- Email: [email protected]
Modification of auditory pathway functions in patients with hearing improvement after middle ear surgery
Modification of auditory pathway functions in patients with hearing improvement after middle ear surgery JUHA-PEKKA VASAMA, MD, PhD, JYRKI P. MÄKELÄ, MD, PhD, and HANS A. RAMSAY, MD, PhD, Helsinki, Finland
We recorded auditory-evoked magnetic responses with a whole-scalp 122-channel neuromagnetometer from seven adult patients with unilateral conductive hearing loss before and after middle ear surgery. The stimuli were 50-msec 1-kHz tone bursts, delivered to the healthy, nonoperated ear at interstimulus intervals of 1, 2, and 4 seconds. The mean preoperative pure-tone average in the affected ear was 57 dB hearing level; the mean postoperative pure-tone average was 17 dB. The 100-msec auditory-evoked response originating in the auditory cortex peaked, on average, 7 msecs earlier after than before surgery over the hemisphere contralateral to the stimulated ear and 2 msecs earlier over the ipsilateral hemisphere. The contralateral response strengths increased by 5% after surgery; ipsilateral strengths increased by 11%. The variation of the response latency and amplitude in the patients who underwent surgery was similar to that of seven control subjects. The postoperative source locations did not differ noticeably from preoperative ones. These findings suggest that temporary unilateral conductive hearing loss in adult patients modifies the function of the auditory neural pathway. (Otolaryngol Head Neck Surg 1998;119:125-30.)
Unilateral sound deprivation early in life is known to affect the developing auditory system in experimental animals. Conductively sound-deprived rats and cats show abnormal ipsilateral inhibition in the inferior colliculus to stimulation of the healthy ear.1,2 In cats reared with neonatal cochlear ablation, the mean thresholds of neurons in the auditory cortex ipsilateral to the stimu-
From the Department of Otolaryngology (Drs. Vasama and Ramsay), University Central Hospital; and Department of Neurology (Dr. Mäkelä), Central Military Hospital. Presented at the Annual Meeting of the American Academy of Otolaryngology–Head and Neck Surgery, Washington, D.C., Sept. 29–Oct. 2, 1996. Reprint requests: J. P. Mäkelä, MD, PhD, Department of Neurology, Central Military Hospital, PL 50, 00301 Helsinki, Finland. Copyright © 1998 by the American Academy of Otolaryngology– Head and Neck Surgery Foundation, Inc. 0194-5998/98/$5.00 + 0 23/77/83126
lated ear are the same as those in the contralateral hemisphere, whereas in intact animals the threshold to ipsilateral stimulation is 17 dB higher.3 In human beings, however, congenital unilateral conductive hearing loss appears to have only a slight effect on the development of the central auditory pathways.4,5 Information regarding the effects of temporary unilateral conductive hearing loss on the adult mammalian auditory system is sparse. The ipsilateral inferior colliculus of cats with 3 to 4 months’ ligation of the external meatus receives diminished inhibitory input from the nonligated ear.1 In adult, awake guinea pigs, the evoked response thresholds decrease, amplitudes increase, and latencies shorten in the auditory cortex ipsilateral to the stimulation of the healthy ear after hair cell destruction in the opposite cochlea.6 In human beings, behavioral evidence for binaural plasticity is seen after plugging of one ear. The subjects initially center a sound image created by binaural tones on the basis of equal loudness; after several days, the adjustment is based more on the actual intensity of the tone.7 The possible changes in auditoryevoked potentials caused by unilateral hearing loss have not been studied. The human primary auditory cortex lies embedded in Sylvius’ fissure and is not easily studied even when the skull is opened during surgery. New methods, such as magnetoencephalography (MEG) have, however, made the study of auditory cortical processing feasible in awake human beings, including patients with hearing deficits. MEG records the magnetic field of small synchronous electrical currents that flow perpendicular to the surface of the activated cortex; only the fields tangential to the head surface, flowing in the fissural cortex, produce MEG signals outside the head, whereas more widely used electroencephalography (EEG) is sensitive to both radial and tangential currents. Changes in the electrical conductivity in scalp and skull do not affect MEG, whereas EEG is distorted by them. These properties aid in MEG source analysis; algorithms needed for source calculations are simpler in MEG than in EEG. One method for modeling the source areas of evoked potential and evoked field is calculation of equivalent current dipoles (ECDs). In such a model the cortical region producing the signal is modeled by a pointlike source, whose location is the center of gravi125
126 VASAMA et al.
Fig. 1. Subject positioned for measurement under the instrument. (Courtesy Neuromag Ltd.)
ty of the cortical area that contributes to the recorded potentials and magnetic fields.8-11 Transient sounds elicit auditory-evoked magnetic fields (AEFs) in the human supratemporal auditory cortex. The most prominent deflection, the N100m, has a peak with a latency of about 100 msecs after sound onset.9,12 N100m latency, amplitude, and source location have been found to be highly reproducible in 30 successive recordings during a 2-month period.13 We studied the effect of correction of temporary unilateral conductive hearing loss, caused by otosclerosis or ossicular chain disruption, on the function of the auditory pathway in adult human beings by comparing preoperative and postoperative AEFs. The measurements were made with a whole-scalp neuromagnetometer, which allows simultaneous recording of the activity from the auditory cortexes of both hemispheres. METHODS AND MATERIAL AEFs were recorded from seven adult patients (4 female, 3 male; age range, 26 to 51 years; mean age, 38 years) with unilateral conductive hearing loss before and after middle ear surgery: five with otosclerosis (patients 1, 3, 4, 5, and 7) and two with disrupted ossicular chains (patients 2 and 6). Five patients had right-sided (patients 1 through 5) and two had left-sided (patients 6 and 7) hearing loss. The mean (±SD)
Otolaryngology– Head and Neck Surgery July 1998
preoperative pure-tone average (PTA; the average of 0.5, 1, and 2 kHz) in the affected ear was 57 ± 12 dB hearing level; the mean postoperative PTA was 17 ± 8 dB. The mean preoperative bone conduction threshold (for frequencies 0.5, 1, and 2 kHz) in the affected ears was 17 dB; the corresponding postoperative value was 15 dB. The duration of hearing loss before surgery varied from 6 to 14 years. The patients were otherwise healthy, and each had a normal auricle, external ear canal, and tympanic membrane on the affected side. The audiogram findings were normal (hearing threshold ≤20 dB for frequencies ≤8 kHz) in the healthy ears. In patients with otosclerosis, stapedioplasty was performed with a Teflon prosthesis. An autologous incus body was sculptured to fit between the manubrium of the malleus and stapes head in patient 2 and between the manubrium of the malleus and stapes footplate in patient 6. The control group consisted of seven healthy adults (2 female, 5 male; age range, 25 to 53 years; mean age, 38 years) without history of hearing deficits. The study was approved by the local ethical committee, and the subjects gave informed consent to be in the study. Tone bursts (50 msec, 1 kHz, 20-msec linear rise and fall times) were delivered to the subject’s healthy, nonoperated ear through plastic tubes, and earpieces were inserted at interstimulus intervals (ISIs) of 1, 2, and 4 seconds. The sound intensity was 80 dB sound pressure level at the earpiece. MEG measurements were performed 1 to 4 months before and 2 months after the operation. MEG signals were measured noninvasively outside the head with a whole-scalp 122-channel neuromagnetometer (Neuromag-122; Biomag Laboratory, Medical Engineering Centre, Helsinki University Hospital, Helsinki, Finland). The measurements were carried out in a magnetically shielded room. The subject was seated with the eyes open and with the head leaning against the helmet-shaped bottom surface of the neuromagnetometer (Fig. 1). Each sensor unit contains a pair of orthogonal gradiometers that measure the tangential derivatives in two orthogonal directions of the magnetic field component normal to the helmet surface at the sensor location. The planar gradiometers pick up the strongest signals from source currents just below the sensor. The exact location of the head, with respect to the superconducting sensors, was determined by measuring magnetic signals produced by indicator currents, led through coils placed at known locations on the scalp. The positions of the coils with respect to the outer landmarks of the head were obtained by means of a threedimensional digitizer. This procedure provided reliable comparison between successive measurements and allowed alignment of the MEG and MRI coordinates, necessary for combination of functional and anatomic structural information. An MRI was obtained from patient 2 by use of a 1-T Siemens Magnetom instrument. The recording passband was 0.03 to 100 Hz, and the sam-
Otolaryngology– Head and Neck Surgery Volume 119 Number 1
A
VASAMA et al. 127
B
Fig. 2. A, AEFs of patient 2 to left-ear stimulation at 2-second ISI. In each response pair, the upper trace illustrates the field derivative along the latitude (latit.) and the lower trace along the longitude (longit.). The head is viewed from above, and the nose points upwards. The passband is 0.03 to 40 Hz, and the traces show a 500-msec time period starting 50 msecs before the stimulus onset. B, Source strengths (Q) and goodness-of-fit curves [g (%)] from responses to left-ear stimulation at 4-second ISI from the left (LH) and right (RH) hemispheres in the same patient.
pling rate was 0.4 kHz. The 500-msec analysis period included a 50-msec prestimulus baseline. The vertical and horizontal electro-oculograms were used to reject data contaminated by eye movements and blinks. For all ISIs, 65 to 110 single responses were averaged to obtain an adequate signal-tonoise ratio. Two ECDs, one in each hemisphere, were used to explain the magnetic field patterns during the N100m. Each ECD, modeling synchronous firing of neurons in a cortical area of less than a few square centimeters, was first found separately for the left- and right-hemisphere data by a least-squares search; the calculations resulted in the three-dimensional location, orientation, and strength (dipole moment; Unit = Ampere • Meter) of the ECD. The goodness of fit of the dipole model, which tells in percentage how much the dipole accounts for the measured field pattern, was calculated according to the method described earlier.10,14 Differences in peak latencies and source strengths were evaluated by use of paired, two-tailed t tests. The test-retest variability of the peak latencies of N100m to tones given monaurally once every 2.5 seconds has been previously estimated with the same recording equipment in a different population of five healthy subjects. In two sets of measurements 1 to 3 weeks apart, the contralateral N100m peaked at 103 ± 11 msecs in the first session and at 103 ± 11 msecs in the second one; corresponding ipsilateral values were 109 ± 10 msecs and 108 ± 8 msecs, respectively (J. Virtanen, Personal communication).
RESULTS
The preoperative whole-scalp AEFs to left-ear stimulation, presented at a 2-second ISI in patient 2 (who had right-sided hearing loss), shows responses over both temporal areas (Fig. 2A). The N100m peak occurs with a latency of 95 msecs over the right hemisphere and 2 msecs later over the left. The estimated strengths of the left- and right-hemisphere ECDs to left-ear stimulation with a 4-second ISI in patient 2 show that the postoperative dipoles reach the maximum amplitude earlier than preoperative ones (Fig. 2B). In the largest responses of all patients from both hemispheres to stimulation of the healthy ear at the 2second ISI, N100m peaks at 76 to 113 msecs over the left and at 72 to 98 msecs over the right hemisphere (Fig. 3). Patients 1, 3, 4, 5, and 6 have shorter postoperative than preoperative N100m latencies over both hemispheres. In patients 2 and 7, the postoperative N100m latency is shorter than the preoperative one over the left hemisphere. Patients 1, 3, 4, and 5 have larger postoperative than preoperative N100m amplitudes over both hemispheres. Patient 2 has a larger postoperative than preoperative N100m amplitude over the right hemisphere, and patient 6 over the left hemisphere, respectively. After the operation, the mean N100m peak latency to nonoperated left-ear stimulation (patients 1 through 5) was 10 msecs shorter over the right (p < 0.01) and 4
128 VASAMA et al.
Otolaryngology– Head and Neck Surgery July 1998
Fig. 3. Responses of patients to preoperative and postoperative stimulation of the healthy ear at 2-second ISI. Channels showing the largest amplitudes over each hemisphere were chosen for display. L, Left ear stimulation; R, right ear stimulation; P, patient number.
msecs shorter over the left hemisphere (nonsignificant). Such clear latency differences were not seen to rightear stimulation in patients 6 and 7 (Table 1). To left-ear stimulation, mean postoperative source strength was 31% stronger (p < 0.01) over the right hemisphere than the preoperative one and 21% stronger over the left
Fig. 4. The preoperative source location for N100m over both hemispheres, superimposed on the MRI scan of patient 2. Circles indicate the source, and tails illustrate the current orientation. The left ear was stimulated at 2second ISI. L, Left hemisphere; R, right hemisphere.
Otolaryngology– Head and Neck Surgery Volume 119 Number 1
VASAMA et al. 129
Table 1. Preoperative and postoperative latencies and source strengths of equivalent current dipoles for N100m, pooled over all ISIs Left ear stimulation Parameters
Latency (msec) Pre healthy Post healthy Post affected Control Q (nAm) Pre healthy Post healthy Post affected Control
RH
Right ear stimulation
LH
95 ± 13 (5) 85 ± 7* 84 ± 6 (2) 87 ± 8 (7)
95 ± 17 (5) 91 ± 5 95 ± 8 (2) 102 ± 5 (7)
22 ± 10 (5) 29 ± 14* 52 ± 15 (2) 24 ± 11 (7)
16 ± 7 (5) 20 ± 9* 20 ± 12 (2) 19 ± 11 (7)
RH
83 ± 8 (2) 83 ± 9 102 ± 6 (5) 103 ± 7 (7) 29 ± 11 (2) 29 ± 11 15 ± 5 (5) 19 ± 7 (7)
LH
80 ± 4 (2) 79 ± 8 94 ± 11 (5) 96 ± 6 (7) 36 ± 11 (2) 25 ± 6* 21 ± 11 (5) 25 ± 13 (7)
Data expressed as mean ± SD or range (no. of subjects). RH, Right hemisphere; LH, left hemisphere; Q, source strength; nAM, nanoamperemeters; Pre, preoperative and Post, postoperative values to stimulation of healthy or affected ear. *Statistically significant difference between preoperative and postoperative values.
hemisphere (p < 0.01). However, to right-ear stimulation the mean postoperative source strength was 31% weaker over the left hemisphere (p < 0.01) than the preoperative one. Both preoperative and postoperative latency and source strength values of the patients were within the normal variance of the control group (Table 1). We also measured the postoperative N100m peak latencies and source strengths to stimulation of the operated ear with improved hearing (i.e., right ear in patients 1 through 5 and left ear in patients 6 and 7) (Table 1). The mean values did not differ significantly from those in controls. Preoperative ECDs for N100m at the 2-second ISI, superimposed on the MRI of patient 2, show one source in each supratemporal cortex (Fig. 4), implying activation of the auditory cortexes within Sylvius’ fissures. There were no significant differences between preoperative and postoperative source locations in patients or between the source locations of the controls and the patients. DISCUSSION
The responses to stimulation of the healthy, nonoperated ear changed after the operation in all patients, suggesting that there was reorganization of the binaural interaction subsequent to resolving the unilateral conductive hearing loss. Either the latencies were shorter or the source strengths were stronger after the middle ear surgery and hearing improvement. Although the difference between preoperative and postoperative latency and source strength values of the patients was obvious, they were within the variance of the normal-hearing control group. Other studies have
shown that congenital unilateral conductive hearing loss up to 70 dB has little effect on AEFs,4 and it is therefore not surprising that the preoperative responses of our patients with unilateral hearing loss were within normal limits. However, the postoperative changes were statistically significant and probably reflect the ability of the auditory system to modify its functions to maintain a balance of input from the two ears. Animal experiments have shown that unilateral cochlear destruction decreases ipsilateral-contralateral latency differences.6 Analogously, in previous studies, three of six patients with unilateral congenital conductive hearing loss and four of eight patients with unilateral idiopathic sudden sensorineural hearing loss had shorter N100m latencies and stronger source strengths in the hemisphere ipsilateral to the stimulated healthy ear.4,15 Furthermore, ipsilateral-contralateral latency differences decreased in some patients during recovery from sudden unilateral hearing loss because of acoustic neuroma removal.16 As expected from these results, the restoration of normal hearing increased the ipsilateralcontralateral latency difference in our patients. A continuous tone of moderate intensity may enhance the amplitude of cortical responses evoked by brief acoustic stimuli.17 In line with this, N100m amplitude to brief, monaural tones is enhanced during a continuous tone presented to the other ear.18 Improved hearing of environmental sounds after operation could cause analogous enhancement in the hemisphere contralateral to the operated ear. However, the N100m amplitude enhancement by the continuous tone is accompanied by an increase, not by a decrease, of latency.18 Obviously, additional mechanisms are needed to explain the latency decrease observed in this study.
Otolaryngology– Head and Neck Surgery July 1998
130 VASAMA et al.
In animal experiments, cortical and subcortical maps reorganize after restricted damage of cochlear receptors, and the cortical area normally representing frequencies of the lesioned cochlear portion is occupied by neighboring frequencies. Nevertheless, this does not mean central compensation of peripheral hearing loss; sensitivity to frequencies normally activating the damaged cochlear region is not restored.19 On the other hand, enhanced mean responses of single cells and increase of the coherence between activities of neurons participating in the same processing assembly during recovery from lesion would lead to behavioral improvement by means of increased reliability and shortened duration of neural processing.20 The increase of N100m amplitude and shortening of its latency after surgery could reflect an increase of coherence of the neural events underlying N100m, and a shortening of the auditory processing time. Thus they may be related to the improved auditory capacity. This is supported by our previous findings in patients with unilateral profound deafness after an acoustic neuroma operation. Their N100ms were initially smaller in amplitude and delayed in both hemispheres and recovered toward normal values during the first postoperative year. The alteration was attributed to enhanced coherence of postsynaptic potentials underlying N100m during recovery.16 This study shows the adaptability of the adult human auditory system after unilateral middle ear surgery to improve hearing. It also shows the benefit to otolaryngologists from a fast and noninvasive MEG method for investigation of auditory cortical functions in normal subjects and in hearing-impaired patients. We thank S. Heikkilä for help in patient preparation for MEG measurements, A. Korvenoja and S. Martinkauppi for MRIs, and J. Virtanen for test-retest data of normal subjects. R. Hari and A. Møller made valuable comments on the manuscript. REFERENCES 1. Moore DR, Irvine DRF. Plasticity of binaural interaction in the cat inferior colliculus. Brain Res 1981;208:198-202. 2. Silverman MS, Clopton BM. Plasticity of binaural interaction I. Effect of early auditory deprivation. J Neurophysiol 1977;40:1266-74. 3. Reale RA, Brugge JF, Chan CK. Maps of auditory cortex in cats
4. 5. 6.
7. 8.
9.
10.
11.
12.
13.
14.
15. 16. 17. 18. 19. 20.
reared after unilateral cochlear ablation in the neonatal period. Dev Brain Res 1987;34:281-90. Vasama J-P, Mäkelä JP, Parkkonen L, Hari R. Auditory cortical responses in humans with congenital unilateral conductive hearing loss. Hear Res 1994;78:91-7. Wilmington D, Gray L, Jahrsdoerfer R. Binaural processing after corrected congenital unilateral conductive hearing loss. Hear Res 1994;74:99-114. Popelar J, Erre J-P, Aran JM, Cazals Y. Plastic changes in ipsicontralateral differences of auditory cortex and inferior colliculus evoked potentials after injury to one ear in the adult guinea pig. Hear Res 1994;72:125-34. Florentine M. Relation between lateralization and loudness in asymmetrical hearing loss. J Am Audiol Soc 1976;1:243-51. Vasama J-P, Mäkelä JP. Auditory pathway reorganization in humans with congenital or acquired unilateral hearing loss. In: Salvi R, Henderson D, Fiorino F, Colletti V, editors. Auditory system plasticity and reorganization. New York: Thieme; 1996. p. 359-70. Hari R. The neuromagnetic method in the study of the human auditory cortex. In: Grandori F, Hoke M, Romani GL, editors. Auditory evoked magnetic fields and potentials. Advances in audiology. Vol. 6. Basel: Karger; 1990. p. 222-82. Hämäläinen M, Hari R, Ilmoniemi R, Knuutila J, Lounasmaa OV. Magnetoencephalography—theory, instrumentation and applications to noninvasive studies of the working human brain. Rev Mod Phys 1993;65:413-97. Mäkelä JP, Hari R. Long-latency auditory evoked magnetic fields. In: Sato S, editor. Magnetoencephalography: clinical applications and comparison of EEG. New York: Raven Press; 1990. p. 177-91. Mäkelä JP, Ahonen A, Hämäläinen M, et al. Functional differences between auditory cortices of the two hemispheres revealed by whole-head neuromagnetic recordings. Hum Brain Mapping 1993;1:48-56. Pantev C, Gallen C, Hampson S, Buchanan S, Sobel D. Reproducibility and validity of neuromagnetic source localization using a large array biomagnetometer. Am J EEG Technol 1991;31:83-101. Kaukoranta E, Hämäläinen M, Sarvas J, Hari R. Mixed and sensory nerve stimulations activate different cytoarchitectonic areas in the human primary somatosensory cortex SI. Exp Brain Res 1986;63:60-6. Vasama J-P, Mäkelä JP. Auditory pathway plasticity in adult humans after idiopathic sudden unilateral sensorineural hearing loss. Hear Res 1995;87:132-40. Vasama J-P, Mäkelä JP, Pyykkö I, Hari R. Abrupt unilateral deafness modifies function of human auditory pathways. Neuroreport 1995;6:961-4. Gerken GM. Alteration of central auditory processing of brief stimuli: a review and a neural model. J Acoust Soc Am 1993;93:2038-49. Loveless N, Vasama J-P, Mäkelä JP, Hari R. Human auditory cortical mechanisms of sound lateralisation: III. Monaural and binaural shift responses. Hear Res 1994;81:91-9. Irvine DRF, Rajan RR. Plasticity in the frequency organization of auditory cortex of adult mammals with restricted cortical lesions. Biomed Res 1993;14:55-9. Ahissar E, Ahissar M. Plasticity in auditory cortical circuitry. Curr Opin Neurobiol 1994;4:580-7.
We recorded auditory-evoked magnetic responses with a whole-scalp 122-channel neuromagnetometer from seven adult patients with unilateral conductive hearing loss before and after middle ear surgery. The stimuli were 50-msec 1-kHz tone bursts, delivered to the healthy, nonoperated ear at interstimulus intervals of 1, 2, and 4 seconds. The mean preoperative pure-tone average in the affected ear was 57 dB hearing level; the mean postoperative pure-tone average was 17 dB. The 100-msec auditory-evoked response originating in the auditory cortex peaked, on average, 7 msecs earlier after than before surgery over the hemisphere contralateral to the stimulated ear and 2 msecs earlier over the ipsilateral hemisphere. The contralateral response strengths increased by 5% after surgery; ipsilateral strengths increased by 11%. The variation of the response latency and amplitude in the patients who underwent surgery was similar to that of seven control subjects. The postoperative source locations did not differ noticeably from preoperative ones. These findings suggest that temporary unilateral conductive hearing loss in adult patients modifies the function of the auditory neural pathway. (Otolaryngol Head Neck Surg 1998;119:125-30.)
Unilateral sound deprivation early in life is known to affect the developing auditory system in experimental animals. Conductively sound-deprived rats and cats show abnormal ipsilateral inhibition in the inferior colliculus to stimulation of the healthy ear.1,2 In cats reared with neonatal cochlear ablation, the mean thresholds of neurons in the auditory cortex ipsilateral to the stimu-
From the Department of Otolaryngology (Drs. Vasama and Ramsay), University Central Hospital; and Department of Neurology (Dr. Mäkelä), Central Military Hospital. Presented at the Annual Meeting of the American Academy of Otolaryngology–Head and Neck Surgery, Washington, D.C., Sept. 29–Oct. 2, 1996. Reprint requests: J. P. Mäkelä, MD, PhD, Department of Neurology, Central Military Hospital, PL 50, 00301 Helsinki, Finland. Copyright © 1998 by the American Academy of Otolaryngology– Head and Neck Surgery Foundation, Inc. 0194-5998/98/$5.00 + 0 23/77/83126
lated ear are the same as those in the contralateral hemisphere, whereas in intact animals the threshold to ipsilateral stimulation is 17 dB higher.3 In human beings, however, congenital unilateral conductive hearing loss appears to have only a slight effect on the development of the central auditory pathways.4,5 Information regarding the effects of temporary unilateral conductive hearing loss on the adult mammalian auditory system is sparse. The ipsilateral inferior colliculus of cats with 3 to 4 months’ ligation of the external meatus receives diminished inhibitory input from the nonligated ear.1 In adult, awake guinea pigs, the evoked response thresholds decrease, amplitudes increase, and latencies shorten in the auditory cortex ipsilateral to the stimulation of the healthy ear after hair cell destruction in the opposite cochlea.6 In human beings, behavioral evidence for binaural plasticity is seen after plugging of one ear. The subjects initially center a sound image created by binaural tones on the basis of equal loudness; after several days, the adjustment is based more on the actual intensity of the tone.7 The possible changes in auditoryevoked potentials caused by unilateral hearing loss have not been studied. The human primary auditory cortex lies embedded in Sylvius’ fissure and is not easily studied even when the skull is opened during surgery. New methods, such as magnetoencephalography (MEG) have, however, made the study of auditory cortical processing feasible in awake human beings, including patients with hearing deficits. MEG records the magnetic field of small synchronous electrical currents that flow perpendicular to the surface of the activated cortex; only the fields tangential to the head surface, flowing in the fissural cortex, produce MEG signals outside the head, whereas more widely used electroencephalography (EEG) is sensitive to both radial and tangential currents. Changes in the electrical conductivity in scalp and skull do not affect MEG, whereas EEG is distorted by them. These properties aid in MEG source analysis; algorithms needed for source calculations are simpler in MEG than in EEG. One method for modeling the source areas of evoked potential and evoked field is calculation of equivalent current dipoles (ECDs). In such a model the cortical region producing the signal is modeled by a pointlike source, whose location is the center of gravi125
126 VASAMA et al.
Fig. 1. Subject positioned for measurement under the instrument. (Courtesy Neuromag Ltd.)
ty of the cortical area that contributes to the recorded potentials and magnetic fields.8-11 Transient sounds elicit auditory-evoked magnetic fields (AEFs) in the human supratemporal auditory cortex. The most prominent deflection, the N100m, has a peak with a latency of about 100 msecs after sound onset.9,12 N100m latency, amplitude, and source location have been found to be highly reproducible in 30 successive recordings during a 2-month period.13 We studied the effect of correction of temporary unilateral conductive hearing loss, caused by otosclerosis or ossicular chain disruption, on the function of the auditory pathway in adult human beings by comparing preoperative and postoperative AEFs. The measurements were made with a whole-scalp neuromagnetometer, which allows simultaneous recording of the activity from the auditory cortexes of both hemispheres. METHODS AND MATERIAL AEFs were recorded from seven adult patients (4 female, 3 male; age range, 26 to 51 years; mean age, 38 years) with unilateral conductive hearing loss before and after middle ear surgery: five with otosclerosis (patients 1, 3, 4, 5, and 7) and two with disrupted ossicular chains (patients 2 and 6). Five patients had right-sided (patients 1 through 5) and two had left-sided (patients 6 and 7) hearing loss. The mean (±SD)
Otolaryngology– Head and Neck Surgery July 1998
preoperative pure-tone average (PTA; the average of 0.5, 1, and 2 kHz) in the affected ear was 57 ± 12 dB hearing level; the mean postoperative PTA was 17 ± 8 dB. The mean preoperative bone conduction threshold (for frequencies 0.5, 1, and 2 kHz) in the affected ears was 17 dB; the corresponding postoperative value was 15 dB. The duration of hearing loss before surgery varied from 6 to 14 years. The patients were otherwise healthy, and each had a normal auricle, external ear canal, and tympanic membrane on the affected side. The audiogram findings were normal (hearing threshold ≤20 dB for frequencies ≤8 kHz) in the healthy ears. In patients with otosclerosis, stapedioplasty was performed with a Teflon prosthesis. An autologous incus body was sculptured to fit between the manubrium of the malleus and stapes head in patient 2 and between the manubrium of the malleus and stapes footplate in patient 6. The control group consisted of seven healthy adults (2 female, 5 male; age range, 25 to 53 years; mean age, 38 years) without history of hearing deficits. The study was approved by the local ethical committee, and the subjects gave informed consent to be in the study. Tone bursts (50 msec, 1 kHz, 20-msec linear rise and fall times) were delivered to the subject’s healthy, nonoperated ear through plastic tubes, and earpieces were inserted at interstimulus intervals (ISIs) of 1, 2, and 4 seconds. The sound intensity was 80 dB sound pressure level at the earpiece. MEG measurements were performed 1 to 4 months before and 2 months after the operation. MEG signals were measured noninvasively outside the head with a whole-scalp 122-channel neuromagnetometer (Neuromag-122; Biomag Laboratory, Medical Engineering Centre, Helsinki University Hospital, Helsinki, Finland). The measurements were carried out in a magnetically shielded room. The subject was seated with the eyes open and with the head leaning against the helmet-shaped bottom surface of the neuromagnetometer (Fig. 1). Each sensor unit contains a pair of orthogonal gradiometers that measure the tangential derivatives in two orthogonal directions of the magnetic field component normal to the helmet surface at the sensor location. The planar gradiometers pick up the strongest signals from source currents just below the sensor. The exact location of the head, with respect to the superconducting sensors, was determined by measuring magnetic signals produced by indicator currents, led through coils placed at known locations on the scalp. The positions of the coils with respect to the outer landmarks of the head were obtained by means of a threedimensional digitizer. This procedure provided reliable comparison between successive measurements and allowed alignment of the MEG and MRI coordinates, necessary for combination of functional and anatomic structural information. An MRI was obtained from patient 2 by use of a 1-T Siemens Magnetom instrument. The recording passband was 0.03 to 100 Hz, and the sam-
Otolaryngology– Head and Neck Surgery Volume 119 Number 1
A
VASAMA et al. 127
B
Fig. 2. A, AEFs of patient 2 to left-ear stimulation at 2-second ISI. In each response pair, the upper trace illustrates the field derivative along the latitude (latit.) and the lower trace along the longitude (longit.). The head is viewed from above, and the nose points upwards. The passband is 0.03 to 40 Hz, and the traces show a 500-msec time period starting 50 msecs before the stimulus onset. B, Source strengths (Q) and goodness-of-fit curves [g (%)] from responses to left-ear stimulation at 4-second ISI from the left (LH) and right (RH) hemispheres in the same patient.
pling rate was 0.4 kHz. The 500-msec analysis period included a 50-msec prestimulus baseline. The vertical and horizontal electro-oculograms were used to reject data contaminated by eye movements and blinks. For all ISIs, 65 to 110 single responses were averaged to obtain an adequate signal-tonoise ratio. Two ECDs, one in each hemisphere, were used to explain the magnetic field patterns during the N100m. Each ECD, modeling synchronous firing of neurons in a cortical area of less than a few square centimeters, was first found separately for the left- and right-hemisphere data by a least-squares search; the calculations resulted in the three-dimensional location, orientation, and strength (dipole moment; Unit = Ampere • Meter) of the ECD. The goodness of fit of the dipole model, which tells in percentage how much the dipole accounts for the measured field pattern, was calculated according to the method described earlier.10,14 Differences in peak latencies and source strengths were evaluated by use of paired, two-tailed t tests. The test-retest variability of the peak latencies of N100m to tones given monaurally once every 2.5 seconds has been previously estimated with the same recording equipment in a different population of five healthy subjects. In two sets of measurements 1 to 3 weeks apart, the contralateral N100m peaked at 103 ± 11 msecs in the first session and at 103 ± 11 msecs in the second one; corresponding ipsilateral values were 109 ± 10 msecs and 108 ± 8 msecs, respectively (J. Virtanen, Personal communication).
RESULTS
The preoperative whole-scalp AEFs to left-ear stimulation, presented at a 2-second ISI in patient 2 (who had right-sided hearing loss), shows responses over both temporal areas (Fig. 2A). The N100m peak occurs with a latency of 95 msecs over the right hemisphere and 2 msecs later over the left. The estimated strengths of the left- and right-hemisphere ECDs to left-ear stimulation with a 4-second ISI in patient 2 show that the postoperative dipoles reach the maximum amplitude earlier than preoperative ones (Fig. 2B). In the largest responses of all patients from both hemispheres to stimulation of the healthy ear at the 2second ISI, N100m peaks at 76 to 113 msecs over the left and at 72 to 98 msecs over the right hemisphere (Fig. 3). Patients 1, 3, 4, 5, and 6 have shorter postoperative than preoperative N100m latencies over both hemispheres. In patients 2 and 7, the postoperative N100m latency is shorter than the preoperative one over the left hemisphere. Patients 1, 3, 4, and 5 have larger postoperative than preoperative N100m amplitudes over both hemispheres. Patient 2 has a larger postoperative than preoperative N100m amplitude over the right hemisphere, and patient 6 over the left hemisphere, respectively. After the operation, the mean N100m peak latency to nonoperated left-ear stimulation (patients 1 through 5) was 10 msecs shorter over the right (p < 0.01) and 4
128 VASAMA et al.
Otolaryngology– Head and Neck Surgery July 1998
Fig. 3. Responses of patients to preoperative and postoperative stimulation of the healthy ear at 2-second ISI. Channels showing the largest amplitudes over each hemisphere were chosen for display. L, Left ear stimulation; R, right ear stimulation; P, patient number.
msecs shorter over the left hemisphere (nonsignificant). Such clear latency differences were not seen to rightear stimulation in patients 6 and 7 (Table 1). To left-ear stimulation, mean postoperative source strength was 31% stronger (p < 0.01) over the right hemisphere than the preoperative one and 21% stronger over the left
Fig. 4. The preoperative source location for N100m over both hemispheres, superimposed on the MRI scan of patient 2. Circles indicate the source, and tails illustrate the current orientation. The left ear was stimulated at 2second ISI. L, Left hemisphere; R, right hemisphere.
Otolaryngology– Head and Neck Surgery Volume 119 Number 1
VASAMA et al. 129
Table 1. Preoperative and postoperative latencies and source strengths of equivalent current dipoles for N100m, pooled over all ISIs Left ear stimulation Parameters
Latency (msec) Pre healthy Post healthy Post affected Control Q (nAm) Pre healthy Post healthy Post affected Control
RH
Right ear stimulation
LH
95 ± 13 (5) 85 ± 7* 84 ± 6 (2) 87 ± 8 (7)
95 ± 17 (5) 91 ± 5 95 ± 8 (2) 102 ± 5 (7)
22 ± 10 (5) 29 ± 14* 52 ± 15 (2) 24 ± 11 (7)
16 ± 7 (5) 20 ± 9* 20 ± 12 (2) 19 ± 11 (7)
RH
83 ± 8 (2) 83 ± 9 102 ± 6 (5) 103 ± 7 (7) 29 ± 11 (2) 29 ± 11 15 ± 5 (5) 19 ± 7 (7)
LH
80 ± 4 (2) 79 ± 8 94 ± 11 (5) 96 ± 6 (7) 36 ± 11 (2) 25 ± 6* 21 ± 11 (5) 25 ± 13 (7)
Data expressed as mean ± SD or range (no. of subjects). RH, Right hemisphere; LH, left hemisphere; Q, source strength; nAM, nanoamperemeters; Pre, preoperative and Post, postoperative values to stimulation of healthy or affected ear. *Statistically significant difference between preoperative and postoperative values.
hemisphere (p < 0.01). However, to right-ear stimulation the mean postoperative source strength was 31% weaker over the left hemisphere (p < 0.01) than the preoperative one. Both preoperative and postoperative latency and source strength values of the patients were within the normal variance of the control group (Table 1). We also measured the postoperative N100m peak latencies and source strengths to stimulation of the operated ear with improved hearing (i.e., right ear in patients 1 through 5 and left ear in patients 6 and 7) (Table 1). The mean values did not differ significantly from those in controls. Preoperative ECDs for N100m at the 2-second ISI, superimposed on the MRI of patient 2, show one source in each supratemporal cortex (Fig. 4), implying activation of the auditory cortexes within Sylvius’ fissures. There were no significant differences between preoperative and postoperative source locations in patients or between the source locations of the controls and the patients. DISCUSSION
The responses to stimulation of the healthy, nonoperated ear changed after the operation in all patients, suggesting that there was reorganization of the binaural interaction subsequent to resolving the unilateral conductive hearing loss. Either the latencies were shorter or the source strengths were stronger after the middle ear surgery and hearing improvement. Although the difference between preoperative and postoperative latency and source strength values of the patients was obvious, they were within the variance of the normal-hearing control group. Other studies have
shown that congenital unilateral conductive hearing loss up to 70 dB has little effect on AEFs,4 and it is therefore not surprising that the preoperative responses of our patients with unilateral hearing loss were within normal limits. However, the postoperative changes were statistically significant and probably reflect the ability of the auditory system to modify its functions to maintain a balance of input from the two ears. Animal experiments have shown that unilateral cochlear destruction decreases ipsilateral-contralateral latency differences.6 Analogously, in previous studies, three of six patients with unilateral congenital conductive hearing loss and four of eight patients with unilateral idiopathic sudden sensorineural hearing loss had shorter N100m latencies and stronger source strengths in the hemisphere ipsilateral to the stimulated healthy ear.4,15 Furthermore, ipsilateral-contralateral latency differences decreased in some patients during recovery from sudden unilateral hearing loss because of acoustic neuroma removal.16 As expected from these results, the restoration of normal hearing increased the ipsilateralcontralateral latency difference in our patients. A continuous tone of moderate intensity may enhance the amplitude of cortical responses evoked by brief acoustic stimuli.17 In line with this, N100m amplitude to brief, monaural tones is enhanced during a continuous tone presented to the other ear.18 Improved hearing of environmental sounds after operation could cause analogous enhancement in the hemisphere contralateral to the operated ear. However, the N100m amplitude enhancement by the continuous tone is accompanied by an increase, not by a decrease, of latency.18 Obviously, additional mechanisms are needed to explain the latency decrease observed in this study.
Otolaryngology– Head and Neck Surgery July 1998
130 VASAMA et al.
In animal experiments, cortical and subcortical maps reorganize after restricted damage of cochlear receptors, and the cortical area normally representing frequencies of the lesioned cochlear portion is occupied by neighboring frequencies. Nevertheless, this does not mean central compensation of peripheral hearing loss; sensitivity to frequencies normally activating the damaged cochlear region is not restored.19 On the other hand, enhanced mean responses of single cells and increase of the coherence between activities of neurons participating in the same processing assembly during recovery from lesion would lead to behavioral improvement by means of increased reliability and shortened duration of neural processing.20 The increase of N100m amplitude and shortening of its latency after surgery could reflect an increase of coherence of the neural events underlying N100m, and a shortening of the auditory processing time. Thus they may be related to the improved auditory capacity. This is supported by our previous findings in patients with unilateral profound deafness after an acoustic neuroma operation. Their N100ms were initially smaller in amplitude and delayed in both hemispheres and recovered toward normal values during the first postoperative year. The alteration was attributed to enhanced coherence of postsynaptic potentials underlying N100m during recovery.16 This study shows the adaptability of the adult human auditory system after unilateral middle ear surgery to improve hearing. It also shows the benefit to otolaryngologists from a fast and noninvasive MEG method for investigation of auditory cortical functions in normal subjects and in hearing-impaired patients. We thank S. Heikkilä for help in patient preparation for MEG measurements, A. Korvenoja and S. Martinkauppi for MRIs, and J. Virtanen for test-retest data of normal subjects. R. Hari and A. Møller made valuable comments on the manuscript. REFERENCES 1. Moore DR, Irvine DRF. Plasticity of binaural interaction in the cat inferior colliculus. Brain Res 1981;208:198-202. 2. Silverman MS, Clopton BM. Plasticity of binaural interaction I. Effect of early auditory deprivation. J Neurophysiol 1977;40:1266-74. 3. Reale RA, Brugge JF, Chan CK. Maps of auditory cortex in cats
4. 5. 6.
7. 8.
9.
10.
11.
12.
13.
14.
15. 16. 17. 18. 19. 20.
reared after unilateral cochlear ablation in the neonatal period. Dev Brain Res 1987;34:281-90. Vasama J-P, Mäkelä JP, Parkkonen L, Hari R. Auditory cortical responses in humans with congenital unilateral conductive hearing loss. Hear Res 1994;78:91-7. Wilmington D, Gray L, Jahrsdoerfer R. Binaural processing after corrected congenital unilateral conductive hearing loss. Hear Res 1994;74:99-114. Popelar J, Erre J-P, Aran JM, Cazals Y. Plastic changes in ipsicontralateral differences of auditory cortex and inferior colliculus evoked potentials after injury to one ear in the adult guinea pig. Hear Res 1994;72:125-34. Florentine M. Relation between lateralization and loudness in asymmetrical hearing loss. J Am Audiol Soc 1976;1:243-51. Vasama J-P, Mäkelä JP. Auditory pathway reorganization in humans with congenital or acquired unilateral hearing loss. In: Salvi R, Henderson D, Fiorino F, Colletti V, editors. Auditory system plasticity and reorganization. New York: Thieme; 1996. p. 359-70. Hari R. The neuromagnetic method in the study of the human auditory cortex. In: Grandori F, Hoke M, Romani GL, editors. Auditory evoked magnetic fields and potentials. Advances in audiology. Vol. 6. Basel: Karger; 1990. p. 222-82. Hämäläinen M, Hari R, Ilmoniemi R, Knuutila J, Lounasmaa OV. Magnetoencephalography—theory, instrumentation and applications to noninvasive studies of the working human brain. Rev Mod Phys 1993;65:413-97. Mäkelä JP, Hari R. Long-latency auditory evoked magnetic fields. In: Sato S, editor. Magnetoencephalography: clinical applications and comparison of EEG. New York: Raven Press; 1990. p. 177-91. Mäkelä JP, Ahonen A, Hämäläinen M, et al. Functional differences between auditory cortices of the two hemispheres revealed by whole-head neuromagnetic recordings. Hum Brain Mapping 1993;1:48-56. Pantev C, Gallen C, Hampson S, Buchanan S, Sobel D. Reproducibility and validity of neuromagnetic source localization using a large array biomagnetometer. Am J EEG Technol 1991;31:83-101. Kaukoranta E, Hämäläinen M, Sarvas J, Hari R. Mixed and sensory nerve stimulations activate different cytoarchitectonic areas in the human primary somatosensory cortex SI. Exp Brain Res 1986;63:60-6. Vasama J-P, Mäkelä JP. Auditory pathway plasticity in adult humans after idiopathic sudden unilateral sensorineural hearing loss. Hear Res 1995;87:132-40. Vasama J-P, Mäkelä JP, Pyykkö I, Hari R. Abrupt unilateral deafness modifies function of human auditory pathways. Neuroreport 1995;6:961-4. Gerken GM. Alteration of central auditory processing of brief stimuli: a review and a neural model. J Acoust Soc Am 1993;93:2038-49. Loveless N, Vasama J-P, Mäkelä JP, Hari R. Human auditory cortical mechanisms of sound lateralisation: III. Monaural and binaural shift responses. Hear Res 1994;81:91-9. Irvine DRF, Rajan RR. Plasticity in the frequency organization of auditory cortex of adult mammals with restricted cortical lesions. Biomed Res 1993;14:55-9. Ahissar E, Ahissar M. Plasticity in auditory cortical circuitry. Curr Opin Neurobiol 1994;4:580-7.