Threshold and frequency- and temporal resolutions of distantly presented bone-conducted sound in the audible-frequency range

In this study, air-conducted sounds were insulated strongly by two types of earplugs and an earmuff to evaluate only the bone-conduction components. First, urethane earplugs (DKSH Japan Silencia) were inserted into the external auditory canals of the subjects. Subsequently, ear conchae were covered with silicone earplugs (Insta-Mold Products Insta-Putty). Finally, except when the stimulus was presented to the mastoid process, the participants wore an ear-muff (3 M PELTOR X5A) over the earplugs (Fig. 1). To examine the insulating effect, we measured the sound pressure level of the air-conducted sounds radiated from the vibrator at the opening of the external auditory canal and the sound attenuation value of the insulating method used in this study based on ISO 4869. ²⁶⁾

Zoom In Zoom Out Reset image size

2.1.1. Measurement of the sound pressure level of air-conducted sounds at the opening of the external auditory canal

2.1.2. Estimation of the sound attenuation value based on ISO 4869

Seven males (21–26 years) with no history of deficits in their hearing functions participated. Bone-conduction stimuli were presented to the left side of the following body parts (Fig. 2): (a) the mastoid process of the temporal bone, (b) sternocleidomastoid muscle, (c) the sternal end of the clavicle, and (d) acromion (only for the hearing threshold measurement). Bone-conduction stimuli in the audible-frequency range (audible stimuli) were presented using a bone-conduction vibrator [RadioEar B-81, Fig. 3(a)], and bone-conduction stimuli in the ultrasonic range (ultrasonic stimulus) were presented using a piezoelectric ceramic vibrator [Murata Manufacturing MA40E7S, Fig. 3(b)]. Figure 4 shows the frequency response for B-81 used in this study between 100 and 10 000 Hz at a constant input voltage of 1 V_RMS.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The audible stimuli were generated with a personal computer using MATLAB (MathWorks) at a sampling frequency of 48 000 Hz and fed to the vibrator via a 24 bit D/A converter (Echo Digital Audio Audiofire 12). On the other hand, the ultrasonic stimuli were generated with a personal computer using MATLAB at a sampling frequency of 192 000 Hz and fed to the vibrator via a 24 bit D/A converter and amplifier (Mess-Tek M-2629B). The bone-conduction vibrators were pressed onto the mastoid process and sternocleidomastoid muscle using an elastic band with foam [Figs. 5(a) and 5(b)]. An icing supporter was used to fix the vibrator to the clavicle and acromion [Figs. 5(c) and 5(d)].

Zoom In Zoom Out Reset image size

Just prior to the experimental sessions in each measurement, the stimulus was presented at the stimulus intensity of the measurement (described later) with the vibrator slightly away from each stimulus location. Then it was confirmed that the air-conducted sounds radiated from the vibrator were not perceived by participants subjectively at any frequency. All the measurements were conducted in an anechoic room.

2.2.1. Measurement of hearing thresholds

2.2.2. Measurement of DLFs

2.2.3. Measurement of TMTFs

The TMTFs were measured when the bone-conduction stimuli were presented to the head (mastoid process), neck (sternocleidomastoid), and trunk (clavicle) to evaluate the temporal resolution of the distantly presented bone-conduction hearing.

The TMTF shows the threshold of sinusoidal amplitude modulation (SAM) detection as a function of the modulation frequency. The SAM detection threshold was determined systematically by measuring the modulation depth. A double sideband amplitude modulation was applied to 1000 Hz Hz and 4000 Hz sinusoidal carriers.

The SAM stimuli are expressed as follows:

$$\begin{eqnarray}&&f\left(t\right)=A\left\{1+m\times \,\sin \left[2\pi {f}_{m}t\right]\right\}\times \,\sin \left(2\pi {f}_{c}t\right).\end{eqnarray} \tag{ 1 }$$

Here, $A$ is the amplitude of the entire stimulus, $m$ is the modulation depth, and ${f}_{m}$ and ${f}_{c}$ are the modulation and carrier frequencies, respectively. The duration of the stimuli was 800 ms, which included 75 ms rising/falling ramps.

The TMTFs were measured using a two-up one-down 3AFC adaptive procedure. The stimulus output levels were set to a level at which each participant was able to detect modulation sufficiently based on a 25 dB SL.

In each trial, three bone-conduction stimuli were presented sequentially. One of the three bone-conduction stimuli was modulated, whereas the others were unmodulated. The participants were requested to respond to the modulation interval. The TMTF measurements were started from the modulation depth $\left(m\right)$ at which the modulation could be detected sufficiently by the participants, and the modulation depth in the modulation intervals was varied adaptively in the same manner as in the measurement of the hearing thresholds. The step size of the modulation depth initially corresponded to 4 dB (in units of 20 log) and was reduced to 2 dB after two reversals and finally to 1 dB after four reversals. The mean of the last eight reversals in a block of 11 reversals was used as the modulation threshold. The modulation frequencies were set to 10, 20, 50, 100, 125, 150, 175, and 200 Hz for the 1000 Hz carrier and 10, 20, 50, 100, 125, 150, 175, 200, 300, 400, 500, 600, 700, 800, 1000, and 1600 Hz for the 4000 Hz carrier, and the TMTF measurements were conducted for each modulation frequency.

In the current study, the intensities of the bone-conduction stimuli were lower (25 dB SL) than former studies ^20,28) to prevent the perception of air-conducted emission from the vibrator. The TMTFs of the air-conducted sounds were measured as a function of the carrier level to examine the relationship between the TMTF and the stimulus intensity.

Four males (21–25 years) with no history of deficits in their hearing functions participated. The carrier and modulation frequencies were the same as those described in Sect. 2.2.3. The carrier levels were set to 15, 30, 45, 60, and 75 dB SL, and the stimuli were presented using a headphone (Sennheiser HD 660 S). The TMTFs were estimated for each participant in the same manner as described in Sect. 2.2.3.

Figures 6(a) and 6(b) show the sound pressure level of the air-conducted sounds at the opening of the external auditory canal and the sound attenuation value by the earplugs and earmuff, respectively. The clavicle shows relatively larger sound pressure levels of approximately 50, 62, and 56 dB SPL, when a 2000, 3000, and 4000 Hz tone bursts were presented to the clavicle at 15 dB SL, respectively. However, the mastoid process and sternocleidomastoid shows sound pressure levels lower than 50 dB SPL. On the other hand, the sound attenuation values are larger than 43 dB at any frequency.

Zoom In Zoom Out Reset image size

Figure 7 shows the relative thresholds of the distantly presented bone-conducted sound with the hearing threshold for the mastoid process in each participant serving as the reference (0 dB). Although the hearing thresholds increased as the stimulus placements moved further away from the head, all the participants were able to perceive bone-conducted sounds even in the distal parts of the body. In the audible-frequency range, although the peak (approx. 5–10 dB) was observed at 500 Hz, the relative thresholds increased as the stimulus frequency increased, particularly at 1000–4000 Hz in the clavicle and acromion. On the other hand, in the ultrasonic range, the relative thresholds were lower than those in the audible-frequency range at all the stimulus placements.

Zoom In Zoom Out Reset image size

The two-way analysis of variance (ANOVA) of the stimulus placements and frequencies showed the significant main effects of these two variables, and a significant interaction between them (P < 0.05). The post-hoc test (Tukey HSD test) showed that the increases in the hearing thresholds relative to the mastoid process in the audible high-frequency range were significantly higher than those in the low-frequency range (sternocleidomastoid: 250, 500, 750, and 1000–4000 Hz, clavicle: 250, 500, 750, and 1000–3000 Hz and 250, 500, 750, and 1000–4000 Hz, acromion: 250–3000 Hz and 250, 500, and 750–4000 Hz). In the clavicle and acromion, those in the high-frequency range were significantly higher than those in the middle-frequency range (clavicle: 2000–4000 Hz, acromion: 2000 and 3000–4000 Hz). Moreover, those in the ultrasonic range (30000 Hz) were significantly lower than those in the audible-frequency range (250–4000 Hz). In the ultrasonic range, the relative threshold increased significantly at the acromion compared with the other locations, whereas in the audible-frequency range, the relative threshold decreased significantly at the sternocleidomastoid muscle compared with the other locations.

Figure 8 shows the DLFs at each stimulus placement. All the participants were able to discriminate the frequency at all the stimulus placements and CFs. Generally, the DLFs were approximately 0.3% at all the CFs at the mastoid process and sternocleidomastoid muscle. In the clavicle, the DLF was equivalent to that of the mastoid process and sternocleidomastoid muscle at 500–2000 Hz, whereas it was higher at 250 and 4000 Hz. The two-way ANOVA of the stimulus placements and CFs showed the significant main effects of these two variables, and a significant interaction between them (P < 0.05). The post-hoc test (Tukey HSD test) showed that the DLF increased significantly at 250 and 4000 Hz in comparison with that at the other frequency ranges in the clavicle.

Zoom In Zoom Out Reset image size

Figures 9(a) and 9(b) show the modulation thresholds of the bone-conducted sounds at each stimulus placement using the 1000 Hz and 4000 Hz sinusoidal carriers, respectively. In the case of the 1000 Hz carrier, no significant variation depending on the increase in the modulation frequency was observed at any of the stimulus placements. Additionally, no significant increase was observed as the stimulus placements moved further away from the head. In the case of the 4000 Hz carrier, the modulation thresholds increased at a modulation frequency of 100 Hz or higher. In the modulation frequency range, although the modulation thresholds of the mastoid process were higher than those of the sternocleidomastoid muscle, no significant difference was observed.

Zoom In Zoom Out Reset image size

Figures 10(a) and 10(b) show the modulation thresholds of the air-conducted sounds at each carrier level using the 1000 Hz and 4000 Hz sinusoidal carriers, respectively. At high carrier levels, the modulation thresholds decreased above a modulation frequency of 100 Hz for the case of the 1000 Hz carrier and 400 Hz for the case of the 4000 Hz carrier, whereas the shapes of the TMTFs were similar to those of the bone-conducted sounds obtained in this study at low carrier levels.

Zoom In Zoom Out Reset image size

When a 3000 Hz tone burst was presented to the clavicle at 15 dB SL, the sound pressure level was the highest at approximately 62 dB SPL. On the other hand, the sound attenuation value at 3000 Hz was approximately 44 dB. In this case, it is considered that the air-conducted sound entering the external auditory canal was perceived as a low loudness sound of about 10 dB SL because the absolute threshold of human hearing at 3000 Hz is approximately 8 dB SPL. ²⁹⁾ Therefore, we need to consider that air-conducted sound may affect the threshold at 3000 Hz on the clavicle, however, it seems that the radiation of air-conducted sounds from the vibrator does not produce significant changes on the current results.

At 2000 Hz, the sound pressure levels were approximately 34–50 dB SPL, whereas the sound attenuation value was approximately 45 dB. So, it is considered that the amount of air-conducted sound sufficient to be perceived by the participants did not enter the external auditory canal because the absolute threshold at 2000 Hz is equivalent to that at 3000 Hz. ²⁹⁾ The same is true for the other frequency ranges, except for 3000 Hz. These results indicate that the air-conducted sounds were sufficiently insulated using the urethane and silicone earplugs and earmuff to evaluate only the perception of bone conduction.

The bone-conducted sounds presented to the distal parts of the body were perceived in all the vibrator locations by all the participants. This result suggests the possibility of developing a novel bone-conduction device using "distant presentation." The hearing thresholds increased as the stimulus placements moved further away from the head, as in our previous study. ²⁴⁾ On the other hand, a dip in the hearing threshold of approximately 5–10 dB at 1000 Hz, shown in the previous study, was not observed. This discrepancy in the hearing threshold might be caused by a difference in the vibrator (previous study: RadioEar B-71) or the method of holding the vibrator; however, further studies are required to elucidate the detailed mechanism.

In this study, the increase in the hearing threshold from the mastoid process in the audible-frequency range was approximately 20–40 dB in the sternocleidomastoid muscle, and 35–65 dB in the clavicle and acromion, which are larger than those in the ultrasonic range. In addition, our previous study on the hearing threshold of distantly presented BCU has shown that the increase in the hearing threshold was only 20 dB in the upper arm and 27 dB in the lower arm. ²⁰⁾ These results indicate that the distance attenuation of bone-conducted sound in the human body is much larger in the audible-frequency range than that in the ultrasonic range. Furthermore, in the ultrasonic range, significant differences in the relative threshold between the clavicle and the acromion, which are at equivalent distances from the head, were observed, whereas no significant difference was observed in the audible-frequency range. This result suggests that the BCU perception is affected not only by the propagation distance but also by other factors, such as the differences in the biological tissue present in the propagation process. On the other hand, the perception of distantly presented bone-conducted sound in the audible-frequency range is strongly dependent on the propagation distance.

At the mastoid process and sternocleidomastoid muscle, the DLF was approximately 0.3% for all the CFs, which was the same as that of the air-conducted sound. ^28,30–32) This result indicates that the frequency resolution of distantly presented bone-conduction hearing is equivalent to that of air-conducted sound when bone-conducted sound is presented to a body part relatively close to the head. In addition, no significant difference was observed in the DLF among stimulus placements for almost all the CFs, whereas the hearing thresholds increased as the stimulus placements moved further away from the head. These results indicate that the degradation of frequency information that occurs during the propagation process of bone-conducted sounds is small.

No significant difference among stimulus placements was observed at any of the modulation frequencies, regardless of the carrier frequency, when bone-conducted sound was presented to the distal parts of the body. This result indicates that no degradation of the temporal information of bone-conducted sounds occurs during the propagation process in the human body.

According to a previous study on the TMTF of air-conducted sound, ³³⁾ the sidebands caused by amplitude modulation were perceived as another pitch above a certain modulation frequency, and the modulation threshold subsequently decreased. In contrast, in this study, no significant decrease in the modulation threshold of bone-conducted sound was observed as the modulation frequency increased. Moreover, it was confirmed that the shapes of the TMTFs of the air-conducted sounds were different from those of the bone-conducted sounds obtained in this study at high carrier levels, whereas they were similar at low carrier levels. These results indicate that the difference between the TMTF of bone-conducted sound obtained in this study and that of air-conducted sound obtained in the previous study occurred because no significant sideband exceeding the level of the noise floor was caused because the stimulus intensity was set to a smaller level (25 dB SL) to prevent the perception of air-conducted sounds radiated from the vibrator. Therefore, it is necessary to consider a more effective insulation method for air-conducted sounds and a method to eliminate the effect of vibratory sensation, which may be generated due to the increase in stimulus intensity, to elucidate the detailed perceptual mechanisms of distantly presented bone-conducted sound in the audible-frequency range.