- Email: [email protected]
The role of temporal speech cues in facilitating the fluency of adults who stutter
Accepted Manuscript Title: The role of temporal speech cues in facilitating the fluency of adults who stutter Author: Jin Park Kenneth J. Logan PII: DOI: Reference:
S0094-730X(15)00057-1 http://dx.doi.org/doi:10.1016/j.jfludis.2015.07.001 JFD 5590
To appear in:
Journal of Fluency Disorders
Received date: Revised date: Accepted date:
11-7-2013 1-7-2015 22-7-2015
Please cite this article as: Park, J., and Logan, K. J.,The role of temporal speech cues in facilitating the fluency of adults who stutter, Journal of Fluency Disorders (2015), http://dx.doi.org/10.1016/j.jfludis.2015.07.001 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
The Role of Temporal Cues
1
Ac ce
pt
ed
M
an
us
cr
ip t
Highlights Adults who stutter spoke more fluently in temporally-altered choral conditions than they did when speaking solo. Adults who stutter also spoke slower and exhibited more temporal entrainment with the choral signal than the adults who did not stutter. Adults who stutter seem to make greater use of choral speech signals than adults who do not stutter. Adults who stutter do not require accurate temporal models of phonetic targets in order to enhance fluent speech.
Page 1 of 50
The Role of Temporal Cues
2
The Role of Temporal Speech Cues in Facilitating the Fluency of Adults who Stutter Jin Park Kenneth J. Logan
ip t
University of Florida
Author Note:
cr
Jin Park, Department of Linguistics, Chungnam National University, Daejeon,
us
Korea; Kenneth J. Logan, Department of Speech, Language, and Hearing Sciences, University of Florida, Gainesville, FL, USA.
Ac ce
pt
ed
M
an
Correspondence concerning this manuscript should be addressed to Jin Park, Department of Linguistics, Chungnam National University, Daejeon, Korea, 305-764. Email: [email protected] telephone: 82-42-821-6391
Page 2 of 50
The Role of Temporal Cues
3
Abstract Purpose: Adults who stutter speak more fluently during choral speech contexts than they do during solo speech contexts. The underlying mechanisms for this effect remain unclear, however. In this study, we examined the extent to which the choral speech effect depended
stutter followed choral signals more closely than typical speakers did.
ip t
on presentation of intact temporal speech cues. We also examined whether speakers who
cr
Method: 8 adults who stuttered and 8 adults who did not stutter read 60 sentences aloud
us
during a solo speaking condition and three choral speaking conditions (240 total sentences), two of which featured either temporally altered or indeterminate word duration patterns.
an
Effects of these manipulations on speech fluency, rate, and temporal entrainment with the choral speech signal were assessed.
M
Results: Adults who stutter spoke more fluently in all choral speaking conditions than they did when speaking solo. They also spoke slower and exhibited closer temporal entrainment
ed
with the choral signal during the mid- to late-stages of sentence production than the adults who did not stutter. Both groups entrained more closely with unaltered choral signals than
pt
they did with altered choral signals.
Ac ce
Conclusions: Findings suggest that adults who stutter make greater use of speech-related information in choral signals when talking than adults with typical fluency do. The presence of fluency facilitation during temporally altered choral speech and conversation babble suggests that temporal/gestural cueing alone cannot account for fluency facilitation in speakers who stutter. Other potential fluency enhancing mechanisms are discussed.
Page 3 of 50
The Role of Temporal Cues
4
The Role of Temporal Speech Cues in Facilitating the Fluency of Adults who Stutter 1.0. Introduction 1.1. Stuttering and speech timing Many researchers have suggested that stuttering is a speech timing/motor sequencing
ip t
disorder (e.g., Alm, 2004; Kent, 1984; MacKay & MacDonald, 1984; Packman, Code, & Onslow, 2007; Perkins, Kent, & Curlee, 1991; Smits-Bandstra & De Nil, 2007; Van Riper,
cr
1982; Zimmerman, 1980). Within this relatively broad theoretical framework, the disorder
us
has been examined in various ways: (a) as a disruption of the temporal coordination of the phonatory, respiratory, and articulatory systems that underlie speech production (e.g.,
an
Conture, Colton, & Gleason, 1988; Max & Gracco, 2005; Van Riper, 1982; Zimmerman, 1980), (b) as a breakdown in the temporal alignment of segmental and prosodic
M
representations prior to speech initiation (Perkins et al., 1991), (c) as a temporal dyssynchrony between language and motor planning and its execution (Au-Yeung, Howell,
ed
& Pilgrim, 1998; Howell, 2004; Howell, Au-Yeung, & Sackin, 1999), (d) as an impairment in the capacity to generate temporal programs that underlie the sequential movements associated
pt
with spoken language (Kent, 1984; Max & Yudman, 2003; Smits-Bandstra & De Nil, 2007)
Ac ce
and (e) as an impairment in the ability to execute speech motor plans once they have been generated (e.g., Kleinow & Smith, 2000; Packman, et al., 2007; Packman, Onslow, Richard, & van Doorn, 1996; Smits-Bandstra & De Nil, 2007). Despite the relatively large amount of empirical research and theoretical conjecture on
matters such as these, the exact role of temporal factors in the speech fluency of speakers who stutter remains unclear. Consequently, one purpose of the present study was to investigate the role of temporal cueing as a means of fluency facilitation with adults who stutter. This was accomplished by examining changes in speech fluency during different types of choral speaking conditions.
Page 4 of 50
The Role of Temporal Cues
5
1.2. Choral speech and fluency enhancement It has been well demonstrated that speakers who stutter exhibit marked improvement in fluency when speaking chorally with others (Andrews, Howie, Dozsa, & Guitar, 1982; Barber, 1939; Bloodstein, 1950; Freeman & Armson, 1998; Guntupalli Kalinowski,
ip t
Saltuklaroglu, & Nanjundeswaran, 2005; Howell, & Powell, 1987; Ingham, Bothe, Jang, Yates, Cotton, & Seybold , 2009; Ingham & Packman, 1979; Ingham, Warner, Byrd, &
cr
Cotton, 2006; Johnson & Rosen, 1937; Kiefte & Armson, 2008; Rami, Kalinowski, Rastatter,
us
Holbert, & Allen, 2005). Early attempts at explaining the fluency enhancing effects associated with choral speech occurred within a psychological framework. For example,
an
Barber (1939) proposed that choral reading is a novel condition that distracts speakers who stutter from their fluency difficulties, thus allowing them to talk more smoothly and with less
M
effort. The distraction hypothesis subsequently was questioned by a number of researchers (e.g., Fransella, 1967; Fransella & Beech, 1965; Stuart, 1999; Wingate, 1969) as being overly
ed
vague and hard to verify empirically, and findings from several studies have not supported a strong form of the distraction hypothesis (see, for example, Arends, Povel, & Kolk, 1988;
pt
Fransella, 1967; Fransella & Beech, 1965; Mallard & Webb, 1980; Stuart, 1999). For
Ac ce
instance, it has been shown that stuttering is not significantly reduced in certain dual task conditions that presumably involve distraction. These include discerning a pattern in an arrhythmic beat while reading aloud (Fransella & Beech, 1965), writing numbers while reading aloud (Fransella, 1967), and turning a light on and off while reading aloud (Mallard & Webb, 1980).
A second possible explanation for improved fluency under choral reading is that choral speaking leads speakers who stutter to focus on the action of speaking instead of focusing on aspects of communication. The attention-based hypothesis is essentially the opposite of the distraction hypothesis that was presented by Barber (1939) and others. That is,
Page 5 of 50
The Role of Temporal Cues
6
it may be that instead of taking one’s mind off speaking, as would be the case with the distraction hypothesis, speakers who stutter, instead, focus their attention more often and/or more fully on speaking. Several neuro-imaging studies (e.g., Boberg, Yeudall, Schopflocher, & Bo-Lassen, 1983; De Nil, Kroll, Lafaille, & Houle, 2003; Kroll, De Nil, Kapur, & Houle,
ip t
1997) have shown that speakers who stutter exhibit post-treatment increases in left hemisphere activation in the auditory cortex relative to pre-treatment baseline levels. In such
cr
studies, the increased left hemisphere activity is associated with the use of newly learned
us
methods of controlled fluency such as regulated speech rate, but not with increased activation of brain regions associated with attention (c.f., De Nil, Kroll, Lafaille, & Houle, 2003;
an
Neumann et al., 2003). In some studies with speakers who stutter (Neumann et al. 2003; Neumann et al., 2005), cortical regions associated with attention and error monitoring (e.g.,
M
the anterior cingulate cortex) seem to be “de-activated” during both pre- and post-treatment speech, while in others (e.g., De Nil et al. 2000), areas that were overly active in a pre-
ed
treatment context show decreased activation following treatment. A third possibility is that fluency-enhancing conditions like choral speech help
pt
speakers who stutter formulate key components of spoken messages and, in doing so, help
Ac ce
speakers synchronize the various neural regions that are necessary for fluent speech (see Neumann et al., 2003 for additional discussion). For example, Kent (1984) proposed a motor modeling explanation for the choral speech effect. Essentially, he suggested that speakers who stutter monitor the choral speech signal and, in doing so, are able to generate the temporal patterns that are necessary for fluent speech. Kent argued that the mechanisms underlying choral speech are similar to those observed in metronome-paced speech (i.e., fluency inducement in response to an external rhythmic signal). In this view, speakers who stutter might use another speaker’s voice as an external model from which they extract temporal information about ongoing speech. It was proposed that speakers then use the
Page 6 of 50
The Role of Temporal Cues
7
extracted information to generate utterances that are more fluent than they would be without such information. In a similar vein, Saltuklaroglu, Kalinowski, and Guntupalli (2004) proposed that choral speech signals activate a speaker’s mirror neuron system. In this view, the
ip t
accompanying choral speech signal provides speakers who stutter with cues about ongoing articulatory gestures that are necessary to produce speech smoothly and with relatively little
cr
effort. Following the Revised Motor Theory of Speech Perception (Liberman & Mattingly,
us
1985; Liberman & Whalen, 2000), Saltuklaroglu et al. argued that articulatory gestures are invariant objects of both speech perception and speech production, and that speech
an
production and speech perception mainly involve the encoding or decoding of shared articulatory gestures. According to Saltuklaroglu et al., another speaker’s voice provides “an
M
external matrix of speech gestures that is rich in redundant speech cues” (p. 342). In this view, choral speech models provide cues that contain information about the spatial and temporal
ed
properties of the gestural patterns that constitute spoken utterances. The speaker then matches these external cues to the developing speech production plan, which facilitates smooth
pt
production of the intended speech utterances. Under a strong form of this hypothesis, one
Ac ce
would expect the facilitative effects of choral speech to diminish when the quality of the choral signal is either altered or degraded. 1.3. Manipulating choral speech signals The effects of choral speech signal manipulation on fluency enhancement have been
examined in a range of studies (e.g., Barber, 1939; Guntupalli et al., 2005; Guntupalli, Nanjundeswaran, Kalinowski, & Dayalu, 2011; Hudock, Dayalu, Saltuklaroglu, Stuart, Zhang, & Kalinowski, 2011; Kiefte & Armson, 2008; Rami et al., 2005). Research findings suggest that fluency facilitation occurs even when the choral speech signal is mismatched with what the speaker is saying (see Barber, 1939; Cherry & Sayers, 1956, for early
Page 7 of 50
The Role of Temporal Cues
8
explorations of this issue). Further, fluency enhancement also occurs when speakers who stutter are provided with only visual images of targeted articulatory movements, and fluency is enhanced further when the visual images lag the speaker’s articulation in a manner analogous to delayed auditory feedback (Hudock et al., 2011).
ip t
Rami et al. (2005) manipulated the acoustic characteristics of choral speech signals to investigate which components of the choral signal are most relevant to the attainment of
cr
fluent speech. Using a low pass filtering technique, Rami et al. created five different versions
us
of a choral speech signal. Speakers who stutter then were asked to read several printed passages while listening to the various choral signals. Results showed a marked reduction of
an
stuttering relative to a solo reading condition during the following four conditions: typical choral speech, choral speech that was filtered at 500Hz, choral speech that was filtered at
M
1000Hz, and whispered choral speech. In contrast, the participants showed no significant reduction in stuttering frequency during choral speech that was filtered at 100 Hz. Based on
ed
the findings, Rami et al. proposed that acoustic cues arising from articulatory events in the vocal tract are necessary for enhancing fluency in speakers who stutter during choral reading.
pt
Guntupalli et al. (2005) compressed and expanded choral speech signals to examine
Ac ce
the effects of temporal cuing on stuttering frequency. Participants who stuttered read passages aloud while listening to choral speech signals in three conditions: real time, temporally compressed speech (-20%, -40%, -60%, -80% of original duration), and temporally expanded speech (+40%, +80% of original duration). Guntupalli et al. hypothesized that choral speech aids in the recovery of gestural (i.e., articulatory) information that pertains to a target utterance and that recovery of gestural information would be affected by the rate at which gestural cues are presented to speaker. Results were generally consistent with this hypothesis, as fluency facilitation was greater when the choral signals were expanded, presented in real time, or compressed by no more than 20% than it was when
Page 8 of 50
The Role of Temporal Cues
9
choral signals were compressed by 40%, 60%, or 80%. Taken together, findings from studies such as these suggest that choral speech is generally a reliable and robust means of enhancing fluency among people who stutter; however, its effectiveness depends to some extent upon the informational properties of choral signal itself.
ip t
1.4. Purpose and rationale The present study extended the existing line of research by examining whether
cr
speakers who stutter depend on an accurate or faithful rendering of temporal information
us
within a choral signal in order to benefit from it. In contrast to the study by Guntupalli et al. (2005), where the timing structure of the choral signal was temporally compressed but still
an
proportionately congruent with typical speech, the durational characteristics of words within some of the choral signals used in the present study were altered such that some words had a
M
longer-than-normal duration relative to surrounding words and other words had a shorterthan-normal duration relative to surrounding words. Consequently, temporal information
ed
within the choral signal was, to some extent, dissociated from the signal’s syntactic structure and, consequently, its expected prosodic structure. It was thought that this manipulation
pt
would reduce moment-to-moment predictability about the temporal structure of an utterance
Ac ce
and might lead to less fluency facilitation for speakers who stutter in comparison to conditions in which speakers were exposed to unaltered, temporally-faithful choral signals (see Barber, 1939, and Hudock et al., 2011 for discussion of this issue). The present study also incorporated multi-speaker conversational babble as another type of temporally unpredictable choral signal. The main rationale for using the conversational babble is that it allows for the concurrent presentation of a speech-based signal that lacks coherent speechrelated information, particularly temporal cues that relate to syllable or word durational patterns. A second purpose of the present study was to examine the extent to which the
Page 9 of 50
The Role of Temporal Cues
10
utterances of speakers who stutter are temporally entrained to choral speech signals and whether the extent of that entrainment differs from what is observed in speakers who do not stutter. This issue has not been examined in past studies of fluency facilitation under choral speech in such a way that allows for determination of how closely the speaker is following
ip t
the choral signal. If speakers who stutter are indeed using choral speech signals to extract critical information about speech timing and/or articulatory configurations (Kent, 1984;
cr
Saltuklaroglu et al., 2004), speakers who stutter might be expected to show more precise
us
entrainment to a choral speech signal when compared to speakers who do not stutter. If speakers who stutter are extracting speech production cues from choral signals, the degree of
an
entrainment that they exhibit should be greatest in conditions where the temporal structure of the auditory signals closely matches the structure of the utterances speakers are attempting to
M
say.
The primary research questions that were addressed in the present study were these:
ed
(a) does the accuracy or availability of temporal cues in a choral speech signal affect the extent to which the signal facilitates the fluency of a speaker who stutters?, and (b) do
pt
participants who stutter show more entrainment to choral speech signals than speakers who
Ac ce
do not stutter? It was predicted that adults who stutter would speak less fluently when talking with temporally altered choral signals than they would when speaking with unaltered choral signals, and that adults who stutter would exhibit more temporal entrainment with the choral speech than adults who do not stutter. 2.0. Method
2.1. Participants Eight adults who stuttered (seven males, one female, M = 22 years, SD = 12.27 years) and eight adults who did not stutter (six males, two females, M = 23 years, SD = 13.75 years) participated in this study. Each of the participants was at least 18 years of age, spoke
Page 10 of 50
The Role of Temporal Cues
11
American English with native competence, and reported having had no exposure to altered auditory feedback (e.g., delayed auditory feedback) during the two years prior to data collection for the present study. In addition, none of the participants reported having participated in fluency interventions that involved altered auditory feedback.
ip t
None of the participants had any known or reported concomitant impairments (other than stuttering for the adults who stutter) that might have influenced their speech fluency
cr
performance during the study. All participants attained a standard score of 85 or higher on the
us
Peabody Picture Vocabulary Test-III (PPVT-III, Dunn & Dunn, 1997) and obtained a standard score of 85 or higher on the reading subtests associated with the Wide Range
an
Achievement Test-3 (WRAT-3; Jastak & Jastak, 1984). In addition, all participants passed a hearing screening. In the stuttering group, all of the participants reported having either no
M
history of formal fluency treatment or no recent history of formal fluency treatment (i.e., within the three years preceding data collection) and, at the time of data collection, each of
ed
the participants in the stuttering group reported feeling dissatisfied with their ability to speak fluently. Stuttering severity was assessed for the participants who stuttered using the
pt
Stuttering Severity Instrument-4 (Riley, 2009). Based on this analysis, it was determined that
Ac ce
two participants exhibited mild stuttering, five exhibited moderate stuttering, and one exhibited severe stuttering. 2.2. Materials
2.2.1. Sentence stimuli. The authors created 60 sentences for use in the experiment.
The 60 sentences featured a variety of syntactic and prosodic patterns but all were of similar length (i.e., 10 to 11 words, 17 to 21 syllables). The sentences were used to elicit speech samples from the participants under various auditory conditions (see the Appendix for examples of sentences). Within the set of 60 sentences was a subset of 20 sentences, each of which contained 10 words, 18 to 20 syllables, and similar syntactic structure. The phrase
Page 11 of 50
The Role of Temporal Cues
12
structures within these 20 sentences were similar as well, with each consisting of one or more phrases that contained a function word (e.g., the, and, by) followed by a multi-syllable content word. Thus, the prosodic patterns within the 20 sentences were highly similar as well. Each of the content words within the 20 sentences began with a consonant and contained at
ip t
least two syllables. Multi-syllable words have been shown to have a higher probability of featuring overt symptoms of stuttering than monosyllabic words (Howell et al., 1999). The 20
cr
structurally matched sentences were designed primarily for use with the entrainment analysis,
us
which is described below.
Stimulus materials were created by having an American-English-speaking adult
an
female with normal speech fluency read each of the 60 sentences aloud into a digital voice recorder (Sony, ICD-P320) while seated in a sound attenuated booth. A computer-based
M
software program (WavePad, version 4.26) was used for capturing and editing the woman’s speech, and all of the audio files were stored in WAV format. These digital recordings were
ed
used to create the choral speech stimuli used in the experiment. 2.2.2. Typical and altered choral speech signals. Two types of choral stimuli were
pt
created from the sentence recordings. The first type, typical choral speech (TCS), consisted
Ac ce
of the recorded sentences that the adult female speaker originally had created. The second type, altered choral speech (ACS) was created using the PRAAT software program (Boersma & Weenink, version 4.3.01), which made it possible to alter the temporal properties of words within the original recordings. With respect to alteration of temporal patterns, each of the words contained in each of the experimental sentences was analyzed individually and its original duration was manipulated. Stress patterns in multi-word utterances are determined largely by the speaker’s communicative goal and the respective grammatical roles that words in the utterance assume in relation to that goal (Wingate, 1984). For instance, pitch-accented words (e.g., content words) tend to have a longer duration than unaccented words (e.g.,
Page 12 of 50
The Role of Temporal Cues
13
function words). In connected speech, word duration also has been shown to depend upon where a word appears in an utterance (Ferreira, 1993). For example, a word generally will have a longer duration when it occurs in a phrase- or clause-final position than when it occurs in a non-final position.
ip t
For each of the experimental sentences, manipulation of word duration was accomplished by using the acoustic signal to identify individual words within the sentences,
cr
and then elongating the duration of the function words (which normally are relatively short)
us
and shortening the duration of the content words (which normally are relatively long). Pause durations between words and phrases were not manipulated. Word duration was manipulated
an
such that function words and content words had similar durations. That is, the mean length of function words was 430 ms and the mean length of content words was 410 ms. The
M
percentage of word expansion (or elongation) with respect to the baseline was 6% and the percentage of compression (or shortening) was 7% with respect to the baseline. The overall
ed
duration of each of the temporally altered sentences, however, remained identical to the duration of the original sentences (i.e., M duration = 4.30 s). Consequently, the speaking rate
pt
(i.e., syllables per second) of the sentences also remained the same. The overall pitch contour
Ac ce
(i.e., intonation) of the temporally altered sentences was not intentionally altered. Thus, it remained relatively consistent with that of the original unaltered versions. The temporal manipulations did not appear to significantly affect sentence
intelligibility. Prior to data collection for the study, the first author presented 20 of the temporally-altered versions of the sentences to five American-English-speaking adults who were unfamiliar with the stimuli and asked them to identify each of the sentences, which each of them did with 100% accuracy. During the course of data collection for the study, participants were regularly reminded to report any difficulties in hearing or understanding the choral sentences, but no such difficulties were reported by any participants.
Page 13 of 50
The Role of Temporal Cues
14
The choral stimuli were finalized by pasting a recorded carrier phrase (The sentence you have to read is…) into the sound files associated with the sentences. The same woman who had read the stimulus sentences also recorded the carrier phrase, and it was positioned within the sound file such that a 500 ms pause was created between end of the carrier phrase
ip t
and the start of the sentence. To minimize the potential for order effects, a unique presentation order for each participant was created in the sentence sets. To minimize the
cr
potential for carryover effects within a sentence set, the 20 sentences that were matched for
us
syntax and prosody were interspersed among the 40 other sentences such that none of the 20 matched sentences was presented consecutively to participants.
an
2.2.3 Multi-speaker babble as a choral signal. Multi-speaker babble noise, which was developed by the International Collegium of Rehabilitative Audiology (ICRA; Dreschler,
M
Vershuure, Ludvigsen, & Westermann, 2001), was used as another type of choral signal. The main rationale for using the multi-speaker babble signal was that it allowed for the concurrent
ed
presentation of a speech-based signal that lacked coherent prosodic (i.e., temporal or durational) and segmental cues. In other words, even though a listener can easily recognize
pt
conversational babble as human speech, a listener is not readily able to extract precise
Ac ce
speech-based information such as temporal cues that relate to word durational patterns from it. The babble signal featured the combination of 6 speakers’ voices. The length of the
original babble recording was one hour and from that, 60 different excerpts, each with a length equal to that of the choral sentence recordings, were selected for presentation as another type of choral signal. To provide a way for speakers to match response initiation with the choral signal, each excerpt of multi-speaker babble noise was preceded by the carrier phrase (The sentence you have to read is…), with the 500 ms pause interposed between the end of the carrier phrase and the start of the babble signal. 2.3. Procedure
Page 14 of 50
The Role of Temporal Cues
15
Each of the participants was seen individually for two sessions. During the first session, participants provided background information by completing a questionnaire and an oral interview, both of which included questions related to basic demographic data, vision and hearing status, medical and educational histories, and communication skills. At this time,
ip t
participants also completed the PPVT-3 and WRAT-3 tests and produced a conversational speech sample. Regarding the latter, participants were asked to talk for at least 3 minutes on
cr
various topics (e.g., a recent movie or television program they had seen, the types of hobbies
us
they have, things that they do on a typical week day, a favorite vacation). The conversational sample was used to verify the diagnosis of stuttering and to estimate stuttering severity.
an
After completing the various preliminary tasks, participants were invited to participate in the second session, one week later, during which the experimental task was presented. The
M
experimental task consisted of reading aloud the set of 60 sentences under the four different auditory conditions. Presentation order was randomized for the four conditions across
ed
participants to control for a possible order effect. Participants were seated in a comfortable chair in front of a laptop computer (Dell Inspiron 2200) in a quiet room. They were asked to
pt
read each of the sentences aloud. The sentences were presented individually on the computer
Ac ce
screen in 14-point Arial Black font.
The sentence reading task was divided into four blocks of 60 sentences (i.e., 240 total
sentences), during which a participant heard either his or her own voice (i.e., Solo condition), or his or her own voice together with one of the following: typical choral speech (i.e., TCS condition), temporally-altered choral speech (i.e., ACS condition), or conversational babble (i.e., BAB condition). Participants heard all of the auditory signals through earphones (Logitech Premium Notebook Headset) at a comfortable listening level (i.e., less than 75 dB SPL, after Rami et al., 2005). All speech responses obtained from the participants were collected with a microphone (Shure, SM 48), which was placed approximately 20 cm below
Page 15 of 50
The Role of Temporal Cues
16
the mouth. The microphone output was sent to a digital audio recorder (Sony DTC-ZA5ES). A digital video camera (Canon, NTSC 2R65MC) was used to record the participants as they completed the sentence production tasks. A backup audio recording of participants’ responses was made using a portable digital recorder (Sony, ICD-P320).
ip t
Prior to commencing the experiment, a research assistant described the tasks to the participants, and then presented them with several practice items. After a participant
cr
successfully responded to the practice items, he or she was asked to commence the main part
us
of the experiment. Participants were asked to wait for completion of the introductory carrier phrase (The sentence you have to read is…) before saying the sentences. Participants were
an
told to say the sentences at whatever pace was most comfortable for them, and they were instructed to speak as naturally as possible and to refrain from using any self-devised
M
techniques (e.g., finger tapping) to reduce or inhibit stuttering-related disfluency. Following completion of each block of experimental sentences, participants were asked to remove the
ed
headphones and read one of several reading passages aloud under normal auditory feedback. The latter activity was performed to minimize carryover effects that might occur across
pt
consecutive experimental blocks. The reading passages dealt with an assortment of topics that
Ac ce
were unrelated to the experimental sentences. They contained an average of 255 words and were written at a middle school level. 2.4. Data analysis
2.4.1. Stuttering frequency. After the speech samples were collected, the first author
replayed the audio/video recordings to analyze each of the syllables within each of the sentences for the presence of stuttering-related behavior such as part-word repetition, audible or inaudible prolongation, and nonverbal behavior such as facial grimacing that appeared to be symptomatic of stuttered speech (see Conture & Kelly, 1991). The number of stuttered syllables per sentence was tallied and used to compute the percent of syllables stuttered
Page 16 of 50
The Role of Temporal Cues
17
(%SS) in the task. 2.4.2. Entrainment with choral signal. The temporal alignment (i.e., entrainment) of selected speech sound segments in the choral model version of a sentence and a participant’s production of the sentence were compared by analyzing 10 randomly selected fluent
ip t
responses from the subset of 20 sentences that featured highly similar syntactic and prosodic structure. The analysis was limited to the TCS and ACS conditions (10 sentences per
cr
condition), because these conditions allowed for a comparison of altered versus unaltered
us
temporal cues in the choral signal. Basic features of the analysis are schematized in Figure 1. The entrainment analysis was conducted by comparing the onset times for five specific
an
reference phones (one sentence-initial phone, three sentence-medial phones, one sentencefinal phone) in a participant’s randomly selected fluent sentences to the onset times for the
M
same five phones in the corresponding choral sentences. Phone onset times were computed by subtracting the time code associated with the end of the standard carrier phrase that
ed
preceded each sentence from the time code associated with the onsets of each of the five reference phones in the sentences. This approach allowed for tracking of a participant’s
pt
entrainment over the course of sentence production. The three sentence-medial phones were
Ac ce
located in standard positions across the syntactically similar sentences (e.g., the onset of the second noun in the subject noun phrase) and each of the sentence-medial words began with a stop consonant to facilitate ease of onset phone identification and accuracy of entrainment value measurement during the acoustic analysis. Temporal measurements for the entrainment analysis were accomplished by recording
the choral speech signals and the participants’ corresponding responses on separate channels in a digital audio recorder through the use of a line-splitter. Differences between the time codes for the choral and participants’ speech signals at each reference point were compared between participant groups to determine whether the speakers who stutter differed from
Page 17 of 50
The Role of Temporal Cues
18
speakers with typical fluency in the extent to which their responses were aligned temporally with the choral models. 2.4.3. Analysis of speech rate. Speech rate was computed for participant responses in each of the four auditory conditions using procedures described by Logan, Byrd, Mazzocchi,
ip t
& Gillam (2011). That is, the number of syllables spoken in the target sentences was summed and then divided by the total amount of time needed to produce each of the target sentences.
cr
This yielded the number of syllables spoken per second. The rate computations were based
us
on the same 20 syntactically and prosodically matched sentences described above in section 2.2.1. The speech rate analysis was conducted to obtain additional information about both the
an
nature of fluency enhancement in the speakers who stutter and the overall entrainment patterns of the nonstuttering and stuttering groups.
M
2.5. Statistical analysis
Because of the relatively small sample size, the distributional pattern of the
ed
participants’ data, and resulting concerns about meeting the statistical assumptions of parametric statistical tests (Conover, 1980), nonparametric statistical methods were used for
pt
analyzing the stuttering frequency data. Two different nonparametric statistics were used.
Ac ce
Within-subject differences across the auditory conditions for speakers who stutter were analyzed using the Friedman test, a nonparametric statistic that utilizes rank-based procedures. To further examine whether there were significant differences in stuttering frequency among the auditory conditions, planned follow-up comparisons were performed using the paired-sample Wilcoxon signed rank test. Mixed-model analysis of variance (ANOVA) was performed to see if there were significant between- or within-subject differences for analyses related to speech entrainment. To control for Type I error rate across these comparisons, the overall alpha level (α = .05) was divided by the number of comparisons in each family of tests (Marasculio &
Page 18 of 50
The Role of Temporal Cues
19
McSweeney, 1977). Effect sizes were reported as follows: partial eta2 (for ANOVA), Cohen’s d (for t-tests), and r (for Wilcoxon signed ranks tests). Guidelines presented by Cohen (1988) and Leech, Barrett, and Morgan (2008) were used to interpret the effect sizes. 2.6. Intra-and Inter-Judge Reliability
ip t
The intra- and inter-judge reliability of stuttering frequency measurement was determined by re-measuring responses from two randomly selected participants who stutter
cr
during two experimental conditions (i.e., Solo, TCS). The intra-judge reliability was assessed
us
by having the first author compute the stuttering frequencies for the participants’ data, and the inter-judge reliability assessed by having the second author re-analyze the same
an
participant data that previously were analyzed by the first author. The average mean difference score across the sentences for intra-judge reliability was -0.02 stuttered syllables
M
per 100 syllables (SD = 0.49). The average mean difference score across the sentences for inter-judge reliability was -0.004 stuttered syllables per 100 syllables (SD = 0.55).
ed
The intra- and inter-judge reliability for the entrainment measurements was determined by re-analyzing data from the TCS and ACS conditions of four randomly selected
pt
participants (two from the stuttering group and two from the nonstuttering group). There were
Ac ce
no significant differences in either the intra-judge measurements, t(199) = 1.051, p = .294, or the inter-judge measurements, t(199) = .600, p = .549. In addition, strong positive Pearson r correlations were found for both the intra-judge measurements, r = .978; p < .001, and interjudge measurements, r = .945; p < .001. 3.0. Results
3.1. Stuttering frequency across the auditory conditions Figure 2 shows the mean frequency of stuttering in the four auditory conditions for the participants who stuttered. The Friedman test showed a significant within-subjects effect for auditory condition (2(3) = 17.625; p = .001), meaning that the frequency of stuttered
Page 19 of 50
The Role of Temporal Cues
20
syllables differed significantly across the four auditory conditions. Results of follow-up testing (adjusted α = .05/3 = .017) for planned comparisons revealed that the participants who stuttered produced less stuttering during the TCS condition than during the Solo condition (Z = -2.524; p = .012, r = -.63). There was no significant difference (Z = -2.060; p = .039, r = -
ip t
.52) in the frequency of stuttering frequency between the TCS and ACS conditions, which suggested that alterations in the accuracy of temporal information did not markedly reduce
cr
the effect of choral auditory signals upon fluency enhancement for adults who stutter.
us
Similarly, no significant differences were found in the frequency of stutter-like disfluencies in the TCS versus BAB comparison (Z = -0.647; p = .518, r = -.16), suggesting that
necessary for fluency enhancement.
M
3.2. Speech rate across auditory conditions
an
presentation of segmental and temporal information to speakers was sufficient, but not
Figure 3 shows the mean speech rate values across the four auditory conditions for the
ed
two participant groups. As indicated in the figure, the sample means for the participants who stuttered appeared to be slower than those for the participants who did not stutter. Results of
pt
mixed-model ANOVA indicated a significant main effect for the between-subjects factor,
Ac ce
fluency group, F(1, 14) = 21.253, p < .001, partial eta2 = .606. Thus, the speakers who stuttered had slower speech rates than speakers who do not stutter. There was, however, no significant effect for the within-subjects factor, auditory condition, F(1.631, 22.837) = 1.870, p =.182, partial eta2 = .118. Further, there was no significant fluency group x auditory condition interaction, F(1.631, 22.837) = 3.153, p = .071, partial eta2 = .184. Thus, participants who stutter did not show marked differences in speech rates across the auditory feedback conditions. 3.3. Entrainment with choral signal Results of mixed-model ANOVA (fluency group x auditory condition, i.e., TCS vs.
Page 20 of 50
The Role of Temporal Cues
21
ACS, x entrainment point) indicated significant main effects for the between-subjects factor, fluency group, F(1,14) = 5.047, p =.041, partial eta2 = .265, and the within-subjects factor, entrainment point, F(2.857, 39.991) = 65.092, p< .001, partial eta2 = .823. However, no significant main effect was found for a second within-subjects factor, auditory condition, F(1,
ip t
14) = 1.456, p = .248, partial eta2 = .094. A significant effect was also found for the fluency group x entrainment point interaction, F(2.857, 39.991) = 4.713, p = .013, partial eta2 = .230,
cr
but no significant effect was found for the fluency group x auditory condition interaction, F(1,
us
14) = .191, p = .579, partial eta2 = .013. A significant effect was found for the entrainment point x fluency group x auditory condition interaction, F(1.290, 18.063) = 7.901, p = .008,
an
partial eta2 = .361.
Because of the presence of the three-way interaction among entrainment point,
M
fluency group, and auditory condition, two-way ANOVA was performed to examine patterns of speaker entrainment across the five reference points. The results for this portion of the
ed
analysis are summarized in Table 1 and Figure 4. At reference points one and two, results of the ANOVA indicated no significant main effects for the between-subjects factor, fluency
pt
group. Thus, there was no significant difference between groups in entrainment during the
Ac ce
early portions of sentence production between groups. There were, however, significant differences between fluency groups for each of the later entrainment points. That is, at reference points three, four, and five, a significant main effect was found for the betweensubject factor, fluency group (see Table 1). Thus, there was a significant difference between fluency groups such that speakers who stuttered exhibited closer entrainment with the choral signals in the mid- to late-stages of choral sentence production than the typically fluent speakers did. Furthermore, significant auditory condition effects were found at reference points two, three, four, and five (see Table 1). In this case, the results indicated that, for both participant groups, more temporal entrainment was found during TCS condition than during
Page 21 of 50
The Role of Temporal Cues
22
the ACS condition. As shown in Table 1, no significant interaction was observed between fluency group and auditory condition across the five reference points; thus, both groups appeared to entrain more closely with typical choral signals than they did with atypical choral signals.
ip t
3.3. Relationships between fluency improvement and entrainment To obtain insight into whether those who showed the most fluent improvement in
cr
response to choral speech models were those who entrained most closely to choral speech
us
models, an additional correlation analysis was conducted. The amount of reduction in stuttering frequency was computed for participants who stuttered by subtracting percentage
an
of syllables stuttered (%SS) during the TCS condition from percentage of syllables stuttered (%SS) during the Solo condition and then dividing the result by the percentage of syllables
M
stuttered (%SS) during the Solo condition and multiplying by 100. The association between this number (amount of reduction in stuttering frequency) and the average of the five
ed
entrainment-point values was examined, with results showing a significant, but modest
4.0. Discussion
pt
positive correlation (rho = .355, p < .05) between the two variables.
Ac ce
4.1 Fluency enhancement: Patterns and possible mechanism As expected, the participants who stuttered spoke less fluently under normal auditory
feedback (i.e., solo speech) than they did when talking along with another speaker’s voice. The fluency enhancing effect of choral speech observed in the present study replicates findings reported in previous research (e.g., Andrews et al., 1982; Guntupalli et al., 2005; Howell, & Powell, 1987; Ingham & Packman, 1979; Kiefte & Armson, 2008; Rami et al., 2005). In the present study, the fluency enhancing effect of choral speech also was observed when temporal information in the choral speech signal was altered substantially to create a mismatch with normal speech prosody (i.e., the TCS condition). The latter finding is broadly
Page 22 of 50
The Role of Temporal Cues
23
consistent with results from studies that have incorporated filtering or compression of choral signals (e.g., Guntupalli et al., 2005; Rami et al., 2005). It also is consistent with findings from Barber (1939), wherein fluency enhancement occurred when speakers who stuttered heard choral signals that were mismatched in content with the passage they read aloud. In the
ip t
present study, evidence of fluency enhancement also was present in the conversational babble condition, a finding which is broadly consistent with results from other studies in which
us
white noise (Andrews et al., 1982; Brayton & Conture, 1978).
cr
enhancement in the speech fluency of speakers who stutter has occurred under exposure to
As discussed in section 1.1, stuttering has been modeled by some as a symptom of a
an
speech-timing deficit (Alm, 2004; Hickock, Houde, & Rong, 2011; Kalinowski, Saltuklaroglu, Guntupalli, & Stuart, 2004; Kent, 1984; Packman, et al., 2007; Van Riper,
M
1982), and it has been proposed that the act of monitoring another person’s speech, as in a choral speaking condition, helps speakers who stutter generate the temporal patterns and/or
ed
articulatory gestures that are necessary for fluent speech. Overall, the findings from the present study do not support a strong view of either the temporal modeling or the articulatory
pt
gesture modeling hypotheses, as fluency enhancement was observed during the temporally
Ac ce
altered choral speech condition, the conversational babble condition, and the typical choral speech condition. The enhancement of fluency during the altered choral speech condition – wherein speech timing and articulatory cues were not precisely matched to lexical information – suggests that if the participants who stuttered did use the choral signal as an informational source for speech production in that context, they did so in a limited or general way (e.g., by utilizing information about the duration of prominent syllables). The enhancement of fluency in the conversational babble condition – wherein speech timing and articulatory cues were unavailable – suggests that fluency facilitation can occur via other mechanisms, as well.
Page 23 of 50
The Role of Temporal Cues
24
The preceding interpretation of the results is generally consistent with that of other researchers (e.g., Howell, 2007; Howell, Powell, & Kahn, 1983; Howell, & Sackin, 2002; Max, Guenther, Gracco, Ghosh, & Wallace, 2004; Stager, Jeffries, & Braun, 2003), who have noted that speakers who stutter can attain improved fluency in conditions such as syllable-
ip t
timed speech, metronomic pacing, and exposure to white noise, wherein externally based segmental and suprasegmental cues are absent. On this basis, these researchers have proposed
cr
that fluency enhancing conditions such as choral speech, syllable-timed speech, and white
us
noise facilitate fluency in speakers who stutters via a general, common mechanism that enhances multimodal sensorimotor integration, particularly in neural systems related to
an
auditory and speech motor system processing. In this view, a speaker’s extraction of information from choral speech signals would be one of possibly several ways to enhance
M
sensorimotor functioning. In choral speech, speech-related information is readily available, flooding the auditory system with external speech input that presumably facilitates a
ed
speaker’s ability to match the auditory targets in a speech motor plan with corresponding somatosensory targets and, in doing so, allows the speaker to compensate for a deficit in the
pt
sensorimotor integration circuit (Hickock et al., 2011).
Ac ce
As shown in a number of brain imaging studies (e.g., Bhatnagar & Buckingham, 2010; Brown, Ingham, Ingham, Laird, & Fox, 2005; De Nil, Kroll, Kapur, & Houle, 2000; Fox, Ingham, Ingham, Hirsch, Down, & Martin, 1996; Fox, Ingham, Ingham, Zamarripa, Xiong, & Lancaster, 2000), speakers who stutter tend to exhibit significant underactivation of the auditory cortex during speech production. The auditory association areas (i.e., the anterior middle temporal gyrus, anterior superior temporal sulcus) are activated by human voice (Belin, Zatorre, Lafaille, Ahad, & Pike, 2000) and intelligible speech signals (Scott, Blank, Rosen, & Wise, 2000), including recordings of one’s own speech (Curio, Neuloh, Numminen, Jousmäki, & Haki, 2000; Houde, Nagarajan, Sekihara, & Merznich, 2002). To the extent that
Page 24 of 50
The Role of Temporal Cues
25
enhanced activation of the auditory cortex facilitates fluency, findings such as these may be at heart of Bloodstein’s (1995) observation that “…virtually any change in stutterers’ accustomed way of hearing themselves speak is likely to alleviate their speech difficulty” (p. 352).
ip t
There is some evidence to suggest that non-speech signals such as white noise lead to activation increases in the secondary auditory cortex, but that they do not directly facilitate
cr
activation of the auditory regions that process speech and voice (Paus, Marrett, Worsley, &
us
Evans, 1996; Paus, Perry, Zatorre, Worsely, & Evans, 1996). In light of this, it is interesting to speculate about whether the fluency enhancement during the conversational babble
an
condition might yield similar cortical activation patterns, and whether any such patterns are accompanied by facilitation of the speech-related regions of the auditory cortex. Additional
M
research is necessary to examine this possibility. Another potential means for altering sensorimotor relationships would be to reduce the extent to which a speaker relies upon
ed
typical, feed-forward processing for motor commands by triggering compensatory activation within the somatosensory mapping system. This approach has been examined by researchers
pt
(e.g., Civier, Tasko, & Guenther, 2010; Max et al., 2004) who have demonstrated that
Ac ce
introduction of novel or unanticipated perturbations of the speech articulators during the course of an ongoing utterance is sufficient to trigger activation of the sensory cortex, and in doing so, effectively alter one’s approach to speech production. Overall, findings such as these support the idea that there may be multiple mechanisms for enhancing sensorimotor integration and, with it, speech fluency. 4.2. Entrainment to choral signals and fluency enhancement Although exposure to speech-based choral signals did not seem to be necessary for fluency facilitation to occur in participants who stuttered, the results from the entrainment analyses provided partial support for idea that the participants who stuttered made use of the
Page 25 of 50
The Role of Temporal Cues
26
choral signals when they were available. For instance, during the mid- to late-stages of sentence production, their speech was more closely aligned to the temporal structure of the choral reference signal than was the speech of the participants who did not stutter. There also was a significant, albeit modest, relationship between the amount of stuttering reduction that
ip t
a speaker who stuttered exhibited under choral speech and the extent to which he or she entrained with a choral signal. In contrast, the participants with typical fluency seemed to stay
cr
with the choral signal for only the first few syllables of a sentence in the TCS and ACS
us
conditions, after which they spoke at a rate that exceeded the one in the choral signal. Both participant groups followed the choral signal more closely over the course of an utterance
an
when it offered a faithful rendering of the target utterance (the TCS condition) than they did when it offered a distorted rendering of the target utterance (the ACS condition). Overall, the
M
differences between participant groups in both entrainment patterns and speech rate suggest that speakers who stutter followed the choral signals during speech production, while
ed
speakers with typical fluency did not.
4.3. Limitations and future directions
pt
Some limitations of the present study should be highlighted. First, this study featured
Ac ce
a small sample size. Also, two of the participants who stutter fell into the mild range of stuttering severity, and thus had less room for fluency enhancement in comparison to participants with moderate and severe stuttering. As such, the main findings should be viewed with caution. Additional research with a larger and more diverse range of severity ratings is necessary for future experiments concerning the role of temporal cues in fluency facilitation. Second, the main experiment consisted of a series of sentence reading tasks. Although many sentences (i.e., 60 sentences in each of the auditory conditions) were used and the sentences were fairly lengthy, the speech production processes used in reading are not exactly the same as those used in spontaneous conversational speech. For example, sentence
Page 26 of 50
The Role of Temporal Cues
27
reading would be expected to place fewer formulation demands on speakers than conversation. Several participants exhibited relatively little stuttering during the sentence task. This may have been related to the use of a sentence-reading task for speech sample elicitation. Use of a reading task was deemed necessary for the present study because of the need
ip t
to systematically manipulate temporal information in the choral signals. Such manipulation would not have been possible had spontaneous conversational speech been used. In future
cr
research, however, it might be feasible to use non-reading speaking tasks such as modeled
us
sentences or production of rehearsed sentences, both of which provide some control over sentence form and content. Both Logan (2001) and Tsiamtsiouris and Cairns (2013) provide
an
examples of such tasks.
Thus far, the discussion has focused primarily on the role of external speech signals as
M
a means of facilitating sensorimotor integration and, with it, speech fluency in speakers who stutter. Based on findings from the present study concerning fluency enhancement in the
ed
context of conversational babble, it is interesting to speculate about how speakers who stutter attain fluency enhancement when cues about the temporal and/or auditory structure of a
pt
target utterance are absent. In addition to the mechanisms mentioned earlier in the discussion,
Ac ce
another possibility is that conditions such as conversational babbling prompt speakers to attend (i.e., self-monitor, self-regulate) more often or more fully to speech articulation, which, in turn, facilitates the sensorimotor processing necessary to support highly fluent speech. As noted earlier in this section and in the introduction, the role of focused attention in facilitating fluency during conditions like conversational babbling is unclear. In speakers who stutter, fluency enhancement under auditory masking (a context with similarities to conversational babble) is accompanied by the perception that speech becomes less effortful than it is when speaking with normal auditory feedback -- though still not equivalent to that of a typical speaker (Ingham et al., 2006; Ingham et al., 2009). The relationships among variables such as
Page 27 of 50
The Role of Temporal Cues
28
speaking effort, implementation of newly learned motor-based fluency management skills, exposure to auditory signals like auditory masking and conversational babble, and attentionrelated mechanisms are not well understood, though. Further research in this area may offer additional insights into both the role of temporal cuing in fluency enhancement and the
ip t
mechanisms of fluency enhancement in situations where cues about the temporal and/or
Ac ce
pt
ed
M
an
us
cr
auditory structure of a target utterance are absent.
Page 28 of 50
The Role of Temporal Cues
29
Acknowledgement
Ac ce
pt
ed
M
an
us
cr
ip t
The authors would like to thank Kristen Rodriguez for her assistance with data collection.
Page 29 of 50
The Role of Temporal Cues
30
References Alm, P.A. (2004). Stuttering and the basal ganglia circuits: A critical review of possible relations. Journal of Communication Disorders, 37, 325-369.
ip t
Andrews, G., Howie, P.M., Dozsa, M., & Guitar, B.E. (1982). Stuttering: Speech pattern characteristics under fluency-inducing conditions. Journal of Speech and Hearing
cr
Research, 25, 208-216.
Journal of Fluency Disorders, 13, 141-151.
us
Arends, N., Povel, D.-J., & Kolk, H. (1988). Stuttering as an attentional phenomenon.
an
Au-Yeung, J., Howell, P., & Pilgrim, L. (1998). Phonological words and stuttering on function words. Journal of Speech, Language, and Hearing Research, 41, 1019-
M
1030.
Barber, V. (1939). Studies in the psychology of stuttering: XV. Chorus reading as a
ed
distraction in stuttering. Journal of Speech Disorders, 4, 371-383. Belin P., Zatorre R.J., Lafaille P., Ahad P. & Pike B. (2000) Voice-selective areas inhuman
pt
auditory cortex. Nature, 403, 309-312.
Ac ce
Bhatnagar, S.C., & Buckingham, H. (2010). Neurogenic stuttering: Its reticular modulation. Current Neurology and Neuroscience Reports, 10, 491-498.
Bloodstein, O. (1950). Conditions under which stuttering is reduced or absent: A review of literature. Journal of Speech Hearing Disorders, 15, 142-153.
Bloodstein, O. (1995). A handbook on stuttering. San Diego, CA: Singular Publishing Group. Boberg, E., Yeudall, L., Schopflocher, D. & Bo-Lassen, P. (1983). The effect of an intensive behavioral program on the distribution of EEG alpha power in stutterers
Page 30 of 50
The Role of Temporal Cues
31
during the processing of verbal and visuospatial information. Journal of Fluency Disorders, 8, 245-273. Boersma, P., & Weenink, D. Praat: Doing phonetics by computer. [Computer Program]. Retrieved from http://www.praat.org.
ip t
Brayton, E.R., & Conture, E.G. (1978). Effects of noise and rhythmic stimulation on the speech of stuttering. Journal of Speech and Hearing Research, 21, 285-294.
cr
Brown, S., Ingham, R., Inghram, J., Laird, A., & Fox, P. (2005). Stuttered and fluent speech
us
production: An ALE meta-analysis of functional neuroimaging studies. Human Brain Mapping, 25, 105-117.
an
Cherry, E., & Sayers, B. (1956).Experiments upon the total inhibition of stammering by external control and some clinical results. Journal of Psychosomatic Research, 1,
M
233-246.
Civier, O., Tasko, S., Guenther, F.H. (2010) Overreliance on auditory feedback may lead to
ed
sound/syllable repetitions: Simulations of stuttering and fluency-inducing
35(3), 246-279
pt
conditions with a neural model of speech production. Journal of Fluency Disorders,
Ac ce
Cohen, J. (1988). Statistical power analysis for the behavioral sciences. (2nd ed.).Hillsdale, NJ: Lawrence Erlbaum Associates, Publishers.
Conover, W.J. (1980). Practical nonparametric statistics (2nd ed.). Wiley: New York. Conture, E. G., Colton, R. H., & Gleason, J. R. (1988). Selected temporal aspects of coordination during fluent speech of young stutterers. Journal of Speech and Hearing Research, 31, 640–653. Conture, E.G., & Kelly, E.M. (1991). Young stutterers' nonspeech behavior during stuttering. Journal of Speech and Hearing Research, 34, 1041–1056.
Page 31 of 50
The Role of Temporal Cues
32
Curio, G., Neuloh, G., Numminen, J., Jousmäki, V., & Hari, R. (2000). Speaking modifies voice-evoked activity in the human auditory cortex. Human Brain Mapping, 9, 183–191. De Nil, L.F., Kroll, R.M., Kapur, S., & Houle, S. (2000). A positron emission tomography
ip t
study of silent and oral single word reading in stuttering and nonstuttering adults. Journal of Speech, Language and Hearing Research, 43, 1038-1054.
cr
De Nil, L.F., Kroll, R.M., Lafaille, S.J., & Houle, S. (2003). A positron emission
us
tomography study short- and long-term treatment effects on functional brain activity in adults who stutter. Journal of Fluency Disorders, 28, 357-379.
an
Dreschler, W. A., Vershuure, H., Ludvigsen, C., & Westermann, S. (2001). ICRA noises: Artificial noise signals with speech-like spectral and temporal properties for
M
hearing instrument assessment. Audiology, 40, 148-157.
Dunn, L.M., & Dunn, L.M. (1997). Peabody Picture Vocabulary Test-Third Edition.
ed
Bloomington, MN: Pearson Assessments. Dworzynski, K., & Howell, P. (2004). Predicting stuttering from phonetic complexity in
pt
German. Journal of Fluency Disorders, 29, 149-173.
Ac ce
Ferreira, F. (1993). The creation of prosody during sentence processing. Psychological Review,100, 233-253.
Fox, P.T., Ingham, R.J., Ingham, J.C. Hirsch, T.B., Downs, J.H., Martin, C. (1996). A PET study of the neural systems of stuttering. Nature, 382, 158-161.
Fox, P.T., Ingham, R.J., Ingham, J.C., Zamarripa, F., Xiong, J.H., & Lancaster, J.L. (2000). Brain correlates of stuttering and syllable production: A PET performancecorrelation analysis, Brain, 123, 1985-2004. Fransella, F. (1967). Rhythm as a distractor in the modification of stuttering. Behavior Research Therapy, 5, 253-255.
Page 32 of 50
The Role of Temporal Cues
33
Fransella, F., & Beech, H.R. (1965). An experimental analysis of the effect of rhythm on the speech of stutterers. Behavior Research Therapy, 3, 195-201. Freeman, K., & Armson, J. (1998). Extent and stability of stuttering reduction during choral reading. Journal of Speech-Language Pathology and Audiology, 22, 188-202.
ip t
Guntupalli, V. K., Kalinowski, G., Saltuklaroglu, T., & Nanjundeswaran, C. (2005). The effects of temporal modification of second speech signals on stuttering inhibition at
cr
two speech rates in adults. Neuroscience Letters, 385, 7–12.
us
Guntupalli, V. K., Nanjundeswaran, C., Kalinowski, J., & Dayalu, V. N. (2011). The effect of static and dynamic visual gestures on stuttering inhibition. Neuroscience Letters,
an
492, 39-42.
Hickok, G., Houde, J., & Rong, F. (2011). Sensorimotor Integration in Speech Processing:
M
Computational Basis and Neural Organization. Neuron, 69, 407-422. Houde, J. F., Nagarajan, S. S., Sekihara, K., & Merzenich, M. M. (2002). Modulation of the
14, 1125–1138.
ed
auditory cortex during speech: An MEG study. Journal of Cognitive Neuroscience,
pt
Howell, P. (2004). Assessment of some contemporary theories of stuttering that apply to
Ac ce
spontaneous speech. Contemporary Issues in Communication Science and Disorders, 31, 123-140.
Howell, P. (2007). The effects of gated speech on the fluency of speakers who stutter. Folia Phoniatrica et Logopaedica, 59, 250-255.
Howell, P., Au-Yeung, J., & Sackin, S. (1999). Exchange of stuttering from function words to content words with age. Journal of Speech, Language, and Hearing Research, 42, 345-354. Howell, P., & Powell, D.J. (1987). Delayed auditory feedback with delayed sounds varying in duration. Perception and Psychophysics, 42, 166–172.
Page 33 of 50
The Role of Temporal Cues
34
Howell, P., Powell, D. J., & Khan, I. (1983). Amplitude contour of the delayed signal and interference in delayed auditory feedback tasks. Journal of Experimental Psychology: Human Perception and Performance, 9, 772-784. Howell, P., & Sackin, S. (2002). Timing interference to speech in altered listening
ip t
conditions. Journal of Acoustical Society of America, 111, 2842-2852. Hudock, D., Dayalu, V. N., Saltuklaroglu, T., Stuart, A., Zhang, J., & Kalinowski, J. (2011).
cr
Stuttering inhibition via visual feedback at normal and fast speech rates.
us
International Journal of Language & Communication Disorders, 46, 169-178. Ingham, R. J., Bothe, A. K., Jang, E., Yates, L., Cotton, J., Seybold, I. (2009). Measurement
an
of speech effort during fluency-inducing conditions in adults who do and do not stutter. Journal of Speech, Language, and Hearing Research, 52, 1286-1301.
M
Ingham, R.J., & Packman, A. (1979). A further evaluation of the speech of stutterers during chorus-and nonchorus reading conditions. Journal of Speech and Hearing
ed
Research, 22, 784-793.
Ingham, R. J., Warner, A., Byrd, A., & Cotton, J. (2006). Speech effort measurement and
pt
stuttering: Investigating the chorus reading effect. Journal of Speech, Language,
Ac ce
and Hearing Research, 49, 660-670. Jastak, J., & Jastak, S. (1984). Wide-Range Achievement Test—Revised. Wilmington, DE: Jastak Associates.
Johnson, W., & Rosen, L. (1937). Studies in the psychology of stuttering VII: effect of certain changes in speech pattern upon frequency of stuttering. Journal of Speech Disorders, 2, 105-110. Kalinowski, J., Saltuklaroglu, T., Guntupalli, V., & Stuart, A. (2004). Gestural recovery and the role of forward and reversed syllabic repetitions as stuttering inhibitors in adults. Neuroscience Letters, 363, 144-49.
Page 34 of 50
The Role of Temporal Cues
35
Kent, R. D. (1984). Stuttering as a temporal programming disorder. In R. Curlee & W. Perkins (Eds.), Nature and treatment of stuttering: New directions (pp. 283-301). San Diego, CA: College Hill Press. Kiefte, M., & Armson, J. (2008). Dissecting choral speech: Properties of the accompanist
ip t
critical to stuttering reduction. Journal of Communication Disorders, 41, 33-48. Kleinow, J., & Smith, A. (2000). Influence of length and syntactic complexity on the speech
cr
motor stability of the fluent speech of adults who stutter. Journal of Speech,
us
Language, and Hearing Research, 43, 548-559.
Kroll, R. M., De Nil, L. F., Kapur, S., & Houle, S. (1997). A positron emission tomography
an
investigation of post-treatment brain activation in stutterers. In H. F. M. Peters & W. Hulstijn (Eds.), Proceedings of the third international conference on speech
M
motor production and fluency disorders (pp. 307-320). Amsterdam, The Netherlands: Elsevier Science Publishers.
ed
Leech, N. L., Barrett, K. C., & Morgan, G. A. (2008). SPSS for intermediate statistics: Use and interpretation (3rd ed.). New York: Lawrence Erlbaum Associates.
pt
Liberman, A. M., & Mattingly, I.G. (1985). The motor theory of speech perception revised.
Ac ce
Cognition, 21, 1-36.
Liberman, A. M., & Whalen, D. H. (2000). On the relation of speech to language. Trends in Cognitive Sciences, 4,187-196.
Logan, K. J. (2001). The effect of syntactic complexity upon the speech fluency of adolescents and adults who stutter. Journal of Fluency Disorders, 26, 85-106.
Logan, K. J., Byrd, C. T., Mazzocchi, E. M., & Gillam, R. B. (2010). Speaking rate characteristics of elementary-school-aged children who do and do not stutter. Journal of Communication Disorders, 44, 130-147.
Page 35 of 50
The Role of Temporal Cues
36
MacKay, D. G., & MacDonald, M. C. (1984). Stuttering as a sequencing and timing disorder. In R.F. Curlee & W.H. Perkins (Eds.), Nature and treatment of stuttering: New directions (pp. 261-82). San Diego: College-Hill Press. Mallard, A. R., & Webb, W.G. (1980). The effects of auditory and visual “distractors” on
ip t
the frequency of stuttering. Journal of Communication Disorders, 13, 207-212.
for the social sciences. Monterey, CA: Brooks/Cole.
cr
Marasculio, L. A., & McSweeney, N. (1977). Nonparametric and distribution-free methods
us
Max, L., & Gracco, V. (2005). Coordination of oral and laryngeal movements in the perceptually fluent speech of adults who stutter. Journal of Speech, Language, and
an
Hearing Research, 48, 524-542.
Max, L., Guenther, F., Gracco, V., Ghosh, S., & Wallace, M. (2004). Unstable or
M
insufficiently activated internal models and feedback-biased motor control as sources of dysfluency: A theoretical model of stuttering. Contemporary Issues in
ed
Communication Sciences and Disorders, 31, 105-122. Max L, & Yudman, E. (2003). Accuracy and variability of isochronous rhythmic timing
pt
access motor systems in stuttering versus nonstuttering individuals. Journal of
Ac ce
Speech, Language and Hearing Research, 46, 146–163. Neumann, K., Euler, H. A., von Gudenberg, A., Giraud, A., Lanfermann, H., Gall, V., & Preibisch, C. (2003). The nature and treatment of stuttering as revealed by fMRI: A within- and between-group comparison. Journal of Fluency Disorders, 28, 381-410. Neumann, K., Preibisch, C., Euler, H. A., von Gudenberg, A., Lanfermann, H., Gall, V., & Giraud, A. (2005). Cortical plasticity associated with stuttering therapy. Journal of Fluency Disorders, 30, 23-39. Packman, A., Code, C., & Onslow, M. (2007). On the cause of stuttering: Integrating theory with brain and behavioral research. Journal of Neurolinguistics, 20, 253-362.
Page 36 of 50
The Role of Temporal Cues
37
Packman, A., Onslow, M., Richard, F., & van Doorn, J. (1996). Syllabic stress and variability: A model of stuttering. Clinical Linguistics and Phonetics, 10, 235-263. Paus, T., Marrett, S., Worsley, K., & Evans, A. (1996). Imaging motor-to-sensory discharges in the human brain: An experimental tool for the assessment of
ip t
functional connectivity. NeuroImage, 4, 78–86. Paus, T., Perry, D. W., Zatorre, R. J., Worsley, K. J., & Evans, A. C. (1996). Modulation of
cr
cerebral blood flow in the human auditory cortex during speech: Role of motor-to-
us
sensory discharges. European Journal of Neuroscience, 8, 2236–2246.
Perkins, W.H., Kent, R.D., & Curlee, R.F. (1991). A theory of neuropsycholinguistic
an
function in stuttering. Journal of Speech and Hearing Research, 34, 734-52. Rami, M.K., Kalinowski, J., Rastatter, M.P., Holbert, D., & Allen, M. (2005). Choral
M
reading with filtered speech: effect on stuttering. Perceptual and Motor Skills, 100, 421-431.
ed
Riley, G.D. & Bakker, K. (2009). Stuttering Severity Instrument-4 (SSI-4). Pro-Ed. Saltuklaroglu, T., Kalinowski, J., & Guntupalli, V.K. (2004). Towards a common neural
pt
substrate in the immediate and effective inhibition of stuttering. International
Ac ce
Journal of Neuroscience, 114, 435-450. Scott, S.K., Blank, C.C., Rosen, S., & Wise, J.R. (2000). Identification of a pathway for intelligible speech in the left temporal lobe, Brain, 123(12), 2400-2406.
Smits-Bandstra, S., & De Nil, L. F. (2007). Sequence skill learning in persons who stutter: implications for cortico-striato-thalamo-cortical dysfunction. Journal of Fluency
Disorders, 32, 251-278. Stager, S.V., Jeffries, K.J., & Braun, A.R. (2003). Common features of fluency-evoking conditions studied in stuttering subjects and controls: An H215OPET study. Journal of Fluency Disorders, 28, 319–336.
Page 37 of 50
The Role of Temporal Cues
38
Stuart, A. (1999). The distraction hypothesis and the practice of pseudoscience: A reply to Bloodstein. Journal of Speech, Language, and Hearing Research, 42, 913-914. Tsiamtsiouris, J., & Cairns, H. S. (2013). Effects of sentence-structure complexity on speech initiation time and disfluency. Journal of Fluency Disorders, 38, 30-44.
ip t
Van Riper, C. (1982). The nature of stuttering. (2nd ed.). Englewood Cliffs, NJ: PrenticeHall.
cr
Wingate, M. E. (1969). Sound and pattern in “artificial” fluency. Journal of Speech and
us
Hearing Research, 12, 677-86.
Wingate, M. E. (1984). Stutter events and linguistic stress. Journal of Fluency Disorders, 9,
an
295-300.
Zimmerman, G. (1980). Articulatory dynamics of fluent utterances of stutterers and
Ac ce
pt
ed
M
nonstutterers. Journal of Speech and Hearing Research, 23, 108-121.
Page 38 of 50
The Role of Temporal Cues
39
Appendix Listed below are examples of the stimulus sentences that were used to elicit speech samples. Sentences 11 – 20 are examples of the subset of stimuli with similar syntax and prosody that were used for speech sample elicitation as well as the speech entrainment analysis.
cr
2. Is the price of the parking decal for motorcycles reasonable?
ip t
1. The athlete who ran for student government selected his running mates.
3. The neighborhood residents will be talking about their primary concerns.
us
4. The journalism majors decided to form a new student newspaper.
5. Philadelphia scored eleven straight points in the fourth quarter and won.
an
6. Many graduate students will receive scholarships to pay for tuition.
M
7. Mr. Peters argued that children deserve quality health insurance coverage. 8. Christopher promised me he would review the problem before Saturday.
ed
9. The police officer ordered the crowd to step back immediately. 10. Jonathan asked, "Was Catherine on the soccer team last semester?"
pt
11. The policeman and detective had ticketed the Canadian by mistake. 12. The composer and teacher had directed the musical from September.
Ac ce
13. The historian from Kentucky will publish the biography on Tuesday. 14. The chancellor from Cambridge could disappoint the people by resigning. 15. The passenger from Baltimore will compliment the captain from Boston. 16. The governor of Tennessee will continue the discussion on Saturday. 17. The technician from Toshiba is purchasing the materials from Gateway. 18. The director of Patriot has donated the valuables to Benjamin. 19. The candidate for Congress was discussing his concerns with Democrats. 20. The governor of Colorado will pardon the lawbreaker by Tuesday.
Page 39 of 50
The Role of Temporal Cues
40
Ac ce
pt
ed
M
an
us
cr
ip t
Figure Captions Fig.1. Sample screen shot showing the extent of entrainment between a choral speech model (top panel) and a participant’s response (bottom panel) at five points within the utterance (dashed lines). The target sentence is The director of Patriot has donated the valuables to Benjamin. The participant’s response is closely aligned in time to the choral signal at entrainment points 1 and 2. At entrainment points 3, 4, and 5, however, the participant’s response is slightly ahead of the choral speech model, indicating that the participant’s articulation rate is faster than that of the choral model. The screen shot was taken from Praat [Boersma & Weenink, 2009]. Fig. 2. Mean frequency of stuttering-like disfluencies (SLDs) for participants who stutter across speaking conditions (Solo = normal auditory feedback; TCS = typical choral signal; ACS = temporally-altered choral signal; BAB = conversational babble feedback; Error bars = standard error.) Fig. 3. Speaking rate (mean number of syllables uttered per second) for participants who stutter and participants who do not stutter across the experimental speaking conditions (Solo = normal auditory feedback, no choral speech; TCS = typical choral signal; ACS = temporally-altered choral signal; BAB = conversational babble feedback; Error bars = standard error.) Fig. 4. Mean differences in phone onset time (in seconds) for speakers who stutter and speakers who do not stutter during the typical (TCS) and altered (ACS) choral speech conditions. Phone onset times for each participant were computed by subtracting the amount of elapsed time from the end of the standard carrier phrase that preceded each sentence from the onset of each of five predetermined reference points within the sentences. Participant phone onset times were then subtracted from the phone onset times in associated choral signal to provide information on the extent to which speaker’s articulation was entrained with the choral signal from the TCS and ACS conditions. Bars in the positive range indicate that the speaker’s articulation lags the choral signal. Alternately, bars in the negative range indicate the speaker’s articulation precedes the choral signal. Bar height indicates the extent of difference between the speaker’s output and the choral signal (RP1 = sentence-initial point; RP2, RP3, RP4 = three sentence-medial points, RP5= sentence-final point; Error bars = standard error)
Page 40 of 50
The Role of Temporal Cues
4
5
pt
ed
M
an
us
cr
ip t
3
Ac ce
12
41
Page 41 of 50
42
Ac ce
pt
ed
M
an
us
cr
ip t
The Role of Temporal Cues
Page 42 of 50
43
Ac ce
pt
ed
M
an
us
cr
ip t
The Role of Temporal Cues
Page 43 of 50
44
Ac ce
pt
ed
M
an
us
cr
ip t
The Role of Temporal Cues
Page 44 of 50
The Role of Temporal Cues
45
Biographical Sketch Jin Park, Ph.D. is a visiting professor in Department of Linguistics, Chungnam National University, Daejeon, Korea. In recent years, Dr. Park has taught courses in psycholinguistics and speech disorders at several universities in Korea. His current research examines the associations between stuttering, prosody, and neuromotor performance.
Ac ce
pt
ed
M
an
us
cr
ip t
Kenneth J. Logan, Ph.D., CCC-SLP, is an Associate Professor and Program Director in the Department of Speech, Language, and Hearing Sciences at the University of Florida, where he teaches courses, conducts research, and provides clinical services in the area of stuttering.
Page 45 of 50
The Role of Temporal Cues
46
ip t
CE Questions 1. What was NOT the main purpose of the present study? a. To investigate the role of temporal cueing in the speech fluency of adults who stutter b. To examine the mechanisms that underlie fluency enhancement during choral speaking c. To investigate whether adults who stutter depend on an accurate rendering of temporal information to benefit from choral speech d. To examine whether adults who stutter fluently under choral reading condition than speaking under delayed auditory feedback e. To examine whether adults who stutter follow choral signals more closely than adults who do not stutter do
Ac ce
pt
ed
M
an
us
cr
2. What was the rationale for using multi-speaker babble noise in the present study? a. It allowed for the concurrent presentation of a speech-like signal that contained coherent segmental cues b. It allowed for the concurrent presentation of a speech-like signal that contained coherent temporal cues c. It allowed for the presentation of a speech-like signal that lacked coherent temporal and segmental cues d. It allowed for examination of the “distraction hypothesis” e. It allowed for presentation of the “auditory masking effect” 3. Results from the present study DOES NOT indicate that: a. Adults who stutter spoke more fluently in temporally-altered choral speaking condition than they did when speaking solo. b. Adults who stutter spoke less fluently while listening to multi-speaker conversational babble than they did when speaking solo. c. Adults who stutter spoke more fluently in temporally-unaltered choral speaking condition than they did when speaking solo. d. Adults who stutter spoke slower in both of temporally-altered and unaltered choral speaking conditions than they did when speaking solo. e. Adults who stutter exhibited more temporal entrainment with the choral speech in both of altered and unaltered choral speaking conditions. 4. Results from the present study support the idea that: a. Adults who stutter make greater use of choral speech signals than adults who do not stutter b. Adults who stutter may not benefit from choral speech signals to attain the enhanced fluency c. Adults who stutter require accurate segmental models of lexical targets in order to enhance speech fluency d. Adults who stutter require accurate temporal models of lexical targets in order to enhance speech fluency e. Adults who stutter require accurate gestural models of phonetic targets in order to enhance speech fluency 5. According to information discussed in the discussion, what are possible explanations for the patterns of fluency facilitation observed during choral speech with speakers who stutter? a. Use of both a “temporal modeling strategy” and a “focused attention” . b. Use of a “relaxation” strategy. c. Use of a “rehearsal” strategy d. Use of both a “relaxation” and “rehearsal” strategy
Page 46 of 50
The Role of Temporal Cues
47
Ac ce
pt
ed
M
an
us
cr
ip t
e. Use of “regulated breathing” strategy.
Page 47 of 50
The Role of Temporal Cues
48
Ac ce
pt
ed
M
an
us
cr
ip t
CE Answers Key: 1 – d; 2 – c; 3 – b; 4 – a; 5 – a.
Page 48 of 50
The Role of Temporal Cues
49
Educational Objectives:
Ac ce
pt
ed
M
an
us
cr
ip t
The reader will be able to (a) summarize competing views on stuttering as a speech timing disorder, (b) describe the extent to which adults who stutter depend on an accurate rendering of temporal information in order to benefit from choral speech, and (c) discuss possible explanations for fluency facilitation in the presence of inaccurate or indeterminate temporal cues.
Page 49 of 50
cr
ip t
Table(s)
2 p .21 <.001 .22
F(1, 14) 8.02 62.96 .29
4 p .01 <.001 .60
2
P.E. .36 .82 .02
F(1, 14) 6.91 9.78 .003
5 p .02 .007 .96
P.E.2 .33 .41 0
M
P.E. .11 .86 .11
an
2
Reference Point 3 F(1, 14) p P.E.2 5.49 .01 .36 62.96 <.001 .83 .64 .44 .04
Ac c
ep te
d
1 Source F(1, 14) p P.E.2 F(1, 14) Group .31 .59 .02 1.78 Task 1.94 .19 .12 84.79 G x Ta 2.05 .17 .13 1.69 2 2 a ( P.E. = partial eta , G = Group; T = Task)
us
Table 1. Results of 2-way Analyses of Variance (Fluency Group x Speaking Task) Across the 5 Reference Points)
Page 50 of 50
S0094-730X(15)00057-1 http://dx.doi.org/doi:10.1016/j.jfludis.2015.07.001 JFD 5590
To appear in:
Journal of Fluency Disorders
Received date: Revised date: Accepted date:
11-7-2013 1-7-2015 22-7-2015
Please cite this article as: Park, J., and Logan, K. J.,The role of temporal speech cues in facilitating the fluency of adults who stutter, Journal of Fluency Disorders (2015), http://dx.doi.org/10.1016/j.jfludis.2015.07.001 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
The Role of Temporal Cues
1
Ac ce
pt
ed
M
an
us
cr
ip t
Highlights Adults who stutter spoke more fluently in temporally-altered choral conditions than they did when speaking solo. Adults who stutter also spoke slower and exhibited more temporal entrainment with the choral signal than the adults who did not stutter. Adults who stutter seem to make greater use of choral speech signals than adults who do not stutter. Adults who stutter do not require accurate temporal models of phonetic targets in order to enhance fluent speech.
Page 1 of 50
The Role of Temporal Cues
2
The Role of Temporal Speech Cues in Facilitating the Fluency of Adults who Stutter Jin Park Kenneth J. Logan
ip t
University of Florida
Author Note:
cr
Jin Park, Department of Linguistics, Chungnam National University, Daejeon,
us
Korea; Kenneth J. Logan, Department of Speech, Language, and Hearing Sciences, University of Florida, Gainesville, FL, USA.
Ac ce
pt
ed
M
an
Correspondence concerning this manuscript should be addressed to Jin Park, Department of Linguistics, Chungnam National University, Daejeon, Korea, 305-764. Email: [email protected] telephone: 82-42-821-6391
Page 2 of 50
The Role of Temporal Cues
3
Abstract Purpose: Adults who stutter speak more fluently during choral speech contexts than they do during solo speech contexts. The underlying mechanisms for this effect remain unclear, however. In this study, we examined the extent to which the choral speech effect depended
stutter followed choral signals more closely than typical speakers did.
ip t
on presentation of intact temporal speech cues. We also examined whether speakers who
cr
Method: 8 adults who stuttered and 8 adults who did not stutter read 60 sentences aloud
us
during a solo speaking condition and three choral speaking conditions (240 total sentences), two of which featured either temporally altered or indeterminate word duration patterns.
an
Effects of these manipulations on speech fluency, rate, and temporal entrainment with the choral speech signal were assessed.
M
Results: Adults who stutter spoke more fluently in all choral speaking conditions than they did when speaking solo. They also spoke slower and exhibited closer temporal entrainment
ed
with the choral signal during the mid- to late-stages of sentence production than the adults who did not stutter. Both groups entrained more closely with unaltered choral signals than
pt
they did with altered choral signals.
Ac ce
Conclusions: Findings suggest that adults who stutter make greater use of speech-related information in choral signals when talking than adults with typical fluency do. The presence of fluency facilitation during temporally altered choral speech and conversation babble suggests that temporal/gestural cueing alone cannot account for fluency facilitation in speakers who stutter. Other potential fluency enhancing mechanisms are discussed.
Page 3 of 50
The Role of Temporal Cues
4
The Role of Temporal Speech Cues in Facilitating the Fluency of Adults who Stutter 1.0. Introduction 1.1. Stuttering and speech timing Many researchers have suggested that stuttering is a speech timing/motor sequencing
ip t
disorder (e.g., Alm, 2004; Kent, 1984; MacKay & MacDonald, 1984; Packman, Code, & Onslow, 2007; Perkins, Kent, & Curlee, 1991; Smits-Bandstra & De Nil, 2007; Van Riper,
cr
1982; Zimmerman, 1980). Within this relatively broad theoretical framework, the disorder
us
has been examined in various ways: (a) as a disruption of the temporal coordination of the phonatory, respiratory, and articulatory systems that underlie speech production (e.g.,
an
Conture, Colton, & Gleason, 1988; Max & Gracco, 2005; Van Riper, 1982; Zimmerman, 1980), (b) as a breakdown in the temporal alignment of segmental and prosodic
M
representations prior to speech initiation (Perkins et al., 1991), (c) as a temporal dyssynchrony between language and motor planning and its execution (Au-Yeung, Howell,
ed
& Pilgrim, 1998; Howell, 2004; Howell, Au-Yeung, & Sackin, 1999), (d) as an impairment in the capacity to generate temporal programs that underlie the sequential movements associated
pt
with spoken language (Kent, 1984; Max & Yudman, 2003; Smits-Bandstra & De Nil, 2007)
Ac ce
and (e) as an impairment in the ability to execute speech motor plans once they have been generated (e.g., Kleinow & Smith, 2000; Packman, et al., 2007; Packman, Onslow, Richard, & van Doorn, 1996; Smits-Bandstra & De Nil, 2007). Despite the relatively large amount of empirical research and theoretical conjecture on
matters such as these, the exact role of temporal factors in the speech fluency of speakers who stutter remains unclear. Consequently, one purpose of the present study was to investigate the role of temporal cueing as a means of fluency facilitation with adults who stutter. This was accomplished by examining changes in speech fluency during different types of choral speaking conditions.
Page 4 of 50
The Role of Temporal Cues
5
1.2. Choral speech and fluency enhancement It has been well demonstrated that speakers who stutter exhibit marked improvement in fluency when speaking chorally with others (Andrews, Howie, Dozsa, & Guitar, 1982; Barber, 1939; Bloodstein, 1950; Freeman & Armson, 1998; Guntupalli Kalinowski,
ip t
Saltuklaroglu, & Nanjundeswaran, 2005; Howell, & Powell, 1987; Ingham, Bothe, Jang, Yates, Cotton, & Seybold , 2009; Ingham & Packman, 1979; Ingham, Warner, Byrd, &
cr
Cotton, 2006; Johnson & Rosen, 1937; Kiefte & Armson, 2008; Rami, Kalinowski, Rastatter,
us
Holbert, & Allen, 2005). Early attempts at explaining the fluency enhancing effects associated with choral speech occurred within a psychological framework. For example,
an
Barber (1939) proposed that choral reading is a novel condition that distracts speakers who stutter from their fluency difficulties, thus allowing them to talk more smoothly and with less
M
effort. The distraction hypothesis subsequently was questioned by a number of researchers (e.g., Fransella, 1967; Fransella & Beech, 1965; Stuart, 1999; Wingate, 1969) as being overly
ed
vague and hard to verify empirically, and findings from several studies have not supported a strong form of the distraction hypothesis (see, for example, Arends, Povel, & Kolk, 1988;
pt
Fransella, 1967; Fransella & Beech, 1965; Mallard & Webb, 1980; Stuart, 1999). For
Ac ce
instance, it has been shown that stuttering is not significantly reduced in certain dual task conditions that presumably involve distraction. These include discerning a pattern in an arrhythmic beat while reading aloud (Fransella & Beech, 1965), writing numbers while reading aloud (Fransella, 1967), and turning a light on and off while reading aloud (Mallard & Webb, 1980).
A second possible explanation for improved fluency under choral reading is that choral speaking leads speakers who stutter to focus on the action of speaking instead of focusing on aspects of communication. The attention-based hypothesis is essentially the opposite of the distraction hypothesis that was presented by Barber (1939) and others. That is,
Page 5 of 50
The Role of Temporal Cues
6
it may be that instead of taking one’s mind off speaking, as would be the case with the distraction hypothesis, speakers who stutter, instead, focus their attention more often and/or more fully on speaking. Several neuro-imaging studies (e.g., Boberg, Yeudall, Schopflocher, & Bo-Lassen, 1983; De Nil, Kroll, Lafaille, & Houle, 2003; Kroll, De Nil, Kapur, & Houle,
ip t
1997) have shown that speakers who stutter exhibit post-treatment increases in left hemisphere activation in the auditory cortex relative to pre-treatment baseline levels. In such
cr
studies, the increased left hemisphere activity is associated with the use of newly learned
us
methods of controlled fluency such as regulated speech rate, but not with increased activation of brain regions associated with attention (c.f., De Nil, Kroll, Lafaille, & Houle, 2003;
an
Neumann et al., 2003). In some studies with speakers who stutter (Neumann et al. 2003; Neumann et al., 2005), cortical regions associated with attention and error monitoring (e.g.,
M
the anterior cingulate cortex) seem to be “de-activated” during both pre- and post-treatment speech, while in others (e.g., De Nil et al. 2000), areas that were overly active in a pre-
ed
treatment context show decreased activation following treatment. A third possibility is that fluency-enhancing conditions like choral speech help
pt
speakers who stutter formulate key components of spoken messages and, in doing so, help
Ac ce
speakers synchronize the various neural regions that are necessary for fluent speech (see Neumann et al., 2003 for additional discussion). For example, Kent (1984) proposed a motor modeling explanation for the choral speech effect. Essentially, he suggested that speakers who stutter monitor the choral speech signal and, in doing so, are able to generate the temporal patterns that are necessary for fluent speech. Kent argued that the mechanisms underlying choral speech are similar to those observed in metronome-paced speech (i.e., fluency inducement in response to an external rhythmic signal). In this view, speakers who stutter might use another speaker’s voice as an external model from which they extract temporal information about ongoing speech. It was proposed that speakers then use the
Page 6 of 50
The Role of Temporal Cues
7
extracted information to generate utterances that are more fluent than they would be without such information. In a similar vein, Saltuklaroglu, Kalinowski, and Guntupalli (2004) proposed that choral speech signals activate a speaker’s mirror neuron system. In this view, the
ip t
accompanying choral speech signal provides speakers who stutter with cues about ongoing articulatory gestures that are necessary to produce speech smoothly and with relatively little
cr
effort. Following the Revised Motor Theory of Speech Perception (Liberman & Mattingly,
us
1985; Liberman & Whalen, 2000), Saltuklaroglu et al. argued that articulatory gestures are invariant objects of both speech perception and speech production, and that speech
an
production and speech perception mainly involve the encoding or decoding of shared articulatory gestures. According to Saltuklaroglu et al., another speaker’s voice provides “an
M
external matrix of speech gestures that is rich in redundant speech cues” (p. 342). In this view, choral speech models provide cues that contain information about the spatial and temporal
ed
properties of the gestural patterns that constitute spoken utterances. The speaker then matches these external cues to the developing speech production plan, which facilitates smooth
pt
production of the intended speech utterances. Under a strong form of this hypothesis, one
Ac ce
would expect the facilitative effects of choral speech to diminish when the quality of the choral signal is either altered or degraded. 1.3. Manipulating choral speech signals The effects of choral speech signal manipulation on fluency enhancement have been
examined in a range of studies (e.g., Barber, 1939; Guntupalli et al., 2005; Guntupalli, Nanjundeswaran, Kalinowski, & Dayalu, 2011; Hudock, Dayalu, Saltuklaroglu, Stuart, Zhang, & Kalinowski, 2011; Kiefte & Armson, 2008; Rami et al., 2005). Research findings suggest that fluency facilitation occurs even when the choral speech signal is mismatched with what the speaker is saying (see Barber, 1939; Cherry & Sayers, 1956, for early
Page 7 of 50
The Role of Temporal Cues
8
explorations of this issue). Further, fluency enhancement also occurs when speakers who stutter are provided with only visual images of targeted articulatory movements, and fluency is enhanced further when the visual images lag the speaker’s articulation in a manner analogous to delayed auditory feedback (Hudock et al., 2011).
ip t
Rami et al. (2005) manipulated the acoustic characteristics of choral speech signals to investigate which components of the choral signal are most relevant to the attainment of
cr
fluent speech. Using a low pass filtering technique, Rami et al. created five different versions
us
of a choral speech signal. Speakers who stutter then were asked to read several printed passages while listening to the various choral signals. Results showed a marked reduction of
an
stuttering relative to a solo reading condition during the following four conditions: typical choral speech, choral speech that was filtered at 500Hz, choral speech that was filtered at
M
1000Hz, and whispered choral speech. In contrast, the participants showed no significant reduction in stuttering frequency during choral speech that was filtered at 100 Hz. Based on
ed
the findings, Rami et al. proposed that acoustic cues arising from articulatory events in the vocal tract are necessary for enhancing fluency in speakers who stutter during choral reading.
pt
Guntupalli et al. (2005) compressed and expanded choral speech signals to examine
Ac ce
the effects of temporal cuing on stuttering frequency. Participants who stuttered read passages aloud while listening to choral speech signals in three conditions: real time, temporally compressed speech (-20%, -40%, -60%, -80% of original duration), and temporally expanded speech (+40%, +80% of original duration). Guntupalli et al. hypothesized that choral speech aids in the recovery of gestural (i.e., articulatory) information that pertains to a target utterance and that recovery of gestural information would be affected by the rate at which gestural cues are presented to speaker. Results were generally consistent with this hypothesis, as fluency facilitation was greater when the choral signals were expanded, presented in real time, or compressed by no more than 20% than it was when
Page 8 of 50
The Role of Temporal Cues
9
choral signals were compressed by 40%, 60%, or 80%. Taken together, findings from studies such as these suggest that choral speech is generally a reliable and robust means of enhancing fluency among people who stutter; however, its effectiveness depends to some extent upon the informational properties of choral signal itself.
ip t
1.4. Purpose and rationale The present study extended the existing line of research by examining whether
cr
speakers who stutter depend on an accurate or faithful rendering of temporal information
us
within a choral signal in order to benefit from it. In contrast to the study by Guntupalli et al. (2005), where the timing structure of the choral signal was temporally compressed but still
an
proportionately congruent with typical speech, the durational characteristics of words within some of the choral signals used in the present study were altered such that some words had a
M
longer-than-normal duration relative to surrounding words and other words had a shorterthan-normal duration relative to surrounding words. Consequently, temporal information
ed
within the choral signal was, to some extent, dissociated from the signal’s syntactic structure and, consequently, its expected prosodic structure. It was thought that this manipulation
pt
would reduce moment-to-moment predictability about the temporal structure of an utterance
Ac ce
and might lead to less fluency facilitation for speakers who stutter in comparison to conditions in which speakers were exposed to unaltered, temporally-faithful choral signals (see Barber, 1939, and Hudock et al., 2011 for discussion of this issue). The present study also incorporated multi-speaker conversational babble as another type of temporally unpredictable choral signal. The main rationale for using the conversational babble is that it allows for the concurrent presentation of a speech-based signal that lacks coherent speechrelated information, particularly temporal cues that relate to syllable or word durational patterns. A second purpose of the present study was to examine the extent to which the
Page 9 of 50
The Role of Temporal Cues
10
utterances of speakers who stutter are temporally entrained to choral speech signals and whether the extent of that entrainment differs from what is observed in speakers who do not stutter. This issue has not been examined in past studies of fluency facilitation under choral speech in such a way that allows for determination of how closely the speaker is following
ip t
the choral signal. If speakers who stutter are indeed using choral speech signals to extract critical information about speech timing and/or articulatory configurations (Kent, 1984;
cr
Saltuklaroglu et al., 2004), speakers who stutter might be expected to show more precise
us
entrainment to a choral speech signal when compared to speakers who do not stutter. If speakers who stutter are extracting speech production cues from choral signals, the degree of
an
entrainment that they exhibit should be greatest in conditions where the temporal structure of the auditory signals closely matches the structure of the utterances speakers are attempting to
M
say.
The primary research questions that were addressed in the present study were these:
ed
(a) does the accuracy or availability of temporal cues in a choral speech signal affect the extent to which the signal facilitates the fluency of a speaker who stutters?, and (b) do
pt
participants who stutter show more entrainment to choral speech signals than speakers who
Ac ce
do not stutter? It was predicted that adults who stutter would speak less fluently when talking with temporally altered choral signals than they would when speaking with unaltered choral signals, and that adults who stutter would exhibit more temporal entrainment with the choral speech than adults who do not stutter. 2.0. Method
2.1. Participants Eight adults who stuttered (seven males, one female, M = 22 years, SD = 12.27 years) and eight adults who did not stutter (six males, two females, M = 23 years, SD = 13.75 years) participated in this study. Each of the participants was at least 18 years of age, spoke
Page 10 of 50
The Role of Temporal Cues
11
American English with native competence, and reported having had no exposure to altered auditory feedback (e.g., delayed auditory feedback) during the two years prior to data collection for the present study. In addition, none of the participants reported having participated in fluency interventions that involved altered auditory feedback.
ip t
None of the participants had any known or reported concomitant impairments (other than stuttering for the adults who stutter) that might have influenced their speech fluency
cr
performance during the study. All participants attained a standard score of 85 or higher on the
us
Peabody Picture Vocabulary Test-III (PPVT-III, Dunn & Dunn, 1997) and obtained a standard score of 85 or higher on the reading subtests associated with the Wide Range
an
Achievement Test-3 (WRAT-3; Jastak & Jastak, 1984). In addition, all participants passed a hearing screening. In the stuttering group, all of the participants reported having either no
M
history of formal fluency treatment or no recent history of formal fluency treatment (i.e., within the three years preceding data collection) and, at the time of data collection, each of
ed
the participants in the stuttering group reported feeling dissatisfied with their ability to speak fluently. Stuttering severity was assessed for the participants who stuttered using the
pt
Stuttering Severity Instrument-4 (Riley, 2009). Based on this analysis, it was determined that
Ac ce
two participants exhibited mild stuttering, five exhibited moderate stuttering, and one exhibited severe stuttering. 2.2. Materials
2.2.1. Sentence stimuli. The authors created 60 sentences for use in the experiment.
The 60 sentences featured a variety of syntactic and prosodic patterns but all were of similar length (i.e., 10 to 11 words, 17 to 21 syllables). The sentences were used to elicit speech samples from the participants under various auditory conditions (see the Appendix for examples of sentences). Within the set of 60 sentences was a subset of 20 sentences, each of which contained 10 words, 18 to 20 syllables, and similar syntactic structure. The phrase
Page 11 of 50
The Role of Temporal Cues
12
structures within these 20 sentences were similar as well, with each consisting of one or more phrases that contained a function word (e.g., the, and, by) followed by a multi-syllable content word. Thus, the prosodic patterns within the 20 sentences were highly similar as well. Each of the content words within the 20 sentences began with a consonant and contained at
ip t
least two syllables. Multi-syllable words have been shown to have a higher probability of featuring overt symptoms of stuttering than monosyllabic words (Howell et al., 1999). The 20
cr
structurally matched sentences were designed primarily for use with the entrainment analysis,
us
which is described below.
Stimulus materials were created by having an American-English-speaking adult
an
female with normal speech fluency read each of the 60 sentences aloud into a digital voice recorder (Sony, ICD-P320) while seated in a sound attenuated booth. A computer-based
M
software program (WavePad, version 4.26) was used for capturing and editing the woman’s speech, and all of the audio files were stored in WAV format. These digital recordings were
ed
used to create the choral speech stimuli used in the experiment. 2.2.2. Typical and altered choral speech signals. Two types of choral stimuli were
pt
created from the sentence recordings. The first type, typical choral speech (TCS), consisted
Ac ce
of the recorded sentences that the adult female speaker originally had created. The second type, altered choral speech (ACS) was created using the PRAAT software program (Boersma & Weenink, version 4.3.01), which made it possible to alter the temporal properties of words within the original recordings. With respect to alteration of temporal patterns, each of the words contained in each of the experimental sentences was analyzed individually and its original duration was manipulated. Stress patterns in multi-word utterances are determined largely by the speaker’s communicative goal and the respective grammatical roles that words in the utterance assume in relation to that goal (Wingate, 1984). For instance, pitch-accented words (e.g., content words) tend to have a longer duration than unaccented words (e.g.,
Page 12 of 50
The Role of Temporal Cues
13
function words). In connected speech, word duration also has been shown to depend upon where a word appears in an utterance (Ferreira, 1993). For example, a word generally will have a longer duration when it occurs in a phrase- or clause-final position than when it occurs in a non-final position.
ip t
For each of the experimental sentences, manipulation of word duration was accomplished by using the acoustic signal to identify individual words within the sentences,
cr
and then elongating the duration of the function words (which normally are relatively short)
us
and shortening the duration of the content words (which normally are relatively long). Pause durations between words and phrases were not manipulated. Word duration was manipulated
an
such that function words and content words had similar durations. That is, the mean length of function words was 430 ms and the mean length of content words was 410 ms. The
M
percentage of word expansion (or elongation) with respect to the baseline was 6% and the percentage of compression (or shortening) was 7% with respect to the baseline. The overall
ed
duration of each of the temporally altered sentences, however, remained identical to the duration of the original sentences (i.e., M duration = 4.30 s). Consequently, the speaking rate
pt
(i.e., syllables per second) of the sentences also remained the same. The overall pitch contour
Ac ce
(i.e., intonation) of the temporally altered sentences was not intentionally altered. Thus, it remained relatively consistent with that of the original unaltered versions. The temporal manipulations did not appear to significantly affect sentence
intelligibility. Prior to data collection for the study, the first author presented 20 of the temporally-altered versions of the sentences to five American-English-speaking adults who were unfamiliar with the stimuli and asked them to identify each of the sentences, which each of them did with 100% accuracy. During the course of data collection for the study, participants were regularly reminded to report any difficulties in hearing or understanding the choral sentences, but no such difficulties were reported by any participants.
Page 13 of 50
The Role of Temporal Cues
14
The choral stimuli were finalized by pasting a recorded carrier phrase (The sentence you have to read is…) into the sound files associated with the sentences. The same woman who had read the stimulus sentences also recorded the carrier phrase, and it was positioned within the sound file such that a 500 ms pause was created between end of the carrier phrase
ip t
and the start of the sentence. To minimize the potential for order effects, a unique presentation order for each participant was created in the sentence sets. To minimize the
cr
potential for carryover effects within a sentence set, the 20 sentences that were matched for
us
syntax and prosody were interspersed among the 40 other sentences such that none of the 20 matched sentences was presented consecutively to participants.
an
2.2.3 Multi-speaker babble as a choral signal. Multi-speaker babble noise, which was developed by the International Collegium of Rehabilitative Audiology (ICRA; Dreschler,
M
Vershuure, Ludvigsen, & Westermann, 2001), was used as another type of choral signal. The main rationale for using the multi-speaker babble signal was that it allowed for the concurrent
ed
presentation of a speech-based signal that lacked coherent prosodic (i.e., temporal or durational) and segmental cues. In other words, even though a listener can easily recognize
pt
conversational babble as human speech, a listener is not readily able to extract precise
Ac ce
speech-based information such as temporal cues that relate to word durational patterns from it. The babble signal featured the combination of 6 speakers’ voices. The length of the
original babble recording was one hour and from that, 60 different excerpts, each with a length equal to that of the choral sentence recordings, were selected for presentation as another type of choral signal. To provide a way for speakers to match response initiation with the choral signal, each excerpt of multi-speaker babble noise was preceded by the carrier phrase (The sentence you have to read is…), with the 500 ms pause interposed between the end of the carrier phrase and the start of the babble signal. 2.3. Procedure
Page 14 of 50
The Role of Temporal Cues
15
Each of the participants was seen individually for two sessions. During the first session, participants provided background information by completing a questionnaire and an oral interview, both of which included questions related to basic demographic data, vision and hearing status, medical and educational histories, and communication skills. At this time,
ip t
participants also completed the PPVT-3 and WRAT-3 tests and produced a conversational speech sample. Regarding the latter, participants were asked to talk for at least 3 minutes on
cr
various topics (e.g., a recent movie or television program they had seen, the types of hobbies
us
they have, things that they do on a typical week day, a favorite vacation). The conversational sample was used to verify the diagnosis of stuttering and to estimate stuttering severity.
an
After completing the various preliminary tasks, participants were invited to participate in the second session, one week later, during which the experimental task was presented. The
M
experimental task consisted of reading aloud the set of 60 sentences under the four different auditory conditions. Presentation order was randomized for the four conditions across
ed
participants to control for a possible order effect. Participants were seated in a comfortable chair in front of a laptop computer (Dell Inspiron 2200) in a quiet room. They were asked to
pt
read each of the sentences aloud. The sentences were presented individually on the computer
Ac ce
screen in 14-point Arial Black font.
The sentence reading task was divided into four blocks of 60 sentences (i.e., 240 total
sentences), during which a participant heard either his or her own voice (i.e., Solo condition), or his or her own voice together with one of the following: typical choral speech (i.e., TCS condition), temporally-altered choral speech (i.e., ACS condition), or conversational babble (i.e., BAB condition). Participants heard all of the auditory signals through earphones (Logitech Premium Notebook Headset) at a comfortable listening level (i.e., less than 75 dB SPL, after Rami et al., 2005). All speech responses obtained from the participants were collected with a microphone (Shure, SM 48), which was placed approximately 20 cm below
Page 15 of 50
The Role of Temporal Cues
16
the mouth. The microphone output was sent to a digital audio recorder (Sony DTC-ZA5ES). A digital video camera (Canon, NTSC 2R65MC) was used to record the participants as they completed the sentence production tasks. A backup audio recording of participants’ responses was made using a portable digital recorder (Sony, ICD-P320).
ip t
Prior to commencing the experiment, a research assistant described the tasks to the participants, and then presented them with several practice items. After a participant
cr
successfully responded to the practice items, he or she was asked to commence the main part
us
of the experiment. Participants were asked to wait for completion of the introductory carrier phrase (The sentence you have to read is…) before saying the sentences. Participants were
an
told to say the sentences at whatever pace was most comfortable for them, and they were instructed to speak as naturally as possible and to refrain from using any self-devised
M
techniques (e.g., finger tapping) to reduce or inhibit stuttering-related disfluency. Following completion of each block of experimental sentences, participants were asked to remove the
ed
headphones and read one of several reading passages aloud under normal auditory feedback. The latter activity was performed to minimize carryover effects that might occur across
pt
consecutive experimental blocks. The reading passages dealt with an assortment of topics that
Ac ce
were unrelated to the experimental sentences. They contained an average of 255 words and were written at a middle school level. 2.4. Data analysis
2.4.1. Stuttering frequency. After the speech samples were collected, the first author
replayed the audio/video recordings to analyze each of the syllables within each of the sentences for the presence of stuttering-related behavior such as part-word repetition, audible or inaudible prolongation, and nonverbal behavior such as facial grimacing that appeared to be symptomatic of stuttered speech (see Conture & Kelly, 1991). The number of stuttered syllables per sentence was tallied and used to compute the percent of syllables stuttered
Page 16 of 50
The Role of Temporal Cues
17
(%SS) in the task. 2.4.2. Entrainment with choral signal. The temporal alignment (i.e., entrainment) of selected speech sound segments in the choral model version of a sentence and a participant’s production of the sentence were compared by analyzing 10 randomly selected fluent
ip t
responses from the subset of 20 sentences that featured highly similar syntactic and prosodic structure. The analysis was limited to the TCS and ACS conditions (10 sentences per
cr
condition), because these conditions allowed for a comparison of altered versus unaltered
us
temporal cues in the choral signal. Basic features of the analysis are schematized in Figure 1. The entrainment analysis was conducted by comparing the onset times for five specific
an
reference phones (one sentence-initial phone, three sentence-medial phones, one sentencefinal phone) in a participant’s randomly selected fluent sentences to the onset times for the
M
same five phones in the corresponding choral sentences. Phone onset times were computed by subtracting the time code associated with the end of the standard carrier phrase that
ed
preceded each sentence from the time code associated with the onsets of each of the five reference phones in the sentences. This approach allowed for tracking of a participant’s
pt
entrainment over the course of sentence production. The three sentence-medial phones were
Ac ce
located in standard positions across the syntactically similar sentences (e.g., the onset of the second noun in the subject noun phrase) and each of the sentence-medial words began with a stop consonant to facilitate ease of onset phone identification and accuracy of entrainment value measurement during the acoustic analysis. Temporal measurements for the entrainment analysis were accomplished by recording
the choral speech signals and the participants’ corresponding responses on separate channels in a digital audio recorder through the use of a line-splitter. Differences between the time codes for the choral and participants’ speech signals at each reference point were compared between participant groups to determine whether the speakers who stutter differed from
Page 17 of 50
The Role of Temporal Cues
18
speakers with typical fluency in the extent to which their responses were aligned temporally with the choral models. 2.4.3. Analysis of speech rate. Speech rate was computed for participant responses in each of the four auditory conditions using procedures described by Logan, Byrd, Mazzocchi,
ip t
& Gillam (2011). That is, the number of syllables spoken in the target sentences was summed and then divided by the total amount of time needed to produce each of the target sentences.
cr
This yielded the number of syllables spoken per second. The rate computations were based
us
on the same 20 syntactically and prosodically matched sentences described above in section 2.2.1. The speech rate analysis was conducted to obtain additional information about both the
an
nature of fluency enhancement in the speakers who stutter and the overall entrainment patterns of the nonstuttering and stuttering groups.
M
2.5. Statistical analysis
Because of the relatively small sample size, the distributional pattern of the
ed
participants’ data, and resulting concerns about meeting the statistical assumptions of parametric statistical tests (Conover, 1980), nonparametric statistical methods were used for
pt
analyzing the stuttering frequency data. Two different nonparametric statistics were used.
Ac ce
Within-subject differences across the auditory conditions for speakers who stutter were analyzed using the Friedman test, a nonparametric statistic that utilizes rank-based procedures. To further examine whether there were significant differences in stuttering frequency among the auditory conditions, planned follow-up comparisons were performed using the paired-sample Wilcoxon signed rank test. Mixed-model analysis of variance (ANOVA) was performed to see if there were significant between- or within-subject differences for analyses related to speech entrainment. To control for Type I error rate across these comparisons, the overall alpha level (α = .05) was divided by the number of comparisons in each family of tests (Marasculio &
Page 18 of 50
The Role of Temporal Cues
19
McSweeney, 1977). Effect sizes were reported as follows: partial eta2 (for ANOVA), Cohen’s d (for t-tests), and r (for Wilcoxon signed ranks tests). Guidelines presented by Cohen (1988) and Leech, Barrett, and Morgan (2008) were used to interpret the effect sizes. 2.6. Intra-and Inter-Judge Reliability
ip t
The intra- and inter-judge reliability of stuttering frequency measurement was determined by re-measuring responses from two randomly selected participants who stutter
cr
during two experimental conditions (i.e., Solo, TCS). The intra-judge reliability was assessed
us
by having the first author compute the stuttering frequencies for the participants’ data, and the inter-judge reliability assessed by having the second author re-analyze the same
an
participant data that previously were analyzed by the first author. The average mean difference score across the sentences for intra-judge reliability was -0.02 stuttered syllables
M
per 100 syllables (SD = 0.49). The average mean difference score across the sentences for inter-judge reliability was -0.004 stuttered syllables per 100 syllables (SD = 0.55).
ed
The intra- and inter-judge reliability for the entrainment measurements was determined by re-analyzing data from the TCS and ACS conditions of four randomly selected
pt
participants (two from the stuttering group and two from the nonstuttering group). There were
Ac ce
no significant differences in either the intra-judge measurements, t(199) = 1.051, p = .294, or the inter-judge measurements, t(199) = .600, p = .549. In addition, strong positive Pearson r correlations were found for both the intra-judge measurements, r = .978; p < .001, and interjudge measurements, r = .945; p < .001. 3.0. Results
3.1. Stuttering frequency across the auditory conditions Figure 2 shows the mean frequency of stuttering in the four auditory conditions for the participants who stuttered. The Friedman test showed a significant within-subjects effect for auditory condition (2(3) = 17.625; p = .001), meaning that the frequency of stuttered
Page 19 of 50
The Role of Temporal Cues
20
syllables differed significantly across the four auditory conditions. Results of follow-up testing (adjusted α = .05/3 = .017) for planned comparisons revealed that the participants who stuttered produced less stuttering during the TCS condition than during the Solo condition (Z = -2.524; p = .012, r = -.63). There was no significant difference (Z = -2.060; p = .039, r = -
ip t
.52) in the frequency of stuttering frequency between the TCS and ACS conditions, which suggested that alterations in the accuracy of temporal information did not markedly reduce
cr
the effect of choral auditory signals upon fluency enhancement for adults who stutter.
us
Similarly, no significant differences were found in the frequency of stutter-like disfluencies in the TCS versus BAB comparison (Z = -0.647; p = .518, r = -.16), suggesting that
necessary for fluency enhancement.
M
3.2. Speech rate across auditory conditions
an
presentation of segmental and temporal information to speakers was sufficient, but not
Figure 3 shows the mean speech rate values across the four auditory conditions for the
ed
two participant groups. As indicated in the figure, the sample means for the participants who stuttered appeared to be slower than those for the participants who did not stutter. Results of
pt
mixed-model ANOVA indicated a significant main effect for the between-subjects factor,
Ac ce
fluency group, F(1, 14) = 21.253, p < .001, partial eta2 = .606. Thus, the speakers who stuttered had slower speech rates than speakers who do not stutter. There was, however, no significant effect for the within-subjects factor, auditory condition, F(1.631, 22.837) = 1.870, p =.182, partial eta2 = .118. Further, there was no significant fluency group x auditory condition interaction, F(1.631, 22.837) = 3.153, p = .071, partial eta2 = .184. Thus, participants who stutter did not show marked differences in speech rates across the auditory feedback conditions. 3.3. Entrainment with choral signal Results of mixed-model ANOVA (fluency group x auditory condition, i.e., TCS vs.
Page 20 of 50
The Role of Temporal Cues
21
ACS, x entrainment point) indicated significant main effects for the between-subjects factor, fluency group, F(1,14) = 5.047, p =.041, partial eta2 = .265, and the within-subjects factor, entrainment point, F(2.857, 39.991) = 65.092, p< .001, partial eta2 = .823. However, no significant main effect was found for a second within-subjects factor, auditory condition, F(1,
ip t
14) = 1.456, p = .248, partial eta2 = .094. A significant effect was also found for the fluency group x entrainment point interaction, F(2.857, 39.991) = 4.713, p = .013, partial eta2 = .230,
cr
but no significant effect was found for the fluency group x auditory condition interaction, F(1,
us
14) = .191, p = .579, partial eta2 = .013. A significant effect was found for the entrainment point x fluency group x auditory condition interaction, F(1.290, 18.063) = 7.901, p = .008,
an
partial eta2 = .361.
Because of the presence of the three-way interaction among entrainment point,
M
fluency group, and auditory condition, two-way ANOVA was performed to examine patterns of speaker entrainment across the five reference points. The results for this portion of the
ed
analysis are summarized in Table 1 and Figure 4. At reference points one and two, results of the ANOVA indicated no significant main effects for the between-subjects factor, fluency
pt
group. Thus, there was no significant difference between groups in entrainment during the
Ac ce
early portions of sentence production between groups. There were, however, significant differences between fluency groups for each of the later entrainment points. That is, at reference points three, four, and five, a significant main effect was found for the betweensubject factor, fluency group (see Table 1). Thus, there was a significant difference between fluency groups such that speakers who stuttered exhibited closer entrainment with the choral signals in the mid- to late-stages of choral sentence production than the typically fluent speakers did. Furthermore, significant auditory condition effects were found at reference points two, three, four, and five (see Table 1). In this case, the results indicated that, for both participant groups, more temporal entrainment was found during TCS condition than during
Page 21 of 50
The Role of Temporal Cues
22
the ACS condition. As shown in Table 1, no significant interaction was observed between fluency group and auditory condition across the five reference points; thus, both groups appeared to entrain more closely with typical choral signals than they did with atypical choral signals.
ip t
3.3. Relationships between fluency improvement and entrainment To obtain insight into whether those who showed the most fluent improvement in
cr
response to choral speech models were those who entrained most closely to choral speech
us
models, an additional correlation analysis was conducted. The amount of reduction in stuttering frequency was computed for participants who stuttered by subtracting percentage
an
of syllables stuttered (%SS) during the TCS condition from percentage of syllables stuttered (%SS) during the Solo condition and then dividing the result by the percentage of syllables
M
stuttered (%SS) during the Solo condition and multiplying by 100. The association between this number (amount of reduction in stuttering frequency) and the average of the five
ed
entrainment-point values was examined, with results showing a significant, but modest
4.0. Discussion
pt
positive correlation (rho = .355, p < .05) between the two variables.
Ac ce
4.1 Fluency enhancement: Patterns and possible mechanism As expected, the participants who stuttered spoke less fluently under normal auditory
feedback (i.e., solo speech) than they did when talking along with another speaker’s voice. The fluency enhancing effect of choral speech observed in the present study replicates findings reported in previous research (e.g., Andrews et al., 1982; Guntupalli et al., 2005; Howell, & Powell, 1987; Ingham & Packman, 1979; Kiefte & Armson, 2008; Rami et al., 2005). In the present study, the fluency enhancing effect of choral speech also was observed when temporal information in the choral speech signal was altered substantially to create a mismatch with normal speech prosody (i.e., the TCS condition). The latter finding is broadly
Page 22 of 50
The Role of Temporal Cues
23
consistent with results from studies that have incorporated filtering or compression of choral signals (e.g., Guntupalli et al., 2005; Rami et al., 2005). It also is consistent with findings from Barber (1939), wherein fluency enhancement occurred when speakers who stuttered heard choral signals that were mismatched in content with the passage they read aloud. In the
ip t
present study, evidence of fluency enhancement also was present in the conversational babble condition, a finding which is broadly consistent with results from other studies in which
us
white noise (Andrews et al., 1982; Brayton & Conture, 1978).
cr
enhancement in the speech fluency of speakers who stutter has occurred under exposure to
As discussed in section 1.1, stuttering has been modeled by some as a symptom of a
an
speech-timing deficit (Alm, 2004; Hickock, Houde, & Rong, 2011; Kalinowski, Saltuklaroglu, Guntupalli, & Stuart, 2004; Kent, 1984; Packman, et al., 2007; Van Riper,
M
1982), and it has been proposed that the act of monitoring another person’s speech, as in a choral speaking condition, helps speakers who stutter generate the temporal patterns and/or
ed
articulatory gestures that are necessary for fluent speech. Overall, the findings from the present study do not support a strong view of either the temporal modeling or the articulatory
pt
gesture modeling hypotheses, as fluency enhancement was observed during the temporally
Ac ce
altered choral speech condition, the conversational babble condition, and the typical choral speech condition. The enhancement of fluency during the altered choral speech condition – wherein speech timing and articulatory cues were not precisely matched to lexical information – suggests that if the participants who stuttered did use the choral signal as an informational source for speech production in that context, they did so in a limited or general way (e.g., by utilizing information about the duration of prominent syllables). The enhancement of fluency in the conversational babble condition – wherein speech timing and articulatory cues were unavailable – suggests that fluency facilitation can occur via other mechanisms, as well.
Page 23 of 50
The Role of Temporal Cues
24
The preceding interpretation of the results is generally consistent with that of other researchers (e.g., Howell, 2007; Howell, Powell, & Kahn, 1983; Howell, & Sackin, 2002; Max, Guenther, Gracco, Ghosh, & Wallace, 2004; Stager, Jeffries, & Braun, 2003), who have noted that speakers who stutter can attain improved fluency in conditions such as syllable-
ip t
timed speech, metronomic pacing, and exposure to white noise, wherein externally based segmental and suprasegmental cues are absent. On this basis, these researchers have proposed
cr
that fluency enhancing conditions such as choral speech, syllable-timed speech, and white
us
noise facilitate fluency in speakers who stutters via a general, common mechanism that enhances multimodal sensorimotor integration, particularly in neural systems related to
an
auditory and speech motor system processing. In this view, a speaker’s extraction of information from choral speech signals would be one of possibly several ways to enhance
M
sensorimotor functioning. In choral speech, speech-related information is readily available, flooding the auditory system with external speech input that presumably facilitates a
ed
speaker’s ability to match the auditory targets in a speech motor plan with corresponding somatosensory targets and, in doing so, allows the speaker to compensate for a deficit in the
pt
sensorimotor integration circuit (Hickock et al., 2011).
Ac ce
As shown in a number of brain imaging studies (e.g., Bhatnagar & Buckingham, 2010; Brown, Ingham, Ingham, Laird, & Fox, 2005; De Nil, Kroll, Kapur, & Houle, 2000; Fox, Ingham, Ingham, Hirsch, Down, & Martin, 1996; Fox, Ingham, Ingham, Zamarripa, Xiong, & Lancaster, 2000), speakers who stutter tend to exhibit significant underactivation of the auditory cortex during speech production. The auditory association areas (i.e., the anterior middle temporal gyrus, anterior superior temporal sulcus) are activated by human voice (Belin, Zatorre, Lafaille, Ahad, & Pike, 2000) and intelligible speech signals (Scott, Blank, Rosen, & Wise, 2000), including recordings of one’s own speech (Curio, Neuloh, Numminen, Jousmäki, & Haki, 2000; Houde, Nagarajan, Sekihara, & Merznich, 2002). To the extent that
Page 24 of 50
The Role of Temporal Cues
25
enhanced activation of the auditory cortex facilitates fluency, findings such as these may be at heart of Bloodstein’s (1995) observation that “…virtually any change in stutterers’ accustomed way of hearing themselves speak is likely to alleviate their speech difficulty” (p. 352).
ip t
There is some evidence to suggest that non-speech signals such as white noise lead to activation increases in the secondary auditory cortex, but that they do not directly facilitate
cr
activation of the auditory regions that process speech and voice (Paus, Marrett, Worsley, &
us
Evans, 1996; Paus, Perry, Zatorre, Worsely, & Evans, 1996). In light of this, it is interesting to speculate about whether the fluency enhancement during the conversational babble
an
condition might yield similar cortical activation patterns, and whether any such patterns are accompanied by facilitation of the speech-related regions of the auditory cortex. Additional
M
research is necessary to examine this possibility. Another potential means for altering sensorimotor relationships would be to reduce the extent to which a speaker relies upon
ed
typical, feed-forward processing for motor commands by triggering compensatory activation within the somatosensory mapping system. This approach has been examined by researchers
pt
(e.g., Civier, Tasko, & Guenther, 2010; Max et al., 2004) who have demonstrated that
Ac ce
introduction of novel or unanticipated perturbations of the speech articulators during the course of an ongoing utterance is sufficient to trigger activation of the sensory cortex, and in doing so, effectively alter one’s approach to speech production. Overall, findings such as these support the idea that there may be multiple mechanisms for enhancing sensorimotor integration and, with it, speech fluency. 4.2. Entrainment to choral signals and fluency enhancement Although exposure to speech-based choral signals did not seem to be necessary for fluency facilitation to occur in participants who stuttered, the results from the entrainment analyses provided partial support for idea that the participants who stuttered made use of the
Page 25 of 50
The Role of Temporal Cues
26
choral signals when they were available. For instance, during the mid- to late-stages of sentence production, their speech was more closely aligned to the temporal structure of the choral reference signal than was the speech of the participants who did not stutter. There also was a significant, albeit modest, relationship between the amount of stuttering reduction that
ip t
a speaker who stuttered exhibited under choral speech and the extent to which he or she entrained with a choral signal. In contrast, the participants with typical fluency seemed to stay
cr
with the choral signal for only the first few syllables of a sentence in the TCS and ACS
us
conditions, after which they spoke at a rate that exceeded the one in the choral signal. Both participant groups followed the choral signal more closely over the course of an utterance
an
when it offered a faithful rendering of the target utterance (the TCS condition) than they did when it offered a distorted rendering of the target utterance (the ACS condition). Overall, the
M
differences between participant groups in both entrainment patterns and speech rate suggest that speakers who stutter followed the choral signals during speech production, while
ed
speakers with typical fluency did not.
4.3. Limitations and future directions
pt
Some limitations of the present study should be highlighted. First, this study featured
Ac ce
a small sample size. Also, two of the participants who stutter fell into the mild range of stuttering severity, and thus had less room for fluency enhancement in comparison to participants with moderate and severe stuttering. As such, the main findings should be viewed with caution. Additional research with a larger and more diverse range of severity ratings is necessary for future experiments concerning the role of temporal cues in fluency facilitation. Second, the main experiment consisted of a series of sentence reading tasks. Although many sentences (i.e., 60 sentences in each of the auditory conditions) were used and the sentences were fairly lengthy, the speech production processes used in reading are not exactly the same as those used in spontaneous conversational speech. For example, sentence
Page 26 of 50
The Role of Temporal Cues
27
reading would be expected to place fewer formulation demands on speakers than conversation. Several participants exhibited relatively little stuttering during the sentence task. This may have been related to the use of a sentence-reading task for speech sample elicitation. Use of a reading task was deemed necessary for the present study because of the need
ip t
to systematically manipulate temporal information in the choral signals. Such manipulation would not have been possible had spontaneous conversational speech been used. In future
cr
research, however, it might be feasible to use non-reading speaking tasks such as modeled
us
sentences or production of rehearsed sentences, both of which provide some control over sentence form and content. Both Logan (2001) and Tsiamtsiouris and Cairns (2013) provide
an
examples of such tasks.
Thus far, the discussion has focused primarily on the role of external speech signals as
M
a means of facilitating sensorimotor integration and, with it, speech fluency in speakers who stutter. Based on findings from the present study concerning fluency enhancement in the
ed
context of conversational babble, it is interesting to speculate about how speakers who stutter attain fluency enhancement when cues about the temporal and/or auditory structure of a
pt
target utterance are absent. In addition to the mechanisms mentioned earlier in the discussion,
Ac ce
another possibility is that conditions such as conversational babbling prompt speakers to attend (i.e., self-monitor, self-regulate) more often or more fully to speech articulation, which, in turn, facilitates the sensorimotor processing necessary to support highly fluent speech. As noted earlier in this section and in the introduction, the role of focused attention in facilitating fluency during conditions like conversational babbling is unclear. In speakers who stutter, fluency enhancement under auditory masking (a context with similarities to conversational babble) is accompanied by the perception that speech becomes less effortful than it is when speaking with normal auditory feedback -- though still not equivalent to that of a typical speaker (Ingham et al., 2006; Ingham et al., 2009). The relationships among variables such as
Page 27 of 50
The Role of Temporal Cues
28
speaking effort, implementation of newly learned motor-based fluency management skills, exposure to auditory signals like auditory masking and conversational babble, and attentionrelated mechanisms are not well understood, though. Further research in this area may offer additional insights into both the role of temporal cuing in fluency enhancement and the
ip t
mechanisms of fluency enhancement in situations where cues about the temporal and/or
Ac ce
pt
ed
M
an
us
cr
auditory structure of a target utterance are absent.
Page 28 of 50
The Role of Temporal Cues
29
Acknowledgement
Ac ce
pt
ed
M
an
us
cr
ip t
The authors would like to thank Kristen Rodriguez for her assistance with data collection.
Page 29 of 50
The Role of Temporal Cues
30
References Alm, P.A. (2004). Stuttering and the basal ganglia circuits: A critical review of possible relations. Journal of Communication Disorders, 37, 325-369.
ip t
Andrews, G., Howie, P.M., Dozsa, M., & Guitar, B.E. (1982). Stuttering: Speech pattern characteristics under fluency-inducing conditions. Journal of Speech and Hearing
cr
Research, 25, 208-216.
Journal of Fluency Disorders, 13, 141-151.
us
Arends, N., Povel, D.-J., & Kolk, H. (1988). Stuttering as an attentional phenomenon.
an
Au-Yeung, J., Howell, P., & Pilgrim, L. (1998). Phonological words and stuttering on function words. Journal of Speech, Language, and Hearing Research, 41, 1019-
M
1030.
Barber, V. (1939). Studies in the psychology of stuttering: XV. Chorus reading as a
ed
distraction in stuttering. Journal of Speech Disorders, 4, 371-383. Belin P., Zatorre R.J., Lafaille P., Ahad P. & Pike B. (2000) Voice-selective areas inhuman
pt
auditory cortex. Nature, 403, 309-312.
Ac ce
Bhatnagar, S.C., & Buckingham, H. (2010). Neurogenic stuttering: Its reticular modulation. Current Neurology and Neuroscience Reports, 10, 491-498.
Bloodstein, O. (1950). Conditions under which stuttering is reduced or absent: A review of literature. Journal of Speech Hearing Disorders, 15, 142-153.
Bloodstein, O. (1995). A handbook on stuttering. San Diego, CA: Singular Publishing Group. Boberg, E., Yeudall, L., Schopflocher, D. & Bo-Lassen, P. (1983). The effect of an intensive behavioral program on the distribution of EEG alpha power in stutterers
Page 30 of 50
The Role of Temporal Cues
31
during the processing of verbal and visuospatial information. Journal of Fluency Disorders, 8, 245-273. Boersma, P., & Weenink, D. Praat: Doing phonetics by computer. [Computer Program]. Retrieved from http://www.praat.org.
ip t
Brayton, E.R., & Conture, E.G. (1978). Effects of noise and rhythmic stimulation on the speech of stuttering. Journal of Speech and Hearing Research, 21, 285-294.
cr
Brown, S., Ingham, R., Inghram, J., Laird, A., & Fox, P. (2005). Stuttered and fluent speech
us
production: An ALE meta-analysis of functional neuroimaging studies. Human Brain Mapping, 25, 105-117.
an
Cherry, E., & Sayers, B. (1956).Experiments upon the total inhibition of stammering by external control and some clinical results. Journal of Psychosomatic Research, 1,
M
233-246.
Civier, O., Tasko, S., Guenther, F.H. (2010) Overreliance on auditory feedback may lead to
ed
sound/syllable repetitions: Simulations of stuttering and fluency-inducing
35(3), 246-279
pt
conditions with a neural model of speech production. Journal of Fluency Disorders,
Ac ce
Cohen, J. (1988). Statistical power analysis for the behavioral sciences. (2nd ed.).Hillsdale, NJ: Lawrence Erlbaum Associates, Publishers.
Conover, W.J. (1980). Practical nonparametric statistics (2nd ed.). Wiley: New York. Conture, E. G., Colton, R. H., & Gleason, J. R. (1988). Selected temporal aspects of coordination during fluent speech of young stutterers. Journal of Speech and Hearing Research, 31, 640–653. Conture, E.G., & Kelly, E.M. (1991). Young stutterers' nonspeech behavior during stuttering. Journal of Speech and Hearing Research, 34, 1041–1056.
Page 31 of 50
The Role of Temporal Cues
32
Curio, G., Neuloh, G., Numminen, J., Jousmäki, V., & Hari, R. (2000). Speaking modifies voice-evoked activity in the human auditory cortex. Human Brain Mapping, 9, 183–191. De Nil, L.F., Kroll, R.M., Kapur, S., & Houle, S. (2000). A positron emission tomography
ip t
study of silent and oral single word reading in stuttering and nonstuttering adults. Journal of Speech, Language and Hearing Research, 43, 1038-1054.
cr
De Nil, L.F., Kroll, R.M., Lafaille, S.J., & Houle, S. (2003). A positron emission
us
tomography study short- and long-term treatment effects on functional brain activity in adults who stutter. Journal of Fluency Disorders, 28, 357-379.
an
Dreschler, W. A., Vershuure, H., Ludvigsen, C., & Westermann, S. (2001). ICRA noises: Artificial noise signals with speech-like spectral and temporal properties for
M
hearing instrument assessment. Audiology, 40, 148-157.
Dunn, L.M., & Dunn, L.M. (1997). Peabody Picture Vocabulary Test-Third Edition.
ed
Bloomington, MN: Pearson Assessments. Dworzynski, K., & Howell, P. (2004). Predicting stuttering from phonetic complexity in
pt
German. Journal of Fluency Disorders, 29, 149-173.
Ac ce
Ferreira, F. (1993). The creation of prosody during sentence processing. Psychological Review,100, 233-253.
Fox, P.T., Ingham, R.J., Ingham, J.C. Hirsch, T.B., Downs, J.H., Martin, C. (1996). A PET study of the neural systems of stuttering. Nature, 382, 158-161.
Fox, P.T., Ingham, R.J., Ingham, J.C., Zamarripa, F., Xiong, J.H., & Lancaster, J.L. (2000). Brain correlates of stuttering and syllable production: A PET performancecorrelation analysis, Brain, 123, 1985-2004. Fransella, F. (1967). Rhythm as a distractor in the modification of stuttering. Behavior Research Therapy, 5, 253-255.
Page 32 of 50
The Role of Temporal Cues
33
Fransella, F., & Beech, H.R. (1965). An experimental analysis of the effect of rhythm on the speech of stutterers. Behavior Research Therapy, 3, 195-201. Freeman, K., & Armson, J. (1998). Extent and stability of stuttering reduction during choral reading. Journal of Speech-Language Pathology and Audiology, 22, 188-202.
ip t
Guntupalli, V. K., Kalinowski, G., Saltuklaroglu, T., & Nanjundeswaran, C. (2005). The effects of temporal modification of second speech signals on stuttering inhibition at
cr
two speech rates in adults. Neuroscience Letters, 385, 7–12.
us
Guntupalli, V. K., Nanjundeswaran, C., Kalinowski, J., & Dayalu, V. N. (2011). The effect of static and dynamic visual gestures on stuttering inhibition. Neuroscience Letters,
an
492, 39-42.
Hickok, G., Houde, J., & Rong, F. (2011). Sensorimotor Integration in Speech Processing:
M
Computational Basis and Neural Organization. Neuron, 69, 407-422. Houde, J. F., Nagarajan, S. S., Sekihara, K., & Merzenich, M. M. (2002). Modulation of the
14, 1125–1138.
ed
auditory cortex during speech: An MEG study. Journal of Cognitive Neuroscience,
pt
Howell, P. (2004). Assessment of some contemporary theories of stuttering that apply to
Ac ce
spontaneous speech. Contemporary Issues in Communication Science and Disorders, 31, 123-140.
Howell, P. (2007). The effects of gated speech on the fluency of speakers who stutter. Folia Phoniatrica et Logopaedica, 59, 250-255.
Howell, P., Au-Yeung, J., & Sackin, S. (1999). Exchange of stuttering from function words to content words with age. Journal of Speech, Language, and Hearing Research, 42, 345-354. Howell, P., & Powell, D.J. (1987). Delayed auditory feedback with delayed sounds varying in duration. Perception and Psychophysics, 42, 166–172.
Page 33 of 50
The Role of Temporal Cues
34
Howell, P., Powell, D. J., & Khan, I. (1983). Amplitude contour of the delayed signal and interference in delayed auditory feedback tasks. Journal of Experimental Psychology: Human Perception and Performance, 9, 772-784. Howell, P., & Sackin, S. (2002). Timing interference to speech in altered listening
ip t
conditions. Journal of Acoustical Society of America, 111, 2842-2852. Hudock, D., Dayalu, V. N., Saltuklaroglu, T., Stuart, A., Zhang, J., & Kalinowski, J. (2011).
cr
Stuttering inhibition via visual feedback at normal and fast speech rates.
us
International Journal of Language & Communication Disorders, 46, 169-178. Ingham, R. J., Bothe, A. K., Jang, E., Yates, L., Cotton, J., Seybold, I. (2009). Measurement
an
of speech effort during fluency-inducing conditions in adults who do and do not stutter. Journal of Speech, Language, and Hearing Research, 52, 1286-1301.
M
Ingham, R.J., & Packman, A. (1979). A further evaluation of the speech of stutterers during chorus-and nonchorus reading conditions. Journal of Speech and Hearing
ed
Research, 22, 784-793.
Ingham, R. J., Warner, A., Byrd, A., & Cotton, J. (2006). Speech effort measurement and
pt
stuttering: Investigating the chorus reading effect. Journal of Speech, Language,
Ac ce
and Hearing Research, 49, 660-670. Jastak, J., & Jastak, S. (1984). Wide-Range Achievement Test—Revised. Wilmington, DE: Jastak Associates.
Johnson, W., & Rosen, L. (1937). Studies in the psychology of stuttering VII: effect of certain changes in speech pattern upon frequency of stuttering. Journal of Speech Disorders, 2, 105-110. Kalinowski, J., Saltuklaroglu, T., Guntupalli, V., & Stuart, A. (2004). Gestural recovery and the role of forward and reversed syllabic repetitions as stuttering inhibitors in adults. Neuroscience Letters, 363, 144-49.
Page 34 of 50
The Role of Temporal Cues
35
Kent, R. D. (1984). Stuttering as a temporal programming disorder. In R. Curlee & W. Perkins (Eds.), Nature and treatment of stuttering: New directions (pp. 283-301). San Diego, CA: College Hill Press. Kiefte, M., & Armson, J. (2008). Dissecting choral speech: Properties of the accompanist
ip t
critical to stuttering reduction. Journal of Communication Disorders, 41, 33-48. Kleinow, J., & Smith, A. (2000). Influence of length and syntactic complexity on the speech
cr
motor stability of the fluent speech of adults who stutter. Journal of Speech,
us
Language, and Hearing Research, 43, 548-559.
Kroll, R. M., De Nil, L. F., Kapur, S., & Houle, S. (1997). A positron emission tomography
an
investigation of post-treatment brain activation in stutterers. In H. F. M. Peters & W. Hulstijn (Eds.), Proceedings of the third international conference on speech
M
motor production and fluency disorders (pp. 307-320). Amsterdam, The Netherlands: Elsevier Science Publishers.
ed
Leech, N. L., Barrett, K. C., & Morgan, G. A. (2008). SPSS for intermediate statistics: Use and interpretation (3rd ed.). New York: Lawrence Erlbaum Associates.
pt
Liberman, A. M., & Mattingly, I.G. (1985). The motor theory of speech perception revised.
Ac ce
Cognition, 21, 1-36.
Liberman, A. M., & Whalen, D. H. (2000). On the relation of speech to language. Trends in Cognitive Sciences, 4,187-196.
Logan, K. J. (2001). The effect of syntactic complexity upon the speech fluency of adolescents and adults who stutter. Journal of Fluency Disorders, 26, 85-106.
Logan, K. J., Byrd, C. T., Mazzocchi, E. M., & Gillam, R. B. (2010). Speaking rate characteristics of elementary-school-aged children who do and do not stutter. Journal of Communication Disorders, 44, 130-147.
Page 35 of 50
The Role of Temporal Cues
36
MacKay, D. G., & MacDonald, M. C. (1984). Stuttering as a sequencing and timing disorder. In R.F. Curlee & W.H. Perkins (Eds.), Nature and treatment of stuttering: New directions (pp. 261-82). San Diego: College-Hill Press. Mallard, A. R., & Webb, W.G. (1980). The effects of auditory and visual “distractors” on
ip t
the frequency of stuttering. Journal of Communication Disorders, 13, 207-212.
for the social sciences. Monterey, CA: Brooks/Cole.
cr
Marasculio, L. A., & McSweeney, N. (1977). Nonparametric and distribution-free methods
us
Max, L., & Gracco, V. (2005). Coordination of oral and laryngeal movements in the perceptually fluent speech of adults who stutter. Journal of Speech, Language, and
an
Hearing Research, 48, 524-542.
Max, L., Guenther, F., Gracco, V., Ghosh, S., & Wallace, M. (2004). Unstable or
M
insufficiently activated internal models and feedback-biased motor control as sources of dysfluency: A theoretical model of stuttering. Contemporary Issues in
ed
Communication Sciences and Disorders, 31, 105-122. Max L, & Yudman, E. (2003). Accuracy and variability of isochronous rhythmic timing
pt
access motor systems in stuttering versus nonstuttering individuals. Journal of
Ac ce
Speech, Language and Hearing Research, 46, 146–163. Neumann, K., Euler, H. A., von Gudenberg, A., Giraud, A., Lanfermann, H., Gall, V., & Preibisch, C. (2003). The nature and treatment of stuttering as revealed by fMRI: A within- and between-group comparison. Journal of Fluency Disorders, 28, 381-410. Neumann, K., Preibisch, C., Euler, H. A., von Gudenberg, A., Lanfermann, H., Gall, V., & Giraud, A. (2005). Cortical plasticity associated with stuttering therapy. Journal of Fluency Disorders, 30, 23-39. Packman, A., Code, C., & Onslow, M. (2007). On the cause of stuttering: Integrating theory with brain and behavioral research. Journal of Neurolinguistics, 20, 253-362.
Page 36 of 50
The Role of Temporal Cues
37
Packman, A., Onslow, M., Richard, F., & van Doorn, J. (1996). Syllabic stress and variability: A model of stuttering. Clinical Linguistics and Phonetics, 10, 235-263. Paus, T., Marrett, S., Worsley, K., & Evans, A. (1996). Imaging motor-to-sensory discharges in the human brain: An experimental tool for the assessment of
ip t
functional connectivity. NeuroImage, 4, 78–86. Paus, T., Perry, D. W., Zatorre, R. J., Worsley, K. J., & Evans, A. C. (1996). Modulation of
cr
cerebral blood flow in the human auditory cortex during speech: Role of motor-to-
us
sensory discharges. European Journal of Neuroscience, 8, 2236–2246.
Perkins, W.H., Kent, R.D., & Curlee, R.F. (1991). A theory of neuropsycholinguistic
an
function in stuttering. Journal of Speech and Hearing Research, 34, 734-52. Rami, M.K., Kalinowski, J., Rastatter, M.P., Holbert, D., & Allen, M. (2005). Choral
M
reading with filtered speech: effect on stuttering. Perceptual and Motor Skills, 100, 421-431.
ed
Riley, G.D. & Bakker, K. (2009). Stuttering Severity Instrument-4 (SSI-4). Pro-Ed. Saltuklaroglu, T., Kalinowski, J., & Guntupalli, V.K. (2004). Towards a common neural
pt
substrate in the immediate and effective inhibition of stuttering. International
Ac ce
Journal of Neuroscience, 114, 435-450. Scott, S.K., Blank, C.C., Rosen, S., & Wise, J.R. (2000). Identification of a pathway for intelligible speech in the left temporal lobe, Brain, 123(12), 2400-2406.
Smits-Bandstra, S., & De Nil, L. F. (2007). Sequence skill learning in persons who stutter: implications for cortico-striato-thalamo-cortical dysfunction. Journal of Fluency
Disorders, 32, 251-278. Stager, S.V., Jeffries, K.J., & Braun, A.R. (2003). Common features of fluency-evoking conditions studied in stuttering subjects and controls: An H215OPET study. Journal of Fluency Disorders, 28, 319–336.
Page 37 of 50
The Role of Temporal Cues
38
Stuart, A. (1999). The distraction hypothesis and the practice of pseudoscience: A reply to Bloodstein. Journal of Speech, Language, and Hearing Research, 42, 913-914. Tsiamtsiouris, J., & Cairns, H. S. (2013). Effects of sentence-structure complexity on speech initiation time and disfluency. Journal of Fluency Disorders, 38, 30-44.
ip t
Van Riper, C. (1982). The nature of stuttering. (2nd ed.). Englewood Cliffs, NJ: PrenticeHall.
cr
Wingate, M. E. (1969). Sound and pattern in “artificial” fluency. Journal of Speech and
us
Hearing Research, 12, 677-86.
Wingate, M. E. (1984). Stutter events and linguistic stress. Journal of Fluency Disorders, 9,
an
295-300.
Zimmerman, G. (1980). Articulatory dynamics of fluent utterances of stutterers and
Ac ce
pt
ed
M
nonstutterers. Journal of Speech and Hearing Research, 23, 108-121.
Page 38 of 50
The Role of Temporal Cues
39
Appendix Listed below are examples of the stimulus sentences that were used to elicit speech samples. Sentences 11 – 20 are examples of the subset of stimuli with similar syntax and prosody that were used for speech sample elicitation as well as the speech entrainment analysis.
cr
2. Is the price of the parking decal for motorcycles reasonable?
ip t
1. The athlete who ran for student government selected his running mates.
3. The neighborhood residents will be talking about their primary concerns.
us
4. The journalism majors decided to form a new student newspaper.
5. Philadelphia scored eleven straight points in the fourth quarter and won.
an
6. Many graduate students will receive scholarships to pay for tuition.
M
7. Mr. Peters argued that children deserve quality health insurance coverage. 8. Christopher promised me he would review the problem before Saturday.
ed
9. The police officer ordered the crowd to step back immediately. 10. Jonathan asked, "Was Catherine on the soccer team last semester?"
pt
11. The policeman and detective had ticketed the Canadian by mistake. 12. The composer and teacher had directed the musical from September.
Ac ce
13. The historian from Kentucky will publish the biography on Tuesday. 14. The chancellor from Cambridge could disappoint the people by resigning. 15. The passenger from Baltimore will compliment the captain from Boston. 16. The governor of Tennessee will continue the discussion on Saturday. 17. The technician from Toshiba is purchasing the materials from Gateway. 18. The director of Patriot has donated the valuables to Benjamin. 19. The candidate for Congress was discussing his concerns with Democrats. 20. The governor of Colorado will pardon the lawbreaker by Tuesday.
Page 39 of 50
The Role of Temporal Cues
40
Ac ce
pt
ed
M
an
us
cr
ip t
Figure Captions Fig.1. Sample screen shot showing the extent of entrainment between a choral speech model (top panel) and a participant’s response (bottom panel) at five points within the utterance (dashed lines). The target sentence is The director of Patriot has donated the valuables to Benjamin. The participant’s response is closely aligned in time to the choral signal at entrainment points 1 and 2. At entrainment points 3, 4, and 5, however, the participant’s response is slightly ahead of the choral speech model, indicating that the participant’s articulation rate is faster than that of the choral model. The screen shot was taken from Praat [Boersma & Weenink, 2009]. Fig. 2. Mean frequency of stuttering-like disfluencies (SLDs) for participants who stutter across speaking conditions (Solo = normal auditory feedback; TCS = typical choral signal; ACS = temporally-altered choral signal; BAB = conversational babble feedback; Error bars = standard error.) Fig. 3. Speaking rate (mean number of syllables uttered per second) for participants who stutter and participants who do not stutter across the experimental speaking conditions (Solo = normal auditory feedback, no choral speech; TCS = typical choral signal; ACS = temporally-altered choral signal; BAB = conversational babble feedback; Error bars = standard error.) Fig. 4. Mean differences in phone onset time (in seconds) for speakers who stutter and speakers who do not stutter during the typical (TCS) and altered (ACS) choral speech conditions. Phone onset times for each participant were computed by subtracting the amount of elapsed time from the end of the standard carrier phrase that preceded each sentence from the onset of each of five predetermined reference points within the sentences. Participant phone onset times were then subtracted from the phone onset times in associated choral signal to provide information on the extent to which speaker’s articulation was entrained with the choral signal from the TCS and ACS conditions. Bars in the positive range indicate that the speaker’s articulation lags the choral signal. Alternately, bars in the negative range indicate the speaker’s articulation precedes the choral signal. Bar height indicates the extent of difference between the speaker’s output and the choral signal (RP1 = sentence-initial point; RP2, RP3, RP4 = three sentence-medial points, RP5= sentence-final point; Error bars = standard error)
Page 40 of 50
The Role of Temporal Cues
4
5
pt
ed
M
an
us
cr
ip t
3
Ac ce
12
41
Page 41 of 50
42
Ac ce
pt
ed
M
an
us
cr
ip t
The Role of Temporal Cues
Page 42 of 50
43
Ac ce
pt
ed
M
an
us
cr
ip t
The Role of Temporal Cues
Page 43 of 50
44
Ac ce
pt
ed
M
an
us
cr
ip t
The Role of Temporal Cues
Page 44 of 50
The Role of Temporal Cues
45
Biographical Sketch Jin Park, Ph.D. is a visiting professor in Department of Linguistics, Chungnam National University, Daejeon, Korea. In recent years, Dr. Park has taught courses in psycholinguistics and speech disorders at several universities in Korea. His current research examines the associations between stuttering, prosody, and neuromotor performance.
Ac ce
pt
ed
M
an
us
cr
ip t
Kenneth J. Logan, Ph.D., CCC-SLP, is an Associate Professor and Program Director in the Department of Speech, Language, and Hearing Sciences at the University of Florida, where he teaches courses, conducts research, and provides clinical services in the area of stuttering.
Page 45 of 50
The Role of Temporal Cues
46
ip t
CE Questions 1. What was NOT the main purpose of the present study? a. To investigate the role of temporal cueing in the speech fluency of adults who stutter b. To examine the mechanisms that underlie fluency enhancement during choral speaking c. To investigate whether adults who stutter depend on an accurate rendering of temporal information to benefit from choral speech d. To examine whether adults who stutter fluently under choral reading condition than speaking under delayed auditory feedback e. To examine whether adults who stutter follow choral signals more closely than adults who do not stutter do
Ac ce
pt
ed
M
an
us
cr
2. What was the rationale for using multi-speaker babble noise in the present study? a. It allowed for the concurrent presentation of a speech-like signal that contained coherent segmental cues b. It allowed for the concurrent presentation of a speech-like signal that contained coherent temporal cues c. It allowed for the presentation of a speech-like signal that lacked coherent temporal and segmental cues d. It allowed for examination of the “distraction hypothesis” e. It allowed for presentation of the “auditory masking effect” 3. Results from the present study DOES NOT indicate that: a. Adults who stutter spoke more fluently in temporally-altered choral speaking condition than they did when speaking solo. b. Adults who stutter spoke less fluently while listening to multi-speaker conversational babble than they did when speaking solo. c. Adults who stutter spoke more fluently in temporally-unaltered choral speaking condition than they did when speaking solo. d. Adults who stutter spoke slower in both of temporally-altered and unaltered choral speaking conditions than they did when speaking solo. e. Adults who stutter exhibited more temporal entrainment with the choral speech in both of altered and unaltered choral speaking conditions. 4. Results from the present study support the idea that: a. Adults who stutter make greater use of choral speech signals than adults who do not stutter b. Adults who stutter may not benefit from choral speech signals to attain the enhanced fluency c. Adults who stutter require accurate segmental models of lexical targets in order to enhance speech fluency d. Adults who stutter require accurate temporal models of lexical targets in order to enhance speech fluency e. Adults who stutter require accurate gestural models of phonetic targets in order to enhance speech fluency 5. According to information discussed in the discussion, what are possible explanations for the patterns of fluency facilitation observed during choral speech with speakers who stutter? a. Use of both a “temporal modeling strategy” and a “focused attention” . b. Use of a “relaxation” strategy. c. Use of a “rehearsal” strategy d. Use of both a “relaxation” and “rehearsal” strategy
Page 46 of 50
The Role of Temporal Cues
47
Ac ce
pt
ed
M
an
us
cr
ip t
e. Use of “regulated breathing” strategy.
Page 47 of 50
The Role of Temporal Cues
48
Ac ce
pt
ed
M
an
us
cr
ip t
CE Answers Key: 1 – d; 2 – c; 3 – b; 4 – a; 5 – a.
Page 48 of 50
The Role of Temporal Cues
49
Educational Objectives:
Ac ce
pt
ed
M
an
us
cr
ip t
The reader will be able to (a) summarize competing views on stuttering as a speech timing disorder, (b) describe the extent to which adults who stutter depend on an accurate rendering of temporal information in order to benefit from choral speech, and (c) discuss possible explanations for fluency facilitation in the presence of inaccurate or indeterminate temporal cues.
Page 49 of 50
cr
ip t
Table(s)
2 p .21 <.001 .22
F(1, 14) 8.02 62.96 .29
4 p .01 <.001 .60
2
P.E. .36 .82 .02
F(1, 14) 6.91 9.78 .003
5 p .02 .007 .96
P.E.2 .33 .41 0
M
P.E. .11 .86 .11
an
2
Reference Point 3 F(1, 14) p P.E.2 5.49 .01 .36 62.96 <.001 .83 .64 .44 .04
Ac c
ep te
d
1 Source F(1, 14) p P.E.2 F(1, 14) Group .31 .59 .02 1.78 Task 1.94 .19 .12 84.79 G x Ta 2.05 .17 .13 1.69 2 2 a ( P.E. = partial eta , G = Group; T = Task)
us
Table 1. Results of 2-way Analyses of Variance (Fluency Group x Speaking Task) Across the 5 Reference Points)
Page 50 of 50