International Journal of Audiology 2014; 53: 360–369
Original Article
Bone conduction hearing sensitivity in normal-hearing subjects: Transcutaneous stimulation at BAHA vs BCI position
Int J Audiol Downloaded from informahealthcare.com by Ball State University on 11/20/14 For personal use only.
Sabine Reinfeldt*, Bo Håkansson*, Hamidreza Taghavi* & Måns Eeg-Olofsson† *Department of Signals and Systems, Chalmers University of Technology, Gothenburg, Sweden, and †Department of Otorhinolaryngology, Head and Neck Surgery, Sahlgrenska University Hospital, The Sahlgrenska Academy, University of Gothenburg, Gothenburg, Sweden
Abstract Objective: Bone conduction (BC) stimulation closer to the cochlea has previously been shown to give higher cochlear promontory acceleration measured by laser Doppler vibrometry (LDV). This study is investigating whether stimulation closer to the cochlea also gives improved hearing sensitivity. Furthermore, the study compares shifts in hearing sensitivity (BC thresholds) and ear-canal sound pressure (ECSP). Design: BC hearing thresholds and ECSP have been measured for stimulation at two positions: the existing bone-anchored hearing aid (BAHA) position, and a new bone conduction implant (BCI) position that is located closer to the cochlea. Study sample: The measurements were made on 20 normal-hearing subjects. Results: Depending on frequency, the ipsilateral hearing threshold was 3–14 dB better, and the ipsilateral ECSP was 2–12 dB higher for the BCI than for the BAHA position, with no significant differences between threshold and ECSP shifts at group level for most frequencies, and individually only for some subjects. Conclusions: It was found that both the objective ECSP and the subjective hearing threshold measurements gave similar improvement as previous LDV measurements for stimulation closer to the cochlea. One exception was that the LDV measurements did not show the improved sensitivity for frequencies below 500 Hz found here.
Key Words: Bone conduction; hearing sensitivity; transcutaneous stimulation; bone conduction implant; bone-anchored hearing aid; threshold shift; ear-canal sound pressure; objective measure; normal-hearing subjects
The bone-anchored hearing aid (BAHA) is a viable alternative for patients with conductive or mixed hearing loss and single-sided deafness, and has reached great success worldwide with more than 100 000 patients treated (Cochlear, 2013). However, the percutaneous solution needs life-long daily care and may in some cases give skin complications or even spontaneous implant losses (Battista et al, 2006; Dun et al, 2012; Reyes et al, 2000; Shirazi et al, 2006; Snik et al, 2005; Tjellström et al, 1994, 2012; Wazen et al, 2011). Patients may also reject a BAHA solution for aesthetic reasons. With the new bone conduction implant (BCI) system, these problems can be avoided. The transducer is implanted under the skin in the temporal bone without use of a percutaneous screw. The BCI system consists of an external audio processor and an implanted unit, named the bridging bone conductor (BBC). To keep the skin and subcutaneous tissue intact, an inductive link is used for the sound and power transmission. The BCI system is described in more details in Håkansson et al (2010) and in Taghavi et al (2012a,b). Another
advantage with the BCI system, as compared to the BAHA, is that feedback is decreased (Taghavi et al, 2012c). The BCI system is now in its first clinical trial with results of the first patient described in Eeg-Olofsson et al (2014). According to laser Doppler vibrometry (LDV) measurements on cadavers (Håkansson et al, 2008; 2010), the acceleration of the ipsilateral cochlear promontory increases when changing from a BAHA transducer at the BAHA position (position A in Figure 1) to a BCI transducer at the suggested BCI position 25 mm behind the posterior border of the bony ear-canal (position B in Figure 1), which is closer to the cochlea and located in the mastoid portion of the temporal bone. It should be noted that the LDV only measures the vibration level in one direction, i.e. in line with the ear-canal through which the light beam was pointed, at a reflector at the cochlear promontory. As the cochlea is sensitive to vibrations in all directions this can be a limiting factor when estimating the real hearing sensitivity from the LDV. However, as the stimulation direction in a bone-conduction
Correspondence: Sabine Reinfeldt, Department of Signals and Systems, Chalmers University of Technology, SE-412 96 Gothenburg, Sweden. E-mail: sabine.reinfeldt@ chalmers.se (Received 13 May 2013; accepted 1 January 2014 ) ISSN 1499-2027 print/ISSN 1708-8186 online © 2014 British Society of Audiology, International Society of Audiology, and Nordic Audiological Society DOI: 10.3109/14992027.2014.880813
BC sensitivity – BAHA vs BCI position
Int J Audiol Downloaded from informahealthcare.com by Ball State University on 11/20/14 For personal use only.
Abbreviations AC BAHA BBC BC BCI BEST contra CT ECSP HL ipsi LDV SD TM TT vs
Air conduction Bone-anchored hearing aid Bridging bone conductor Bone conduction Bone conduction implant Balanced electromagnetic separation transducer contralateral Computed tomography Ear-canal sound pressure Hearing level ipsilateral Laser Doppler vibrometry Standard deviation Tympanic membrane Transcranial transmission versus
hearing device is perpendicular to the skull bone and approximately in line with the measurement direction along the ear-canal, this direction might be the most important contributor to the real hearing response using a BAHA or a BCI (Stenfelt et al, 2000, 2005; Reinfeldt et al, 2013). It was also shown in the cadaver studies that the acceleration of the contralateral cochlea with full systems in a sound field was lower with the BCI than with the BAHA. In summary, these LDV results may indicate that the hearing sensitivity increases for ipsilateral stimulation and is slightly lower for contralateral stimulation by using a BCI as compared to using a BAHA. Eeg-Olofsson et al (2008; 2011) investigated the promontory velocity from LDV measurements by stimulating at eight positions
361
between the BAHA position and a position close to the ear-canal, using the same transducer at all positions. The conclusions were that the ipsilateral promontory velocity was higher for closer stimulation, and that the transcranial transmission (TT) differences between the positions were large at positions close to the cochlea and at high frequencies. So far, the LDV results from the cadaver studies have not yet been verified against psychoacoustical measurements to confirm that the hearing sensitivity really is increased when stimulating closer to the cochlea. However, one study has been performed on BAHA patients with a unilateral middle-ear common cavity (Eeg-Olofsson et al, 2013). LDV measurements were compared with BC hearing thresholds, showing similar trends for vibration and threshold measurements on a group level, but low correlation at the individual level. LDV measurements are usually not feasible to perform on live humans since the promontory is not visible from outside. Therefore another objective measurement method would be desirable as a substitute for hearing thresholds, e.g. when examining the performance of a bone-conduction (BC) device. It has previously been shown that relative BC hearing thresholds are similar to relative ear-canal sound pressure (ECSP) at most frequencies (Reinfeldt et al, 2013). That study was done in normal-hearing subjects with open ear-canals and for stimulation positions far from each other. The ECSP reflects the sound radiated into the ear-canal from the surrounding structures. Since the ear-canal is close to the cochlea, it has been hypothesized that a change of the ECSP level is correlated to a BC sensitivity change (Reinfeldt et al, 2013). A simple experiment can be made to investigate if the BC hearing sensitivity increases by moving the stimulation position closer to the cochlea, or more specifically from the BAHA to the BCI position. By pressing a test rod attached to a BAHA device, e.g. Baha Intenso™ and test rod from Cochlear® (Cochlear Bone-Anchored Solutions, Sweden), against the skin at the BAHA and BCI positions (the test rod should not touch the pinna) with closed ear-canals and listen to some music, you will clearly hear a difference in sound level. Generally, a louder sound is heard when the test rod is pressed at the BCI position (approximately 25 mm behind the center of the ear-canal) as compared to the BAHA position (55 mm behind the ear-canal). This result is in line with the previous LDV measurements on cadavers and the method may be used to verify hearing sensitivity differences between the two positions. A first aim of this study is to investigate the BC hearing sensitivity difference by measuring BC tone thresholds for stimulation at the BAHA position and a position close to the suggested BCI position on normal-hearing subjects. A second aim is to measure the ECSP for the same stimulation positions and to compare the two measurement methods. The results will also be compared to previous LDV measurements on cadavers in Håkansson et al (2008; 2010), and Eeg-Olofsson et al (2008; 2011) to verify the validity of these LDV measurements regarding hearing sensitivity differences.
Materials and Methods Subjects Figure 1. Position A (BAHA position) is 55 mm from the entrance of the ear-canal at the line extending from the lateral margin of the eye to the upper border of the pinna; and position B (BCI position) is on a line between the ear-canal and position A, about 25 mm from the ear-canal entrance as close as possible behind the pinna.
All measurements were performed on 20 voluntary subjects (10 male and 10 female, ages 23 to 44 years, with an average of 30 years) with otologically-normal ears (ISO 226, 2003). Their air-conduction (AC) hearing thresholds were no worse than 20 dB HL for frequencies 125 to 8000 Hz and the interaural sensitivity difference was maximum 15 dB at any frequency
362
S. Reinfeldt et al.
measured (125, 250, 500, 750, 1000, 1500, 2000, 3000, 4000, 6000, and 8000 Hz). One test subject was included in the study despite that his right ear hearing threshold was 25 dB HL at 8000 Hz (that frequency was excluded from the analysis).
Int J Audiol Downloaded from informahealthcare.com by Ball State University on 11/20/14 For personal use only.
Measurement setup A sound insulated room of 16 m3 was used for all measurements. To determine the subjects’ baseline hearing tone thresholds for inclusion in the study, the Hughson-Westlake procedure according to ISO 8253-1 (2010) was used with TDH39 headphones and a digital audiometer (AC40; Interacoustics A/S, Assens, Denmark). For BC stimulation, a modified BEST transducer (Håkansson, 2003), was attached to the skin over the skull bone with a steel spring headband. In order to reduce the skin dampening effect, a higher static pressure (6–10 N) and a smaller contact surface area (38 mm2) than normal in BC audiometry (175 mm2 and 5.4 N according to ISO 389-3 (1994)) was used. The adapter has a tip diameter of 7 mm (effective diameter of 6 mm as there is a radius of 0.5 mm in the edge towards the skin; the real BCI attachment area to the skull bone is 6 mm), and a height of 4 mm as shown in Figure 2. The transducer was pressed against the skin at the two positions illustrated in Figure 1 and driven by a signal analyzer (Agilent 35670A, Agilent Technologies Inc., USA) via a 10 Ω current limiting resistor. Position A was approximately 55 mm from the entrance of the ear-canal at a line extending from the lateral margin of the eye to the upper border of the pinna (position A is where a BAHA normally is positioned). Position B is on a line between the ear-canal and position A, about 25 mm from the center of the ear-canal entrance, closest possible behind the pinna and as close to the suggested position for a BCI as possible. The BC hearing thresholds were measured using pure tones according to the Békésy procedure in the AC40 audiometer. A pulsed sequence (instead of constant) was used to make it easier for the patient to distinguish the tone from the masking noise. Frequencies were 125, 250, 500, 750, 1000, 1500, 2000, 3000, 4000, 6000, and 8000 Hz, and the step level was 1 dB. When measuring BC hearing thresholds, masking must be used to ensure perception in the test-ear and avoid false thresholds from the non-test ear. In this study, the masking noise was a broadband noise spectrally designed to be of equal level above the normal hearing thresholds (0 dB HL) for all 1/3-octave frequency bands
(see Reinfeldt et al (2010) for more details). It was fed from a CD player via a power amplifier (Sony TA-N220; Sony Corp., Tokyo, Japan) to an insert earphone (Etymotic Research ER-2; Etymotic Research Inc., USA). The foam part of the insert earphone was modified to avoid occlusion effect (Stenfelt & Reinfeldt, 2007). The sound pressure in the ear-canals (ECSP) was measured by a microphone (Knowles EM3046; Knowles, USA), originally used in BAHA devices. This microphone was chosen instead of the Etymotic ER-7 probe tube microphone as it provided a lower noise floor. The same probe tube was used as in the Etymotic ER-7 system. The inlet to the EM3046 fits perfectly to this probe tube with no leakage that could have altered the measurement results. The microphone was connected via a battery operated low noise linear preamplifier (Gennum LC506; Semtech Corp., USA), which provided the signal to the Agilent signal analyser in order to measure the frequency response between the ECSP output and the input voltage to the BC transducer. Measurements of both ipsilateral and contralateral ECSP were performed. To avoid both AC sound transmission and occlusion effect when measuring the ECSP, an ear-plug was deeply inserted in the ear-canal. The AC sound was attenuated with 20 to 40 dB, depending on frequency, by the ear-plug (Reinfeldt et al, 2007) and could therefore be neglected. Stenfelt and Reinfeldt (2007) showed that the deeper an ear-plug is inserted, the less occlusion effect is obtained. For the ECSP measurements, the length of the ear-canal was first determined by entering a probe tube until it touched the tympanic membrane (TM). Then, the probe tube was inserted into an ear-plug (E-A-R Classic; 3M Corp., USA) by using a catheter; it was tightly sealed by the foam of the ear-plug. The ear-plug was then placed in the ear-canal with the probe tube tip approximately 3 mm from the TM and the end of the ear-plug close to the probe tube tip. A contact between the transducer housing and the pinna could give a false ECSP caused by vibrations transmitted by soft tissues and the ear-plug into the remaining open part of the ear-canal. To place the BC transducer at position B without this contact, the pinna was folded by using a wide sport tape (Leukoplast).
Measurements BC hearing thresholds were obtained ipsi- and contralaterally for stimulation at positions A and B. Correspondingly, ECSP were measured for the same conditions. BC stimulation was always done on the same side of the head (left), while ECSP measurements and hearing thresholds were performed in both ears. In total eight measurements were performed; they are presented in Figure 3 and listed as follows: • • • •
Figure 2. BEST with adapter having a tip diameter of 7 mm (effective diameter of 6 mm as there is a radius of 0.5 mm in the edge towards the skin) and height of 4 mm.
Threshold ipsi A and B ECSP ipsi A and B Threshold contra A and B ECSP contra A and B
The measurements were varied in ten different orders for the test subjects to avoid order effects. The orders were constructed to give as few changes as possible. The masking ear-phone and the ear-plug were switched between hearing thresholds and ECSP measurements. Since the static pressure towards the skin of the BC transducer was relatively high, the transducer was temporarily removed after each measurement and the stimulation position was switched for the next measurement. It was important to reattach the transducer at the same
BC sensitivity – BAHA vs BCI position
(a)
(a) A
A
B
B
ECSP ipsi A and B
Masking for thresholds
‘Contra ECSP, B vs A’
‘Ipsi ECSP, B vs A’ ‘Ipsi thresholds, B vs A’
Threshold ipsi A and B
(b)
(b)
A
‘Contra thresholds, B vs A’
A B
B
Int J Audiol Downloaded from informahealthcare.com by Ball State University on 11/20/14 For personal use only.
363
Masking for thresholds
ECSP contra A and B
‘A thresholds, ipsi vs contra’ ‘B thresholds, ipsi vs contra’ ‘A ECSP, ipsi vs contra’ ‘B ECSP, ipsi vs contra’
Threshold contra A and B
Figure 3. (a) Ipsilateral hearing thresholds and ECSP for stimulation positions A and B. (b) Contralateral hearing thresholds and ECSP for stimulation positions A and B. Solid lines are thresholds and dashed lines are ECSP measurements.
Figure 4. (a) Calculations of hearing threshold and ECSP shifts for position B vs position A. (b) Calculations of hearing threshold and ECSP shifts for contralateral vs ipsilateral response.
positions and therefore the positions were marked with a pen before the measurements started.
calibrated using a Brüel & Kjær 4230 sound level calibrator. The probe tube connected to the microphone was then positioned as close as possible to the reference microphone in a Brüel & Kjær anechoic test chamber type 4222 with the assumption that the microphones measured the same sound pressure. By this procedure, a frequency-dependent microphone sensitivity was determined and applied to the microphone output voltage to convert to sound pressure. The output from the BC transducer was calibrated according to ISO 389-3 (1994) using a Brüel & Kjær 4930 artificial mastoid, and the dynamic output force (in dB relative 1 Newton) from the BC transducer against the pad of Brüel & Kjær 4930 was related to the output of the artificial mastoid by using a pad correction. The audiometer (AC40) was calibrated according to ISO 389-1 (1998) by the manufacturer for threshold measurements with earphones. Since all measurements were comparative, there was no need for calibration, except for the baseline hearing thresholds, described above. However, the calibrations were done before and after all measurements to ensure that there were no changes in the performance of the equipment during the study period.
Calculations The comparisons between different measurements are shown in Figure 4. Positive shifts for position B vs A means better hearing or higher ECSP for stimulation at position B than at position A. Positive shifts for ipsilateral vs contralateral response means better hearing or higher ECSP ipsilaterally than contralaterally. For all measures, the averages of all subjects were calculated together with the standard deviation (SD). To assess whether the hearing threshold and ECSP shifts for the stimulation positions were separated from zero, a Wilcoxon signed rank test was performed between the positions both for thresholds and ECSP. The null-hypothesis was that there were no differences between the positions. A two-tailed paired t-test was used to test similarities between the hearing threshold and ECSP shifts with the null-hypothesis that there were no differences between the methods. The test was performed only at frequencies where thresholds were measured and where the microphone was above the noise floor (nine frequencies between 125 and 4000 Hz).To correct for the probability that a significant result occurs purely by chance when performing independent t-tests at each of the nine tested frequencies, a Sidak correction was applied. Pearson correlation was used to calculate the correlation between hearing threshold shift and ECSP shift at individual level, and also Wilcoxon signed rank test was used to compare ipsilateral threshold shift and ECSP shift B vs A for each individual over all frequencies.
Calibrations The microphone was calibrated using a reference microphone (Brüel & Kjær 4134; Brüel & Kjær Sound & Vibration Measurement A/S, Denmark). The sensitivity of the reference microphone was
Results Ipsilateral and contralateral response, stimulating in position B vs A To illustrate the individual differences in the hearing threshold and ECSP shifts for position B vs A, the shifts for each individual, averages, and SDs are shown in Figure 5. ‘Ipsi thresholds, B vs A’ is shown in Figure 5 (a). The individual differences were large, but generally there is an improved sensitivity at position B relative position A. Also the ‘Ipsi ECSP, B vs A’, in Figure 5 (b), gave a similar pattern. Figure 5 (c) and (d), show the threshold and ECSP shifts for the contralateral side. Here the general trend was a slight
364
S. Reinfeldt et al. 40
40
(c)
30
30
20
20
10 0
10 0
−10
−10
−20
−20
−30
125
250
500
1000
2000
4000
−30
8000
125
250
500 1000 2000 Frequency [Hz]
4000
8000
250
500 1000 2000 Frequency [Hz]
4000
8000
Frequency [Hz] 40
40
(b)
(d)
30
30
20
20 ECSP shift [dB]
ECSP shift [dB]
Int J Audiol Downloaded from informahealthcare.com by Ball State University on 11/20/14 For personal use only.
Threshold shift[dB]
Threshold shift [dB]
(a)
10 0
10 0
−10
−10
−20
−20
−30
125
250
500
1000
2000
4000
8000
Frequency [Hz]
−30
125
Figure 5. Individual (thin solid) and average (thick solid) curves together with SD (bars) for (a) ‘Ipsi thresholds, B vs A’, (b) ‘Ipsi ECSP, B vs A’, (c) ‘Contra thresholds, B vs A’, and (d) ‘Contra ECSP, B vs A’. A positive value means that position B gives a higher sensation than position A.
improvement for position B. The SD’s for all shifts were between 4 and 13 dB. Test-retests were performed and in Figure 6, the hearing threshold shifts (a) and ECSP shifts (b) are shown for one subject measured three times. The measurement and stimulation equipment were removed and reattached between each measurement. The variability is expected in both hearing thresholds and ECSP shifts, e.g. in Reinfeldt et al (2013), the standard deviations for these measures were in general between 4 and 12 dB. In this study, the threshold shifts show a dynamic range of 20 dB at some frequencies and almost no variability at other frequencies. The ECSP shifts had lower variability than the threshold shifts. The average threshold and ECSP shifts between position B and A are summarized in Figure 7, both for the ipsilateral and contralateral sides. Positive values demonstrate better hearing or higher ECSP for stimulation at position B than at position A. The average ‘Ipsi thresholds, B vs A’ were positive for all frequencies. The hearing thresholds showed a shift of between 3 and 14 dB in favor of position B. Since the microphone (together with preamplifier) signal was in its noise floor above 4000 Hz, the ECSP shifts have been disregarded at these high frequencies. ‘Ipsi ECSP, B vs A’ was between 2 and 12 dB. The contralateral shifts were lower than the ipsilateral ones, both for threshold and ECSP shifts, but still mainly positive. ‘Contra
thresholds, B vs A’ were between ⫺3 and 6 dB, and ‘Contra ECSP, B vs A’ was between ⫺2 and 11 dB. To investigate if there was a statistical support for improved hearing when moving the stimulation from position A to position B, a Wilcoxon signed rank test was performed ipsilaterally both for thresholds and ECSP. The test showed that thresholds vary with position for α ⫽ 0.05 at all frequencies 125 – 8000 Hz except at 500 Hz, and that ECSP varies with position for α ⫽ 0.05 at all frequencies 125 – 4000 Hz except at 4000 Hz. Hence, it is a regarded as a statistically significant improvement to move the stimulation from position A to position B. The corresponding Wilcoxon signed rank test for contralateral responses showed that thresholds vary with position for α ⫽ 0.05 at 1000 – 3000 Hz, and ECSP at 250 – 500 and 1000 – 2000 Hz. To compare the hearing threshold shifts with the ECSP shifts, a two-tailed paired t-test was performed. The null-hypothesis that there are no differences between the hearing threshold and ECSP shifts for stimulation in position B vs A was confirmed (nonsignificant differences) at most frequencies. The only differences found were ipsilaterally at 750 and 4000 Hz, and contralaterally at 1000 Hz with p ⬍ 0.05. At the individual level, the correlation between the methods was low with an average correlation coefficient
BC sensitivity – BAHA vs BCI position
365
20
40
‘Ipsi thresholds, B vs A’ ‘Ipsi ECSP, B vs A’ ‘Contra thresholds, B vs A’ ‘Contra ECSP, B vs A’
(a) 30 15
B vs A [dB]
Threshold shift [dB]
20 10 0
10
5
−10 0
−20
125
250
500 1000 2000 Frequency [Hz]
4000
8000
(b) 30 20 10 0 −10 −20 −30
125
−5
125
250
500
1000 2000 Frequency [Hz]
4000
8000
Figure 7. Hearing threshold and ECSP shifts, stimulating in position B vs A. Solid line is ‘Ipsi thresholds, B vs A’; dashed line is ‘Ipsi ECSP, B vs A’; dotted line is ‘Contra thresholds, B vs A’; and dashdotted line is ‘Contra ECSP, B vs A’. A positive value means that position B gives a higher sensation than position A.
40
ECSP shift [dB]
Int J Audiol Downloaded from informahealthcare.com by Ball State University on 11/20/14 For personal use only.
−30
250
500 1000 2000 Frequency [Hz]
4000
8000
Figure 6. Test-retest on one subject for (a) ‘Ipsi thresholds, B vs A’, and (b) ‘Ipsi ECSP, B vs A’. A positive value means that position B gives a higher sensation than position A. of 0.25 for ipsilateral comparison and 0.41 for contralateral comparison. However, a Wilcoxon signed rank test for each individual and ipsilateral comparison showed that the null hypothesis could be rejected for 18 out of the 20 subjects with α ⫽ 0.05, hence showing that no statistical differences could be shown between hearing threshold shifts and ECSP shift for 18 of 20 subjects. For the contralateral comparison, no statistical differences could be shown between hearing threshold shifts and ECSP shifts for 16 of 20 subjects.
Ipsilateral vs contralateral response, stimulating in position A and B Figure 8 shows the individual differences in the hearing threshold and ECSP shifts between ipsilateral and contralateral responses (also termed transcranial attenuation) with the shifts for each individual, averages, and SD. Positive values mean higher hearing sensitivity or higher ECSP level on the ipsilateral than on the contralateral side. ‘A thresholds, ipsi vs contra’ are shown in Figure 8 (a), and ‘A ECSP, ipsi vs contra’ in Figure 8 (b). They were around 0 dB at low frequencies and a few dB’s positive at higher frequencies. ‘B thresholds, ipsi vs contra’ and ‘B ECSP, ipsi vs contra’ are shown in Figure 8 (c) and (d), respectively, showing higher shifts than for position A.
SDs of all measures were between 6 and 12 dB. The individual differences were large, but generally stimulation at position B gave higher shifts than position A, and hence lower hearing sensitivity and lower ECSP level contralaterally than ipsilaterally. In Figure 9, the averages of the hearing threshold and ECSP shifts between ipsilateral and contralateral measurements are shown for position A and for position B. ‘A thresholds, ipsi vs contra’ was low, between ⫺ 3 and 8 dB. ‘A ECSP, ipsi vs contra’ was between ⫺ 3 and 10 dB. For position B, the difference between ipsilateral and contralateral measurements was higher: between 2 and 14 dB for ‘B thresholds, ipsi vs contra’, and between 3 and 10 dB for ‘B ECSP, ipsi vs contra’. To statistically investigate if shifting between ipsilateral and contralateral stimulation gives a difference in hearing, the Wilcoxon signed rank test was used. The null hypothesis that there were no differences between the sides was rejected at 2000 – 3000 and 8000 Hz for thresholds at position A, and at 1000 and 3000 – 4000 Hz for ECSP at position A. At position B, the null hypothesis was rejected at 250 and 750 – 6000 Hz for thresholds, and at 125 – 250 and 750 – 4000 Hz for ECSP. Hence, the frequency ranges with significant differences were wider for position B. The t-test for the transcranial measurements indicated that the hearing threshold and ECSP shifts are similar. There were non-significant differences for all frequencies except for position A at 4000 Hz and for position B at 125 Hz, both with p ⬍ 0.05. At the individual level, the correlation between the methods was low with an average correlation coefficient of 0.29 for position A and 0.23 for position B. When performing the Wilcoxon signed rank test for each individual, there was significant difference with α ⫽ 0.05 for 17 of the 20 subjects in position A, and for 14 of the 20 subjects in position B.
Discussion Despite general large individual variability, there are certain trends identified among the results. First of all, position B gives a higher hearing sensitivity than position A for ipsilateral stimulation, which may be of significant importance for the rehabilitation with a BCI device. Secondly, it seems that the ECSP can be used as an
S. Reinfeldt et al.
366
50
50 40
(a)
40 30
20 10 0 −10
10 0 −10 −20
−30
−30
40
125
250
500 1000 2000 Frequency [Hz]
4000
−40
8000
40
30
30
20
20
10 0
250
500 1000 2000 Frequency [Hz]
4000
8000
250
500 1000 2000 Frequency [Hz]
4000
8000
0 −10
−20
−20
−30
−30 125
250
500 1000 2000 Frequency [Hz]
4000
8000
(d)
10
−10
−40
125
50
(b)
ECSP shift [dB]
ECSP shift [dB]
20
−20
50
−40
125
Figure 8. Individual (thin solid) and average (thick solid) curves together with SD (bars) for (a) ‘A thresholds, ipsi vs contra’, (b) ‘A ECSP, ipsi vs contra’, (c) ‘B thresholds, ipsi vs contra’, and (d) ‘B ECSP, ipsi vs contra’. A positive dB value means that the ipsilateral side is more sensitive than the contralateral side, i.e. it is a measure of transcranial attenuation.
14 12 10 ipsi vs contra [dB]
Int J Audiol Downloaded from informahealthcare.com by Ball State University on 11/20/14 For personal use only.
Threshold shift [dB]
Threshold shift [dB]
30
−40
(c)
8 6 4 2 0 −2
‘A thresholds, ipsi vs contra’ ‘A ECSP, ipsi vs contra’ ‘B thresholds, ipsi vs contra’ ‘B ECSP, ipsi vs contra’
−4 −6 125
250
500
1000 2000 Frequency [Hz]
4000
8000
Figure 9. Hearing threshold and ECSP shifts, ipsilateral vs contralateral response, stimulating in position A and B. Solid line is ‘A thresholds, ipsi vs contra’; dashed line is ‘A ECSP, ipsi vs contra’; dotted line is ‘B thresholds, ipsi vs contra’; and dash-dotted line is ‘B ECSP, ipsi vs contra’. A positive dB value means that the ipsilateral side is more sensitive than the contralateral side, i.e. it is a measure of transcranial attenuation.
objective measure instead of the subjective BC hearing thresholds, which may be of importance in situations where thresholds are not suitable (e.g. patient anaesthetized), or in long term investigations requiring lower variability (e.g. follow up of the BCI implant’s performance). However, an additional correlation analysis showed that individual ECSP and threshold shifts do not correlate in the same way as the average shifts do. On the other hand, Wilcoxon signed rank tests for each individual over all frequencies showed that the threshold shift and ECSP shift for a majority of the subjects were not significantly different. For ipsilateral response, B vs A, the threshold and ECSP shifts only differed significantly for 2 of the 20 subjects. In summary, caution should be taken when measuring ECSP instead of thresholds on individual level.
Ipsilateral and contralateral response, stimulating in position B vs A According to Figure 7 and to Wilcoxon’s signed rank tests, better hearing thresholds and higher ECSP were found ipsilaterally for position B compared to position A for all frequencies. This finding supports what was indicated in the studies by Håkansson et al (2008; 2010) and Eeg-Olofsson et al (2008) by using LDV, that the hearing sensitivity is higher for stimulation closer to the cochlea, or more specifically at the BCI position compared to the BAHA
BC sensitivity – BAHA vs BCI position 25
(a)
20
Ipsi, B vs A [dB]
15 10 5 0 ‘Ipsi thresholds, B vs A’ ‘Ipsi ECSP, B vs A’ Eeg−Olofsson et al. (2008) Håkansson et al. (2010)
−5
125
250
500 1000 2000 Frequency [Hz]
4000
8000
10 8
(b)
A, ipsi vs contra [dB]
6 4 2 0 −2 −4 −6
‘A thresholds, ipsi vs contra’ ‘A ECSP, ipsi vs contra’ Eeg−Olofsson et al. (2011) Stenfelt (2012)
−8 −10
20
125
250
(c)
500 1000 2000 Frequency [Hz]
4000
8000
4000
8000
‘B thresholds, ipsi vs contra’ ‘B ECSP, ipsi vs contra’ Eeg−Olofsson et al. (2011)
15 B, ipsi vs contra [dB]
Int J Audiol Downloaded from informahealthcare.com by Ball State University on 11/20/14 For personal use only.
−10
10
5
0
−5
−10
125
250
500 1000 2000 Frequency [Hz]
Figure 10. (a) Comparison between ‘Ipsi thresholds, B vs A’ (solid line); ‘Ipsi ECSP, B vs A’ (dashed line); Ipsi promontory acceleration shift position 4 vs position 1 (Eeg-Olofsson et al, 2008 (dotted line); and Ipsi promontory acceleration shift B vs A (Håkansson et al, 2010) (dash-dotted line). Håkansson et al (2010) used different transducers at position A and B. (b) Comparison between ‘A thresholds, ipsi vs contra’ (solid line); ‘A ECSP, ipsi vs contra’ (dashed line); A promontory acceleration shift ipsi vs contra derived from Eeg-Olofsson et al, 2011 (dotted line); and A thresholds ipsi vs contra from Stenfelt, 2012 (dash-dotted line). (c) Comparison of ‘B thresholds, ipsi vs contra’ (solid line), ‘B ECSP, ipsi vs contra’ (dashed line), and B promontory acceleration shift ipsi vs contra derived from Eeg-Olofsson et al, 2011: dotted line.
367
position. In this study, the ipsilateral hearing thresholds were better and the ECSP was higher when stimulation was closer to the cochlea. The previous LDV studies did not reveal the sensitivity difference for frequencies below 500 Hz. The reason might be that the LDV only measured in one direction. On the other hand, at higher frequencies this direction is either equal to or is dominating the cochlear response when stimulating at the BCI position. According to Stenfelt et al (2005), vibration measurements of the cochlea in one direction are appropriate for low frequencies (below 3000 Hz), while three-dimensional vibration data give better estimates at higher frequencies. Given that assumption, to measure the vibration in one direction would not be the cause of the different response at low frequencies. Other reasons may be possible, but not known to the authors at present. One possible explanation for the improvement in sensitivity by changing from position A to position B is the difference in mechanical point impedance of the skull bone at those positions. Position A is located in the thin parietal bone, while position B is in the thicker temporal bone that also comprises the cochlear promontory. There are differences in mechanical point impedance curves on cadavers between position A and B down to approximately 250 Hz (position 1 and 4 in Eeg-Olofsson et al (2008)), where the point impedance in position B is higher. This may indicate that position B can be regarded as a more rigid and direct drive of the cochlea, while position A has a more compliant structure that will attenuate the vibration transmission to the cochlea especially at higher frequencies. The test-retests (Figure 6) showed a large variability for some frequencies in the ipsilateral threshold shift, while the variability was a bit lower in the ipsilateral ECSP shift. The large variability was expected since especially subjective hearing thresholds are known to have a large intrasubject variability, and here, the threshold shift is considered. The fact that the objective ECSP shift has lower intrasubject variability implies that the objective ECSP is a more accurate measure than the subjective hearing threshold. The contralateral hearing threshold and ECSP shifts between position B and A were lower than the ipsilateral shifts (see Figure 7). The hearing was slightly improved by changing the stimulation position from A to B contralaterally for most frequencies up to 4000 Hz, significantly at 1000 – 3000 Hz for thresholds, and for 250 – 500 and 1000 – 2000 Hz for ECSP. This finding differs from the results in Eeg-Olofsson et al (2011) where a slight decrease in contralateral cochlear stimulation was shown when stimulating at position B compared to position A. It might be explained by other still unknown factors, such as different routes of BC sound transmission and differences in mechanical point impedance. In Håkansson et al (2010), an approximately 10 dB decrease in the LDV response was shown contralaterally by changing from a BAHA device at the BAHA position to a BCI device at the BCI position. This finding is most likely due to the fact that Håkansson et al (2010) compared whole systems, including the inductive link for the BCI device. Another aspect was whether the hearing threshold shifts and the ECSP shifts between position B and A were similar; according to Figure 7 and the performed t-test they are, except at a few frequencies (750 and 4000 Hz ipsilaterally, and 1000 Hz contralaterally). Hence, to estimate the hearing ability between these two stimulation positions objectively, ECSP measurements may be used. Of course, the use of ECSP measurements is only possible for patients with normal outer ears, where it is possible to insert a probe tube microphone. To further investigate the similarities between hearing
Int J Audiol Downloaded from informahealthcare.com by Ball State University on 11/20/14 For personal use only.
368
S. Reinfeldt et al.
threshold shifts and ECSP shift, a correlation analysis was performed for all shifts, and the Wilcoxon signed rank test was used over all frequencies for each subject. The results showed that caution should be taken when measuring on individual level. The similarity between hearing threshold and ECSP shifts was also shown in Reinfeldt et al (2013) for most of the tested frequencies; relative hearing thresholds can be estimated by relative ECSP at most frequencies. In Reinfeldt et al (2013), BC hearing thresholds and ECSP measurements were obtained for stimulation at the ipsilateral and contralateral mastoids and at the center of the forehead, both for open and occluded ear-canals. Occluded earcanals were used in both this study and in Reinfeldt et al (2013), however the occlusion was not equally deep. Reinfeldt et al (2013) inserted the ear-plug 18 mm from tragus, which gave an occlusion effect. That, in turn, gave an influence on the BC hearing and even dominated the BC hearing at low frequencies. In this study, the ear-plug (and probe tube) was inserted as deep as possible to minimize the occlusion effect. Another difference is that in this investigation, the comparison for the ipsilateral hearing threshold and ECSP shifts is between nearby stimulation positions, while in Reinfeldt et al (2013) the stimulation positions were at longer distances from each other. The method of using ECSP measurements could be an easy way of showing the expected gain of a BCI in a patient, provided that the patient has a normal outer ear. In case the ECSP is not suitable, the nasal sound pressure could be an alternative. This will be investigated and presented in a future paper.
Ipsilateral vs contralateral response, stimulating in position A and B Figure 9 shows that the hearing threshold and ECSP shifts between ipsilateral and contralateral responses (transcranial attenuation) for position A were around zero and increasing for higher frequencies, and that they were higher for position B than for position A (maximum 10 dB higher). The t-test comparisons between hearing threshold and ECSP shifts showed only two statistical significant differences (position A: 4000 Hz, and position B: 125 Hz) with p ⬍ 0.05. Hence, it could be considered justified to use relative ECSP as an objective measure of relative thresholds between ipsilateral and contralateral sides. The estimated shifts between ipsilateral and contralateral responses both for hearing thresholds and ECSP were higher for position B than for position A. This implies that the sound transmitted to the other ear is relatively lower if stimulation is at position B. If a patient would wear bilateral bone conduction devices, the relative difference in cochlear stimulation would be larger when using the BCI position than when using the BAHA position. Hence, it would imply an improved binaural hearing.
is one reason for the higher difference in promontory acceleration seen in Figure 10 (a) in the high frequency range. Another reason for this difference is that position B in Håkansson et al (2010) is closer to the cochlea (deeper in the bone) than the current position B. In this study, the same BC transducer was used transcutaneously for both stimulation positions, and hence the hearing threshold and ECSP shifts should be compensated for the difference in linear frequency response between the devices and the difference between transcutaneous and direct BC stimulation with screw or flat surface attachment, in order to make a proper comparison to the results in Håkansson et al (2010). Then the high frequency boost of the BCI would be included, and the sensitivity difference obtained would be higher around 4000 – 5000 Hz than in the result of this study. The comparison in Figure 10 (a) shows that the three types of measurements (thresholds, ECSP, promontory acceleration) gave approximately the same estimated shift above 500 Hz, except for somewhat lower shift in Eeg-Olofsson et al (2008). Hence, hearing threshold shifts can be estimated by ECSP shifts at almost all frequencies and by acceleration shifts of the cochlear promontory above 500 Hz. Parts of this conclusion was also made in Reinfeldt et al (2013), where relative hearing thresholds and relative ECSP were compared for stimulation at the forehead and both mastoids. It was found that for most frequencies, relative hearing thresholds could be estimated by relative ECSP. In Figure 10 (b) a comparison is made between ‘A thresholds, ipsi vs contra’; ‘A ECSP, ipsi vs contra’; A promontory acceleration shift ipsi vs contra (inverted TT from position 1) from Eeg-Olofsson et al (2011); and transcranial attenuation boneconduction hearing-aid position thresholds in Stenfelt (2012). All methods show that the ipsilateral and contralateral responses are almost equal, or differ by a maximum 10 dB. Most methods show a few dB higher (or better) responses at the ipsilateral ear for higher frequencies. Stenfelt (2012) made a comparison between his perceptual result and promontory acceleration results from previous studies by Stenfelt and Goode (2005) and Eeg-Olofsson et al (2011), and found similar results at all frequencies except above 6000 Hz. In Figure 10 (c), a comparison is made between ‘B thresholds, ipsi vs contra’; ‘B ECSP, ipsi vs contra’; and B promontory acceleration shift ipsi vs contra (inverted TT from position 4 from Eeg-Olofsson et al (2011). Except for shifts at frequencies below 500 Hz and lower attenuation at around 1500 – 2000 Hz for Eeg-Olofsson et al (2011), a similar shift between ipsilateral and contralateral ears can be seen for all methods. This further strengthens the conclusion that relative BC hearing thresholds can be estimated by relative ECSP and by relative promontory acceleration for most frequencies.
Conclusions Comparison with previous studies In Figure 10 (a) a comparison is made for ‘Ipsi thresholds, B vs A’, ‘Ipsi ECSP, B vs A’, Ipsi promontory acceleration shift position 4 vs position 1 (Eeg-Olofsson et al, 2008), and Ipsi promontory acceleration shift position B vs position A (Håkansson et al, 2010). In the study by Håkansson et al (2010), a Baha Classic was used at position A and a modified BEST was used at position B. A difference between the Baha Classic and the modified BEST is that the modified BEST has a high frequency boost at around 4000 – 5000 Hz. This feature will be used in the BCI and
The aim of the study was to compare the transcutaneous hearing sensitivity by measuring hearing thresholds and ear-canal sound pressure (ECSP) when stimulating at the typical BAHA site and when stimulating close to the position for a BCI device with the same transducer at both positions. The hearing sensitivity at the ipsilateral side was 3 – 14 dB higher from the BCI position than from the BAHA position. This result confirmed previous cochlear promontory acceleration measurements by using laser Doppler vibrometry (LDV) (Håkansson et al, 2008, 2010; Eeg-Olofsson et al, 2008, 2011); however, the hearing sensitivity was also found
BC sensitivity – BAHA vs BCI position to be higher for the low frequencies, which was not shown in the LDV measurements. The correlation was low between hearing threshold shifts and ECSP shifts at an individual level, but the Wilcoxon signed rank test for each subject showed no statistical differences between the shifts for most subjects. At a group level, the comparison showed no significant differences for most frequencies, which imply that the objective method of ECSP shift could be used as an estimate of the subjective hearing threshold shift between different stimulation positions.
Int J Audiol Downloaded from informahealthcare.com by Ball State University on 11/20/14 For personal use only.
Declaration if interest: The authors report no declaration of interest. The study was supported by a grant from Vinnova (Swedish Innovation Agency).
References Battista R.A. & Littlefield P.D. 2006. Revision BAHA surgery. Otolaryngol Clin North Am, 39, 801–813. Cochlear Bone Anchored Solutions celebrate 100 000 patients at Osseo 2013 meeting. Newcastle, UK. Dun C.A., Faber H.T., de Wolf M.J., Mylanus E.A., Cremers C.W. et al. 2012. Assessment of more than 1000 implanted percutaneous boneconduction devices: Skin reaction and implant survival. Otol Neurotol, 33, 192–198. Eeg-Olofsson M., Stenfelt S., Tjellström A. & Granström G. 2008. Transmission of bone-conducted sound in the human skull measured by cochlear vibrations. Int J Audiol, 47, 761–769. Eeg-Olofsson M., Stenfelt S. & Granström G. 2011. Implications for contralateral bone-conducted transmission as measured by cochlear vibrations. Otol Neurotol, 32, 192–198. Eeg-Olofsson M., Stenfelt S., Taghavi H., Reinfeldt S., Håkansson B. et al. 2013. Transmission of bone-conducted sound: Correlation between hearing perception and cochlear vibration. Hear Res, 306, 11–20. Eeg-Olofsson M., Håkansson B., Reinfeldt S., Taghavi H., Lund H. et al. 2014. The bone-conduction implant: First implantation, surgical and audiologic aspects. Otol Neurotol, 35, 679–685. Håkansson B. 2003. The balanced electromagnetic separation transducer: A new bone-conduction transducer. J Acoust Soc Am, 113, 818–825. Håkansson B., Eeg-Olofsson M., Reinfeldt S., Stenfelt S. & Granström G. 2008. Percutaneous vs. transcutaneous bone-conduction implant system: A feasibility study on a cadaver head. Otol Neurotol, 29(8), 1132–1139. Håkansson B., Reinfeldt S., Eeg-Olofsson M., Östli P., Taghavi H. et al. 2010. A novel bone conduction implant (BCI): Engineering aspects and pre-clinical studies. Int J Audiol, 49, 203–215. Håkansson B. 2011. The future of bone-conduction hearing devices. In: M. Kompis, M-D Caversaccio. Implantable Bone Conduction Hearing Aids. Basel: Karger Press, pp. 140–152. ISO:226. 2003. Acoustics - Normal equal loudness-level contours. Geneva: International Organization for Standardization. ISO 389-1. 1998. Acoustics - Reference zero for the calibration of audiometric equipment. Part 1: Reference equivalent threshold sound pressure levels for pure tones and supra-aural earphones. Geneva: International Organization for Standardization.
369
ISO:389-3.1994. Reference zero for the calibration of audiometric equipment. Part 3: Reference equivalent threshold force levels for pure tones and bone vibrators. Geneva: International Organization for Standardization. ISO:8253-1. 2010. Acoustics - Audiometric test methods. Part 1: Pure-tone air- and bone-conduction audiometry. Geneva: International Organization for Standardization. Reinfeldt S., Stenfelt S., Good T. & Håkansson B. 2007. Examination of bone-conducted transmission from sound field excitation measured by thresholds, ear-canal sound pressure, and skull vibrations. J Acoust Soc Am, 121, 1576–1587. Reinfeldt S., Östli P., Håkansson B. & Stenfelt S. 2010. Hearing one’s own voice during phoneme vocalization: Transmission by air- and boneconduction. J Acoust Soc Am, 128, 75–762. Reinfeldt S., Stenfelt S. & Håkansson B. 2013. Estimation of boneconduction skull transmission by hearing thresholds and ear-canal sound pressure. Hear Res, 299, 19–28. Reyes R.A., Tjellström A. & Granström G. 2000. Evaluation of implant losses and skin reaction around extraoral bone-anchored hearing implants: A 0- to 8-year follow-up. Otolaryngol Head Neck Surg, 122, 272–276. Snik A., Mylanus E., Proops D., Wolfaardt J., Hodgetts W. et al. 2005. Consensus statements on the BAHA system: Where do we stand at present? Ann Otol Rhinol Laryngol Suppl, 195, 2–12. Shirazi M., Marzo S. & Leonetti J. 2006. Perioperative complications with the bone-anchored hearing aid. Otolaryngol Head Neck Surg, 134, 236–9. Stenfelt S., Håkansson B. & Tjellström A. 2000. Vibration characteristics of bone-conducted sound in vitro. J Acoust Soc Am, 107, 422–431. Stenfelt S. & Goode R.L. 2005. Transmission properties of boneconducted sound: Measurements in cadaver heads. J Acoust Soc Am, 118, 2373–2391. Stenfelt S. & Reinfeldt S. 2007. A model of the occlusion effect with boneconducted stimulation. Int J Audiol, 46, 595–608. Taghavi H., Håkansson B. & Reinfeldt S. 2012a. A novel bone-conduction implant system: Analog radio frequency data and power link design. In: Proceedings of the 9th IASTED international conference on biomedical engineering, Innsbruck, Austria, 327–335. Taghavi H., Håkansson B. & Reinfeldt S. 2012b. Analysis and design of RF power and data link using amplitude modulation of class-E for a novel bone-conduction implant. IEEE Trans Biomed Eng, 59, 3050–3059. Taghavi H., Håkansson B., Reinfeldt S., Eeg-Olofsson M. & Akhshijan S. 2012c. Feedback analysis in percutaneous bone-conduction device and bone-conduction implant on a dry cranium. Otol Neurotol, 33, 413–420. Tjellström A. & Granström G. 1994. Long-term follow-up with the bone anchored hearing aid: A review of the first 100 patients between 1977 and 1985. ENT J, 73, 21–23. Tjellström A. & Håkansson B. 1995. The bone-anchored hearing aid: Design principles, indications, and long-term clinical results. Otolaryngologic Clinics of North America, 28, 53–72 Tjellström A. & Stalfors J. 2012. Bone-anchored hearing device surgery: A 3- to 6-year follow-up with life table and worst-case scenario calculation. Otol Neurotol, 33, 891–894. Wazen J., Wycherly B. & Daugherty J. 2011. Complications of bone-anchored hearing devices. In: M. Kompis, M-D. Caversaccio, Implantable Bone Conduction Hearing Aids. Basel: Karger Press, pp. 63–72.
Original Article
Bone conduction hearing sensitivity in normal-hearing subjects: Transcutaneous stimulation at BAHA vs BCI position
Int J Audiol Downloaded from informahealthcare.com by Ball State University on 11/20/14 For personal use only.
Sabine Reinfeldt*, Bo Håkansson*, Hamidreza Taghavi* & Måns Eeg-Olofsson† *Department of Signals and Systems, Chalmers University of Technology, Gothenburg, Sweden, and †Department of Otorhinolaryngology, Head and Neck Surgery, Sahlgrenska University Hospital, The Sahlgrenska Academy, University of Gothenburg, Gothenburg, Sweden
Abstract Objective: Bone conduction (BC) stimulation closer to the cochlea has previously been shown to give higher cochlear promontory acceleration measured by laser Doppler vibrometry (LDV). This study is investigating whether stimulation closer to the cochlea also gives improved hearing sensitivity. Furthermore, the study compares shifts in hearing sensitivity (BC thresholds) and ear-canal sound pressure (ECSP). Design: BC hearing thresholds and ECSP have been measured for stimulation at two positions: the existing bone-anchored hearing aid (BAHA) position, and a new bone conduction implant (BCI) position that is located closer to the cochlea. Study sample: The measurements were made on 20 normal-hearing subjects. Results: Depending on frequency, the ipsilateral hearing threshold was 3–14 dB better, and the ipsilateral ECSP was 2–12 dB higher for the BCI than for the BAHA position, with no significant differences between threshold and ECSP shifts at group level for most frequencies, and individually only for some subjects. Conclusions: It was found that both the objective ECSP and the subjective hearing threshold measurements gave similar improvement as previous LDV measurements for stimulation closer to the cochlea. One exception was that the LDV measurements did not show the improved sensitivity for frequencies below 500 Hz found here.
Key Words: Bone conduction; hearing sensitivity; transcutaneous stimulation; bone conduction implant; bone-anchored hearing aid; threshold shift; ear-canal sound pressure; objective measure; normal-hearing subjects
The bone-anchored hearing aid (BAHA) is a viable alternative for patients with conductive or mixed hearing loss and single-sided deafness, and has reached great success worldwide with more than 100 000 patients treated (Cochlear, 2013). However, the percutaneous solution needs life-long daily care and may in some cases give skin complications or even spontaneous implant losses (Battista et al, 2006; Dun et al, 2012; Reyes et al, 2000; Shirazi et al, 2006; Snik et al, 2005; Tjellström et al, 1994, 2012; Wazen et al, 2011). Patients may also reject a BAHA solution for aesthetic reasons. With the new bone conduction implant (BCI) system, these problems can be avoided. The transducer is implanted under the skin in the temporal bone without use of a percutaneous screw. The BCI system consists of an external audio processor and an implanted unit, named the bridging bone conductor (BBC). To keep the skin and subcutaneous tissue intact, an inductive link is used for the sound and power transmission. The BCI system is described in more details in Håkansson et al (2010) and in Taghavi et al (2012a,b). Another
advantage with the BCI system, as compared to the BAHA, is that feedback is decreased (Taghavi et al, 2012c). The BCI system is now in its first clinical trial with results of the first patient described in Eeg-Olofsson et al (2014). According to laser Doppler vibrometry (LDV) measurements on cadavers (Håkansson et al, 2008; 2010), the acceleration of the ipsilateral cochlear promontory increases when changing from a BAHA transducer at the BAHA position (position A in Figure 1) to a BCI transducer at the suggested BCI position 25 mm behind the posterior border of the bony ear-canal (position B in Figure 1), which is closer to the cochlea and located in the mastoid portion of the temporal bone. It should be noted that the LDV only measures the vibration level in one direction, i.e. in line with the ear-canal through which the light beam was pointed, at a reflector at the cochlear promontory. As the cochlea is sensitive to vibrations in all directions this can be a limiting factor when estimating the real hearing sensitivity from the LDV. However, as the stimulation direction in a bone-conduction
Correspondence: Sabine Reinfeldt, Department of Signals and Systems, Chalmers University of Technology, SE-412 96 Gothenburg, Sweden. E-mail: sabine.reinfeldt@ chalmers.se (Received 13 May 2013; accepted 1 January 2014 ) ISSN 1499-2027 print/ISSN 1708-8186 online © 2014 British Society of Audiology, International Society of Audiology, and Nordic Audiological Society DOI: 10.3109/14992027.2014.880813
BC sensitivity – BAHA vs BCI position
Int J Audiol Downloaded from informahealthcare.com by Ball State University on 11/20/14 For personal use only.
Abbreviations AC BAHA BBC BC BCI BEST contra CT ECSP HL ipsi LDV SD TM TT vs
Air conduction Bone-anchored hearing aid Bridging bone conductor Bone conduction Bone conduction implant Balanced electromagnetic separation transducer contralateral Computed tomography Ear-canal sound pressure Hearing level ipsilateral Laser Doppler vibrometry Standard deviation Tympanic membrane Transcranial transmission versus
hearing device is perpendicular to the skull bone and approximately in line with the measurement direction along the ear-canal, this direction might be the most important contributor to the real hearing response using a BAHA or a BCI (Stenfelt et al, 2000, 2005; Reinfeldt et al, 2013). It was also shown in the cadaver studies that the acceleration of the contralateral cochlea with full systems in a sound field was lower with the BCI than with the BAHA. In summary, these LDV results may indicate that the hearing sensitivity increases for ipsilateral stimulation and is slightly lower for contralateral stimulation by using a BCI as compared to using a BAHA. Eeg-Olofsson et al (2008; 2011) investigated the promontory velocity from LDV measurements by stimulating at eight positions
361
between the BAHA position and a position close to the ear-canal, using the same transducer at all positions. The conclusions were that the ipsilateral promontory velocity was higher for closer stimulation, and that the transcranial transmission (TT) differences between the positions were large at positions close to the cochlea and at high frequencies. So far, the LDV results from the cadaver studies have not yet been verified against psychoacoustical measurements to confirm that the hearing sensitivity really is increased when stimulating closer to the cochlea. However, one study has been performed on BAHA patients with a unilateral middle-ear common cavity (Eeg-Olofsson et al, 2013). LDV measurements were compared with BC hearing thresholds, showing similar trends for vibration and threshold measurements on a group level, but low correlation at the individual level. LDV measurements are usually not feasible to perform on live humans since the promontory is not visible from outside. Therefore another objective measurement method would be desirable as a substitute for hearing thresholds, e.g. when examining the performance of a bone-conduction (BC) device. It has previously been shown that relative BC hearing thresholds are similar to relative ear-canal sound pressure (ECSP) at most frequencies (Reinfeldt et al, 2013). That study was done in normal-hearing subjects with open ear-canals and for stimulation positions far from each other. The ECSP reflects the sound radiated into the ear-canal from the surrounding structures. Since the ear-canal is close to the cochlea, it has been hypothesized that a change of the ECSP level is correlated to a BC sensitivity change (Reinfeldt et al, 2013). A simple experiment can be made to investigate if the BC hearing sensitivity increases by moving the stimulation position closer to the cochlea, or more specifically from the BAHA to the BCI position. By pressing a test rod attached to a BAHA device, e.g. Baha Intenso™ and test rod from Cochlear® (Cochlear Bone-Anchored Solutions, Sweden), against the skin at the BAHA and BCI positions (the test rod should not touch the pinna) with closed ear-canals and listen to some music, you will clearly hear a difference in sound level. Generally, a louder sound is heard when the test rod is pressed at the BCI position (approximately 25 mm behind the center of the ear-canal) as compared to the BAHA position (55 mm behind the ear-canal). This result is in line with the previous LDV measurements on cadavers and the method may be used to verify hearing sensitivity differences between the two positions. A first aim of this study is to investigate the BC hearing sensitivity difference by measuring BC tone thresholds for stimulation at the BAHA position and a position close to the suggested BCI position on normal-hearing subjects. A second aim is to measure the ECSP for the same stimulation positions and to compare the two measurement methods. The results will also be compared to previous LDV measurements on cadavers in Håkansson et al (2008; 2010), and Eeg-Olofsson et al (2008; 2011) to verify the validity of these LDV measurements regarding hearing sensitivity differences.
Materials and Methods Subjects Figure 1. Position A (BAHA position) is 55 mm from the entrance of the ear-canal at the line extending from the lateral margin of the eye to the upper border of the pinna; and position B (BCI position) is on a line between the ear-canal and position A, about 25 mm from the ear-canal entrance as close as possible behind the pinna.
All measurements were performed on 20 voluntary subjects (10 male and 10 female, ages 23 to 44 years, with an average of 30 years) with otologically-normal ears (ISO 226, 2003). Their air-conduction (AC) hearing thresholds were no worse than 20 dB HL for frequencies 125 to 8000 Hz and the interaural sensitivity difference was maximum 15 dB at any frequency
362
S. Reinfeldt et al.
measured (125, 250, 500, 750, 1000, 1500, 2000, 3000, 4000, 6000, and 8000 Hz). One test subject was included in the study despite that his right ear hearing threshold was 25 dB HL at 8000 Hz (that frequency was excluded from the analysis).
Int J Audiol Downloaded from informahealthcare.com by Ball State University on 11/20/14 For personal use only.
Measurement setup A sound insulated room of 16 m3 was used for all measurements. To determine the subjects’ baseline hearing tone thresholds for inclusion in the study, the Hughson-Westlake procedure according to ISO 8253-1 (2010) was used with TDH39 headphones and a digital audiometer (AC40; Interacoustics A/S, Assens, Denmark). For BC stimulation, a modified BEST transducer (Håkansson, 2003), was attached to the skin over the skull bone with a steel spring headband. In order to reduce the skin dampening effect, a higher static pressure (6–10 N) and a smaller contact surface area (38 mm2) than normal in BC audiometry (175 mm2 and 5.4 N according to ISO 389-3 (1994)) was used. The adapter has a tip diameter of 7 mm (effective diameter of 6 mm as there is a radius of 0.5 mm in the edge towards the skin; the real BCI attachment area to the skull bone is 6 mm), and a height of 4 mm as shown in Figure 2. The transducer was pressed against the skin at the two positions illustrated in Figure 1 and driven by a signal analyzer (Agilent 35670A, Agilent Technologies Inc., USA) via a 10 Ω current limiting resistor. Position A was approximately 55 mm from the entrance of the ear-canal at a line extending from the lateral margin of the eye to the upper border of the pinna (position A is where a BAHA normally is positioned). Position B is on a line between the ear-canal and position A, about 25 mm from the center of the ear-canal entrance, closest possible behind the pinna and as close to the suggested position for a BCI as possible. The BC hearing thresholds were measured using pure tones according to the Békésy procedure in the AC40 audiometer. A pulsed sequence (instead of constant) was used to make it easier for the patient to distinguish the tone from the masking noise. Frequencies were 125, 250, 500, 750, 1000, 1500, 2000, 3000, 4000, 6000, and 8000 Hz, and the step level was 1 dB. When measuring BC hearing thresholds, masking must be used to ensure perception in the test-ear and avoid false thresholds from the non-test ear. In this study, the masking noise was a broadband noise spectrally designed to be of equal level above the normal hearing thresholds (0 dB HL) for all 1/3-octave frequency bands
(see Reinfeldt et al (2010) for more details). It was fed from a CD player via a power amplifier (Sony TA-N220; Sony Corp., Tokyo, Japan) to an insert earphone (Etymotic Research ER-2; Etymotic Research Inc., USA). The foam part of the insert earphone was modified to avoid occlusion effect (Stenfelt & Reinfeldt, 2007). The sound pressure in the ear-canals (ECSP) was measured by a microphone (Knowles EM3046; Knowles, USA), originally used in BAHA devices. This microphone was chosen instead of the Etymotic ER-7 probe tube microphone as it provided a lower noise floor. The same probe tube was used as in the Etymotic ER-7 system. The inlet to the EM3046 fits perfectly to this probe tube with no leakage that could have altered the measurement results. The microphone was connected via a battery operated low noise linear preamplifier (Gennum LC506; Semtech Corp., USA), which provided the signal to the Agilent signal analyser in order to measure the frequency response between the ECSP output and the input voltage to the BC transducer. Measurements of both ipsilateral and contralateral ECSP were performed. To avoid both AC sound transmission and occlusion effect when measuring the ECSP, an ear-plug was deeply inserted in the ear-canal. The AC sound was attenuated with 20 to 40 dB, depending on frequency, by the ear-plug (Reinfeldt et al, 2007) and could therefore be neglected. Stenfelt and Reinfeldt (2007) showed that the deeper an ear-plug is inserted, the less occlusion effect is obtained. For the ECSP measurements, the length of the ear-canal was first determined by entering a probe tube until it touched the tympanic membrane (TM). Then, the probe tube was inserted into an ear-plug (E-A-R Classic; 3M Corp., USA) by using a catheter; it was tightly sealed by the foam of the ear-plug. The ear-plug was then placed in the ear-canal with the probe tube tip approximately 3 mm from the TM and the end of the ear-plug close to the probe tube tip. A contact between the transducer housing and the pinna could give a false ECSP caused by vibrations transmitted by soft tissues and the ear-plug into the remaining open part of the ear-canal. To place the BC transducer at position B without this contact, the pinna was folded by using a wide sport tape (Leukoplast).
Measurements BC hearing thresholds were obtained ipsi- and contralaterally for stimulation at positions A and B. Correspondingly, ECSP were measured for the same conditions. BC stimulation was always done on the same side of the head (left), while ECSP measurements and hearing thresholds were performed in both ears. In total eight measurements were performed; they are presented in Figure 3 and listed as follows: • • • •
Figure 2. BEST with adapter having a tip diameter of 7 mm (effective diameter of 6 mm as there is a radius of 0.5 mm in the edge towards the skin) and height of 4 mm.
Threshold ipsi A and B ECSP ipsi A and B Threshold contra A and B ECSP contra A and B
The measurements were varied in ten different orders for the test subjects to avoid order effects. The orders were constructed to give as few changes as possible. The masking ear-phone and the ear-plug were switched between hearing thresholds and ECSP measurements. Since the static pressure towards the skin of the BC transducer was relatively high, the transducer was temporarily removed after each measurement and the stimulation position was switched for the next measurement. It was important to reattach the transducer at the same
BC sensitivity – BAHA vs BCI position
(a)
(a) A
A
B
B
ECSP ipsi A and B
Masking for thresholds
‘Contra ECSP, B vs A’
‘Ipsi ECSP, B vs A’ ‘Ipsi thresholds, B vs A’
Threshold ipsi A and B
(b)
(b)
A
‘Contra thresholds, B vs A’
A B
B
Int J Audiol Downloaded from informahealthcare.com by Ball State University on 11/20/14 For personal use only.
363
Masking for thresholds
ECSP contra A and B
‘A thresholds, ipsi vs contra’ ‘B thresholds, ipsi vs contra’ ‘A ECSP, ipsi vs contra’ ‘B ECSP, ipsi vs contra’
Threshold contra A and B
Figure 3. (a) Ipsilateral hearing thresholds and ECSP for stimulation positions A and B. (b) Contralateral hearing thresholds and ECSP for stimulation positions A and B. Solid lines are thresholds and dashed lines are ECSP measurements.
Figure 4. (a) Calculations of hearing threshold and ECSP shifts for position B vs position A. (b) Calculations of hearing threshold and ECSP shifts for contralateral vs ipsilateral response.
positions and therefore the positions were marked with a pen before the measurements started.
calibrated using a Brüel & Kjær 4230 sound level calibrator. The probe tube connected to the microphone was then positioned as close as possible to the reference microphone in a Brüel & Kjær anechoic test chamber type 4222 with the assumption that the microphones measured the same sound pressure. By this procedure, a frequency-dependent microphone sensitivity was determined and applied to the microphone output voltage to convert to sound pressure. The output from the BC transducer was calibrated according to ISO 389-3 (1994) using a Brüel & Kjær 4930 artificial mastoid, and the dynamic output force (in dB relative 1 Newton) from the BC transducer against the pad of Brüel & Kjær 4930 was related to the output of the artificial mastoid by using a pad correction. The audiometer (AC40) was calibrated according to ISO 389-1 (1998) by the manufacturer for threshold measurements with earphones. Since all measurements were comparative, there was no need for calibration, except for the baseline hearing thresholds, described above. However, the calibrations were done before and after all measurements to ensure that there were no changes in the performance of the equipment during the study period.
Calculations The comparisons between different measurements are shown in Figure 4. Positive shifts for position B vs A means better hearing or higher ECSP for stimulation at position B than at position A. Positive shifts for ipsilateral vs contralateral response means better hearing or higher ECSP ipsilaterally than contralaterally. For all measures, the averages of all subjects were calculated together with the standard deviation (SD). To assess whether the hearing threshold and ECSP shifts for the stimulation positions were separated from zero, a Wilcoxon signed rank test was performed between the positions both for thresholds and ECSP. The null-hypothesis was that there were no differences between the positions. A two-tailed paired t-test was used to test similarities between the hearing threshold and ECSP shifts with the null-hypothesis that there were no differences between the methods. The test was performed only at frequencies where thresholds were measured and where the microphone was above the noise floor (nine frequencies between 125 and 4000 Hz).To correct for the probability that a significant result occurs purely by chance when performing independent t-tests at each of the nine tested frequencies, a Sidak correction was applied. Pearson correlation was used to calculate the correlation between hearing threshold shift and ECSP shift at individual level, and also Wilcoxon signed rank test was used to compare ipsilateral threshold shift and ECSP shift B vs A for each individual over all frequencies.
Calibrations The microphone was calibrated using a reference microphone (Brüel & Kjær 4134; Brüel & Kjær Sound & Vibration Measurement A/S, Denmark). The sensitivity of the reference microphone was
Results Ipsilateral and contralateral response, stimulating in position B vs A To illustrate the individual differences in the hearing threshold and ECSP shifts for position B vs A, the shifts for each individual, averages, and SDs are shown in Figure 5. ‘Ipsi thresholds, B vs A’ is shown in Figure 5 (a). The individual differences were large, but generally there is an improved sensitivity at position B relative position A. Also the ‘Ipsi ECSP, B vs A’, in Figure 5 (b), gave a similar pattern. Figure 5 (c) and (d), show the threshold and ECSP shifts for the contralateral side. Here the general trend was a slight
364
S. Reinfeldt et al. 40
40
(c)
30
30
20
20
10 0
10 0
−10
−10
−20
−20
−30
125
250
500
1000
2000
4000
−30
8000
125
250
500 1000 2000 Frequency [Hz]
4000
8000
250
500 1000 2000 Frequency [Hz]
4000
8000
Frequency [Hz] 40
40
(b)
(d)
30
30
20
20 ECSP shift [dB]
ECSP shift [dB]
Int J Audiol Downloaded from informahealthcare.com by Ball State University on 11/20/14 For personal use only.
Threshold shift[dB]
Threshold shift [dB]
(a)
10 0
10 0
−10
−10
−20
−20
−30
125
250
500
1000
2000
4000
8000
Frequency [Hz]
−30
125
Figure 5. Individual (thin solid) and average (thick solid) curves together with SD (bars) for (a) ‘Ipsi thresholds, B vs A’, (b) ‘Ipsi ECSP, B vs A’, (c) ‘Contra thresholds, B vs A’, and (d) ‘Contra ECSP, B vs A’. A positive value means that position B gives a higher sensation than position A.
improvement for position B. The SD’s for all shifts were between 4 and 13 dB. Test-retests were performed and in Figure 6, the hearing threshold shifts (a) and ECSP shifts (b) are shown for one subject measured three times. The measurement and stimulation equipment were removed and reattached between each measurement. The variability is expected in both hearing thresholds and ECSP shifts, e.g. in Reinfeldt et al (2013), the standard deviations for these measures were in general between 4 and 12 dB. In this study, the threshold shifts show a dynamic range of 20 dB at some frequencies and almost no variability at other frequencies. The ECSP shifts had lower variability than the threshold shifts. The average threshold and ECSP shifts between position B and A are summarized in Figure 7, both for the ipsilateral and contralateral sides. Positive values demonstrate better hearing or higher ECSP for stimulation at position B than at position A. The average ‘Ipsi thresholds, B vs A’ were positive for all frequencies. The hearing thresholds showed a shift of between 3 and 14 dB in favor of position B. Since the microphone (together with preamplifier) signal was in its noise floor above 4000 Hz, the ECSP shifts have been disregarded at these high frequencies. ‘Ipsi ECSP, B vs A’ was between 2 and 12 dB. The contralateral shifts were lower than the ipsilateral ones, both for threshold and ECSP shifts, but still mainly positive. ‘Contra
thresholds, B vs A’ were between ⫺3 and 6 dB, and ‘Contra ECSP, B vs A’ was between ⫺2 and 11 dB. To investigate if there was a statistical support for improved hearing when moving the stimulation from position A to position B, a Wilcoxon signed rank test was performed ipsilaterally both for thresholds and ECSP. The test showed that thresholds vary with position for α ⫽ 0.05 at all frequencies 125 – 8000 Hz except at 500 Hz, and that ECSP varies with position for α ⫽ 0.05 at all frequencies 125 – 4000 Hz except at 4000 Hz. Hence, it is a regarded as a statistically significant improvement to move the stimulation from position A to position B. The corresponding Wilcoxon signed rank test for contralateral responses showed that thresholds vary with position for α ⫽ 0.05 at 1000 – 3000 Hz, and ECSP at 250 – 500 and 1000 – 2000 Hz. To compare the hearing threshold shifts with the ECSP shifts, a two-tailed paired t-test was performed. The null-hypothesis that there are no differences between the hearing threshold and ECSP shifts for stimulation in position B vs A was confirmed (nonsignificant differences) at most frequencies. The only differences found were ipsilaterally at 750 and 4000 Hz, and contralaterally at 1000 Hz with p ⬍ 0.05. At the individual level, the correlation between the methods was low with an average correlation coefficient
BC sensitivity – BAHA vs BCI position
365
20
40
‘Ipsi thresholds, B vs A’ ‘Ipsi ECSP, B vs A’ ‘Contra thresholds, B vs A’ ‘Contra ECSP, B vs A’
(a) 30 15
B vs A [dB]
Threshold shift [dB]
20 10 0
10
5
−10 0
−20
125
250
500 1000 2000 Frequency [Hz]
4000
8000
(b) 30 20 10 0 −10 −20 −30
125
−5
125
250
500
1000 2000 Frequency [Hz]
4000
8000
Figure 7. Hearing threshold and ECSP shifts, stimulating in position B vs A. Solid line is ‘Ipsi thresholds, B vs A’; dashed line is ‘Ipsi ECSP, B vs A’; dotted line is ‘Contra thresholds, B vs A’; and dashdotted line is ‘Contra ECSP, B vs A’. A positive value means that position B gives a higher sensation than position A.
40
ECSP shift [dB]
Int J Audiol Downloaded from informahealthcare.com by Ball State University on 11/20/14 For personal use only.
−30
250
500 1000 2000 Frequency [Hz]
4000
8000
Figure 6. Test-retest on one subject for (a) ‘Ipsi thresholds, B vs A’, and (b) ‘Ipsi ECSP, B vs A’. A positive value means that position B gives a higher sensation than position A. of 0.25 for ipsilateral comparison and 0.41 for contralateral comparison. However, a Wilcoxon signed rank test for each individual and ipsilateral comparison showed that the null hypothesis could be rejected for 18 out of the 20 subjects with α ⫽ 0.05, hence showing that no statistical differences could be shown between hearing threshold shifts and ECSP shift for 18 of 20 subjects. For the contralateral comparison, no statistical differences could be shown between hearing threshold shifts and ECSP shifts for 16 of 20 subjects.
Ipsilateral vs contralateral response, stimulating in position A and B Figure 8 shows the individual differences in the hearing threshold and ECSP shifts between ipsilateral and contralateral responses (also termed transcranial attenuation) with the shifts for each individual, averages, and SD. Positive values mean higher hearing sensitivity or higher ECSP level on the ipsilateral than on the contralateral side. ‘A thresholds, ipsi vs contra’ are shown in Figure 8 (a), and ‘A ECSP, ipsi vs contra’ in Figure 8 (b). They were around 0 dB at low frequencies and a few dB’s positive at higher frequencies. ‘B thresholds, ipsi vs contra’ and ‘B ECSP, ipsi vs contra’ are shown in Figure 8 (c) and (d), respectively, showing higher shifts than for position A.
SDs of all measures were between 6 and 12 dB. The individual differences were large, but generally stimulation at position B gave higher shifts than position A, and hence lower hearing sensitivity and lower ECSP level contralaterally than ipsilaterally. In Figure 9, the averages of the hearing threshold and ECSP shifts between ipsilateral and contralateral measurements are shown for position A and for position B. ‘A thresholds, ipsi vs contra’ was low, between ⫺ 3 and 8 dB. ‘A ECSP, ipsi vs contra’ was between ⫺ 3 and 10 dB. For position B, the difference between ipsilateral and contralateral measurements was higher: between 2 and 14 dB for ‘B thresholds, ipsi vs contra’, and between 3 and 10 dB for ‘B ECSP, ipsi vs contra’. To statistically investigate if shifting between ipsilateral and contralateral stimulation gives a difference in hearing, the Wilcoxon signed rank test was used. The null hypothesis that there were no differences between the sides was rejected at 2000 – 3000 and 8000 Hz for thresholds at position A, and at 1000 and 3000 – 4000 Hz for ECSP at position A. At position B, the null hypothesis was rejected at 250 and 750 – 6000 Hz for thresholds, and at 125 – 250 and 750 – 4000 Hz for ECSP. Hence, the frequency ranges with significant differences were wider for position B. The t-test for the transcranial measurements indicated that the hearing threshold and ECSP shifts are similar. There were non-significant differences for all frequencies except for position A at 4000 Hz and for position B at 125 Hz, both with p ⬍ 0.05. At the individual level, the correlation between the methods was low with an average correlation coefficient of 0.29 for position A and 0.23 for position B. When performing the Wilcoxon signed rank test for each individual, there was significant difference with α ⫽ 0.05 for 17 of the 20 subjects in position A, and for 14 of the 20 subjects in position B.
Discussion Despite general large individual variability, there are certain trends identified among the results. First of all, position B gives a higher hearing sensitivity than position A for ipsilateral stimulation, which may be of significant importance for the rehabilitation with a BCI device. Secondly, it seems that the ECSP can be used as an
S. Reinfeldt et al.
366
50
50 40
(a)
40 30
20 10 0 −10
10 0 −10 −20
−30
−30
40
125
250
500 1000 2000 Frequency [Hz]
4000
−40
8000
40
30
30
20
20
10 0
250
500 1000 2000 Frequency [Hz]
4000
8000
250
500 1000 2000 Frequency [Hz]
4000
8000
0 −10
−20
−20
−30
−30 125
250
500 1000 2000 Frequency [Hz]
4000
8000
(d)
10
−10
−40
125
50
(b)
ECSP shift [dB]
ECSP shift [dB]
20
−20
50
−40
125
Figure 8. Individual (thin solid) and average (thick solid) curves together with SD (bars) for (a) ‘A thresholds, ipsi vs contra’, (b) ‘A ECSP, ipsi vs contra’, (c) ‘B thresholds, ipsi vs contra’, and (d) ‘B ECSP, ipsi vs contra’. A positive dB value means that the ipsilateral side is more sensitive than the contralateral side, i.e. it is a measure of transcranial attenuation.
14 12 10 ipsi vs contra [dB]
Int J Audiol Downloaded from informahealthcare.com by Ball State University on 11/20/14 For personal use only.
Threshold shift [dB]
Threshold shift [dB]
30
−40
(c)
8 6 4 2 0 −2
‘A thresholds, ipsi vs contra’ ‘A ECSP, ipsi vs contra’ ‘B thresholds, ipsi vs contra’ ‘B ECSP, ipsi vs contra’
−4 −6 125
250
500
1000 2000 Frequency [Hz]
4000
8000
Figure 9. Hearing threshold and ECSP shifts, ipsilateral vs contralateral response, stimulating in position A and B. Solid line is ‘A thresholds, ipsi vs contra’; dashed line is ‘A ECSP, ipsi vs contra’; dotted line is ‘B thresholds, ipsi vs contra’; and dash-dotted line is ‘B ECSP, ipsi vs contra’. A positive dB value means that the ipsilateral side is more sensitive than the contralateral side, i.e. it is a measure of transcranial attenuation.
objective measure instead of the subjective BC hearing thresholds, which may be of importance in situations where thresholds are not suitable (e.g. patient anaesthetized), or in long term investigations requiring lower variability (e.g. follow up of the BCI implant’s performance). However, an additional correlation analysis showed that individual ECSP and threshold shifts do not correlate in the same way as the average shifts do. On the other hand, Wilcoxon signed rank tests for each individual over all frequencies showed that the threshold shift and ECSP shift for a majority of the subjects were not significantly different. For ipsilateral response, B vs A, the threshold and ECSP shifts only differed significantly for 2 of the 20 subjects. In summary, caution should be taken when measuring ECSP instead of thresholds on individual level.
Ipsilateral and contralateral response, stimulating in position B vs A According to Figure 7 and to Wilcoxon’s signed rank tests, better hearing thresholds and higher ECSP were found ipsilaterally for position B compared to position A for all frequencies. This finding supports what was indicated in the studies by Håkansson et al (2008; 2010) and Eeg-Olofsson et al (2008) by using LDV, that the hearing sensitivity is higher for stimulation closer to the cochlea, or more specifically at the BCI position compared to the BAHA
BC sensitivity – BAHA vs BCI position 25
(a)
20
Ipsi, B vs A [dB]
15 10 5 0 ‘Ipsi thresholds, B vs A’ ‘Ipsi ECSP, B vs A’ Eeg−Olofsson et al. (2008) Håkansson et al. (2010)
−5
125
250
500 1000 2000 Frequency [Hz]
4000
8000
10 8
(b)
A, ipsi vs contra [dB]
6 4 2 0 −2 −4 −6
‘A thresholds, ipsi vs contra’ ‘A ECSP, ipsi vs contra’ Eeg−Olofsson et al. (2011) Stenfelt (2012)
−8 −10
20
125
250
(c)
500 1000 2000 Frequency [Hz]
4000
8000
4000
8000
‘B thresholds, ipsi vs contra’ ‘B ECSP, ipsi vs contra’ Eeg−Olofsson et al. (2011)
15 B, ipsi vs contra [dB]
Int J Audiol Downloaded from informahealthcare.com by Ball State University on 11/20/14 For personal use only.
−10
10
5
0
−5
−10
125
250
500 1000 2000 Frequency [Hz]
Figure 10. (a) Comparison between ‘Ipsi thresholds, B vs A’ (solid line); ‘Ipsi ECSP, B vs A’ (dashed line); Ipsi promontory acceleration shift position 4 vs position 1 (Eeg-Olofsson et al, 2008 (dotted line); and Ipsi promontory acceleration shift B vs A (Håkansson et al, 2010) (dash-dotted line). Håkansson et al (2010) used different transducers at position A and B. (b) Comparison between ‘A thresholds, ipsi vs contra’ (solid line); ‘A ECSP, ipsi vs contra’ (dashed line); A promontory acceleration shift ipsi vs contra derived from Eeg-Olofsson et al, 2011 (dotted line); and A thresholds ipsi vs contra from Stenfelt, 2012 (dash-dotted line). (c) Comparison of ‘B thresholds, ipsi vs contra’ (solid line), ‘B ECSP, ipsi vs contra’ (dashed line), and B promontory acceleration shift ipsi vs contra derived from Eeg-Olofsson et al, 2011: dotted line.
367
position. In this study, the ipsilateral hearing thresholds were better and the ECSP was higher when stimulation was closer to the cochlea. The previous LDV studies did not reveal the sensitivity difference for frequencies below 500 Hz. The reason might be that the LDV only measured in one direction. On the other hand, at higher frequencies this direction is either equal to or is dominating the cochlear response when stimulating at the BCI position. According to Stenfelt et al (2005), vibration measurements of the cochlea in one direction are appropriate for low frequencies (below 3000 Hz), while three-dimensional vibration data give better estimates at higher frequencies. Given that assumption, to measure the vibration in one direction would not be the cause of the different response at low frequencies. Other reasons may be possible, but not known to the authors at present. One possible explanation for the improvement in sensitivity by changing from position A to position B is the difference in mechanical point impedance of the skull bone at those positions. Position A is located in the thin parietal bone, while position B is in the thicker temporal bone that also comprises the cochlear promontory. There are differences in mechanical point impedance curves on cadavers between position A and B down to approximately 250 Hz (position 1 and 4 in Eeg-Olofsson et al (2008)), where the point impedance in position B is higher. This may indicate that position B can be regarded as a more rigid and direct drive of the cochlea, while position A has a more compliant structure that will attenuate the vibration transmission to the cochlea especially at higher frequencies. The test-retests (Figure 6) showed a large variability for some frequencies in the ipsilateral threshold shift, while the variability was a bit lower in the ipsilateral ECSP shift. The large variability was expected since especially subjective hearing thresholds are known to have a large intrasubject variability, and here, the threshold shift is considered. The fact that the objective ECSP shift has lower intrasubject variability implies that the objective ECSP is a more accurate measure than the subjective hearing threshold. The contralateral hearing threshold and ECSP shifts between position B and A were lower than the ipsilateral shifts (see Figure 7). The hearing was slightly improved by changing the stimulation position from A to B contralaterally for most frequencies up to 4000 Hz, significantly at 1000 – 3000 Hz for thresholds, and for 250 – 500 and 1000 – 2000 Hz for ECSP. This finding differs from the results in Eeg-Olofsson et al (2011) where a slight decrease in contralateral cochlear stimulation was shown when stimulating at position B compared to position A. It might be explained by other still unknown factors, such as different routes of BC sound transmission and differences in mechanical point impedance. In Håkansson et al (2010), an approximately 10 dB decrease in the LDV response was shown contralaterally by changing from a BAHA device at the BAHA position to a BCI device at the BCI position. This finding is most likely due to the fact that Håkansson et al (2010) compared whole systems, including the inductive link for the BCI device. Another aspect was whether the hearing threshold shifts and the ECSP shifts between position B and A were similar; according to Figure 7 and the performed t-test they are, except at a few frequencies (750 and 4000 Hz ipsilaterally, and 1000 Hz contralaterally). Hence, to estimate the hearing ability between these two stimulation positions objectively, ECSP measurements may be used. Of course, the use of ECSP measurements is only possible for patients with normal outer ears, where it is possible to insert a probe tube microphone. To further investigate the similarities between hearing
Int J Audiol Downloaded from informahealthcare.com by Ball State University on 11/20/14 For personal use only.
368
S. Reinfeldt et al.
threshold shifts and ECSP shift, a correlation analysis was performed for all shifts, and the Wilcoxon signed rank test was used over all frequencies for each subject. The results showed that caution should be taken when measuring on individual level. The similarity between hearing threshold and ECSP shifts was also shown in Reinfeldt et al (2013) for most of the tested frequencies; relative hearing thresholds can be estimated by relative ECSP at most frequencies. In Reinfeldt et al (2013), BC hearing thresholds and ECSP measurements were obtained for stimulation at the ipsilateral and contralateral mastoids and at the center of the forehead, both for open and occluded ear-canals. Occluded earcanals were used in both this study and in Reinfeldt et al (2013), however the occlusion was not equally deep. Reinfeldt et al (2013) inserted the ear-plug 18 mm from tragus, which gave an occlusion effect. That, in turn, gave an influence on the BC hearing and even dominated the BC hearing at low frequencies. In this study, the ear-plug (and probe tube) was inserted as deep as possible to minimize the occlusion effect. Another difference is that in this investigation, the comparison for the ipsilateral hearing threshold and ECSP shifts is between nearby stimulation positions, while in Reinfeldt et al (2013) the stimulation positions were at longer distances from each other. The method of using ECSP measurements could be an easy way of showing the expected gain of a BCI in a patient, provided that the patient has a normal outer ear. In case the ECSP is not suitable, the nasal sound pressure could be an alternative. This will be investigated and presented in a future paper.
Ipsilateral vs contralateral response, stimulating in position A and B Figure 9 shows that the hearing threshold and ECSP shifts between ipsilateral and contralateral responses (transcranial attenuation) for position A were around zero and increasing for higher frequencies, and that they were higher for position B than for position A (maximum 10 dB higher). The t-test comparisons between hearing threshold and ECSP shifts showed only two statistical significant differences (position A: 4000 Hz, and position B: 125 Hz) with p ⬍ 0.05. Hence, it could be considered justified to use relative ECSP as an objective measure of relative thresholds between ipsilateral and contralateral sides. The estimated shifts between ipsilateral and contralateral responses both for hearing thresholds and ECSP were higher for position B than for position A. This implies that the sound transmitted to the other ear is relatively lower if stimulation is at position B. If a patient would wear bilateral bone conduction devices, the relative difference in cochlear stimulation would be larger when using the BCI position than when using the BAHA position. Hence, it would imply an improved binaural hearing.
is one reason for the higher difference in promontory acceleration seen in Figure 10 (a) in the high frequency range. Another reason for this difference is that position B in Håkansson et al (2010) is closer to the cochlea (deeper in the bone) than the current position B. In this study, the same BC transducer was used transcutaneously for both stimulation positions, and hence the hearing threshold and ECSP shifts should be compensated for the difference in linear frequency response between the devices and the difference between transcutaneous and direct BC stimulation with screw or flat surface attachment, in order to make a proper comparison to the results in Håkansson et al (2010). Then the high frequency boost of the BCI would be included, and the sensitivity difference obtained would be higher around 4000 – 5000 Hz than in the result of this study. The comparison in Figure 10 (a) shows that the three types of measurements (thresholds, ECSP, promontory acceleration) gave approximately the same estimated shift above 500 Hz, except for somewhat lower shift in Eeg-Olofsson et al (2008). Hence, hearing threshold shifts can be estimated by ECSP shifts at almost all frequencies and by acceleration shifts of the cochlear promontory above 500 Hz. Parts of this conclusion was also made in Reinfeldt et al (2013), where relative hearing thresholds and relative ECSP were compared for stimulation at the forehead and both mastoids. It was found that for most frequencies, relative hearing thresholds could be estimated by relative ECSP. In Figure 10 (b) a comparison is made between ‘A thresholds, ipsi vs contra’; ‘A ECSP, ipsi vs contra’; A promontory acceleration shift ipsi vs contra (inverted TT from position 1) from Eeg-Olofsson et al (2011); and transcranial attenuation boneconduction hearing-aid position thresholds in Stenfelt (2012). All methods show that the ipsilateral and contralateral responses are almost equal, or differ by a maximum 10 dB. Most methods show a few dB higher (or better) responses at the ipsilateral ear for higher frequencies. Stenfelt (2012) made a comparison between his perceptual result and promontory acceleration results from previous studies by Stenfelt and Goode (2005) and Eeg-Olofsson et al (2011), and found similar results at all frequencies except above 6000 Hz. In Figure 10 (c), a comparison is made between ‘B thresholds, ipsi vs contra’; ‘B ECSP, ipsi vs contra’; and B promontory acceleration shift ipsi vs contra (inverted TT from position 4 from Eeg-Olofsson et al (2011). Except for shifts at frequencies below 500 Hz and lower attenuation at around 1500 – 2000 Hz for Eeg-Olofsson et al (2011), a similar shift between ipsilateral and contralateral ears can be seen for all methods. This further strengthens the conclusion that relative BC hearing thresholds can be estimated by relative ECSP and by relative promontory acceleration for most frequencies.
Conclusions Comparison with previous studies In Figure 10 (a) a comparison is made for ‘Ipsi thresholds, B vs A’, ‘Ipsi ECSP, B vs A’, Ipsi promontory acceleration shift position 4 vs position 1 (Eeg-Olofsson et al, 2008), and Ipsi promontory acceleration shift position B vs position A (Håkansson et al, 2010). In the study by Håkansson et al (2010), a Baha Classic was used at position A and a modified BEST was used at position B. A difference between the Baha Classic and the modified BEST is that the modified BEST has a high frequency boost at around 4000 – 5000 Hz. This feature will be used in the BCI and
The aim of the study was to compare the transcutaneous hearing sensitivity by measuring hearing thresholds and ear-canal sound pressure (ECSP) when stimulating at the typical BAHA site and when stimulating close to the position for a BCI device with the same transducer at both positions. The hearing sensitivity at the ipsilateral side was 3 – 14 dB higher from the BCI position than from the BAHA position. This result confirmed previous cochlear promontory acceleration measurements by using laser Doppler vibrometry (LDV) (Håkansson et al, 2008, 2010; Eeg-Olofsson et al, 2008, 2011); however, the hearing sensitivity was also found
BC sensitivity – BAHA vs BCI position to be higher for the low frequencies, which was not shown in the LDV measurements. The correlation was low between hearing threshold shifts and ECSP shifts at an individual level, but the Wilcoxon signed rank test for each subject showed no statistical differences between the shifts for most subjects. At a group level, the comparison showed no significant differences for most frequencies, which imply that the objective method of ECSP shift could be used as an estimate of the subjective hearing threshold shift between different stimulation positions.
Int J Audiol Downloaded from informahealthcare.com by Ball State University on 11/20/14 For personal use only.
Declaration if interest: The authors report no declaration of interest. The study was supported by a grant from Vinnova (Swedish Innovation Agency).
References Battista R.A. & Littlefield P.D. 2006. Revision BAHA surgery. Otolaryngol Clin North Am, 39, 801–813. Cochlear Bone Anchored Solutions celebrate 100 000 patients at Osseo 2013 meeting. Newcastle, UK. Dun C.A., Faber H.T., de Wolf M.J., Mylanus E.A., Cremers C.W. et al. 2012. Assessment of more than 1000 implanted percutaneous boneconduction devices: Skin reaction and implant survival. Otol Neurotol, 33, 192–198. Eeg-Olofsson M., Stenfelt S., Tjellström A. & Granström G. 2008. Transmission of bone-conducted sound in the human skull measured by cochlear vibrations. Int J Audiol, 47, 761–769. Eeg-Olofsson M., Stenfelt S. & Granström G. 2011. Implications for contralateral bone-conducted transmission as measured by cochlear vibrations. Otol Neurotol, 32, 192–198. Eeg-Olofsson M., Stenfelt S., Taghavi H., Reinfeldt S., Håkansson B. et al. 2013. Transmission of bone-conducted sound: Correlation between hearing perception and cochlear vibration. Hear Res, 306, 11–20. Eeg-Olofsson M., Håkansson B., Reinfeldt S., Taghavi H., Lund H. et al. 2014. The bone-conduction implant: First implantation, surgical and audiologic aspects. Otol Neurotol, 35, 679–685. Håkansson B. 2003. The balanced electromagnetic separation transducer: A new bone-conduction transducer. J Acoust Soc Am, 113, 818–825. Håkansson B., Eeg-Olofsson M., Reinfeldt S., Stenfelt S. & Granström G. 2008. Percutaneous vs. transcutaneous bone-conduction implant system: A feasibility study on a cadaver head. Otol Neurotol, 29(8), 1132–1139. Håkansson B., Reinfeldt S., Eeg-Olofsson M., Östli P., Taghavi H. et al. 2010. A novel bone conduction implant (BCI): Engineering aspects and pre-clinical studies. Int J Audiol, 49, 203–215. Håkansson B. 2011. The future of bone-conduction hearing devices. In: M. Kompis, M-D Caversaccio. Implantable Bone Conduction Hearing Aids. Basel: Karger Press, pp. 140–152. ISO:226. 2003. Acoustics - Normal equal loudness-level contours. Geneva: International Organization for Standardization. ISO 389-1. 1998. Acoustics - Reference zero for the calibration of audiometric equipment. Part 1: Reference equivalent threshold sound pressure levels for pure tones and supra-aural earphones. Geneva: International Organization for Standardization.
369
ISO:389-3.1994. Reference zero for the calibration of audiometric equipment. Part 3: Reference equivalent threshold force levels for pure tones and bone vibrators. Geneva: International Organization for Standardization. ISO:8253-1. 2010. Acoustics - Audiometric test methods. Part 1: Pure-tone air- and bone-conduction audiometry. Geneva: International Organization for Standardization. Reinfeldt S., Stenfelt S., Good T. & Håkansson B. 2007. Examination of bone-conducted transmission from sound field excitation measured by thresholds, ear-canal sound pressure, and skull vibrations. J Acoust Soc Am, 121, 1576–1587. Reinfeldt S., Östli P., Håkansson B. & Stenfelt S. 2010. Hearing one’s own voice during phoneme vocalization: Transmission by air- and boneconduction. J Acoust Soc Am, 128, 75–762. Reinfeldt S., Stenfelt S. & Håkansson B. 2013. Estimation of boneconduction skull transmission by hearing thresholds and ear-canal sound pressure. Hear Res, 299, 19–28. Reyes R.A., Tjellström A. & Granström G. 2000. Evaluation of implant losses and skin reaction around extraoral bone-anchored hearing implants: A 0- to 8-year follow-up. Otolaryngol Head Neck Surg, 122, 272–276. Snik A., Mylanus E., Proops D., Wolfaardt J., Hodgetts W. et al. 2005. Consensus statements on the BAHA system: Where do we stand at present? Ann Otol Rhinol Laryngol Suppl, 195, 2–12. Shirazi M., Marzo S. & Leonetti J. 2006. Perioperative complications with the bone-anchored hearing aid. Otolaryngol Head Neck Surg, 134, 236–9. Stenfelt S., Håkansson B. & Tjellström A. 2000. Vibration characteristics of bone-conducted sound in vitro. J Acoust Soc Am, 107, 422–431. Stenfelt S. & Goode R.L. 2005. Transmission properties of boneconducted sound: Measurements in cadaver heads. J Acoust Soc Am, 118, 2373–2391. Stenfelt S. & Reinfeldt S. 2007. A model of the occlusion effect with boneconducted stimulation. Int J Audiol, 46, 595–608. Taghavi H., Håkansson B. & Reinfeldt S. 2012a. A novel bone-conduction implant system: Analog radio frequency data and power link design. In: Proceedings of the 9th IASTED international conference on biomedical engineering, Innsbruck, Austria, 327–335. Taghavi H., Håkansson B. & Reinfeldt S. 2012b. Analysis and design of RF power and data link using amplitude modulation of class-E for a novel bone-conduction implant. IEEE Trans Biomed Eng, 59, 3050–3059. Taghavi H., Håkansson B., Reinfeldt S., Eeg-Olofsson M. & Akhshijan S. 2012c. Feedback analysis in percutaneous bone-conduction device and bone-conduction implant on a dry cranium. Otol Neurotol, 33, 413–420. Tjellström A. & Granström G. 1994. Long-term follow-up with the bone anchored hearing aid: A review of the first 100 patients between 1977 and 1985. ENT J, 73, 21–23. Tjellström A. & Håkansson B. 1995. The bone-anchored hearing aid: Design principles, indications, and long-term clinical results. Otolaryngologic Clinics of North America, 28, 53–72 Tjellström A. & Stalfors J. 2012. Bone-anchored hearing device surgery: A 3- to 6-year follow-up with life table and worst-case scenario calculation. Otol Neurotol, 33, 891–894. Wazen J., Wycherly B. & Daugherty J. 2011. Complications of bone-anchored hearing devices. In: M. Kompis, M-D. Caversaccio, Implantable Bone Conduction Hearing Aids. Basel: Karger Press, pp. 63–72.
