Tapping the head activates the vestibular system

Loud clicks can activate vestibular receptors in normal subjects. We have shown that brief (0.1 msec), loud (more than 95 dB normal hearing level [NHL]), square-wave clicks produce a short-latency, surface-positive potential in the ipsilateral sternocleido-mastoid (SCM) muscle. These click-evoked potentials appear to be generated by a vestibulocollic reflex possibly originating in the saccule. ^[1,2]

Although click-evoked vestibular potentials can be present from ears with profound sensorineural deafness, they are attenuated by even mild to moderate conductive deafness. ^[3] We wondered whether vestibular activation might be produced by tapping the skull, thus circumventing the normal middleear conductive mechanism in a manner analogous to that for bone conduction tests of hearing. We studied whether a gentle head tap with a standard clinical reflex hammer would reliably evoke shortlatency EMG potentials with characteristics similar to those elicited by loud clicks.

Methods.

We studied 15 normal subjects and 20 patients. Ten patients had undergone therapeutic unilateral vestibular neurectomy for intractable vertigo caused by Meniere's disease, six had severe unilateral conductive hearing loss with no vestibular abnormalities, and four had total bilateral sensorineural hearing loss with no vestibular abnormalities. For recording, the patients reclined, raising their heads to activate the SCM muscles to a target level of EMG activity. Surface EMG recordings were made from an active electrode over the uppermost part of each SCM muscle. This recording technique results in slightly shorter latencies for click-evoked potentials than previously reported. ^[2] The unrectified EMG responses were averaged (n equals 128 to 512) following either clicks to one ear or skull taps delivered manually, through a pad at Fz, with a small nylon-handled reflex hammer (Keeler, London, UK) Figure 1. The hammer was fitted with an inertial trigger switch that produced a delay of less than 3 msec. The deceleration at impact was measured to be about 90 g over 2.5 msec.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1. The technique for eliciting tap-evoked vestibular myogenic potentials. The subject reclines and elevates the head about 30 degrees to activate both sternocleidomastoid muscles. The active surface electrode is placed over the upper end of each muscle and the reference electrode over the sternoclavicular joint. The forehead is tapped, through a small pad at Fz, with a nylon-handled reflex hammer fitted with an inertial trigger switch.

Results.

Figure 2 shows a typical result from a normal subject. About 7 msec after the tap stimulus, the response began simultaneously in the two SCM muscles, with a positive potential peak at about 10 msec followed by a negative peak at 17 msec. In the normal subjects, the mean latency of the initial positivity was 9.8 msec; the mean latency of the initial negativity was 16.5 msec. The average peak-to-peak amplitude was 243 mu V, which was almost always larger than the initial response to 100 dB NHL clicks in the same subjects. Additional waves followed, and these were not abolished by selective vestibular neurectomy. The initial positive-negative waves were absent on the side ipsilateral to the selective vestibular neurectomy in all 10 patients who had undergone this procedure but were preserved in all six with severe conductive hearing loss Figure 3. The early potentials were present bilaterally in the four patients with total bilateral sensorineural deafness.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2. Normal tap-evoked vestibular myogenic potentials recorded from the right (top trace) and left (bottom trace) sternocleidomastoid muscles of a normal subject in response to taps at Fz. I equals initial positive peak; II equals initial negative peak. Time scale divisions equals 5 msec; amplitude scale divisions equals 200 mu V.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3. (Left column) Results from a patient after a selective left vestibular neurectomy. Taps caused a normal potential from the right sternocleidomastoid muscle (SCM), shown in the third trace, with an initial positivity (I) at 7.9 msec and an initial negativity (II) at 13.4 msec (I to II amplitude equals 125 mu V); there was no comparable response from the left SCM (fourth trace). Click stimulation of the right ear was followed by a normal potential from the right SCM (first trace) with an initial positivity (I) at 10.9 msec and an initial negativity (II) at 17.8 msec (I to II amplitude equals 169 mu V). There were no comparable potentials from the left SCM (second trace) in response to stimulation of the left ear. Triggering delay probably explains the shorter latency to taps than to clicks (see text). (Right column) Results from a patient with a 40-dB left-sided conductive hearing loss, no vestibular symptoms, and normal caloric tests. In response to taps, the latency of the initial positivity (I) is 10.2 msec from the right SCM (third trace) and 10.7 msec from the left SCM (fourth trace); the latency of the initial negativity (II) is 19.1 msec from the right SCM (I to II amplitude equals 278 mu V) and 19.9 msec from the left (I to II amplitude equals 282 mu V). There were no click-evoked potentials from the left SCM in response to stimulation of the left ear (third trace), but there were normal potentials from the right SCM in response to stimulation of the right ear (first trace; initial positivity equals 10.3 msec; initial negativity equals 18.4 msec; amplitude equals 139 mu V). Time scale divisions equals 5 msec.

Discussion.

Primary vestibular afferents in the monkey can be activated by vibration as well as by sound, ^[4] possibly in the same way as bone-conducted sound activates cochlear receptors. We showed that gentle tapping of the skull evokes short-latency, bilateral EMG responses in the SCM muscles that resemble the click-evoked vestibulocollic reflex. Stimulation in the midline on the frontal bone, like bone conduction of sound to the cochlea, ^[5] gave the most reproducible results, even though this means that the vestibular end-organs on both sides are simultaneously activated.

The tap-evoked response, like the click-evoked response, consists of an initial positive-negative wave and begins less than 10 msec after the trigger pulse. The slightly shorter latencies of the tapevoked responses compared with the latencies of the click-evoked responses are probably caused by the mechanical delay of the trigger. Tap-evoked responses are generally larger in amplitude than the corresponding click-evoked responses, a result that could be due either to more effective activation of the same vestibular afferents as are excited by clicks or to activation of additional vestibular afferents. Although saccular primary afferents are more sensitive to sound than are other vestibular afferents, there is no such selectivity in their sensitivity to vibration ^[4]; thus, caution must be exercised in directly equating the click-evoked responses with the tap-evoked ones.

The tap-evoked response, like the click-evoked response, is also followed by additional waves that are unlikely to be of vestibular origin. In particular, the negative wave of the vestibular-dependent response can be indistinguishable from the later, often larger, nonvestibular negativity. In practice, we accept only latencies very close to those described, and prefer to be able to identify two separate peaks before we feel confident that a negative potential is truly vestibular dependent.

Because of impairment of transmission through the middle ear, the click-evoked vestibular response is attenuated by conductive hearing loss. ^[3] Bone conduction is a complex process but depends in part on a direct compression wave. ^[6] As predicted, we found that tap-evoked vestibular responses, unlike the click-evoked responses, were not attenuated by middle ear conduction abnormalities. On the other hand, with taps, unlike with clicks, selective activation of one vestibular apparatus is not possible. Clinically, it is likely that the two techniques will prove more useful together than separately.