Age and Sex Comparison of Aerodynamic Phonation Measurements Using Noninvasive Assessment

Abstract

Purpose

The goal of this study was to present vocal aerodynamic measurements from pediatric and adult participant pools. There are a number of anatomical changes involving the larynx and vocal folds that occur as children age and become adults. Data were collected using two methods of noninvasive aerodynamic assessment: mechanical interruption and labial interruption.

Method

A total of 154 participants aged 4–24 years old took part in this study. Ten trials were performed for both methods of airway interruption. To perform mechanical interruption, participants phonated /α/ for 10 s trials while a balloon valve interrupted phonation 5 times. For labial interruption, participants said /pα/ 5 times at comfortable and quiet volumes. Aerodynamic measures included subglottal pressure, phonation threshold pressure, mean airflow, laryngeal resistance, and others.

Results

One hundred one participants (51 females) successfully completed testing with both methods. Eight out of 20 measurements were found to have a statistically significant effect of participant age on measurements. Sex alone had a significant effect on vocal efficiency for the labial quiet method.

Conclusions

The data discussed here can be used to view age and sex trends in vocal aerodynamic measurements. When using either method of mechanical or labial interruption, participant age needs to be taken into account to properly interpret several aerodynamic parameters. A participant's sex is not as important when using these methods.

Approximately 1.4% of children between the ages of 3 and 17 years had reported a voice problem over the previous 12 months in a 2012 survey of a public health database (Black et al., 2015). In a survey of adults, nearly one third reported a voice disorder at some point in their lifetime, including 6.6% who indicated they were dealing with a vocal problem at the time of the survey (Roy et al., 2005). With a large number of people developing a vocal disorder during their lives, an easy way to detect vocal disorders is needed to effectively treat patients. Vocal efficiency (VE) is the ratio of acoustic output to aerodynamic input (Titze, 1992) and is a value that tends to decrease with the presence of a vocal disorder such as vocal nodules, polyps, edema, nerve paralysis, or cancer (Jiang et al., 2004; Tanaka & Gould, 1985). This study aims to find normative values for the following aerodynamic parameters that are used for calculating VE to determine any age- or sex-related trends in the data of healthy participants so that it may be compared to disordered patients in order to help diagnose disorders in the future.

The aerodynamic input to the VE calculation is aerodynamic power (PW; also called phonation power). PW is the product of subglottal pressure (Ps), which is the driving force behind phonation, and mean airflow rate (MFR), the rate at which air moves through the vocal tract. Laryngeal resistance (LR) can be calculated as Ps divided by MFR. This is similar to Ohm's law for an electric circuit, where resistance equals voltage over current. LR is affected by the size and shape of the airway as well as the degree of vocal fold adduction (Netsell et al., 1991). It has been shown to increase in patients who have been diagnosed with muscle tension dysphonia (Zheng et al., 2012). Increased Ps has been measured in adults with disorders, such as nodules, polyps, edema, lesions, vocal fold thickening, and muscle tension dysphonia (Gillespie et al., 2013; Giovanni et al., 2000; Yiu et al., 2004). Phonation threshold flow (PTF), a measure of MFR at the onset of phonation, has been used to differentiate patients with mass lesions and vocal fold mobility disorders from healthy controls but was not able to individually specify the individual disorders (Zhuang et al., 2009). Corresponding to PTF is phonation threshold power (PTW), the PW at phonation onset. PTW has been used to differentiate patients with polyps from healthy controls (Zhuang et al., 2013).

The studies mentioned above have shown the value in the usage of aerodynamic measurements for determining the presence of vocal disorders. This type of capability could prove useful for monitoring vocal disorders during treatment to ensure that treatment is working properly. However, in order to compare disordered subjects to healthy ones, a database of normal, healthy parameters is required. In this study, two noninvasive measurement methods using airway interruption were used to acquire measurements of healthy individuals. The first technique was the labial method, where the participant repeated the labial plosive /pα/ (Smitheran & Hixon, 1981). This is the operating principle of devices such as the Phonatory Aerodynamic System (PENTAX Medical). The other method uses complete airway interruption by use of a mechanically controlled balloon valve—mechanical interruption (Jiang et al., 1999).

Another advantage of this study employing two measurement methods is the chance to find any trends in vocal fold hysteresis. Vocal fold hysteresis is the ratio of offset phonation threshold pressure (PTP) to onset PTP (Lucero, 1999). This is essentially the ratio of phonatory energy required to continue phonating against the energy required to start phonating. Between the two methods studied here, both onset and offset PTPs are measurable. The labial method can be used to measure onset PTP, and the mechanical method can measure offset PTP. Theoretical calculations (Lucero, 1999) and excised larynx models (Regner et al., 2008) have been used to show that onset PTP should be higher than offset PTP, with a ratio of offset/onset PTP being somewhere between 0.5 and 1. In other words, it should take less energy to keep the vocal folds in motion than it should to make them start vibrating. However, a study done by Plant et al. (2004) showed wide variability in measurements, including participants with hysteresis ratios greater than 1.

Another important note from the above studies is that they almost exclusively studied the adult population. However, from birth until adulthood, the vocal folds are in a constant state of development (Eckel et al., 1999; Hartnick et al., 2005). At birth, the vocal folds resemble an acellular monolayer and become a bilayered structure within the first several months of life. A trilaminar structure develops by 7 years of age, and the molecular composition begins to resemble an adult vocal fold around 13 years of age (Boudoux et al., 2009; Hartnick et al., 2005). Also near this age, the vocal ligament begins to develop (Hartnick et al., 2005). Previous research done with children noticed potential age-related trends in pressure measurements (Hoffman et al., 2019) as well as MFR and LR (Scholp et al., 2019). These last two studies indicate that there could be age-related trends that are missing due to a lack of studies that cover the entire developmental spectrum.

The goal of this study is to present measurements of aerodynamic variables measured noninvasively across a range of ages from children to young adults, while also comparing males and females. Previous research has done limited work finding age-related trends in the pediatric population. This study will expand on that knowledge, while also comparing the pediatric results to those of young adults in order to measure the entire developmental spectrum from ages 4 to 24 years. Due to the vocal folds transitioning from a monolayer to a trilaminar structure and the development of the vocal ligament, it is expected that there will be several age-related trends within this data set. In addition, the data will also be split into male and female groups to determine if any of these developmental changes are also affected by sex.

Method

Study Population

This study was approved by the institutional review board at the University of Wisconsin–Madison, and all participants provided written consent for their participation. For pediatric participants, written consent was obtained from both the participant and their parent/guardian. The study was performed at the University of Wisconsin Laryngeal Physiology Laboratory between July 2017 and March 2020. A total of 154 people were recruited. Their ages ranged from 4 to 24 years old (M = 13.7 ± 6.3 years). Participants were recruited as either a pediatric (ages 4–17 years) or an adult (ages 18+ years) volunteer. For the entire study population, the minimum age allowed to participate was 4 years old, and the oldest individual included in the study was 24 years old. Adult subjects were asked to complete a phone screening prior to study enrollment to ensure no history of dysphonia or smoking habits. Parents/guardians completed phone screenings for pediatric participants. Prior to testing, all pediatric participants performed the Consensus Auditory-Perceptual Evaluation of Voice protocol (American Speech-Language-Hearing Association, 2002), and the audio samples were analyzed by a speech-language pathologist to confirm vocal health.

Data Collection

A custom airflow interruption device based on Jiang et al.'s (1999) work and an updated version used in recent studies (Hoffman et al., 2019; Lamb et al., 2020; Scholp et al., 2019) were used for all data collection. The device used was a polyvinyl chloride (PVC) tube containing several sensors. To record audio signals, an AKG CK77 condenser microphone (Harman International Industries, Inc.) was connected to an MP13 preamplifier (Rolls Co.). A pneumotach amplifier (Series 1110, Hans Rudolph, Inc.) was used to measure pressure and airflow. To perform airflow interruptions, a balloon valve (Series 9340, Hans Rudolph, Inc.) was attached to one end of the PVC tube. The data were recorded at 10 kHz using a custom LabVIEW program via a USB-6341 data acquisition board (National Instruments, Inc.). The entire setup can be seen in Figure 1.

Figure 1.

Depiction of the interruption device. The mouthpiece or facemask is placed on the left end, and airflow moves from left to right through the tube. NI DAQ Board = National Instruments Data Acquisition Board.

Labial Method

An intraoral tube 0.066 in. in diameter is extended from the pressure measurement port of the PVC tube to the participant's mouth. The distal end of the intraoral tube ran just past their lips, reaching the tip of their tongue. A silicone anesthesia mask was connected to the distal end (the end opposite the balloon valve) of the PVC tube. Participants were instructed to hold this mask tightly to their face in order to create an airtight seal. For pediatric participants, parents were allowed to help hold the mask in place if necessary. There were multiple-sized masks available as well. To perform this task, participants were instructed to say the labial plosive /pα/. First, they did this at a comfortable volume (to measure Ps). Then, they repeated the task but were told to phonate “as quietly as possible, but without whispering” (to measure PTP). Participants performed 10 trials of each variation, saying “pa” 5 times per trial. During this phase of testing, the balloon valve was not used and stayed deflated.

Mechanical Method

For this method, there was no intraoral tube, and instead, a silicon mouthpiece (similar to a snorkel or athletic mouthguard) was attached to the distal end of the PVC tube. Participants placed their lips around this mouthpiece and were instructed to plug their nose. They phonated a constant /α/ for approximately 10 s at a comfortable volume for each trial. During the trial, the balloon inflated 5 times for 250 ms, completely occluding the airway and interrupting phonation. Ps and PTP were measured during this interruption. To ensure consistent interruptions, balloon inflation was automated by a computer that works on the order of gigahertz.

In order to maintain consistency across trials and methods, participants were provided visual feedback of their sound pressure level (SPL). The microphone-to-mouth distance was 45 mm for the mechanical method and 65 mm for the labial method. To account for learning bias, trial order was randomized. During practice trials for whichever method they did first, participants established a baseline SPL, and that was used as a reference for the remaining trials. No reference SPL was required to perform labial at quiet volume, since that was expected to have a lower SPL than the mechanical or normal volume labial trials.

Data Analysis

Custom LabVIEW and MATLAB programs were used to analyze the raw data, as done previously (Hoffman et al., 2019). For trials performed with the labial method, Ps and PTP were recorded as the maximum pressure achieved prior to phonation of the plosive for both comfortable and quiet volumes. MFR was averaged during the vowel phase of the /pα/. For trials performed with the mechanical method, Ps was marked as the pressure 150 ms after the onset of the interruption (Hoffman et al., 2009), and PTP was measured as Ps minus the pressure measured at phonation offset. MFR was averaged as the airflow during the constant phonation between interruptions.

To determine LR, a participant's average Ps across all trials was divided by their average MFR, again from across all trials. PW was calculated the same way, except by multiplying Ps and MFR. PTW uses the same calculation, but specifically from the quiet labial trials. To calculate the hysteresis ratio, offset PTP, as measured with mechanical interruption, was divided by onset PTP, as measured with the quiet labial interruption. SPL was calculated based on the audio signal recorded via the microphone; VE was calculated by the following formula: $\mathrm{VE}=\frac{4\pi r^2{I}_0{10}^{\mathrm{SPL}/10}}{\mathrm{Ps}\times \mathrm{MFR}}$ , where r is the microphone-to-mouth distance and I ₀ is the standard reference intensity (10⁻¹² W/m²; Titze, 1994). A snapshot of all the included variables and their abbreviations is in Table 1. Everything measured by the quiet labial method is at phonation onset. While mechanical interruption is able to measure offset PTP, it is unable to measure PTF or PTW. The labial method performed at a comfortable volume is unable to measure any of the phonation threshold parameters.

Table 1.

The different variables that were measured in this study.

Parameter	Measurement	Mechanical	Labial comfortable	Labial quiet
Subglottal pressure	Pressure during normal phonation	mPs	cPs	—
Phonation threshold pressure	Pressure taken at phonation onset or offset	mPTP	—	qPTP
Mean flow rate	Airflow during phonation	mMFR	cMFR	—
Phonation threshold flow	Airflow at phonation onset or offset	—	—	PTF
Aerodynamic power	Ps × MFR	mPW	cPW	—
Phonation threshold power	Power at phonation onset or offset	—	—	PTW
Laryngeal resistance	Ps / MFR	mLR	cLR	qLR
Hysteresis ratio	mPTP / qPTP	—	—	—
Sound pressure level	Phonation intensity	mSPL	cSPL	qSPL
Vocal efficiency	Ratio of radiated acoustic power to aerodynamic input	mVE	cVE	qVE

Note. Labial quiet by definition measures everything at phonation onset. While mechanical interruption can measure PTP, the method does not allow for measurement of PTF or PTW. Since labial comfortable is always performed at a comfortable volume, it does not measure any of the phonation threshold parameters.

Statistical Analysis

SPSS Version 26 (IBM) was used for all statistical analysis. Multifactor analysis of variance (ANOVA) with post hoc Tukey's tests was used to determine any age, sex, or Age x Sex effects. Linear regression was performed with each of the aerodynamic parameters as a dependent variable and age as the independent variable in order to quantify the trend of each parameter with respect to age. To determine if there were any differences in parameters based on measurement method, paired t tests were used to compare Ps, MFR, LR, PW, and VE for mechanical and labial methods at a comfortable volume. A significance level of α = .05 was used. A total of 101 participants were used in the data analysis. Four pediatric participants were excluded for having a dysphonic rating on the Consensus Auditory-Perceptual Evaluation of Voice protocol. Thirty-four pediatric participants and seven adult participants were excluded for having incomplete data on either the mechanical and/or labial interruption protocol. The excluded participants totaled 53. A final age and sex breakdown of the included participants is in Table 2.

Table 2.

Breakdown of counts for the final study population by age and sex.

Age group (years)	Male	Female	Total
4–7	6	8	14
8–13	14	13	27
14–17	8	6	14
18–24	22	24	46
Total	50	51	101

Results

Of the 20 parameters examined in this study, eight of them showed a significant effect with respect to age with the ANOVA testing. One parameter (qVE) had a significant effect with respect to sex. Two parameters (cLR, cVE, and qVE) had a significant effect of Age x Sex. A full summary of results is in Table 3, and summary graphs are included in Figures 2 –9. For the regression analysis, the following parameters generally increased with respect to participant age: mMFR, cMFR, cPW, cSPL, and PTF. The next ones generally decreased: mPs, mPTP, mLR, cLR, qPTP, and qLR. These last 9 showed no change: mPW, mSPL, mVE, cPs, cVE, PTW, qSPL, qVE, and hysteresis ratio. mPs, mPW, cPs, cPW, and PTW did not have a significant age effect with the ANOVA test but showed significant trends with the regression tests. cVE and qVE had significant effects due to age in the ANOVA test but showed no significant trends with respect to age in the regression analysis. For results of all statistical tests and raw data based on age, please see the Appendix.

Table 3.

Summary of results of the statistical tests.

Variable	ANOVA			Regression
	Age	Sex	Age x Sex	Male	Female	Total
Mechanical interruption				−		−
mPs	X			−		−
mPTP	X			+	+	+
mMFR					+	+
mPW	X			−	−	−
mLR
mSPL
mVE
Labial comfortable					−	−
cPs	X			+	+	+
cMFR				+		+
cPW	X		X	−	−	−
cLR				+
cSPL			X
cVE
Labial quiet				−	−	−
qPTP	X			+	+	+
PTF						+
PTW	X			−	−	−
qLR
qSPL	X	X	X
qVE					+
Hysteresis

Note. Blank boxes had nonsignificant results. The linear regression was performed with the entire subject pool with respect to subject age. A significance level of α = .05 was used for statistical testing. X indicates a significant result in the ANOVA testing, + indicates a significant increasing trend with respect to age, and − indicates a significant decreasing trend with respect to age. Ps = subglottal pressure; PTP = phonation threshold pressure; MFR = mean flow rate; PW = aerodynamic power; LR = laryngeal resistance; SPL = sound pressure level; VE = vocal efficiency; PTF = phonation threshold flow; PTW = phonation threshold power.

Figure 2.

Subglottal pressure (Ps) versus participant age. cPs = labial measurements at a comfortable volume; mPs = mechanical measurements.

Figure 3.

Sound pressure level (SPL) versus participant age. cSPL = labial measurements at a comfortable volume; mSPL = mechanical measurements; qSPL = labial measurements at a quiet volume.

Figure 4.

Phonation threshold pressure (PTP) versus participant age. mPTP = mechanical measurements; qPTP = labial measurements at a quiet volume.

Figure 5.

Laryngeal resistance (LR) versus participant age. cLR = labial measurements at a comfortable volume; mLR = mechanical measurements; qLR = labial measurements at a quiet volume.

Figure 6.

Mean airflow rate (MFR) versus participant age. cMFR = labial measurements at a comfortable volume; mMFR = mechanical measurements; PTF = phonation threshold flow, measured with labial at a quiet volume.

Figure 7.

Aerodynamic power (PW) versus participant age. cPW = labial measurements at a comfortable volume;mPW = mechanical measurements; PTW = phonation threshold power, measured with labial at a quiet volume.

Figure 8.

Vocal efficiency (VE) versus participant age. cVE = labial measurements at a comfortable volume; mVE = mechanical measurements; qVE = labial measurements at a quiet volume.

Figure 9.

Hysteresis ratio of mechanical phonation threshold pressure (mPTP) to labial PTP at a quiet volume (qPTP) plotted against participant age.

Table 4 shows the data comparing measurements of Ps, MFR, LR, PW, and VE between the mechanical and labial comfortable methods. Ps, MFR, and LR were all significantly different when measured by the two methods. PW and VE did not have a significantly different result.

Table 4.

Comparison of subglottal pressure (Ps), mean airflow rate (MFR), laryngeal resistance (LR), aerodynamic power (PW), and vocal efficiency (VE%) for the mechanical and labial comfortable methods of interruption.

Variable	Mechanical	Labial	t	p
Ps (cm H2O)	6.44 ± 2.10	7.07 ± 2.20	2.24	.0273*
MFR (L/min)	11.07 ± 4.85	9.28 ± 4.81	3.86	< .001**
LR (cm H2O/L/min)	0.69 ± 0.38	1.08 ± 0.87	4.91	< .001**
PW (cm H2O L/min)	72.98 ± 51.57	65.97 ± 43.72	1.48	.1423
VE% × 1,000	2.30 ± 8.41	3.27 ± 12.05	1.27	.2098

Note. VE% is presented as the value multiplied by 1,000 to avoid very small decimals with many leading zeros.

^*Significance at α = .05.

^**Significance at α = .01.

Discussion

The goal of this study was to find any age-related trends for a variety of aerodynamic pressures that could be obtained through noninvasive airflow interruption techniques. To ensure that the values measured here were reasonable, we compared our results to some previous research and found they were comparable (see Table 5; Awan et al., 2013; Goozée et al., 1998; Netsell et al., 1994; Weinrich et al., 2013). This comparison was kept to Ps, MFR, LR, and PW because those parameters have more prior research. Age-related effects will be discussed below. Because there were no significant effects for sex and only cLR had a significant effect for Age x Sex, the majority of the discussion will focus on participant age.

Table 5.

General ranges of normal values for subglottal pressure, mean flow rate, laryngeal resistance, and aerodynamic power compared to the male and female averages from this study.

Measurement	Normal range	Parameter	Male	Female
Subglottal pressure (cm H2O)	5–10	mPs	6.21 ± 2.38	6.67 ± 2.10
		cPS	7.29 ± 2.17	6.85 ± 2.25
Mean flow rate (L/min)	10–18	mMFR	11.25 ± 4.77	10.90 ± 5.01
		cMFR	10.10 ± 5.09	8.41 ± 9.28
Laryngeal resistance (cm H2O/L/min)	0.3–1.3	mLR	0.65 ± 0.40	0.72 ± 0.37
		cLR	1.06 ± 0.96	1.11 ± 0.78
Aerodynamic power (cm H2O L/min)	50–180	mPW	71.27 ± 46.33	74.70 ± 57.26
		cPW	72.92 ± 47.22	58.47 ± 39.41

Note. These are not exact ranges but are general guidelines based on published research.

Ps

Overall, there were not many changes noted for Ps, measured with either the mechanical method or the labial method (see Figure 2). cPs had no significant effects for age, as well as no significant regression trends with respect to participant age. mPs only showed a negative relationship with age for the male population. There are several studies that have demonstrated Ps to have a positive relationship with SPL (Björklund & Sundberg, 2016; Traser et al., 2017). This trend was slightly observed with our data, but one of the limitations of this study was that SPL was not controlled for, which could confound any conclusions made. All the participants were allowed to phonate at their own comfortable level for the Ps calculations. Even though SPL was not explicitly controlled for, there were no significant differences noted based on participant age or sex (see Figure 3).

PTP

There was a significant age effect for mPTP. Based on the regression statistics, PTP tends to decrease as people age (see Figure 4). Both males and females had significant negative trends for qPTP. For mPTP, the negative trend was significant in male participants. While not significant, females actually had a positive slope for mPTP (r = .032). If the data are broken down by age group, the mean female group mPTP decreases from young children through older children but then increases again for adult females. This trend is not seen with qPTP or with male participants. In fact, mPTP with adults is one of the few age groups, among any of the parameters measured in this study, where there was a significant difference due to sex (p = .007). It is unclear why this would occur only with adult females and only when using mechanical interruption. Other than that outlier scenario, PTP tends to decrease with age.

LR

mLR and cLR showed a significant effect with age, and regression showed that LR decreases with age (see Figure 5). This occurs due to the increasing size of the airway during development (i.e., adults have a larger airway than children; Stathopoulos & Sapienza, 1993; Weinrich et al., 2013). The increase in area decreases the resistance. However, LR, in general, is not dependent only on glottal anatomy. This is demonstrated by the airway resistance measured for labial at a comfortable volume (cLR) being higher than the resistance measured at a quiet volume (qLR; paired t test, t = 5.88, p < .001). This indicates that the relationship between pressure and airflow is not constant and that resistance has some dependence on the level of phonation, aspects of which have been studied previously (Nishida, 1967; Van den Berg, 1956).

MFR

Both mMFR and cMFR had significant age effects for MFR, and all groups had a significant positive regression with respect to age (see Figure 6). Given that Ps did not change very much with age and because LR saw a significant decrease with age, MFR needed to increase since Ps = MFR × LR. From labial trials at a quiet volume, PTF also had a significant age effect and also increased with age, following a similar trend to the MFR measures recorded during normal, comfortable phonation.

PW

While neither mPW nor cPW had a significant age effect from the ANOVA testing, both increased with respect to age for the total study population (see Figure 7). An increase in power as age increases would be expected since PW = Ps × MFR. Ps stayed constant with regard to age, but MFR showed a significant increase; therefore, PW would also be expected to increase. At phonation threshold, PTW did not have a significant effect for age, but the regression model showed an increase in PTW as age increases. In general, PW appears to increase with participant age.

VE

It must first be noted that the VE results reported here need to be understood within the context of this study (see Figure 8). Since VE is the ratio of radiated acoustic power to input PW, both quantities are needed. In this study, acoustic power was measurable through SPL recordings; however, SPL was not explicitly controlled for during this study. The main function of the SPL recordings was to attempt to have participants be consistent across trials for the purpose of the aerodynamic measurements. Having mentioned that, the VE results of this study did not have many significant results. cVE and qVE both had significant results in the ANOVA calculations but did not show any significant trends with respect to age in the regression analysis. The r values showed a weak, negative correlation between age and VE, so while the result is not highly significant (due to the conflicting test results), it appears that VE slowly decreases with age.

Hysteresis Ratio

The only significant result for the ratio of offset/onset PTP (mPTP/qPTP) was that linear regression for all female participants showed an increasing ratio with respect to age (see Figure 9). This trend likely has to do with the interesting trend in mPTP that was noted above. As with mPTP, the hysteresis ratio decreases as pediatric participants age (not at statistical significance), but then the value increases for female adults. This trend is not seen with male participants. In fact, the average adult female ratio was greater than 1 (ratio = 1.03), indicating that offset PTP was higher than onset PTP. This would appear to violate theoretical principles of PTP calculations (Lucero, 1999). However, this discrepancy is likely due to an error during data collection. Figure 10A displays the distribution of the hysteresis ratios (mPTP/qPTP) for the different female age groups and shows that there were no statistical outliers within the adult female group when considering each age group separately. However, when looking at the entire study population, there were six outliers out of 101 participants, and five of them were in the adult female group. Figure 10B shows the ratio of Ps measurements (mPs/cPs). The Ps ratios show that there are two outliers (out of 24 adult females) for the ratio of mPs to cPs for the 18–24 age group. Given that these outliers are not present in any other age group, it is reasonable to think there were undetected errors during data collection that are affecting the overall results. We attempted to determine if other measurements in this study were affecting the hysteresis measurement by removing participants from the analysis who had an outlier value in any other parameter measurement (Ps, MFR, LR, etc.). There were 36 participants who were outliers in at least one other parameter, which left 65 participants for the final outlier-free analysis. The results of this analysis are displayed in Table 6, and it is seen that the adult female average decreases and appears more in the range of the rest of the values. Because we do not have any physiological rationale to remove these participants from the data set, these results should not be viewed out of context, and further research could be warranted into the relationship between hysteresis and other aerodynamic parameters. The adult age group was the only group to show any noticeable difference in the hysteresis ratio when comparing sexes; the three pediatric age groups had p values of 1.000 in the post hoc tests, indicating no presence of a difference. Outside of that, the hysteresis ratio does not vary much with age. This agrees with both mPTP and qPTP generally decreasing, which would keep the ratio consistent.

Figure 10.

(A) Ratio of onset/offset phonation threshold pressure (PTP) for females in each age group. This is the hysteresis ratio. (B) Ratio of mPs to cPs for females in each age group. The center line of each box represents the median, and the boxes contain the data between the first (q1) and third (q3) quartiles. Any values marked with a “+” are outliers determined as being greater than q3 + 1.5 × (q3 − q1). There were only outliers above the upper bound. Ps = subglottal pressure.

Table 6.

Comparison of hysteresis results pre- and postremoval of outliers.

Age group (years)	Hysteresis including outliers			Hysteresis with outliers removed
	Male	Female	Total	Male	Female	Total
4–7	0.73 ± 0.38	0.72 ± 0.33	0.73 ± 0.34	0.46 ± 0.01	0.74 ± 0.38	0.67 ± 0.34
8–13	0.70 ± 0.52	0.64 ± 0.28	0.67 ± 0.42	0.60 ± 0.26	0.72 ± 0.31	0.65 ± 0.28
14–17	0.65 ± 0.29	0.53 ± 0.28	0.61 ± 0.28	0.56 ± 0.23	0.68 ± 0.18	0.61 ± 0.21
18–24	0.59 ± 0.27	1.03 ± 0.53	0.82 ± 0.48	0.61 ± 0.27	0.78 ± 0.26	0.70 ± 0.27
Total	0.65 ± 0.37	0.82 ± 0.46	0.74 ± 0.42	0.59 ± 0.24	0.74 ± 0.28	0.67 ± 0.27

Note. Outliers were considered for any of the parameters measured in this study (subglottal pressure, mean airflow rate, laryngeal resistance, phonation threshold pressure, phonation threshold flow, aerodynamic power, and phonation threshold power). Outliers were calculated using a quartiles method: values less than q1 − 1.5 × (q3 − q1) or greater than q3 + 1.5 × (q3 − q1), where q1 and q3 are the first and third quartiles of the data, respectively. There were 36 participants that had at least one outlier parameter, leaving 65 participants for the final analysis.

Differences in Methods

Both methods used in this study have been validated, compared, and believed to be measuring the same quantities (Chapin et al., 2011; Jiang et al., 1999; Smitheran & Hixon, 1981). However, the results for Ps, MFR, and LR differed for the two methods. It is likely that these varied results come from differences in MFR measurements for the two methods. When using the mask, it is often more difficult to get a good, airtight seal around the mouth and nose, causing air to leak out. This would cause the MFR to decrease for the labial method, which is what was observed. Because LR = Ps / MFR, a lowered MFR would lead to an increase in LR, which is also what was observed. It is unclear what would have caused the Ps measurements to change, but with adult subjects, the Ps values have shown to be comparable (Lamb et al., 2020). It is interesting to note that the change in MFR was enough to cause a change in LR, but not PW or VE, even though those are also both reliant on MFR. There may have been some cancellation of the effects of the changing Ps as well that allowed those values to not differ as much between methods.

Study Limitations

Because participants were told to phonate comfortably for all study tasks, SPL and fundamental frequency (F0) were not controlled for in this study. The only restriction on SPL was that participants were told to keep their own SPL consistent between trials. F0 was not measured in real time and could only be calculated from the participant's audio recording after testing, and not all participants had data where a value for F0 was able to be extracted. A relationship between F0 and parameters such as Ps has been discussed previously (Ladefoged & McKinney, 1963; Titze, 1989). From the results that could be extracted, this study found fairly large ranges in F0 (210 ± 62 Hz for both mechanical and labial methods). Entire future studies could be dedicated to controlling either SPL or F0 to obtain further results, particularly in the pediatric population.

Conclusions

This study has used two methods of airflow interruption (complete airflow interruption and labial interruption) to noninvasively measure phonatory aerodynamic parameters. Ps, PTP, MFR, PTF, LR, PW, PTW, and hysteresis ratio were compared with respect to participant age and sex. Very few differences in measurements were found with respect to sex, while participant age played a role in several of the metrics. As such, participant age would appear to be more important when considering the results of aerodynamic tests than a participant's sex.

Knowing how these parameters change (or do not change) in healthy pediatric subjects, the next step is to research what trends emerge in a disordered population. Previous research has been done with adults for measurements of Ps, LR, PTF, and PTW, for example. That data can be translated to pediatric subjects now and compared to a healthy data set. Having this information will help clinicians make more informed diagnoses.

Appendix

Table A1.

Results of ANOVA testing.

Variable	Age		Sex		Age × Sex
	F	p	F	p	F	p
Mechanical interruption	1.399	.160	0.592	.444	0.886	.587
mPs	2.063	.017*	0.808	.372	0.966	.503
mPTP	2.358	.006**	0.827	.367	0.595	.867
mMFR	1.217	.275	0.113	.737	0.342	.988
mPW	4.025	< .001**	2.889	.094	1.453	.153
mLR	1.114	.366	0.100	.754	1.214	.300
mSPL	1.207	.289	0.100	.754	0.999	.464
mVE
Labial comfortable	0.850	.642	0.014	.906	1.034	.435
cPs	2.529	.005**	2.394	.128	1.129	.357
cMFR	1.036	.441	0.769	.385	1.090	.389
cPW	5.136	< .001**	0.441	.510	1.995	.038*
cLR	1.307	.222	1.403	.242	0.736	.710
cSPL	1.760	.065	1.098	.301	2.506	.016*
cVE
Labial quiet	1.719	.056	1.376	.245	1.113	.363
qPTP	2.102	.018*	0.087	.769	0.720	.736
PTF	1.777	.052	0.000	.995	1.713	.086
PTW	2.907	.001**	0.278	.601	0.844	.613
qLR	1.520	.119	1.186	.281	1.078	.398
qSPL	4.419	< .001**	7.228	.010*	7.264	< .001**
qVE
Hysteresis	1.198	.288	0.156	.694	0.378	.378

^*Significance at α = .05.

^**Significance at α = .01.

Table A2.

Results of linear regression for males and females of all ages and the entire subject population.

Variable	Male		Female		Everyone
	r	p	r	p	r	p
Mechanical interruption	−.409	.003**	−.152	.287	−.288	.003**
mPs	−.432	.002**	.032	.826	−.197	.048*
mPTP	.416	.003**	.658	< .001**	.540	< .001**
mMFR	.071	.618	.414	.003**	.263	.010*
mPW	−.562	< .001**	−.707	< .001**	−.628	< .001**
mLR	.286	.073	−.071	.660	.077	.476
mSPL	.000	.922	−.200	.193	−.151	.170
mVE
Labial comfortable	−.148	.299	−.345	.013*	−.251	.012*
cPs	.676	< .001**	.434	.005**	.549	< .001**
cMFR	.484	.001*	.276	.081	.378	< .001**
cPW	−.584	< .001**	−.528	< .001**	−.554	< .001**
cLR	.439	.005**	.000	.930	.190	.088
cSPL	−.361	.516	−.205	.222	−.170	.152
cVE
Labial quiet	−.326	.021*	−.370	.007**	−.351	< .001**
qPTP	.412	.006**	.439	.003**	.423	< .001**
PTF	.247	.111	.245	.114	.247	.022*
PTW	−.457	.002**	−.621	< .001**	−.551	< .001**
qLR	.224	.166	−.161	.297	.000	.869
qSPL	−.105	.542	−.210	.200	−.141	.225
qVE
Hysteresis	−.173	.228	.324	.020*	.105	.290

^*Significance at α = .05.

^**Significance at α = .01.

Table A3.

Comparison of parameters for each age group based on sex.

Variable	4–7 years	8–13 years	14–17 years	18–24 years
Mechanical interruption	0.668	0.975	1.000	0.504
mPs	0.992	0.999	1.000	0.022*
mPTP	1.000	0.783	0.945	0.985
mMFR	1.000	0.996	0.969	0.703
mPW	0.858	0.073	0.992	0.871
mLR	1.000	0.977	0.937	0.977
mSPL	1.000	0.997	0.509	1.000
mVE
Labial comfortable	1.000	1.000	1.000	0.802
cPs	1.000	1.000	1.000	0.023*
cMFR	1.000	1.000	1.000	0.157
cPW	0.982	0.987	0.876	0.777
cLR	1.000	1.000	0.983	0.438
cSPL	1.000	1.000	0.608	1.000
cVE
Labial quiet	1.000	0.996	0.970	0.993
qPTP	1.000	1.000	0.997	0.861
PTF	1.000	1.000	0.927	0.960
PTW	0.997	0.919	1.000	1.000
qLR	1.000	1.000	0.859	0.292
qSPL	1.000	0.971	0.945	1.000
qVE
Hysteresis	1.000	1.000	1.000	0.007*

Note.  Only mPTP for adults and cMFR for adults showed any difference in sex. All values presented are p values of post hoc ANOVA testing using Tukey's test.

^*Significance at α = .05.

Table A4.

Summary statistics for the main metrics of the study broken down by age group and gender.

Group	mPs (cm H2O)			cPs (cm H2O)
	Male	Female	Total	Male	Female	Total
4–7	8.78 ± 4.13	6.94 ± 1.43	7.73 ± 2.92	7.39 ± 1.82	7.77 ± 2.45	7.60 ± 2.13
8–13	6.67 ± 1.71	7.42 ± 2.14	7.03 ± 1.93	7.71 ± 2.03	7.61 ± 2.30	7.66 ± 2.12
14–17	5.87 ± 2.93	5.45 ± 1.23	5.74 ± 2.22	7.59 ± 1.35	7.65 ± 2.83	7.59 ± 1.94
18–24	5.34 ± 1.28	6.49 ± 1.70	5.94 ± 1.61	6.89 ± 2.59	5.92 ± 1.72	6.39 ± 2.22
Total	6.21 ± 2.38	6.67 ± 1.80	6.44 ± 2.10	7.29 ± 2.17	6.85 ± 2.25	7.07 ± 2.20
	mMFR (L/min)			cMFR (L/min)
Group	Male	Female	Total	Male	Female	Total
4–7	6.67 ± 2.00	6.53 ± 1.46	6.60 ± 1.69	3.92 ± 2.79	4.04 ± 0.84	3.97 ± 2.04
8–13	10.42 ± 4.01	7.83 ± 1.95	9.13 ± 3.36	7.48 ± 2.85	7.28 ± 4.89	7.38 ± 3.79
14–17	12.07 ± 2.76	9.34 ± 2.71	10.82 ± 2.96	8.03 ± 4.87	9.04 ± 1.17	8.55 ± 3.43
18–24	12.72 ± 5.45	13.86 ± 5.31	13.32 ± 5.35	13.63 ± 3.88	9.50 ± 4.77	11.52 ± 4.79
Total	11.25 ± 4.77	10.90 ± 5.01	11.07 ± 4.85	10.10 ± 5.09	8.41 ± 4.45	9.28 ± 4.81
	mPW (cm H 2 O L/min)			cPW (cm H 2 O L/min)
Group	Male	Female	Total	Male	Female	Total
4–7	61.49 ± 34.24	46.21 ± 15.01	54.54 ± 27.20	29.98 ± 21.82	34.94 ± 7.96	32.18 ± 16.39
8–13	71.79 ± 44.20	57.03 ± 18.65	64.41 ± 34.08	55.79 ± 32.09	53.10 ± 39.45	54.58 ± 34.63
14–17	79.49 ± 58.70	50.13 ± 16.44	66.41 ± 43.72	62.61 ± 43.58	68.01 ± 23.81	65.30 ± 33.20
18–24	71.02 ± 48.72	96.34 ± 73.64	84.23 ± 63.58	95.55 ± 49.17	62.34 ± 45.42	78.56 ± 49.65
Total	71.27 ± 46.33	74.70 ± 57.26	72.98 ± 51.57	72.92 ± 47.22	58.47 ± 39.41	65.97 ± 43.72
	mLR (cm H 2 O/L/min)			cLR (cm H 2 O/L/min)
Group	Male	Female	Total	Male	Female	Total
4–7	1.34 ± 0.51	1.10 ± 0.25	1.23 ± 0.42	2.74 ± 1.54	2.31 ± 1.03	2.55 ± 1.28
8–13	0.68 ± 0.21	1.02 ± 0.41	0.85 ± 0.36	1.08 ± 0.38	1.34 ± 0.80	1.20 ± 0.60
14–17	0.52 ± 0.18	0.65 ± 0.33	0.59 ± 0.25	1.38 ± 1.08	0.87 ± 0.37	1.14 ± 0.81
18–24	0.50 ± 0.30	0.50 ± 0.15	0.50 ± 0.23	0.54 ± 0.26	0.86 ± 0.56	0.71 ± 0.47
Total	0.65 ± 0.40	0.72 ± 0.37	0.69 ± 0.38	1.06 ± 0.96	1.11 ± 0.78	1.08 ± 0.87
	mPTP (cm H 2 O)			qPTP (cm H 2 O)
Group	Male	Female	Total	Male	Female	Total
4–7	3.41 ± 2.09	2.97 ± 1.11	3.16 ± 1.55	4.51 ± 0.76	4.42 ± 1.75	4.46 ± 1.37
8–13	2.50 ± 0.98	2.74 ± 1.01	2.62 ± 0.98	4.19 ± 1.34	4.52 ± 1.14	4.35 ± 1.24
14–17	2.03 ± 1.02	1.83 ± 0.64	1.99 ± 0.83	3.28 ± 0.80	3.95 ± 1.55	3.59 ± 1.14
18–24	1.86 ± 0.56	2.88 ± 1.08	2.39 ±1.00	3.49 ± 1.03	3.21 ± 1.21	3.34 ± 1.13
Total	2.25 ± 1.11	2.74 ± 1.05	2.50 ± 1.10	3.77 ± 1.13	3.82 ± 1.42	3.80 ± 1.27
	PTF (L/min)			PTW (cm H 2 O L/min)
Group	Male	Female	Total	Male	Female	Total
4–7	4.58 ± 1.88	3.57 ± 0.94	4.02 ± 1.43	20.87 ± 8.23	16.43 ± 7.12	18.40 ± 7.50
8–13	7.18 ± 4.43	7.77 ± 6.49	7.49 ± 5.38	32.79 ± 29.84	32.26 ± 25.78	32.54 ± 27.29
14–17	7.27 ± 4.03	9.11 ± 5.84	8.16 ± 4.63	23.91 ± 13.56	41.67 ± 45.20	32.27 ± 30.81
18–24	11.02 ± 6.05	8.97 ± 3.47	9.97 ± 4.95	41.23 ± 28.49	32.57 ± 23.57	36.80 ± 26.15
Total	8.83 ± 5.47	8.08 ± 4.71	8.46 ± 5.06	34.36 ± 26.26	31.89 ± 26.66	33.14 ± 26.18
	qLR (cm H 2 O/L/min)			Hysteresis (offset/onset PTP)
Group	Male	Female	Total	Male	Female	Total
4–7	1.27 ± 0.81	1.51 ± 1.07	1.40 ± 0.91	0.73 ± 0.38	0.72 ± 0.33	0.73 ± 0.34
8–13	0.78 ± 0.51	1.07 ± 0.84	0.92 ± 0.69	0.70 ± 0.52	0.64 ± 0.28	0.67 ± 0.42
14–17	0.69 ± 0.59	0.63 ± 0.59	0.66 ± 0.54	0.65 ± 0.29	0.53 ± 0.28	0.61 ± 0.28
18–24	0.43 ± 0.36	0.40 ± 0.24	0.41 ± 0.30	0.59 ± 0.27	1.03 ± 0.53	0.82 ± 0.48
Total	0.64 ± 0.54	0.71 ± 0.69	0.68 ± 0.61	0.65 ± 0.37	0.82 ± 0.46	0.74 ± 0.42
	mSPL (dB)			cSPL (dB)
Group	Male	Female	Total	Male	Female	Total
4–7	70.70 ± 2.57	70.79 ± 5.85	70.75 ± 4.42	61.12 ± 3.47	64.16 ± 8.42	62.81 ± 6.52
8–13	71.58 ± 8.40	74.46 ± 7.63	73.18 ± 7.88	65.15 ± 10.46	66.20 ± 10.20	65.71 ± 10.01
14–17	69.07 ± 4.78	73.49 ± 10.53	71.25 ± 7.48	64.85 ± 9.11	69.71 ± 12.54	67.14 ± 10.04
18–24	73.81 ± 5.03	72.02 ± 5.48	72.88 ± 5.29	70.94 ± 7.01	65.61 ± 8.46	68.16 ± 8.17
Total	72.34 ± 5.74	72.60 ± 6.57	72.49 ± 6.12	67.89 ± 8.40	66.04 ± 9.10	66.94 ± 8.72
	qSPL (dB)			qVE% × 1,000
Group	Male	Female	Total	Male	Female	Total
4–7	56.50 ± 4.85	57.82 ± 6.65	57.24 ± 5.61	0.23 ± 0.29	0.48 ± 0.90	0.38 ± 0.71
8–13	58.41 ± 11.03	59.30 ± 9.19	58.88 ± 9.78	3.71 ± 9.64	1.85 ± 4.06	2.78 ± 7.17
14–17	54.21 ± 7.46	60.86 ± 9.95	57.36 ± 8.52	0.09 ± 0.11	2.46 ± 5.26	1.17 ± 3.55
18–24	60.81 ± 6.87	55.37 ± 7.36	57.98 ± 7.57	0.36 ± 0.76	0.27 ± 0.83	0.31 ± 0.79
Total	58.91 ± 7.90	57.12 ± 7.99	57.99 ± 7.90	0.94 ± 4.20	0.86 ± 2.59	0.90 ± 3.45
	mVE% × 1,000			cVE% × 1,000
Group	Male	Female	Total	Male	Female	Total
4–7	0.39 ± 0.20	1.29 ± 2.35	0.89 ± 1.73	0.53 ± 0.48	1.09 ± 3.37	1.26 ± 2.50
8–13	3.50 ± 9.15	6.29 ± 17.05	5.09 ± 13.78	10.88 ± 26.49	8.14 ± 19.22	9.51 ± 22.11
14–17	0.35 ± 0.36	10.25 ± 22.24	4.85 ± 14.99	0.71 ± 0.98	13.77 ± 30.22	6.65 ± 20.31
18–24	1.16 ± 1.22	0.63 ± 0.84	0.89 ± 1.06	0.99 ± 1.04	0.98 ± 1.39	0.98 ± 1.21
Total	1.43 ± 4.13	3.09 ± 10.95	2.30 ± 8.41	2.55 ± 10.74	3.96 ± 13.31	3.27 ± 12.05

Note.  Statistics are presented as mean ± standard deviation. mVE%, cVE%, and qVE% are presented as the values multiplied by 1,000 to avoid showing numbers with many leading zeros.

Funding Statement

Funding for this research was provided by National Institute on Deafness and Other Communication Disorders Grant R01 DC009153, awarded to Jack J. Jiang.