Article | Open | Published:

# A feasibility study on non-invasive oxidative metabolism detection and acoustic assessment of human vocal cords by using optical technique

## Introduction

Voice disorder such as the vocal fatigue is a complex clinical phenomenon that is recognized as potentially debilitating presents a significant challenge to clinical practice1,2. Professional voice users such as teachers, singers and actors are particularly susceptible to vocal fatigue3,4,5,6. Certainly, vocal fatigue can seriously affect the social and occupational functioning of such professions. Although vocal fatigue has been defined as a feeling of vocal tiredness and weak voice after the voice is overused or abused7,8,9. Unfortunately, the definition, critical identifying features, and fundamental mechanisms of vocal fatigue remain either unclear. Therefore, vocal fatigue is an important vocal health issue that needs to be further investigated. Numerous studies have adopted the vocal loading test to induce vocal fatigue in voice-healthy participants, and then to track the changes in vocal function that has potential worth to investigate the progression of fatigue as it affects the normal voice10,11,12,13,14,15,16,17. The several current clinical detection methods such as auditory perceptual analysis, vocal cords imaging (laryngoscope and videolaryngostroboscopy), aerodynamic analysis, acoustic analysis, and self-evaluation have been widely used for clinical applications of the assessment of voice problems18,19,20,21,22,23. These methods have greatly increased our knowledge about the mechanisms of vocal fatigue. However, vocal fatigue is most likely multifaceted physiological and biomechanical mechanisms. The mechanisms include neuromuscular fatigue, increased vocal fold viscosity, reduced blood circulation, non-muscular tissue strain, and respiratory muscle fatigue. Therefore, the more multidimensional assessment of vocal fatigue is very important for clinical applications.

Functional near-infrared spectroscopy (fNIRS) is a non-invasive technique that provides continuous recording of local changes in tissue oxygenation and perfusion with detection of oxy- and deoxy-hemoglobin concentration changes with at least dual-wavelength (around 800 nm) near-infrared illumination24,25. Additionally, fNIRS has several benefits such as less expensive, non-ionizing radiation, real-time measurement, long-time monitoring, easy operation, and completely patient-oriented measurement with high temporal resolution (ms). The photons in the spectral window of 600–1000 nm wavelength (also known as “optical window” or “therapeutic window”) can penetrate several centimeters into human tissue. Therefore, fNIRS has generated a lot of scientific interest and has been applied in various deep tissue applications such as imaging of the brain, breast, limb, muscle, and joint24,25,26,27,28,29,30,31. According to the type of source modulation, fNIRS can be classified into three modes as continuous wave (CW), frequency domain (FD), and time domain (TD)25,31. In the CW system, the constant illumination is used for intensity measurement. For this reason, the CW system is solely used to detect changes in absorption coefficient that cannot be used to determine the absolute value of the concentration of oxygenated and deoxygenated hemoglobin. Contrarily, the information of absorption and scattering coefficients of the medium can be determined by using FD and TD system to measure the intensity and phase or time delay of the received light. Therefore, the absolute value of the tissue optical parameters can be estimated to obtain the absolute value of the concentration of oxygenated and deoxygenated hemoglobin. However, in the clinical studies, the analysis of statistically significant difference before and after some specific test is more important than quantification of absolute value. The CW system can be used to detect the concentration changes of oxygenated (Δ[HbO2]) and deoxygenated hemoglobin (Δ[Hb]). Additionally, the CW system provides higher time resolution, relatively low-cost, and easy transportability. Therefore, the CW based system was developed for detection of vocal cords in this study. Besides, the optical detection method such as fiber-optic acoustic sensing is able to measure vibrations from the surface of the skin during vocalization, providing intensity of time series and frequency information that can be processed for acoustic analysis. The fiber-optic acoustic sensing also called optical microphone has many benefits such as high sensitivity, anti-RFI, anti-EMI, high safety and credibility for medical diagnosis32,33,34,35. Consequently, the optical detection methods can provide not only oxidative metabolism detection but also acoustic assessment for vocal cords diagnosis.

## Results

Likewise, the acoustic parameters related to amplitude perturbation were also calculated to evaluate the differences between male and female groups after and before vocal loading task. Figure 5 shows the variation of the parameters related to amplitude perturbation. The result indicated that all the parameters were not significantly different between male and female groups except the parameter of APQ. Besides, there were only ShdB and Shim presented significantly different before and after vocal loading task in the male group. It is also noteworthy that all the value of parameters of amplitude perturbation increased after vocal loading task.

## Discussion

Acoustic analysis is the non-invasive and most used approach to voice quality assessment of voice disorder in clinical research and application36,37,38,39,40. Unlike previous studies of acoustic analysis that used the traditional microphone to record voice sample, the present study developed and used an optical technique (as an optical microphone) to record voice sample for acoustic analysis. This method was used to receive the vibration signals transmitted from the vocal cords to the surface of the tissue. The greatest benefit of the optical microphone is to avoid ambient sound and electromagnetic interference.

## Methods

### Participants

In this study, 60 healthy adults (30 males and 30 females), age between 20 to 25 years old (the mean age was 22 ± 2.15 years old) without throat or vocal cord diseases at least one month, were recruited from National Chiao Tung University, Taiwan. All participants provided written informed consent. The study was in accordance with the latest version of the Declaration of Helsinki, and approved by the Institutional Review Board (IRB) in Research Ethics Committee for Human Subject Protection, National Chiao Tung University, Taiwan.

### Experimental Setup

In this study, the non-invasive oxidative metabolism detection and acoustic assessment were integrated into a novel system by using the fNIRS-DOT, that we called it a multi-functional vocal cords detection system. Therefore, the information of oxidative metabolism and acoustics of the vocal cords could be captured simultaneously. Figure 7 shows the scheme of multi-functional vocal cords detection system. The system setup included a pair of laser diodes (785 nm and 850 nm) (WSLR-785/850–050 m-M-PD, Wavespectrum, Beijing, China), an independent laser diode (785 nm) and two silicon photo detectors (PDA100A, Thorlabs, Newton, New Jersey, U.S.). In function of the oxidative metabolism detection, the bifurcated fiber bundles, also called Y-type fibers (BFY400HS02, Thorlabs, New Jersey, US) was used to introduce laser light from the two laser heads into one optical fiber as the same emission point. The source-detector separation is 3 cm that can provide adequate depth of penetration (the penetration depth into tissue is approximately 1.5 cm or larger26,27,28,29) for oxidative metabolism detection of the vocal cords. In function of the acoustic assessment, the Y-type fiber was used to transmit and receive the optical signal as an optical microphone to detect the signal of the tissue surface vibration. All the backscattered optical signals from vocal cords and tissue surface vibration were detected with two silicon photo detectors (Si-PD), respectively. The amplified analog signals from the Si-PD were converted to digital data by using an A/D converter (NI myRIO-1900, National Instruments, Austin, Texas, U.S.) with 12-bit resolution. The A/D converter was also used to regulate the optimum laser diode drive current with the technique of Time-Division Multiplexing (TDM) to obtain stable and accurate measurements. The optical power of laser diode was regulated at 5 mW. The optical fibers were fixed on a black flexible probe that was brought in close contact with the skin to prevent noise from the environmental and surface backscattering.

## Data Analysis

### Oxidative metabolism detection

In this function, the backscattered optical signals of the dual-wavelength (785 nm and 850 nm) from human tissue were received for calculating the concentration changes in Δ[HbO2] and Δ[Hb] with technique of continuous-wave and a sampling rate of 50 Hz. According to the modified Beer-Lambert Law (MBLL)25,50,51, the optical density (OD) can be defined as follows:

$$O{D}_{\lambda }=-\,{\mathrm{log}}_{10}(\frac{I}{{I}_{0}})={\varepsilon }_{\lambda }CL{B}_{\lambda }+{G}_{\lambda }$$
(1)

where I o and I are the intensities of incident light and detected light, respectively; The OD λ is the optical density for wavelength λ that means the attenuation of near-infrared light intensity in tissue; ε λ is the extinction coefficient of the chromophore; C is the concentration of the chromophore; L is the distance between the light entry and exit the tissue; B λ is a pathlength factor that related to tissue scattering; G λ is defined as a geometric factor that related to tissue geometry. Assuming that G λ remains constant during a measurement, the change in optical density ΔOD λ can be obtained as follows:

$${\rm{\Delta }}O{D}_{\lambda }=O{D}_{\lambda }(t)-O{D}_{\lambda }({t}_{0})=-\,{\mathrm{log}}_{10}(\frac{I(t)}{I({t}_{0})})={\varepsilon }_{\lambda }{\rm{\Delta }}CL{B}_{\lambda }$$
(2)

where OD λ (t 0) and OD λ (t) are the initial and instantaneous values of the optical density from the tissue, respectively; I(t 0) and I(t) are the measured intensities at initial and instantaneous time; ΔC is the change in concentration of the chromophore. In human tissue, changes in concentration were dominated with the chromophore of HbO2 and Hb. Therefore, the change in optical density can be obtained as follows:

$${\rm{\Delta }}O{D}_{\lambda }=({\varepsilon }_{\lambda }^{Hb{O}_{2}}\cdot {\rm{\Delta }}[Hb{O}_{2}]+{\varepsilon }_{\lambda }^{Hb}\cdot {\rm{\Delta }}[Hb])\cdot L{B}_{\lambda }$$
(3)

The concentration changes in Δ[HbO2] and Δ[Hb] could be obtained by solving Eq. (3) with the optical signals of dual-wavelength (785 nm and 850 nm). Assuming that L and B λ remains constant during a measurement, the change in optical density can be obtained as follows:

$${\rm{\Delta }}O{D}_{785}=({\varepsilon }_{785}^{Hb{O}_{2}}\cdot {\rm{\Delta }}[Hb{O}_{2}]+{\varepsilon }_{785}^{Hb}\cdot {\rm{\Delta }}[Hb])$$
(4)
$${\rm{\Delta }}O{D}_{850}=({\varepsilon }_{850}^{Hb{O}_{2}}\cdot {\rm{\Delta }}[Hb{O}_{2}]+{\varepsilon }_{850}^{Hb}\cdot {\rm{\Delta }}[Hb])$$
(5)

Finally, the description can be rewritten from Eq. (4) and Eq. (5) as follows:

$${\rm{\Delta }}[Hb{O}_{2}]=\frac{{\varepsilon }_{785}^{Hb}\cdot {\rm{\Delta }}O{D}_{850}-{\varepsilon }_{850}^{Hb}\cdot {\rm{\Delta }}O{D}_{785}}{{\varepsilon }_{785}^{Hb}\cdot {\varepsilon }_{850}^{Hb{O}_{2}}-{\varepsilon }_{850}^{Hb}\cdot {\varepsilon }_{785}^{Hb{O}_{2}}}$$
(6)
$${\rm{\Delta }}[Hb]=\frac{{\varepsilon }_{850}^{Hb{O}_{2}}\cdot {\rm{\Delta }}O{D}_{850}-{\varepsilon }_{785}^{Hb{O}_{2}}\cdot {\rm{\Delta }}O{D}_{785}}{{\varepsilon }_{785}^{Hb}\cdot {\varepsilon }_{850}^{Hb{O}_{2}}-{\varepsilon }_{850}^{Hb}\cdot {\varepsilon }_{785}^{Hb{O}_{2}}}$$
(7)

In this study, the concentration changes of Δ[HbO2] and Δ[Hb] were used to analyze the oxidative metabolism of vocal cords before and after vocal loading task.

### Acoustic assessment with optical microphone

In this function, the backscattered optical signal of the independent laser diode (785 nm) from the tissue surface vibration was detected with a sampling rate of 50 kHz during subjects voiced the sustained vowel /a/. The sustained vowels /a/, /u/, /o/, and /i/ are usually used in clinical acoustic assessments. However, the sustained vowel /a/ would enhance measurement reliability43. The spectrum distribution was calculated with the algorithm of fast Fourier transform (FFT). Additionally, the software of Multi-Dimensional Voice Program (MDVP) was utilized to analyze voice parameters included average fundamental frequency, frequency perturbation (Jita, Jitt, PPQ, and RAP), amplitude perturbation (ShdB, Shim, and APQ), and noise-to-harmonic ratio (NHR)52,53,54,55. The choice of these voice parameters were designated by the clinician. The description of each of the extracted parameter was listed as follows: Average fundamental frequency (f0, Hz): f0 is an average value of all extracted period-to-period fundamental frequency value that is calculated from the extracted period-to-period pitch data; Absolute jitter (Jita, µsec): To calculate the period-to-period variability of the pitch period that dependent on the average fundamental frequency. Therefore, Jita is typically related to hoarse voices; Jitter percent (Jitt, %): Jitter percent is a relative measure of a very short-term variability of the pitch period. It also related to hoarse voices. The higher value of Jita and Jitt means unstable of voice quality; Pitch perturbation quotient (PPQ, %): PPQ is a relative measure of pitch perturbation in short term cycle (smoothing cycle of 5 periods) of voice analysis. Hoarse and breathy voices may cause PPQ to increase; Relative average perturbation (RAP, %): RAP is a relative measure of average pitch perturbation in a short-term cycle (smoothing cycle of 3 periods) of voice analysis. Hoarse and breathy voices also may cause PPQ to increase; Shimmer in dB (ShdB, dB): To calculate in dB of the very short term variability of the peak-to-peak amplitude of the voice that is very sensitive to amplitude variations between consecutive pitch periods. It is also related to hoarse and breathy voices; Shimmer percent (Shim, %): Shim is a relative measure of a very short-term variability of the peak-to-peak amplitude of the voice. It is typically related to hoarse and breathy voices; Amplitude perturbation quotient (APQ, %): APQ is a relative measure of a short term (smoothing cycle-to-cycle of 11 periods) irregularity of the peak-to-peak amplitude of the voice. Hoarse and breathy voices may also cause APQ to increase; Noise-to-harmonics ratio (NHR, A.U.): NHR is defined as the ratio of the energy of noise and the energy of harmonic spectral during the range of 70–4200 Hz. The lower value of NHR indicates the better voice quality. Contrarily, the higher value of NHR is interpreted as more spectral noise that may be caused by the frequency and amplitude variations, turbulent noise, subharmonic components or voice breaks.

In the both two functions, the average data was obtained for group-level analysis to reduce the effects of individual differences. The results were expressed as the mean ± SD. The significant differences analysis between healthy male and female group was made with a two-sample t-test. Additionally, the significant difference analysis before and after vocal loading task was also made with a two-sample t-test. Analyses were performed with software of LabVIEW (Version 2017, National Instruments, Austin, Texas, U.S.), MATLAB (Version 7.11.0.584 R2010b, MathWorks Inc., Natick, MA, U. S.) and Multi-Dimensional Voice Program (Version MDVP model 5150, Lincoln Park, NJ, U.S.). A more stringent p value of <0.01 was considered as statistically significant in two-sample t-test.

## Conclusion

In this study, the multi-functional vocal cords detection system was developed by using optical technique. This system can provide a novel and non-invasive detection approach for human vocal fold oxidative metabolism detection and acoustic assessment simultaneously. Our results demonstrated that the physiological analysis of vocal cords (include oxidative metabolism detection and acoustic assessment) before and after vocal loading task could be successfully measured by using an optical method. Therefore, this optical method could be a potential tool for clinical application of voice disorder. Although there are still several limitations31, fNIRS could provide relevant information on key mechanisms of oxidative metabolism of different organs by performing a clear and reliable hypothesis of the test protocol. In the future study, the system will be optimized for more clinical applications such as postoperative prognosis and monitoring. We hope that the proposed method can provide more clinical information to help ENT physician to develop the treatment strategy and therapeutic monitoring of vocal disorders.

Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

## References

1. 1.

Nanjundeswaran, C., Jacobson, B. H., Gartner-Schmidt, J. & Verdolini, A. K. Vocal fatigue index (VFI): Development and validation. J. Voice. 29, 433–440 (2015).

2. 2.

Caratya, M.-J. & Montacié, C. Vocal fatigue induced by prolonged oral reading: Analysis anddetection. Comput. Speech. Lang. 28, 453–466 (2014).

3. 3.

Costa, V. D., Prada, E., Roberts, A. & Cohen, S. Voice disorders in primary school teachers and barriers to care. J. Voice. 26, 69–76 (2015).

4. 4.

Charn, T. C. & Mok, P. K. H. Voice problems amongst primary school teachers in Singapore. J. Voice. 26, 141–147 (2012).

5. 5.

Chen, S. H., Chiang, S.-C., Chung, Y.-M., Hsiao, L.-C. & Hsiao, T.-Y. Risk factors and effects of voice problems for teachers. J. Voice. 24, 183–190 (2010).

6. 6.

Williams, N. R. Occupational groups at risk of voice disorders: a review of the literature. Occup. Med. 53, 456–460 (2003).

7. 7.

Vilkman, E. Occupational safety and health aspects of voice and speech professions. Folia. Phoniatr. Logop. 56, 220–253 (2004).

8. 8.

Yiu, E. M.-L. et al. Quantitative high-speed laryngoscopic analysis of vocal fold vibration in fatigued voice of young karaoke singers. J. Voice. 27, 753–761 (2013).

9. 9.

Schwartz, S. R. et al. Clinical practice guideline: hoarseness (dysphonia). Otolaryngol Head Neck Surg. 141, S1–S31 (2009).

10. 10.

Hanschmann, H., Gaipl, C. & Berger, R. Preliminary results of a computer-assisted vocal load test with 10-min test duration. Eur Arch Otorhinolaryngol. 268, 309–313 (2011).

11. 11.

Remacle, A., Finck, C., Roche, A. & Morsomme, D. Vocal impact of a prolonged reading task at two intensity levels: objective measurements and subjective self-ratings. J. Voice. 26, e177–e186 (2012).

12. 12.

Dehqan, A. & Scherer, R. C. Acoustic analysis of voice: Iranian teachers. J. Voice. 27, e17–e21 (2013).

13. 13.

Echternach, M., Nusseck, M., Dippold, S., Spahn, C. & Richter, B. Fundamental frequency, sound pressure level and vocal dose of a vocal loading test in comparison to a real teaching situation. Eur Arch Otorhinolaryngol. 271, 3263–3268 (2014).

14. 14.

Ingle, J. W. et al. Role of steroids in acute phonotrauma: A basic science investigation. Laryngoscope. 124, 921–927 (2014).

15. 15.

Whitling, S., Rydell, R. & Lyberg Åhlander, V. Design of a clinical vocal loading test with long-time measurement of voice. J. Voice. 29, e13–e27 (2015).

16. 16.

Ben-David, B. M. & Icht, M. Voice changes in real speaking situations during a day, with and without vocal loading: Assessing call center operators. J. Voice. 30, e1–e11 (2016).

17. 17.

Gorham-Rowan, M., Berndt, A., Carter, M. & Morris, R. The effect of a vocal loading task on vocal function before and after 24 hours of thickened liquid use. J. Speech Pathol. Ther. 1, 1–5 (2016).

18. 18.

Welham, N. V. & Maclagan, M. A. Vocal fatigue: current knowledge and future directions. J. Voice. 17, 21–30 (2003).

19. 19.

Gillivan-Murphy, P., Drinnan, M. J., O’Dwyer, T. P., Ridha, H. & Carding, P. The effectiveness of a voice treatment approach for teachers with self-reported voice problems. J. Voice. 20, 423–431 (2006).

20. 20.

Chen, S. H., Hsiao, T. Y., Hsiao, L. C., Chung, Y. M. & Chiang, S. C. Outcome of resonant voice therapy for female teachers with voice disorders: perceptual, physiological, acoustic, aerodynamic, and functional measurements. J. Voice. 21, 415–425 (2007).

21. 21.

Boucher, V. J. & Ayad, T. Physiological attributes of vocal fatigue and their acoustic effects: a synthesis of findings for a criterion-based prevention of acquired voice disorders. J. Voice. 24, 324–336 (2010).

22. 22.

Jiang, J. J. & Maytag, A. L. Aerodynamic measures of glottal function: what extra can they tell us and how do they guide management? Curr. Opin. Otolaryngol Head Neck Surg. 22, 450–454 (2014).

23. 23.

Behlau, M., Madazio, G. & Oliveira, G. Functional dysphonia: strategies to improve patient outcomes. Patient Related Outcome Measures. 6, 243–253 (2015).

24. 24.

Bale, G., Elwell, C. E. & Tachtsidis, I. From Jöbsis to the present day: a review of clinical near-infrared spectroscopy measurements of cerebral cytochrome-c-oxidase. J. Biomed. Opt. 21(091307), 1–18 (2016).

25. 25.

Scholkmann, F. et al. A review on continuous wave functional near-infrared spectroscopy and imaging instrumentation and methodology. Neuroimage. 85, 6–27 (2014).

26. 26.

Kravari, M., Angelopoulos, E., Vasileiadis, I., Gerovasili, V. & Nanas, S. Monitoring tissue oxygenation during exercise with near infrared spectroscopy in diseased populations-A brief review. Int. Jind. Ergonom. 40, 223–227 (2010).

27. 27.

Hamaoka, T., McCully, K. K., Niwayama., M. & Chance, B. Phil. The use of muscle near-infrared spectroscopy in sport, health and medical sciences: recent developments. Trans. R. Soc. 28, 4591–4604 (2011).

28. 28.

Binzoni, T. & Spinelli, L. Near-infrared photons: a non-invasive probe for studying bone blood flow regulation in humans. J. Physiol Anthropol. 34, 1–6 (2015).

29. 29.

Koga, S. et al. Effects of increased skin blood flow on muscle oxygenation/deoxygenation: comparison of time-resolved and continuous-wave near-infrared spectroscopy signals. Eur. J. Appl. Physiol. 115, 335–343 (2015).

30. 30.

Seong, M. et al. Simultaneous blood flow and blood oxygenation measurements using a combination of diffuse speckle contrast analysis and near-infrared spectroscopy. J. Biomed Opt. 21(027001), 1–6 (2016).

31. 31.

Grassia, B. & Quaresima, V. Near-infrared spectroscopy and skeletal muscle oxidative function in vivo in health and disease: a review from an exercise physiology perspective. J. Biomed Opt. 21(091313), 1–20 (2016).

32. 32.

Bucaroa, J. A. & Lagakos, N. Lightweight fiber optic microphones and accelerometers. Rev. Sci. Instrum. 72, 2816–2821 (2001).

33. 33.

Kadirvel, K. et al. 42nd AIAA 2004-1310 (2004).

34. 34.

NessAiver, M. S., Stone, M., Parthasarathy, V., Kahana, Y. & Paritsky, A. Recording high quality speech during tagged cine-MRI studies using a fiber optic microphone. J. Magn Reson Imaging. 23, 92–97 (2006).

35. 35.

Teixeira, J. G. V., Leite, I. T., Silva, S. & Frazão., O. Advanced fiber-optic acoustic sensors. Photonic sensors. 4, 198–208 (2014).

36. 36.

Gillespie, A. I., Dastolfo, C., Magid, N. & Gartner-Schmidt, J. Acoustic analysis of four common voice diagnoses: moving toward disorder-specific assessment. J. Voice. 28, 582–588 (2014).

37. 37.

Teixeira, J. P. & Fernandes, P. O. Acoustic analysis of vocal dysphonia. Procedia Computer Science. 65, 466–473 (2015).

38. 38.

Batra, K., Bhasin, S. & Singh, A. Acoustic analysis of voice samples to differentiate healthy and asthmatic persons. IJECS. 4, 13161–13164 (2015).

39. 39.

Lin, F. C., Chen, S. H., Chen, S. C., Wang, C. T. & Kuo, Y. C. Correlation between acoustic measurements and self-reported voice disorders among female teachers. J. Voice. 30, 460–465 (2016).

40. 40.

Gillespie, A. I., Gartner-Schmidt, J., Lewandowski, A. & Awan, S. N. An examination of pre- and posttreatment acoustic versus auditory perceptual analyses of voice across four common voice disorders. J. Voice. Accepted for publication April 19, 2017. Article in press (2017).

41. 41.

Ting, H. N., Chia, S. Y., Abdul Hamid, B. & Mukari, S. Z. Acoustic characteristics of vowels by normal Malaysian Malay young adults. J. Voice. 25, e305–e309 (2011).

42. 42.

Ting, H. N., Chia, S. Y., Kim, K. S., Sim, S. L. & Abdul Hamid, B. Vocal fundamental frequency and perturbation measurements of vowels by normal Malaysian Chinese adults. J. Voice. 25, e311–e317 (2011).

43. 43.

Brockmann, M., Drinnan, M. J., Storck, C. & Carding, P. N. Reliable jitter and shimmer measurements in voice clinics: the relevance of vowel, gender, vocal intensity, and fundamental frequency effects in a typical clinical task. J. Voice. 25, 44–53 (2011).

44. 44.

Yamauchi, A. et al. Age- and gender-related difference of vocal fold vibration and glottal configuration in normal speakers: analysis with glottal area waveform. J. Voice. 28, 525–531 (2014).

45. 45.

Yamauchi, A. et al. Quantitative analysis of digital videokymography: a preliminary study on age- and gender-related difference of vocal fold vibration in normal speakers. J. Voice. 29, 109–119 (2015).

46. 46.

Lovato, A. et al. Multi-dimensional voice program (MDVP) vs Praat for assessing euphonic subjects: A preliminary study on the gender-discriminating power of acoustic analysis software. J. Voice. 30, 765.e1–765.e5 (2015).

47. 47.

Su, M. C. et al. Measurement of adult vocal fold length. J. Laryngol Otol. 116, 447–449 (2002).

48. 48.

Borkowska, B. & Pawlowski, B. Female voice frequency in the context of dominance and attractiveness perception. Anim. Behav. 82, 55–59 (2011).

49. 49.

Fang, R., Jiang, J. J., Smith, B. L. & Wu, D. Expression of hypoxia inducible factor-1α and vascular endothelia growth factor in vocal polyps. Laryngoscope. 123, 2184–2188 (2013).

50. 50.

Boas, D. A. et al. The accuracy of near infrared spectroscopy and imaging during focal changes in cerebral hemodynamics. Neuroimage. 13, 76–90 (2001).

51. 51.

Kocsis, L., Herman, P. & Eke, A. The modified Beer-Lambert law revisited. Phys. Med. Biol. 51, N91–N98 (2006).

52. 52.

Kent, R. D., Vorperian, H. K., Kent, J. F. & Duffy, J. R. Voice dysfunction in dysarthria: application of the Multi-Dimensional Voice Program. J. Commun. Disord. 36, 281–306 (2003).

53. 53.

Nicastri, M., Chiarella, G., Gallo, L. V., Catalano, M. & Cassandro, E. Multidimensional Voice Program (MDVP) and amplitude variation parameters in euphonic adult subjects. Normative study. Acta. Otorhinolaryngol Ital. 24, 337–341 (2004).

54. 54.

Godino-Llorente, J. I. et al. Acoustic analysis of voice using WPCVox: a comparative study with Multi Dimensional Voice Program. Eur. Arch. Otorhinolaryngol. 265, 465–476 (2008).

55. 55.

Maryn, Y., Corthals, P., De Bodt, M., Van Cauwenberge, P. & Deliyski, D. Perturbation measures of voice: a comparative study between Multi-Dimensional Voice Program and Praat. Folia. Phoniatr. Logop. 61, 217–226 (2009).

## Acknowledgements

This work was supported in part by the Taiwan National Science Council under Grant Nos. MOST 104-2221-E-009 -192 -MY3, and a grant from Ministry of Education, Aim for the Top University Plan in National Chiao-Tung University 106W970.

## Author information

### Affiliations

1. #### Institute of Biomedical Engineering, National Chiao Tung University, Hsinchu, 30010, Taiwan

• Jung-Chih Chen
•  & Ching-Cheng Chuang
2. #### Department of Otolaryngology, China Medical University Hospital, Taichung, 40447, Taiwan

• Tzu-Chieh Lin
•  & Yung-An Tsou
3. #### Department of Otolaryngology, Hsinchu Cathay General Hospital, Hsinchu, 30060, Taiwan

• Chih-Hsien Liu
4. #### Department of Electrical Engineering, National United University, Miaoli, 36063, Taiwan

• Chia-Yen Lee

### Contributions

C.-C. Chuang and T.-C. Lin wrote the main manuscript text; J.-C. Chen and C.-H. Liu recruited the participants and collected the data; C.-C. Chuang and C.-Y. Lee developed the system and algorithm for signal processing and prepared the figures; C.-C. Chuang drew the Figure. 7; Y.-A. Tsou collected the literatures to provide discussion and clinical advice; C.-C. Chuang also performed the statistical analyses. All authors have reviewed and agreed to all of the content in the manuscript.

### Competing Interests

The authors declare that they have no competing interests.

### Corresponding author

Correspondence to Ching-Cheng Chuang.