- Email: [email protected]
Effects of Coaching and Repeated Trials on Maximum Phonational Frequency Range in Children
ARTICLE IN PRESS Effects of Coaching and Repeated Trials on Maximum Phonational Frequency Range in Children Estella P.-M. Ma and Trista K.-Y. Li, Hong Kong Summary: Purpose. Maximum phonational frequency range (MPFR) is the frequency range from the lowest to the highest pitch that an individual can produce. This study investigated the effects of coaching and repeated trials on MPFR in a group of school-age children. Methods. Thirty girls aged 6–11 years were randomly assigned into two groups: coaching and non-coaching. All of the participants produced the lowest and the highest phonational frequency for 10 times each. The participants in the coaching group were prompted by the clinician with verbal encouragement and a visual cue (hand-sweeping) to produce their maximum performance. The participants in the non-coaching group were simply asked to repeat the task 10 times. Results. The clinician’s coaching helped the participants in the coaching group reach their MPFR in fewer trials. The MPFRs elicited in 10 trials were significantly greater than those elicited in fewer trials. Conclusions. These findings suggested that coaching and repeated trials could facilitate the elicitation of MPFR more efficiently. Key Words: pediatric voice–voice assessment–vocal functions–coaching–repeated trials.
INTRODUCTION Maximum phonational frequency range (MPFR) is a test of maximum vocal performance. It refers to the frequency range from the lowest pitch in the modal register to the highest pitch in the falsetto register that an individual can produce.1 Glottal fry is commonly excluded because it is not continuously used in speech.1 MPFR reflects the vocal capacity of an individual. It reveals the biomechanical and physiological limits of the respiratory and phonatory systems.2–4 Clinically, a reduction in MPFR can be a sign of voice problems.5 MPFR is one of the most frequently obtained measures in voice evaluation.6,7 Currently, there is no standardized procedure for eliciting MPFR, which makes the comparison of MPFR data across voice clinics and research difficult.4,8,9 The variability of MPFR reported in the literature is rather large.10 Van Oordt and Drost11 reported that children aged 6–16 years could present MPFRs from 1.5 to 3 octaves. Reich et al8 showed that the MPFRs in children aged 6–13 years could range from 1 to 3.6 octaves. Different task variables can lead to MPFR variability across individuals. These variables include instructions to clients, elicitation tasks,3,4 time of day,12 coaching by the clinician, visual feedback, and repeated trials.3,4,9,13 Cooper and Yanagihara12 examined the influences of the time of day on the lowest phonational frequency in a group of vocally healthy adults. Their results showed that the lowest phonational frequency varied from one to three semitones (STs) throughout the day. Zraick et al4 compared the effects of two elicitation procedures, mid-basal-to-ceiling and mid-ceiling-to-basal, on MPFR in adults and found no significant difference. Reich et al3 investigated the MPFRs of 40 children from grades 3 to 6. They found that discrete step produced sig-
nificantly smaller MPFRs than elicitation tasks such as slow and fast steps, and slow and fast glissando. Because MPFR production requires an individual’s maximum vocal effort, factors that enhance motivation are expected to promote better MPFR performance. Examples of such factors include verbal encouragement and coaching provided by the clinician.3,13 Coleman13 and Kent et al10 suggested that the presence of clinician coaching may increase one’s motivation. According to McClelland,14 extrinsic motivation created by external sources such as encouragement and incentives offered by others can increase one’s self-confidence and intent to achieve one’s goal. Early studies suggest positive effects of coaching in the form of verbal encouragement on maximum phonation time elicitation.15,16 However, whether a similar positive influence of coaching can be applied for MPFR has not yet been studied. Practice through repeated trials is necessary for motor performance improvement.17 Superior performance elicited through repeated trials has been reported for maximum phonation time, with more than three trials needed to elicit a representative maximum phonation time in children.18 Some speakers have required up to 15 trials to achieve their maximum phonation time.19,20 Currently, three trials are commonly used to elicit MPFR in children3,21 and adults.8,9,12 One earlier study implemented as many trials as were needed to satisfy the researcher and the subject.1 Whether the use of repeated trials promotes larger MPFR has yet to be proven. The present study aimed to investigate the effects of coaching and repeated trials on MPFR in children. It was hypothesized that coaching and repeated trials could promote larger MPFR. METHOD
Accepted for publication May 16, 2016. From the Voice Research Laboratory, Division of Speech and Hearing Sciences, The University of Hong Kong, Hong Kong. Address correspondence and reprint requests to Estella P.-M. Ma, Voice Research Laboratory, Division of Speech and Hearing Sciences, The University of Hong Kong, 7/F Meng Wah Complex, Pokfulam, Hong Kong. E-mail: [email protected] Journal of Voice, Vol. ■■, No. ■■, pp. ■■-■■ 0892-1997 © 2016 Published by Elsevier Inc. on behalf of The Voice Foundation. http://dx.doi.org/10.1016/j.jvoice.2016.05.013
Participants Thirty girls between the ages of 6 and 11 years (mean age = 8.97 years, SD = 2.00) were recruited. The lower age limit was chosen to ensure that the participants could comprehend and follow the instructions. The upper age limit was chosen to exclude voice changes secondary to puberty. All of the participants were Cantonese native speakers who had normal voice quality, as judged
ARTICLE IN PRESS 2
Journal of Voice, Vol. ■■, No. ■■, 2016
TABLE 1. Participants’ Age, Weight, and Height in the Coaching and the Non-coaching Groups Age (in years) Group Coaching (N = 15) Non-coaching (N = 15)
Weight (in kg)
Height (in cm)
Mean
(SD)
Mean
(SD)
Mean
(SD)
8.97 9.00
(2.00) (1.61)
26.93 31.87
(6.74) (8.85)
131.60 133.60
(13.42) (12.19)
Abbreviations: N, number of participants; SD, standard deviation.
perceptually by the researchers; had not received any previous voice and singing training; and had normal hearing according to parental reports. Participants were excluded from the study if they had a previous history of or presented with a respiratory disorder or allergy, or had a previous history of or presented with any form of speech or language disorder or delay. The participants in both groups were similar in age, weight, and height (Table 1). Equipment All of the recordings used in this study were recorded at the Voice Research Laboratory, The University of Hong Kong. The background noise was measured by a sound level meter (Quest Electronics, Model 210, Oconomowoc, WI, USA) and kept below 55 dBA throughout the recording process. Swell Real-time DSP Phonetograph version 2.0 (Phog 2.0, AB Nyvalla DSP, Stockholm, Sweden) with a Dell Pentium III 500-MHz PC computer (Dell Inc., Round Rock, TX, USA) was used to capture the recordings. Procedures The participants were randomly assigned to the coaching group or the non-coaching group. MPFRs were elicited using glissando. Under this procedure, the participants were instructed to sustain the vowel /a/ at their most comfortable pitch and loudness level. They were then asked to glide from the comfortable pitch to the lowest or to the highest frequency. The lowest and highest frequencies were elicited for 10 trials each. All of the participants sat directly in front of the computer screen for visual feedback on their vocal performance. Visual feedback was provided to facilitate maximum performance for the participants who did not have any prior musical training.22 A headmounted condenser microphone (AKG Acoustics C420, Vienna, Austria) was placed on each participant’s head with a mouth-tomicrophone distance kept at 5 cm throughout the recording. Before the actual recording, the participants were allowed to practice the pitch-gliding task three times as vocal warm-up.9 There was no instruction provided for the participants during the glissando practice. The voice samples were recorded directly into the Soundswell phonetogram Phog 2.0 (Hitech Development AB, Sweden). The phonation registration duration of the Phog 2.0 program was 25 ms. The program captured and displayed the signals in real time on a computer screen as augmented visual feedback for the participants in both groups. During the recording process, the participants in the coaching group were prompted by the clinician with verbal encouragement and a hand-sweeping gesture that traveled up or down. After each trial, they were encour-
aged to perform better in the following trial. The verbal instructions provided for the participants in the coaching group were as follows: First trial: I want to know how high/low a pitch you can produce. Take a deep breath and then say /a/ from your most comfortable pitch to the highest/lowest pitch you can produce. Remember to go as high/low as possible. Ready? Go! [while the child was performing] Good! Keep going! From the second to tenth trials: You did a great job in the previous trial. Now, I want you to do that again. See if you can produce an even higher/lower pitch. Remember to take a deep breath and go as high/low as possible! Ready? Go! [while the child was performing] Good! Keep going! The feedback provided for the participants in the noncoaching group was as follows: First trial: I want to see how high/low a pitch you can produce. Take a deep breath and then say /a/ from your most comfortable pitch to the highest/lowest pitch you can produce. Ready? Go! From the second to tenth trials: Okay. Now, do it again. . . The lowest frequency was elicited before the highest frequency to avoid vocal fatigue.13 All of the participants repeated 10 downward trials before producing 10 upward trials. The whole data collection process took about 30 minutes. Six of the participants (20% of the 30 participants) underwent the same procedure 2 weeks after the first data collection. This was to evaluate the test-retest reliability of the recording procedure.
Data analysis Three measures were derived for each participant: highest phonational frequency, lowest phonational frequency, and MFPR. For each participant, the lowest frequency across all of the trials was regarded as the lowest phonational frequency. Similarly, the highest frequency across all of the trials was regarded as the highest phonational frequency. The MFPR was calculated as the difference between the highest and the lowest frequencies. Because the fundamental frequency values in hertz are linear in scale, but the perception of sound is logarithmic, the frequency range data were converted to a logarithmic scale in STs. This provided a standard comparison of the frequency ranges of the coaching and non-coaching groups.23 The frequency range was converted from hertz to STs using the following algorithm: MPFR
ARTICLE IN PRESS Estella P.-M. Ma and Trista K.-Y. Li
3
Effects of Coaching and Repeated Trials on MPFR in Children
TABLE 2. Test-retest, Inter- and Intra-rater Reliabilities, and Agreement in the Lowest and Highest Frequencies Percentage (%) Measure Lowest frequency Intra-rater Inter-rater Test-retest Highest frequency Intra-rater Inter-rater Test-retest
Pearson r
Exact Agreement
Within One ST
Within Two STs
0.962* 0.811* 0.824*
81.7 53.3 26.7
91.7 65.0 45.0
100.0 88.3 75.0
0.998* 0.974* 0.800*
91.7 66.7 13.3
95.0 78.3 25.5
98.3 95.0 58.3
* Significance level at .0001. Abbreviation: ST, semitone.
in STs = (log10 [the highest frequency in hertz/the lowest frequency in hertz])/log102 × 12.
MPFR across trials Table 3 lists the means and standard deviations of the frequency measures in both groups across the 10 trials. Table 4 lists the means, standard deviations, and the ranges of the frequency measures in both groups.
Effects of repeated trials on MPFR Table 5 lists the means and standard deviations of the lowest frequency (Hz), the highest frequency (Hz), and MPFR (Hz/ST) in the first, third, fifth, and tenth trials. To investigate the main effects of repeated trials and their interaction with the effects of coaching and repeated trials, a two-way mixed analysis of variance was used with the group as the between-subject factor (coaching versus non-coaching) and the number of trials (1, 3, 5, or 10 trials) as the within-subject factors for each measure. The Mauchly test of sphericity for within-subject factors was significant (P = 0.0001), indicating that the assumption of compound symmetry had been violated. Thus, the results of the withinsubject effects with a Greenhouse-Geisser epsilon correction, which corrected the degree of freedom, were reported. The significant main effects of trials were found for the lowest frequency (F[1,3] = 18.60, P = 0.0001), the highest frequency (F[1,3] = 14.67, P = 0.001), MPFR in hertz (F[1,2] = 17.51, P = 0.0001), and MPFR in ST (F[1,2] = 28.71, P = 0.0001). A follow-up repeated one-way analysis of variance was performed within each measure to further evaluate the main effect of repeated trials. For all of the frequency measures, those elicited in 10 trials were significantly better than those elicited in one or three trials (all P < 0.01). The lowest and highest frequencies elicited in 10 trials were significantly lower and higher, respectively, and the MPFRs were significantly greater in 10 trials. All of the frequency measures elicited from three and five trials were similar (all P > 0.05), except for the MPFR in STs (P = 0.005). There were no significant group main effects or interaction effects for any of the frequency measures (P values > 0.05).
Effects of coaching on MPFR The effects of coaching were evaluated by comparing the frequency measures between the coaching and the non-coaching groups using independent t tests. Four repeated t tests were conducted for data analysis, and the alpha level was adjusted to 0.0125 (0.05/4) using the Bonferroni adjustment to avoid any potential type II errors. There was no statistical significance in any of all the frequency measures between groups.
Minimum number of trials required to elicit maximum performance Figure 1 shows the percentage of the participants who achieved the lowest and highest frequencies in each trial. To better demonstrate how closely the participants’ performances approximated the maximum frequency values, the mean percentages of the lowest and highest frequencies achieved in each trial were also analyzed (Figure 2). For the lowest frequency, the majority (80%)
Reliability of data analysis procedure As the determination of frequency measures required the visual judgment of frequency data (lowest and highest frequencies) from the display on the computer screen, we needed to establish the reliability of this procedure. The voice recordings of six of the participants (20%) were reanalyzed by the clinician for the lowest and the highest frequencies. This was to evaluate the intrarater reliability. The same recordings of these six participants were then analyzed by another clinician to evaluate the interrater reliability. RESULTS Reliability of data analysis procedure Pearson correlation coefficients were used to evaluate the reliability of the data analysis procedure used in this study. The intra-rater reliability coefficients were above 0.96 (P = 0.0001). The inter-judge reliability coefficients were above 0.81 (P = 0.0001). Table 2 lists the results of the test-retest, interrater and intra-rater reliabilities, and agreement of the lowest and highest frequencies.
ARTICLE IN PRESS 4
Journal of Voice, Vol. ■■, No. ■■, 2016
TABLE 3. Means (and Standard Deviations) for Frequency-related Measures Across 10 Trials in the Coaching and Non-coaching Groups Number of Trials Group
1
2
Lowest frequency in Hz C 171.05 162.85 (34.81) (38.00) NC 184.29 179.43 (45.39) (34.40) Highest frequency in Hz C 1786.45 2000.53 (723.40) (680.06) NC 1586.85 1523.85 (624.59) (616.19) Frequency range in Hz C 1615.39 1867.15 (739.69) (680.02) NC 1402.56 1502.42 (644.11) (669.69) Frequency range in semitones C 39.13 43.87 (10.76) (8.39) NC 36.46 38.40 (9.38) (8.84)
3
4
5
6
7
8
9
10
158.01 (33.02) 184.26 (39.92)
164.53 (27.63) 179.17 (31.29)
162.01 (33.95) 175.19 (34.20)
164.93 (26.49) 184.82 (28.10)
165.61 (26.29) 173.47 (52.96)
169.53 (31.19) 183.30 (40.48)
164.93 (30.88) 172.67 (37.46)
164.12 (26.12) 166.00 (31.78)
1685.60 (744.66) 1519.31 (580.58)
1936.86 (788.99) 1592.66 (682.92)
1815.99 (683.55) 1568.25 (644.97)
2034.95 (619.24) 1668.65 (627.26)
1895.79 (571.38) 1586.09 (631.51)
1928.73 (535.07) 1762.39 (729.40)
1932.27 (505.15) 1732.28 (712.47)
2003.81 (484.10) 1744.06 (729.50)
1904.30 (696.17) 1526.35 (659.50)
2010.40 (719.23) 1591.14 (708.05)
2026.63 (700.36) 1595.01 (705.45)
2168.21 (647.51) 1623.23 (711.86)
2181.39 (651.82) 1638.55 (712.66)
2200.04 (645.39) 1670.58 (727.47)
2206.68 (636.24) 1728.96 (767.36)
2206.68 (636.24) 1754.06 (765.18)
44.99 (7.75) 39.00 (8.92)
46.46 (6.71) 38.07 (9.08)
46.92 (6.61) 40.40 (8.59)
48.39 (5.86) 40.93 (8.60)
48.48 (5.84) 41.20 (8.56)
48.66 (5.83) 41.53 (8.71)
48.99 (5.84) 42.13 (8.88)
48.99 (5.84) 42.73 (8.83)
Note: Standard deviations are in parentheses. Abbreviations: C, coaching; NC, non-coaching.
of the participants in the coaching group could achieve their best performance by the third trial, which already represented 95% of the participants’ lowest frequency values. However, about onefourth (27%) of the non-coaching group could achieve their lowest frequency by the third trial. For the highest frequency, six trials were needed for both groups to achieve at least 95% of their highest frequency values (Figure 2B). However, by the sixth trial, 80% of the participants in the coaching group had achieved the highest frequency, compared with only 46.7% of the participants in the non-coaching group.
DISCUSSION The aim of the present study was to investigate the effects of coaching and repeated trials on determining MPFR in children. The results revealed no significant coaching effect on the lowest frequency, highest frequency, and MPFR values. However, coaching did help the participants reach their MPFR in fewer trials. In addition, the MPFRs elicited in one, three, and five trials were all significantly smaller than those elicited in 10 trials. The frequency measures (lowest frequency, highest frequency, and MPFR) obtained from the coaching and non-coaching
TABLE 4. Means (Standard Deviations) and Ranges of the Four Frequency Measures in the Coaching and Non-coaching Groups Coaching Measures (unit) Lowest frequency (Hz) Highest frequency (Hz) MPFR (Hz) MPFR (ST)
Non-coaching
Independent t Tests
Mean (SD)
Range
Mean (SD)
Range
t
df
P
135.18 (26.20) 2341.85 (636.48) 2206.67 (630.24) 48.99 (5.84)
103.8–196.0
151.79 (29.22) 1905.85 (758.23) 1754.06 (765.18) 42.73 (9.82)
110.0–207.7
−1.64
28
.11
987.8–2960
1.71
28
.10
832.2–2691.2
1.84
28
.77
28.0–56.0
2.29
28
.03
1179.7–3520 1074.7–3396.5
Note: None of the above P values are smaller than 0.017.
35.0–58.0
ARTICLE IN PRESS Estella P.-M. Ma and Trista K.-Y. Li
5
Effects of Coaching and Repeated Trials on MPFR in Children
TABLE 5. Means and Standard Deviations for the Lowest Frequency (Hz), Highest Frequency (Hz), and MPFR (Hz/ST) on First, Third, Fifth, and Tenth Trials Across the Coaching and Non-coaching Groups Measure Number of Trials
Group
First trial
C NC
Third trial
C NC
Fifth trial
C NC
Tenth trial
C NC
Lowest F0 (Hz)
Highest F0 (Hz)
MPFR (Hz)
MPFR (ST)
171.05 (34.81) 184.29 (45.39) 145.01 (31.13) 169.97 (36.07) 138.94 (26.98) 160.59 (32.83) 135.18 (26.20) 151.79 (29.23)
1786.45 (723.40) 1586.85 (624.59) 2049.31 (698.04) 1696.32 (648.31) 2165.57 (699.60) 1755.60 (700.92) 2341.85 (636.48) 1905.85 (758.2)
1615.39 (739.69) 1402.56 (644.11) 1904.30 (696.17) 1526.35 (659.50) 2020.63 (700.36) 1595.01 (705.45) 2206.67 (636.24) 1754.06 (765.18)
39.13 (10.76) 36.46 (9.38) 44.99 (7.75) 39.00 (8.92) 46.92 (6.61) 40.40 (8.59) 48.99 (5.84) 42.73 (9.82)
Note: Standard deviations are in parentheses. Abbreviations: C, coaching; NC, non-coaching.
groups were similar, suggesting no clear coaching effect on eliciting MPFR. The results were not consistent with other studies, which have shown significant coaching effect on maximum vocal performance tasks, namely, maximum phonation time15 and maximum sound pressure level.24 The lack of coaching effect in this study might be due to the provision of visual feedback during the elicitation of MPFR. Zelaznik25 suggested that any real-time feedback can facilitate one’s performance. Letting the participants visually refer to their own performances could have motivated them to perform better. In addition to such intrinsic motivation, visual feedback can simultaneously assist a participant’s learning during a task. Finnegan26 advocated that providing children with visual feedback helped them self-monitor and enhance their performance in a task measuring maximum phonation time. It is apparent that the effects of visual feedback on maximum vocal performance cannot be underestimated. Future research studies should examine the effects of coaching on MPFR between participants who are provided with visual feedback of their performances and those who are not. The frequency measures elicited in the first trial were significantly poorer than those elicited in subsequent trials for both groups. The MPFRs elicited in 10 trials were significantly greater than those elicited in one, three, or five trials. These results could be explained by practice effect on maximal performance. Repeated practice can increase one’s familiarity with the methods and procedures of the tasks.27 Bless and Hirano28 suggested that optimal vocal performances could only be elicited when sufficient practice was allowed. As Figure 1B shows, although a clinician’s coaching was provided, less than 50% of the participants in the coaching group exhibited their highest frequency by the third trial. The participants in the coaching group achieved only 90% of the highest
frequency values by the third trial (Figure 2B). The present results suggest that optimal frequency performance may not be exhibited in three trials. A radical rethinking of using only three trials to obtain MPFR in contemporary voice assessment protocols is warranted. The present results suggest that with a clinician’s coaching, three trials are recommended for eliciting the lowest frequency, and six trials are needed to elicit the highest frequency. Our data suggest that the majority of the participants came very close to reaching their lowest and highest frequencies in these trials. Such repeated trials take less than 10 minutes to complete, an acceptable amount of time in regular clinical acoustic data collection. Comparison with the literature on MPFR in children The MPFR data from the present study (mean MPFR in the coaching group: 48.99 ST; non-coaching group: 42.73 ST) were greater in magnitude than those reported in previous studies. Most early studies used three trials to elicit MPFR. However, considering the frequency values obtained by the third trial, the mean MPFRs obtained in the present study (coaching: 44.99 ST; noncoaching: 39.0 ST) were still greater than those in early reports. Reich et al3 found a mean MPFR of 27.8 ST (collapsed across gender and elicitation tasks) in three trials. McAllister et al29 reported a mean MPFR of 25.0 ST in 24 vocally healthy children aged 10 years. Heylen and colleagues22 reported a mean MPFR of 26.4 ST in 94 vocally healthy children aged 6–11 years. The variability of MPFRs elicited across studies might possibly be due to methodological differences, namely, number of repeated trials, provision of real-time visual feedback, provision of clinician’s coaching, phonation registration time, eliciting procedures (eg, manual versus automatic recording), and recording software used.30 In the present study, the provision of clin-
ARTICLE IN PRESS 6
Journal of Voice, Vol. ■■, No. ■■, 2016
FIGURE 1. Percentage of participants reaching (A) the lowest frequency and (B) the highest frequency in each trial (solid line = coaching group; dotted line = non-coaching group). ician’s coaching and visual feedback of performance might have facilitated participants to approach their true physiological vocal limits. Nevertheless, the great variability in MPFR values across studies reflects the need for standardized elicitation procedures for valid MPFR comparisons in clinical and research settings. Such elicitation procedures should also reveal children’s true physiological vocal limits.
Future research directions In the present study, only girls with normal voices were recruited. Boys and girls can differ in their learning styles, personalities, and motivations in achieving optimal performance in maximum vocal tasks. It would be interesting to compare how boys and girls react to coaching and repeated trials. Moreover, performance in children with normal voices
ARTICLE IN PRESS Estella P.-M. Ma and Trista K.-Y. Li
Effects of Coaching and Repeated Trials on MPFR in Children
7
FIGURE 2. Mean percentage of (A) the lowest frequency and (B) the highest frequency as a function of the trials (solid line = coaching group; dotted line = non-coaching group).
might differ from that in children with dysphonia. A dysphonic voice would increase the variability and hence reduce the precision of an investigation.15 Future studies should also look into the effects of coaching and repeated trials in children with dysphonia. A similar methodology could also be extended for use in adults with and without voice disorders, as the learning
curve, motivation, comorbidities, and other factors may differ substantially between children and adults. CONCLUSION The present findings suggest that coaching and repeated trials could make the elicitation of MPFR more efficient. The results
ARTICLE IN PRESS 8 suggest that with a clinician’s coaching, three trials are recommended to elicit the lowest frequency, whereas six trials are needed to elicit the highest frequency. REFERENCES 1. Hollien H, Dew D, Philips P. Phonational frequency ranges of adults. J Speech Hear Res. 1971;14:755–760. 2. Baken RJ, Orlikoff RF. Clinical Measurement of Speech and Voice. 2nd ed. San Diego: Singular Publishing Group; 2000. 3. Reich AR, Mason JA, Frederickson RR, et al. Factors influencing fundamental frequency range estimates in children. J Speech Hear Disord. 1989;54:429–438. 4. Zraick RI, Keyes MP, Montague JC, et al. Mid-basal-to-ceiling versus mid-ceiling-to-basal elicitation of maximum phonational frequency range. J Voice. 2002;16:317–322. 5. Colton RH, Casper JK, Leonard R. Understanding Voice Problems: A Physiological Perspective for Diagnosis and Treatment. 3rd ed. Philadelphia: Lippincott Williams & Wilkins; 2006. 6. Behrman A. Common practices of voice therapists in the evaluation of patients. J Voice. 2005;19:454–469. 7. Hirano M. Objective evaluation of the human voice: clinical aspects. Folia Phoniatr (Basel). 1989;41:89–144. 8. Reich AR, Frederickson RR, Mason JA, et al. Methodological variables affecting phonational frequency range in adults. J Speech Hear Disord. 1990;55:124–131. 9. Zraick RI, Nelson JL, Montague JC, et al. The effect of task on determination of maximum phonation frequency range. J Voice. 2000;14:154–160. 10. Kent RD, Kent JF, Rosenbek JC. Maximum performance tests of speech production. J Speech Hear Disord. 1987;52:367–387. 11. Van Oordt HWA, Drost HA. Development of the frequency range of the voice in children. Folia Phoniatr Logop. 1963;15:289–298. 12. Cooper M, Yanagihara N. A study of the basal pitch level variations found in the normal speaking voices of males and females. J Commun Disord. 1971;3:261–266. 13. Coleman RF. Sources of variation in phonetogram. J Voice. 1993;7:1–14. 14. McClelland DC. Human Motivation. Cambridge: Cambridge University Press; 1987. 15. Reich AR, Mason JA, Polen SB. Task administration variables affecting phonation-time measures in third-grade girls with normal voice quality. Lang Speech Hear Serv Sch. 1986;17:262–269.
Journal of Voice, Vol. ■■, No. ■■, 2016 16. Soman B. The effect of variations in method of elicitation on maximum sustained phoneme duration. J Voice. 1997;11:285–294. 17. Williams AM, Hodges NJ. Skill Acquisition in Sport: Research, Theory and Practice. London: Routledge; 2004. 18. Lewis K, Casteel L, McMahon J. Duration of sustained /a/ related to the number of trials. Folia Phoniatr (Basel). 1982;34:41–48. 19. Neiman GS, Edeson B. Procedural aspects of eliciting maximum phonation time. Folia Phoniatr (Basel). 1981;33:285. 20. Stone RE. Issues in clinical assessment of laryngeal function: contraindications for subscribing to maximum phonation time and optimum fundamental frequency. In: Vocal Fold Physiology: Contemporary Research and Clinical Issues. 1983:410–424. 21. Austin MD, Leeper HA. Basal pitch and frequency level variation in male and female children: a preliminary investigation. J Commun Disord. 1975;8:309–316. 22. Heylen L, Wuyts FL, Mertens F, et al. Evaluation of the vocal performance of children using a voice range profile index. J Speech Lang Hear Res. 1998;41:232–238. 23. Ma EP-M, Robertson J, Radford C, et al. Reliability of speaking and maximum voice range measures in screening for dysphonia. J Voice. 2007;21:397–406. 24. Coleman RF, Mabis JH, Hinson JK. Fundamental frequency—sound pressure level profiles of adult male and female voices. J Speech Lang Hear Res. 1977;20:197–204. 25. Zelaznik HN. Advances in Motor Learning and Control. Champaign: Human Kinetics; 1996. 26. Finnegan DE. Maximum phonation time for children with normal voices. J Commun Disord. 1984;17:309–317. 27. Boyle MO, Ackerman PL. Individual differences in skill acquisition. In: Skill Acquisition in Sport. 2004:84–102. 28. Bless D, Hirano M. Verbal instructions: a critical variable in obtaining optimal performance for maximum phonation time. Paper presented at the Annual Convention of the American Speech-Language-Hearing Association, Toronto, 1982. 29. McAllister A, Sederholm E, Sundberg J, et al. Relations between voice range profiles and physiological and perceptual voice characteristics in ten-year-old children. J Voice. 1994;8:230–290. 30. Hallin AE, Fröst K, Holmberg EB, et al. Voice and speech range profiles and Voice Handicap Index for males—methodological issues and data. Logoped Phoniatr Vocol. 2012;37:47–61.
INTRODUCTION Maximum phonational frequency range (MPFR) is a test of maximum vocal performance. It refers to the frequency range from the lowest pitch in the modal register to the highest pitch in the falsetto register that an individual can produce.1 Glottal fry is commonly excluded because it is not continuously used in speech.1 MPFR reflects the vocal capacity of an individual. It reveals the biomechanical and physiological limits of the respiratory and phonatory systems.2–4 Clinically, a reduction in MPFR can be a sign of voice problems.5 MPFR is one of the most frequently obtained measures in voice evaluation.6,7 Currently, there is no standardized procedure for eliciting MPFR, which makes the comparison of MPFR data across voice clinics and research difficult.4,8,9 The variability of MPFR reported in the literature is rather large.10 Van Oordt and Drost11 reported that children aged 6–16 years could present MPFRs from 1.5 to 3 octaves. Reich et al8 showed that the MPFRs in children aged 6–13 years could range from 1 to 3.6 octaves. Different task variables can lead to MPFR variability across individuals. These variables include instructions to clients, elicitation tasks,3,4 time of day,12 coaching by the clinician, visual feedback, and repeated trials.3,4,9,13 Cooper and Yanagihara12 examined the influences of the time of day on the lowest phonational frequency in a group of vocally healthy adults. Their results showed that the lowest phonational frequency varied from one to three semitones (STs) throughout the day. Zraick et al4 compared the effects of two elicitation procedures, mid-basal-to-ceiling and mid-ceiling-to-basal, on MPFR in adults and found no significant difference. Reich et al3 investigated the MPFRs of 40 children from grades 3 to 6. They found that discrete step produced sig-
nificantly smaller MPFRs than elicitation tasks such as slow and fast steps, and slow and fast glissando. Because MPFR production requires an individual’s maximum vocal effort, factors that enhance motivation are expected to promote better MPFR performance. Examples of such factors include verbal encouragement and coaching provided by the clinician.3,13 Coleman13 and Kent et al10 suggested that the presence of clinician coaching may increase one’s motivation. According to McClelland,14 extrinsic motivation created by external sources such as encouragement and incentives offered by others can increase one’s self-confidence and intent to achieve one’s goal. Early studies suggest positive effects of coaching in the form of verbal encouragement on maximum phonation time elicitation.15,16 However, whether a similar positive influence of coaching can be applied for MPFR has not yet been studied. Practice through repeated trials is necessary for motor performance improvement.17 Superior performance elicited through repeated trials has been reported for maximum phonation time, with more than three trials needed to elicit a representative maximum phonation time in children.18 Some speakers have required up to 15 trials to achieve their maximum phonation time.19,20 Currently, three trials are commonly used to elicit MPFR in children3,21 and adults.8,9,12 One earlier study implemented as many trials as were needed to satisfy the researcher and the subject.1 Whether the use of repeated trials promotes larger MPFR has yet to be proven. The present study aimed to investigate the effects of coaching and repeated trials on MPFR in children. It was hypothesized that coaching and repeated trials could promote larger MPFR. METHOD
Accepted for publication May 16, 2016. From the Voice Research Laboratory, Division of Speech and Hearing Sciences, The University of Hong Kong, Hong Kong. Address correspondence and reprint requests to Estella P.-M. Ma, Voice Research Laboratory, Division of Speech and Hearing Sciences, The University of Hong Kong, 7/F Meng Wah Complex, Pokfulam, Hong Kong. E-mail: [email protected] Journal of Voice, Vol. ■■, No. ■■, pp. ■■-■■ 0892-1997 © 2016 Published by Elsevier Inc. on behalf of The Voice Foundation. http://dx.doi.org/10.1016/j.jvoice.2016.05.013
Participants Thirty girls between the ages of 6 and 11 years (mean age = 8.97 years, SD = 2.00) were recruited. The lower age limit was chosen to ensure that the participants could comprehend and follow the instructions. The upper age limit was chosen to exclude voice changes secondary to puberty. All of the participants were Cantonese native speakers who had normal voice quality, as judged
ARTICLE IN PRESS 2
Journal of Voice, Vol. ■■, No. ■■, 2016
TABLE 1. Participants’ Age, Weight, and Height in the Coaching and the Non-coaching Groups Age (in years) Group Coaching (N = 15) Non-coaching (N = 15)
Weight (in kg)
Height (in cm)
Mean
(SD)
Mean
(SD)
Mean
(SD)
8.97 9.00
(2.00) (1.61)
26.93 31.87
(6.74) (8.85)
131.60 133.60
(13.42) (12.19)
Abbreviations: N, number of participants; SD, standard deviation.
perceptually by the researchers; had not received any previous voice and singing training; and had normal hearing according to parental reports. Participants were excluded from the study if they had a previous history of or presented with a respiratory disorder or allergy, or had a previous history of or presented with any form of speech or language disorder or delay. The participants in both groups were similar in age, weight, and height (Table 1). Equipment All of the recordings used in this study were recorded at the Voice Research Laboratory, The University of Hong Kong. The background noise was measured by a sound level meter (Quest Electronics, Model 210, Oconomowoc, WI, USA) and kept below 55 dBA throughout the recording process. Swell Real-time DSP Phonetograph version 2.0 (Phog 2.0, AB Nyvalla DSP, Stockholm, Sweden) with a Dell Pentium III 500-MHz PC computer (Dell Inc., Round Rock, TX, USA) was used to capture the recordings. Procedures The participants were randomly assigned to the coaching group or the non-coaching group. MPFRs were elicited using glissando. Under this procedure, the participants were instructed to sustain the vowel /a/ at their most comfortable pitch and loudness level. They were then asked to glide from the comfortable pitch to the lowest or to the highest frequency. The lowest and highest frequencies were elicited for 10 trials each. All of the participants sat directly in front of the computer screen for visual feedback on their vocal performance. Visual feedback was provided to facilitate maximum performance for the participants who did not have any prior musical training.22 A headmounted condenser microphone (AKG Acoustics C420, Vienna, Austria) was placed on each participant’s head with a mouth-tomicrophone distance kept at 5 cm throughout the recording. Before the actual recording, the participants were allowed to practice the pitch-gliding task three times as vocal warm-up.9 There was no instruction provided for the participants during the glissando practice. The voice samples were recorded directly into the Soundswell phonetogram Phog 2.0 (Hitech Development AB, Sweden). The phonation registration duration of the Phog 2.0 program was 25 ms. The program captured and displayed the signals in real time on a computer screen as augmented visual feedback for the participants in both groups. During the recording process, the participants in the coaching group were prompted by the clinician with verbal encouragement and a hand-sweeping gesture that traveled up or down. After each trial, they were encour-
aged to perform better in the following trial. The verbal instructions provided for the participants in the coaching group were as follows: First trial: I want to know how high/low a pitch you can produce. Take a deep breath and then say /a/ from your most comfortable pitch to the highest/lowest pitch you can produce. Remember to go as high/low as possible. Ready? Go! [while the child was performing] Good! Keep going! From the second to tenth trials: You did a great job in the previous trial. Now, I want you to do that again. See if you can produce an even higher/lower pitch. Remember to take a deep breath and go as high/low as possible! Ready? Go! [while the child was performing] Good! Keep going! The feedback provided for the participants in the noncoaching group was as follows: First trial: I want to see how high/low a pitch you can produce. Take a deep breath and then say /a/ from your most comfortable pitch to the highest/lowest pitch you can produce. Ready? Go! From the second to tenth trials: Okay. Now, do it again. . . The lowest frequency was elicited before the highest frequency to avoid vocal fatigue.13 All of the participants repeated 10 downward trials before producing 10 upward trials. The whole data collection process took about 30 minutes. Six of the participants (20% of the 30 participants) underwent the same procedure 2 weeks after the first data collection. This was to evaluate the test-retest reliability of the recording procedure.
Data analysis Three measures were derived for each participant: highest phonational frequency, lowest phonational frequency, and MFPR. For each participant, the lowest frequency across all of the trials was regarded as the lowest phonational frequency. Similarly, the highest frequency across all of the trials was regarded as the highest phonational frequency. The MFPR was calculated as the difference between the highest and the lowest frequencies. Because the fundamental frequency values in hertz are linear in scale, but the perception of sound is logarithmic, the frequency range data were converted to a logarithmic scale in STs. This provided a standard comparison of the frequency ranges of the coaching and non-coaching groups.23 The frequency range was converted from hertz to STs using the following algorithm: MPFR
ARTICLE IN PRESS Estella P.-M. Ma and Trista K.-Y. Li
3
Effects of Coaching and Repeated Trials on MPFR in Children
TABLE 2. Test-retest, Inter- and Intra-rater Reliabilities, and Agreement in the Lowest and Highest Frequencies Percentage (%) Measure Lowest frequency Intra-rater Inter-rater Test-retest Highest frequency Intra-rater Inter-rater Test-retest
Pearson r
Exact Agreement
Within One ST
Within Two STs
0.962* 0.811* 0.824*
81.7 53.3 26.7
91.7 65.0 45.0
100.0 88.3 75.0
0.998* 0.974* 0.800*
91.7 66.7 13.3
95.0 78.3 25.5
98.3 95.0 58.3
* Significance level at .0001. Abbreviation: ST, semitone.
in STs = (log10 [the highest frequency in hertz/the lowest frequency in hertz])/log102 × 12.
MPFR across trials Table 3 lists the means and standard deviations of the frequency measures in both groups across the 10 trials. Table 4 lists the means, standard deviations, and the ranges of the frequency measures in both groups.
Effects of repeated trials on MPFR Table 5 lists the means and standard deviations of the lowest frequency (Hz), the highest frequency (Hz), and MPFR (Hz/ST) in the first, third, fifth, and tenth trials. To investigate the main effects of repeated trials and their interaction with the effects of coaching and repeated trials, a two-way mixed analysis of variance was used with the group as the between-subject factor (coaching versus non-coaching) and the number of trials (1, 3, 5, or 10 trials) as the within-subject factors for each measure. The Mauchly test of sphericity for within-subject factors was significant (P = 0.0001), indicating that the assumption of compound symmetry had been violated. Thus, the results of the withinsubject effects with a Greenhouse-Geisser epsilon correction, which corrected the degree of freedom, were reported. The significant main effects of trials were found for the lowest frequency (F[1,3] = 18.60, P = 0.0001), the highest frequency (F[1,3] = 14.67, P = 0.001), MPFR in hertz (F[1,2] = 17.51, P = 0.0001), and MPFR in ST (F[1,2] = 28.71, P = 0.0001). A follow-up repeated one-way analysis of variance was performed within each measure to further evaluate the main effect of repeated trials. For all of the frequency measures, those elicited in 10 trials were significantly better than those elicited in one or three trials (all P < 0.01). The lowest and highest frequencies elicited in 10 trials were significantly lower and higher, respectively, and the MPFRs were significantly greater in 10 trials. All of the frequency measures elicited from three and five trials were similar (all P > 0.05), except for the MPFR in STs (P = 0.005). There were no significant group main effects or interaction effects for any of the frequency measures (P values > 0.05).
Effects of coaching on MPFR The effects of coaching were evaluated by comparing the frequency measures between the coaching and the non-coaching groups using independent t tests. Four repeated t tests were conducted for data analysis, and the alpha level was adjusted to 0.0125 (0.05/4) using the Bonferroni adjustment to avoid any potential type II errors. There was no statistical significance in any of all the frequency measures between groups.
Minimum number of trials required to elicit maximum performance Figure 1 shows the percentage of the participants who achieved the lowest and highest frequencies in each trial. To better demonstrate how closely the participants’ performances approximated the maximum frequency values, the mean percentages of the lowest and highest frequencies achieved in each trial were also analyzed (Figure 2). For the lowest frequency, the majority (80%)
Reliability of data analysis procedure As the determination of frequency measures required the visual judgment of frequency data (lowest and highest frequencies) from the display on the computer screen, we needed to establish the reliability of this procedure. The voice recordings of six of the participants (20%) were reanalyzed by the clinician for the lowest and the highest frequencies. This was to evaluate the intrarater reliability. The same recordings of these six participants were then analyzed by another clinician to evaluate the interrater reliability. RESULTS Reliability of data analysis procedure Pearson correlation coefficients were used to evaluate the reliability of the data analysis procedure used in this study. The intra-rater reliability coefficients were above 0.96 (P = 0.0001). The inter-judge reliability coefficients were above 0.81 (P = 0.0001). Table 2 lists the results of the test-retest, interrater and intra-rater reliabilities, and agreement of the lowest and highest frequencies.
ARTICLE IN PRESS 4
Journal of Voice, Vol. ■■, No. ■■, 2016
TABLE 3. Means (and Standard Deviations) for Frequency-related Measures Across 10 Trials in the Coaching and Non-coaching Groups Number of Trials Group
1
2
Lowest frequency in Hz C 171.05 162.85 (34.81) (38.00) NC 184.29 179.43 (45.39) (34.40) Highest frequency in Hz C 1786.45 2000.53 (723.40) (680.06) NC 1586.85 1523.85 (624.59) (616.19) Frequency range in Hz C 1615.39 1867.15 (739.69) (680.02) NC 1402.56 1502.42 (644.11) (669.69) Frequency range in semitones C 39.13 43.87 (10.76) (8.39) NC 36.46 38.40 (9.38) (8.84)
3
4
5
6
7
8
9
10
158.01 (33.02) 184.26 (39.92)
164.53 (27.63) 179.17 (31.29)
162.01 (33.95) 175.19 (34.20)
164.93 (26.49) 184.82 (28.10)
165.61 (26.29) 173.47 (52.96)
169.53 (31.19) 183.30 (40.48)
164.93 (30.88) 172.67 (37.46)
164.12 (26.12) 166.00 (31.78)
1685.60 (744.66) 1519.31 (580.58)
1936.86 (788.99) 1592.66 (682.92)
1815.99 (683.55) 1568.25 (644.97)
2034.95 (619.24) 1668.65 (627.26)
1895.79 (571.38) 1586.09 (631.51)
1928.73 (535.07) 1762.39 (729.40)
1932.27 (505.15) 1732.28 (712.47)
2003.81 (484.10) 1744.06 (729.50)
1904.30 (696.17) 1526.35 (659.50)
2010.40 (719.23) 1591.14 (708.05)
2026.63 (700.36) 1595.01 (705.45)
2168.21 (647.51) 1623.23 (711.86)
2181.39 (651.82) 1638.55 (712.66)
2200.04 (645.39) 1670.58 (727.47)
2206.68 (636.24) 1728.96 (767.36)
2206.68 (636.24) 1754.06 (765.18)
44.99 (7.75) 39.00 (8.92)
46.46 (6.71) 38.07 (9.08)
46.92 (6.61) 40.40 (8.59)
48.39 (5.86) 40.93 (8.60)
48.48 (5.84) 41.20 (8.56)
48.66 (5.83) 41.53 (8.71)
48.99 (5.84) 42.13 (8.88)
48.99 (5.84) 42.73 (8.83)
Note: Standard deviations are in parentheses. Abbreviations: C, coaching; NC, non-coaching.
of the participants in the coaching group could achieve their best performance by the third trial, which already represented 95% of the participants’ lowest frequency values. However, about onefourth (27%) of the non-coaching group could achieve their lowest frequency by the third trial. For the highest frequency, six trials were needed for both groups to achieve at least 95% of their highest frequency values (Figure 2B). However, by the sixth trial, 80% of the participants in the coaching group had achieved the highest frequency, compared with only 46.7% of the participants in the non-coaching group.
DISCUSSION The aim of the present study was to investigate the effects of coaching and repeated trials on determining MPFR in children. The results revealed no significant coaching effect on the lowest frequency, highest frequency, and MPFR values. However, coaching did help the participants reach their MPFR in fewer trials. In addition, the MPFRs elicited in one, three, and five trials were all significantly smaller than those elicited in 10 trials. The frequency measures (lowest frequency, highest frequency, and MPFR) obtained from the coaching and non-coaching
TABLE 4. Means (Standard Deviations) and Ranges of the Four Frequency Measures in the Coaching and Non-coaching Groups Coaching Measures (unit) Lowest frequency (Hz) Highest frequency (Hz) MPFR (Hz) MPFR (ST)
Non-coaching
Independent t Tests
Mean (SD)
Range
Mean (SD)
Range
t
df
P
135.18 (26.20) 2341.85 (636.48) 2206.67 (630.24) 48.99 (5.84)
103.8–196.0
151.79 (29.22) 1905.85 (758.23) 1754.06 (765.18) 42.73 (9.82)
110.0–207.7
−1.64
28
.11
987.8–2960
1.71
28
.10
832.2–2691.2
1.84
28
.77
28.0–56.0
2.29
28
.03
1179.7–3520 1074.7–3396.5
Note: None of the above P values are smaller than 0.017.
35.0–58.0
ARTICLE IN PRESS Estella P.-M. Ma and Trista K.-Y. Li
5
Effects of Coaching and Repeated Trials on MPFR in Children
TABLE 5. Means and Standard Deviations for the Lowest Frequency (Hz), Highest Frequency (Hz), and MPFR (Hz/ST) on First, Third, Fifth, and Tenth Trials Across the Coaching and Non-coaching Groups Measure Number of Trials
Group
First trial
C NC
Third trial
C NC
Fifth trial
C NC
Tenth trial
C NC
Lowest F0 (Hz)
Highest F0 (Hz)
MPFR (Hz)
MPFR (ST)
171.05 (34.81) 184.29 (45.39) 145.01 (31.13) 169.97 (36.07) 138.94 (26.98) 160.59 (32.83) 135.18 (26.20) 151.79 (29.23)
1786.45 (723.40) 1586.85 (624.59) 2049.31 (698.04) 1696.32 (648.31) 2165.57 (699.60) 1755.60 (700.92) 2341.85 (636.48) 1905.85 (758.2)
1615.39 (739.69) 1402.56 (644.11) 1904.30 (696.17) 1526.35 (659.50) 2020.63 (700.36) 1595.01 (705.45) 2206.67 (636.24) 1754.06 (765.18)
39.13 (10.76) 36.46 (9.38) 44.99 (7.75) 39.00 (8.92) 46.92 (6.61) 40.40 (8.59) 48.99 (5.84) 42.73 (9.82)
Note: Standard deviations are in parentheses. Abbreviations: C, coaching; NC, non-coaching.
groups were similar, suggesting no clear coaching effect on eliciting MPFR. The results were not consistent with other studies, which have shown significant coaching effect on maximum vocal performance tasks, namely, maximum phonation time15 and maximum sound pressure level.24 The lack of coaching effect in this study might be due to the provision of visual feedback during the elicitation of MPFR. Zelaznik25 suggested that any real-time feedback can facilitate one’s performance. Letting the participants visually refer to their own performances could have motivated them to perform better. In addition to such intrinsic motivation, visual feedback can simultaneously assist a participant’s learning during a task. Finnegan26 advocated that providing children with visual feedback helped them self-monitor and enhance their performance in a task measuring maximum phonation time. It is apparent that the effects of visual feedback on maximum vocal performance cannot be underestimated. Future research studies should examine the effects of coaching on MPFR between participants who are provided with visual feedback of their performances and those who are not. The frequency measures elicited in the first trial were significantly poorer than those elicited in subsequent trials for both groups. The MPFRs elicited in 10 trials were significantly greater than those elicited in one, three, or five trials. These results could be explained by practice effect on maximal performance. Repeated practice can increase one’s familiarity with the methods and procedures of the tasks.27 Bless and Hirano28 suggested that optimal vocal performances could only be elicited when sufficient practice was allowed. As Figure 1B shows, although a clinician’s coaching was provided, less than 50% of the participants in the coaching group exhibited their highest frequency by the third trial. The participants in the coaching group achieved only 90% of the highest
frequency values by the third trial (Figure 2B). The present results suggest that optimal frequency performance may not be exhibited in three trials. A radical rethinking of using only three trials to obtain MPFR in contemporary voice assessment protocols is warranted. The present results suggest that with a clinician’s coaching, three trials are recommended for eliciting the lowest frequency, and six trials are needed to elicit the highest frequency. Our data suggest that the majority of the participants came very close to reaching their lowest and highest frequencies in these trials. Such repeated trials take less than 10 minutes to complete, an acceptable amount of time in regular clinical acoustic data collection. Comparison with the literature on MPFR in children The MPFR data from the present study (mean MPFR in the coaching group: 48.99 ST; non-coaching group: 42.73 ST) were greater in magnitude than those reported in previous studies. Most early studies used three trials to elicit MPFR. However, considering the frequency values obtained by the third trial, the mean MPFRs obtained in the present study (coaching: 44.99 ST; noncoaching: 39.0 ST) were still greater than those in early reports. Reich et al3 found a mean MPFR of 27.8 ST (collapsed across gender and elicitation tasks) in three trials. McAllister et al29 reported a mean MPFR of 25.0 ST in 24 vocally healthy children aged 10 years. Heylen and colleagues22 reported a mean MPFR of 26.4 ST in 94 vocally healthy children aged 6–11 years. The variability of MPFRs elicited across studies might possibly be due to methodological differences, namely, number of repeated trials, provision of real-time visual feedback, provision of clinician’s coaching, phonation registration time, eliciting procedures (eg, manual versus automatic recording), and recording software used.30 In the present study, the provision of clin-
ARTICLE IN PRESS 6
Journal of Voice, Vol. ■■, No. ■■, 2016
FIGURE 1. Percentage of participants reaching (A) the lowest frequency and (B) the highest frequency in each trial (solid line = coaching group; dotted line = non-coaching group). ician’s coaching and visual feedback of performance might have facilitated participants to approach their true physiological vocal limits. Nevertheless, the great variability in MPFR values across studies reflects the need for standardized elicitation procedures for valid MPFR comparisons in clinical and research settings. Such elicitation procedures should also reveal children’s true physiological vocal limits.
Future research directions In the present study, only girls with normal voices were recruited. Boys and girls can differ in their learning styles, personalities, and motivations in achieving optimal performance in maximum vocal tasks. It would be interesting to compare how boys and girls react to coaching and repeated trials. Moreover, performance in children with normal voices
ARTICLE IN PRESS Estella P.-M. Ma and Trista K.-Y. Li
Effects of Coaching and Repeated Trials on MPFR in Children
7
FIGURE 2. Mean percentage of (A) the lowest frequency and (B) the highest frequency as a function of the trials (solid line = coaching group; dotted line = non-coaching group).
might differ from that in children with dysphonia. A dysphonic voice would increase the variability and hence reduce the precision of an investigation.15 Future studies should also look into the effects of coaching and repeated trials in children with dysphonia. A similar methodology could also be extended for use in adults with and without voice disorders, as the learning
curve, motivation, comorbidities, and other factors may differ substantially between children and adults. CONCLUSION The present findings suggest that coaching and repeated trials could make the elicitation of MPFR more efficient. The results
ARTICLE IN PRESS 8 suggest that with a clinician’s coaching, three trials are recommended to elicit the lowest frequency, whereas six trials are needed to elicit the highest frequency. REFERENCES 1. Hollien H, Dew D, Philips P. Phonational frequency ranges of adults. J Speech Hear Res. 1971;14:755–760. 2. Baken RJ, Orlikoff RF. Clinical Measurement of Speech and Voice. 2nd ed. San Diego: Singular Publishing Group; 2000. 3. Reich AR, Mason JA, Frederickson RR, et al. Factors influencing fundamental frequency range estimates in children. J Speech Hear Disord. 1989;54:429–438. 4. Zraick RI, Keyes MP, Montague JC, et al. Mid-basal-to-ceiling versus mid-ceiling-to-basal elicitation of maximum phonational frequency range. J Voice. 2002;16:317–322. 5. Colton RH, Casper JK, Leonard R. Understanding Voice Problems: A Physiological Perspective for Diagnosis and Treatment. 3rd ed. Philadelphia: Lippincott Williams & Wilkins; 2006. 6. Behrman A. Common practices of voice therapists in the evaluation of patients. J Voice. 2005;19:454–469. 7. Hirano M. Objective evaluation of the human voice: clinical aspects. Folia Phoniatr (Basel). 1989;41:89–144. 8. Reich AR, Frederickson RR, Mason JA, et al. Methodological variables affecting phonational frequency range in adults. J Speech Hear Disord. 1990;55:124–131. 9. Zraick RI, Nelson JL, Montague JC, et al. The effect of task on determination of maximum phonation frequency range. J Voice. 2000;14:154–160. 10. Kent RD, Kent JF, Rosenbek JC. Maximum performance tests of speech production. J Speech Hear Disord. 1987;52:367–387. 11. Van Oordt HWA, Drost HA. Development of the frequency range of the voice in children. Folia Phoniatr Logop. 1963;15:289–298. 12. Cooper M, Yanagihara N. A study of the basal pitch level variations found in the normal speaking voices of males and females. J Commun Disord. 1971;3:261–266. 13. Coleman RF. Sources of variation in phonetogram. J Voice. 1993;7:1–14. 14. McClelland DC. Human Motivation. Cambridge: Cambridge University Press; 1987. 15. Reich AR, Mason JA, Polen SB. Task administration variables affecting phonation-time measures in third-grade girls with normal voice quality. Lang Speech Hear Serv Sch. 1986;17:262–269.
Journal of Voice, Vol. ■■, No. ■■, 2016 16. Soman B. The effect of variations in method of elicitation on maximum sustained phoneme duration. J Voice. 1997;11:285–294. 17. Williams AM, Hodges NJ. Skill Acquisition in Sport: Research, Theory and Practice. London: Routledge; 2004. 18. Lewis K, Casteel L, McMahon J. Duration of sustained /a/ related to the number of trials. Folia Phoniatr (Basel). 1982;34:41–48. 19. Neiman GS, Edeson B. Procedural aspects of eliciting maximum phonation time. Folia Phoniatr (Basel). 1981;33:285. 20. Stone RE. Issues in clinical assessment of laryngeal function: contraindications for subscribing to maximum phonation time and optimum fundamental frequency. In: Vocal Fold Physiology: Contemporary Research and Clinical Issues. 1983:410–424. 21. Austin MD, Leeper HA. Basal pitch and frequency level variation in male and female children: a preliminary investigation. J Commun Disord. 1975;8:309–316. 22. Heylen L, Wuyts FL, Mertens F, et al. Evaluation of the vocal performance of children using a voice range profile index. J Speech Lang Hear Res. 1998;41:232–238. 23. Ma EP-M, Robertson J, Radford C, et al. Reliability of speaking and maximum voice range measures in screening for dysphonia. J Voice. 2007;21:397–406. 24. Coleman RF, Mabis JH, Hinson JK. Fundamental frequency—sound pressure level profiles of adult male and female voices. J Speech Lang Hear Res. 1977;20:197–204. 25. Zelaznik HN. Advances in Motor Learning and Control. Champaign: Human Kinetics; 1996. 26. Finnegan DE. Maximum phonation time for children with normal voices. J Commun Disord. 1984;17:309–317. 27. Boyle MO, Ackerman PL. Individual differences in skill acquisition. In: Skill Acquisition in Sport. 2004:84–102. 28. Bless D, Hirano M. Verbal instructions: a critical variable in obtaining optimal performance for maximum phonation time. Paper presented at the Annual Convention of the American Speech-Language-Hearing Association, Toronto, 1982. 29. McAllister A, Sederholm E, Sundberg J, et al. Relations between voice range profiles and physiological and perceptual voice characteristics in ten-year-old children. J Voice. 1994;8:230–290. 30. Hallin AE, Fröst K, Holmberg EB, et al. Voice and speech range profiles and Voice Handicap Index for males—methodological issues and data. Logoped Phoniatr Vocol. 2012;37:47–61.