Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Hearing Research, 55 (1991) 9-23 0 1991 Elsevier Science Publishers HEARES B.V. All rights reserved 037%5955/91/$03.50 01596 Effect of high-frequency hearing loss on compound action potentials recorded from the intracranial portion of the human eighth nerve Aage R. Mgller and Hae Dong Jho Department of Neurological Surgery, University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania, U.S.A (Received 12 September 1990; accepted 30 January 1991) Compound action potentials (CAP) were recorded from the exposed intracranial portion of the eighth nerve to stimulation with click sounds in patients with sensorineural high-frequency hearing loss who underwent microvascular decompression (MVD) operations to treat trigeminal neuralgia (TN). In patients with normal hearing the CAP recorded in that way is characterized by a negative peak, preceded by a small positivity and followed by a positivity and sometimes a second negative peak. In patients with high-frequency hearing loss the CAP also usually had an initial sharp negative peak in response to clicks of high intensity (105 to 110 dB Pe SPL), similar to findings in patients with normal hearing, but in patients with high-frequency hearing loss the initial negative peak was often followed by a slow negative deflection. The latency of the initial negative peak in the CAP in patients with high-frequency hearing loss was longer than the latency of this peak in patients with normal hearing, but the difference in latencies of this peak to condensation and rarefaction clicks was small. When the stimulus intensity was lowered the amplitude of the initial peak decreased, and the CAP became dominated by a broad negative peak with a latency of 6 to 8 ms. In 11 of 15 patients with severe high-frequency hearing loss, a series of quasiperiodic waves was superimposed on the CAP. The frequency of these waves varied between 500 and 1200 Hz, and the waves could be detected between 6 and 16 ms after presentation of the click stimulus. These waves were usually present in the response to stimuli in the intensity range from 75 to 110 dB Pe SPL. Only 4 of 17 patients with normal hearing had similar waves. Human auditory nerve; Compound action potentials; Intraoperative Introduction Studies of the neural generators of the brainstem auditory evoked potentials (BAEP) have shown that it is difficult to draw conclusions about the human auditory system from the results of studies in animals. Thus, although the results of animal studies have advanced our understanding of the anatomy and physiology of the auditory system in general, the performance of studies in man has been crucial for the identification of the neural generators of the BAEP in man. This is also true for studies of the pathophysiology of the human auditory system. Studies of the neural generators of the BAEP, on the basis of intracranially recorded evoked potentials, have also mostly been performed in patients with normal hearing (Moller et al., 1981a,b; Hashimoto et al., 1981). In previous studies (Moller and Jho, 1988, 1989a,b, 19911, we showed how the compound action potential (CAP) that can be recorded from the exposed intracranial portion of the eighth nerve in patients with normal hearing varies as a function of stimulus intensity and Correspondence to: Aage R. Moller, Surgery, 9402 Presbyterian-University Pittsburgh, PA 15213, U.S.A. Department of Neurological Hospital, 230 Lothrop Street, recordings; Active cochlear processes click polarity. These studies were performed in patients who underwent operation for vascular compression of the fifth and seventh cranial nerves by the microvascular decompression (MVD) technique (Jannetta 1977, 1981a,b). We have also shown that in some patients a series of more or less periodic waves appears superimposed on the recorded potentials (Moller and Jho, 1990). Because patients who undergo operation for microvascular decompression of cranial nerves may have normal hearing or various forms of hearing loss, related or unrelated to the disorder for which they are undergoing operation, recording CAP directly from the exposed intracranial portion of the eighth nerve during such operations provides an opportunity to study these potentials in patients with various types of hearing loss as well as in patients with normal hearing. We showed earlier (Moller et al., 1991) that the CAP recorded in response to broadband clicks from patients with various degrees of hearing loss have a multitude of different shapes. In the present paper we present the results of recording CAP from the exposed intracranial portion of the eighth nerve in patients with high-frequency hearing loss, while the patients underwent MVD operations of the fifth cranial nerve to treat trigeminal neuralgia (TN). We show how the CAP recorded from the exposed eighth nerve changes as a function of the stimulus intensity, and we analyze the waves that are superimposed on the late slow components of the CAP in some of these patients. Methods The results reported in the present paper are based on recordings obtained intraoperatively in patients undergoing microvascular decompression (MVD) of the fifth cranial nerve to treat trigeminal neuralgia (TN). In all patients a complete audiological and otological evaluation was performed before the operation. Testing included pure tone audiometry, speech audiometry using recorded lists (NW61, tympanometry, and recording of the acoustic middle ear reflexes. The audiological tests were performed in a sound-insulated booth by a certified audiologist using modern, calibrated equipment (Msller and Moller, 1989): Patients whose pure tone thresholds were elevated by 20 dB or more at more than 2 frequencies of presumed cochlear origin were included in the present study, but those with a conductive hearing loss or a hearing loss that was presumed to be a result of an auditory nerve lesion were excluded. Thus, only patients who had nearly the same hearing loss on both sides and who had speech discrimination scores and acoustic middle ear reflex thresholds within the range expected with regard to their pure tone audiograms were included in the study (although one patient had not mastered the English language sufficiently to permit speech discrimination testing to be performed). These selection criteria were chosen to exclude patients who might have had a lesion on the auditory nerve. Earlier reports describe the operative techniques used (Jannetta, 1977, 1981a,b) and the technique that was used to record brainstem auditory evoked potentials (BAEP) and compound action potentials (CAP) from the exposed eighth cranial nerve (Moller and k Jannetta, 1983; Moller, 1988a; Msller and Jho, 1988, 1989a,b), and therefore only a brief description of these methods will be given here. BAEP were recorded intraoperatively from the time that the patient was anesthetized to the time that the wound was closed. Click sounds were generated by applying 100~PS duration rectangular waves (delivered by a Grass Instrument Co., audio-stimulator, Type SlO CTCMA) to miniature stereo earphones (Radio Shack, Realistic) that were fitted into the patient’s outer ear and kept in place with adhesive tape. The ambient noise level in the operating room is a limiting factor as to how low sound levels can be studied. In order to obtain the best possible sound insulation in the ear that was tested, the earphone was placed so that it fit tightly over the entrance to the patient’s ear canal, and it was sealed in the outer ear in an water-tight fashion, with several layers of adhesive plastic tape being applied so that tape covered the entire outer ear. All the sounds were presented at a rate of 19 pps. The earphones were calibrated by attaching them to a 2.5-cc slightly damped cavity to which a l/4-inch condenser microphone (Bruel and Kjaer, Type 4135) was attached. The sound pressure of the click sounds was also measured at the entrance of the ear canal using the same l/4-inch condenser microphone. The click sounds presented in that way produced a waveform characteristic of a slight ringing (Fig. 1A) and had a spectrum that included a broad peak around 600 Hz and a smaller peak around 5 kHz (Moller and Jho, 1991). The spectrum of the lOO-l.~srectangular pulse used to drive the earphone has a ‘null’ at 10 kHz, which also is reflected in the spectrum of the sound (Fig. 1B). The clicks of opposite polarity had identical waveforms (except that the waveforms were also inverted) and their spectra were not noticeably different. The mean threshold of click sounds presented at a rate of 19 pps was determined in 4 persons with normal hearing (threshold equal to or better than 10 dB be- !I Fig. l.(A) Waveform of the click sound recorded at the entrance of the ear canal. A l/&inch condenser microphone (Bruel and Kjaer. Type 4135) had been placed in the recess between the tragus and the antitragus, and the earphone had been positioned for intraoperative recording. Solid lines, condensation click. Dashed lines, rarefaction click. Peak equivalent sound pressure level (Pe SPL) was 105 dB. (B) Spectrum of the sound (condensation click), the waveform of which is shown in A. 11 tween 250 and 4000 Hz and 20 dB or better between 6 and 8 kHz) with the earphone placed in the same way as in the operating room. This threshold was 39.5 f 1.9 dB Pe SPL (range, 36 to 41 dB). The CAP from the exposed eighth nerve were recorded using a monopolar electrode made from a multistrand Teflon-insulated silver wire (Medwire Corp., Type Ag 7/40) to the uninsulated tip of which a small cotton wick was sutured (Mgller and Jannetta, 1983; Moller, 1988a). The cotton wick was placed on the exposed eighth nerve, and the reference electrode for the recording of CAP from NV111 was placed on the opposite earlobe. The potentials were amplified (Grass Instrument Co., ac amplifiers, Type 12) at a gain of 20,000 x . Filter settings were 3 Hz highpass (6 dB point with 6 dB/octave rolloff) and 3000 Hz lowpass (6 dB point with 24 dB/octave rolloff). The recorded potentials were averaged using an LSI 11/73 microprocessor or an Apollo DM 3000 computer with a sampling interval of 40 or 80 &S and 256 or 512 datapoints in each record. In the operating room a baseline record was displayed together with a display of the actual record for easy detection of changes in the recorded potentials. BAEP were always recorded simultaneously with CAP recorded from the eighth nerve, and the averaged responses were stored together on computer disks for further processing. The differences between responses recorded to condensation and rarefaction clicks were obtained, and these differences were highpass filtered, to eliminate low-frequency components, with a zero-phase digital highpass filter. The power spectra of these difference records were obtained by computing (fast) Fourier transforms (using a Hanning window) of autocorrelograms of the highpass filtered difference records (see Blackman and Tukey, 1959). The results presented in this paper were obtained during intraoperative monitoring of evoked potentials that is performed routinely in patients undergoing MVD operations at our institution. The procedure was approved by the Biomedical Institutional Review Board of the University of Pittsburgh School of Medicine, and the patients gave their informed consent to participation in monitoring and use of the results for scientific purposes. Results In patients with moderate degrees of high-frequency hearing loss (Fig. 2A), the shape of the CAP (Fig. 3) is rather similar to what is seen in patients with normal hearing (Fig. 4), namely a large and sharp negative peak, preceded by a small positivity. The large negative peak is followed by a small negative peak. When the stimulus intensity is decreased the latency of the initial Fig. 2. Audiograms (A through E) of the 5 patients studied. Speech discrimination score is given in insert (the patient without a speech discrimination score had not mastered the English language sufficiently for this testing to be performed). Lower data points are acoustic middle ear reflex thresholds. 110 105 creased, but its latency remained almost independent of the stimulus intensity. When stimulus intensity was less than 90 dB there was no discernible initial peak, and the responses consisted of a single broad, negative peak. The response to llO-dB clicks also showed a broad negative deflection that directly followed the initial negative peak and lasted for about 10 ms. When the stimulus intensity was decreased this negative deflection shifted towards longer latencies, and it appeared as a separate peak at stimulus intensities of 100 dB and below. There were periodic waves superimposed on this broad peak, and the phase of these waves reversed exactly when the polarity of the clicks was reversed. Such waves were not seen in the patient whose results are shown in Fig. 3. d0 dB 110 dB ‘,, , 65 75 dB 105 d6 100 dE dB I I I I I 1 I I I 0 2 4 6 6 10 12 14 16 Fig. 3. CAP obtained shows a high-frequency upward deflection. in the patient hearing dashed lines. The amplitude the calibration to condensation calibrations I 16 whose audiogram loss. Negativity The responses to rarefaction solid lines and the responses ‘I 20 ms in Fig. 2,A is shown as an clicks are shown as clicks are shown as apply to the records above bars, and the sound pressure 95 dt? 65 dB is given in dB Pe SPL. The same conventions are used in all subsequent figures. negative peak increases and the peak broadens. There is only a slight difference between the responses to condensation and rarefaction clicks. In patients with severe high-frequency hearing loss (Figs. 2B and 5) there is a distinct initial peak in the response to high-intensity clicks (110 to 90 dB), similar to the peak seen in patients with normal hearing except that its latency (4.2 ms) is longer than the latency of the peak in patients with normal hearing (Msller et al., 1988). The amplitude of this initial negative peak decreased rapidly when the stimulus intensity was de- I I I I I I I I I I I 0 1 2 3 4 5 6 7 6 9 10 Fig. 4. Typical CAP obtained in a patient with normal hearing ms 13 110 dB 105 dB waves superimposed on this broad peak. The patient whose results are shown in Fig. 6 is the only patient reported on in this paper who had a BAEP component that led to a suspicion of retrocochlear involvement. However, this patient’s acoustic middle ear reflex and speech discrimination scores were normal and his hearing loss was almost equal on both sides. For the patient whose results are shown in Fig. 7, an initial negative peak was present in the CAP in response to high-intensity clicks, but this peak was different for condensation and rarefaction clicks. This patient also had waves superimposed on the slow negative deflection that followed the initial peak, the phase ’ ‘:, ‘L, ~j\,> 100 dB h“l__ \ \ 85 110 d0 A.--- \ -. dB i \ 80 dB -\ 105 dB 95 dB 65 dB 75 dB \ \ ; : / -\ 60 _;_ .L \, dB _- \ f-_-. ‘-_ _--- ’ I-_-u -_- --._. .- 76 -_ ---_ .--. 17_>” - d0 I I 0 2 4 I I I I I I I I 6 6 10 12 14 16 16 20 Ills Fig. 5. Results similar to those shown in Figs. 3 and 4, but from another patient. This patient’s audiogram is shown in Fig. 2B. Similar results were obtained in two other patients with severe high-frequency hearing losses (Figs. 2C and 6 and Figs. 2D and 7). In the patient whose results are shown in Fig. 6 there was almost no initial negative peak in the CAP, even at 110 dB, and the response consisted of a broad negative peak with a latency between 8 and 10 ms that resembled the response at 95 dB of the patient illustrated in Fig. 5. Also in the patient illustrated in Fig. 6 there were several distinct I I 1 0 2 4 I I I B 6 10 _ _ I I I I I 12 14 16 16 20 tn8 Fig. 6. Recordings similar to those shown in Figs. 3, 4. and 5. but obtained in another patient. This patient’s audiogram is shown in Fig. 2C. 13 110 dt3 curves. And, in fact, these early components in the difference curves decrease rapidly in amplitude when the stimulus intensity is decreased, although clear oscillations may be seen to occur over a large range of stimulus intensities. These oscillations occur between 6 and 16 ms after the click stimulus, and when examined closely these oscillations are found to be of slightly higher frequency in the early portion than in the portions with longer latencies. Corresponding multiple peaks are seen between fiO0 and 1200 Hz in the power spectra of these difference functions. It is also notable that the individual waves appear with almost the same latency at the different stimulus intensities. ,,^_. ‘/ 105 95 dB dtl 110 dB ..I 85 i_.- dB I_. .-i. --. :’ - _.,“.Z r i ‘J 105 df ‘. 1 1 0 2 I 4 1 I I I I I I I 6 6 10 12 14 16 18 20 ,- -- /lOpV / ins Fig. 7. Recordings similar to those shown in Figs. 3. 4, 5, and 6, but obtained in another patient. This patient’s audiogram is shown in Fig. 2D. of which reversed exactly when the polarity of the click stimuli were reversed, In a patient with both high- and low-frequency hearing loss, the waveform of the CAP had the normaf triphasic shape in response to cIicks of high intensity (105 dB), and there was little difference in the CAP in response to condensation and rarefaction clicks (Figs. 2E and 8). When the stimulus intensity was decreased to 90 dB or beIow in this patient, the CAP changed to a broad negative deflection on which numerous waves were superimposed, and the phase of these waves reversed when the click polarity was reversed. The nature of the waves that are superimposed on the CAP in patients with hearing loss can be understood better by examining Figs. 9-11, which show curves representing the difference between the responses to clicks of opposite polarity, together with the power spectra of these difference functions. The initial deflection in these curves reflects the difference between responses to rarefaction and condensation clicks in the initial triphasic component of the CAP. A difference in the latencies of the initial components in the responses to condensation and rarefaction clicks appears as a large, early deflection in the difference 05 dB I ,i ;. 80 dEi 0 2 8 6 6 10 $2 14 16 18 20 ms to those shown in Figs. 3, 4, 5, 6, and 7, Fig. 8. Recordings imiiar but obtained in a patient with both high- and low-frequency hearing loss. This patient’s audiogram is shown in Fig. 2E. 15 It may be seen by reviewing Figs. 9-12 that the late components of the CAP increased with increasing stimulus intensity. This is illustrated in Fig. 13, which shows the root mean square (RMS) values of the difference between the responses to condensation and rarefaction cficks within the interval 4.5-14.5 ms (Fig. 13A). Also shown is the RMS value of the sum of the responses to clicks of opposite polarity (Fig. 13B). Fig. 13 shows that both these measures increase at about the same rate with increasing stimulus intensity, but the RMS value of the sum of the responses is about 5 times larger than the RMS value of the sum of the differences. There was usually little change in the waveform of the recorded potentials during the intradural portion of the operation, but the amplitudes of the potentials changed in relation to how wet the recording site became (although the recording electrode was never covered by fluid when recordings were being made). A comparison of the late components obtained at different times after the beginning of the intracranial por- 105 d8 75 d8 tion of the operation shows little difference in these components over time, except for a slight change in amplitude (Fig. 14). When the recording electrode was moved from a middle position on the intracranial portion of the eighth nerve to a more distal location near the porus acusticus, the fatency of ail components decreased, which is in good agreement with the assumption that the recorded potentials are a result of propagated neural activity in the auditory nerve. Of the 4 patients with high-frequency hearing losses whose results were reported on in the present paper, 3 had such late quasiperiodic waves in their CAP. Of 6 other patients with high-frequency hearing losses who were operated upon to relieve trigeminal neuralgia (TN), 4 had such waves in their CAP, and of 5 patients with both high- and Iow-frequency hearing losses 4 had similar waves in their CAP. In contrast, only 4 of 17 patients with normal hearing who were operated upon to relieve TN had detectable waves in their CAP. Thus, although the late waves occur more frequently Frewuency in kHz Fig. 9. Difference between the responses to condensation and rarefaction clicks (lefthand curves) in the patient whose audiogram and CAP are shown in Figs. 2B and 5, and the power spectra (righthand graphs) of these differences. The power spectrum was computed from a 10.24-ms-long autocorrelation of fhe response shown to its left. in patients with high-frequency hearing loss, they are also seen in patients with normal hearing. In these patients with normal hearing or a mild hearing loss the late components often consist of a single wave (Fig. 1%. A typical example of a recording from a patient who had a noticeable high-frequency hearing loss but no discernible late components is shown in Fig. 16. It is seen that the initial negative peak is similar to what is 110 dB 105 dB 100 dB 95 dB 90 65 60 75 seen in patients with normal hearing. However, there is no discernible response at 90 dB Pe SPL and below, due to the patient’s hearing loss. Discussion The results of the present study show that the waveform of the CAP that can be recorded from the 110 dB 105 dB 100 dB dB 95 dB 90 dB 65 dB 80 dB 75 dB dB dB dB _-_- 5 I 0 I I I I I I I I I I 2 4 6 a 10 12 14 16 16 20 ms 0.02 0.05 0.1 0.2 0.5 Frequency Fig. IO. Results similar to results shown in Fig. 9, but for the patient whose audiogram 0.7 1 in 2 3 5 7 dB 10 kHz and CAP are shown in Figs. 2C and 6. The small peak! seen in the spectra at about 1500 Hz are probably a result of electrical interference. 17 intracranial portion of the eighth nerve in patients with different degrees of high-frequency hearing loss is more complex than the waveform of the response in patients with normal hearing. The CAP recorded from the eighth nerve in patients with high-frequency hearing loss has a broad peak with a latency between 6 and 10 ms in the response to clicks presented at an intensity of 90 dB or less. In a high proportion of such patients, quasiperiodic oscillations occur in the interval between 6 and 16 ms after the stimulus. Similar waves are seen less often in patients with normal hearing. Based on the latency and the frequency of oscillation of these components in the CAP in response to broadband clicks, they are believed to be a result of excitatory vibrations of the apical portion of the basilar membrane. Because the CAP recorded from a nerve is the convolution of a single nerve impulse with the weighted sum of the temporal distribution of nerve impulses in all the nerve fibers contributing to the response (Goldstein, 19601, the waveform of the CAP recorded from the eighth nerve reflects the distribution in time of 110 dB nerve impulses in the population of nerve fibers that are activated. When the waveform of a single action potential is known, the distribution of nerve impulses can be estimated by deconvolution of the recorded CAP and a single nerve action potential. Deconvolution of two waveforms in the time domain is equivalent to dividing the spectra of the two waveforms. Because the spectrum of a single action potential has a bandpass characteristic, rapid changes in the distribution of nerve impulses will appear more prominently in the CAP than slow changes, and steady-state discharge rates are not reflected at all in the CAP. The sharp initial wave of the CAP may therefore reflect the distribution of nerve impulses in time more accurately than the broad wave in the CAP, and the distribution of nerve impulses in time is likely to be even broader than the broad peak in the CAP. This is naturally predicated on the assumption that the waveforms of single nerve impulses are all the same and not noticeably different in patients with hearing loss from waveforms of nerve impulses in patients with normal hearing. That this is indeed the case is indicated by the fact n 105 dB 95 dB 110 65 75 I dB dB dB I I I I I I I I I 95 dB 55 d0 75 dB I Frequency Fig. 11. Results similar to results shown in kHz in Fig. 9, but for the patient whose audiogram and CAP are shown in Figs. 2D and 7. The sharp seen above about 1800 Hz are the result of electrical interference. peaks that the initial negative peak in the CAP elicited by high-intensity sounds is similar in patients with hearing loss and those with normal hearing. Thus, it seems reasonable to assume that the hearing loss of the patients presented in this study was a result of cochlear impairment and thus was unlikely to have caused any change in the waveforms of single nerve impulses. The CAP that can be recorded from the inner ear using the technique known as electrocochleography (ECoG) is in many ways similar to the CAP that can be recorded from the intracranial portion of the eighth nerve. Late and broad peaks in the CAP from the eighth nerve, similar to those seen in the present study, have been reported in the ECoG response of a patient with high-frequency hearing loss (Elberling, 1974). In these recordings from the ear canal, the initial sharp negative peak that was present at high sound intensities was lost when the stimulus intensity was lowered to 75 dB Pe SPL, and below this stimulus intensity only a slow and broad peak remained. Elberling (1974) used masking with highpass-filtered noise, similar to a technique used in animal experiments (Teas et al., 1962), to determine derived action potentials in normal-hearing 110 dB 105 dB 95 human subjects. These derived responses were assumed to represent the responses of specific areas of the basilar membrane, and Elberling (1974) concluded that this broad peak was the result of excitation of the 1500-Hz region of the basilar membrane, thus the area immediately below the sharp cutoff of the patient’s pure tone audiogram. At 95 dB Pe SPL, the latencies of these derived action potentials were 6.5 ms at 0 to 500 Hz, 4.8 ms for the range 500 to 1000 Hz, 3 ms at 1 to 2 kHz, 2.5 ms for 2 to 4 kHz, and about 1.8 ms for the range above 4 kHz. The conduction time in the auditory nerve from the cochlea to the location on the eighth nerve from which we recorded can be assumed to be about 1 ms on the basis of its conduction velocity (20 to 40 m/set, Lazorthes et al., 1961; Spoendlin and Schrott, 1989) and the distance (which is about 2 cm, Lang, 1981, 1983). The latency of the broad peak in the CAP in the patient in our study who had a hearing loss above 1000 Hz (Fig. 2D) was 8 to 10 ms, which indicates that this peak originated from the 500 to 1000 Hz region of the basilar membrane, a situation similar to the one reported by Elberling (1974). The findings of the present study are in agreement dB , 85 110 dB 105 dB dB I I I I I I I I I I 0 2 4 6 8 10 12 14 16 18 dB 85 dB I I 20 95 ms 0.02 5 I I111111 0.05 0.1 I 0.2 I I II III 0.5 Frequency Fig. 12. Results similar to those shown in Fig. 0, but from the patient whose audiogram 0.7 1 ,n I I lllllll 2 3 5 ktir and CAP are showm,ln Figs. 2E and 8 7 dB 10 19 with those of Coats and Martin (1977) that the latenties of the nerve action potentials recorded from the ear canal are generally prolonged in patients with high-frequency hearing loss and that in these patients there is a greater difference between the responses to clicks of opposite polarity than is found for patients with normal hearing. Although we studied only a small number of patients and their audiograms represent only a portion of the spectrum of results that may be expected from those with high-frequency hearing loss, our results also agree with those of Coats and Martin (1977) in that they show that there is poor correlation between deviation of the CAP from normal and the shape of the audiogram in individual patients. Coats and Martin (1977) also found that the components following the initial negative peak in the response recorded from the ear canal seemingly reversed in phase when click polarity was reversed. Because this component consists of only one wave it is difficult to distinguish such a shift from a shift in latency. We see similar components in our recordings from the exposed intracranial-portion of the eighth nerve (c.f., Fig. 3 at 110 dB). These components are different from the much more prolonged quasiperiodic oscillations that we have shown occur in recordings from the exposed intracranial portion of the eighth nerve in the present study as well as in a previous study (Moller and Jho, 1990). These oscillations occur in the time interval between 5 and 15 ms after the click sound. In the present study we showed that such waves are indeed common in patients with severe high-frequency hearing loss, although they do occur in individuals with normal or near-normal hearing. We earlier pointed out that there are similarities in latency and frequency range between the oscillations we observed in the CAP recorded from the eighth nerve and evoked otoacoustic emissions (Moller and Jho, 1990). However, there is at least one important difference, namely that otoacoustic emissions are a nearly constant phenomenon and most pronounced in people with normal hearing, while the late waves in the CAP occur rarely in individuals with normal hearing and preferentially in patients with high-frequency hearing losses of cochlear origin. The long duration of these oscillations and the fact that they reverse precisely when the polarity of the click sound is reversed indicate that they are the result of a damped oscillation of a bandpass filter that has a UV 0 6.0 r DIFFERENCE 4.0 - 3.5 - 2.5 - 0.25 70 75 00 05 90 95 100 dB 105 110 70 PeSPL I I I I I 75 60 65 90 95 dB I 100 I 105 I 110 PeSPL Fig. 13. (A) RMS values of the difference between the late waves in response to condensation and rarefaction clicks determined in the interval 4.5 to 14.5 ms after the stimulus and shown as a function of the stimulus intensity. (B) RMS values of the sum of the responses to condensation and rarefaction clicks in the interval 0 to 20 ms after the stimulus. In both graphs, open circles represent data from the patient illustrated in Fig. 2A, crosses show data illustrated in Fig. 2B, filled circles show data illustrated in Fig. 2C, squares show data illustrated in Fig. 2D, and triangles show data illustrated in Fig. 2E. narrow bandwidth. The fact that the latencies of the individual waves of this oscillation are nearly independent of the stimulus intensity indicates that these components are the neural responses to continuous excitation. In this respect, these oscillations are similar to responses caused by stimulation with low-frequency tonebursts, where the latency of the individual waves in the response also is nearly independent of the stimulus intensity (Moller and Jho, 1989a). Similar results have been obtained in animal experiments using noise stimuli (Msller, 1975, 1983). The relatively high amplitude of these oscillations (about 20% of the total amplitude of the CAP) is an indication that these late waves represent neural activity in a large number of nerve fibers that is phase-locked to frequencies within a very narrow range. If the frequencies of the damped oscillations to which the individual nerve fibers were phase-locked differed by any noticeable amount from each other, the resulting gross response would not contain such sustained oscillations but rather would have the shape of a single deflection marking the beginning of the damped oscillations. In order to account for the results of this study and the 105 dB 105 dB 95 d6 .",--)\; a5 dB /_ 75 d6 ,__- -. : 6._- d’ 13:26 ,’ 65 dB $1 i, 13136 DISTAL I( ‘1 I 11 I Fig. 15. CAP recorded (normal I I I I I I I I I I I 0 2 4 6 a IO 12 14 16 ia 20 Fig. 14. Difference rarefaction between clicks obtained bers) during the intradural the responses at different portion to condensation to relieve and TN. The three top recordings were made from the middle portion of the intracranial obtained ings were portion of the eighth nerve, and the bottom record was from a distal location near the porus acusticus. The recordobtained from the same patient whose atidiogram CAP are shown in Figs. 2D and 7, respectively. I I L 11 12 13 14 from a patient with only a slight hearing 15 Ills loss to 4000 Hz, 25 and 20 dB at 6000 and 8000 Hz, respectively). ms times (given by legend num- of an operation I 10 0123456789 and earlier one (Moller and Jho, 19901, it seems necessary to assume that as many as 20% of the nerve fibers that are activated by a high-intensity click sound are phaselocked to low frequencies within a very narrow range to produce these oscillations. If we accept that the basilar membrane is the source of the oscillation to which the neural activity is phase-locked and that the basilar 21 110 dB 105 dB 95 dB 90 d0 / 2.5~” 1 0.5pv I 0 I 1 I I I I I I 2 3 4 5 6 7 I I I 8 9 10 ms Fig. 16. CAP recorded from a patient with a rather severe hearing loss but no discernible late waves. (Normal to 1000 Hz, down 55 dB at 2000 and 4000 Hz, and 85 dB at 6000 and 8000 Hz.) membrane is continuously tuned along its length to frequencies within the audible range, then it is difficult to understand how so many nerve fibers could be tuned to nearly the same frequency. The oscillations in the CAP could be explained in another way. Perhaps there are one or a few narrow regions of the basilar membrane with extreme degrees of frequency selectivity. This would result in prolonged ringing when the ear is excited by a broadband transient sound. Such a high degree of frequency selectivity could occur at localized spots on the basilar membrane as a result of large discontinuities in the properties of the basilar membrane. In explaining the mechanism of otoacoustic emission, it has been put forward that positive feedback systems could produce very high degrees of spectral selectivity at a certain point on the basilar membrane, which would account for the prolonged oscillations. The long latency of the otoacoustic emissions was explained by a slowing in propagation of the traveling wave as it approaches the region of high selectivity (Kemp, 1978). While this is a plausible explanation for the otoacoustic emission, it is difficult to understand how a sufficient number of nerve fibers could be activated by the vibration of such a small area of the basilar membrane that would account for the large amplitude of the oscillations in the CAP. There are two reasons why it does not seem likely that these waves are a result of the loss of hair cells in the basilar region of the cochlea per se. First, the low-frequency nature of the oscillations and their long latency indicate that the waves originate from apical cochlear regions, which on the basis of the normal low-frequency hearing threshold may be assumed to be normal. Second, the long latency of these waves seems to indicate that the traveling wave has already passed the region of the cochlea that represents the frequency region of impaired hearing threshold. It thus seems that noise exposure, aging, or whatever other factor caused the (selective) high-frequency hearing loss in these patients also in one way or another injured or affected the low-frequency region of the basilar membrane in such a way that it produced these abnormal oscillations, without affecting the threshold for pure tones to a noticeable degree. Studies of changes in the cochlea that occur as a result of aging or exposure to noise have shown a loss of hair cells in the basal portion of the cochlea but no noticeable pathologies in the more apical portion of the cochlea (Hawkins and Johnsson, 1976). It is naturally difficult to speculate about what effect these late waves in the CAP of patients with highfrequency hearing losses may play in auditory function, but in view of the large amplitude of these oscillations and the fact that they occur over a large range of sound intensities, one would expect that they would have perceptual correlates. However, so far the only such abnormal phenomenon observed in patients with highfrequency hearing loss is tinnitus, and this is often associated with high-frequency hearing loss. It is surprising that, in the numerous animal experiments that have been carried out on noise-induced hearing loss and hearing loss caused by ototoxic drugs, no such waves have been reported. If similar waves do not exist in animals in which hearing loss has been induced by exposure to noise or ototoxic drugs, other explanations for their existence in man must be sought. One difference between most studies of noise-induced 22 hearing loss in animals and those in man is that most animals are studied relatively soon after a hearing loss has been induced, whereas the patients we have studied have in most cases had the hearing loss for many years. Perhaps there are changes that occur over time that are not noticeable in the acute or relatively acute time frame of the animal experiments performed to date. In the present study we only included data from patients who were operated upon for relief of TN, although we have data from patients operated upon for relief of hemifacial spasm (HFS) or disabling positional vertigo (DPV). We did not include patients with HFS or DPV in this study because of the possibility that their auditory nerves might have been affected by the vascular compression of the facial nerve or the vestibular nerve that causes these diseases. In fact it has been shown (Moller and Moller, 1985) that some patients with HFS have audiometric abnormalities that can be associated with vascular compression of the auditory nerve, in addition to the vascular compression of the facial nerve that is the cause of their disease, and that these audiometric abnormalities occur only on the side with HFS. Because the anatomic distance between the eighth nerve and the trigeminal nerve is quite large, it is unlikely that the blood vessel that compresses the trigeminal nerve could also affect the eighth nerve. Also, the fact that the patients we studied had similar hearing loss on both sides while only the trigeminal nerve on one side was affected by vascular compression further indicates that the hearing loss was not a result of the vascular abnormality that gave rise to TN. Finally, the operating room is not an ideal environment for recording CAP from the eighth nerve because it is impossible to shield the patient fully from electrical and acoustic noise. The effects of electrical interference can be reduced by averaging more responses, but ambient sounds will always mask the test sounds to some extent. This masking prevents measurement of responses to test sounds of low intensities, and thus limits the range of sound intensities over which responses can be studied. In addition, the ambient noise in the operating room may affect the responses in one way or another, but since the noise changes constantly in nature it is difficult to obtain a valid estimate of its character. Nevertheless, we have attempted to reduce the noise that reaches the patient’s ear by careful placement and taping of the earphone in the ear, and the fact that we can obtain responses from patients with normal hearing down to 25 to 30 dB above the perceptual threshold of the clicks indicates that the degree to which ambient noise in the operating room masks the test tones is moderate. Thus, we believe that such ambient noise has not affected the results presented in this paper except for limiting the intensity range over which potentials could be studied. Acknowledgements This work was supported by the National Institutes of Health (Grants No. ROl-NS21378 and No. ROlDC00272). The authors are grateful to Peter J. Jannetta, M.D., for making patients under his care available for this study, and to Margareta B. Moller, M.D., D.Med.Sci., for audiological evaluation of the patients in the present study. References Blackman, R.B. and Tukey, J.W. (1959) Measurement of Power Spectra. Dover Publications, Inc., New York. Coats, A.C. and Martin, J.L. (1977) Human auditory nerve action potentials and brainstem evoked responses. Arch. Otolaryngol. 203, 605-622. Elberling, C. (1974) Action potentials along the cochlear partition recorded from the ear canal in man. Stand. Audiol. 3, 13-19. Goldstein, M.H. (1960) A statistical model for interpreting neuroelectric responses. Inf. Contr. 3, 1-17. Hashimoto, I.. Ishiyama, Y., Yoshimoto, T. and Nemoto, S. (1981) Brainstem auditory evoked potentials recorded directly from human brain-stem and thalamus. Brain 104. 841-859. Hawkins, J.E. and Johnsson, L.G. (1976) Patterns of sensorineural degeneration in human ears exposed to noise. In: D. Henderson. R.P. Hamernik, D.S. Dosanjh and J.H. Mills (Eds.). Effects of Noise on Hearing. Raven Press, New York, pp. 91-110. Jannetta, P.J. (1977) Observations on the etiology of trigeminal neuralgia, hemifacial spasm, acoustic nerve dysfunction, and glossopharyngeal neuralgia. Definite microsurgical treatment and results in 117 patients. Neurochirurgie (Paris) 20, 145-154. Jannetta. P.J. (1981a) Hemifacial spasm. In: M. Samii and P.J. Jannetta (Eds.). The Cranial Nerves. Springer-Verlag, New York, pp. I-X. Jannetta. P.J. (1981b) Vascular decompression in trigeminal neuralgia. In: M. Samii and P.J. Jannetta (Eds.), The Cranial Nerves. Springer-Verlag, New York, pp. 331-340. Kemp, D.T. tlY78) Stimulated acoustic emissions from within the human auditory system. J. Acoust. Sot. Am. 64: 1386-1391. Lang, J. (1981) Facial and vestibulocochlear nerve, topographic anatomy and variations. In: M. Samii and P.J. Jannetta (Eds.), The Cranial Nerves. Springer-Verlag, New York, pp. 363-377. Lang, J. (1983) Clinical Anatomy of the Head. Neurocranium Orbit Craniocervical Regions. Springer-Verlag Publishing Co., New York Lazorthes, G., Lacomme, Y., Gaubert. J. and Planel, H. (1961) La constitution du nerf auditif. La Presse Medicale 69: 1067-1068. Moller, A.R. (1975) Latency of unit responses in the cochlear nucleus determined in two different ways. J. Neurophysiol. 38, 812-821. Moller, A.R. (1983) Frequency selectivity of phase-locking of complex sounds in the auditory nerve of the rat. Hear. Res. 1 I, 267-284. Moller, A.R. (lY88a) Evoked Potentials in Intraoperative Monitoring. Williams and Wilkins, Baltimore, MD. Moller, A.R. and Jannetta, P.J. (1983) Monitoring auditory functions during cranial nerve microvascular decompression operations by direct recording from the eighth nerve. J. Neurosurg. 59, 493-499. Moller. A.R. and Jho, H.D. (1988) Responses from the brainstem at the entrance of the eighth nerve in human to contralateral stimulation. Hear. Res. 37, 477.52. Moller, A.R. and Jho, H.D. (lY89a) Response from the exposed 23 intracranial human auditory nerve to low-frequency tones: Basic characteristics. Hear. Res. 38, 163-175. Moller. A.R. and Jho, H.D. (1989b) Responses from the exposed human auditory nerve to pseudorandom noise. Hear. Res. 42, 237-252. Moller, A.R. and Jho, H.D. (1990) Late components in the compound action potentials (CAP) recorded from the intracranial portion of the human eighth nerve. Hear. Res. 45, 75-86. Moller, A.R. and Jho, H.D. (1991) Compound action potentials recorded from the intracranial portion of the auditory nerve in man: Effects of stimulus intensity and polarity. Audiology (in press). Moller, M.B. and Moller, A.R. (1985) Audiometric abnormalities in hemifacial spasm. Audiology 24, 396-405. Moller, A.R. and Moller, M.B. (1989) Does intraoperative monitoring of auditory evoked potentials reduce incidence of hearing loss as a complication of microvascular decompression of cranial nerves? Neurosurgery 24, 257-263. Moller, A.R., Jannetta, P.J., Bennett, M. and Moller, M.B. (1981a) Intracranially recorded responses from the human auditory nerve: New insights into the origin of brain stem evoked potentials (BSEPs). Electroenceph. Clin. Neurophysiol. 52, 18-27. Moller, A.R., Jannetta, P.J. and Moller, M.B. (1981b) Neural generators of brainstem evoked potentials. Results from human intracranial recordings. Ann. Otol. Rhinol. Laryngol. 90, 591-596. Moller, A.R., Jannetta, P.J. and Sekhar, L.N. (1988) Contributions from the auditory nerve to the brainstem auditory evoked potentials (BAEPsJ: Results of intracranial recording in man. Electroenceph. Clin. Neurophysiol. 71, 198-211. Meller, A.R., Moller, M.B., Jannetta, P.J. and Jho, H.D. (1991) Auditory nerve compound action potentials and brainstem auditory evoked potentials in patients with varying degrees of hearing loss. Ann. Otol. Rhinol. Laryngol. (in press). Spoendlin, H. and Schrott, A. (1989) Analysis of the human auditory nerve. Hear. Res. 43: 25-38. Teas, D.C., Eldredge, D.H. and Davis, H. (1962) Cochlear responses to acoustic transients: An interpretation of whole-nerve action potentials. J. Acoust. Sot. Am. 34, 1438-1459.