To investigate the contribution of body vibrations to the vibratory sensation induced by high-level, complex low-frequency noise, we conducted two experiments. In Experiment 1, eight male subjects were exposed to seven types of low-frequency noise stimuli: two pure tones [a 31.5-Hz, 100-dB(SPL) tone and a 50-Hz, 100-dB(SPL) tone] and five complex noises composed of the pure tones. For the complex noise stimuli, the sound pressure level of one tonal component was 100 dB(SPL) and that of another one was either 90, 95, or 100 dB(SPL). Vibration induced on the body surface was measured at five locations, and the correlation with the subjective rating of the vibratory sensation at each site of measurement was examined. In Experiment 2, the correlation between the body surface vibration and the vibratory sensation was similarly examined using seven types of noise stimuli composed of a 25-Hz tone and a 50-Hz tone. In both the experiments, we found that at the chest and the abdomen, the rating of the vibratory sensation was in close correlation with the vibration acceleration level (VAL) of the body surface vibration measured at each corresponding location. This was consistent with our previous results and suggested that at the trunk of the body (the chest and the abdomen), the mechanoreception of body vibrations plays an important role in the experience of the vibratory sensation in persons exposed to high-level low-frequency noise. At the head, however, no close correlation was found between the rating of the vibratory sensation and the VAL of body surface vibration. This suggested that at the head, the perceptual mechanisms of vibration induced by high-level low-frequency noise were different from those in the trunk of the body.
Keywords: Body surface, complex low-frequency noise, low-frequency noise, noise-induced vibration, vibratory sensation
|How to cite this article:|
Takahashi Y. A study on the contribution of body vibrations to the vibratory sensation induced by high-level, complex low-frequency noise. Noise Health 2011;13:2-8
|How to cite this URL:|
Takahashi Y. A study on the contribution of body vibrations to the vibratory sensation induced by high-level, complex low-frequency noise. Noise Health [serial online] 2011 [cited 2020 Apr 5];13:2-8. Available from: http://www.noiseandhealth.org/text.asp?2011/13/50/2/73993
| Introduction|| |
Low-frequency noise and noise with a large low-frequency component are prevalent in living and working environments. ,,,,, Some household appliances, such as air-conditioning systems and ventilating systems, generate low- or moderate-level low-frequency noises. Noises from cars, trains, ships, and aircraft have large low-frequency components. In working environments, various industrial machines, such as blowers, exhaust fans, air compressors, large engines, and the like, generate low-frequency noises at high sound pressure levels that occasionally exceed 100 dB (SPL).
The hearing sensitivity of a human being deteriorates at lower frequencies.  In spite of its low audibility, however, low-frequency noise often causes a person to experience a vibratory sensation. Mψller and Lydolf  reported that many persons perceived vibration in their bodies when exposed to low-frequency noise. Inukai and his collaborators  revealed that "vibration", as well as "sound pressure" and "loudness", is one of the chief contributors to human psychological responses to low-frequency noise. The induction of a vibratory sensation is a unique and intrinsic characteristic of low-frequency noise.
It is well known, on the other hand, that high-level low-frequency noise induces actual vibrations in the human body. ,, The levels of these vibrations, which we call "noise-induced vibrations" in this article, are not especially high. Our previous study,  however, showed that subjective ratings of vibration perceived in the chest or the abdomen were in close correlation with the vibration acceleration levels (VALs) of noise-induced vibrations measured at the corresponding location on the body surface.  This suggested that in the chest and the abdomen of persons exposed to high-level low-frequency noise, mechanoreception contributes to the perception of vibration. However, the previous study was carried out using only pure tones within a narrow frequency range (20-50 Hz). To confirm the results mentioned above, it was necessary to carry out additional studies using more varied low-frequency noise stimuli.
The aim of this study, as a follow-up of the previous study, was to investigate the contribution of noise-induced vibration to the vibratory sensation induced by high-level, complex low-frequency noise. For this purpose, we carried out two experiments in which subjects were exposed to seven types of low-frequency noise stimuli consisting of two pure tones and five complex noises composed of the pure tones. In the experiments, we measured noise-induced vibrations at five locations on the body surface and examined the correlation between the VALs of the measured noise-induced vibration and the subjective ratings of the magnitude of vibration in the corresponding part of the body.
Definition of key terms
In this study, we defined the head as the body part above the neck. The chest was defined as the trunk of the body between the neck and the diaphragm, and the abdomen was defined as the trunk of the body between the diaphragm and the pelvis.
We defined "vibration perceived in the head" as the subjective perception of vibration in either the whole of the head or any part of the head, independent of other sensations such as the hearing sensation and the perception of vibration in any other part of the body. "Vibration perceived in the chest" and "vibration perceived in the abdomen" were defined similarly.
| Methods|| |
Eight male subjects (19-25 years, mean ± SD = 22.8 ± 1.9 years) with normal hearing participated in Experiment 1. The experiment was carried out in a sound-insulated test chamber [3.16 m (W) × 2.85 m (L) × 2.80 m (H)] equipped with 12 loudspeakers (TL-1801, Pioneer , Tokyo, Japan).  All the loudspeakers were installed in a wall located in front of the subject. The background noise in the test chamber had been found to be adequately low in a previous study  and was assumed not to affect our findings. The temperature in the test chamber was set at 25°C.
Seven types of low-frequency noise stimuli were used in Experiment 1. Two of them were 31.5- and 50-Hz pure tones with sound pressure levels of 100 dB(SPL). The other five stimuli were complex noises composed of the pure tones. For the complex noise stimuli, the sound pressure level of one tonal component was 100 dB(SPL) and that of another component was either 90, 95, or 100 dB(SPL). The sources for the two pure tonal stimuli were WAV-type data generated on a personal computer, and the source data for the five complex noises were generated by synthesizing the two source data for the pure tonal stimuli with a phase difference of 0°. After being D/A converted, the frequency spectrum of the source data was modified through a digital filter (CP4, Lake Technology, Sydney, Australia) to compensate for the frequency response inherent in the test chamber.  Then, the sound source was power-amplified and fed to the loudspeakers.
The subject's body vibrations were measured on the body surface by the same method as reported previously.  Briefly, using a small (3.56 mm × 6.86 mm × 3.56 mm) and lightweight (0.5 g) accelerometer (EGA-125-10D, Entran Devices, Fairfield, USA) attached to each measuring location by means of double-sided adhesive tape, we detected the vibration perpendicular to the body surface, amplified it, passed it through a low-pass filter, and then recorded it on Digital Audio Tape (DAT) using a multi-channel data recorder (PC216Ax, Sony Precision Technology, Tokyo, Japan). Off-line analysis by a fast Fourier transform (FFT) analyzer (HP3566A, Hewlett Packard, Washington, USA) yielded the power spectrum of the detected vibration. At 31.5 and 50 Hz, the spectral component in the power spectrum was transformed to a VAL defined as follows:
VAL = 20 × log 10 (ameas/aref ) (dB),
where ameas was the measured acceleration [m/second 2 (r.m.s.)] and aref was the reference acceleration equal to 10-6 m/second 2 . We then calculated the total vibration acceleration level (VAL total ), defined as the power summation of the 31.5-Hz VAL and the 50-Hz VAL. We assumed that the measured VALs included the contribution of vibrations inherent in the human body. In the above transformation, however, we did not separate the inherent vibration from the total vibration measured, because the phase relationship between the inherent vibration and the noise-induced vibration was unknown. Although the inherent vibration originating in a strong heartbeat and respiration could contribute to the perception of vibration, these sensations were expected to be perceivable at very low frequencies. Within the 25- to 50-Hz frequencies used in the present study, the magnitudes of the inherent vibrations were not large and, therefore, we supposed that the inherent vibration hardly contributed to the subjective perception of vibration.
The experiment comprised three sessions. In the first session, using two accelerometers, we simultaneously measured the subject's body surface vibrations at the right anterior (2 cm above the right nipple) and left anterior chest (2 cm above the left nipple) with the subject in a standing position. At the beginning of the session, the inherent vibrations were recorded (1 minute) without presenting a noise stimulus. Then, seven types of noise stimuli were presented in a random order. The duration of each noise exposure was 1 minute, during which the noise-induced vibrations were recorded. After each noise exposure, participants had a 1-minute-long rest period with no noise stimulus. During the rest period, the subject rated the magnitude of "vibration perceived in the chest" experienced in the preceding exposure period as either "not sensed" [rating score (RS) = 1], "slightly sensed" (RS = 2), "mildly sensed" (RS = 3), "strongly sensed" (RS = 4), or "very strongly sensed" (RS = 5). No reference noise was presented during the rating process.
During the second session, in a manner similar to that used in the first session, the subject's inherent and noise-induced vibrations were measured simultaneously at the right anterior (5 cm below the pit of the stomach and 5 cm to the right of the midline) and left anterior abdomen (5 cm below the pit of the stomach and 5 cm to the left of the midline). The subject rated the magnitude of "vibration perceived in the abdomen" using the same rating scale as in the first session.
In the last session, during which the subject sat on a stool, the subject's inherent and noise-induced vibrations were measured at the forehead (2 cm above the level of the eyebrows and on the midline) using one accelerometer, and "vibration perceived in the head" was rated by the same rating scale as that used in the preceding two sessions. Throughout the three sessions, the subject wore no clothes on the upper half of the body to allow the accelerometers to be attached, and had no hearing protection so that he could be exposed to the low-frequency noise stimuli under his normal hearing condition.
Eight male subjects (20-27 years, mean ± SD = 23.4 ± 2.5 years) with normal hearing, who were not identical to the group used in Experiment 1, were enrolled in Experiment 2. We used seven types of low-frequency noise stimuli in which the 31.5-Hz tone was replaced with a 25-Hz tone. Except for the subjects and the noise stimuli, we used the same experimental methods and procedures as in Experiment 1.
For statistical analysis, we used a statistical software package (SPSS for Windows 17, SPSS Japan, Tokyo, Japan ) and adopted a P value below 0.05 as the criterion for statistical significance.
The protocol of the study was approved in advance by the Research Ethics Committee of the National Institute of Industrial Health, Japan (National Institute of Occupational Safety and Health, Japan, presently), and informed consent for study participation was obtained from each subject before the measurement.
| Results|| |
[Figure 1] shows the VALs (means ± SD) and the VAL total values (means) of the inherent and noise-induced vibrations measured at the right chest in the two experiments. In both the experiments, all the VALs of the two components (31.5 and 50 Hz in Experiment 1, 25 and 50 Hz in Experiment 2) in the noise-induced vibration were significantly higher than those in the inherent vibration (P < 0.05, by the Wilcoxon signed-rank test), except in one case [the 25-Hz vibration induced by the stimulus composed of a 25-Hz, 90-dB(SPL) tone and a 50-Hz, 100-dB(SPL) tone in Experiment 2]. This suggested that the present treatment of the inherent vibration in calculating VAL caused no serious difficulty. For all the noise stimuli, the VAL total value of the noise-induced vibration was approximately equal to the higher VAL in the two components. Characteristics similar to those mentioned above were found in the noise-induced vibrations measured at the head and the abdomen, though the VALs at the head and the abdomen were lower than those at the chest. For the detailed characteristics of the noise-induced vibrations measured in this study, the previous article can be referred,  in which part of the experimental data (for six of the eight subjects) in Experiment 1 were analyzed.
|Figure 1: The VALs (means ± SD) and the VALtotal values (means) of the inherent and noise-induced vibrations measured at the right chest (a) in Experiment 1 and (b) in Experiment 2. "No exposure" denotes the inherent vibrations|
Click here to view
[Figure 2] shows the rating scores of three types of vibratory sensation (means ± SD) in the two experiments. In Experiment 1 [Figure 2]a, we found a statistically significant difference in only 1 of the 21 cases (seven noise stimuli × three combinations of the vibratory sensations) between any two rating scores of different vibratory sensations (by the Wilcoxon signed-rank test). In Experiment 2 [Figure 2]b, we found a statistically significant difference in only 4 of the 21 cases. Four of the five significant differences in the two experiments were the differences between the "vibration perceived in the head" and the "vibration perceived in the abdomen". That is, there was a slight tendency that the "vibration perceived in the head" was rated more highly than the "vibration perceived in the abdomen".
|Figure 2: The rating scores (means ± SD) of three types of vibratory sensation obtained (a) in Experiment 1 and (b) in Experiment 2|
Click here to view
As shown in the large SD in [Figure 2], the inter-individual difference in the subjective perception of vibration was quite large, which implied that it was difficult to derive a clear relationship between the rating score of the vibratory sensation and the VAL of the noise-induced vibration. Therefore, to find a general tendency, we examined the correlation between the mean rating score of the vibratory sensation and the mean VAL of the noise-induced vibration. [Figure 3] shows the correlations between the mean ratings of the "vibration perceived in the chest" and the mean VAL total values of the noise-induced vibration measured at the chest in the two experiments. In the correlations, two mean VAL total values measured at the right and left sides of the body are plotted against the mean subjective rating because no clear difference was found between the VALs measured at the right and left half sides of the body.  The solid line in the figure is a regression line, while the "r" in the figure denotes the Pearson correlation coefficient. We expected that calculating the correlations using the mean values would minimize the effect of the variance in the raw data. Therefore, we considered that the Pearson correlation coefficient could be applied validly, though this treatment was artificial. In both the experiments, the correlations at the chest were found to be statistically significant [r = 0.772 (P < 0.01) in Experiment 1, and r = 0.901 (P < 0.01) in Experiment 2]. [Figure 4] shows the correlations at the abdomen obtained by the same method as that used in [Figure 3]. The correlations at the abdomen were also found to be statistically significant in both the experiments [r = 0.833 (P < 0.01) in Experiment 1, and r = 0.813 (P < 0.01) in Experiment 2]. At the head, however, no statistically significant correlation between the subjective ratings and VALs was obtained in either experiment.
|Figure 3: The correlation between the mean ratings of the "vibration perceived in the chest" and the mean VALtotal values measured at the chest (a) in Experiment 1 and (b) in Experiment 2|
Click here to view
|Figure 4: The correlation between the mean ratings of the "vibration perceived in the abdomen" and the mean VALtotal values measured at the abdomen (a) in Experiment 1 and (b) in Experiment 2|
Click here to view
At the chest, the slope of the regression line was calculated to be 0.063 [95% confidence interval (CI): 0.030-0.090] in Experiment 1 [Figure 3]a, while the slope was calculated to be 0.094 (95% CI: 0.065-0.122) in Experiment 2 [Figure 3]b. These two slopes were close to each other. At the abdomen, the slopes of the regression line in Experiments 1 and 2 were 0.180 (95% CI: 0.105-0.255) and 0.152 (95% CI: 0.084-0.221), respectively [Figure 4]. They were close to each other and steeper than the slopes obtained at the chest. The similarity in the regression lines was indicative of the consistency between the results of the two experiments which were carried out independently of each other with different subjects.
[Table 1] summarizes the Pearson correlation coefficients obtained at the three locations. At the chest and the abdomen, all the four Pearson correlation coefficients calculated for the VAL total were found to be statistically significant (P < 0.01). For comparison, in the table, we have also listed the coefficients for the correlation between the mean subjective rating of the vibratory sensation and the total sound pressure level (SPL total ), which was equal to the sound pressure level of the stimulus, for the tonal stimuli, or the power summation of the sound pressure levels of the two tonal components in the stimulus, for the complex noise stimuli. No statistical significance was found in either experiment in the Pearson correlation coefficient for the SPL total . This result indicated that at the trunk of the body, the subjective perception of vibration was related more closely to the magnitude of noise-induced vibration than to the sound pressure level of the noise.
|Table 1: Coefficients for the correlation between the mean ratings of the vibratory sensation and the mean VALtotal and SPLtotal values|
Click here to view
The 50-Hz VALs of the noise-induced vibration were higher than the 31.5- and 25-Hz VALs [Figure 1]. To verify whether the 50-Hz vibration contributed dominantly to the perception of vibration at the trunk of the body, we examined the Pearson correlation coefficient for the VAL total while applying a test frequency-weighting to the VAL. In this examination, hypothesizing that the slope of the test frequency-weighting was constant in the 25- to 50-Hz range, the slope varied between -20 and +20 dB/oct. by a 0.5-dB/oct. step. The correction at 50 Hz was temporarily set to be 0 dB. In trying a slope of -6.0 dB/oct., for example, corrections for the VALs were +6.0 dB at 25 Hz, +4.0 dB at 31.5 Hz, and 0.0 dB at 50 Hz. [Figure 5] shows the trace of the Pearson correlation coefficient obtained at the chest. In both the experiments, the Pearson correlation coefficient was almost at its maximum when the slope of the test frequency-weighting was approximately -10 dB/oct. or higher. Applying a frequency-weighting with a negative slope to the VAL means attaching importance to the 31.5- or 25-Hz component in the noise-induced vibration. At the chest, therefore, we speculated that the 31.5- [Figure 5]a and 25-Hz [Figure 5]b components contributed effectively to the perception of vibration of the subjects in both the experiments. At the abdomen, on the other hand, the maximum in the trace of the Pearson correlation coefficients in the two experiments appeared when the slope of the test frequency-weighting was negative but nearly equal to 0.0 dB/oct. Therefore, the effective contribution of the 31.5- and 25-Hz components at the abdomen could also be assumed, though the contribution might be weaker than at the chest.
|Figure 5: The trace of the correlation coefficient for the VALtotal at the chest (a) in Experiment 1 and (b) in Experiment 2. The correlation coefficient was calculated while the slope of the test frequency-weighting was changed|
Click here to view
| Discussion|| |
In our previous study using pure tonal stimuli within the 20-50 Hz range,  we measured the subjects' ratings of the same three types of vibratory sensation as in the present study, and examined the relationship between the ratings and the VALs of the noise-induced vibrations measured at the same locations as in the present study. We found that the "vibration perceived in the head" tended to be rated more highly than the other vibratory sensations (the "vibration perceived in the chest" and the "vibration perceived in the abdomen"). At the chest and the abdomen, the ratings of the vibratory sensation were in closer correlation with the VALs of the noise-induced vibration than with the sound pressure levels of the noise stimuli. At the head, the subjective rating of the vibratory sensation was also in significant correlation with the VALs measured at the head, but the correlation seemed to be looser than those obtained at the chest and the abdomen. In addition, the slope of the regression line for the correlation was steepest at the abdomen and gentlest at the head. The vibratory sensations in the previous study were rated on a scale of 1-3, which was different from the rating scale (1-5) used in this study. Therefore, quantitative comparisons between the two results should be approached with caution. Qualitative comparisons between them, however, are considered to be possible. The present results [Figure 3] and [Figure 4] are very similar to and consistent with the previous results qualitatively. The similarity and consistency in the two sets of results support the idea that at the trunk of the body, mechanoreception plays an important role in the perception of vibration induced by high-level low-frequency noise.
At the head, on the other hand, the two results indicate that mechanoreception does not contribute much to the perception of vibration induced by high-level low-frequency noise, and suggest that the mechanisms involved in the perception of vibration in the head are different from those in the trunk of the body. One common feature in the two sets of results was that in spite of the comparatively low VALs of the noise-induced vibration at the head, the "vibration perceived in the head" was rated highly. Increasing pressure in the ear may be one cause of the vibratory sensation experienced in the head. The effects of auditory stimulation by high-level low-frequency noise on the perception of vibration in the head remain to be investigated.
In this study, only high-level noise stimuli inducing vibrations that could be detected by our accelerometer were used. It is expected, however, that low-frequency noise at a lower sound pressure level can also cause a person to feel vibration. The ranges of frequencies and sound pressure levels that contribute to the perception of vibratory sensation via noise-induced vibration remain as a subject of future investigation.
Smith previously found that chest resonance occurred around the 63- to 100-Hz range in response to high-level low-frequency noises from aircraft engines.  In our study, however, we did not use noise stimuli at frequencies higher than 50 Hz. This was because the spatial uniformity of the sound pressure levels in the test chamber deteriorated at such frequencies.  If noise stimuli at frequencies higher than 50 Hz could have been used, some useful results on the contribution of the chest resonance to the perception of vibration might have been obtained. This interesting point also remains to be investigated.
In recent years, "vibroacoustic disease", which has been chiefly found in crew and mechanical engineers working in or with airplanes, has drawn increasing attention to the effects of low-frequency noise on humans. According to Castelo Branco and his collaborators, , the disease is a whole-body, systemic pathology caused by long-term exposure to high-level low-frequency noise that results in thickening of the cardiovascular structures, airway flow limitation, abnormal proliferation of extra-cellular matrices, and so on. Although the detailed mechanisms causing the disease have not been clarified, it is possible that noise-induced vibration is associated with the occurrence of the disease. It is reasonable to expect that vibrations appearing on the body surface reflect the vibrations occurring in the inner body. If the detailed relationship between noise-induced vibration and the vibratory sensation is clarified, the subjective magnitude of the vibratory sensation could be used as an indicator of the risk of adverse physical effects, including vibroacoustic disease, caused by high-level low-frequency noise.
| Conclusions|| |
The results of the present study, together with our previous results, support the idea that in the trunk of the body, mechanoreception plays an important role in the perception of the vibratory sensation induced by high-level low-frequency noise. The data suggest that the mechanisms of vibration perception in the head may be different from those in the trunk of the body.
It should be noted, however, that the present results were obtained under limited experimental conditions including a small number of subjects, noise stimuli within a limited frequency range, and a narrow range of sound pressure levels. To confirm the present results and obtain more useful results, more detailed studies should be carried out.
| Acknowledgment|| |
The present study was supported in part by a Grant-in-Aid for Encouragement of Young Scientists (No. 13780699), founded by the Japan Society for the Promotion of Science.
| References|| |
|1.||Berglund B, Hassmén P, Job RFS. Sources and effects of low-frequency noise. J Acoust Soc Am 1996;99:2985-3002. |
|2.||Leventhall HG. Low frequency noise and annoyance. Noise Health 2004;6:59-72. |
|3.||Persson Waye K, Bengtsson J, Agge A, Björkman M. A descriptive cross-sectional study of annoyance from low frequency noise installations in an urban environment. Noise Health 2003;5:35-46. |
|4.||Mirowska M. An investigation and assessment of annoyance of low frequency noise in dwellings. J Low Freq Noise Vib Active Control 1998;17:119-26. |
|5.||Tubbs RL. Noise exposure to airline ramp employees. Appl Occup Environ Hyg 2000;15:657-63. |
|6.||Pawlaczyk-Luszczynska M. Occupational exposure to infrasonic noise in Poland. J Low Freq Noise Vib Active Control 1998;17:71-83. |
|7.||Moller H, Pedersen CS. Hearing at low and infrasonic frequencies. Noise Health 2004;6:37-57. |
|8.||Moller H, Lydolf M. A questionnaire survey of complaints of infrasound and low-frequency noise. J Low Freq Noise Vib Active Control 2002;21:53-64. |
|9.||Inukai Y, Taya H, Miyano H, Kuriyama H. A multidimensional evaluation method for the psychological effects of pure tones at low and infrasonic frequencies. J Low Freq Noise Vib 1986;5:104-12. |
|10.||Smith SD. Characterizing the effects of airborne vibration on human body vibration response. Aviat Space Environ Med 2002;73:36-45. |
|11.||Takahashi Y, Kanada K, Yonekawa Y. Some characteristics of human body surface vibration induced by low frequency noise. J Low Freq Noise Vib Active Control 2002;21:9-20. |
|12.||Takahashi Y, Maeda S. Measurement of human body surface vibrations induced by complex low-frequency noise composed of two pure tones. J Low Freq Noise Vib Active Control 2003;22:209-23. |
|13.||Takahashi Y, Kanada K, Yonekawa Y. The relationship between vibratory sensation and body surface vibration induced by low-frequency noise. J Low Freq Noise Vib Active Control 2002;21:87-100. |
|14.||Takahashi Y, Yonekawa Y, Kanada K, Maeda S. An infrasound experiment system for industrial hygiene. Ind Health 1997;35:480-8. |
|15.||Castelo Branco NA, Alves-Pereira M. Vibroacoustic disease. Noise Health 2004;6:3-20. |
|16.||Alves-Pereira M, Castelo Branco NA. Vibroacoustic disease: Biological effects of infrasound and low-frequency noise explained by mechanotransduction cellular signalling. Prog Biophys Mol Biol 2007;93:256-79. |
6-21-1 Nagao, Tama-ku, Kawasaki 214-8585
Source of Support: Supported in part by a Grant-in-Aid for Encouragement of Young Scientists (No. 13780699), founded by the Japan Society for the Promotion of Science, Conflict of Interest: None
[Figure 1], [Figure 2], [Figure 3], [Figure 4], [Figure 5]