Effects of warm-up exercises on self-assessed vocal effort

Abstract

Purpose

An elevated sense of vocal effort due to increased vocal demand is frequently reported by patients with voice disorders. However, effects of vocal warm-up on self-assessed vocal effort have not been thoroughly examined. A recently developed version of the Borg CR-10 Scale facilitates vocal effort assessments, following different vocal warm-up tasks.

Methods

Effects of a short (5 min) vocal warm-up on self-assessed vocal effort was evaluated using the Borg CR-10. Twenty-six vocally healthy participants (13F, 13M, mean age 22.6), in two randomised groups, underwent sessions of either reading aloud or semi-occluded vocal tract exercises (SOVTE). Vocal effort was evaluated at four times: pre to post vocal warm-up and two silence periods. Non-parametric analyses for repeated measures and calculations for within-subject standard deviation were applied in group comparisons.

Results

Following vocal warm-up, vocal effort ratings were increased to a statistically significant degree in both intervention groups compared to baseline ratings. After a 5-min rest in silence following completion of the vocal warm-up, vocal effort ratings returned to baseline levels in both groups. The drop in ratings immediately post warm-up compared to 5 min later was statistically significant for the SOVTE group.

Conclusions

Five minutes of vocal warm-up caused increased self-perceived vocal effort in vocally healthy individuals. The increased sense of effort dissipated faster following warm-up for the SOVTE group. When using the Borg CR-10 scale to track vocal effort, it may be beneficial to apply experience-based anchors.

Keywords:

• Vocal effort
• vocal warm-up
• self-evaluation, Borg CR-10
• SOVT

Introduction

Increased vocal effort is frequently reported by patients suffering with hyperfunctional voice disorders. However, vocal effort is challenging to report in exact terms and is often described in varying terms by clinicians and patients [1]. Recently, Hunter et al., refined the meaning of vocal effort. They suggest a narrowed definition for vocal effort as a talker’s self-assessed vocal exertion in response to a specific communicative context [2]. Obtaining objective evidence of self-assessed vocal effort is multifaceted and demanding. Good attempts have been made to find a correlation between increased vocal effort and acoustic outcome and/or morphological changes in or around the larynx during phonation [3–7]. Still, more evidence is needed to further the knowledge on how any vocal intervention, for example, vocal warm-up, would affect self-assessed vocal effort.

According to the above definition, vocal effort is closely related to for example., vocal intensity, depending on the communication context. Following this rationale, as the sound pressure level of vocal output increases, so would self-assessed vocal effort [3,4,6,8–10]. Different contexts will pose different vocal demands. There is a specific association between increased self-assessed vocal effort and increased subglottal pressure with simultaneous laryngopharyngeal compression [5]. This association is in line with Chang and Karnell who have shown vocal effort to co-vary with phonation threshold pressure [11]. The ability to produce soft voice thus decreases with increased vocal effort.

Increased vocal intensity, subglottal pressure and phonation threshold pressure are also in line with findings that show how vocal effort increases in relation to the Lombard effect. F0 and F1 follow increased vocal intensity [8] and harmonics-to-noise ratio will predict vocal effort levels perceived by listeners and speakers alike [6].

Due to the complexity of the vocal apparatus, including internal sensorimotor feedback and external auditory feedback, self-assessment is the preferred mode of evaluation for vocal effort [12,13], as is suggested in the definition by Hunter et al. [2]. The Voice Handicap Index [14] shows closer association to self-assessments of vocal effort than, for example, those of voice quality [15]. Recently, administering self-assessment of vocal effort has been made more accessible by the development of a vocal effort scale (based on the Borg CR-10 Scale) as designed by van Leer and van Mersbergen [16].

This paper aims to further the knowledge on (1) how self-assessment of self-perceived vocal effort may be evaluated in vocally healthy speakers and (2) how two vocal warm-up exercises with ensuing voice rest may affect ratings of self-assessed perception of vocal effort. The aim of the study was to examine any changes in self-assessed vocal effort caused by two contrasting vocal warm-up interventions in vocally healthy individuals with the following research questions:

1. How does vocal warm-up through reading aloud compare to semi-occluded vocal tract exercise (SOVTE) with a short plastic straw regarding self-assessed vocal effort in vocally healthy individuals?

2. If vocal effort ratings are affected by any vocal warm-up intervention, how does controlled vocal rest impact ratings?

The running hypothesis was that a vocal warm-up session with SOVTE would reduce self-assessed vocal effort as compared to reading aloud. SOVTE should provide the speaker with instant feedback of positive changes to vocal production and performance, thereby alleviating effort associated with phonation.

Methods

Participants

Twenty-seven cisgender participants took part, 13 females and 14 males, with gender as a biological factor only. The age range was 18–30 (Mean: 22.6, SD 3.3). None reported previous voice problems. None were smokers. Each participant was randomised to take part in one of two vocal warm-up interventions: n = 13 participants underwent reading aloud and n = 14 SOVTE. One participant was excluded from the study post hoc, as he did not follow instructions not to speak during silent periods. Participants were unaware of group comparisons and study aims at the time of the experiment. Before taking part, all participants completed an online survey, including general demographic information and the following standardised indices: Vocal Fatigue Index (VFI) [17], Big Five Inventory 10 (BFI-10) [18], and Voice Handicap Index-10 (VHI-10) [19]. All participants were self-reported vocally healthy individuals, and likewise, the vocal health related indices confirmed overall good vocal health and are therefore not presented in the results.

Ethical considerations

Each participant gave informed written consent prior to participating in this study, agreeing to the use of voice recordings to be disclosed in reports and presentations. Participants were free to decline or terminate involvement in the study at any time, without any explanation or sanctions for taking part in future studies. Participants were not told the absolute rationale of the study before partaking. There were no personal ties between participants and test leaders. All participants were recruited through an online recruit and scheduling system at Michigan State University in the College of Communication Arts and Sciences. Compensation in the form of course credit was provided for their time. Human subject participation was approved under review from the Michigan State University’s Human Research Protection Programs Human Subject’s Review Board (FWA #00004556, Study Protocol #13-1149).

Procedure

Participants were asked to rate perceived vocal effort at four times: before and after two separate silence periods and before and after assigned vocal warm-up sessions between silence periods. See process flow chart in Figure 1. The rationale for the first silence period was to give all participants a chance to relax their phonatory systems and to validate the repeated vocal effort rating with no vocal activity in between. The second silence period investigated any short-term carry-over effects of vocal effort rating. During the vocal silence periods, participants were played soft, ambient instrumental music. When assessing levels of vocal effort, participants were prompted to utter the following statement aloud “my current vocal effort level is” and continue with the corresponding number of their sense of vocal effort during phonation according to the Borg CR-10 scale (Figure 2). Sessions took place in a double-walled attenuation and isolation booth (53sf, 7 ft high, 33 dBA ambient noise). Sessions were presented using a semi-automated protocol in Microsoft PowerPoint, with participants seeing the presentation on a monitor in front of them, as they progressed. Spoken self-assessment of vocal effort was recorded (44100 Hz, 16 bit, wav) on a laptop using a digital audio workstation (Reaper: https://www.reaper.fm/). Participants were fitted with a head-worn omnidirectional microphone (Shure Beta 53, 20-20kHz, SNR 55 dB at 94 dB SPL) that was connected to the laptop via a USB audio interface (Focusright Scarlett 2i2).

Figure 1. Flow chart of trial process with placement for vocal effort ratings (VER 1-4), vocal warm-up interventions (loud reading or SOVTE) and silence periods. The procedure was prefaced with a practice session and baseline recordings of vocal effort levels.

Flow chart showing trial process.

Figure 2. (Left) self-evaluation of vocal effort as designed by van Leer and van Mersbergen, first published in Journal of Voice, 2017 [16]. (Right) self-evaluation of vocal effort as presented in the experimental set-up.

Left: Borg CR-10 scale as originally published by van Leer and van Mersbergen, 2017. Right: Borg CR-10 scale presented at a slant, in the current study.

Vocal effort ratings

Vocal effort ratings were carried out using the Borg category ratio scale, which was originally developed for evaluating perceived exertion in medicine, somatic ergonomics and sports. The scale combines subjective categories with ratio scaling, thus making comparisons and repeated measures more robust and computational [20]. The Borg CR-10 scale, as adapted to evaluate vocal demand response by van Leer and van Mersbergen [16] (Figure 2) was applied in the current study. Participants categorised their current vocal effort on the 12-point scale and were allowed to use intermediate steps if they wished. Portraying changes in vocal demand in the experimental set-up, the scale was presented at an incline, ad modum Morishita et al. [21].

The rating scale was anchored allegorically only, comparing any change in vocal effort to muscular effort associated to lifting weights at the gym. Participants gave spoken ratings based on allegorical scale references, where 0 = no vocal effort, with steps up to 10 = maximum vocal effort. Descriptive experiential anchors of the scale were presented for the extreme ends, with a comparison of “effortless soft conversation in a quiet room” (0/no effort) to “trying to talk during severe laryngitis or trying to make yourself heard when standing close to a jet engine” (10/maximum vocal effort).

Vocal warm-up sessions

Two short vocal warm-up sessions were compared: 5 min of reading aloud and 5 min of SOVTE using a short, small diameter (100 mm by 5 mm) stirring straw. The two types were selected to represent more intuitive/later stages (reading) and less intuitive/earlier stages (SOVTE) of voice training that could both be carried out in a short, controlled intervention. For the former, more intuitive phonatory warm-up, reading aloud was chosen as it was expected to have a lower effect on self-assessed vocal effort than the latter, having omitted vital parts of a vocal training program, and merely implementing existing phonatory patterns [22]. A 5-min period of reading aloud was applied, as is integrated in the later stages of Lessac-Madsen Voice Therapy. Reading aloud is considered a natural mode of phonation for employing existing vocal technique [23,24]. Participants were instructed to read short texts on Christmas traditions from Reader’s Digest at a comfortable, habitual pace, pitch, and loudness [25]. They were not required to recall what they had read, nor to pay any attention to whether they misread anything, just to continue reading aloud for the 5 min. For the less intuitive phonatory warm-up, SOVTE was chosen and was expected to have a bigger impact on self-assessed vocal effort than reading aloud, with expected changes to pre-existing phonatory patterns. SOVTE in a narrow, short straw is a method known for its sensory effect, but implies less natural phonation, thus representing earlier stages of voice training [26]. The set-up for the SOVTE warm-up session was for the greater part obtained from a publicly available YouTube channel from The National Center for Voice and Speech (username NCVS456), in which Professor Ingo Titze demonstrates a short SOVTE program consisting of the following exercises, together not exceeding 5 min and with a couple of minutes’ of rest in between each:

1. Six slowly ascending and descending pitch glides, covering pitch range from bottom to top and back again (Figure 3, left).

2. Six sets of three ascending hills/accents, then gliding back down, again covering pitch range (Figure 3, right).

3. Reading Goldilocks and the Three Bears aloud through the straw. Instructions were to read using exaggerated dynamics, as if reading to a child. This part replaced a singing exercise [26], as all participants were naïve voice users, unaccustomed to comfortably sing for a recording.

Figure 3. (Left) visual representation as shown prompt for ascending pitch glides (descending not shown). (Right) visual representation as shown prompt for ascending hills/accents for recorded pitch range during SOVTE (own material).

Two computer screens showing visual prompts for participants.

Vocal warm-up sessions were preceded and succeeded by 5 min of silence, to validate the vocal effort rating (first silence) and to evaluate the duration of any change in vocal effort (Figure 1).

Analyses

Statistical analyses were performed with IBM SPSS Statistics 27 (SPSS Inc., Chicago, IL). Komolgorov–Smirnov and Shapiro–Wilks tests showed p<.05, for all ratings, thus group comparisons were carried out using non-parametric statistical analyses. Effort ratings were explored a priori using Friedman’s test for repeated measures. Each rating was compared post hoc to each other using Wilcoxon’s signed rank test. Each rating was compared for group differences using Mann–Whitney U tests. Significance level was set at p<.05 for a priori analyses. Post hoc tests were adjusted using Bonferroni correction, adjusting probability values to p<.025.

To track both groups’ true changes in vocal effort due to vocal warm-up, within-subject standard deviation were extracted for later statistical analysis ad modum Atkinson et al. [27] as adapted by Vogel and Maruff [28]. Standardised vocal effort scores were subsequently determined by subtracting the baseline score from each following effort rating and dividing by the within-subject standard deviation (WSSD).

Results

Vocal effort was generally rated low in both groups throughout trials, with mean baseline scores of 1, i.e. very slight vocal effort. Figure 4 visualises changes in vocal effort ratings at the four time points. VER at each given time point and WSSD are shown in Table 1. Repeated measures showed statistically significant changes in vocal effort ratings for both groups: Reading: (χ²(3) =11.792, p=.008), SOVTE: (χ²(3) =10.156, p=.017). Post hoc tests showed that the initial silence period, i.e. vocal effort rating (VER) 1 to VER 2, did not affect vocal effort ratings in either group. The ratings made for VER 1 and VER 2 were used along with an additional baseline rating completed in connection to participant surveys just preceding the experimental process shown in Figure 1 to form the basis for within-subject standard deviation calculations.

Figure 4. Absolute (left) and relative (right) changes in self-evaluated vocal effort. Bars represent standard error.

Left: results from absolute vocal effort ratings. Right: results from standardized vocal effort ratings.

Albeit slightly, vocal effort ratings increased following vocal warm-up, VER 2 to VER 3, in both groups, moving towards moderate vocal effort. Though the SOVTE group reached higher levels (and greater standard deviation) of vocal effort ratings following vocal warm-up, the increase was statistically significant in the reading group (Z = −2.384, p=.017). The second silence period caused vocal effort ratings to drop from VER 3 to VER 4, in both groups, to a statistically significant degree in the SOVTE group only (Z = −2.539, p= .011). There were no statistically significant differences between groups for any time point. See full account of vocal effort ratings in Tables 2 and 3. Changes to vocal effort with WSSD taken into account showed an increase in vocal effort due to vocal warm-up in both groups, as shown in Figure 4. The Mann–Whitney U test showed no statistically significant differences in WSSD between groups (U = 39.000, p=.399). Friedman’s test for repeated measures showed overall statistically significant changes to vocal effort ratings within both groups. Reading (χ²=30.360, p<.001) and SOVTE (χ²=27.440, p<.001). Post hoc tests with Wilcoxon’s signed rank test of scores between time points showed statistically significant changes to vocal effort evaluation in both groups, however not at all time points, according to Table 3.

Table 1. Individual vocal effort ratings (VER) at each given time point: VER 1–4 and within-subject standard deviation (WSSD).

			VER1		VER2	VER3	VER4
Group	Participant	Baseline rating	Pre silence I	Post silence I	Pre intervention	Post intervention	Post silence II	WSSD (VER 1–4)
Reading	F01_R1	1	1	1	1	2	0.5	0
	F05_R3	0	1	0	0.5	1	0.5	0.58
	F10_R1	1	2	2	3	3	3	0.58
	F12_R2	1	1	1	1	1	1	0
	F13_R3	1	2	1	2	2	1	0.58
	F17_R1	0.5	0.5	0.5	0.5	2	2	0
	F27_R2	0.5	1	0.5	1	2	0.5	0.29
	M03_R2	2	4	4	5	7	4	1.15
	M15_R4	0	0	0	0	0	0	0
	M19_R1	1	1	1	1	1	1	0
	M20_R3	0.7	1	1	1	1,7	0.7	0.17
	M21_R4	2	2	2	2	3	3	0
	M22_R2	2	2	2	2	2	2	0
SOVTE	F02_S1	0	0	0	0.5	0,5	0.5	0
	F04_S2	0	0.5	0	0.5	2	0	0.29
	F08_S1	1	3	0	2	7	0	1.53
	F14_S3	0.5	2	1	1	3	3	0.76
	F18_S1	2	3	2	3	2	2	0.58
	F24_S3	3	2	2	3	2	2	0.58
	F25_S4	0	0.5	0	0.5	1	0	0.29
	F26_S2	0	0	0	0	0,5	0	0
	M06_s3	2	2	2	2	5	3	0
	M07_S4	1	2	1	2	3	2	0.58
	M11_S2	0.5	1.2	0.5	1	1	0.5	0.4
	M16_S4	2	2	2	3	4	3	0
	M23_S1	2	2	2	2	2	2	0

Table 2. Aggregated group results, explaining mean and standard deviation (SD) of vocal effort ratings (VER) at four time points: VER 1–4 with time points compared within groups in total with Friedman’s test for repeated measures as well as at each time change with Wilcoxon’s signed rank test.

Repeated measures	Reading	SOVTE	Mann–Whitney U post hoc test reading vs. SOVTE
VER 1 mean (SD)	1.42 (.9968)	1.55 (1.0228)	U = 73.000, p = .541 NS
VER 2 mean (SD)	1.54 (1.3144)	1.58 (1.0576)	U = 79.000, p = .772 NS
Wilcoxon’s signed rank post hoc test VER 1 vs. VER 2	Z = −1.089, p = .276 NS	Z = −.108, p = .914 NS
VER 3 mean (SD)	2.12 (1.6800)	2.54 (1.8870)	U = 73.500 p = .562 NS
Wilcoxon’s signed rank post hoc test VER 2 vs. VER 3	Z = −2.384, p = .017*	Z = −1.899, p = .058 NS
VER 4 mean (SD)	1.48 (1.2208)	1.39 (1.2442)	U = 76.000, p = .657 NS
Wilcoxon’s signed rank post hoc test VER 3 vs. VER 4	Z = −2.214, p = .027 NS	Z = −2.539, p = .011*
Friedman’s test a priori for repeated measures	χ^2(3) = 11.792, p = .008*	χ ^2(3) = 10.156, p = .017*

Group comparisons were made at each time point with Mann–Whitney U tests. All significance testing has been Bonferroni corrected. Statistically significant results p<.025 are marked with *.

Table 3. Aggregated group results, explaining changes to self-evaluated scores of vocal effort with within-subject standard deviation (WSSD, see Table 2) taken into account at four time points: VER 1–4.

Repeated measures standardised scores with WSSD	Reading	SOVTE	Mann–Whitney U post hoc test reading vs. SOVTE
WSSD by group	.25	.39	U = 39.000, p = .399 NS
ΔVER 1	0	0	NA
ΔVER 2	2.23	2.30	U = 75.500, p = .643 NS
Wilcoxon’s signed rank post hoc test VER 1 vs. VER 2	Z = −3.065, p = .002*	Z = −3.063, p = .002*
ΔVER 3	3.41	3.50	U = 71.500, p = .504 NS
Wilcoxon’s signed rank post hoc test VER 2 vs. VER 3	Z = −2.524, p = .012*	Z = −1.993, p = .046 NS
ΔVER 4	2.04	1.86	U = 79.500, p = .801 NS
Wilcoxon’s signed rank post hoc test VER 3 vs. VER 4	Z = −2.201, p = .028 NS	Z = −2.521, p = .012*
Friedman’s test a priori for repeated measures	χ^2 = 30.360, p < .001*	χ^2 = 27.440, p < .001*

Time points are compared in total within groups with Friedman’s test for repeated measures as well as at each time change with Wilcoxon’s signed rank test. Group comparisons were made at each time point with Mann–Whitney U tests. All significance testing has been Bonferroni corrected. Statistically significant results p<.025 are marked with *.

Discussion

The first research question examined how vocal warm-up sessions comprised of either reading aloud or SOVTE affected self-assessments of vocal effort in vocally healthy individuals. Although there were no statistically significant differences between groups, they both increased their absolute scores of self-assessed vocal effort following vocal warm-up sessions—including a statistically significant increase in the reading group. The changes were slight, only ever reaching towards moderate vocal effort on the Borg CR-10 scale.

An increased sense of vocal effort may reflect an increased self-awareness of the voice, although this is an area in need of more investigation. Participants may suddenly have been mindful of their voice being monitored and of preciously unfamiliar vibratory feedback, especially for the SOVTE group. This mode of phonation in untrained speakers may have led to increased “phonatory self-awareness.” Increased awareness of the voice and factors potentially threatening vocal health, would verify the hypothesis from Thomas et al., who implore teacher students to take heed of risk factors in their work environment, reflected in worryingly high VHI scores [29].

The second research question focussed on the mode of eliciting evaluations of vocal effort by use of the Borg CR-10 scale, by comparing silence periods to sessions of vocal warm-up. A short rest in silence following warm-up sessions (VER 3 to VER 4) lead to a decrease of vocal effort in both groups—statistically significant in the SOVTE group. A decline was expected due to cessation of the warm-up activity. The fact that the difference was statistically significant for the SOVTE group may indicate that an increased sense of effort may wear off sooner following SOVTE than reading aloud. The initial silence period, between VER 1 and VER 2 did not notably affect absolute vocal effort ratings in either group, indicating scale solidity. However, the standardised scores showed a different trend, with increased vocal effort ratings during the first silence period (VER1 to VER2). The increase may be interpreted as increased vocal awareness in an unfamiliar setting as discussed above. It could on the other hand be viewed as a suggested dispersion for baseline ratings of vocal effort in vocally healthy individuals. This notion would require further exploration.

Waiting for 5 min is fundamentally different to phonating in a new, previously unfamiliar setting as the latter places a far heavier cognitive burden on participants. In a previous study, van Mersbergen et al. discuss the possible influences of cognitive strain on self-assessed vocal effort—that ratings of cognitive load and vocal effort are interrelated and that depletion of mental resources during a task may hasten increased vocal effort [30]. This hypothesis backs up the results of increased vocal effort due to vocal intervention in the current study. The running hypothesis was that vocal warm-up sessions would reduce vocal effort. Vocal effort was also hypothesised to decrease to a higher degree in the SOVTE group than the reading group, as it offers instant feedback of vocal tract occlusion and a sensation of vibration or buzzing, much as resonance tube phonation in water would [31]. Potentially, sensory feedback and softer vocal fold adduction would lessen the phonation threshold pressure, rendering a sense of ease to phonation [26]. Instead, results showed a clearer decline in vocal effort following the vocal warm-up for the SOVTE group as compared to the reading group. Vocal effort ratings possibly increased following warm-up due to improved biofeedback from a widened vocal tract, certainly in the case of SOVTE [31]. The finding is in line with Meerschman et al., who compared three separate modes of SOVTE during a rigorous three-week program in patients with dysphonia. They found improved objective acoustic ratings following straw phonation (using a straw double our length at 21 mm, but with the same width), but no corresponding subjective improvement. Self-assessment of voice quality even deteriorated following three weeks of straw phonation [32]. It may also be hypothesised that the narrow straw used in the current study may increase back pressure to an unnecessarily high extent during phonation, leading to higher levels of vocal effort. Our findings indicate that low levels of self-assessed vocal effort, or self-perceived exertion, need to be further explored as they may not necessarily be opposed to vocal health.

Clinical implications

While increased awareness of vocal production as shown in increased sense of effort may be part of a somatic defence mechanism, beneficial to vocal health, there is another side of the coin for voice patients. Any intended change of vocal behaviour in the voice clinic, requires a patient withstanding a certain level of vocal effort, or else a new vocal technique will not be long lasting. Finding and calibrating an appropriate level of vocal effort is crucial for patients to be able to make themselves heard while using a durable mode of phonation. The notion relates to vocal demand (or loading) capacity, which is an important part of a healthy voice. Echternach et al. have examined how to test for vocal loading (or vocal demand) capacity in a controlled environment, showing that as little as 10 min of a properly monitored vocal loading task will correspond to the capacity needed for 45 min of teaching [33]. This type of controlled testing may be useful in voice therapy when outcome goals relate to increasing a patient’s ability to cope with manageable levels of vocal effort. As vocal warm-up is deemed to be a part of the vocal demand or loading process, it is not altogether surprising that slightly increased levels of vocal effort would be an effect of vocal warm-up [34]. From a clinician’s perspective, vocal warm-up is described as a subjective, positive part of an increased vocal demand, which in turn ultimately will lead to vocal fatigue, via increased vocal effort [23]. Even small changes to vocal effort may be the goal for patients aiming to manage higher doses of phonation, thus coping with higher levels of vocal effort while using a healthy voice.

Study limitations and suggestions for future work

The current study did not control for changes in vocal intensity nor in subglottal pressure following vocal warm-up. Non-invasive evaluation of intraoral pressure cm H₂O [11], preceding and following vocal warm-up would be a good complement. Although variable in nature, objective markers may enable confirmed changes to self-assessed perceived vocal effort. This was, however, beyond the scope of the current research aims. Expecting vocal effort to decrease following vocal warm-up is reasonable. It is important to control for the amount of vocal dose accumulated before a participant attends the experiment. In vocally healthy young adults, it can be assumed that a fair number of vocal demands have already been placed on the phonatory system, depending on what time of day the experiment was carried out. As would also be in a clinical assessment, previous vocal dose was not controlled in this study and is indeed difficult to control for without appropriate modes of voice accumulation, as merely asking someone how much they have used their voice in a day, may differ significantly from reality [35].

Another factor that may vary in individuals is the ability to self-assess perceived vocal effort at all. This ability needs to not be taken for granted and may have to be studied when examining self-assessed vocal effort, certainly in patients with hyperfunctional voice disorders. As vocal effort may be related to sensorimotor feedback of the voice function, which in turn is increased by vocal warm-up, Crow et al. suggest taking individual ability for interoception, i.e. a person’s ability to perceive internal somatic processes, into account when investigating vocal self-assessment of vocal function, whether it be vocal ability or acoustic output [36]. Crow’s new hypotheses are compelling and would be important to investigate further in relation to variability of within-subject standard deviation. It may provide answers to what variability is expected in self-assessments of vocal effort and by whom, not assuming all vocally healthy people (or for that matter voice patients) to have a common reaction to increased vocal demands. The short warm-up sessions applied in the current study could, on the other hand, indicate the start of the vocal warm-up process, which may adequately involve a slightly increased sense of vocal effort.

Finally, the Borg CR-10 effort scale used in this study not only indicated significant differences between time points but has previously also been instrumental in indicating significant effects for cognitive effort (mentioned above) [32], mood [37,38], and post-therapy outcomes [16]. However, with only 10 specified levels of effort in the Borg CR-10, the scale delineation in these studies may lack the sensitivity to be clinically significant [16]. In response, a revised scale has been recently presented in a study in which participants responded to target specific vocal effort levels to estimate the use of the Borg CR-100 scale. The magnification of the scale from the Borg CR-10, seems to have helped vocally healthy speakers use the scale more appropriately, as the authors were able to validate that phonatory sound pressure level output matched to vocal effort level [10]. Future iterations of the current study could employ this new scale. Nevertheless, the overall results and conclusions should be unaffected with only a higher resolution of small differences potentially illustrated.

Conclusion

A short vocal warm-up session may cause the sense of vocal effort to increase in vocally healthy young adults. The increase is hypothetically related to sensory feedback from the vocal tract during phonation along with increased awareness of the work involved in voicing, hypothetically related to increased awareness of one’s own efforts connected to phonation. Participants of the current study had a significantly greater reduction in vocal effort following SOVTE and a short period of rest, than when reading out loud with rest. This may be because SOVTE aims to improve the ease of phonation and may have possible mitigating effects of effort over time. A within-subject standard deviation in repeated measures of vocal effort ratings using the Borg CR-10 scale in vocally healthy young adults of 0.32 is to be expected. The use of the Borg CR-10 to track vocal effort may benefit from experience-based anchors, not only allegorical or anecdotal anchors. It may also be more useful when used with a more defined resolution, i.e. Borg CR-100.

Acknowledgements

The authors express their gratitude to all participants’ participation in this study. We would also like to thank the two independent reviewers, whose comments improved this manuscript considerably.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

The research reported in this publication was supported by the National Institute of Deafness and Other Communication Disorders of the National Institutes of Health under Award Number R01DC012315 (P.I. Eric Hunter). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Notes on contributors

Susanna Whitling

Susanna Whitling, SLP, PhD is a registered Speech and Language Pathologist and a researcher and lecturer of Speech and Language Pathology at the Department of Logopedics, Phoniatrics and Vocology, Faculty of Medicine, Lund University, Lund, Sweden. She specializes in occupational voice use, treatment and risk management.

Qin Wan

Qin Wan, PhD is a faculty and researcher in the Department of Rehabilitation at East China Normal University in Shanghai, People's Republic of China. She specializes in speech production in children with specialty vocal onset.

Mark L. Berardi

Mark l. Berardi, PhD is a Postdoc researcher in the Vocal Control and Vocal Well-Being Laboratory at University Hospital Bonn, Bonn, Germany. He specializes in acoustics, voice acoustics, and machine learning.

Eric J. Hunter

Eric J. Hunter, MS, PhD is the Associate Dean for Research in the College of Communication Arts and Sciences as well as a Professor in the Department of Communicative Sciences and Disorders at Michigan State University, East Lansing, Michigan, United States of America. He specializes in voice signal processing, voice production, and voice modeling.