Training Slow Speaking Rate to Treat Motor Speech Disorders

1
Training Slow Speaking Rate to Treat Motor Speech
Disorders
Authors Laura Clinton Greg Turner, Ph.D.,
CCC-SLP Robert de Jonge, Ph.D., CCC-A
2
INTRODUCTION
Motor speech disorders (dysarthria and apraxia of
speech) result from neurological impairment in
the areas of the brain responsible for motor
planning, motor programming, neuromuscular
control, and execution of speech and they
account for 41 of all acquired communication
disorders (Yorkston, Beukelman, Strand, Bell,
1999 Duffy, 2005). Dysarthria can cause a
reduction in speaking rate, abnormal prosodic
features, and significant intelligibility
deficits resulting in a reduced ability to
communicate successfully (Ansel Kent, 1992
Kent, 1992 Duffy, 2005).
3
Apraxia of speech is commonly characterized by
groping, inconsistent articulation errors,
especially in longer utterances, and increased
errors in repetition based tasks therefore, the
apraxic speaker exhibits problems of speech
initiation, speech sequencing, speech selection,
reduced rate of speech, disturbances in
articulation and rhythm, abnormal prosody, and
decreased intelligibility (Dworkin, 1991 Chapey,
1994 McNeil, 1997 Yorkston, et al., 1999
Wambaugh Nessler, 2004).
4
Treatment of motor speech disorders may consist
of behavioral, medical, and assistive treatment
techniques. (Duffy, 2005). One behavioral
technique to improve communication is slowing
speaking rate. Duffy (2005) states, Rate may
be the most powerful single, behaviorally
modifiable variable for improving
intelligibility (p. 479).
5
Purpose
Slowing speaking rate has been found to be an
effective therapy technique for both dysarthria
and apraxia of speech (Simmons, 1978 Yorkston
and Beukelman, 1981 Yorkston et al., 1999). One
of the outcomes of slowing speaking rate is an
improvement in speech intelligibility (Southwood,
1987 Wambaugh and Martinez, 2000) however,
few studies have documented the effectiveness of
methods for the control of speaking rate in
individuals with motor speech disorders. The
purpose of this study is to evaluate a training
program aimed at slowing speaking rate for read
sentences in an individual with both dysarthria
and apraxia of speech through the use of visual,
auditory, and verbal feedback.
6
METHODS
Participant The participant in this investigation
was a forty-nine-year-old male who has
developmental speech and language deficits (see
Table 1). He reported that though he had the
communication disorder since childhood, he
received no speech remediation services during
the years of his formal education. The
participant was diagnosed with mild-moderate
apraxia of speech and dysarthria. The
participant had not received speech therapy for
two years prior to this study.
7
Table 1 Assessment Results
8
Experimental Design The experimental design used
in this study consisted of a single-subject ABAB
research design. The first A phase is the
baseline phase and was continued until speaking
rate stability was achieved. Stability was
defined as average sentence duration of the 18
stimulus sentences falling within /-5 of each
other over three consecutive baseline sessions.
The first B phase consists of the treatment phase
where the participant was trained to reduce
speaking rate. The second A phase represents the
withdrawal phase. During this phase the
participant did not receive treatment.
Reinstatement of treatment represents the second
B phase and was completed in the same way as the
first treatment phase. Sentence duration and
speech intelligibility was acquired on a weekly
basis during the last three phases and every
other day during the first phase.
9
Choosing Target Speaking Rate

The speaking rate that maximized the
participants intelligibility was used as the
target rate for this study. To find this rate,
the participant produced the 18 stimulus
sentences (each seven syllables in length) at his
habitual rate and at 90, 80, 70, 60, and 50
of the habitual rate. The participant was
presented with a prerecorded auditory model of
each of the 18 sentences at a targeted sentence
duration associated with the targeted speaking
rate (e.g., 90). The intensity waveform of each
model was presented both auditorily and visually
to the participant through the use of the
Real-Time Pitch program within the Sona-Speech
II, model 3650 program (KayPentax, 2006). In
addition, the participant was trained to use the
visual feedback of his own sentence productions
from the Real-Time Pitch program to match the
model production.
10
The investigator put two vertical lines on both
the top and bottom panels of the Real-Time Pitch
screen indicating the target duration range (i.e.
mean plus and minus five percent plus reaction
time). The participant was asked to produce the
different target rates by slowing articulation
time and not increasing pause time. The
participant produced each sentence until one
production fell within the target duration range.
This resulted in a total of 108 sentence
productions (18 sentences X 6 speaking
rates). Three listeners listened to a digitized
recording of each of the 18 sentences for a given
speaking rate and orthographically transcribed
each sentence. The speaking rate judged as
having the highest intelligibility rating was 60
of the participants habitual speaking rate.
11
Treatment
The treatment phase of this study focused
strictly on slowing rate of the nine training
sentences to a target durational range of
4.39-4.85 seconds. The target sentence duration
range represents /- 5 of the average duration
of the 60 target speaking rate. This range was
chosen based on a study by Hyland and Weismer
(1988). The participants target duration range
was changed from /- 5 to /- 7.5 of the
participants habitual speaking rate during week
six of the first treatment phase. This was done
based on the observation of the experimenter that
the participant was consistently short of the
smaller target range and was unable to make
progress.
12
For each sentence produced by the participant the
investigator first displayed and played one of
her prerecorded sentences in the upper panel of
the computer screen and noted how the end of the
time history of her sentence ended between the
vertical lines indicating target duration range.
The participant was directed to the lower screen
containing the same set of vertical lines. When
the participant was prepared, he was asked, Are
you ready by the instructor, he replied Yes
and the instructor then said, Begin and began
recording at that time while the participant then
produced the same sentence. The sentence was
recorded and displayed as an intensity time
history trace in the lower panel of the computer
screen. This was repeated three times for each
sentence regardless of accuracy.
13
The participant received verbal, auditory, and
visual feedback after his production. Verbal
feedback consisted of knowledge of performance
and knowledge of results. Treatment sessions
consisted of distributed practice and random
practice of the training sentences within and
across treatment sessions, with the focus on
slowing rate by increasing articulation time and
not by inserting pauses. Treatment continued
until the participant reached the target speaking
rate of 60 of his habitual rate with 80
accuracy over three consecutive treatment
sessions. After reaching the criterion, the
withdrawal phase began. The Second B Phase was
conducted in the same way as the First B Phase.
The last three phases of the study continued for
seven weeks.
14
One time each week, throughout the treatment
phase, withdrawal of treatment phases, and
reinstatement of treatment phase, both the
trained and untrained sentences were recorded in
a single walled sound booth. One time a week, on
random days, the participant was tape recorded
using the Tascam DAT tape recorder and Crown
CM311A head mounted microphone while
spontaneously producing the 9 trained sentences
and the 9 untrained sentences to evaluate
generalization. Each tape recorded sentence was
digitized at 44.1 kHz via the TF32 program
(Milenkovic, 2002) and displayed on the screen in
the form of a sound pressure waveform. Cursors
were placed at the beginning and end of the
sentence for measurement of sentence duration.
15
In order to collect weekly intelligibility
measures, the digitized sentences were normalized
to -3 dB and presented to three listeners through
a customized software program developed in
Revolution 2.8.1 (www.runrev.com). The PC-based
program used Quick Time to play .wav files
through the computers sound card. The listeners
had the opportunity to listen to each sentence
twice (Yorkston Beukelman, 1981). Each
listener listened to the sentences through
loudspeakers at a self-chosen most comfortable
listening level. The listeners orthographically
transcribed the sentences on a keyboard. The
investigator calculated the percent correct
(total words correctly identified) for each of
the three listeners transcription for the
trained and untrained sentences and calculated an
average percent correct across the three
listeners. The intelligibility data will not be
presented in this poster.
16
RESULTS
Accuracy Results Within the Treatment
Program Accuracy data was collected during the
treatment phases (Refer to Figure 1). The first
treatment phase was terminated after the
participant achieved 80 accuracy of the target
duration/speaking rate (60 of his habitual) for
three consecutive sessions. It took seven weeks
to meet exit criterion during the first treatment
phase. The withdrawal phase and second treatment
phase also continued for seven weeks in order to
equalize the effects of extraneous variables
(McReynolds and Kearns, 1983). During the second
treatment phase the participant achieved
criterion by the 4th session and maintained an
average of 93 accuracy after criterion was
achieved.
17
(No Transcript)
18
Duration Results of Trained Sentences Results
indicate the participant reached a stable average
sentence duration/speaking rate of 2844 ms during
baseline measurements (final 3 data collection
points) of the trained sentences (Refer to Figure
2). Sentence duration was variable during the
first treatment phase for trained sentences. It
ranged from 2698-3967 ms The participant was not
successful in slowing sentence production to his
goal of 60 of his habitual speaking rate. The
participant, however, was able to slow speaking
rate to a maximum of 70 of his habitual speaking
rate. Results indicate the participant was able
to slow his habitual speaking rate (increase
sentence duration) in 6 out of seven sessions
during the first treatment phase.
19
(No Transcript)
20
Withdrawal data indicates the participant
produced a stable sentence duration ranging from
2962-3117 ms compared to range of 2698-3967 ms
for the first treatment phase, indicating an
increase in speaking rate. The data collected
during the second treatment phase was also
variable. The client was able to alter speaking
rate however, he was not successful in achieving
his goal of 60 of habitual speaking rate.
Average sentence duration in the second treatment
phase ranged from 2967-3975 ms indicating an
increase in sentence duration relative to the
withdrawal phase. Results indicate the
participant was able to slow his speaking rate
relative to his average habitual speaking rate
for the baseline phase in all 7 of the data
collection sessions during the second treatment
phase.
21
Duration Results of Untrained Sentences Duration
results of the untrained sentences were similar
to the results of the trained sentences (Refer to
Figure 3). The participants average habitual
speaking rate for the untrained sentences was
determined to be 2620 ms during baseline
measurements. The data collected during the first
treatment phase indicated that the participant
was able to slow his speech, but unable to reach
a slowed rate of 60 of his habitual speaking
rate. Sentence duration in the first treatment
phase was variable and ranged from 2754-3829 ms.
The participant was only able to reach a maximum
slowed speaking of 70 of his habitual speaking
rate relative to average habitual speaking rate
for the baseline phase. The participant slowed
his speaking rate (increased sentence duration)
in all 7 data collection sessions in the first
treatment phase.
22
(No Transcript)
23
Withdrawal data indicates the participant
produced a stable sentence duration range of
2908-3136 ms compared to range of 2754-3829 ms
for the first treatment phase, indicating an
increase in speaking rate. The data collected
during the second treatment phase was also
variable. The client was able to alter speaking
rate however, he was not successful in achieving
his goal of 60 of habitual speaking rate.
Average sentence duration in the second treatment
phase ranged from 2698-3843 ms, indicating an
increase in sentence duration relative to the
withdrawal phase. The participant was able to
slow his speaking rate relative to average
habitual speaking rate for the baseline phase,
and increase sentence duration in all 7 of the
data collection sessions in the second treatment
phase.
24
CONCLUSIONS
The participant was able to alter his speaking
rate and produce the trained sentences at 60 of
his habitual speaking rate with 80 accuracy
during therapy, but was unable to transfer the
new speaking rate to data collection outside of
the therapy room. That is, the participants
performance during data collection did not
reflect performance during treatment. However, on
average the participant did slow his habitual
speaking rate for both treatment phases relative
to the baseline and withdrawal phase. Speakers
with AOS do not consistently adjust speaking rate
for sentences when instructed to speak at given
rates (Robin, Bean and Folkins, 1989). This was
true for the present speaker and was the reason
for expanding the target range.
25
Perhaps the client would have produced a slower
speaking rate if he was told specifically during
data collection to reduce his speech to exactly
what he had practiced during the treatment
session. This was attempted after the treatment
program had concluded and verbal instruction by
itself was found to be unsuccessful. Another
possibility for lack of generalization of the new
speaking rate outside of he therapy room may be
due to the participants deficits in the area of
memory as indicated in the results of the Ross
Information Processing Assessment (See Table
1).
26
The participants ability to generalize was
observed to be dependent on the time of data
collection within a session (i.e., beginning or
end of session). If data was collected at the
beginning of the therapy session, speaking rate
was faster than if it was collected at the end of
the therapy session. These findings support the
influence of the participants memory deficits on
his ability to generalize. A final reason for
the lack of generalization may be due to the fact
that the generalization task involved removal of
all cues (i.e., visual, auditory and verbal).
Articulatory methods for treatment of AOS focus
on gradual removal of cueing (Yorkston et al.
1999). A more programmatic decrease in cueing may
have lead to better generalization of rate.
27
Future Research Replication of this study should
be completed with different participants
exhibiting motor speech disorders. The
evaluation of the effectiveness of different
feedback modes (i.e., visual, verbal, and
auditory feedback) on acquisition and
generalization of a slowed speaking rate should
be documented. Investigation of the interaction
between different modes of feedback in the
training of a slowed speaking rate along with
participant characteristics should be addressed.
28
Selected References Ansel, B. M., Kent, R. D.
(1992). Acoustic-Phonetic Contrasts and
intelligibility in the dysarthria associated with
mixed cerebral palsy. Journal of Speech and
Hearing Research, 35, 296-308.   Chapey, R., Ed.
(1994). Language intervention strategies in adult
aphasia (3rd ed.). Brooklyn, NY Williams
Wilkins.   Dworkin, J. P. (1991). Motor speech
disorders a treatment guide. St. Louis, MO
Mosby. Duffy, J. R. (1995). Motor speech
disorders substrates, differential diagnosis,
and management. St. Louis, MO Elsevier.    
29
Kent, R. D. (1992). Intelligibility in speech
disorders Theory, measurement, and management.
Philadelphia, PA John Benjamins
Publishing.   McNeil, M. R. (1997) Clinical
management of sensorimotor speech disorders. New
York, NY Theime.   Simmons, N. N. (1978). Finger
counting as an intersystemic reorganizer in
apraxia of speech. Clinical Aphasiology
Conference Proceedings,174-179. Southwood, H.
(1987). The use of prolonged speech in the
treatment of apaxia of speech. Clinical
Aphasiology, 15, 277-287.
30
 Wambaugh, J. L. Martinez, A. L. (2000).
Effects of rate and rhythm control treatment on
consonant production accuracy in apraxia of
speech. Aphasiology, 14,851-871.   Wambaugh, J.
L. Nessler, C. (2004). Modification of sound
production treatment for apraxia of speech
acquisition and generalization effects.
Aphasiology, 18,407-427.   Yorkston, K. M.,
Beukelman, D. R. (1981b). Ataxic dysarthria
Treatment sequences based on intelligibility and
prosodic considerations. Journal of Speech and
Hearing Disorders, 46, 398-404.   Yorkston, K.
M., Beukelman, D. R., Strand, E. A., Bell, K.
R. (1999). Management of motor speech disorders
in children and adults (2nd ed). Austin, TX
ProEd.