Stuttering has been described as a neurophysiological speech disorder. The Effect of Frequency Altered Feedback on Stuttering Duration and Type

Transcription

1 The Effect of Frequency Altered Feedback on Stuttering Duration and Type Andrew Stuart Carmen L. Frazier Joseph Kalinowski Paul W. Vos East Carolina University, Greenville Purpose: The effect of frequency altered feedback (FAF) on stuttering type (i.e., prolongation, repetition, or silent block) and stuttering duration (i.e., average duration of stuttering event and total stuttering time) was examined. Method: Retrospective analyses of previously collected data from 12 adult persons who stutter who participated in an ABA time-series design while reading orally was undertaken. It was hypothesized that stuttering duration would decrease and there would be a differential reduction in the type of stuttering during FAF, concurrent with previously confirmed reduction of stuttering episodes. A total of 2,971 stuttered syllables were analyzed. Results: The total stuttering duration (s/min) was significantly reduced by approximately 50% irrespective of stuttering type (p =.0014). Although significant differences in the average duration(s) of the 3 stuttering types (p =.0064) existed, FAF significantly reduced each type of stuttering by approximately 20% (p =.0055). There was no differential effect on the reduction of proportion of stuttering type during FAF (p =.36). Conclusions: FAF positively affects the speech of persons who stutter by reducing the proportion of stuttered events with a concomitant decrease in duration of residual stuttering and total stuttering time during oral reading. KEY WORDS: stuttering, duration, altered auditory feedback Stuttering has been described as a neurophysiological speech disorder that causes an involuntary disruption of a continuing attempt to produce a spoken utterance (Perkins, 1990, p. 184). Speech disruptions of people who stutter (PWS) typically include sound or syllable repetitions, sound prolongations, monosyllabic whole-word repetitions, and/or within-word pauses (Conture, 2001). Stuttering severity is typically related to the frequency and duration of these overt behaviors. Clinical assessment instruments (e.g., Stuttering Severity Instrument for Children and Adults Third Edition [SSI-3], Riley, 1994; Iowa Scale of Severity of Stuttering, Sherman, 1952) associate increasing frequency and duration of stuttering with categorical increases in the severity of the disorder. The frequency with which PWS display stuttering is on average about 10% of words during oral reading but varies largely within and between PWS across a number of situational variables (Bloodstein, 1995). The cause of stuttering remains elusive. As such, treatment is generally directed at alleviating symptoms of the disorder. Stuttering Disclosure As per the American Speech-Language-Hearing Association (2004) Conflicts of Professional Interest [Issues in Ethics], we acknowledge a potential conflict in that the first and third authors of the manuscript are co-inventors of a device that utilizes altered auditory feedback to alleviate stuttering (i.e., SpeechEasy). These authors disclose a financial interest in the form of minority ownership of non-voting common stock in Janus Development Group, Inc., the company that manufactures the device. Journal of Speech, Language, and Hearing Research Vol August 2008 D American Speech-Language-Hearing Association /08/

2 treatment outcomes are typically focused at reducing the overall number of stuttering episodes; producing natural and automatic sounding speech and speech rate; normalizing abnormal speech attitudes; and reducing the duration of stuttering episodes (Bloodstein, 1995). Presently, therapeutic alternatives include prolonged speech variants or speech restructuring treatments that incorporate training to reduce speech rate via specific articulatory/vocal targets (Van Riper, 1973; Wingate, 1976), pharmaceutical agents (Maguire, Yu, Franklin, & Riley, 2004), and self-contained ear-level devices delivering altered auditory feedback (AAF; Stuart, Kalinowski, Rastatter, Saltuklaroglu, & Dayalu, 2004; Stuart, Kalinowski, Saltuklaroglu, & Guntupalli, 2006; Stuart et al. 2003). Although reports of these therapeutic alternatives in reducing the frequency of stuttering events can be found, evidence of any effects on the duration of residual stuttering events during treatment have only been documented for pharmaceutical agents (Maguire, Riley, Franklin, & Gottschalk, 2000; Maguire, Riley, et al., 2004; Murray, Kelly, Campbell, & Stefanik, 1977; Prins, Mandelkorn, & Cerf, 1980; Rosenberger, Wheelden, & Kalotkin, 1976; Rustin, Kuhr, Cook, & James, 1981; Stager et al., 2005; Wells & Malcolm, 1971), with equivocal results, and AAF. To date, only Martin and Haroldson (1979) have reported the effect of AAF on stuttering duration in PWS. Specifically, they examined the effects of five treatments in 20 adults who stuttered, including delayed auditory feedback (DAF), in an AB design in which PWS spoke spontaneously for 20 and 30 min during baseline and treatment conditions, respectively. The DAF delay interval was 250 ms. Martin and Haroldson reported a significant 85% reduction in stuttering frequency and a concomitant significant 38% reduction in the duration of stuttered events ( p <.01). Mean duration of stuttering, measured manually, was 1.37 s and 0.85 s for baseline and treatment conditions, respectively. Frequency altered feedback (FAF), which entails spectral shifting of the speaker s voice either up or down, is another form of AAF that has proven to be effective in reducing disfluency in PWS (Armson, Foote, Witt, Kalinowski, & Stuart, 1997; Armson & Stuart, 1998; Hargrave, Kalinowski, Stuart, Armson, & Jones 1994; Howell, El-Yaniv, & Powell, 1987; Howell, Sackin, & Williams, 1999; Kalinowski, Armson, Roland-Mieszkowski, Stuart, & Gracco, 1993; Kalinowski, Stuart, Wamsley, & Rastatter, 1999; MacLeod, Kalinowski, Stuart, & Armson, 1995; Stuart, Kalinowski, Armson, Stenstrom, & Jones, 1996; Stuart, Kalinowski, & Rastatter, 1997; Zimmerman, Kalinowski, Stuart, & Rastatter, 1997). Stuttered syllables are typically reduced by 50% 90% during FAF relative to nonaltered auditory feedback (NAF) during oral reading. Three observations are worth noting with regard to stuttering reduction during FAF: Less robust effects have been observed with spontaneous speech (Armson & Stuart, 1998; Ingham, Moglia, Frank, Costello-Ingham, & Cordes, 1997; Natke, Grosser, & Kalveram, 2001); considerable individual variability exists; and stuttering is typically not completely eliminated (e.g., Armson et al., 1997; Armson & Stuart, 1998; Hargrave et al., 1994; Kalinowski et al., 1993; MacLeod et al., 1995; Stuart et al., 1996, 1997, 2004, 2006). These observations are interesting, in light of the documented positive self-report measures from PWS who wear devices that utilize AAF (Kalinowski, Guntupalli, Stuart, & Saltuklaroglu, 2004; Stuart et al., 2006). Positive self-reported measures may be related to not only overall reduction in stuttering but also positive changes in 10% 50% of residual stuttering episodes. For example, even though residual stuttering persists during FAF, the duration of residual stuttering episodes may decrease, contributing to the self-reported reduction in the perception of struggle, avoidance, and expectancy. Further, the variability in individual responsiveness to FAF may be related to the typology of stuttering. It has been our impression from observations of many research participants and clinical patients that PWS who have stuttering predominately characterized by silent blocks are less likely to benefit from AAF. We have speculated that this is due to the fact that when one experiences a silent block there is no auditory alteration in their speech signal to enhance fluency. Consequently, there would be a smaller reduction in stuttering episodes during AAF relative to someone whose stuttering was not dominated by silent blocks. In addition, it stands to reason, there would be greater reductions in types of stuttering that were audible (i.e., prolongations and repetitions) where alteration in the speech signal could be heard by PWS. The purpose of this study was to explore these speculations by examining duration and stuttering type as a function of auditory feedback (i.e., FAF vs. NAF) in adult PWS during oral reading. We performed retrospective analyses of previously collected data from 12 adults who stutter who had participated in an ABA time-series design (Armson & Stuart, 1998). It was hypothesized that stuttering duration would decrease as a result of FAF concurrent with the reduction of the frequency of stuttering episodes, consistent with the previous report of Martin and Haroldson (1979). It was also hypothesized that the proportion of types of disfluencies would differ during FAF relative to NAF (i.e., a greater amount of silent blocks vs. prolongations and repetitions). Method Participants The participants were 12 adults who stutter (Armson & Stuart 1998). They included 10 males and 2 females with an average age of 35 years (SD = 8.3; 890 Journal of Speech, Language, and Hearing Research Vol August 2008

3 range years). Using the SSI-3 (Riley, 1994), the severity of 5, 6, and 1 of the participants were assessed as mild, moderate, and severe, respectively. Participants did not report any other speech or language problems nor were any reported in previous clinical reports. Participants presented with normal hearing sensitivity, defined as having hearing thresholds of 25 db HL or better at octave frequencies from 250 to 8000 Hz and normal bilateral middle ear function. Four participants were receiving therapy at the time of data collection, whereas the remaining 8 had received therapy in the past ranging from as recently as 1 month to as long as 15 years previous. Apparatus In the original study (Armson & Stuart, 1998), participants spoke into a microphone (JVC Model M-510) whose output was sent in series to an audio mixer (Studiomaster Model Session Mix), a digital signal processor (Yamaha Model DSP-1), and an amplifier (Sansui Model AU-55900). In the FAF condition, the signal was shifted in frequency up one-quarter octave for one half of participants chosen randomly and down one-quarter octave for the remaining participants by the processor before being returned to the participants ears via insert earphones (EAR Tone Model 3A). The digital processing for the FAF effect introduced a negligible and imperceptible delay of 0.1 ms. These shifts were chosen originally, as frequency shifts of the same magnitude are equally effective regardless of the direction of the shift (Hargrave et al., 1994; Stuart et al., 1996). Armson and Stuart reported no significant effect of the direction of the FAF shift on the total number of syllables produced, number of stuttering events, or percentage of stuttering during reading or monologue. Earphone output was calibrated to approximate real-ear average conversation sound pressure levels of speech outputs consistent with auditory selfmonitoring during normal conversation (see Kalinowski et al., 1993, for a detailed description of this calibration procedure). Speech samples were video recorded with a camera (JVC Model S-62U) and video stereo cassette recorder (JVC Model BR-6400U). The original analog videotapes from Armson and Stuart (1998) were digitized for subsequent analyses. Using a videocassette recorder ( VCR; Panasonic Model AG 6200), the original tapes were input into a digital VCR (Sony Model DRS-20). The two-channel audio videotapes were sampled at 48 khz. The video analogue signal was D1 compressed with a 4:1:1 column sampling rate. The digitized signal was input to a personal computer (Power Mac G4 dual 1.42-GHz PowerPC, 512 MB RAM, GB hard drive; Apple Computer) via a FireWire peripheral cable. Again, the sampling rate was 48 khz. Final video editing was completed with imovie (Apple Computer, 2003). Final audio editing was undertaken with BIAS. Peak 4.5 DV digital audio editing software. Procedure In the original study (Armson & Stuart 1998), each participant read for 20 min in an ABA time-series design (i.e., 5 min of NAF [NAF-1], 10 min of FAF, and 5 min of NAF [NAF-2]). Twenty-two passages from junior-highlevel texts of approximately 300 syllables served as text. During the NAF conditions, participants did not wear insert earphones. Participants were instructed to speak with a normal loudness and rate. Participants were also instructed not to control or reduce stuttering by using strategies or techniques that they had learned in therapy. Two trained research assistants identified stuttering episodes in the original study (Armson & Stuart, 1998). 1 Armson and Stuart defined stuttering episodes as partword prolongations, part-word repetitions, or inaudible postural fixations (i.e., silent blocks). In this study, the originally identified stuttering episodes were subsequently measured in duration and categorized with respect to stuttering type. The same constructs (i.e., prolongations, repetitions, and silent blocks) were used in an effort to remain consistent with the original identification of stuttering events. Prolongations (e.g., mmmmy name is ) and repetitions ( my-my-my-my name is ) occurred on single syllables or single-syllable whole words (i.e., the universal unit of speech production; Kent, Adams, & Turner, 1996; MacNeilage, 1998). Duration, in seconds, was operationally defined for repetitions as the onset of the acoustic waves associated with the initial disfluent sound in the word to the cessation of the acoustic wave for the iteration (Throneburg & Yairi, 2001, p. 42) of the repetition. Prolongation duration, in seconds, was determined as the onset of acoustic wave associated with the prolonged sound to the cessation of its acoustic wave (Throneburg & Yairi, p. 42). The duration of inaudible postural fixations was examined by playing the digitized video recording in conjunction with the acoustic signal. The duration of inaudible postural fixations was operationally defined as the onset of a forced unnatural block identified in the visual modality to the cessation of the acoustic wave for the stuttered token identified in the auditory modality. Participants produced a total of 45,567 syllables during their respective 20 min of reading (i.e., 10,490, 1 Armson and Stuart (1998) reported an interjudge syllable-by-syllable agreement of.78, as indexed by Cohen s kappa. Stuart et al.: Effect of Frequency Altered Feedback 891

4 24,381, and 10,696 during NAF-1, FAF, and NAF-2 conditions, respectively). A total of 2,971 stuttered syllables (i.e., 938, 1,104, and 929 during NAF-1, FAF, and NAF-2 conditions, respectively) were categorized for type and measured for duration by the second author. A trained research assistant also examined 10% (i.e., 300 tokens) of the data set chosen at random to assess interobserver agreement. An independent samples t test was utilized to examine the averaged mean durations of stuttering episodes as a function of observer. There was no significant difference between observers, t(598) =.82, p =.41 (mean difference = 0.43 s, SE =.053, 95% confidence interval of the difference = 0.15 to 0.060). There was also a statistically significant positive Pearson correlation between the duration measures of each observer (r =.90, p <.0001). Interobserver syllable-by-syllable agreement for categorization of stuttering episodes was 0.50 ( p <.0001) as indexed by Cohen s kappa.this Cohen s kappa value represents a good (Fleiss, Levin, & Paik, 2003) or moderate (Landis & Koch, 1977) strength of agreement between observers beyond chance. Results Because participants contributed a different amount of stuttering tokens, individual mean values for total duration of stuttering events, duration of stuttering events, and proportion of stuttering events were first calculated. Mean group values were subsequently derived from these individual values. Individual average total stuttering duration was calculated by summing the duration of all stuttering events for each stuttering type and then dividing by the number of minutes per feedback condition (i.e., 5 min for NAF-1 and NAF-2 vs. 10 min for FAF; hereafter referred to as s/min). Individual average duration values were calculated by dividing the sum of stuttered syllable durations for each stuttering type by the number of stuttered syllables during each feedback condition. Individual average proportional values were calculated by dividing the number of stuttered syllables for each stuttering type by the total number of syllables produced during each feedback condition. Scatterplots of the individual averaged total stuttered syllable durations (s/min), averaged stuttered syllable durations (in seconds), and averaged proportions of stuttered syllables, as a function of feedback and stuttering type, are displayed in Figures 1, 2, and 3, respectively. Diagonal lines represent total stuttered syllable durations, averaged stuttered syllable durations, and proportions of stuttered syllables of equal magnitude in Figures 1, 2, and 3, respectively. The relative effect or improvement in the stuttering index is reflected in the amount of data points that fall below the diagonal in each scatterplot. That is, the magnitude of total Figure 1. Scatterplots of individual averaged total stuttered syllable durations (s/min) as a function of feedback and stuttering type (N = 36). Diagonal line represents total stuttered syllable durations of equal magnitude. FAF = frequency altered feedback. NAF1 = first nonaltered auditory feedback condition. NAF2 = second nonaltered auditory feedback condition. durations of stuttered syllables (s/min), averaged durations, and proportions was greater in NAF versus FAF. In each scatterplot, it is evident that the overwhelming majority of participants saw a positive improvement during FAF. Figure 4 displays the group mean total stuttering duration (s/min) as a function of stuttering type and feedback condition collapsed across participants. A compound within-blocks nonadditive two-factor repeated analysis of variance (ANOVA) with participant as a nuisance factor (Cobb, 1998) was undertaken to investigate differences in mean total durations of stuttered syllables as a function of type and feedback. A significant main effect of feedback was observed, F(2, 22) = 8.98, p = There was no significant main effect of type, F(2, 22) = 2.81, p =.082, or Feedback Type interaction, F(4, 44) = 1.87, p =.13. Two orthogonal single degree-of-freedom comparisons were completed to explore the source of the Figure 2. Scatterplots of individual averaged stuttered syllable durations (in seconds) as a function of feedback and stuttering type (N = 36). Diagonal line represents averaged stuttered syllable durations of equal magnitude. 892 Journal of Speech, Language, and Hearing Research Vol August 2008

5 Figure 3. Scatterplots of individual averaged proportions of stuttered syllables as a function of feedback and stuttering type (N = 36). Diagonal line represents proportions of stuttered syllables of equal magnitude. Figure 5. Group average mean durations (in seconds) of stuttered syllables as a function of feedback and stuttering type. Error bars represent ±1 SE of the mean. significant main effect of feedback. There was no significant difference between the durations of stuttered syllables in the two NAF conditions (estimate = 0.067, SE = 0.74, t = 0.090, p =.93; 95% confidence interval = 1.47 to 1.60). A significantly lower duration of stuttered syllables in the FAF condition relative to the two NAF conditions was evident (estimate = 2.72, SE =0.64, t =4.24,p =.00034; 95% confidence interval = 1.39 to 4.05). Figure 5 displays the group mean stuttering duration average (in seconds) as a function of stuttering type and feedback condition collapsed across participants. A compound within-blocks nonadditive two-factor repeated ANOVA with participant as a nuisance factor (Cobb, 1998) was undertaken to investigate differences in mean durations of stuttering events as a function of feedback and stuttering type. Significant main effects of feedback, F(2, 22) = 6.66, p =.0055, and stuttering type, F(2, 22) = 10.47, p =.00064, were found. The Feedback Type interaction was not statistically significant, F(4, 44) = 0.45, p =.77. Again, two orthogonal single degree-of-freedom comparisons were completed to explore the source of the significant main effect of feedback. There was no significant difference between the durations of stuttered syllables in the two NAF conditions (estimate = , SE = 0.066, t = , p =.99; 95% confidence interval = 1.14 to 0.14), whereas there was a significantly lower duration of stuttered syllables in the FAF condition relative to the two NAF conditions (estimate = 0.21, SE = 0.057, t = 3.65, p =.0014; 95% confidence interval = 0.09 to 0.33). Two orthogonal single degree-of-freedom comparisons were also undertaken to examine the source of the significant main effect of stuttering type. There was a significant difference between the durations of silent blocks and repetitions in both conditions (estimate = 0.32, SE =0.089,t =3.60,p =.0016; 95% confidence interval = 0.14 to 0.50). There was a significant difference between the durations of silent blocks plus repetitions versus prolongations in both conditions (estimate = 0.29, SE = 0.10, Figure 4. Group total mean durations (s/min) of stuttered syllables as a function of feedback and stuttering type. Error bars represent ±1 SE of the mean. Figure 6. Group mean proportions of stuttered syllables as a function of feedback and stuttering type. Error bars represent ±1 SE of the mean. Stuart et al.: Effect of Frequency Altered Feedback 893

6 t = 2.82, p =.0010; 95% confidence interval = 0.50 to 0.77). In other words, the average durations of all three types of stuttering were significantly different in both conditions with the duration of repetitions the greatest and prolongations the shortest. Figure 6 displays the group mean proportions of stuttered syllables as a function of stuttering type and feedback condition collapsed across participants. A compound within-blocks nonadditive two-factor repeated ANOVA with participant as a nuisance factor (Cobb, 1998) was undertaken to investigate differences in mean proportions of stuttered syllables as a function of feedback and stuttering type. 2 The participants proportional scores were transformed by an arcsine function prior to subjecting them to inferential statistical analysis. A significant main effect of feedback was observed, F(2, 22) = 4.10, p =.031. There was no significant main effect of type, F(2, 22) = 1.08, p =.36, or Feedback Type interaction, F(4, 44) = 1.28, p =.293. Two orthogonal single degree-of-freedom comparisons were completed to explore the source of the significant main effect of feedback. It was found that there was no significant difference between the proportions of stuttered syllables in the two NAF conditions (estimate = 0.29, SE = 0.57, t = 0.51, p =.62; 95% confidence interval = 1.46 to 0.89), whereas there were a significantly lower proportion of stuttered syllables in the FAF condition relative to the two NAF conditions (estimate = 1.39, SE = 0.49, t = 2.18, p =.0010; 95% confidence interval = 0.37 to 2.41). Discussion The fluency enhancing effects of FAF while reading orally have been previously established both in isolation and in combination with other forms of AAF (Armson & Stuart, 1998; Hargrave et al., 1994; Howell et al., 1987, 1999; Kalinowski et al., 1993; MacLeod et al., 1995; Stuart et al., 1996, 1997, 2004, 2006). Three additional findings have been revealed herein: Namely, FAF significantly reduced the total stuttering duration during oral reading, although there was no significant difference in the 2 In the previous study, Armson and Stuart (1998) presented data (i.e., number of syllables produced and percentage stuttering) for each of 20 successive 1-min intervals. In their omnibus analyses, however, data were collapsed and averaged for each baseline and experimental segment. We chose to do the same in this study. Consequently, an examination of individual or group trends across time was not possible with any degree of statistical probability or confidence. This type of statistical analysis is only achievable in time-series designs with autoregressive integrated moving average (ARIMA) models (Box & Jenkins, 1970). This analysis depends on a large number (e.g., approximately 100) of data points or observations (Box & Jenkins). One could have increased the number of data points to generate the preferred number of observations for ARIMA analysis by reducing the size of the 1-min bins (e.g., to 10 s). Doing so, however, would have reduced the variance of stuttering events within bins and in some occasions reduced them completely, making the identification of patterns within the time series impossible. total duration of stuttering as a function of type. Second, FAF significantly reduced the average duration of all types of stuttering during oral reading despite significant differences in the durations of the three stuttering types. Finally, there was no differential effect on the reduction of proportion of stuttering type during FAF while participants were reading orally. That is, prolongations, repetitions, and silent blocks evidenced the same proportional reduction during FAF while participants were reading orally. As hypothesized, stuttering duration decreased during FAF while reading orally. Not only was the proportion of stuttering episodes reduced during FAF while reading orally, the duration of residual stuttering also decreased. Contrary to what was hypothesized, the proportion of types of disfluencies did not differ during FAF. In general, the average proportion of stuttering, average duration of stuttering episodes, and total duration of stuttering were reduced by 50%, 20%, and 50% relative to the NAF conditions, respectively (see Figures 4 6). Individual variability in the ameliorative effect of FAF during oral reading to reduce stuttering (i.e., some PWS show greater reductions in the frequency of stuttering, whereas a small percentage show no effect at all) has been recognized previously (e.g., Armson et al., 1997; Armson & Stuart, 1998; Hargrave et al., 1994; Ingham et al., 1997; Kalinowski et al., 1993; MacLeod et al., 1995; Stuart et al., 1996, 1997, 2004, 2006). This is also evident in the total duration of stuttering and reduction of stuttering duration (see Figures 1 and 2). Although the amount of reduction in stuttering duration and total duration of stuttering was variable, the overwhelming percentage of participants experienced positive reductions during oral reading. These findings are consistent with previously reported findings of stuttering duration reduction under DAF (Martin & Haroldson, 1979). Martin and Haroldson reported a greater reduction, however, under DAF than evidenced with this study (cf. 38% vs. 20%). Herein, the average duration of stuttering events collapsed across NAF conditions and stuttering type was s, which was slightly less than that reported by Martin and Haroldson of s. The average duration of stuttering events collapsed across stuttering type in FAF was s compared with Martin and Haroldson s duration under DAF of s. These differences may be related to differences in AAF conditions, participant characteristics, and/or measurement techniques. In addition, Martin and Haroldson collected spontaneous speech samples, and in this study oral reading samples were examined. The findings of average stuttering durations are consistent with previously reported literature. Collapsed across NAF conditions, the average durations of stuttering events were s, s, and s for prolongations, silent blocks, and repetitions, respectively. 894 Journal of Speech, Language, and Hearing Research Vol August 2008

7 First, the fact that repetitions were longer than prolongations and silent blocks is consistent with that found in children (Conture & Kelly, 1991; Throneburg & Yairi, 2001; Zebrowski, 1991, 1994). Generally, most average durations of prolongations and silent blocks are less than 1 s. Multiple unit part-word and monosyllabic repetitions tend to be longer in duration than 1 s, whereas single part-word and monosyllabic repetitions tend to be shorter in duration than 1 s (Throneburg & Yairi, 1994, 2001). This distinction was not examined in this investigation, as that fine grain of analyses was not undertaken. That is, all part-word, whole-word, and phrase repetitions were lumped together. These findings may have some clinical relevance. Not only can FAF reduce the amount of disfluencies among PWS during oral reading, the duration of the residual stuttering is significantly reduced as well as the total amount of time stuttering. These findings may help explain why self-report measures on the efficacy of AAF devices are so positive (Kalinowski et al., 2004; Stuart et al., 2006) despite the fact that stuttering reduction is not as robust during monologue compared with oral reading. This is speculation, however, as these findings have yet to be confirmed during spontaneous speech. In addition, stuttering severity was not taken into account as a variable in our analyses. It remains to be determined if FAF has a differential effect on the proportion of stuttering events, duration of stuttering events, and total duration of stuttering events as a function of stuttering severity for both oral reading and spontaneous speech. The effect of other forms of AAF on the proportion of stuttering events, duration of stuttering events, and total duration of stuttering events for both oral reading and spontaneous speech needs to be evaluated. Finally, when stuttering occurs, it appears that prolongations are more efficient than repetitions or silent blocks, as their durations are temporally shorter and, consequently, forward following speech returns faster. One may speculate that prolongation strategies may be more effective as a therapeutic tool regardless of feedback type when durational matters are a consideration. We were surprised by the observation that FAF did not have a differential effect on stuttering type. It was our clinical impression that those individuals who have stuttering predominately characterized by silent blocks were less likely to benefit from AAF. That is, when one is locked in a silent block there is no auditory alteration in their speech signal to push them through their stuttering episode. It appears that this is not the case with FAF. It may be that some carryover fluency transfers to their silent blocks from preceding forward flowing speech. Investigations into other forms of AAF are also warranted (e.g., DAF). It is likely that the concomitant reduction in stuttering events and a reduction in duration of residual stuttering episodes and total time stuttering contributes to the perceived naturalness of speech produced by PWS while experiencing FAF (Stuart & Kalinowski, 2004; Stuart et al., 2004, 2006). In summary, there was no differential effect on the reduction of proportion of stuttering type (i.e., prolongations, repetitions, and silent blocks) during FAF. The average stuttering episode duration was significantly reduced by approximately 20% irrespective of stuttering type. Further, a significant decrease in the total duration of stuttering occurred during FAF by approximately 50%. FAF ameliorates stuttering by both reducing the proportion of stuttering events and decreasing the duration of residual stuttering events and total stuttering time. Acknowledgments This work was presented in part at the 2005 American Speech-Language-Hearing Association Annual Convention, San Diego, CA, on November 18, References Armson, J., Foote, S., Witt, C., Kalinowski, J., & Stuart, A. (1997). Effect of frequency altered feedback and audience size on stuttering. European Journal of Disorders of Communication, 32, Armson, J., & Stuart, A. (1998). Effect of extended exposure to frequency-altered feedback on stuttering during reading and monologue. Journal of Speech, Language, and Hearing Research, 41, Bloodstein, O. (1995). A handbook on stuttering (5th ed.). San Diego, CA: Singular Publishing Group. Box, G. E. P., & Jenkins, G. M. (1970). Time series analysis: Forecasting and control. San Francisco: Holden-Day. Cobb, G. W. (1998). Introduction to design and analysis of experiments. New York: Springer-Verlag. Conture, E. G. (2001). Stuttering: Its nature, diagnosis and treatment. Boston: Allyn and Bacon. Conture, E. G., & Kelly, E. M. (1991). Young stutterers nonspeech behaviors during stuttering. Journal of Speech and Hearing Research, 34, Fleiss, J. L., Levin, B., & Paik, M. C. (2003). Statistical methods for rates and proportions (3rd ed.). Hoboken, NJ: John Wiley & Sons. Hargrave, S., Kalinowski, J., Stuart, A., Armson, J., & Jones, K. (1994). Effect of frequency altered feedback on stutterers fluency at normal and fast speech rates. Journal of Speech and Hearing Research, 37, Howell, P., El-Yaniv, N., & Powell, D. J. (1987). Factors affecting fluency in stutterers. In H. F. M. Peters & W. Hulstijin (Eds.), Speech motor dynamics in stuttering (pp ). New York: Springer-Verlag. Howell, P., Sackin, S., & Williams, R. (1999). Differential effects of frequency-shifted feedback between child and adult stutterers. Journal of Fluency Disorders, 24, Stuart et al.: Effect of Frequency Altered Feedback 895

8 Ingham, R. J., Moglia, R. A., Frank, P., Costello-Ingham, J., & Cordes, A. (1997). Experimental investigation of the effects of frequency altered feedback on the speech of adults who stutter. Journal of Speech, Language, and Hearing Research, 40, Kalinowski, J., Armson, J., Roland-Mieszkowski, M., Stuart, A., & Gracco, V. (1993). The effects of alterations in auditory feedback on stuttering frequency. Language and Speech, 36, Kalinowski, J., Guntupalli, V., Stuart, A., & Saltuklaroglu, T. (2004). Self-reported efficacy of an ear-level prosthetic device that delivers altered auditory feedback for the management of stuttering. International Journal of Rehabilitation Research, 27, Kalinowski, J., Stuart, A., Wamsley, L., & Rastatter, M. P. (1999). Effect of monitoring condition and altered auditory feedback on stuttering frequency. Journal of Speech, Language, and Hearing Research, 42, Kent, R. D., Adams, S. G., & Turner, G. S. (1996). Models of speech production. In N. J. Lass (Ed.), Principles of experimental phonetics (pp. 3 45). St. Louis, MO: Mosby. Landis, J. R., & Koch, G. G. (1977). The measurement of observer agreement for categorical data. Biometrics, 33, MacLeod, J., Kalinowski, J., Stuart, A., & Armson, J. (1995). Effect of single and combined altered auditory feedback on stuttering frequency at two speech rates. Journal of Communication Disorders, 28, MacNeilage, P. F. (1998). The frame/content theory of evolution of speech production. Behavioral and Brain Sciences, 21, Maguire, G. A., Riley, G. D., Franklin, D. L., & Gottschalk, L. A. (2000). Risperidone for treatment of stuttering. Journal of Clinical Psychopharmacology, 20, Maguire, G. A., Riley, G. D., Franklin, D. L., Maguire, M. E., Nguyen, C. T., & Brojeni, P. H. (2004). Olanzapine in the treatment of developmental stuttering: A double-blind, placebo-controlled trial. Annals of Clinical Psychiatry, 16, Maguire, G. A., Yu, B. P., Franklin, D. L., & Riley, G. D. (2004). Alleviating stuttering with pharmacological interventions. Expert Opinion on Pharmacotherapy, 5, Martin, R., & Haroldson, S. K. (1979). Effects of five experimental treatments on stuttering. Journal of Speech and Hearing Research, 22, Murray, T. J., Kelly, P., Campbell, L., & Stefanik, K. (1977). Haloperidol in the treatment of stuttering. British Journal of Psychiatry, 130, Natke, U., Grosser, J., & Kalveram, K. T. (2001). Fluency, fundamental frequency, and speech rate under frequencyshifted auditory feedback in stuttering and nonstuttering persons. Journal of Fluency Disorders, 26(3), Perkins, W. H. (1990). What is stuttering? Journal of Speech and Hearing Disorders, 55, Prins, D., Mandelkorn, T., & Cerf, F. A. (1980). Principal and differential effects of haoperidol and placebo treatments upon speech disfluencies in stutterers. Journal of Speech and Hearing Research, 23, Riley, G. D. (1994). Stuttering Severity Instrument for Children and Adults (SSI-3; 3rd ed.). Austin, TX: Pro-Ed. Rosenberger, P. B., Wheelden, J. A., & Kalotkin, M. (1976). The effect of haloperidol on stuttering. American Journal of Psychiatry, 133, Rustin, L., Kuhr, A., Cook, P. J., & James, I. M. (1981). Controlled trial of speech therapy versus oxprenolol for stammering. British Medical Journal (Clinical Research Ed.), 283, Sherman, D. (1952). Clinical and experimental use of the Iowa Scale of Severity of Stuttering. Journal of Speech and Hearing Disorders, 17, Stager, S. V., Calis, K., Grothe, D., Bloch, M., Berensen, N. M., et al. (2005). Treatment with medications affecting dopaminergic and serotonergic mechanisms: Effects on fluency and anxiety in persons who stutter. Journal of Fluency Disorders, 30, Stuart, A., & Kalinowski, J. (2004). The perception of speech naturalness of post-therapeutic and altered auditory feedback speech of adults with mild and severe stuttering. Folia Phoniatrica et Logopaedica, 56, Stuart, A., Kalinowski, J., Armson, J., Stenstrom, R., & Jones, K. (1996). Fluency effect of frequency alterations of plus/minus one-half and one-quarter octave shifts in auditory feedback of people who stutter. Journal of Speech and Hearing Research, 39, Stuart, A., Kalinowski, J., & Rastatter, M. P. (1997). Effects of monaural and binaural altered auditory feedback on stuttering frequency. The Journal of the Acoustical Society of America, 101, Stuart, A., Kalinowski, J., Rastatter, M. P., Saltuklaroglu, T., & Dayalu, V. N. (2004). Investigations of the impact of altered auditory feedback in-the-ear devices on the speech of people who stutter: Initial fitting and four-month follow-up. International Journal of Language & Communication Disorders, 39, Stuart, A., Kalinowski, J., Saltuklaroglu, T., & Guntupalli, V. K. (2006). Investigations of the impact of altered auditory feedback in-the-ear devices on the speech of people who stutter: One-year follow-up. Disability and Rehabilitation, 28, Stuart, A., Xia, S., Jiang, Y., Jiang, T., Kalinowski, J., & Rastatter, M. P. (2003). Self-contained in-the-ear device to deliver altered auditory feedback: Applications for stuttering. Annals of Biomedical Engineering, 31, Throneburg, R. N., & Yairi, E. (1994). Temporal dynamics of repetitions during the early stage of childhood stuttering: An acoustic study. Journal of Speech and Hearing Research, 37, Throneburg, R. N., & Yairi, E. (2001). Durational, proportionate, and absolute frequency characteristics of disfluencies: A longitudinal study regarding persistence and recovery. Journal of Speech, Language, and Hearing Research, 44, Van Riper, C. (1973). The treatment of stuttering. Englewood Cliffs, NJ: Prentice-Hall. Wells, P. G., & Malcolm, M. T. (1971). Controlled trial of the treatment of 36 stutterers. British Journal Of Psychiatry, 119, Journal of Speech, Language, and Hearing Research Vol August 2008

9 Wingate, M. E. (1976). Stuttering: Theory and treatment. New York: Irvington. Zebrowski, P. M. (1991). Duration of the speech disfluencies of beginning stutterers. Journal of Speech and Hearing Research, 34, Zebrowski, P. M. (1994). Duration of sound prolongation and sound/syllable repetition in children who stutter: Preliminary observations. Journal of Speech and Hearing Research, 37, Zimmerman, S., Kalinowski, J., Stuart, A., & Rastatter, M. P. (1997). Effect of altered auditory feedback on people who stutter during scripted telephone conversations. Journal of Speech, Language, and Hearing Research, 40, Received July 9, 2007 Revision received October 26, 2007 Accepted October 29, 2007 DOI: / (2008/065) Contact author: Andrew Stuart, College of Allied Health Sciences, Department of Communication Sciences and Disorders, Health Sciences Building, East Carolina University, Greenville, NC Carmen L. Frazier is now at the University of Pennsylvania School of Medicine, Philadelphia. Stuart et al.: Effect of Frequency Altered Feedback 897