Upload
others
View
2
Download
0
Embed Size (px)
Citation preview
1
Taitelbaum-Swead, R., Avivi, M., Gueta, B., Fostick, L. (2019). The effect of
delayed auditory feedback (DAF) and frequency altered feedback (FAF) on speech
production: cochlear implanted versus normal hearing individuals. Clinical
Linguistics & Phonetics.
***This is a self-archiving copy and does not fully replicate the published
version***
The Effect of Delayed Auditory Feedback (DAF) and Frequency Altered Feedback
(FAF) on Speech Production: Cochlear Implanted vs Normal Hearing Individuals
Running Title: Auditory Perturbation in CI and NH
2
Abstract
Normal auditory feedback contributes to moment to moment control of speech
production. Effects of auditory feedback’s absence on hearing-impaired individuals
are widely documented but auditory perturbation has not been investigated. Our
objective was to evaluate the effect of Delayed Auditory Feedback (DAF) and
Frequency Altered Feedback (FAF) on speech production among prelingual cochlear
implant (CI) users and normal hearing (NH) individuals, to evaluate CI users’ reliance
on auditory feedback. Twenty young adults (10 CI, 10 NH), without developmental
and cognitive impairments, participated in the study. Under variable auditory
feedback conditions, speech production (spontaneous or reading aloud) was measured
using speech rate, percentage of interruptions, fundamental frequency (F0), and
relative intensity. Results showed that: (1) Both DAF and FAF caused slower speech
rates and more interruptions while reading aloud, with DAF having larger effect, (2)
Altered feedback produced no differences between groups, except an increase in F0
for CI users during DAF, and (3) CI users’ ability to understand speech via phone and
without lip-reading was positively correlated with performance under DAF. These
findings suggest auditory perturbation similarly affects speech production among
prelingual CI users and NH individuals, indicating CI users depend on auditory
feedback to the same degree as normal hearing individuals.
Key words: speech production, Delayed Auditory Feedback, Frequency Altered
Feedback, auditory perturbation, cochlear implants
3
Introduction
Many speech perception theories presuppose a tight link between speech
perception and production (Liberman, Cooper, Shankweiler, & Studdert-Kennedy,
1967). This presupposition has been strengthened by neurobiological evidence.
Transcranial magnetic stimulation of the motor cortex has shown activation of
speech-related muscle areas during speech perception (Fadiga, Craighero, Buccino, &
Rizzolatti, 2002). In addition, fMRI studies have shown overlapping activated cortical
areas during speech production and passive listening to speech (Wilson, Saygin,
Sereno, & Iacoboni, 2004).
Further supporting the proposed speech perception/production association is
evidence gathered from studying the absence of auditory feedback on speech
production. There is compelling evidence that absence of auditory feedback during
language acquisition affects the development of speech production abilities (Blamey
et al., 2001). Among adults that have already acquired speech, the absence of auditory
feedback is also impactful, but in a different way (Goehl & Kaufman, 1984): It is
associated with deterioration in speech production over time. Studies that examined
the deterioration of segmental and supra-segmental features of speech in
adventitiously deaf adults ( Lane & Webster, 1991) found that the absence of auditory
feedback (due to losing the ability to hear) affected the accuracy of the acoustic-
phonetic features of some speech sounds such as vowels, voiced and voiceless
consonants (Waldstein, 1990), and substitution of affricates and fricatives (Leder &
Spitzer, 1990). At the supra-segmental level, adventitiously deaf individuals showed
changes in fundamental frequency, increases in voice intensity, and lengthening of
speech utterances (Leder & Spitzer, 1993). These changes caused their speech to
deteriorate further with time.
4
Auditory feedback affects speech not only when it is absent, but also when it
is altered. Studies have examined the effect of immediately altered auditory feedback
on speech production of normal hearing adults. These alterations included: Adding
background noise known as the Lombard effect (Lane & Tranel, 1971), Frequency
Alteration Feedback (FAF), (Elman, 1981), and delayed auditory feedback (DAF)
(Fairbanks & Guttman, 1958). These studies also found changes in speech production
as a result of altered feedback.
Among the variety of techniques for altering auditory feedback, Delayed
Auditory Feedback has been the most extensively studied technique for its effect on
speech production. As early as 1950, it was already reported that when a subject hears
himself with a DAF, he decreases his speech rate and fluency of speech, and increases
vocal intensity (Fairbanks & Guttman, 1958). In specific, a delay of 200 ms has been
reported to have a maximum negative effect on the speaker: This length delay
approximately matches the average syllable length, thus creating an interfering,
disruptive rhythm that makes monitoring the speech signal difficult for the listener
(Fairbanks, 1955). After these pioneering findings, many subsequent studies found
changes in speech production resulting from DAF. These changes include slowing of
the speech rate, prolongations of vowels, disfluencies and misarticulations
(Sasisekaran, 2012). These speech disruptions are thought to be corrective actions to
overcome the discrepancies between intended output and conflicting sensory
feedback.
Another type of auditory perturbation, FAF, in which the voice pitch is
unexpectedly shifted upward or downward and presented to the speaker while
speaking (Behroozmand, Korzyukov, Sattler, & Larson, 2012), also creates disruption
5
in speech production. Variations in the direction of fundamental frequency (F0) as a
result of pitch perturbation is reported in the literature (Petersen, 1986). Usually,
talkers show a compensatory pattern: When the perceived pitch is higher than the
intended pitch, the talker decreases F0 in order to compensate for the disparity.
Conversely, when the feedback pitch is downward, the F0 is increased (Behroozmand
et al., 2012). However, some talkers change their pitch in the same direction as the
feedback, thus making the F0 more distant from the original value (Jones & Munhall,
2003). The FAF technique is a useful method to directly investigate the relationship
between auditory feedback and pitch control during ongoing vocalizations.
While many of the previous studies examined the impact of altered auditory
feedback on normal hearing individuals, less attention has been paid to the impact on
those with cochlear implants. A large number of studies have demonstrated the
success of cochlear implants in providing better sound accessibility and enabling
better speech perception and production, among both children and adults (De Raeve,
Vermeulen, & Snik 2015; Svirsky, Robbins, Kirk, Pisoni, & Miyamoto 2000). The
literature has also shown evidence that speech production accuracy is closely related
to speech perception abilities in cochlear implants (CI) users. Studies found that early
speech intelligibility is strongly correlated with speech perception and language
abilities (Blamey et al., 2001; Tobey, Geers, Brenner, Altuna, & Gabbert, 2003).
Moreover, speech production proficiency also serves as a significant predictor of
development in speech perception and language skills after 10 years of CI use
(Casserly & Pisoni, 2013; Tobey, Geers, Sundarrajan, & Lane, 2011). Testing CI
users of different ages and onsets of deafness (such as prelingual children, prelingual
adults, and postlingual adults) can show how restored hearing affects speech
production differently among these groups (Blamey et al., 2001; Kishon-Rabin,
6
Taitelbaum, Tobin, & Hildesheimer 1999; Tobey et al., 2003). Longitudinal studies
have reported improvements in segmental and supra-segmental aspects of speech,
such as shifts in various acoustic and perceptual variables, generally toward the range
of normal (Kishon-Rabin et al., 1999; Svirsky, Jones, Osberger, & Miyamoto 1998;
Tye-Murray, Spencer, & Woodworth 1995). Some studies evaluated the short-term
effect of CIs on speech production of the hearing-impaired by turning off the CI in
order to remove auditory feedback (Lane et al., 2007; Svirsky et al., 1998); the
absence of auditory feedback resulted in speech production changes in vowel
formants (Lane et al., 2007; Svirsky & Tobey, 1991), nasalization (Svirsky et al.,
1998), production of sibilants (Lane et al., 2007), fundamental frequency and intensity
(Higgins, McCleary, & Schulte, 2001). These short-term effect studies suggest that CI
recipients who demonstrate better speech perception and production as a result of CI
usage may be relying on moment to moment auditory feedback, which could explain
the immediate deterioration of some speech contrasts when the CI is turned off (Lane
et al., 2007; Svirsky & Tobey, 1991). In an interesting study, Casserly (2015)
simulated the experience of hearing using a CI with normal hearing (NH) listeners.
Using this simulation, they evaluated the effect of real-time CI auditory feedback on
speech production in NH subjects, and found significant changes in the first formant
(F1) of vowels. This finding may reflect strategies to maximize kinesthetic feedback.
While the impact of the absence of auditory feedback on CI users is well-
established, the effect of auditory perturbations is less well-studied. The effect of
DAF has been compared utilizing individuals with different types of hearing
impairments (Barac-Cikoja, 2004; Tye-Murray, 1992); CI users were found to be
affected by DAF differently than participants with hearing aids, as were prelingual CI
children compared to postlingual CI children. However, these studies were limited by
7
the fact that none of these groups were compared to normal hearing participants. A
study that did, compare postlingual CI users to a normal hearing group, used a
variation of the Lombard effect, testing the impact of adding different levels of
background noise on speech production (Perkell et al., 2007). They found that sibilant
contrasts were susceptible to noise. These sibilant contrasts decreased over the entire
range of increasing noise levels for NH and was variable for CI users. Moreover, CI
users were less able to increase contrasts during noise, compared to their NH peers.
The results of this study suggest that CI users rely on auditory feedback differently
than NH individuals. However, the manipulation was subjected only to the addition of
background noise. Therefore, it is still unknown whether prelingual CI users will be
sensitive to altered temporal and spectral auditory feedback, compared to normal
hearing participants.
The current study
To our knowledge, this is the first study to evaluate the effect of altered
auditory feedback on speech production of prelingual CI user adults compared to NH
adults. This study paradigm enabled us to test whether CI users that were implanted
early in their life, rely on auditory feedback to the same degree as their NH peers,
shedding light on the role of hearing in speech production among adults.
From a clinical perspective, this paradigm is an additional method to measure
the reliance on auditory feedback provided by the CI. In order to estimate how much
this reliance on auditory feedback is currently reflected in day to day life, we asked CI
participants about their functional hearing (understanding speech via phone, without
lip-reading, and in multiple-speaker situations). We set out to analyze the correlation
between this self-assessment of functional hearing and the speech production that
8
would occur as a result of altered feedback. This correlation was expected to give
further insight into the nature of the relationship between the clinical measurements of
the impact of auditory feedback and the subjective impression of its role in day-to-day
life. The results of the current study are anticipated to shed light on the extent of the
relationship between speech perception and production among CI users.
Method
Participants
Two groups of subjects were enrolled in the study: Individuals with CIs and
those with normal hearing. All were native Hebrew speakers and had no
developmental and cognitive impairments.
CI users
The CI group consisted of 10 implanted young adults (8 women and 2 men)
who met the following inclusion criteria: (1) onset of severe to profound hearing
impairment before 3 years of age; (2) hearing aid usage prior to implantation; (3)
mainstream education and oral communication; and (4) usage of multichannel
cochlear implants. The mean chronological age of the CI group, their hearing devise,
age at implantation, etiology of hearing loss, and type of implant are described in
Table 1.
INSERT TABLE 1
NH participants
9
The NH group consisted of 10 young women (n=8) and men (n=2),
undergraduate students aged 20-25 years, who had normal hearing thresholds pure-
tone air-conduction thresholds less than 15 dB HL bilaterally at octave frequencies
from 250 - 4,000 Hz.
Task and stimuli
The stimuli were delivered in 70 dB SPL under three different conditions: (1)
Normal Auditory Feedback (NAF)- a natural condition (for NH) and with the implant
on (for CI); (2) Delay Auditory Feedback (DAF) – delay of 200 milliseconds (found
to be most impactful on speech production) with no change in spectral characteristics
of the speech; (3) Frequency Altered Feedback (FAF) – Increase of 200 cents (0.167
octaves) in F0 with no change in temporal characteristics of the stimuli
(Behroozmand et al., 2012). Stimuli consisted of: (1) reading aloud from three
different passages (one for each condition) of 180-200 syllables that included all
consonants and vowels in the Hebrew language, and (2) spontaneous speech – the first
fifteen seconds of participants’ answers to everyday questions delivered by the
researchers (three questions- one for each condition: Describe your living
environment, the route from your home to the nearest grocery store, and your job).
The CI participants also answered three questions about their subjective level
of functioning with the implant in daily activities. The questions asked about the CI’s
hearing ability (1) via phone, (2) listening to speaker in the same room, but without
the availability of lip-reading (such as when not seeing the speaker), and (3) in a
conversation with multiple talkers. Each question was answered on a 1-4 scale.
Apparatus
10
Auditory perturbation was done using application developed by Kosti. This
application implemented delayed auditory feedback with a scale unit of milliseconds
and fundamental frequency changes of semitones (100 cents). The application was
operated by a mini Ipad 2013. The stimuli were delivered to the NH group by insert
earphones for the Ipad and to the CI group by a mobile induction hook (music and
mobile) connecting to the CI with electromagnetic telecoil. Recordings were done
using a Sennheiser MZH 3072 microphone positioned 10 cm from the speaker’s
mouth and external sound card (Asus Xonar U7). All recordings were saved onto a
laptop (Dell). Recordings were edited using Sound Forge 11 software which digitized
(16-bit) the stimuli at a sampling rate of 44 kHz and all acoustic analyses were done
by Praat software.
Acoustic Analyses
Speech rate was determined by calculating the number of syllables spoken per
second, divided by the length of the passage in seconds; silent pauses were not
included. In addition, the number of interruptions was also calculated: Including
omissions, substitutions and duplications of syllables or words. The number of
interruptions was divided by the number of syllables per condition.
The fundamental frequency and the relative intensity were measured only
during the reading of passages. A fundamental frequency extraction algorithm, a
built-in feature of the Praat program based on an autocorrelation function, was used to
estimate the mean F0. Mean relative intensity was calculated automatically by the
Praat software.
Procedure
11
The study was approved by the Institutional Review Board and was conducted
in accordance with Good Clinical Practice (GCP) guidelines. All participants received
a full explanation about the study and signed an informed consent document. All
potential candidates for the NH group were screened for hearing levels prior to
participation in the study.
Participants were required to read a passage and to answer the everyday
questions. The order of the conditions and the order of the speech stimuli (reading
passage and spontaneous speech) were randomly intermixed across participants. Each
participant was recorded under the three different conditions (NAF, DAF, and FAF)
in a single session.
Statistical analyses
Repeated measures analyses of variance (ANOVAs) were performed on each
of the dependent variables. Analysis for speech rate was carried out with all types of
feedback (NAF, DAF, FAF) and speech (spontaneous, reading) as within-subjects
effects, and group (CI, NH) as a between subjects effect. Analyses for percentage of
interruptions, fundamental frequency, and relative intensity, were carried out only on
reading scores due to their fixed number of syllables; therefore, analyses for these
variables were done with only the type of feedback (NAF, DAF, FAF) as within-
subjects effects, and group (CI, NH) as a between-subjects effect. Post-hoc analysis
was done using Least Significant Difference (LSD) tests. Pearson product-moment
correlations were used to measure association between dependent variables and self-
report measures of speech intelligibility among the CI users: (1) during phone
conversation, (2) when lip reading is not available, and (3) in multiple-speaker
situations.
12
Results
Table 2 presents means and SDs of CI and NH participants for all dependent
variables, across feedback and speech types.
INSERT TABLE 2
Speech rate
Significant main effects were found for feedback type (F(2,36) = 13.782, p <
.001, partial η2 = .434) and speech type (F(1,18) = 13.697, p = .002, partial η2 = .432).
Speech rate was the slowest with DAF, compared to both NAF (LSD = 1.058, p <
.001) and FAF (LSD = .527, p = .001). Speech rate with FAF was also slower than
NAF (LSD = .531, p = .032). Speech rate in reading (Mean = 4.357, SD = .161) was
faster than in spontaneous speech (Mean = 3.616, SD = .181). Figure 1 presents
means and SD of speech rates in spontaneous speech (in all conditions: NAF, FAF,
DAF) for CI and NH groups. No effect was found for group on speech rate (F(1,18) =
2.221, p = .153, partial η2 = .110), nor interactions for group X feedback type (F(2,36) =
.416, p = .663, partial η2 = .023), group X speech type (F(1,18) = .427, p = .522, partial
η2 = .023), feedback type X speech type (F(2,36) = .499, p = .611, partial η2 = .027), or
group X feedback type X speech type (F(2,36) = .585, p = .562, partial η2 = .031.
INSERT FIGURE 1
Interruptions
13
Figure 2 presents the mean percentage of interruptions in spontaneous speech
(in all conditions: NAF, FAF, DAF) for CI and NH groups. A significant main effect
was found for feedback type (F(2,32) = 7.683, p = .002, partial η2 = .299), but not for
group (F(1,16) = 2.209, p = .155, partial η2 = .109). A higher percentage of
interruptions occurred with DAF than NAF (LSD = .032, p = .008) or FAF (LSD =
.025, p = .022), which also had more interruptions than NAF (LSD =.007, p = .002).
No interaction of group X feedback type (F(2,32) = .656, p = .525, partial η2 = .035) was
observed.
INSERT FIGURE 2
Fundamental frequency and relative intensity
No main effects in fundamental frequency were found for group (F(1,18) =
2.277, p = .150, partial η2 = .118) and feedback type (F(2,36) = 2.165, p = .130, partial
η2 = .113), but a group X feedback type interaction was found (F(2,36) = 3.232, p =
.052, partial η2 = .160). Post-hoc repeated measures analysis showed a main effect for
feedback type only for the CI group (F(2,16) = 4.294, p = .042, partial η2 = .349), but
not for the control group (F(2,16) = 2.514, p = .127, partial η2 = .218). CI users’
fundamental frequency significantly increased following DAF, as compared to NAF
(LSD = 18.078, p = .035). No difference was observed between FAF and NAF (LSD
= 12.411, p = .103) or DAF and FAF (LSD = 5.667, p = .269).
No main effects in relative intensity were found for group (F(1,18) = .513, p =
.483, partial η2 = .029) and feedback type (F(2,36) = 1.932, p = .160, partial η2 =.102),
and no group X feedback type interaction (F(2,36) = 2.601, p = .089, partial η2 = .133)
was found.
14
3.4 Self-report measures
Table 3 presents correlations for self-report intelligibility measures (phone
conversation, without lip-reading, multiple-talkers), with study dependent variables.
The ability to understand phone conversation was positively correlated with reading
aloud fundamental frequency raw score, during DAF. That is, the better CI users
understand phone conversation, the more the fundamental frequency increases while
reading aloud, during DAF. Speech intelligibility in the no lip-reading condition was
positively correlated with interruptions while reading aloud during DAF. No
significant correlations were found for the intelligibility of the multiple-talker
condition.
INSERT TABLE 3
Discussion
The current study compared the effect of altered auditory feedback on the
speech production of both adult prelingual CI users and NH individuals, in order to
evaluate the degree of reliance on auditory feedback among CI users. Both DAF and
FAF were found to affect speech production as evidenced by slower speech rates and
more interruptions while reading aloud. DAF had a larger negative effect on speech
production than FAF. The most significant study finding was the lack of difference
between CI users and NH individuals resulting from altered feedback, except for
increase in F0 during the DAF condition that appeared for CI users only. Also, of note
15
was the finding that CI users’ self-reported ability to understand phone conversation
and speech without lip-reading was positively correlated with the effects of DAF.
This finding suggests that CI users rely on auditory feedback in their daily hearing
functioning.
Significantly, this is the first study to compare CI users to NH participants on
the effects of DAF. Previous studies that tested the effect of DAF on the speech
production of hearing-impaired individuals did not compare them to the NH
population (Barac-Cikoja, 2004; Tye-Murray, 1992). Our finding that altered auditory
feedback similarly influences NH and CI users for most speech production features
(speech rate, interruptions, and relative intensity) can be considered null effect, or to
suggest that prelingual CI users have learned to rely on auditory feedback while
producing speech. This means that, like NH individuals, the speech motor system of
CI users utilizes relevant information from auditory feedback in order to fine-tune its
speech movement to achieve end goals. Our finding demonstrates a direct effect of
delayed auditory feedback on overall movement coordination in continuous speech of
CI users. Indeed, studies that previously tested CI users suggested that their reliance
on the CI increases with time (Dettman et al., 2016; Ertmer & Goffman, 2011; Tobey
et al., 2011), showing a growth in speech production abilities within the first six years
of implant use (Blamey et al., 2001; Tomblin, Peng, Spencer, & Lu, 2008). It is also
important to note that previous studies usually tested the effect of CI on prelingual
children up until their high school years and used speech intelligibility measures
(Tobey et al., 2011). Therefore, the added contribution of the present study to the
existing literature is the testing of prelingual implantees in their twenties who have
used their implants for a long period of time. Moreover, our method of using altered
feedback to assess CI users’ reliance on auditory feedback from their implants (as
16
evidenced by disrupted speech production) is a unique feature of the study design.
The results indicate that, for young adults who are prelingual CI users, reliance on
auditory feedback through their implants is already well-developed.
Although CI users, as a group, were affected by altered feedback similarly to
the NH participants in most features of speech production, we found a positive
correlation between the CI group’s daily functional hearing self-report of hearing
ability via phone, and their speech features (interruptions and F0) under DAF. This
result is in line with previous studies showing such a relationship among cochlear
implant users (Kishon-Rabin, Gehtler, Taitelbaum, Kronenberg, Muchnik, &
Hildesheimer, 2002; Tye-Murray et al., 1995). This finding is supported by the
Directions Into Velocities of Articulators (DIVA) model (Guenther, Ghosh, &
Tourville, 2006) suggests that lexical retrieval of strings of words leads to sequential
activation of speech sound map cells that send signals to cells in the auditory,
somatosensory, and primary motor cortical areas; these signals lead to production of
speech sounds through feed-forward and feedback systems. The feed-forward system
produces skilled, rapid movements that do not rely on external (auditory,
somatosensory) feedback. The feedback system teaches, refines, and updates the
movements based on error detection and correction. When auditory feedback is
removed, the production of speech relies only on the feed-forward system, along with
some somatosensory feedback. The results of the present study suggest that CI users’
feed-forward subsystem is well-tuned, similarly to that of NH individuals. The
resemblance of the experimental design to phone conversation can explain the
correlation found with this function and not the others (hearing without the
availability of lip-reading and multiple talkers’ conversation).
17
In the present study, significant changes in speech production among both
groups occurred when DAF was applied. The delay provided was 200 ms, which was
previously found to have a maximum negative effect on the speaker by making self-
monitoring of the speech signal significantly difficult for the listener (Fairbanks,
1955). Studies that previously evaluated the impact of delayed feedback on normal
hearing participants also found lower speech rates and an increase in the number of
interruptions (Kort, Nagarajan, & Houde,2014; Sasisekaran, 2012). Thus, the effect of
temporal alterations on speech production as a result of DAF was more pronounced in
the current study than the effect of spectral alterations as a result of FAF
manipulation. This finding is different from those reported in the literature showing
FAF also to have noteworthy effect on speech production, particularly in fundamental
frequency (Behroozmand et al., 2012). The discrepancy between the results of the
current and previous studies might be due to the methods of speech production
utilized in our study design. In the present study, we used continuous speech (either
reading text aloud or spontaneous speech) while most of the previous studies that
examined FAF used sustained vowels (Behroozmand et al., 2012). It may be that
when using spontaneous speech, the speaker compensates more quickly for changes
spectral pitch, relative to sustained vowels. The discrepancy also might be explained
by our study’s greater emphasis on measures of temporal changes (speech rate and
interruptions) than on spectral changes (fundamental frequency only). Thus, the
results of the current study may reflect larger sensitivity to temporal alterations.
Regarding spectral changes, the CI group showed pitch variations as a result
of DAF manipulation: Most of the CI participants showed an increase in mean
fundamental frequency under DAF during the reading aloud condition. This is in line
with the traditionally-reported finding that the pitch of hearing-impaired individuals'
18
voices are characterized as “too high” compared to normal-hearing speakers (Gilbert
& Campbell, 1980). Moreover, studies that tested short-term effects of turning off
implants also showed immediate elevation in the F0 as a result of auditory feedback
concealing (Higgins et al., 2001). Taken together with the findings of the present
study, it seems that CI users may be using this pattern of increasing their F0 when
auditory feedback is altered or nonexistent. Surprisingly, however, the NH group did
not change their F0 as a result of DAF. It may be that a delay of 200 milliseconds is
not enough to produce such a pitch change among NH participants, but for CI users,
who are more prone to increasing their F0, it is sufficient. Further studies should
explore this direction.
Declaration of Interest
The authors have no declaration of interest to report. The study was not supported by
any sponsor or funding agency.
Acknowledgment
The author would like to thank Shira Chana Bienstock for her thorough editorial
review of this manuscript.
19
References
Barac-Cikoja, D. (2004). Effects of temporal and spectral alterations of speech
feedback on speech production by persons with hearing loss. Archives of
Otolaryngology–Head & Neck Surgery, 130, 598-603.
Behroozmand, R., Korzyukov, O., Sattler, L., & Larson, C. R. (2012). Opposing
and following vocal responses to pitch-shifted auditory feedback: evidence for
different mechanisms of voice pitch control. The Journal of the Acoustical Society of
America, 132, 2468-2477.
Blamey, P. J., Sarant, J. Z., Paatsch, L. E., Barry, J. G., Bow, C. P., Wales, R. J.,
... & Tooher, R. (2001). Relationships among speech perception, production,
language, hearing loss, and age in children with impaired hearing. Journal of Speech,
Language, and Hearing Research, 44, 264-285.
Casserly, E. D. (2015). Effects of real-time cochlear implant simulation on
speech production. The Journal of the Acoustical Society of America, 137, 2791-2800.
Casserly, E. D., & Pisoni, D. B. (2013). Nonword repetition as a predictor of
long-term speech and language skills in children with cochlear implants. Otology and
Neurotology, 34, 460.
De Raeve, L., Vermeulen, A., & Snik, A. (2015). Verbal cognition in deaf
children using cochlear implants: effect of unilateral and bilateral stimulation.
Audiology and Neurotology, 20, 261-266.
Dettman, S. J., Dowell, R. C., Choo, D., Arnott, W., Abrahams, Y., Davis, A.,
... & Briggs, R. J. (2016). Long-term communication outcomes for children receiving
cochlear implants younger than 12 months: A multicenter study. Otology and
Neurotology, 37, e82-e95.
20
Elman, J. L. (1981). Effects of frequency‐shifted feedback on the pitch of
vocal productions. The Journal of the Acoustical Society of America, 70, 45-50.
Ertmer, D. J., & Goffman, L. A. (2011). Speech production accuracy and
variability in young cochlear implant recipients: comparisons with typically
developing age-peers. Journal of Speech, Language, and Hearing Research, 54, 177-
189.
Fadiga, L., Craighero, L., Buccino, G., & Rizzolatti, G. (2002). Speech
listening specifically modulates the excitability of tongue muscles: a TMS study.
European journal of Neuroscience, 15, 399-402.
Fairbanks, G. (1955). Selective vocal effects of delayed auditory feedback.
Journal of Speech and Hearing Disorders. 20,333-346.
Fairbanks, G., & Guttman, N. (1958). Effects of delayed auditory feedback
upon articulation. Journal of Speech & Hearing Research. 1, 12-22.
Gilbert, H. R., & Campbell, M. I. (1980). Speaking fundamental frequency in
three groups of hearing-impaired individuals. Journal of Communication Disorders,
13, 195-205.
Goehl, H., & Kaufman, D. K. (1984). Do the effects of adventitious deafness
include disordered speech? Journal of Speech and Hearing Disorders, 49, 58-64.
Guenther, F.H., Ghosh, S.S., Tourville, J.A. 2006. Neural modeling and
imaging of the cortical interactions underlying syllable production. Brain and
Language, 96, 280-301.
Higgins, M. B., McCleary, E. A., & Schulte, L. (2001). Articulatory changes
with short-term deactivation of the cochlear implants of two prelingually deafened
children. Ear and Hearing, 22, 29-41.
21
Jones, J. A., & Munhall, K. G. (2003). Learning to produce speech with an
altered vocal tract: The role of auditory feedback. The Journal of the Acoustical
Society of America, 113, 532-543.
Kishon-Rabin, L., Gehtler, I., Taitelbaum, R., Kronenberg, J., Muchnik, C., &
Hildesheimer, M. (2002). Development of speech perception and production in
children with cochlear implants. Annals of Otology, Rhinology & Laryngology,
111(5_suppl), 85-90.
Kishon-Rabin, L., Taitelbaum, R., Tobin, Y., & Hildesheimer, M. (1999). The
effect of partially restored hearing on speech production of postlingually deafened
adults with multichannel cochlear implants. The Journal of the Acoustical Society of
America, 106, 2843-2857.
Kort, N. S., Nagarajan, S. S., & Houde, J. F. (2014). A bilateral cortical
network responds to pitch perturbations in speech feedback. Neuroimage, 86, 525-
535.
Lane, H., Matthies, M. L., Guenther, F. H., Denny, M., Perkell, J. S.,
Stockmann, E., ... & Zandipour, M. (2007). Effects of short-and long-term changes in
auditory feedback on vowel and sibilant contrasts. Journal of Speech, Language, and
Hearing Research, 50, 913-927.
Lane, H., & Tranel, B. (1971). The Lombard sign and the role of hearing in
speech. Journal of Speech, Language, and Hearing Research, 14, 677-709.
Lane, H., & Webster, J. W. (1991). Speech deterioration in postlingually
deafened adults. The Journal of the Acoustical Society of America, 89, 859-866.
Leder, S. B., & Spitzer, J. B. (1990). A perceptual evaluation of the speech of
adventitiously deaf adult males. Ear and Hearing, 11, 169-175.
22
Leder, S. B., & Spitzer, J. B. (1993). Speaking fundamental frequency,
intensity, and rate of adventitiously profoundly hearing‐impaired adult women. The
Journal of the Acoustical Society of America, 93, 2146-2151.
Liberman, A. M., Cooper, F. S., Shankweiler, D. P., & Studdert-Kennedy, M.
(1967). Perception of the speech code. Psychological Review, 74, 431-461.
Perkell, J. S., Denny, M., Lane, H., Guenther, F., Matthies, M. L., Tiede, M.,
... & Burton, E. (2007). Effects of masking noise on vowel and sibilant contrasts in
normal-hearing speakers and postlingually deafened cochlear implant users. The
Journal of the Acoustical Society of America, 121, 505-518.
Petersen, N. R. (1986). Perceptual compensation for segmentally conditioned
fundamental frequency perturbation. Phonetica, 43, 31-42.
Sasisekaran, J. (2012). Effects of delayed auditory feedback on speech
kinematics in fluent speakers. Perceptual and Motor Skills, 115, 845-864.
Svirsky, M. A., Jones, D., Osberger, M. J., & Miyamoto, R. T. (1998). The
effect of auditory feedback on the control of oral-nasal balance by pediatric cochlear
implant users. Ear and Hearing, 19(5), 385-393.
Svirsky, M. A., Robbins, A. M., Kirk, K. I., Pisoni, D. B., & Miyamoto, R. T.
(2000). Language development in profoundly deaf children with cochlear implants.
Psychological Science, 11(2), 153-158.
Svirsky, M. A., & Tobey, E. A. (1991). Effect of different types of auditory
stimulation on vowel formant frequencies in multichannel cochlear implant users. The
Journal of the Acoustical Society of America, 89, 2895-2904.
Tobey, E. A., Geers, A. E., Brenner, C., Altuna, D., & Gabbert, G. (2003).
Factors associated with development of speech production skills in children implanted
by age five. Ear and Hearing, 24, 36S-45S.
23
Tobey, E. A., Geers, A. E., Sundarrajan, M., & Lane, J. (2011). Factors
influencing elementary and high-school aged cochlear implant users. Ear and
hearing, 32, 27S-38S.
Tomblin, J. B., Peng, S. C., Spencer, L. J., & Lu, N. (2008). Long-term
trajectories of the development of speech sound production in pediatric cochlear
implant recipients. Journal of Speech, Language, and Hearing Research, 51, 1353-
1368.
Tye‐Murray, N. (1992). Young cochlear implant users’ response to delayed
auditory feedback. The Journal of the Acoustical Society of America, 91, 3483-3486.
Tye-Murray, N., Spencer, L., & Woodworth, G. G. (1995). Acquisition of
speech by children who have prolonged cochlear implant experience. Journal of
Speech, Language, and Hearing Research, 38, 327-337.
Waldstein, R. S. (1990). Effects of postlingual deafness on speech production:
implications for the role of auditory feedback. The Journal of the Acoustical Society
of America, 88, 2099-2114.
24
Table 1. Characteristics of Cochlear Implant (CI) participants
Chronological age
Mean (SD) 22.5 (4.2)
Hearing device
Two implants n=6
One implant and hearing aid n=2
One implant n=2
Age at implantation
Mean (SD) of first 4.4 (5.02)
Mean (SD) of second 13 (5.2)
Etiology of hearing loss
Genetic n=1
Cytomegalovirus (CMV) n=2
Usher syndrome n=1
Unknown n=6
Type of implant
Nucleus cochlear n=8
Advanced Bionics n=2
25
Table 2. Means and SDs for CI and NH participants on all study measures
Spontaneous Speech Reading Aloud
NH CI NH CI
Mean SD Mean SD Mean SD Mean SD
Speech rate
NAF 4.34 0.97 4.00 1.68 5.00 0.62 4.72 0.90
DAF 3.25 .99 2.77 0.95 4.29 0.86 3.52 0.97
FAF 3.68 0.80 3.65 1.06 4.59 .96 4.01 0.79
Interruptions
NAF 0% 0.00 1% 0.01
DAF 3% 0.04 5% 0.06
FAF 1% 0.01 2% 0.02
Fundamental frequency
NAF
184.74 51.21 203.97 35.81
DAF
173.39 43.17 222.04 43.47
FAF
200.66 47.16 194.16 75.81
Relative intensity
NAF
66.21 9.81 66.63 13.22
DAF
66.58 11.00 70.61 9.48
FAF
65.40 11.68 71.19 9.09
NAF = Normal Auditory Feedback; DAF = Delayed Auditory Feedback; FAF =
Frequency Altered Feedback; CI = Cochlear Implant; NH = Normal Hearing
26
Table 3. Correlations between self-report speech intelligibility measures (via phone,
without lip-reading, with multiple-talkers) and study dependent variables.
Spontaneous Speech Reading Aloud
Phone No lip-reading Multiple-talkers Phone No lip-reading Multiple-talkers
Speech rate
NAF 0.424 -0.077 0.111 0.44 0.088 0.377
DAF -0.226 -0.544 0.164 0.083 -0.176 0.268
FAF 0.42 0.097 0.373 0.559 0.308 0.451
Interruptions
NAF -0.206 -0.223 -0.574 -0.041 0.387 0.217
DAF -0.055 0.192 -0.117 0.27 .646* 0.224
FAF -0.351 -0.515 -0.662 0.229 0.43 0.152
Fundamental frequency
NAF
0.557 0.274 0.128
DAF
.786* 0.234 0.313
FAF
0.08 0.376 -0.175
Relative intensity
NAF
-0.559 -0.356 -0.274
DAF
-0.398 -0.161 -0.287
FAF
-0.473 -0.442 -0.1
NAF = Normal Auditory Feedback; DAF = Delayed Auditory Feedback; FAF =
Frequency Altered Feedback
*p<.05
27
Figure 1. Mean speech rate and SD in spontaneous speech (in all conditions: NAF,
FAF, DAF) of CI and NH groups.