You are on page 1of 11
Article Changes in Acoustic Characteristics of the Voice Across the Life Span: Measures From Individuals 4-93 Years of Age Elaine T, Stathopoulos,° Jessica E. Huber,” and Joan E, Sussman* Purpose: The purpose of he present investigation wos 1 excmine ‘acoustic voice changes ocross the lfe span. Previous voice production investigations used smoll numbors of participants, had limited age ranges, ond produced contradictory results. ‘Method: Voice recordings were made from 192 male and female parlcipants 4-93 years of age. Acoustic measures of fundamental Frequency (FO), sound pressure lovl (SP), and signal-to-noise ratio (SNR} were made. Cooficions of voriaton for FO, SPL, and SNR served as measures of varicbiliy. Variables were onclyzed seporately with stepwise regresions, using age ond sex as predictors Results: Differances in FO and SNR across the life span ware sex specifi. Male ond femcle participants demonstrated significant nonlineor trends for FO, but the trends were sronger for mele participants. Female poricipants demonstrated a similar nonlinear trond for SNR, whereas male participants demonstrated linear increases in SNR with age. Variability of FO, SPL, and SNR followed nonlinear trends, higher at younger «and alder ages. Conclusions: Changes in voice production accur throughout he lifespan, offen in « nonlinear way and diferent for male ond female individuals. Higher variability of acoustic measures of voice inbeth young end old speakers reflects chonges in anctomic strucute, physiologic mechanisms, and motor conta Key Words: voice, lfe spon, acoustic velopmental process and are appropriately discuss- ing both development and aging as one life-span phenomenon (Gallahue & Ozmun, 1995; Shadden & Toner, 1997). Speech scientists and naive listeners would agree that the sound of an individual's voice changes through out his/herlife span (e.g, Harsberger, Shrivastav, Brown, Rothman, & Hollien, 2008; Linville, 1987; Morris & Brown, 1987). Listeners can successfully identify the sex of pre- pubertal speakers as young as 4 years of age (Bennett & ‘Weinberg, 19795 Ingrisano, Weismer, & Schuckers, 1980) and can ascertain the person's decade of life; they can even distinguish a speaker's chronological age to within R esearchers are altering their perspective of the de- ity at Buffalo, NY Purdue University, West Lafayette, IN Correspondence to Elaine T: Stathopoulos:stathop@butTalo.eda Bulitor: Robert Schlauch Associate Editor: Robert Hillman Received February 8, 2010 Rovision received August 24, 2010 Accepted November 22, 2010 DOI: 10.1044/1092-4388(2010/10-0036) several years (Huntley, Hollien, & Shipp, 1987; Linville & Fisher, 1985). Perceptual indices of voice known to dif- ferentiate young speakers from old speakers include pitch, loudness, breathiness, and—especially for the elderly — ‘tremulousness (Hollien, 1987; Linville, 2002). These per- ceptual indices have been linked to acoustie measures of frequency, vocal intensity, noise, and acoustic variability. Acoustic measurements have been extensively used to make deductions about the underlying speech phys- ology. Acoustic output of the vocal tract is closely asso- ciated with underlying physiological mechanisms, and it forms a bridge to human speech perception (Forrest & ‘Weismer, 2009; Kreiman, Gerratt, & Antofianzas-Barroso, 2007). For example, acoustic measures have been used todemonstrate that manipulations of speaking rate and vocal loudness result in changes to underlying artic latory processes in speakers who are both neurologi- cally disordered/dysarthrie and neurologically normal (Tjaden & Wilding, 2004), Acoustic measurementshave also been used to examine the changes to the physiol- ogy of voiee production across development and typical aging. However, studies spanning the entire life span are lncking. Journal af Spesch, Longuage, and Hearing Ressarch # Val. 54 » 1011-1021 » Augut 2011 » © Amin SpeaclonguageHearing Auocaion 1011 Considering voice fundamental frequency (FO) first, many investigations have substantiated the decrease in, FO from infaney and/or preschool through puberty (Eguchi & Hirsh, 1969; Huber, Stathopoulos, Curione, Ash, & Johnson, 1999; Kent, 1976; Peterson & Barney, 1952; Stathopoulos & Sapienza, 1997). These changes in FO are generally considered to be resultof increased length, and mass of the voeal folds as the human body grows. ‘The change is sharper for male adolescents at puberty because of a larger increase in mass as the hormonal environment changes. Related to aging, Russell, Penny, and Pemberton (1995) as well as Brown, Morris, Hollien, and Howell (1991) found that FO in women and men generally decreased over a 50-year period. Toward the end of that period, FO decreased more in women and increased slightly in men. In both men and women, the changes in FO have been related to hormonal changes, Ieading to thinning of the folds in men and to thickening or edematous folds in women, A socond important acoustic variable reflecting un- derlying processes is sound pressure level (SPL). SPL is controlled primarily through an interaction of respi- ratory and laryngeal mechanisms, requiring coordina tion of muscular (Finnegan, Luschei, & Hoffman, 2000; Isshiki, 1965) and neural (Hirano, Kurita, & Nakashima, 1983; Nagai, Ota, Konopacki, & Connor, 2005) control systems. McAllister, Sederholm, Sundberg, nd Gramming (1994) found that women and older children approaching pubescence are capable of producing a wider dynamic range than younger normal-speaking 10-year-olds. Dy- namie range is not an indication of speaking SPL, so itis unknown whether children spoke at different SPLs than adults, A large eross-sectional study of young children and ‘young adults showed that young children used higher SPL than did the young adults when instructed to phonate a syllable train at comfortable loudness levels Stathopoulos & Sapienza, 1997); however, there were no speaker data, extending into the later years. When examining the olderend of the life-span continuum, some studies have found that voices of older adults tend to be associated with decreased vocal intensity (Linville, Skarin, & Fornatto, 1989), In fact, researchers believe that low vocal inten- sity may be an important indicator of old-sounding voices, and it is used as a target for treatment of aged speakers, (LO. Ramig et al., 2001), However, not all studies have reported higher SPL for young children or lower SPL in, older adult speakers. Data from Huber et al. (1999) did not show differences in SPL for young children versus adults. Several other authors also reported no differences, in SPL between younger and older adults (Biever & Bless, 1989; Huber, 2008; Sapienza & Dutka, 1996) Huber and Spruill (2008) reported that older adults also responded to several cues to increase loudness with the same amount of loudness change as young adults. In 1012 Jeurnal of Speech, longuage, and Hearing Research « Val. 4 © 101 sum, the findings about changes in the control of vocal, intensity across the life span are equivocal. Another variable that impacts how voices sound is the amount of additive noise in the voice signal, wich is, thought to reflect underlying processes of vocal fold clo- sure. Ithas been long known that pereeived breathiness, is closely linked to inereased glottal opening and noise (eg., Deal & Emanuel, 1978; Fritzell, Hammarberg, Gauffin, Karlsson, & Sundberg, 1986; Hollien, 1987; Linville, 2002; Sédersten, Lindestad, & Hammarberg, 1991), Further, acoustic and modeling data support the fact that women produce voice with larger glottal open- {ings and open quotients than men (Hanson & Chuang, 1999; Holmberg, Hillman, & Perkell, 1988; Klatt & Klatt, 1990; Mendoza, Valencia, Muitoz, & Trujillo, 1996; Stathopoulos & Sapienza, 1997; Titze, 1989). Hanson and, Chuang (1999) have provided a detailed analysis of the underlying causes of the increased aspiration noise in the voicing signal of women. Women’s stronger aspira- ‘tion noise results in several acoustic measures, including the following: less well-defined formants, steeper spectral, tilt, and additional high-frequeney noise around the third formant. A strong consensus is that women produce voice with more of a posterior glottal opening than men. In, typically aging adults, increased glottal gaps—in par- ticular, spindle-shaped gaps—have been substantiated by several laryngoscope studies (Biever & Bless, 19895, Honjo & Isshiki, 19805 Linville, 1992; Pontes, Yamasaki, & Behlau, 2006). Along with findings of inereased breath- iness, findings show that the glottal mechanisms of el- derly men and women display some degree of vocal fold edema, atrophy, inereased membranous glottal gap with, ‘vocal fold vibration showing greater vibratory aperio- dicity, altered mucosal wave, and reduced amplitude of vibration. These changes to the voeal folds and glottal closure result in larger amounts of noise in the voicing signal. Signal-to-noise ratio (SNR) and harmonics-to- noise ratio (HINR) are useful for quantifying the amount of noise in the voieing signal (Awan & Frenkel, 1994; Ferrand, 2002; Yumoto, Sasaki, & Okamura, 1984). Ferrand (2002) reported that elderly speakers produced voice with lower HNR (indicating more noise in the voice signal) than middle-age and young participants, where- as jitter values (period-to-period changes in FO) did not, ‘vary across participant groups. Ferrand and Yumoto et al, (1984) coneluded that a measure such as HINR is a ‘more sensitive measure for quantifying the amount of additive acoustic noise in the voicing signal than jitter. ‘There is little agreement on whether children pro- duce voice with more or less noise characteristics than adults, as the data from children with normal voice are searce. McAllister ct al. (1994) reported a high incidence of glottal opening and breathiness and/or hoarseness in, 60 school children who were 10 years of age and reduced 1021 © August 2011 voice range profiles (FO and SPL). Their stroboscopic findings indicate that 10% of the children had vocal nod- ules, and 23% had glottal openings. In children, an ab- normal glottal opening can lead to production of glottal, noise and breathier sounding voices. However, MeAllister cet al did not report HINR or SNR for these children. Data froma glottal airflow investigation of normal-speaking. children (Stathopoulos & Sapienza, 1997) show that chil- dren produce more glottal airflow than men during the closed portion of the phonatory eyele, which could lead ‘to more noise in the voice signal; however, measures of noise were not made in the study. A final factor gaining attention for examining un- derlying processes of voice production is variability. Nu- ‘merous investigators support the concept that increased, variability observed in acousticvoice measures reflects a more variable motor control system. Increased variabil- ity in children’s speech has been found in acoustic mea sures such as FO and voice onset time (Eguchi & Hirsh, 1969; Ohde, 1985; B. L, Smith, 1994; Sussman & Sapienza, 1994) and in aerodynamie measures (Stathopoulos, 1995). Additionally, increased variability of voice measures has also been associated with typical speakers as they age (Linville, 1988; Linville & Korabie, 1986; Max & ‘Mueller, 1996; Orlikoff, 1990). For example, increases inthe SDs of FO and SPL have been found in older men and women and are assoeiated with tremor (Linville, 1988; Linville & Korabic, 1986; Orlikoff, 1990). Dur- ing the developmental process, more variable acoustic ‘measures have led researchers to conclude that speech motor control must change to accommodate changing anatomical and neurophysiological systems (Callan, Kent, Guenther, & Vorperian, 2000). Orlikoff (1990), in relation to acoustic measures, also has suggested that increased variability of speech production in very old participants is not unlike the increased variability of the developing child. Empirical studies of voice characteristics have been lacking, as broad age ranges have not been studied. numbers of participants have been limited, and empi~ ical findings have been contradictory across investiga tions (Linville, 2001). A systematic data set utilizing a large number of participants is needed to clearly charac- terize the acoustie changes that reflect typical life-span changes. The present research design allowed assess- ment of life-span acoustic changes. On the basis of pre- vious literature, we hypothesized the following: Ilbe consistent with previously published data. FO will be highest for the youngest speakers and will decrease until puberty for both sexes be- cause of growth. As speakers approach the fifth de- cade of life, FO will decrease for female participants and will inerease for male participants, reflecting hormonal changes in the body. SPL will remain steady for most of the life span, but it will decline in older adults, There will be no sex differences. 3. SNR will reflect higher noise in voice signals of younger children and older adults. There will be no sex differences in younger children, but there will be sex differences beginning at 20 years of age. 4, Variability of acoustic measures will demonstrate U-shaped curves with higher variability for young children and older adults. There will be no sex differences, Method Participants The sample included 192 (88 male, 104 female) par- ticipants who were 4-93 years of age. Exact numbers of individuals, Ms, SDs, and age ranges for the age groups are presented in Table 1. Participants had normal speech and voice for their age, as determined by a speech- language pathologist, and spoke with a general North American dialect. Participants younger than 30 years of age were required to have hearing within normal limits, ‘whereas those older than 30 years of age were required to have hearing ability adequate to follow instructions. All participants had to be able to perform the speech tasks for the voice recordings. Participants reported the absence of any major health problems, as determined by an interview and a health questionnaire, and freedom Table 1. Portcipant age summary descriptive data by sox Female participants ‘Mole participants Age (years) mn —-M(SD) Range -MSD) Range 4 6 4510.28) 40-48 6 46023) 43-48 6 6 64(024) 61-66 6 621010) 6063 8 6 851035 81-89 6 84(018) 8287 10 6 10.2(0.06) 10.1-103 6 10.4(0.20) 102-107 126 12.4(0.33) 120-128 6 12.8(0.24) 121-128 46 144 (0.28) 141-148 6 143(0.20) 141-147 1S 6 16.3 10.20) 162-167 6 163(0.18) 160-165 18 6 18,610.29) 18.2188 6 18.4(0.28) 180-187 2 6 22.40.82) 209-231 6 21.6(0.55) 208-224 30 6 31.6 (280) 296-372 4 32.4(1.97) 308-361 M6 463/282) A18-485 4 447 (3.73) 409-487 50 6 550/379) 500-582 6 54.4(3.05) 49.8-59.0 8 7 650/307) 600-888 5 65.5(390) 599-694 7 1 752{179) 730-793 6 75.4(3.43) 69.7-793 80 10 8231.54) 800-847 7 8351219) 796-863 90 4 92.2/0.88) 913-994 0 Total 104 88 Stathopouos eal: Acoustic Chavoterisies of Voice Acres the lite Span 1013 from any acute conditions such as allergies, eolds, or flu on the day of testing. Participants were nonsmokers for at least the last 5 years, and they had no professional voice training. Speech Tasks Participants were asked to sustain an [al fora *com- fortable” period of time using their “comfortable every- day pitch and loudness levels.” Three trials of the task were performed, Data collection from participants 4-25 years of age and participants 30-93 years of age was com- pleted at different times, separated by about 4 years. The FO and SPI data from the participants 4-25 years, of age were published previously in Huber et al’s (1999) study. Equipment and Procedure For children and adults 4-25 years of age, voice recordings were obtained using a free-field microphone (Quest condenser microphone with sound-level meter, ‘Model 1700) and filter/amplifier (Frequency Devices, ‘Model 901) using a constant 6-em mouth-to-microphone distance. The microphone signal was directly digitized into an IBM PC (analog-to-digital eonversion board Data ‘Translation, DT2821) using CSpeech at a sampling rate of 20 KHz (Milenkovie, 2001). The signals were low-pass filtered for anti-aliasing at 8 kHz by the Frequency De- vices filter/amplifier. A sound-level meter and the Fre~ quency Devices filter/amplifier, coupled to the microphone, amplified the microphone signal during the study. The sound-level meter was set for C-weighting. The gain ap- plied to the microphone signal by the sound-level meter and the Frequeney Devices filter/amplifier varied de- pending on how loudly an individual spoke and was, factored in when calibrating the acoustic signal for SPL, ‘measurements. Foradults 30-98 years of age, voice recordings were obtained using a free-field, head-mounted microphone (AKG C240) and preamplifier (ART Tube Amplifier) at a constant 4-cm mouth-to-microphone distance. Be~ cause there were two different: mouth-to-mierophone distance recordings, the SPL data from the 30-95-year- old group were corrected from a 4-cm to a 6-em mouth- to-microphone distance using the following formula: (20 x Log(4/6) = a ‘The microphone signal was directly digitized into aDell PC (Sound Card, Model SigmaTel C-Major, 9704) using ‘TF32 (Milenkovic, 2001) at a sampling rate of 44.1 kHz, and was low-pass filtered by the onboard SigmaTel chip at 17.64 Hz, A sound-level meter and the ART Tube amplifier, coupled to the microphone, amplified the 1014 Jeurnat of Speech, Language, and Hearing Research « Val. 4 © 101 ‘microphone signal during the study. The sound-level meter ‘was set for C-weighting. The gain applied to the micro- phone signal by the sound-level meter and the ART Tube amplifier varied depending on how loudly an individual, spoke and was factored in when calibrating the acoustie signal for SPL measurements Measurements ‘The middle interval of each vowel production (500 ms for the 4-25-year-old participants; 2,000 ms for the 30— 98-year-old participants) was extracted for data mea surement, For most samples, all three measurements listed below were based on the entire interval as one uninterrupted sample using TF32 Jitter/Shimmer/ Voice Aperiodicity SNR Analysis (Milenkovie, 2001) Ifa sample contained a voice breals, the voice break was, eliminated, and measurements were made from more than one shorter sample around the voice break and then were averaged to obtain the trial value 1. Average FO in Hz: The average FO was calculated by averaging the values over the selected mid-vowel sample and then by averaging across the three trials for each participant. 2. SPL in dB: SPL provides information reflecting the vocal intensity level of the production, SPL values ‘were calculated by comparing the root-mean-square values of vowel production with calibration tone values. Reported values were averaged across the sample. 3, SNR: SNR is the ratio of the periodic energy in the vowel signal to the energy in the aperiodic compo- nent of the voiced signal. Statistical Analysis Data were analyzed in two ways. First, to examine ‘mean differences across the broad age range, participant ‘means of the three trials were analyzed. Second, to pro- vide an estimate of variability, coefficients of variation, (COVARS) were computed for FO, SPL, and SNR. To ob- tain the COVAR, Ms and SDs seross the three trials for cach participant were computed. COVARs were ealeu- lated for each participant by dividing the SD aeross the three trials by the mean across the three trials and then multiplying the quotient by 100. Each dependent variable was analyzed separately using stepwise regressions. Age (linear regression), age squared (nonlinear regression), and sex were londed into the stepwise regression as predictors. If sex was a sig- nificant predictor, a second set of stepwise regressions ‘was completed on male and female participants indi- vidually; age and age-squared were loaded into the 1021 © August 2011 stepwise regression as predictors. Significance level was set at p <.05. Results FO Sex was a significant factor in the stepwise regres- sion (F = 140.71, p <.001, R? = 22), so further analysis, ‘was completed separately for male and female partic- pants. For female participants, both age (F = 37.60, p< .001, R® = 48) and age-squared (F = 16.87, p <.001, R® = 55) were significant factors in the stepwise regres sion, For male participants, both age (F = 122.14, p <.001, R® = .30) and age-squared (F = 88.85, p <.001, R? = .66) ‘were significant factors in the stepwise regression. The fit for the nonlinear regression was much stronger for both sexes than the linear one, and the relationship be- ‘tween FO and age was somewhat stronger for male than female partieipants (see Figure 1). In male participants, FO declined steadily from 4 to 50 years of age and then began to steadily rise. For female participants, FO de- clined, although with a shallower slope than for men, to 60 years of age and then rose slightly. SPL Sex was not a significant factor in the stepwise re- gression (F = 0.08, p = .77), so the analysis reflects male and female participants combined. The linear trend for age (F = 29.60, p <.001, R® =.13) was significant in the Figure 1. Scaiter plot of lndamental frequency (FO) by age: male participants [open circles] and femole poricipants (filed circles) Lineor end nonlineor trends for both mele and female participants Fite the dota, stepwise regression, but the nonlinear regression, age- squared (F = 0.005, p = .95), was not significant. SPL Tinearly increased with age for both female and male participants (see Figure 2). SNR Sex was a significant factor in the stepwise regres- sion (F = 134.86, p =.01, R°=.18), so farther analysis was, completed separately for male and female participants. For female participants, both age (F = 29.41, p <.001, R? =_05) and age-squared (F = 24.98, p <.001, R® = 24) ‘were significant factors in the stepwise regression, but ‘the nonlinear regression was stronger than the linear one. SNR rose to 50 years of age and then fell after 50 years, of age, indicating more additive noise in the voicing sig- nal at both young and old ages (see Figure 3). For male participants, the linear trend for age (F'= 7.43, p = .008, R® = 08) was significant in the stepwise regression, but the nonlinear regression, age-squared (F = 1.43, p = .23), ‘was not significant. For male participants, SNR linearly increased with age, indicating less additive noise as the male participants aged (see Figure 8). FO-COVAR Sex was not a significant factor in the stepwise re- gression (F = 0.00, p = .99), so the analysis reflects male and female participants combined. Both age (P = 23.24, p=.005, R° = 23) and age-squared (F = 17.86, p <.001, Re = 20) were significant factors in the stepwise regres- sion. For both female and male participants, FO-COVAR Figure 2. Scatter plot of sound pressure level (SPU by age: mole portcipans (open circles} andl femole poricipants (filed circles) Linear trend fio the data ccress sexes, 100: ‘Mean SPL (d8) 0 oo © © Ww Age (years) © Females © Males ‘Unear Femaes) = Unearttates) = = Noivear (Females) — Montes sess o © 8 10 © Females Age (years) Unsar Regression Stathopouos eal: Acoustic Charoterisies of Voice Acres the lite Span 1015 Figure 3. Scaler plot of sgnalto-nose ratio [SNR by age: mole participants (open circles) and female pestcipant filed icles). Linear and nonlinear trends ft forthe Femole paricipants, Linear trend fit forthe mele poricipants SNR ° co & © ‘Age (years) wo Lear Femaes) Uneor ses Nontner (rales) declined steadily from 4 to 30 years of age and then began to steadily rise, indicating higher FO variability for both the young and older ages (see Figure 4). SPL-COVAR Sex was not a significant factor in the stepwise re- gression (F = 0.80 p = 37), so the analysis reflects male and female participants combined. Both age (F = 12.26, p <.001, R® = 09) and age-squared (F = 16.00, p <.001, Figure 4. Scaiter plot of fundamental frequency coeficients of ‘oration (FO-COVARs) by age: male portcipants (open circles) and ferele participants filed cirdes. Linear ond nonlinear trends fi to the date across sexes 14 2 ° 0 —= g s 3 2 °F ‘Age (years) Linear Regression 1016 — Jeurnat of Speech, Language, and Hearing Research « Val. 4 © 101 R? =..03) were significant factors in the stepwise regres- sion, For both female and male participants, SPL-COVAR, declined to about 20 years of age and then started to rise atabout 60years of age, indicating higher SPL variability for young and older speakers (see Figure 5). SNR-COVAR Sex was not a significant factor in the stepwise re- gression (= 0.08, p =.77, so the analysis reflects male and female participants combined. The nonlinear re- gression, age-squared (F = 6.71, p = 01, R? = 03), was signifieantin the stepwise regression, but the linear trend, age (F = 3.61, p = .06), did not add significantly to the ‘model. For both female and male participants, SNR- COVAR declined slightly to about 20 years of age and started to rise at 50 years of age, indicating higher SNR variability for younger and older speakers (see Figure 6). Discussion ‘The purpose of the present cross-sectional investiga tion was to examine age- and sex-related aeoustie voice characteristics across the life span, beginning in early childhood at 4 years of age and ending with the very old at 93 years of age. The present acoustie measures are discussed in terms of the reflected changes to the under- lying anatomical and physiological components involved in voice production, FO was included in this examination, despite the large amount of data already available from previous Figure 5. Scatter plo of sound pressure level coefficients of variation (SPLCOVARs by age: male paticipants (open circles) ond femole participants (filed circles). Linear ond nonlinear trends fio the data SPL-COVAR om 4 6 6 100 ° rena ‘Age (years) NoninaseRegrson 1021 © August 2011 Figure 6. Scaltor plo of signal-to-noise ratio coolicions oF variation (SNR-COVARs) by age: mae participants (pen circles) and female portciponts {filed circles). Nonlinear tend ifthe data across sexes 0 20 ° SNR-COVAR pws Age (years) studies, because of the need for systematic replication of FO data across a broad age range, Our predictions ‘were generally confirmed. In male participants, FO de- clined steadily from 4 to 50 years of age and then began to steadily rise. In female participants, FO declined steadily from 4 to 60 years of age but with a shallower slope than that of male participants. Women’s FO begin to rise slightly but not until 80 years of age. An obvious explanation for young children's steadily lowering FO until puberty is inereased length and mass of the vocal folds, (e.g, Stathopoulos, 2000). In middle-age women, the low- ‘ring of FO has always been explained by menopausal hor- monal changes causing thickened or edematous folds. In middle-age men, the vocal folds thin (Brown et al., 1991 Eguchi & Hirsh, 1969; Huber et al., 1999; Kent, 1976; Peterson & Barney, 1952; Stathopoulos & Sapienza, 1997). Abitbol, Abithol, and Abitbol (1999) studied 100 postmenopausal women who were professional voice users. A small number of the women (17) showed elin- ical level voice symptoms, and another 83 women showed slight changes to the voice, including lower vocal inten- sity, increased vocal fatigue, decreased range of high frequencies, and loss of timbre in their spoken and sing- ing voices. The changes in women’s voices were asso- ciated with a change in their hormonal environment, including increased production of progesterone, which in ‘turn was hypothesized to result in atrophy of the vo- cal fold mucosa and thyroarytenoid muscles of the vocal folds The data from Raj, Gupta, Chowdhury, and Chadha (2010) are in agreement with those of Abitbol etal. (1999); as estrogen decreased and progesterone increased, sub- stantial changes occurred in the women's voices. Close ‘examination of the data shows that FO decreased in both women and men, with most of the FO decline occurring bby 50 years of age. Older men demonstrated a relatively large increase in FO after 50 years of age, potentially reflecting a decrease in the mass of the folds because of changes to the hormonal environment. It is clear that la- ryngeal cartilages and muscle function change as a func- tion ofage. Although length characteristies do not seem to change in the aging larynx, there are other multiple ‘changes to the eover, lamina propria, and body (muscle) of the vocal folds that may account for the changes in FO (Hirano et al., 1983). A more recent investigation ex- plored the reasons for'some of the voeal parameter changes in the elderly male voiee (Gugatschks et al., 2010). Men who displayed increased FO were also found to have lower estrogen levels than age-matched controls, Unlike FO data, there were very little previously published data on SPL and SNR across the life span. The prediction for SPL on the basis of the limited literature was that SPL would romain steady for most of the life span but would decline in older adults. It was, therefore, somewhat surprising that data from the present inves- tigation indicated that SPL increased as speakers aged. Huber et al, (1999) found no differences in SPL between children and young adults. Several other authors also found no differences in SPL between younger and older adult speakers Biever & Bless, 1989; Huber, 2008; Huber & Spruill, 2008; Sapienza & Dutka, 1996). Overall, on the basis of previous literature and the data from the present study, it seems that average SPL does not reflect the declining laryngeal system, However, variability of SPL shows a different result and may be a more sen- sitive measure of aging effects. ‘The third predietion was that SNR would demon- strate a U-shaped curve, with higher noise in voices of younger children and older adults. This prediction was partially supported by the data from the present study. Female participants followed the predicted trend. A lower SNR for the young and older female participants reflects, more additive noise in the voiced signal (see Figure 3). As, men aged, SNR steadily increased, indicating less ad- ditive noise in the voicing signal. The fact that women displayed higher noise characteristics than men begin- ning in their 50s fits with what we know about female hormonal changes. The average age for menopause is 51 years, and it is well-known that the hormonal changes associated with menopause affect bone, skin, and mus- le, as well as other body systems such as the eardio- vascular system (National Institute on Aging, 2008). ‘There is a plausible explanation for why women produce voice with more additive noise than men. Along with the acoustic studies mentioned in the previous section of this, article (Hanson & Chuang, 1999; Mendoza et al., 1996), vidoostroboscopic investigations of both young and older women with normal voices showed that both groups Stathopouos eal: Acoustic Charoterisies of Voice Acres the lite Span 1017 displayed glottal gap configurations during phonation. Younger women tended to display a posterior glottal gap configuration, whereas older women displayed more an- terior gap or spindle-shaped configurations (Linville, 1992; Pontes, Kyrillos, Behlau, De Biase, & Pontes, 2002; Pontes etal., 2006). Hirano, Kiyokawa, and Kurita (1988) provided data linking the larger posterior opening in, women to the morphological structure of the posterior glottis. The vocal fold angle is larger in women than in men, As women age and go through the soft tissue/ cartilage changes that are inevitable, these changes aro coupled with the incomplete glottal closure pattern. The plausible result is an even noisier voice for older women than for older men, The present large cross-sectional study presents new data showing that aging affects men and women differently—in particular, relative to noise characteristies of the voice Children also demonstrated increased SNR, in both. ‘male and female participants. There are several laryn- geal differences in children that may explain this find- ing. These differences include the following: an immature voeal ligament until 20 years of age, differences in the texture of the laryngeal cartilage and the shape of ar- ticular surfaces, and less dense ligamentous fibers such, as the macula flava (Hirano et al., 1983; Kahane, 1978). ‘These anatomic and physiologic changes may cause chil- dren to have reduced vocal fold closure and are likely to result in increased spectral noise for the children. The present data establish that young children produce voice with more additive noise than young and middle-age adults. COVARS were calculated as indicators of variabil- ity for each of the measures. There were important life- span effects for variability, but none were sex specific. ‘All three variability measures showed significant U-shaped curves, indicating more variability at younger and older ages. F0-COVAR declined to 30 years of age and then rose again, These data showed that both children and older adults demonstrated higher FO variability. Higher vari- ability in older adultshas been associated with tremor in, the voice Linville, 1988). Vocal tremor may be caused by both laryngeal and respiratory tremor(Farinella, Hixon, Hoit, Story, & Jones, 2006) and may reflect instability in the underlying neurological substrates in older adults (Aronson & Hartman, 1981; Duffy, 2005). Young children, have been shown to have less control over their vocal, fold tension, possibly because of immature motor be- havior or because of shorter vocal folds. Forexample, in Stathopoulos and Sapienza’s (1997) study, FO increased with increased SPL, and subglottal pressure was much larger for the 4-and 6-year-old children than for the adults ‘The reduced control of vocal fold tension could help ex- plain the present findings for FO variability. Alternatively, 1018 —Jeurnat of Speech, Longuage, and Hearing Research « Val. 4 © 101 the increased variability eould be linked to physical substrates such as reduced differentiation of the vocal, fold layer structure/vocal ligament. ‘Higher SPL and SNR variability at young and older ages can be explained by changes to the laryngeal his- tologie structure for both young children and aging adults. Laryngeal cartilages display different levels of ossifica- tion with additional age-dependent structural changes to the cover and layer structure of the vocal folds (Gray, Hirano, & Sato, 1993; Hirano et al., 1983; Kahane, 1983; Sato, Hirano, & Nakashima, 2002). It is likely that the changes to laryngeal cartilages, soft tissue, and muscle affect the control of adductory forces that are so impor- tant to the overall control of voice produetion. SPL and, SNR show higher variability for the younger and older speakers, reflecting changes to the physical substrates at both ends of the life-span continuum. ‘The variability findings in the present study concur with findings in previous studies, both in acoustic mea sures and kinematic measures. Numerous investigators, have supported the concept that increased acoustic vari- ability of voiee measures reflects a more variable motor control system (Eguchi & Hirsh, 1969; Linville, 1988; Linville & Korabie, 1986; Max & Mueller, 1996; Ohde, 1985; Orlikoff, 1990; B. L. Smith, 1994; Sussman & Sapienza, 1994), Increased variability of nonacoustie mea sures such as the spatiotemporal index in both young children and older adults, as compared with that of young adults, substantiates the acoustie findings of inereased variability at the ends of the life span (A. Smith & Zelaznik, 2004; Walsh & Smith, 2002; Wohlert & Smith, 1998). A. Smith and colleagues as well as other research- ers have hypothesized that increased spatiotemporal, index oceurs in both children and older adults because of changes in motor control and neuroanatomical sys- tems (Callan et al., 2000; L. A. Ramig & Ringel, 19835, A. Smith & Zelaznik, 2004; Wohlert. & Smith, 1998), As mentioned earlier, Orlikoff (1990) has likened in- creased variability of speech production in very old par- ticipants to increased variability of the developing child. ‘This hypothesis was supported in the data from the pres- ent study in which all three variability measures dem- onstrated significant U-shaped curves. Without a life-span, study such as the present one, it would be difficult to discern these kinds of nonlinear trends across the life span. Both child and adult voices change because of ‘their changing anatomy and physiology. Children are undergoing maturation, whereas older adults are un- dergoing a lessening and weakening of their anatomical, and physiological systems (Gallahue & Ozmun, 1995), References Abitbol, J., Abitbol, P., & Abitbol, B. (1998). Sex hormones and the female voice. Journat of Voiee, 13, 424 448, 1021 © August 2011 Aronson, A. E., & Hartman, D. E. (1981), Adductor spastic ysphonia as a sign of essential (voice) tremor. Journal of ‘Speech and Hearing Disorders, 46, 52-58. Awan, 8. N,, & Frenkel, M. L. (1994). Improvements in esti- ‘mating the harmonies-to-noise ratio of the voice, Journal of Voice, 8, 255-262. Bonnett, S, & Woinborg, B. (1979). Soxual charactoristies of preadolescentchildren's voices. The Journal ofthe Acoustical Society of America, 65, ¥79-189. Biever, D. M,, & Bless, D. M. (1989). Vibratory characte tics of the vocal folds in young adult and geriatric women, Journal of Voice, 8, 120-131 Brown, W. 8. Jr, Morris, R. J., Hollion, H., & Howell, E. (1991), Speaking fundamental frequency characteristies a a function of age and professional singing. Journal of Voice, 5, 310-315, Callan, D. E., Kent, R. D., Guenther, F. H., & Vorperian, 1H, K. (2000), An auditory-feedback-based neural network ‘model of speech production that is robust to developmental changes in the size and shape of the articulatory system, Journal of Speech, Language, and Hearing Research, 43, ‘Ta 736. Deal, R. B,, & Emanuel, F, W. (1978), Some waveform and spectral features of vowel roughness. Journal of Speech and Hearing Research, 21, 250-264, Duffy, J. R. (2005). Hyperkinetic dysarthrias. In J. R. Dufly (Ed,), Motor spooch disorders: Substrates, differential diag: nosis, and management (2nd ed., pp. 275-289). St. Louis, MO: Blsevier Mosby. Eguchi, 8, & Hirsh, I. J. (1969). Development of speech sounds in children. Acta Oto-Laryngologiea, 257/Suppl.), 1 51. Farinella, K. A., Hixon, T. J., Hoit, J. D., Story, B.H., & Jones, P. A. (2008). Listener perception of respiratory” induced voice tremor, American Journal of Speech-Language Pathology, 15, 72-84. Ferrand, C. T. (2002). Harmonies-to-noise ratio: An index of vocal aging. Journal of Voice, 16, 480-487. Finnegan, E. M., Luschei, E. S., & Hoffman, H. T: (2000). ‘Modulations in respiratory and laryngeal activity associated with changes in vocal intensity during speech, Journal of Speech, Language, and Hearing Research, 43, 994-950. Forrest, K., & Weismer, G. (2009), Acoustic analysis of motor speech disorders. In M. R. MeNeil (Ba.), Clinioa! manage- ment of sensorimotor speech digorders (2nd ed, pp. 46-63) New York, NY: Thieme Medical Publishers. Fritzell, B., Hammarberg, B., Gauffin, J., Karlsson, 1, & Sundberg, J. (1986). Breathiness and insufficient vocal fold closure. Journal of Phoneties, 14, 549-553, Gallahue, D.L., & Ozmun, J.C.(1995), Understanding motor development: An overview. In 8, Spoolman (Ed), Understand. ing motor development: Infants, children, adolescents, adults (Grd.ed., pp. 2-25). Madison, WE: Brown & Benchmark Gray, S., Hirano, ML, & Sato, K. (1998). Molecular and cellular structure of vocal fold tissue. In I. R. Titze (Ed), Vocal fold physiology: Frontiers in basic science (pp. 1-24), San Diogo, CA: Singular. Gugatschka, M, Kiesler, K,, Obermayer-Pietsch, B., Schoekler, B., Schmid, C., Groselj-Strele, A & Friedrich, G. (2010). Sex hormones and the elderly male voice. Journal of Voice, 24, 369-873. ‘Hanson, H. M., & Chuang, E, 8, (1999), Glottal character isties of male speakers: Acoustic correlates and comparison ‘with female data, The Journal of the Acoustical Society of America, 106, 1064-1077. Harnsberger, J. D. Shrivastay, R., Brown, S,, Rothman, H.,, & Hollien, H. (2008). Speaking rate and fundamental frequency as speoch cues to perceived age. Journal of Voice, 22, BS. 69, Hirano, M,, Kiyokawa, K., & Kurita, S. (1988). Laryngeal ‘muscles and glottal shaping. In O. Fujimura (Bd), Vaca? physiology voice production, mechanisms, and functions (pp. 249. 284). New York, NY: Raven Press, Hirano, M., Kurita, S., & Nakashima, T. (1989). Growth, dovolopment, and aging of the human voeal folds. In D. B. J. Abbs (Bd), Vaca! fold physiology: Contemporary research ‘and clinical issues (pp. 22-43). San Diego, CA: College-Hill, Press, Hollien, H. (1987). *Old voices": What do we really know about them? Journal of Voice, 1, 2-17. Holmberg, E. B., Hillman, R. E., & Perkell, J. 8. (1988). Glottal airflow and transglottal air prossure measurements for male and female speakers in soft, normal, and loud voice. The Journal of the Acoustical Society of America, 84, 51-529. Honjo, ., & Isshiki, N. (1980). Laryngoseopie and voiee characteristies of aged persons, Arehives of Otolaryngology, 106, 149-150. ‘Huber, J. E. (2008). Effects of utterance length and voeal Toudnoss on speech breathing in older adults. Respiratory Physiology & Neurobiology, 164, 323-330. Huber, J. E., & Spruill, J. (2008), Age-related changes to speech breathing with increased vocal loudness. Journal of Speech, Language, and Hearing Research, 51, 851-668, Huber, J. E., Stathopoulos, E. T., Curione, G. M., Ash, TA, & Johnson, K. (1999), Formants of children, women, and men: The effects of vocal intensity variation. The Journal of the Acoustical Society of America, 106, 1532-1542 Huntley, R,, Hollien, IL, & Shipp, T. (1987), Influences of listoner characteristics on perceived age estimations sTournal of Voice, 1, 49-52 Ingrisano, D., Weismer, G., & Schuekers, G. H. (1980). Sex identification of preschool children’s voiees. Folia Phontatrica, 32,61. Ishii, N. (1965). Vocal intensity and air low rate, Folia Phoniatriea, 17, 93-104. Kahane, J. C. (1978), A morphological study of the human prepubertal and pubertal larynx. American Journal of Anatomy, 151, 11-19, Kahane, J. C. (1989), Postnatal development and aging of the human larynx. Seminars in Speech and Language, 4, 189-205. Kent, R. D. (1976), Anatomical and neuromuscular matu- ration of the speech mechanism: Evidence from acoustic studies, Journal of Speech and Hearing Research, 19, 421-447, Klatt, D. HL, & Klatt, L. C. (1990). Analysis, synthesis, and perteption of voice quality variations among female and male talkors. The Journal of the Acoustical Society of America, 87, 820-887 Stathopouos eal: Acoustic Charoterisies of Voice Acres the lite Span 1019 Kreiman, J., Gerratt, B. R,, & Antofianzas-Barroso, N. (2007). Measures of the glottal source spectrum. Journal of Speech, Language, and Hearing Research, 50, 595-610. Linville, S. E. (1987). Acoustic perceptual studies of aging voice in women. Journal of Voiee, I, 44-48. Linville, S. E. (1988). Intraspeaker variability in fundamen- tal frequency stability: An age-related phenomenon? The Journal of the Acoustical Society of America, 89, 741-745, Linville, S. E. (1992). Glottal gap configurations in two age ‘groups of women. Journal of Speech and Hearing Research, 95, 1209-1218, Linville, 8. B. (2001), Acoustic aspects of aging voice. In 8. B. Linville (.), Vocal aging (pp. 169-184), San Diego, CA: Singular. Linville, 8. F. (2002) Source characteristics of aged voice asseseed from long-term average spectra. Journal of Voice, 16, 472-479. Linville, S. E., & Fisher, H. B. (1985). Acoustie charactor. istics of perceived versus actual voeal age in controlled phonation by adult females, The Journal of the Acoustical Society of America, 78, 40-48, Linville, S.E,, & Korabie, E. W. (1986), Elderly listeners’ estimates of voeal age in adult females. The Journal of the Acoustical Society of America, 80, 692-694, Linville, 8. E., Skarin, B. D., & Fornatto, E. (1989). The interrelationship of measures related to vocal function, speech rate, and laryngeal appearance in elderly women. Journal of Speech and Hearing Research, 32, 323-380. ‘Max, La, & Mueller, P. B. (1996). Speaking FO and cepstral periodicity analysis of conversational speech in a 105-year- ‘old woman: Variability of aging effects. Journal of Voice, 10, 245-251. MeAllister, A, Sederholm, E., Sundberg, J.,& Gramming, P, (1994), Relations between voice range profiles and phys- ‘logical and pereeptual voice characterises in ten-year- old children. Journal of Voice, 8, 230-239. Mendoza, E., Valencia, N., Mufior, J., & Trujillo, H. (1996) Differences in voice quality between men and women: Use of the long-term average spectrum (TAS). Journal of Voice, 10, 59-66. Milenkovie, P. H. (2001). TF32 user's manual (Computer software]. Retrieved from http:/faserpages.chorus.net/especch / Morris, R.J., & Brown, W. S. Jr. (1987). Age-related voice ‘measures among adult women. Journal of Votee, I, 38-43. ‘Nagai, H., Ota, F., Konopacki, R., & Connor, N. P. (2005), Discoordination of laryngeal and respiratory movements in aged rats. American Journal of Otolaryngology, 28, 877-382, ‘National Institute on Aging. (2008), Menopause: Time fora change. Retrieved from wwwnia.nih,gov/healthinformation/ publications /menopause/introduction him. Ohde, R.N. (1985). Fundamental frequency correlates of stop consonant voieing and vowel quality in the speech of preadoleseent children. The Journal ofthe Acousticad Society of America, 78, 1554-1561 Orlikoff, R. F. (1990), The relationship of age and cardio- vascular health to certain acoustie characteristics of male voices. Journal of Speech and Hearing Research, 33, 450-457, 1020 Jeurnal af Spach, Language, and Hearing Research + Val SA Peterson, G. E., & Barney, H. L, (1952), Control methods ‘used in a study of the vowels. The Journal of the Acoustical Society of America, 24, 175-184, Pontes, P., Kyrillos, L., Behlau, M., De Biase, N., & Pontes, A. (2002). Voeal nodules and laryngeal morphology. Journal of Voice, 16, 408-414. Pontes, P., Yamasaki, R., & Bohlau, M. (2006). Morpho- logical and functional aspects of the senile larynx. Folia Phoniatrica, 68, 151-158. Raj, A, Gupta, B., Chowdhury, A., & Chadha, S. (2010). ‘A Stull of voice changes in various phases of menstrual yee and in postmenopausal women. Journal of Voice, 24, 363-368, Ramig, L. A., & Ringel, R. L. (1983). Effects of physiological ‘aging on selected acoustic characteristies of voice. Journal of Speech and Hearing Research, 26, 22-30. Ramig, L. 0., Gray, 8, Baker, K., Corbin-Lewis, K Buder, E., Lusehei, F, ... Smith, M. (2001). The aging voice: A review, treatment data and familial and genetic perspectives. Folia Phoniatriea et Logopacdicn: Official Organ of the International Association of Logopedics and Phoniatries (ALP), 53, 952-265. Russell, A., Penny, L., & Pemberton, C. (1995). Speaking fundamental frequency changes over time in women: A longitudinal study. Journal of Speech and Hearing Research, 38, 101-108. Sapienza, C. M., & Dutla, J. (1996), Glottal airflow char: acteristies of women's voice production along an aging continuum, Journal of Speech and Hearing Research, 39, 822-828, Sato, K, Hirano, M,, & Nakashima, T. (2002), Age-related ‘changes ofollagenous fibersin the human voeal fold mucosa, ‘The Annals of Otology, Rhinology, and Laryngology, 111, 15-20, Shadden, B. B., & Toner, M. A. (1997), Introduction: The continuum of life funetions. In B. B. Shadden & M. A. Toner (Bds,), Aging and communication: For clinicians byelinicians (pp. 3-18). Austin, TX: Pro-Ed, ‘Smith, A., & Zelaznik, H. N. (2004). Development of fune- ‘tional synergies for speech motor cocrdination in child- hood and adolescence. Developmental Paychobiologs, 45; 22-88 ‘Smith, B. L. (1994). Effects of experimental manipulations ‘and intrinsic contrast on relationships between duration and temporal variability in children’s and adults speech. Journal of Phonetics, 22, 155-115. Sédersten,M.,Lindestad, P. A., & Hammarberg, B.(1991). ‘Vocal fold closure, perceived breathiness, and acoustic characteristics in normal adult speakers. In J. Gauffin & B. Hammarberg (Rds.), Vaca! fold physiology: Acoustic, perceptual, and physiological aspects of voice mechanisms (pp. 217-224). San Diego, CA: Singular, Stathopoulos, E. T. (1995). Variability revisited: An acoustic, ‘aerodynamic, and respiratory kinematic comparison of children and adults during speech. Journal of Phonetics, 1 67-80. Stathopoulos, E, T. (2000). A review of the development of the child voice: An anatomical and functional perspective. InP. White (Fd.), Child voice (pp. 1-12). Stockholm, Sweden: KTH Voice Research Centre, TOIN-1021 * August 2011 Stathopoulos, E. , & Sapienza, C. M, (1997). Develop- ‘mental changes in laryngeal and respiratory funetion with variations in sound pressure level. Journal of Speech, Language, and Hearing Research, 40, 595-614. Sussman, J. E., & Sapienza, C. (1994), Articulatory, devel opmental, and gender effects on measures of fundamental froquency and jitter. Journal of Voiee, 8, 145-156. Titze, LR. (1989). Physiologic and acoustie differences be- tween male and female voices. The Journal of the Acoustical Society of America, 85, 1699-1707. ‘Tjaden, K., & Wilding, G. E. (2004). Rate and loudness ‘manipulations in dysarthria: Acoustie and perceptual find ings. Journal of Speech, Language, and Hearing Research, 47, 766-783, Walsh, B., & Smith, A. (2002), Articulatory movements in adolescents; Evidence for protracted development of speech ‘motor control processes. Journal of Speech, Language, and Hearing Research, 45, 119-1138, Wohlert, A. B., & Smith, A. (1998), Spatiotemporal stability flip movements in older adult speakers. Journal of Speech, Language, and Hearing Research, 41, 41-50. Yumoto, E, Sasaki, ¥., & Okamura, H. (1984). Harmonies- to-noise ratio and psychophysical measurement of the degree of hoarseness. Journal of Speech and Hearing Research, 27, 2-6. Stathopoulos eal Acoustic Characteristics of Voice Acrss the ile Span 1021

You might also like