end organs is uncertain. Longer duration galvanic stimulation evokes a characteristic electromyographic (EMG) response in the soleus muscles of standing subjects (Britton et al., 1993; Fitzpatrick et al., 1994; Watson and Colebatch, 1997). This response consists of two parts- the short (SL) and medium latency (ML) components. It differs from the galvanicevoked vestibulocollic response in that it is strongly influenced by both postural task and head position, such that changing direction of head rotation leads to a reversal in polarity of both components. Although it is not certain
1388-2457/02/$ - see front matter q 2002 Elsevier Science Ireland Ltd. All rights reserved. doi:10.1016/S1 388-2457(02)00 376-0
CLINPH 2002118
360
A.M. Bacsi et al. / Clinical Neurophysiology 114 (2003) 359–365
that both components travel via the vestibulospinal tract (Britton et al., 1993), for simplicity we will use the term vestibulospinal for both reflexes. Watson and Colebatch (1998a) found that click vestibular activation could evoke vestibulospinal reflexes in some subjects which resembled those following galvanic stimulation. The effects following clicks had a similar appearance and latency to those following cathodal vestibular stimulation, i.e. the effects of clicks were consistent with an excitatory effect on vestibular afferents. However, these authors were unable to compare the relative effectiveness of the two methods given the very different durations of each type of stimulus. Our study was designed to deduce whether similar or distinct populations of vestibular afferents are activated by acoustic and galvanic stimulation by comparing the effectiveness of ‘matched’ stimuli in eliciting vestibulospinal reflexes. Tone bursts excite saccular afferents (McCue and Guinan, 1994, 1997) and evoke short latency reflexes which resemble those to clicks (Welgampola and Colebatch, 2001b). Tone bursts have the advantage that their duration can be safely varied to match that commonly used with galvanic stimulation in evoking vestibulospinal reflexes. Initially, the intensities of the acoustic and galvanic stimuli were adjusted so that the sizes of the corrected reflex amplitudes recorded from the SCM muscles were within 10% of one another. We then tested the relative effectiveness of these ‘matched’ stimulus intensities using longer (but equal) durations in evoking vestibulospinal reflexes. If both forms of vestibular stimulation were to activate a similar population of fibres and, in particular, if saccular fibres provide an important contribution to vestibulospinal reflexes, then the effect of both methods of stimulation should be similar for reflexes recorded in both the neck and the leg. Failure to see similar sized effects in the legs would allow us to conclude that each technique activates different populations of vestibular afferents. A further consequence of such an observation would be that the use of both techniques is likely to provide information about vestibular function additional to either one alone.
2. Methods 2.1. Subjects and study design Twelve healthy adults aged 25–46 (7 females, 5 males) were studied after obtaining informed consent and local ethics committee approval. The study consisted of two parts. Firstly, the initial ‘matching’ was performed by measuring vestibulocollic reflexes to short (2 ms) acoustic and galvanic (direct current) stimuli, with the latter adjusted to evoke the same size of short latency vestibulocollic reflex as the former. Secondly, recordings were made of vestibulospinal reflexes from over the soleus muscles, in standing subjects, using longer (20 ms) duration acoustic and galvanic stimuli. The current intensity was that previously deter-
mined as ‘matching’ the acoustic stimulus for vestibulocollic reflexes. For all acoustic stimulation rise time and fall time ¼ 0 ms. 2.2. Vestibulocollic reflexes Tone bursts (124 dB SPL at 500 Hz, 2 V peak to peak) were generated using a CED 1401 interface (Cambridge Electronic Design, Cambridge, UK) and phase-locked to the trigger pulse. The tone burst was delivered pseudorandomly to either the right or left ear through calibrated headphones (TDH 49, Telephonics Corp., USA). A total of 512 stimuli (256 per ear) were presented at 5/s. Subjects were tested in the recumbent position, on a couch with an adjustable backrest, while their heads were elevated against gravity to activate both SCM muscles. Electromyographic activity was recorded from Ag/AgCl surface electrodes (3 M, St Paul, MN, USA) over symmetrical sites on both sides of the neck, amplified and bandpass filtered (32 Hz–1.6 kHz). Active recording electrodes were placed on the upper third of the muscle belly and reference electrodes on the medial ends of the clavicles. An earth electrode was placed on either the upper end of the sternum or the medial forearm. EMG was sampled at 5000 Hz, from 20 ms before to 100 ms after stimulus onset using a 1401 plus analogue to digital converter and Sigavg software (Cambridge Electronic Design, Cambridge, UK) on a PC. Averages were made of the unrectified and rectified EMG signals. The reflex responses were measured from the unrectified average, while the rectified average was used to quantify the level of tonic activation. The ipsilateral p13n23 response was used for matching but the contralateral n12p20 response (Watson and Colebatch, 1998b) was also measured, if present. ‘Corrected reflex amplitudes’ were calculated by determining the peak to peak amplitude and then dividing by the mean rectified EMG activity level for the 20 ms preceding the stimulus onset to correct for background activation (Colebatch et al., 1994; Welgampola and Colebatch, 2001b). Corrected reflex amplitude or reflex gain is dimensionless and is largely independent of the level of tonic activity during the stimulus presentation (Colebatch et al., 1994). Short duration galvanic stimulation (model DS2 A; Digitimer, UK) was via electrodes of 27–30 cm 2, individually cut from commercially available electrosurgical plating, coated with electrode gel and placed over the mastoids. These were then secured with tape and an elastic bandage was wrapped around the head. An indifferent electrode was placed either over the spine at C7 or on the shoulder opposite to the side of stimulation. By convention we commenced this part of the experiment with the cathode over the left mastoid, followed by the cathode over the right mastoid. For galvanic-evoked vestibulocollic reflexes, the stimulating and recording sites are close together, and an artefact eliminator was used (Watson and Colebatch, 1998b). Two trials were performed, one at rest and the
A.M. Bacsi et al. / Clinical Neurophysiology 114 (2003) 359–365
other during contraction with the subject raising the head against gravity, with a brief rest period (,1 min) between recordings. In each, stimuli 2 ms long were delivered at 3/s to a total of 256 stimuli per trial. The resting trace (artefact only) was subtracted from the active trace (artefact plus response) to eliminate the stimulus artefact (Watson and Colebatch, 1998b). Vestibulocollic reflexes were then recorded beginning with a 4 mA current which was adjusted to produce a corrected p13n23 reflex amplitude within 10% of that for the tone burst-evoked vestibulocollic reflex. Early responses in the contralateral SCM were also measured, but not used for matching. 2.3. Vestibulospinal reflexes Nine of 12 subjects underwent vestibulospinal testing on a subsequent day. Three subjects underwent all testing on the same day. The vestibulospinal part of the experiment was performed with subjects standing, eyes closed, feet together, head rotated 90 degrees to the left and leaning slightly forwards. All subjects received a standard tone burst stimulus of 124 dB SPL at 500 Hz (2 V peak to peak), duration 20 ms and 256 tones per trial (128 for each ear; presented pseudorandomly as for the vestibulocollic part of the experiment). This stimulus is sufficient to activate the stapedial reflex but there is little attenuation after 400 ms (Welgampola and Colebatch, 2001b), thus, we chose a repetition interval of 500 ms. This tone burst is also sufficient to elicit a startle reflex, and these occur earlier and more frequently in standing subjects (Brown et al., 1991). Nevertheless we consider any effect of startle would have been minimal, both because the effects would have rapidly habituated with the rapid repetition of the stimuli and also because the reflex excitability would not have inverted for the two ears, a characteristic feature of vestibulospinal reflexes (Watson and Colebatch, 1997). Only reciprocal effects were measured in our estimation of the SL and ML reflexes. Unipolar cathodal stimulation was given at the intensity previously determined as ‘matching’ the corrected reflex amplitudes from the neck recordings for each subject. These stimuli were also of 20 ms duration, administered at 2 Hz with 128 stimuli per trial. By convention, we also commenced recordings with cathode left. Surface electrodes were placed over the soleus, 4.5–5.5 cm apart. EMG recordings were made from the soleus contralateral to the direction of head rotation (in this case, the right soleus), as this has previously been shown to be the larger response (Britton et al., 1993; Watson and Colebatch, 1997). EMG was amplified and filtered (32 Hz–1.6 kHz) and then sampled at 5000 Hz, from 50 ms before to 250 ms after stimulus onset. Each trial was rectified before averaging. The artefact eliminator was not required for this part of the experiment. Traces were analysed by superimposing the cathode-left and cathode-right recordings after digital filtering and
361
subtraction of prestimulus rectified EMG (Watson and Colebatch, 1997; Welgampola and Colebatch, 2001a). The SL reflex was defined as the response beginning between 40 and 70 ms and the onset taken as the point at which the traces of the two stimulus montages first diverged. The response immediately following the SL and with opposite polarity, commencing at about 90 ms after the stimulus onset, was taken as the end of the short latency response and the beginning of the ML response. The point at which the traces crossed over a second time was taken as the end of the ML response (Britton et al., 1993; Fitzpatrick et al., 1994). The SL and ML reflex amplitudes were expressed as a percentage of mean rectified EMG, after subtraction of the level of prestimulus EMG. This method allowed for differences in prestimulus EMG activity. The average of the two values obtained for the two stimulus polarities was calculated to give a percentage estimate of the overall size of the SL and ML reflexes (Welgampola and Colebatch, 2001a). Theoretically, because our vestibulospinal stimulus was relatively short, a response to its cessation (an ‘off’ response) might have affected the size of the ML reflex. Any such effect should have been constant for a constant SL reflex so should not have affected the hypothesis being tested. In the case of tone bursts, responses were often small and 10 of 12 subjects had repeat trials to verify the presence or absence of a response. If a (small) response was present for only one trial and appeared to be absent in the other, the values for SL and ML amplitudes for the positive trial were halved. 2.4. Statistical methods Student’s paired t tests and Pearson Correlations were used to compare the sizes of vestibulocollic and vestibulospinal reflexes. Chi squared and Fisher’s exact tests were used to analyse contralateral responses to tone burst and galvanic stimuli in the neck. Unless stated otherwise, results are given as mean ^ standard deviation. 3. Results 3.1. Vestibulocollic reflexes All subjects had an ipsilateral positive (p13) followed by a negative (n23) response to tone bursts of 124 dB SPL at 500 Hz (2 V peak to peak). The average latencies for the p13 and n23 peaks were 13.6 ^ 1.75 and 22 ^ 2.28 ms, respectively, for the right side, and 12.6 ^ 0.72 and 21.5 ^ 1.95 ms, respectively, for the left side. The results for right and left corrected reflex amplitudes were, respectively, 0.87 ^ 0.31 (range 0.28–1.31) and 0.86 ^ 0.32 (range 0.39–1.47). The average galvanic current intensity was 4.0 mA (range 2–6 mA; Table 1) and all stimuli were well tolerated. Overall, the galvanic-evoked reflex amplitudes were very closely matched, with corrected values for the right and left sides of 0.85 ^ 0.33 and 0.85 ^ 0.31 (Fig.
362
A.M. Bacsi et al. / Clinical Neurophysiology 114 (2003) 359–365
Table 1 Intensities of galvanic stimulation used to match vestibulocollic reflexes Subject
Right (mA)
Left (mA)
1 2 3 4 5 6 7 8 9 10 11 12
5 2 4.5 3.5 4 4.5 6 5 3 6 5 4.5
4.75 4 5 3 4 4.5 2.5 4 3 6 4 4
uniform – subjects with the largest galvanic-evoked SL reflexes had the greatest reduction in their responses to acoustic stimulation. There was a significant negative correlation between the relative size of the acoustic-evoked SL reflex and the SL reflex amplitude evoked by galvanic stimulation (r ¼ 20:81, P , 0:01). ML reflexes showed a different pattern. ML reflexes were nearly always absent in response to acoustic stimulation, being detectable in only 2 of 24 trials (with amplitudes of 3 and 4%). Overall, both SL and ML amplitudes were significantly smaller for tone burst
1). Matching the reflex evoked by galvanic stimulation to within 10% of the tone burst-evoked reflex was successful in 20 of the 24 ears. In four ears (three subjects) this was not possible; the galvanic stimulus amplitude used evoked smaller responses (by 26, 28 and 29%), or, in one case evoked a larger reflex (by 29%) than the target amplitude. Nevertheless, there was no significant difference between the corrected amplitudes to acoustic and galvanic stimulation on either the right (P ¼ 0:45, paired t test) or left sides (P ¼ 0:68, paired t test). Contralateral responses (i.e. from the left SCM with stimulation on the right and vice-versa) were more common with galvanic stimulation (Fig. 2). This difference was significant for right sided stimulation, (3 crossed responses to acoustic stimulation versus 8 crossed responses to galvanic stimulation; P , 0:05) but not for left sided stimulation (2 versus 6 crossed responses; P ¼ 0:08). Overall, contralateral responses were significantly smaller than ipsilateral responses for both methods of stimulation (tð11Þ ¼ 4:03–8:5; P , 0:002 in all cases). 3.2. Vestibulospinal reflexes Galvanic stimulation with an intensity ‘matching’ that for tone bursts, (i.e. an average of 4.0 ^ 1.0 mA) with a duration of 20 ms produced SL and ML reflexes in all subjects. The mean onset latencies were 53.8 ms (range 47.6–63.5) and 88.5 ms (range 84.6–93.4), respectively. Average SL amplitude (expressed as a percentage of baseline EMG activity) was 18 ^ 11% (range 7–42%) and, for ML, was 18 ^ 9% (range 8–32%). Tone bursts of 20 ms duration (124 dB SPL) produced an SL response in 16 of 24 trials with an average onset at 51.8 ^ 6.0 ms. The mean SL amplitude to the tone burst (as a percentage of baseline EMG activity) was 7 ^ 3% (Fig. 3). The tone burst-evoked SL reflexes had smaller amplitudes than the galvanic-evoked reflexes in 11 of 12 subjects (Fig. 4). In one subject with a small response to galvanic stimulation, there was a small increase (from 8.5 to 10%) in the SL response to tone bursts. The difference between the size of the galvanic-evoked SL reflex and that evoked by acoustic stimulation was not
Fig. 1. Matching of tone burst and galvanic vestibulocollic reflexes in a single subject, consisting of averaged EMG for 256 stimuli per ear for both acoustic and galvanic stimulation. The top pair of tracings was obtained after tone burst stimulation, and the corrected reflex amplitudes are 1.17 for both the right and left sides. The bottom pair of tracings shows the response to galvanic stimulation, where corrected reflex amplitudes are 1.08 for the right side and 1.14 for the left side. The dotted lines indicate stimulus onset.
A.M. Bacsi et al. / Clinical Neurophysiology 114 (2003) 359–365
363
Fig. 3. Vestibulospinal reflexes recorded over the right soleus muscle to galvanic and tone burst stimulation: rectified, averaged EMG traces are shown for 128 acoustic or galvanic stimuli. The DC component has been removed and traces have been digitally filtered. Galvanic stimuli were given in pulses of 20 ms. Intensities were 5 mA for the right side and 4.75 mA for the left side. Tone bursts were administered bilaterally at 124 dB SPL; each tone lasted 20 ms. The galvanic intensities were those which ‘matched’ the vestibulocollic reflex amplitudes. SL size for galvanic stimulation was 29% whereas no response was measured after tone burst stimulation. The dotted lines indicate stimulus onset. For galvanic stimulation, the dark line indicates results for cathode left; lighter line indicates results for cathode right. In the case of tone bursts, the dark line represents the EMG trace for left-sided acoustic stimulation; the lighter line represents the EMG trace for right-sided acoustic stimulation.
activation using acoustic stimuli appears to be selective for the saccule (Townsend and Cody, 1971; Cazals et al., 1983; Didier and Cazals, 1989; Murofushi and Curthoys, 1997). Murofushi and Curthoys (1997) found that click sensitive neurons were predominantly found in the inferior division of the vestibular nerve which supplies the saccule, although some were also present in the posterior part of the superior
Fig. 2. Ipsilateral and contralateral responses to right-sided galvanic and tone burst stimulation consisting of averaged EMG for 256 stimuli per ear for both methods of stimulation. The upper pair of traces shows the responses to galvanic stimulation, with corrected reflex amplitudes of 1.08 on the ipsilateral side and 0.36 for the contralateral n12p20 response. The lower pair of traces shows the responses to tone burst stimulation, where the ipsilateral side had a corrected reflex amplitude of 1.17, and the contralateral n12p20 was absent. The dotted lines indicate stimulus onset.
stimulation than for galvanic stimulation (tð11Þ ¼ 3:37, P , 0:01; tð11Þ ¼ 6:8, P , 0:001 for SL and ML, respectively). There was no significant correlation between the sizes of tone burst or galvanic-evoked vestibulocollic reflex amplitudes and corresponding SL reflexes for either right or left sides (P range 0.148–0.979).
4. Discussion Previous studies have indicated that the site of vestibular
Fig. 4. Comparison of sizes for galvanic and tone burst-evoked vestibulospinal reflexes: SL component. The left column of points corresponds to the size of the SL reflex evoked by galvanic stimulation. The right column shows the sizes of the SL reflexes evoked using tone burst stimulation. Individual subject results are paired by a joining line. Eleven of 12 subjects had a greater response to galvanic stimulation, a difference which was statistically significant (t ¼ 3:37, P , 0:01). Similar findings applied to the ML amplitudes. The dotted lines indicate subjects whose galvanicevoked vestibulocollic responses were not matched to within 10% of tone burst-evoked responses.
364
A.M. Bacsi et al. / Clinical Neurophysiology 114 (2003) 359–365
division of the vestibular nerve. In the same study, biocytin tracing of click sensitive neurons revealed nerve endings throughout the saccular maculae, whereas the utricular maculae were free of labelled nerve endings. Galvanic stimulation, in contrast, appears to work at the level of the vestibular nerve terminals. Spiegel and Scala (1943) found that destruction of the end-organ did not suppress the response to galvanic stimulation, whereas the response was abolished by sectioning the vestibular nerve. Goldberg et al. (1984) identified the site of action as the spike trigger zone. Cathodal stimulation increased, and anodal stimulation decreased, the discharge rate of vestibular afferents. Irregularly firing afferents were six times more sensitive than regularly firing afferents to applied current. Day et al. (1997) proposed that the central nervous system interprets galvanic stimulation as a tilt of the support surface (presumably an otolith effect), with resultant movement of body and head towards the anodal ear. Other studies (Zink et al., 1998; Watson et al., 1998) reported that low intensity (3 mA) galvanic stimulation applied to the mastoid processes may be relatively selective for the otolith, resulting in perceived tilt of the visual field and ocular torsion towards the anode but provoking minimal or no nystagmus. This finding is in contrast to earlier studies, such as Breson et al. (1971), who recorded horizontal nystagmus in 61 normal subjects following galvanic stimulation at 1.6 mA. Zink et al. (1998) recorded 5 s trials, which may have been too short to record sustained nystagmus which can require a ‘buildup’ time of up to 6 s (Karlberg et al., 2000). In addition, the study by Watson et al. (1998) involved a visual fixation point, known to suppress nystagmus, as well as a low sampling frequency of 2 Hz. Inter-individual differences may also influence the pattern of responses. Kleine et al. (1999) found that, after sinusoidal galvanic stimulation, some subjects had a large nystagmic response with little tonic response (indicating predominant semicircular canal stimulation) whereas other subjects had a large tonic response with little nystagmus (indicating predominant otolith stimulation). The evidence overall is compatible with galvanic stimulation having an action on both otolith and semicircular canal afferents, with significant inter-individual variation (MacDougall et al., 2002). Our experiment showed that despite evoking equal-sized vestibulocollic reflexes, there was a clear dissociation between tone burst and galvanic-induced vestibulospinal reflexes. There was also a trend for galvanic stimulation to evoke more crossed vestibulocollic responses, but this was only significant for right-sided stimulation. Future studies with larger sample sizes may clarify this issue. Vestibulospinal responses to galvanic stimuli were significantly larger than those for tone bursts in nearly all subjects. As noted above, acoustically-evoked vestibulocollic reflexes are likely to be the result of saccular excitation. The very limited effectiveness of longer tone burst stimuli in evoking ML reflexes suggests that saccular afferents have only a minor role at most in the production of this reflex.
The findings for the SL component of the vestibulospinal reflex were more complex. Our findings suggest that saccular afferents may be sufficient to evoke small SL reflexes but that additional vestibular afferents contribute in those subjects with large SL reflexes. In contrast to tone bursts, galvanic stimulation evoked prominent SL and ML vestibulospinal responses in all subjects. Previous studies have provided evidence that galvanic stimulation causes activation of otolith and semicircular canal afferents (see above). The SL reflex is thought to be mediated by vestibulospinal pathways (Britton et al., 1993). The main descending vestibular projection to the legs is the lateral vestibulospinal tract, originating from Deiter’s nucleus (Brodal, 1981). This, in turn, receives a major part of its input from otolith receptors. It is likely that galvanicevoked vestibulospinal reflexes are, at least in part, the result of otolith activation. If so, our results suggest that the saccular contribution is less important than afferents arising from the utricle, particularly for the ML component of vestibulospinal reflexes in soleus. The greater frequency of crossed responses to galvanic-evoked vestibulocollic reflexes would also be consistent with an effect on utricular afferents. Utricular afferents, unlike saccular afferents, have been shown to evoke crossed excitation in SCM motoneurons (Kushiro et al., 1999). Our observations indicate that different populations of vestibular afferents are activated by galvanic and tone burst stimulation. This, in turn, implies that tone burst and galvanic-evoked reflexes may probe different aspects of vestibular function. Acknowledgements This work was supported by the National Health and Medical Research Council of Australia. We thank Ms Sally Rosengren for assistance with data collection and editorial comments. References Breson K, Elberling C, Fangel J. Galvanic nystagmography. Acta Otolaryngol 1971;71(6):449–455. Britton TC, Day BL, Brown P, Rothwell JC, Thompson PD, Marsden CD, Postural EMG. responses in the arm and leg following galvanic vestibular stimulation in man. Exp Brain Res 1993;94:143–151. Brodal A. Neurological anatomy in relation to clinical medicine. 3rd ed. Oxford: Oxford University Press, 1981. pp. 202–205. Brown P, Day BL, Rothwell JC, Thompson PD, Marsden CD. The effect of posture on the normal and pathological auditory startle reflex. J Neurol Neurosurg Psychiatry 1991;54:892–897. Cazals Y, Aran J, Erre J, Guilhaume A, Aurousseau C. Vestibular acoustic reception in the guinea pig: a saccular function? Acta Otolaryngol 1983;95:211–217. Cheng P, Murofushi T. The effects of plateau time on vestibular-evoked myogenic potentials triggered by tone bursts. Acta Otolaryngol 2001;121:935–938. Colebatch JG, Halmagyi GM, Skuse NF. Myogenic potentials generated by a click-evoked vestibulocollic reflex. J Neurol Neurosurg Psychiatry 1994;57:190–197.
A.M. Bacsi et al. / Clinical Neurophysiology 114 (2003) 359–365 Day BL, Severac-Cauquil A, Bartolomei L, Pastor MA, Lyon IN. Human body-segment tilts induced by galvanic stimulation: a vestibularly driven balance protection mechanism. J Physiol 1997;500(3):661– 672. Didier A, Cazals Y. Acoustic responses recorded from the saccular bundle on the eighth nerve of the guinea pig. Hearing Res 1989;37:123–128. Fitzpatrick R, Burke D, Gandevia SG. Task-dependent reflex responses and movement illusions evoked by galvanic vestibular stimulation in standing humans. J Physiol 1994;478.2:363–372. Goldberg JM, Smith CE, Ferna´ ndez C. Relation between discharge regularity and responses to externally applied galvanic currents in vestibular nerve afferents of the squirrel monkey. J Neurophysiol 1984;51:1236– 1256. Karlberg M, McGarvie L, Magnusson M, Aw ST, Halmagyi GM. The effects of galvanic stimulation on the human vestibulo-ocular reflex. Neuroreport 2000;11:3897–3901. Kleine JF, Guldin WO, Clarke AH. Variable otolith contribution to the galvanically induced vestibulo-ocular reflex. Neuroreport 1999;10:1143–1148. Kushiro K, Zakir M, Ogawa Y, Sato H, Uchino Y. Saccular and utricular inputs to SCM motoneurons of decerebrate cats. Exp Brain Res 1999;126(3):410–416. MacDougall HG, Brizuela AE, Burgess AM, Curthoys IS. Between-subject variability and within-subject reliability of the human eye-movement response to bilateral galvanic (DC) vestibular stimulation. Exp Brain Res 2002;144:69–78. McCue MP, Guinan JJ. Acoustically responsive fibres in the vestibular nerve of the cat. J Neurosci 1994;14:6058–6070. McCue MP, Guinan JJ. Spontaneous activity and frequency selectivity of acoustically responsive vestibular afferents in the cat. J Neurophysiol 1995;74.4:1563–1572. McCue MP, Guinan JJ. Sound-evoked activity in primary afferent neurons of a mammalian vestibular system. Am J Otol 1997;18:355– 360. Murofushi T, Curthoys IS. Physiological and anatomical study of click-
365
sensitive primary vestibular afferents in the guinea pig. Acta Otolaryngol 1997;117(1):66–72. Murofushi T, Takegoshi H, Ohki M, Ozeki H. Galvanic-evoked myogenic responses in patients with an absence of click-evoked vestibulocollic reflexes. Clin Neurophysiol 2002;113(2):305–309. Spiegel SA, Scala NP. Response of the labyrinthine apparatus to electrical stimulation. Arch Otol 1943;38:131–138. Townsend GL, Cody DTR. The averaged inion response evoked by acoustic stimulation: its relation to the saccule. Ann Otorhinolaryngol 1971;80:121–131. Watson SRD, Colebatch JG. EMG responses in the soleus muscles evoked by unipolar galvanic vestibular activation. Electroenceph clin Neurophysiol 1997;105:476–483. Watson SRD, Colebatch JG. Vestibular-evoked EMG responses in soleus: a comparison between click and galvanic stimulation. Exp Brain Res 1998a;119:504–510. Watson SRD, Colebatch JG. Vestibulocollic reflexes evoked by short duration galvanic stimulation in man. J Physiol 1998b;513(2):587–597. Watson SRD, Brizuela AE, Curthoys IS, Colebatch JG, MacDougall HG, Halmagyi GM. Maintained ocular torsion produced by bilateral and unilateral galvanic (DC) vestibular stimulation in humans. Exp Brain Res 1998;122:453–458. Welgampola MS, Colebatch JG. Vestibulospinal reflexes: quantitative effects of sensory feedback and postural task. Exp Brain Res 2001a;139:345–353. Welgampola MS, Colebatch JG. Characteristics of tone burst-evoked myogenic potentials in SCM muscles. Otol Neurotol 2001b;22:796– 802. Young ED, Ferna´ ndez C, Goldberg JM. Responses of squirrel monkey vestibular neurons to audio frequency sound and head vibration. Acta Otolaryngol 1977;84:352–360. Zink R, Bucher SF, Weiss A, Brandt T, Dieterich M. Effects of galvanic vestibular stimulation on otolithic and semicircular canal eye movements and perceived vertical. Electroenceph clin Neurophysiol 1998;107:200–205.