PreprintPDF Available

Bang for your buck: A single-case experimental design study of practice amount and distribution in treatment for childhood apraxia of speech

Authors:
Preprints and early-stage research may not have been peer reviewed yet.

Abstract and Figures

This single-case experimental design treatment study systematically and orthogonally examined the effects of practice amount and practice distribution (intensity) in six children with childhood apraxia of speech, in the context of an integral stimulation-based treatment approach called ASSIST (Apraxia of Speech Systematic Integral Stimulation Treatment). Findings indicated that overall, (1) more practice resulted in greater learning than less practice, and (2) more massed (more intensive) practice resulted in greater learning than more distributed (less intensive) practice.
Content may be subject to copyright.
For Peer Review
Bang for your buck: A single-case experimental design
study of practice amount and distribution in treatment for
childhood apraxia of speech
Journal:
Journal of Speech, Language, and Hearing Research
Manuscript ID
JSLHR-S-18-0212.R2
Manuscript Type:
Research Article
Date Submitted by the
Author:
11-Mar-2019
Complete List of Authors:
Maas, Edwin; Temple University, Communication Sciences and Disorders
Gildersleeve-Neumann, Christina; Portland State University, Speech and
Hearing Sciences
Jakielski, Kathy; Augustana College, Communication Sciences and
Disorders
Kovacs, Nicolette; Temple University, Communication Sciences and
Disorders
Stoeckel, Ruth; Mayo Clinic, Speech-Language Pathology
Vradelis, Helen; Temple University, Communication Sciences and
Disorders; University of Wisconsin Madison Graduate School,
Communication Sciences and Disorders
Welsh, Mackenzie; Temple University, Communication Sciences and
Disorders
Keywords:
Apraxia of speech, Children, Efficacy, Intervention, Speech production
Journal of Speech, Language, and Hearing Research
For Peer Review
RUNNING HEAD: Bang for your Buck: CAS Tx Amount Distribution
1
Bang for your buck: A single-case experimental design study of practice amount
and distribution in treatment for childhood apraxia of speech
Edwin Maas1,*
Christina Gildersleeve-Neumann2
Kathy Jakielski3
Nicolette Kovacs1
Ruth Stoeckel4
Helen Vradelis1,5
Mackenzie Welsh1
1 Department of Communication Sciences and Disorders, Temple University, Philadelphia, PA
2 Department of Speech and Hearing Sciences, Portland State University, Portland, OR
3 Department of Communication Sciences and Disorders, Augustana College, Rock Island, IL
4 Department of Speech-Language Pathology, Mayo Clinic, Rochester, MN
5 Currently affiliated with the Department of Communication Sciences and Disorders, University
of Wisconson, Madison, WI
Funding
This work was supported by a generous grant from the Childhood Apraxia of Speech Association
of North America (CASANA; currently known as Apraxia Kids) (PI: Maas). The content is
solely the responsibility of the authors and does not necessarily represent the official views of
Apraxia Kids.
Conflict of Interest
Edwin Maas, Christina Gildersleeve-Neumann, and Ruth Stoeckel serve on the Professional
Advisory Board of Apraxia Kids. There are no other conflicts.
* Contact Author:
Edwin Maas
Department of Communication Sciences and Disorders
Temple University
1701 N. 13th St.
Philadelphia, PA 19122
phone: (215) 204-1148
fax: (215) 204-5954
e-mail: emaas@temple.edu
Key Words: Apraxia of Speech; Principles of Motor Learning; Treatment; Practice Amount;
Practice Distribution; Treatment Intensity
Page 1 of 67 Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
RUNNING HEAD: Bang for your Buck: CAS Tx Amount Distribution
2
ABSTRACT
Purpose: To examine two aspects of treatment intensity in treatment for childhood apraxia of
speech (CAS): practice amount and practice distribution.
Method: Using an alternating treatments single-subject design with multiple baselines, we
compared high versus low amount of practice, and massed versus distributed practice, in six
children with CAS. Conditions were manipulated in the context of integral stimulation treatment.
Changes in perceptual accuracy, scored by blinded analysts, were quantified with effect sizes.
Results: Four children showed an advantage for high amount of practice, one showed an opposite
effect, and one showed no condition difference. For distribution, four children showed a clear
advantage for massed over distributed practice post treatment; one showed an opposite pattern
and one showed no clear difference. Follow-up revealed a similar pattern. All children
demonstrated treatment effects (larger gains for treated than untreated items).
Conclusions: High practice amount and massed practice were associated with more robust
speech motor learning in most children with CAS, compared to low amount and distributed
practice, respectively. Variation in effects across children warrants further research to determine
factors that predict optimal treatment conditions. Finally, this study adds to the evidence base
supporting the efficacy of integral stimulation treatment for CAS.
Page 2 of 67Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
RUNNING HEAD: Bang for your Buck: CAS Tx Amount Distribution
3
Parents of children with childhood apraxia of speech (CAS), and speech-language
pathologists (SLPs) who work with these children, often ask how much treatment is needed for
meaningful improvement, and how best to distribute treatment. Children with CAS often show
little or slow progress in standard therapy (ASHA, 2007; Campbell, 1999; Hall, 2000; Shriberg
et al., 1997), leading to recommendations for intensive intervention to optimize outcomes
(ASHA, 2007; Hall, 2000; Strand & Skinder, 1999). Given that resources (money, personnel
time) are limited, it is important to maximize use and impact of these limited resources. The
current study continues and extends a systematic research program that examines effects of
various practice conditions in treatment for CAS to optimize outcomes (Edeal & Gildersleeve-
Neumann, 2011; Maas & Farinella, 2012; Maas, Butalla, & Farinella, 2012). In particular, we
examine two critical aspects of intensity, namely amount and distribution of practice, in six
children with CAS, in the context of an integral stimulation-based intervention.
Our approach to optimizing treatment efficacy for children with CAS is guided by the
motor learning literature, which offers insight into factors that facilitate learning (retention and
transfer) of motor skills (Schmidt & Lee, 2005). This literature indicates that some practice
conditions enhance learning relative to others, effects sometimes referred to as principles of
motor learning (e.g., Strand, Stoeckel, & Baas, 2006; see Maas et al., 2008, for review).
Incorporation of principles of motor learning in treatment for CAS has been recommended (e.g.,
Hall, 2000; Robin, Maas, Sandberg, & Schmidt, 2007) because practice conditions exert their
effect primarily on motor planning and programming (e.g., Wright et al., 2004), the presumed
locus of deficit in CAS (ASHA, 2007; Maas & Mailend, 2017; Nijland et al., 2002; Strand et al.,
2006). The effects of these practice conditions are thought to arise at the level of motor planning
and programming because reaction time evidence suggests that (for example) random practice,
Page 3 of 67 Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
RUNNING HEAD: Bang for your Buck: CAS Tx Amount Distribution
4
but not blocked practice, of multi-component movement sequences results in recoding and
consolidation of the sequence into a single motor plan (Wright et al., 2004). The ability to form
stable motor plans for multi-part sequences (such as multisyllabic words, phrases) is important
for developing accurate, fluent, and efficient speech, and therefore, conditions that facilitate such
recoding would seem particularly relevant for children with CAS. Several treatment approaches
incorporate principles of motor learning (e.g., Strand et al., 2006; Murray, McCabe, & Ballard,
2015; Skelton & Hagopian, 2014). However, the evidence for such recommendations remains
limited because few studies have directly compared conditions in CAS treatment (Edeal &
Gildersleeve-Neumann, 2011; Maas & Farinella, 2012; Maas et al., 2012; Namasivayam et al.,
2015; Preston, Leece, McNamara, & Maas, 2017), and although findings from the motor learning
literature can provide “best guesses” about optimal conditions, ultimately these principles of
motor learning must be studied directly in the populations of interest (e.g., CAS) because
principles of motor learning are largely based on studies of adults with intact motor systems
learning nonspeech motor tasks (see Maas et al., 2008, for further discussion).
The present study focused on two critical aspects of treatment intensity (amount and
distribution), because intensity is frequently cited as a critical treatment variable (Edeal &
Gildersleeve-Neumann, 2011; Hall, 2000; Maas, Gildersleeve-Neumann, Jakielski, & Stoeckel,
2014; Warren, Fey, & Yoder, 2007). In fact, intensive intervention programs for CAS have
emerged in various places across the US in recent years, and although several studies have
examined outcomes of such intensive programs (Murray et al., 2015; Namasivayam et al., 2015;
Preston, Leece, & Maas, 2016; Strand et al., 2006), intensity has received little systematic study
in treatment for CAS. Only Edeal and Gildersleeve-Neumann (2011) and Namasivayam et al.
(2015) have directly compared high and low amounts of practice, and no studies have examined
Page 4 of 67Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
RUNNING HEAD: Bang for your Buck: CAS Tx Amount Distribution
5
intensity while controlling for practice amount. Thus, the primary purpose of the present study
was to provide direct, controlled evidence regarding the role of amount and distribution of
practice in treatment for CAS. A secondary goal was to further examine the efficacy of integral
stimulation treatment approaches for CAS; these approaches incorporate integral stimulation as a
core element (use of multimodal cues to elicit and support a child’s attempt, e.g., simultaneous
visual and auditory models; see The Present Study and the Methods section). Integral
stimulation-based approaches have among the strongest evidence to date (Maas et al., 2014;
Murray et al., 2014), with several studies from independent research groups reporting treatment
effects, despite differences in protocols (e.g., Edeal & Gildersleeve-Neumann, 2011; Maas &
Farinella, 2012; Strand et al., 2006). To address this secondary goal, we used integral
stimulation-based treatment as the vehicle to study intensity.
PRACTICE AMOUNT
Practice amount refers to the number of practice trials and sessions provided throughout
the treatment period (cumulative intervention intensity; Warren et al., 2007). The motor learning
literature indicates that more practice results in greater retention (Shea, Kohl, & Indermill, 1990),
and similar effects have been reported for typical speakers (Kim, LaPointe, & Stierwalt, 2012).
More practice trials provide more opportunity for the learner to determine the relationship
between motor commands (pattern of muscle contractions), initial starting configuration (current
location of articulator), and movement outcome (whether target was achieved). Motor learning,
including speech motor learning, involves establishing such complex relationships over multiple
attempts at a movement (Guenther et al., 2006; Schmidt & Lee, 2005; Wolpert, Ghahramani, &
Flanagan, 2001). The learner can then use this knowledge to perform this and similar movements
in the future by deriving motor commands for the intended goal given the current position of
Page 5 of 67 Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
RUNNING HEAD: Bang for your Buck: CAS Tx Amount Distribution
6
articulators. This view of motor learning accounts for the fact that people can perform a given
motor task despite variability in context. Thus, a larger number of practice trials results in more
accurate and robust knowledge of the relationships that support performance on future attempts.
In the CAS literature, two studies examined practice amount (Edeal & Gildersleeve-
Neumann, 2011; Namasivayam et al., 2015). In an alternating treatments single-subject design
with two children, Edeal and Gildersleeve-Neumann randomly assigned target sounds to two
conditions: moderate (30-40 trials/session) and high production frequency (100+ trials/session).
Sounds were embedded in functional words with simple syllable shapes (CV, VC, CVC, CVCV)
and practiced using an integral stimulation approach. Both children showed greater retention and
transfer for more frequently practiced sounds on probes at the end of treatment sessions. While
these results are encouraging, the study included only two children who received different
practice amounts (cumulative intervention intensity) and different practice distributions. Further,
probe scores only took into account accuracy of the target sounds rather than the entire word.
Namasivayam et al. (2015) used a pre/post design to compare two groups of children
with CAS receiving a motor-based speech intervention (that also included integral stimulation
methods) with different practice amounts. The More-Practice group (N=21) received 20 sessions
over 10 weeks, the Less-Practice group (N=12) received 10 sessions over 10 weeks. SLPs
(N=46) were assigned to one of these conditions using stratified random assignment to account
for clinician skill level. Namasivayam et al. compared performance before and after practice in
each group separately for speech sound accuracy, intelligibility, and functional communication.
Significant changes for speech sound accuracy and functional communication were found only
for the More-Practice group, although both groups showed comparable effect sizes, and no direct
group comparisons were performed. Sentence intelligibility did not improve for either group.
Page 6 of 67Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
RUNNING HEAD: Bang for your Buck: CAS Tx Amount Distribution
7
Thus, available evidence suggests that more practice results in greater learning, including
for children with CAS. However, to date this evidence is limited to two published studies, one of
which (Namasivayam et al., 2015) confounded practice amount and distribution: the More-
Practice group received twice as many sessions per week as the Less-Practice group (i.e. sessions
were more closely spaced). The present study was designed to conduct a within-study
comparison of different practice amounts while controlling for practice distribution.
PRACTICE DISTRIBUTION
While clinicians and researchers generally agree that more practice leads to better
outcomes, it is not clear how to distribute practice across time to maximize learning. Studies
using similar treatment approaches report different magnitudes of improvement (e.g., Maas &
Farinella, 2012; Strand et al., 2006), which may be due in part to different practice distributions
(Edeal & Gildersleeve-Neumann, 2011; Maas et al., 2012, 2014). Practice distribution (dose
frequency; Warren et al., 2007) refers to how a given practice amount is divided over time, and
typically involves a distinction between massed and distributed practice (Baddeley & Longman,
1978; Shea et al., 2000). In massed practice, many trials occur in a short period of time, whereas
in distributed practice the same number of trials are divided over a longer time period.
The motor learning literature shows greater retention for distributed practice than for
massed practice (Baddeley & Longman, 1978; Shea et al., 2000; see Lee & Genovese, 1988, for
meta-analysis). This effect is not specific to motor learning but is also observed in other learning
domains, and is sometimes referred to as the ‘spacing effect’ (e.g., verbal learning: Bahrick &
Hall, 2005; Bjork & Allen, 1970). With respect to motor learning, in a seminal study, Baddeley
and Longman trained four groups of British postal workers on a keyboarding task for mail
sorting. All participants received 60 hours of training and were randomly assigned to one of four
Page 7 of 67 Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
RUNNING HEAD: Bang for your Buck: CAS Tx Amount Distribution
8
schedules applied to a five-day workweek. All groups received practice five days per week, but
the length and number of sessions was systematically varied as follows: (1) one session of one
hour per day (12 weeks of training: 12 x 5 x 1 = 60 hours), (2) two sessions of one hour per day,
separated by at least two hours (6 weeks: 6 x 5 x 2 = 60 hours), (3) one session of two hours per
day (6 weeks: 6 x 5 x 2 = 60 hours), and (4) two sessions of two hours per day, separated by at
least two hours (3 weeks: 3 x 5 x 4 = 60 hours). Retention was examined at 1, 3, and 9 months
following training, using subgroups matched on performance at the end of training. The two
groups receiving 1-hour sessions (groups 1 and 2) had significantly faster rates of correct key
strokes than groups receiving 2-hour sessions (groups 3 and 4), with the group of one 1-hour
session per day (most distributed practice, group 1) showing the fastest performance. The group
receiving two 2-hour sessions per day (most massed practice, group 4) was not only the slowest
but also made significantly more errors than other groups. Although all groups showed some
decrease in speed of correct keystrokes at follow-up, this decrease stabilized after three months
for all groups except for the most massed practice group, which declined further at nine months.
Various explanations have been proposed for the spacing effect. One hypothesis is that
long-term memory formation depends on strengthening (consolidation) of neural and cognitive
representations (Brashers-Krug, Shadmehr, & Bizzi, 1996; Shea et al., 2000), and that successful
consolidation requires some degree of uninterrupted time (‘downtime’) after practice. Massed
practice can be viewed as a disruption of the consolidation process because the relative amount
of ‘downtime’ is less than when practice trials or sessions are spaced further apart. Another
explanation is that distributed practice results in more detailed representations, because it results
in greater forgetting than massed practice (Bjork & Allen, 1970; Lee & Weeks, 1987). The idea
is that in massed practice, the action plan representation is still available in memory, and thus
Page 8 of 67Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
RUNNING HEAD: Bang for your Buck: CAS Tx Amount Distribution
9
requires less reconstructive processing to retrieve and perform again. By contrast, in distributed
practice the learner has forgotten more of the task and has to engage in more processing to
recreate the action plan. It is this additional practice in reconstructing the action plan that results
in more elaborate action plans and ability to retain and transfer the skill (Lee & Weeks, 1987).
Based on this literature, we might predict a benefit of distributed over massed practice in
treatment for CAS. However, whereas the neuroplasticity literature is in agreement with the
motor learning literature regarding benefits of practice amount, evidence from the neuroplasticity
literature suggests the opposite for practice distribution: massed practice results in greater
learning than distributed practice (Kleim & Jones, 2008). For example, Kleim et al. (2002) found
improved reaching behaviors in rats trained on a regimen of 400 trials/day over 10 days, whereas
Luke, Allred, and Jones (2004) found no improvement in reaching skills in rats trained on a
regimen of 60 trials/day over 20 days. This difference led Luke et al. to suggest that massed
(more intense) practice may be needed to induce changes (Kleim & Jones, 2008). However, this
interpretation rests on a between-study comparison (with a confound between amount and
distribution) and on an animal model. A randomized controlled trial with human learners
examined the effect of treatment intensity on recovery of arm and leg function in stroke patients
with limb paresis (Kwakkel et al., 1999). Improvement was greater with intense training than
with less intense training at 6, 12, and 20 weeks post stroke. However, this study and others (e.g.,
Askim et al., 2010; Størvold et al., in press) also conflated amount and distribution. Overall,
recommendations based on the neuroplasticity literature suggest that massed practice results in
better learning than distributed practice, although amount and distribution are often conflated.
To our knowledge, no studies have examined practice distribution in typical speakers, but
several studies have explored the role of practice distribution in speech intervention research for
Page 9 of 67 Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
RUNNING HEAD: Bang for your Buck: CAS Tx Amount Distribution
10
various clinical populations (Allen, 2013; Spielman et al., 2007; Wambaugh et al., 2013; see
Kaipa & Peterson, 2016, for a systematic review), including for CAS (Namasivayam et al., 2015;
Thomas, McCabe, & Ballard, 2014). In Namasivayam et al. (2015; reviewed above), children
with CAS received treatment twice per week (massed) or once per week (distributed) over 10
weeks. The massed group, but not the distributed group, showed significant improvement post
treatment. Thus, massed practice appeared to benefit learning compared to distributed practice.
However, as noted above, this study confounded distribution with amount of practice (massed
practice also involved more practice), making it impossible to separate the contributions of each
factor to the overall effect. Thomas et al. (2014) examined practice distribution of Rapid Syllable
Transition (ReST) treatment (Ballard et al., 2010; Murray et al., 2015). They compared four
children with CAS who received distributed ReST (2 sessions/week for 6 weeks) to massed
(standard) ReST (4 sessions/week for 3 weeks) published in a prior study (Murray et al., 2015).
Thomas et al. found comparable gains immediately post treatment but continued gains only for
standard ReST, suggesting that massed practice is more beneficial than distributed practice.
Taken together, evidence from various literatures result in two competing hypotheses
about optimal practice distribution for CAS: One hypothesis (speech motor learning in CAS
follows principles of motor learning) predicts greater learning with distributed practice; the other
(speech motor learning in CAS follows principles of neuroplasticity) predicts greater learning
with massed practice. Both these literatures have replicated support for their predictions, and it is
unclear a priori which of these is most likely to apply to speech motor learning in CAS. This
uncertainty poses a challenge for SLPs and their clinical decisions regarding how to optimally
structure treatment for children with CAS. Empirical study is needed to resolve this uncertainty
and adjudicate between two plausible but contradictory hypotheses. Some research has begun to
Page 10 of 67Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
RUNNING HEAD: Bang for your Buck: CAS Tx Amount Distribution
11
address this question. However, the number of studies with children with CAS is small, and
given the noted confounds and reliance on cross-study comparisons, further research is
warranted to determine optimal treatment distribution. The current study was designed to
examine practice distribution in a single study whilst controlling for practice amount.
THE PRESENT STUDY
The present study examined practice amount and distribution in the context of an integral
stimulation treatment (“watch me, listen carefully, say what I say”) that incorporates principles
of motor learning, and tactile cues and reduced speech rate as needed (Strand et al., 2006). We
chose this approach because it is considered “probably effective” (Murray, McCabe, & Ballard,
2014) and the only approach with replicated evidence from independent research groups (Maas
et al., 2014). Although details of implementation have varied across these studies (e.g., Edeal &
Gildersleeve-Neumann, 2011; Maas & Farinella, 2012; Strand et al., 2006), all involved a core
reliance on integral stimulation as support and elicitation technique; tactile cues; gradual fading
of cues; and a focus on whole-word movement accuracy during treatment. If treatments
involving the same components show efficacy despite variations in implementation (e.g., with
respect to frequency of sessions, target selection criteria), the support for those shared
components is strengthened by showing generalizability across variations of implementation.
We examined effects of practice amount and distribution on retention of practiced targets
(words/phrases) using a single-case experimental alternating treatments design with six children
with CAS. Practice amount and distribution were manipulated by creating matched sets of targets
for different conditions. The research questions and hypotheses were as follows:
1. Does practice amount influence speech motor learning in treatment for children with CAS?
Page 11 of 67 Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
RUNNING HEAD: Bang for your Buck: CAS Tx Amount Distribution
12
a. Hypothesis 1 (more-is-better hypothesis) is that more practice offers more opportunities
for learning and predicts greater retention for targets that receive more practice.
2. Does distribution influence speech motor learning in treatment for children with CAS?
a. Hypothesis 2a (motor learning hypothesis) is that speech motor learning in CAS follows
motor learning principles and predicts greater retention for targets in distributed practice.
b. Hypothesis 2b (neuroplasticity hypothesis) is that speech motor learning in CAS follows
neuroplasticity principles and predicts greater retention for targets in massed practice.
METHODS
PARTICIPANTS
Six children with CAS (see Table 1) were recruited via postings on the CASANA website
(apraxia-KIDS.org) and at local outreach events. Children were evaluated by an experienced SLP
with expertise in CAS. Evaluation included a case history and formal and informal measures to
evaluate speech, language, and cognitive skills. Four additional children were evaluated but did
not meet our stringent inclusion criteria for CAS diagnosis (see below). All testing and treatment
took place in a quiet, child-friendly room in the Speech, Language, and Brain Lab at Temple
University. The number of sessions required to determine eligibility varied depending how many
sessions were needed to complete the tasks used to diagnose CAS; sessions typically lasted 60
minutes or less. Parent permission was obtained for all children, and all children provided assent.
All study procedures were approved by the Temple University Institutional Review Board.
Inclusionary criteria were: (1) age between 4 and 12 years old, (2) from homes and
educational settings where the primary language spoken was English, (3) verbal output (50+
words) and communicative intent as determined by the SLP and parent report, (4) presence of a
speech sound disorder (SSD), based on a score below the 10th percentile on the Goldman-Fristoe
Page 12 of 67Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
RUNNING HEAD: Bang for your Buck: CAS Tx Amount Distribution
13
Test of Articulation (GFTA-2; Goldman & Fristoe, 2000), (5) CAS as primary speech diagnosis
(see below), (6) normal hearing based on parent report, and (7) typical nonverbal cognition as
determined by a T-score within 1.5 SD of the mean on nonverbal subtests of the Reynolds
Intellectual Assessment Scales (RIAS; Reynolds & Kamphaus, 2003). Exclusionary criteria
were: (1) diagnosis of disorders that significantly affect communication and/or social interactions
(e.g., autism), as per referral diagnosis, (2) uncorrected vision impairments (per parent report)
that would likely interfere with the ability to take advantage of visual cues, (3) significant
impairments of oral structure (e.g., cleft palate) as judged by the SLP based on an oral motor
examination (Robbins & Klee, 1987), and (4) a primary diagnosis of dysarthria, as judged by the
SLP. In addition, children were administered the Peabody Picture Vocabulary Test (PPVT-4;
Dunn & Dunn, 2007), the Expressive Vocabulary Test (EVT-2; Williams, 2007), and the
Clinical Evaluation of Language Fundamentals (CELF-4; Semel, Wiig, & Secord, 2003). These
measures were used for descriptive purposes only. The Dynamic Evaluation of Motor Speech
Skill (DEMSS; Strand et al., 2013) was administered to obtain further information about motor
speech skill from a dynamic assessment developed specifically for this purpose (Strand et al.,
2013). The speech samples elicited from this test were used for clinical judgment; the total score
on the DEMSS was not used to diagnose CAS. None of the children had a history of recurrent
ear infections, feeding or swallowing problems, learning disability, neurological disease, or
neurodevelopmental condition, except where noted.
Children had to meet stringent diagnostic criteria for CAS to qualify for the study. First,
expert SLP judgment is currently the gold standard for CAS (ASHA, 2007; Murray, McCabe,
Heard, & Ballard, 2015). To strengthen confidence in diagnosis, each child was required to be
judged as having CAS by four independent expert SLPs, including the SLP who administered the
Page 13 of 67 Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
RUNNING HEAD: Bang for your Buck: CAS Tx Amount Distribution
14
assessment and three offsite experts (2nd, 3rd, 5th authors) who independently rated children from
video recordings of the assessment (including DEMSS, GFTA, DEAP, conversational speech,
and any other portion deemed relevant by the rater). Each SLP independently rated each child on
a 3-point scale (0 = no CAS, 1 = possible CAS, 2 = CAS) based on signature perceptual speech
features of CAS (inconsistent vowel and consonant errors on repeated productions, difficulties
achieving and transitioning into articulatory configurations, abnormal prosody). Inclusion
required an average rating > 1 and no rating of 0 from any SLP (averages provided in Table 1).
Second, children were required to receive an Apraxia Score of 1 or 2 on the Maximum
Performance Task (MaxPT) protocol developed and prospectively validated by Thoonen et al.
(1996, 1999) and manualized by Rvachew, Hodge, and Ohberg (2005). In this protocol, children
sustain vowels and fricatives as long as possible and perform diadochokinetic tasks (e.g., “say
/pʌtʌkʌ/ as fast and as long as you can”). Thoonen et al. (1999) showed that the combination of
scores differentiates CAS from other SSDs with 100% sensitivity and 91% specificity.
Third, children had to exhibit inconsistent productions on repeated attempts of the same
word, based on a score >40% on the Word Inconsistency subtest of the Diagnostic Evaluation of
Articulation and Phonology (DEAP; Dodd et al., 2006). In this test, children produce 25 words
three times (with other tasks interposed), and each word is scored as consistent (same all three
times) or inconsistent (two or more different productions).
TABLE 1 about here
CAS001. CAS001 was an 11;3 year old right-handed boy who was adopted from South
Korea at age 1 and lived with his White adoptive parents and sister. He was exposed to Korean
as an infant but the language at home was English. Birth history and family history of speech
disorders were unknown. Developmental milestones were slightly delayed, and at age 3 he was
Page 14 of 67Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
RUNNING HEAD: Bang for your Buck: CAS Tx Amount Distribution
15
diagnosed with severe CAS. His speech errors were characterized by distortion, omission, and
substitution errors and included gliding of liquids and cluster reduction. He showed reduced
articulatory accuracy, vowel distortions and substitutions, atypical prosody, and inconsistent
productions on repeated trials on the DEMSS. He performed below norms on all MaxPT tasks;
he produced repetitions of monosyllables but produced no more than two trisyllable sequences in
a row and made several sequencing errors. Oral mechanism examination revealed reduced range
of motion during nonspeech oral motor sequences and limited tongue-jaw dissociation on tongue
protrusion and elevation despite normal isolated nonspeech oral movements. He had occasional
pitch breaks and difficulty sustaining phonation on repeated /hɑ.hɑ.hɑ/ productions. Loudness
was at times reduced, and resonance was hypernasal, suggesting possible mild dysarthria.
Receptive language scores were well below age expectations, although his scores must be
considered with caution, as he exhibited marked apprehension of making errors, especially when
items required a verbal response, and was often reluctant to respond and rehearsed silently before
responding. Nonverbal cognition was in the normal range, and phonological awareness skills
were below the normal range (Phonological Awareness Composite standard score = 67) based on
the Comprehensive Test of Phonological Processing (CTOPP; Wagner et al., 1999).
CAS002. CAS002 was a 7;11 year old left-handed non-Hispanic White boy who lived
with his parents and older brother. English was the only language spoken at home. His mother
reported a family history of speech disorders (his older brother had speech-language delays but
normalized quickly with treatment). He was diagnosed with CAS at age 3. His speech errors
were primarily omission and substitution errors, and included initial voicing, final devoicing,
cluster reduction, and gliding. He also exhibited an unusual but relatively consistent substitution
of retroflex [l] for /ɹ/. On the DEMSS, he demonstrated notably reduced articulatory accuracy,
Page 15 of 67 Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
RUNNING HEAD: Bang for your Buck: CAS Tx Amount Distribution
16
vowel distortions and substitutions, atypical prosody, and inconsistent productions on repeated
attempts. Although his maximum phonation rate and maximum fricative duration rates were
below age norms (Thoonen et al., 1999), his monosyllable repetition rate was in the normal
range. He exhibited great difficulty with the trisyllable: he did not maintain the sequence,
producing both vowel errors and consonant additions (e.g., [pɑptokɑ] for /pʌtʌkʌ/). The oral
mechanism exam revealed normal range of motion and symmetry for isolated movements but
some reduced range of movement, imprecision, and incoordination when alternating movements.
Vocal quality, resonance, and loudness were within normal limits. His language and nonverbal
cognition were in the normal range, and his phonological awareness was in the low average
range (CTOPP Phonological Awareness Composite standard score = 88).
CAS003. CAS003 was a 5;11 year old non-Hispanic right-handed White boy who lived
with his parents and two brothers (including an identical twin, CAS004 below). English was the
only language spoken at home. His father reported a medical history of premature birth (30.5
weeks) and delayed speech and language development. Other developmental milestones were
achieved within expected time frames. His father also reported a family history of speech
disorders: his father did not start fully talking until age 4, and his identical twin brother also
presented with a speech disorder. CAS003’s speech errors were characterized primarily by
inconsistent substitution and omission errors, including voicing errors, stopping, deaffrication,
gliding, and cluster reduction. Atypical errors were also noted, including substitutions of later-
developing sounds for earlier-developing ones (e.g., [ʧ] for /f/) and clusters for singletons (e.g.,
[sn] for /n/). On the DEMSS, he exhibited moderately decreased articulatory accuracy, vowel
distortions and substitutions, atypical prosody, and inconsistent productions. His maximum
phonation duration was slightly below the normative range, but his monosyllable repetition rate
Page 16 of 67Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
RUNNING HEAD: Bang for your Buck: CAS Tx Amount Distribution
17
was in the normal range. Despite producing each monosyllable repeatedly, he did not produce
the trisyllable correctly, with inconsistent consonant substitutions and sequencing errors (e.g.,
[pakaba] for /pʌtʌkʌ/). Oral mechanism evaluation revealed no difficulty with isolated oral
movements but reduced range of movement, slow rate, and limited tongue-jaw dissociation
during repeated alternating movements. Vocal quality, loudness, and resonance were within
normal limits. CAS003 demonstrated age-appropriate language and nonverbal cognitive skills.
CAS004. CAS004 was a 5;11 year old non-Hispanic right-handed White boy, who lived
with his parents and two brothers (including his identical twin brother, CAS003 above). His
developmental and family history were the same as described for CAS003 above. CAS004
mostly made omission and substitution errors, as well as distortions; errors included initial
voicing, gliding, and cluster reduction. He also occasionally substituted later-developing sounds
for earlier-developing sounds (e.g., [ʃ] for /k/). The DEMSS revealed moderately decreased
articulatory accuracy, vowel distortions and substitutions, atypical prosody, and inconsistent
errors. He performed below the normative range for maximum phonation duration on MaxPT,
but monosyllable repetition rate was in the low average range. He had difficulty maintaining /kʌ/
for more than seven repetitions (reverting to [tʌ] thereafter). Despite producing each of the
syllables repeatedly, he did not produce the trisyllable more than twice in a row, switching to and
alternating with [pʌtʌtʌ]. Inconsistent vowel distortions and substitutions were noted (e.g.,
[pʌtitʌ] for /pʌtʌkʌ/), as were occasional consonant additions (e.g., [pʌtiptʌ]). Oral structures
were intact and symmetrical, with possible malocclusion of dentition. Some lateral movement
was noted on jaw opening but he had no difficulty with other isolated movements of lips or
tongue. However, reduced range of motion and incoordination were observed in sequenced and
alternating movements. Vocal quality, loudness, and resonance were within normal limits. His
Page 17 of 67 Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
RUNNING HEAD: Bang for your Buck: CAS Tx Amount Distribution
18
receptive language skills according to CELF subtests was in the low average range. Due to
increased distractibility, the Concepts and Directions subtest was administered over multiple
sessions; given this nonstandardized administration, the subtest score and the Receptive
Language Index were not computed. However, for the Word Classes subtest, CAS004 obtained a
scaled score of 10 (percentile = 50) and for the Sentence Structure subtest he received a scaled
score of 6 (percentile = 16). Finally, his nonverbal cognitive skills were in the normal range.
CAS005. CAS005 was a 6;0 year old right-handed non-Hispanic White boy, who lived
with his parents and older sister. English was the only language spoken at home, and there was
no family history of SSD. His mother reported a history of “severe” recurrent ear infections with
a slight hearing loss and myringotomy tubes at age 2, but intact hearing at the time of the study.
His medical history revealed supraventricular tachycardia in utero at 33 weeks. He was
diagnosed with CAS at age 2;9. Testing at intake revealed a severe SSD characterized primarily
by omissions, distortions, and atypical substitution errors (e.g., [n] for /ʤ/, [bw] for /l/, [gl] for
/f/). The DEMSS revealed moderately decreased articulatory accuracy, vowel distortions and
substitutions, and inconsistent productions. Although atypical prosody was not observed on the
DEMSS, he did exhibit frequent difficulty with transitioning between sounds and syllables in
spontaneous connected speech. His spontaneous speech was slow and effortful. He scored below
normative values on all MaxPT tasks. He produced repetitions of single syllables, although in
most cases fewer than 10 syllables, and often voiced the consonants. He did not produce the
trisyllable sequence correctly. His slow monosyllable repetition rate and reduced maximum
phonation duration resulted in a Dysarthria Score of 2. On the oral mechanism exam, he
inconsistently exhibited an open mouth rest posture, and frequently his tongue rested on his
lower lip. Drooling and pooling of saliva was noted intermittently during the assessment. He had
Page 18 of 67Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
RUNNING HEAD: Bang for your Buck: CAS Tx Amount Distribution
19
no difficulty with isolated lip movements; however, incoordination and reduced range of motion
were noted on repeated alternating movements (protrude-retract). Lip seal was slightly
inadequate and he did not bite his lower lip or elevate his tongue with either verbal directive or
visual model. He sustained phonation on /ɑ/ but had difficulty producing /hɑ.hɑ.hɑ./. Vocal
quality was slightly breathy, with reduced loudness. Hypernasality was also noted throughout the
assessment. His reduced oral motor skills in the absence of structural deficits suggests that his
SSD includes a contribution of dysarthria, characterized primarily by weakness. Clinical
judgment suggested that the relative contribution of CAS to the overall SSD outweighed that of
dysarthria. CAS005 also exhibited a receptive language disorder. He could not complete the
CELF subtest Concepts & Following Directions despite repeated attempts; thus, the Receptive
Language Index could not be computed. He scored in the low average range for Word Classes
(scaled score = 7; percentile = 16) and well below average for Sentence Structure (scaled score =
1; percentile = 0.1). Nonverbal cognitive skills were age-appropriate.
CAS008. CAS008 was a 4;7 year old right-handed non-Hispanic White/African-
American boy who was an only child living with his parents. English was the primary language
spoken at home, with occasional Spanish exposure from grandmother and at school. His father
reported family history of speech disorder, with the father receiving speech therapy until age 12.
CAS008’s speech errors consisted primarily of substitution and omission errors, with occasional
distortions; errors included final consonant deletion, stopping, final devoicing, deaffrication,
gliding, and cluster reduction. Atypical errors included substituting clusters for singletons (e.g.,
[bl] for /n/; [ts] for /s/) and later-developing sounds for earlier ones (e.g., [v] for /w/). The
DEMSS revealed decreased articulatory accuracy, vowel distortions and substitutions, atypical
prosody, and inconsistent productions. Final schwa additions and schwa insertions were also
Page 19 of 67 Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
RUNNING HEAD: Bang for your Buck: CAS Tx Amount Distribution
20
noted. The MaxPT protocol was discontinued on the first attempt due to difficulty, and
completed in a subsequent session. He exhibited difficulty sustaining phonation and frication,
most likely due to difficulty coordinating respiration with phonation onset. His fricatives were
distorted and unstable, and he did not sustain /z/ for more than a few seconds, after which he
sustained a vowel instead. Although maximum phonation duration was below the normative
range, his monosyllabic repetition rate was in the normal range (albeit with occasional voicing
alternations). He had significant difficulty with the trisyllabic sequence and did not produce the
correct sequence more than twice, and only at a slow rate. His errors were mainly sequencing
errors, with occasional voicing alternations and vowel distortions (e.g., [pʌkʌtʌ], [tʌkʌpʌ],
[pɑgɪbʌ], for /pʌtʌkʌ/). Oral structures were intact and symmetrical, with possible malocclusion
(overbite). He had no difficulty with isolated oral movement tasks, but mis-sequenced repeating
alternating movements. Vocal quality, loudness, and resonance were within normal limits. Mild
hypernasality was occasionally noted during the coordinated speech tasks. His vocabulary was at
or above average, and nonverbal cognitive skills were above average.
DESIGN
Design Overview
To compare conditions, we used an alternating treatments design (Kearns, 1986) with
different treatment target sets for each condition, combined with a withdrawal/maintenance
phase and multiple baselines across behaviors (target sets) and participants, with a minimum of
three baseline sessions. Total study duration (excluding initial assessment) was approximately 24
weeks, including 3-6 weeks of initial baseline, 8 weeks (16 sessions) of treatment divided into
two 4-week phases (8 sessions), each followed by a 2-week withdrawal/maintenance phase, and
Page 20 of 67Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
RUNNING HEAD: Bang for your Buck: CAS Tx Amount Distribution
21
two follow-up points, at 4 and 8 weeks post treatment (see Figure 1). Due to scheduling
conflicts, actual study duration ranged from 23 (CAS003 & CAS004) to 27 weeks (CAS008).
FIGURE 1 about here
Both massed and distributed practice conditions included 10 targets (words/phrases; see
Target Selection). All 10 distributed practice items were trained in both treatment phases (16
sessions), whereas massed practice items were divided into two 5-item subsets, with each subset
practiced in only one treatment phase (8 sessions). Treatment sessions included both conditions
for an equal amount of time (25 minutes/condition). Because massed subsets contained half as
many targets as the distributed set, and given equal time per session, targets in massed subsets
received twice as much practice per phase as items in the distributed set. Therefore, comparison
of massed subset 1 to the distributed set after phase 1 addressed practice amount. Because each
massed subset was practiced only in one phase (versus two phases for the larger Distributed set),
the practice amount for the complete massed set and the distributed set was equal after phase 2.
Therefore, comparison of massed and distributed sets after phase 2 addressed practice
distribution. Thus, distribution was manipulated by varying set size with a given amount of time,
rather than by varying number of sessions per unit time. To clarify, the ten massed practice
targets were practiced over four weeks (eight sessions of ~25 minutes) – five in each phase – and
distributed practice targets – all ten items – over eight weeks (16 sessions of ~25 minutes). By
varying target set size (5 vs. 10) and keeping session duration the same, targets in massed
practice received twice as much practice per session, and per phase, as distributed items. But
because distributed items were targeted in two phases (as opposed to one phase for massed
items), the overall amount of practice for massed items was the same as that for distributed
items. For example, if 20 teaching episodes (see below) can be completed in 25 minutes, then
Page 21 of 67 Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
RUNNING HEAD: Bang for your Buck: CAS Tx Amount Distribution
22
each of the five massed practice targets receives on average four teaching episodes per session.
In contrast, the ten distributed practice items would receive on average two teaching episodes per
session. Thus, to achieve the same amount of practice, the distributed practice items received
practice over 16 sessions (eight weeks) versus eight sessions (four weeks) for massed practice
targets. At the level of the complete set, the overall treatment period was also the same (eight
weeks to achieve the same amount of practice for all items in each set).
Order of conditions within each session was pseudorandomized as follows: the child
rolled a die before the first weekly session to determine which condition would be presented first
in that session; the following session would have the reverse order. Thus, the order of conditions
was counterbalanced by week but randomized across weeks, and each condition was presented
an equal number of times in the first and in the second half of a session (8/16 first, 8/16 second).
Target Selection
Target words/phrases were tailored to each child based on personal relevance to enhance
motivation. Possible targets were generated from a list provided by child and parents based on a
questionnaire (Wilson & Gildersleeve-Neumann, 2014), and were administered several times in
direct imitation prior to baseline in order to narrow and refine the list (e.g., changing names to
preferred nicknames) and to gauge pre-treatment accuracy and stability and rule out items that
were consistently correct or improving during baseline (based on judgment by a research
assistant not blinded to time of data collection).1
1 This analysis was performed in an effort to gauge performance and establish stable baselines prior to initiation of
treatment, and did not constitute our primary outcome. The primary outcome data reported below were based on
blinded analysis after collection of the final follow-up data point, so that data could be analyzed blinded to both
treatment status and data collection time point (e.g., pre vs. post treatment). See Data Analysis section below. We
note that one consequence of this decision to rely on blinded outcome assessors for our primary outcome is that
stable baselines cannot be guaranteed prior to initiation of treatment. Establishing stable baselines before initiating
treatment typically involves nonblinded analysis with respect to pre- vs. post-treatment status (including in the
present study), and thus controls neither for assessor bias nor for perceptual drift. As a result, stability of baselines in
such cases may be overestimated. The blinded analyses are more credible and more rigorous methodologically.
Page 22 of 67Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
RUNNING HEAD: Bang for your Buck: CAS Tx Amount Distribution
23
From each child’s list, three sets of ten words/phrases were created as potential treatment
targets for each child. Effort was made to balance sets for interest area (e.g., friends’ names, food
items), and sets were matched for length (number of syllables), complexity (Index of Phonetic
Complexity; Jakielski, 2002, 2016), and for accuracy on the first two baseline sessions and
average baseline accuracy, as scored by a research assistant (all ps > 0.05, 2-tailed 2-sample t-
tests). To eliminate bias, matching of sets was completed prior to assignment to condition.
Once sets were matched, they were randomly assigned (via random number generator) to
one of three conditions: (1) Distributed Practice, (2) Massed Practice, and (3) Control. Next, the
Massed Practice set was divided into two subsets of five items, matched for length, complexity,
and baseline accuracy to each other and to Distributed and Control sets. After matching, the two
subsets were randomly assigned to phase 1 or 2. Item set information is given in Supplement S1.
PROCEDURES
The study took place in a quiet, child-friendly room at Temple University. A parent and a
research assistant were often present in addition to child and SLP. All sessions were both video
recorded (using a Canon Vixia HF10 recorder) and audio recorded (at 44.1 kHz and 16-bit using
an M-Audio Aries condensor microphone connected to a Marantz CDR420 CD recorder).
Treatment Procedures
Treatment phases involved two weekly 1-hour sessions to approximate realistic logistic
constraints in typical settings. Both conditions were administered in each session; sessions were
divided into two halves (25 min./condition, controlled by egg-timer), separated by a 10-minute
break. Treatment was provided by the same SLP who conducted the assessments.
Our integral stimulation-based approach was derived from, and similar to, Dynamic
Temporal and Tactile Cueing (DTTC; Strand et al. 2006), but differs in target selection, less
Page 23 of 67 Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
RUNNING HEAD: Bang for your Buck: CAS Tx Amount Distribution
24
intensive session frequency, and less dynamic implementation. To facilitate replicability,
experimental control, and consistency, we developed a protocol that systematically controls and
varies treatment parameters (Appendix A). For this reason, we refer to this approach as ASSIST
(Apraxia of Speech Systematic Integral Stimulation Treatment). Like other integral stimulation
treatments, ASSIST incorporates motor learning principles, modeling, and various cues. The
SLP initially facilitated speech through immediate imitation (integral stimulation: “watch me,
listen to me, say what I say”), and worked towards independent productions. Articulation was
shaped through multimodal (tactile, visual, auditory; simultaneous production) cueing techniques
to promote accurate movement gestures. Cues were individualized and varied dynamically
depending on the child’s response and motivation. The SLP gradually faded these cues and
varied the interval between her and the child’s production. The goal was for children to produce
the entire target word or phrase correctly, with correctly articulated segments and accurate and
fluent prosody, in order to provide them with a functional, personally meaningful set of
utterances that might be useful in their daily life.
A critical element of the protocol for present purposes is the teaching episode. A teaching
episode refers to the sequence of events from initial elicitation of a target through the child’s
final attempt before switching to another target. If the child’s initial attempt is incorrect, the SLP
provides feedback and cues (e.g., tactile, simultaneous production at slower rate), and up to five
additional partial or complete attempts by the child before eliciting the target again with the same
elicitation method as at the start of the teaching episode. If the child’s initial attempt is correct,
the SLP requests another production (e.g., “can you say that again?”), to reinforce the speech
motor pattern and to maintain a roughly equal number of production attempts across conditions.
Page 24 of 67Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
RUNNING HEAD: Bang for your Buck: CAS Tx Amount Distribution
25
For each condition, targets were written on two index cards (20 for distributed practice,
10 for massed practice). Before each session, decks were shuffled to implement random practice;
targets appeared on two cards to decrease predictability of upcoming targets. Given that children
produced multiple attempts per teaching episode, this practice schedule is considered randomized
block practice (small blocks, presented in random order). Once all cards had been presented, the
SLP reshuffled the deck and continued. To implement reduced feedback frequency (Maas et al.,
2012), verbal feedback on initial and final attempts in a teaching episode was faded, with
feedback on all of the first ten teaching episodes in a session, 9/10 for the next ten, etc. (Ballard
et al., 2007). Feedback included knowledge of results (accuracy, e.g., “That was good!”) and
knowledge of performance (nature of the response, e.g., “Your lips were not closed.”). The SLP
used her clinical judgment to decide the focus of feedback for each attempt, depending on the
error produced, the child’s response and ability to implement the change, etc.. For instance, if a
child was unable after several attempts to approximate a particular segment, the SLP would
focus on prosodic errors or fluency (e.g., “Let’s focus on keeping the sounds together this time”)
before returning to the difficult segment, in order to minimize frustration and maintain
motivation and attention. Thus, while ASSIST is less dynamic than DTTC, it still incorporates
allowances for clinical judgment in treatment delivery for reasons of ecological validity.
To systematically work towards independence in speech production, four different
elicitation methods were used, in order from more to less support (less to more independent):
immediate imitation at slowed rate, immediate imitation at normal rate, delayed imitation at
normal rate, and independent (e.g., in response to a question). Each session half started with
immediate imitation at normal rate, and the elicitation method changed depending on the child’s
Page 25 of 67 Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
RUNNING HEAD: Bang for your Buck: CAS Tx Amount Distribution
26
performance during the session as judged by the treating SLP. Elicitation method changed after
2/2 consecutive correct or incorrect first attempts of a teaching episode (see Appendix A).
Probe Procedures
The primary data to assess treatment effects were productions elicited on a probe task that
included all 30 items (distributed, massed, control). Items were presented in a different random
order each time by the treating SLP using either delayed reading (CAS001, CAS002) or delayed
imitation (CAS003-CAS008). Delayed production tasks were used to reduce reliance on the
SLP’s auditory and visual model. For delayed reading, children were shown an index card with
the written item, read it silently, and produced the item only after the SLP removed the card (to
minimize effects of reading out loud on prosody). No feedback on accuracy or performance was
provided. Probes were administered weekly, except during follow-up (2-week interval and 4-
week interval). During treatment phases, probes were administered at the beginning of a session.
Treatment Fidelity Procedures
To ensure treatment fidelity (Kaderavek & Justice, 2010), the treating SLP followed an
operationalized, manualized protocol (see Appendix A). An independent analyst also assessed
fidelity for a minimum of two randomly selected sessions from video recordings; this analyst
was not involved in treatment or probe data analysis. Rated fidelity aspects included the numbers
and proportions of teaching episodes with reduced feedback frequency (<90%), feedback delay
(1-3 seconds), feedback on both initial and final attempt, same elicitation method for both initial
and final attempt, the number and proportion of teaching episodes with intended initial elicitation
attempt. It became clear during the fidelity analysis that judgments of speaking rate (slow rate vs.
normal rate) were not reliable, and therefore we combined these elicitation methods for fidelity
Page 26 of 67Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
RUNNING HEAD: Bang for your Buck: CAS Tx Amount Distribution
27
purposes. We also report the number of teaching episodes, number of whole-target production
attempts per teaching episode, and amount of time per condition (see Appendix B).
DATA ANALYSIS
The dependent measure was perceptual accuracy of words/phrases on probes. Accuracy
was judged from audio recordings by independent raters blinded to condition (massed,
distributed, control) and time point (e.g., baseline, follow-up). To enable blinding to time point
and control for perceptual drift, recordings were presented in random order (and thus, analysis
could only begin after collection of the final follow-up probe for each child). Clues about time
point or treatment status were removed from files by a research assistant not involved in analysis.
Analysts judged accuracy on a syllable-by-syllable basis by judging whether each
syllable contained error(s) or not. Errors included substitutions, omissions, distortions, additions,
metatheses, and unintelligible syllables.2 Accuracy was judged against the SLP’s model; casual
speech and dialect variations were not considered errors (even if different from SLP’s model).
For example, if the SLP produced the item How are you today? (CAS003) with a full vowel in
today ([tuˈdeɪ]) but the child produced a reduced vowel ([tǝˈdeɪ]), this was considered correct
because vowel reduction is acceptable variation in this context. Based on these syllable-by-
syllable accuracy judgments, each item was awarded a score of 0, 1, or 2 to represent major
errors (0), minor errors (1), and correct responses (2) (similar to Maas et al., 2012; Strand et al.,
2006), in order to account for differences in length and complexity of items. A score of 0 (major
error) indicated that fewer than 50% of the syllables were correct, a score of 1 (minor error)
indicated that 50% or more (but not 100%) of syllables were correct, and a score of 2 (correct)
2 Initially, prosody was also included in the scores, but these judgments were unreliable and are not reported here.
Page 27 of 67 Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
RUNNING HEAD: Bang for your Buck: CAS Tx Amount Distribution
28
indicated that all syllables were error-free.3 For each probe, item scores were averaged by set and
divided by 2 to provide the condition score in percent accuracy. Average inter-rater reliability
based on 19 randomly selected probes (16% of probes) indicated exact agreement on syllable-by-
syllable judgments for 79% of syllables scored by both raters, and exact agreements for 68% of
item scores (with 96% within 1 scale point); intra-rater reliability based on one randomly
selected probe indicated exact agreement on syllable-by-syllable judgments for 91% of syllables
double-scored, and exact agreement for 83% of item scores (100% within 1 scale point).
Analysis involved visual inspection of graphs, supplemented with standardized effect
sizes (d statistic; Beeson & Robey, 2006; Busk & Serlin, 1992) to compare and quantify changes
relative to baseline in each condition for each child. Effect size d was computed as [ (mean
scorepost – mean scorepre) / SDpre ]. We operationally define d > 1 as improvement (pre-post gains
exceeds baseline standard deviation) (Maas et al., 2012; McAllister Byun et al., 2014). Note that
a larger standard deviation during baseline (more variable, less stable baseline) has the effect of
reducing the ability to detect an effect, as it increases the denominator. Thus, increased
variability during baseline sets a higher bar for obtaining an effect. For completeness, we also
report unstandardized gains (percent change). We expected larger effect sizes for treated than for
untreated control items after each treatment phase. In addition to individual analyses, we also
compared conditions across the six children using nonparametric Cochran-Mantel-Haenszel chi-
square 2) tests to detect overall condition differences and nonparametric Wilcoxon signed rank
tests to identify significant pairwise differences = 0.05).
To address the role of practice amount, we compared effect sizes for massed practice
subset 1 to the distributed practice set during maintenance phase 1, averaged across both probes
3 Although this scoring system reduces sensitivity for monosyllabic targets (which can receive only segmental scores
of 0 or 2), this is unlikely to impact the findings because there were very few monosyllabic targets for any child.
Page 28 of 67Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
RUNNING HEAD: Bang for your Buck: CAS Tx Amount Distribution
29
in this phase. Hypothesis 1 (more-is-better hypothesis) predicts a larger effect size for massed
practice subset 1 than for the distributed practice set, because after phase 1, massed practice
subset items had received twice as much practice as distributed practice items.
To address the role of practice distribution, we compared effect sizes for massed practice
items (subsets 1+2 combined) and the distributed practice set in maintenance phase 2 and follow-
up. Hypothesis 2a (motor learning hypothesis) predicts larger effect sizes for the distributed than
for the massed practice set; Hypothesis 2b (neuroplasticity hypothesis) predicts the reverse.
RESULTS
INDIVIDUAL ANALYSES
Summary data for all children and group means are presented in Tables 2 and 3. Data for
children CAS001-CAS003 are plotted in Figure 2, and data for CAS004-CAS008 in Figure 3.
FIGURES 2 and 3 about here
CAS001. CAS001 demonstrated somewhat variable but nonrising baselines prior to
treatment phase 1. On initiation of treatment, Massed subset 1 showed an increase in accuracy.
Distributed items showed a more gradual improvement toward the end of the phase, but this
improvement was comparable to the Control set, and performance for both these sets declined
when treatment was withdrawn. As-yet-untreated Massed subset 2 did not change until treatment
phase 2, although a slight increase was noted on the probe immediately prior to treatment. The
Distributed set again showed a modest increase during phase 2.
Effect sizes for phase 1 indicated treatment effects for both treated conditions, with an
advantage for the Low amount set (d = 2.31, 20.0% gain) compared to the High amount set (d =
1.09, 16.7% gain). Control items did not reveal a significant effect (d = 0.57, 10.0% gain).
Page 29 of 67 Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
RUNNING HEAD: Bang for your Buck: CAS Tx Amount Distribution
30
At maintenance phase 2, CAS001 showed a minimal advantage for massed over
distributed practice in terms of effect size but not in terms of percent gain (Massed: d = 2.74,
15.8% gain; Distributed: d = 2.31, 20.0% gain). At follow-up, both effect size and percent gain
revealed an advantage for distributed practice (d = 3.46, 30.0% gain) compared to massed
practice (d = 2.74, 15.8% gain). Untreated items did not improve either at maintenance (d = 0.24,
4.2% gain) or at follow-up (d = 0.52, 9.2% gain).
CAS002. CAS002 showed stable or slightly declining baselines prior to phase 1
treatment. High amount items (Massed 1) improved during treatment phase 1. During
maintenance phase 1, all sets declined except for the High amount set. During treatment phase 2,
the Distributed set showed improvement. At follow-up, accuracy of the massed set was slightly
higher than accuracy for distributed items.
Both effect sizes and percent gains after phase 1 indicated a benefit for High amount
items (d = 1.83, 17.5% gain) compared to the Low amount items and the Control items, both of
which showed a decline (Low amount: d = -1.79, 11.3% loss; Control: d = -2.98, 18.8% loss).
Thus, only the High amount set showed a gain.
In the immediate two-week post-treatment phase, CAS002 showed a clear advantage for
Massed items (d = 2.53, 11.3% gain), with no effects for Distributed (d = 0.60, 3.8% gain) or
Control items (d = 0.60, 3.7% gain). This advantage was maintained during the follow-up phase,
at which point the Distributed items approached an effect (d = 0.99, 6.3% gain) while the Control
items declined (d = -0.20, 1.3% loss). Thus, only Massed items showed a robust treatment effect.
CAS003. CAS003 showed largely stable baselines before treatment, with a slight increase
for untreated items. At the end of treatment phase 1, both treated sets showed improvement.
During treatment phase 2, a slight decline was noted for the massed sets. After both treatment
Page 30 of 67Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
RUNNING HEAD: Bang for your Buck: CAS Tx Amount Distribution
31
phases, performance was generally maintained at levels comparable to those after treatment
phase 1, and above baseline for Massed subset 1 and Distributed items.
After phase 1, CAS003 showed a larger effect size for High amount items (d = 2.35,
22.5% gain) than for Low amount (d = 1.13, 5.4% gain) and Control items (d = 1.12, 9.6% gain).
During maintenance phase 2, CAS003 showed an advantage for Massed (d = 2.58, 16.3%
gain) over Distributed items (d = 1.83, 8.8% gain); both outperformed Control items (d = 0.73,
6.3% gain). During follow-up, a small advantage for Distributed over Massed items emerged for
effect size (d = 2.35 vs. d = 2.19) but not for percent gain (13.8% vs. 11.3%). Although control
items also improved at follow-up, gains were smaller than for treated items (d= 1.02, 8.8% gain).
CAS004. CAS004 demonstrated stable baselines before initiation of treatment. Following
treatment phase 1, Massed 1 items showed a small improvement that dissipated prior to
treatment phase 2. After phase 2, Distributed items and Massed 2 items showed a small increase.
After phase 1, CAS004 improved only on High amount items (d = 2.83, 14.2% gain),
with slight declines for Low amount (d = -0.41, -1.7%) and Control items (d = -0.60, -3.8%).
CAS004 showed a robust advantage for Massed practice items compared to Distributed
items, both at maintenance 2 (Distributed: d = 3.50, 17.5% gain; Massed: d = 1.84, 7.5% gain)
and during follow-up (Distributed: d = 3.50, 17.5% gain; Massed: d = 0.61, 2.5% gain).
Untreated items did not show improvement at either point (d = 0.20, 1.3% gain at both points).
CAS005. CAS005 demonstrated a rise around the third baseline probe for Control items
and Massed subset 1 items, which subsequently stabilized before treatment onset. Massed subset
2 did show a rising baseline during this initial baseline, which largely stabilized prior to its being
treated. No clear response to treatment was evident during treatment phase 1, but treated items
did show higher accuracy immediately following phase 1. Distributed items remained stable
Page 31 of 67 Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
RUNNING HEAD: Bang for your Buck: CAS Tx Amount Distribution
32
through treatment phase 2, and Massed subset 2 showed an increase toward the end of treatment
phase 2, whereas Massed 1 and Control items declined during phase 2 treatment.
Effect sizes for practice amount revealed greater gains for High amount items (d = 3.18,
28.6% gain) than for Low amount (d = 2.47, 13.2%) or Control items (d = 1.39, 15.0% gain).
CAS005 showed a consistent advantage for Massed compared to Distributed practice
both in the immediate post-treatment phase (Massed: d = 2.32, 18.6% gain; Distributed: d = 1.07,
5.7% gain) and during the follow-up phase (Massed: d = 2.63, 21.1% gain; Distributed: d = 2.00,
10.7% gain). Both treated conditions also consistently outperformed untreated Control items
(maintenance 2: d = -0.23, 2.5% loss; follow-up: d = 0.69, 7.5% gain).
CAS008. CAS008 demonstrated somewhat variable baselines, with indications of rising
trends immediately prior to treatment for both massed subsets; massed subset 2 stabilized prior to
phase 2 treatment. Following treatment phase 1, improvement was noted only for the Distributed
items, and to some extent for the Control items (which returned to baseline levels after the first
maintenance 1 probe). In treatment phase 2, both massed subsets improved while Distributed
items remained stable. In the maintenance 2 phase, Distributed items showed some additional
gains, which largely dissipated during follow-up, unlike massed items which retained gains.
Effect sizes for practice amount indicated no treatment effects; unexpectedly, only
Control items showed an effect (d = 1.38, 9.0% gain). Although percent gain for Low amount
items was of similar magnitude (9.4%), greater baseline variability resulted in a nonsignificant
effect size (d = 0.89); High amount items showed no improvement (d = 0.21, 1.8% gain).
Regarding distribution, CAS008 showed inconsistent effects: At maintenance 2, he
showed an advantage for Distributed (d = 2.06, 21.9% gain) over Massed practice (d = 0.48,
4.8% gain), whereas at follow-up, this pattern was reversed (Distributed d: 0.89, 9.4% gain;
Page 32 of 67Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
RUNNING HEAD: Bang for your Buck: CAS Tx Amount Distribution
33
Massed d = 1.47, 14.8% gain). Control items did not show clear improvement (maintenance 2: d
= 0.61, 4.0% gain; follow-up: d = 1.00, 6.5% gain).
TABLES 2 and 3 about here
GROUP ANALYSES
Group means by condition are provided in Tables 2 and 3. At maintenance phase 1, there
were no significant effects for practice amount for either effect sizes nor for percent gains,
although there was a trend (p < 0.10) in the comparison between High amount and Control sets
(S = 8.5, p = 0.094, for both d and percent gain; all other ps > 0.10, see Appendix C Table C).
For practice distribution, Cochran-Mantel-Haenszel tests indicated significant condition
differences for both effect size and percent change at both M2 and FU (see Appendix C, Table
D). Pairwise comparisons revealed no significant group differences between Massed and
Distributed conditions (ps > 0.20). At M2, both treated sets showed a significant advantage
compared to control items for percent change (S = 10.5, p = 0.031 for both), and a trend for
effect size (Distributed vs. Control: S = 7.5, p = 0.063; Massed vs. Control: S = 9.5, p = 0.063).
At follow-up, both Massed and Distributed sets showed significantly greater gains for percent
change (S = 10.5, p = 0.031 for both) but only Massed items differed significantly from the
control items in terms of effect size (S = 10.5, p = 0.031; Distributed: S = 9.5, p = 0.063).
DISCUSSION
The present study was designed to examine two critical aspects of treatment intensity,
amount and distribution, in treatment for CAS. This study also examined the overall impact of
integral stimulation-based treatment for CAS. Below we discuss findings relative to these goals.
PRACTICE AMOUNT
Page 33 of 67 Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
RUNNING HEAD: Bang for your Buck: CAS Tx Amount Distribution
34
At the individual level, four of the six children showed an advantage of more practice
(CAS002, CAS003, CAS004, CAS005). One child (CAS001) showed an unexpected advantage
for the low amount condition, and one child (CAS008) did not show effects for either treated set
but did improve on control items. These unexpected effects most likely reflect item-specific
effects. Despite our best efforts at matching item sets for difficulty and personal interest, random
assignment to conditions may have led to low amount (CAS001) and control sets (CAS008)
containing more items that were more motivating or used more frequently outside treatment.
Averaged across all six children, there was a numerical advantage for more practice
compared to less practice after treatment phase 1, both in terms of effect sizes and percent gain
(see Table 2). Item sets receiving a low amount of practice (the Distributed set) did not differ
notably from untreated sets, and neither of these sets showed an effect on average (average d <
1), unlike High amount items (average d = 1.45). The numerical advantage for High amount was
not significant in the nonparametric group analyses however, possibly due to the small sample.
Taken together, the individual data and the group data are suggestive of an advantage for
more practice, with 4/6 children showing greater gains and effect sizes than items receiving less
practice, and an average effect size across children that exceeded baseline standard deviations.
These findings are consistent with both the motor learning literature (e.g., Shea et al., 1990) and
the CAS treatment literature (Edeal & Gildersleeve-Neumann, 2011; Namasivayam et al., 2015),
and add support to the notion that children with CAS benefit from more practice. However, this
interpretation must be taken with caution given that the differences were generally modest and
the group analysis failed to detect significant condition differences. One factor that may have
contributed to these modest effects is that the overall amount of practice during phase 1 was
small, with only eight 25-minute sessions per condition. This amount of practice may not have
Page 34 of 67Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
RUNNING HEAD: Bang for your Buck: CAS Tx Amount Distribution
35
been sufficient for condition differences to emerge. Some support for this view comes from the
observations that more robust treatment effects were evident after the second treatment phase,
and that only two children showed positive effects and a clear advantage over the control items
for the low amount condition, (CAS001 and CAS005), compared to five children for the high
amount condition (CAS001, CAS002, CAS003, CAS004, CAS005). Thus, it appears that despite
the relatively few and short sessions, reducing the set size (thus increasing the number of practice
opportunities per item) resulted in speech motor learning. Clearly, further research is needed
with larger sample sizes and greater differences between conditions. Nevertheless, our findings
suggest that more practice is likely to result in greater gains, consistent with Hypothesis 1 and
with previous literature (Edeal & Gildersleeve-Neumann, 2011; Namasivayam et al., 2015).
PRACTICE DISTRIBUTION
For practice distribution, five of the six children showed a numerical benefit in terms of
effect size for massed practice over distributed practice during the two-week maintenance period
immediately following treatment phase 2, whereas only one (CAS008) showed an advantage for
distributed practice (see Table 3). For percent change, the pattern was the same except that
CAS001 showed a slight advantage for distributed rather than massed practice. At follow-up,
four of the six children showed an advantage for massed practice compared to distributed
practice for effect size (CAS002, CAS004, CAS005, CAS008) and two the opposite (CAS003,
CAS008); participant CAS003 did show a numerical massed practice advantage for percent
change. Such ambiguous effects (inconsistent between d and %change) are difficult to interpret
and always involved small differences (d differences < 0.5, %change differences < 5%). To our
knowledge, there are no established criteria for determining condition differences in alternating
treatments designs, hence we rely primarily on numerical differences in standardized effect sizes
Page 35 of 67 Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
RUNNING HEAD: Bang for your Buck: CAS Tx Amount Distribution
36
(our primary outcome measure). However, to exercise reasonable caution in interpretation, we
only consider conditions to differ if both effect size and percent change are in the same direction,
and consider ambiguous effects to be ties. With this more conservative criterion, four of the six
children showed a massed practice advantage at maintenance, and four at follow-up, with one
reverse pattern and one tie at each time period.
Of the two children with a distributed practice advantage that was consistent between
effect size and percent change (CAS008 at M2 and CAS001 at FU), one (CAS008) showed a
reverse pattern at maintenance. For him, the data indicate a steady decline for distributed practice
items over the follow-up period and a steady increase for massed practice items (see Figure 3).
While his data must be interpreted with caution given that he also showed improvements for
untreated items in treatment phase 1, this pattern across the follow-up phase does suggest that
massed practice may have resulted in more robust speech motor learning (longer-term retention)
than distributed practice. Unfortunately we do not have longer-term follow-up data to determine
whether this trend would have continued. In contrast, for CAS001, there appears to be a more
robust advantage for distributed practice: he showed a tie during maintenance (distributed
practice advantage for percent change), and massed practice items declined during the follow-up
period while distributed practice items improved. It is interesting to note that CAS001 was
considerably older than the other children, because as noted in the Introduction, most principles
of motor learning are based on studies with adults. Thus, the observation that the oldest child was
the only one to show a relatively clear advantage for distributed practice (consistent with the
motor learning literature) raises the intriguing possibility that effects of practice conditions may
vary with age, such that massed practice may be optimal for younger children and benefits of
Page 36 of 67Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
RUNNING HEAD: Bang for your Buck: CAS Tx Amount Distribution
37
distributed practice emerge for older children. At this point, given these limited data, this
possibility remains necessarily speculative, but it deserves attention in future research.
When considered as a group, there was a numerical advantage for massed practice
compared to distributed practice both in terms of effect size and percent change, at both
maintenance and follow-up. Although the two treatment conditions did not differ significantly,
likely due to the small sample size, only the massed practice condition differed significantly
from the control condition for effect size at maintenance. In addition, massed practice showed an
effect (d > 1) for 5/6 children at maintenance and 6/6 children at follow-up, compared to 5/6 and
3/6 for distributed practice (and 0/6 and 1/6 for control items).
On the whole, the weight of the evidence of our findings suggests a benefit for massed
over distributed practice for children with CAS. These findings are consistent with the
neuroplasticity literature (Hypothesis 2b), in which massed practice is considered an important
learning principle (Kleim & Jones, 2008), as well as with the CAS treatment literature
(Namasivayam et al., 2015; Thomas et al., 2014). As in Thomas et al., we showed that the
massed practice advantage became more evident at longer retention intervals (with the exception
of CAS001). Thus, our findings provide further support from a well-controlled within-study
comparison for the use of intensive, massed practice in CAS treatment. It should be reiterated
that the manner in which practice distribution was manipulated here differs from that in previous
studies which varied the time period and/or practice amount (Namasivayam et al., 2015; Thomas
et al., 2014). Nevertheless, the findings provide converging evidence from divergent methods to
suggest that massed practice likely leads to greater, more robust improvement for most, though
not all, children with CAS. Although speculative, according to the neuroplasticity literature
(which is largely based on animal models of acquired brain damage), massed practice results in
Page 37 of 67 Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
RUNNING HEAD: Bang for your Buck: CAS Tx Amount Distribution
38
strengthened synaptic connections between neurons in pertinent neural circuits (Kleim & Jones,
2008). By extension, it is conceivable that in children with CAS, massed practice strengthens
synaptic connections in neural circuits underlying speech motor planning, for example
connections between auditory target and corresponding motor command (Terband et al., 2009) or
between somatosensory input and motor command (Terband & Maassen, 2010; Terband et al.,
2014). Future studies using neuroimaging and computational modeling simulations may be able
to shed further light on these possibilities.
ASSIST AND INTEGRAL STIMULATION TREATMENT EFFICACY
All children demonstrated treatment effects for at least one condition after completing
both phases of treatment, with larger gains and effect sizes for treated items than for untreated
control items, for which effect sizes in most cases did not exceed 1. In addition to demonstrating
treatment effects for all children individually, group averages and statistical analysis of group
data confirmed that treated items showed greater gains than untreated items. Thus, the present
study also adds another replication to the literature supporting the efficacy of integral stimulation
treatment for CAS. Despite differences in participants (e.g., DTTC was primarily developed for,
and studied in, younger children with more severe CAS) and differences in implementation
(discussed below), our protocol included the core aspects of integral stimulation treatment (e.g.,
integral stimulation, tactile cues, slowed rate, gradual fading of cues, focus on whole-target
movement accuracy). As such, the fact that we observed treatment effects strengthens the
evidence base for integral stimulation-based treatments, by demonstrating treatment effects for
this combination of shared components despite variations in other treatment parameters. Such
convergence illustrates that the effects in prior studies were not specific to those participants or
Page 38 of 67Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
RUNNING HEAD: Bang for your Buck: CAS Tx Amount Distribution
39
clinicians, or specific target selection criteria. Rather, the growing evidence base suggests that it
is the integral stimulation-based elements that are responsible for emergence of treatment effects.
Nevertheless, despite this replicated evidence, the degree of improvement varies across
studies. In the present study, gains were modest compared to some previous studies using
integral-stimulation-based approaches (Strand & Debertine, 2000; Strand et al., 2006). While the
presence of treatment effects across studies supports the efficacy of integral stimulation-based
treatment regardless of variation in treatment parameters, the magnitude of treatment effects is
likely influenced by variations in treatment protocol and by methodological differences. Below
we briefly discuss some of these differences to inform design of future treatment studies.
One of the most important differences in approach relates to target selection. In the
present study, targets were selected primarily based on personal functional relevance. Previous
studies have shown gains for treated items but little to no generalization to untrained items (Maas
& Farinella, 2012; Maas et al., 2012; Strand et al., 2006). Following recommendations from our
earlier work (Maas & Farinella, 2012; Maas et al., 2012), we chose personally meaningful targets
to enhance motivation and potential for carryover. This choice was also consistent with the
neuroplasticity principle of salience (Kleim & Jones, 2008), which recommends practice on
relevant, meaningful behaviors. However, the relative complexity of our targets likely resulted in
smaller gains compared to studies in which targets were functionally relevant but less complex
(e.g., Strand et al., 2006). Selecting targets that are personally relevant and build on movement
gestures and syllable shapes already mastered may lead to greater gains on those items and may
enhance motivation due to increased success during treatment (Jakielski, 2016). However, a
drawback of this strategy is that it often excludes targets with high personal relevance for a child
(e.g., their name includes a sound not yet mastered). Anecdotal clinical observations also suggest
Page 39 of 67 Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
RUNNING HEAD: Bang for your Buck: CAS Tx Amount Distribution
40
that high motivation may help a child overcome the difficulty of a complex target. Nevertheless,
selection of challenging targets in the present study likely contributed to the relatively modest
overall effect sizes compared to studies using DTTC. As discussed elsewhere (Maas et al., 2014),
target selection is a critical element of any treatment approach, and to our knowledge, there has
not been a systematic comparison of different methods of target selection in treatment for CAS.
Another important difference in approach with previous studies relates to overall amount
and distribution of practice. For example, studies using DTTC typically involved one hour of
treatment per day, five days per week, for six weeks. Instead, in our work here and elsewhere,
treatment was typically administered in two or three weekly 1-hour sessions over five to eight
weeks. Thus, studies using DTTC typically involved more, and more massed, treatment. If
massed practice is indeed more beneficial for children with CAS, then our more distributed
treatment here (in terms of sessions over weeks) may have led to more modest effect sizes.
Finally, ASSIST is less dynamic than DTTC, because in order to adequately study the
intervention and enable replication, we opted to create a more structured treatment protocol,
which necessarily limited some of the dynamic elements of the treatment. For instance, the
current protocol did not include a pre-practice phase, and the choice of elicitation method was
pre-specified (though still based on the child’s performance as judged by the SLP). Similarly,
each teaching episode began and ended with an elicitation of the target, practice was always
random, and a verbal feedback schedule was pre-specified. Thus, while ASSIST retained some
dynamic elements of treatment (e.g., type of cues and support provided in a teaching episode;
choice of elicitation method based on child’s performance), ASSIST did have fewer dynamic
aspects than DTTC. Greater gains might have emerged with a less structured protocol.
Page 40 of 67Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
RUNNING HEAD: Bang for your Buck: CAS Tx Amount Distribution
41
Several methodological differences may also have contributed to the more modest gains
in the present study. One such difference is that we randomly assigned targets to conditions, after
careful matching of target sets. This strategy was important to establish experimental control,
and avoided any potential bias (conscious or not) in selecting the easiest or most motivating
targets which may have contributed to larger gains in previous studies.
Another difference is that we administered probes at the beginning rather than at the end
of a session (as in Edeal & Gildersleeve-Neumann, 2011; Strand & Debertine, 2000). Moreover,
probes involved delayed production tasks (as in Edeal & Gildersleeve-Neumann, 2011), rather
than immediate imitation (as in Maas et al., 2012; Strand et al., 2006). Not all children reliably
achieved delayed elicitation methods during treatment; thus, the probe task represented a more
challenging level (a higher bar) than that experienced during treatment, which may have
depressed performance on probes but reflects a higher (desired) level of skill improvement.
Finally, our data were analyzed by blinded analysts, rather than by the treating clinician
(e.g., Edeal & Gildersleeve-Neumann, 2011; Strand et al., 2006). This blinding eliminated bias in
analysis, and controlled for potential perceptual drift across the treatment period. Some portion
of the gains in previous studies may have been attributable to potential bias or perceptual drift;
by controlling those factors here we may have a more accurate estimate of the true (more
modest) change attributable to the child’s improvements in speech motor skill.
CLINICAL IMPLICATIONS
The findings have several clinical implications for treatment for CAS. These implications
must be considered preliminary given the caveats noted throughout the discussion above.
First, the replication of the efficacy of integral stimulation-based intervention for six new
children with CAS supports the clinical use of such interventions. Although the evidence base
Page 41 of 67 Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
RUNNING HEAD: Bang for your Buck: CAS Tx Amount Distribution
42
remains small, integral stimulation approaches have been supported by seven peer-reviewed
studies to date, with a total of 17 children (Baas et al., 2008; Edeal & Gildersleeve-Neumann,
2011; Gildersleeve-Neumann & Goldstein, 2015; Maas & Farinella, 2012; Maas et al., 2012;
Strand & Debertine, 2000; Strand et al., 2006). The present study expands this to 23 children.
Second, children with CAS show greater speech motor learning with more practice.
Treatment phase 1 findings suggest that eight sessions of 25 minutes may not result in robust
gains when targeting 10 items, at least when using ASSIST. Whereas 5/6 children demonstrated
larger and positive effects for the 5-item set than for untreated items, only 2/6 showed such
effects for the 10-item set. Thus, when the limiting factor is number and duration of sessions,
clinicians can increase practice amount and expect gains by decreasing the number of targets. In
a 25-minute session, children completed on average 25 teaching episodes and 71 whole-word
production attempts, regardless of condition (see Appendix B). For a set of 5 items, this means
approximately 5 teaching episodes and about 14 whole-word production attempts per item,
compared to 2.5 teaching episodes and 7 whole-word attempts per item for a set of 10 items.
Third, most children with CAS may show greater, more robust improvements for items
practiced in a massed regimen. In the present study, distribution of a fixed amount of practice
was manipulated by varying set size and practicing items either in two separate “bursts” of four
weeks or over a period of eight weeks. Thus, for a child with a larger set of target items, our
findings suggest that selecting subsets for more intensive practice may lead to greater gains over
a given period (here, the 10 massed items generally outperformed the 10 distributed items after 8
weeks). To the extent that we observed a massed practice advantage, our findings also provide
some support for intensive service delivery models (e.g., summer camps) in which distribution is
defined in terms of number of sessions over time (e.g., Preston et al., 2016). However, direct
Page 42 of 67Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
RUNNING HEAD: Bang for your Buck: CAS Tx Amount Distribution
43
study is needed to determine the benefits of such intervention models relative to more distributed
service delivery models (e.g., same number of sessions throughout a semester or school year).
FUTURE DIRECTIONS
Future studies should replicate and extend this work, and investigate other aspects of
intensity in treatment for CAS. Given the inter-individual variability in treatment response and
pattern, future research should also examine participant factors (e.g., age, severity) and target
selection factors (e.g. complexity) that may impact treatment factors, in order to move towards
an evidence-based approach to providing optimal and personalized care for children with CAS.
To date, the support for efficacy of integral stimulation treatment has been based
exclusively on speech accuracy measures. However, the success and clinical relevance of a
treatment must not solely be based on impairment-level measures but also on measures of
activity and participation. Very few studies have used activity- and participation-level measures
in treatment for motor-based pediatric speech disorders, including CAS (Kearney et al., 2015).
Future studies should examine other, more functional outcome measures, such as parent-ratings
of intelligibility and participation, as well as direct intelligibility measures obtained from
unfamiliar listeners (as in Namasivayam et al., 2015). A study examining such measures for the
children in the present study is currently in progress and will be reported in a separate report.
CONCLUSIONS
In conclusion, this study showed greater gains with more practice for most children with
CAS. Not all children showed this benefit, possibly due to the limited amount of overall practice
and some unanticipated item-specific effects. In addition, five of the six children revealed a
massed practice advantage, contrary to the motor learning literature but consistent with the
neuroplasticity literature. The massed practice benefit emerged especially at longer-term
Page 43 of 67 Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
RUNNING HEAD: Bang for your Buck: CAS Tx Amount Distribution
44
retention intervals. The study also provided another well-controlled replication of the efficacy of
integral stimulation-based interventions (here, ASSIST). Overall gains were modest, however,
and several important considerations were discussed to help optimize treatment and research
designs for future studies. Future studies should also include more functional outcome measures.
Acknowledgments
We thank Apraxia Kids (formerly known as CASANA: Childhood Apraxia of Speech
Association of North America) for funding. We also thank Carolina Echeverri, Talia
Irgangladen, Leanne Long, Brian Kulsik, and Sarah Rosenberg for their help with data and
fidelity analyses; the SLP for evaluating the children and administering the treatment; and Kyra
Skoog for computing complexity scores for target stimuli. Most importantly, we thank the
children and their families for their time and participation. Portions of this research were
conducted as part of two Masters theses (by Nicolette Kovacs and Mackenzie Welsh).
Preliminary data were presented at the American Speech-Language-Hearing Association
Convention (Los Angeles, CA, November 2017).
Page 44 of 67Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
RUNNING HEAD: Bang for your Buck: CAS Tx Amount Distribution
45
REFERENCES
Allen, M. M. (2013). Intervention efficacy and intensity for children with speech sound disorder.
Journal of Speech, Language, and Hearing Research, 56, 865-877.
American Speech-Language-Hearing Association. (2007). Childhood apraxia of speech
[Technical Report]. Available at www.asha.org/policy.
American Speech-Language-Hearing Association. (2012). 2012 schools survey report: SLP
caseload characteristics. Available at ww.asha.org/research/memberdata/schoolssurvey/
Askim, T., Mørkved, S., Engen, A., Roos, K., Aas, T., & Indredavik, B. (2010). Effects of a
community-based intensive motor training program combined with early supported
discharge after treatment in a comprehensive stroke unit. A randomized, controlled trial.
Stroke, 41, 1697-1703.
Baas, B. S., Strand, E. A., Elmer, L. M., & Barbaresi, W. J. (2008). Treatment of severe
childhood apraxia of speech in a 12-year-old male with CHARGE association. Journal of
Medical Speech-Language Pathology, 16, 181-191.
Baddeley, A. D., & Longman, D. J. A. (1978). The influence of length and frequency of training
session on the rate of learning to type. Ergonomics, 21, 627–635.
Bahrick, H. P., & Hall, L. K. (2005). The importance of retrieval failures to long-term retention:
A metacognitive explanation of the spacing effect. Journal of Memory and Language, 52,
566-577.
Ballard, K. J., Maas, E., & Robin, D. A. (2007). Treating control of voicing in apraxia of speech
with variable practice. Aphasiology, 21, 1195-1217.
Ballard, K. J., Robin, D. A., McCabe, P., & McDonald, J. (2010). A treatment for dysprosody in
childhood apraxia of speech. Journal of Speech, Language, and Hearing Research, 53,
Page 45 of 67 Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
RUNNING HEAD: Bang for your Buck: CAS Tx Amount Distribution
46
1227-1245.
Beeson, P. M., & Robey, R. R. (2006). Evaluating single-subject treatment research: Lessons
learned from the aphasia literature. Neuropsychology Review, 16, 161-169.
Bjork, R. A., & Allen, T. W. (1970). The spacing effect: Consolidation or differential encoding?
Journal of Verbal Learning and Verbal Behavior, 9, 567-572.
Brashers-Krug, T., Shadmehr, R., & Bizzi, E. (1996). Consolidation in human motor memory.
Nature, 382 (6588), 252-255.
Busk, P. L., & Serlin, R. (1992). Meta-analysis for single case research. In T. R. Kratochwill & J.
R. Levin (Eds.), Single-case research design and analysis: New directions for psychology
and education. Hillsdale, NJ: Lawrence Erlbaum Associates.
Campbell, T. F. (1999). Functional treatment outcomes in young children with motor speech
disorders. In A. J. Caruso & E. A. Strand (Eds.), Clinical management of motor speech
disorders in children (pp. 385-396). New York: Thieme.
Dodd, B., Hua, Z., Crosbie, S., Holm, A., & Ozanne, A. (2006). Diagnostic evaluation of
articulation and phonology. San Antonio, TX: PsychCorp.
Dunn, L. M., & Dunn, D. M. (2007). Peabody picture vocabulary test (4th ed.). AGS/Pearson.
Edeal, D. M., & Gildersleeve-Neumann, C. E. (2011). The importance of production frequency
in therapy for childhood apraxia of speech. American Journal of Speech-Language
Pathology, 20, 95-110.
Goldman, R., & Fristoe, M. (2000). The Goldman–Fristoe test of articulation (2nd ed.). Austin,
TX: Pro-Ed.
Hall, P. K. (2000). A letter to the parent(s) of a child with developmental apraxia of speech. Part
IV: Treatment of DAS. Language, Speech, and Hearing Services in Schools, 31, 179-181.
Page 46 of 67Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
RUNNING HEAD: Bang for your Buck: CAS Tx Amount Distribution
47
Jakielski, K. J. (2002). The Index of Phonetic Complexity: Preliminary findings. Paper presented
at the Annual Child Phonology Conference (Wichita, KS, May 2002).
Jakielski, K. J. (2016). Building speech and quantifying complexity. The manual. Pittsburgh,
PA: Childhood Apraxia of Speech Association of North America.
Kaderavek, J. N., & Justice, L. M. (2010). Fidelity: An essential component of evidence-based
practice in speech-language pathology. American Journal of Speech-Language
Pathology, 19, 369-379.
Kaipa, R., & Peterson, A. M. (2016). A systematic review of treatment intensity in speech
disorders. International Journal of Speech-Language Pathology, 18, 507-520.
Kearney, E., Granata, F., Yunusova, Y., van Lieshout, P., Hayden, D., & Namasivayam, A.
(2015). Outcome measures in developmental speech sound disorders with a motor basis.
Current Developmental Disorders Reports, 2, 253-272.
Kearns, K. P. (1986). Flexibility of single-subject experimental designs. Part II: Design selection
and arrangement of experimental phases. Journal of Speech and Hearing Disorders, 51,
204-214.
Kim, I., LaPointe, L. L., & Stierwalt, J. A. G. (2012). The effect of feedback and practice on the
acquisition of novel speech behaviors. American Journal of Speech-Language Pathology,
21, 89-100.
Kleim, J. A., Barbay, S., Cooper, N. R., Hogg, T. M., Reidel, C. N., …, & Nudo, R. J. (2002).
Motor learning-dependent synaptogenesis is localized to functionally reorganized motor
cortex. Neurobiology of Learning and Memory, 77, 63-77.
Page 47 of 67 Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
RUNNING HEAD: Bang for your Buck: CAS Tx Amount Distribution
48
Kleim, J. A., & Jones, T. A. (2008). Principles of experience-dependent neural plasticity:
Implications for rehabilitation after brain damage. Journal of Speech, Language, and
Hearing Research, 51, S225-S239.
Kwakkel, G., Wagenaar, R. C., Twisk, J. W. R., Lankhorst, G. J., & Koetsier, J. C. Intensity of
leg and arm training after primary middle-cerebral artery stroke: A randomised trial. The
Lancet, 354, 191-196.
Lee, T. D., & Genovese, E. D. (1988). Distribution of practice in motor skill acquisition:
Learning and performance effects reconsidered. Research Quarterly for Exercise and
Sport, 59, 277-287.
Lee, T. D., & Weeks, D. J. (1987). The beneficial influence of forgetting on short-term retention
of movement information. Human Movement Science, 6, 233-245.
Luke, L. M., Allred, R. P., & Jones, T. A. (2004). Unilateral ischemic sensorimotor cortical
damage induces contralesional synaptogenesis and enhances skilled reaching with the
ipsilateral forelimb in adult male rats. Synapse, 54, 187-199.
Maas, E., Butalla, C.E., & Farinella, K.A. (2012). Feedback frequency in treatment for childhood
apraxia of speech. American Journal of Speech-Language Pathology, 21, 239-257
Maas, E., & Farinella, K. A. (2012). Random versus blocked practice in treatment for childhood
apraxia of speech. Journal of Speech, Language, and Hearing Research, 55, 561-578.
Maas, E., Gildersleeve-Neumann, C. E., Jakielski, K. J., & Stoeckel, R. (2014). Motor-based
intervention protocols in treatment of childhood apraxia of speech (CAS). Current
Developmental Disorders Reports, 1 (3), 197-206.
Page 48 of 67Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
RUNNING HEAD: Bang for your Buck: CAS Tx Amount Distribution
49
Maas, E., Robin, D. A., Austermann Hula, S. N., Freedman, S. E., Wulf, G., … , & Schmidt, R.
A. (2008). Principles of motor learning in treatment of motor speech disorders. American
Journal of Speech-Language Pathology, 17, 277-298.
McAllister Byun, T., Hitchcock, E. R., & Swartz, M. T. (2014). Retroflex versus bunched in
treatment for rhotic misarticulation: Evidence from ultrasound biofeedback intervention.
Journal of Speech, Language, and Hearing Research, 57, 2116-2130.
McAllister Byun, T., Swartz, M. T., Halpin, P. F., Szeredi, D., & Maas, E. (2016). Direction of
attentional focus in biofeedback treatment for /r/ misarticulation. International Journal of
Language and Communication Disorders, 51, 384-401.
McCauley, R. J., & Strand, E. A. (2008). A review of standardized tests of nonverbal oral and
speech motor performance in children. American Journal of Speech-Language
Pathology, 17, 81-91.
Murray, E., McCabe, P., & Ballard, K. J. (2014). A systematic review of treatment outcomes for
children with Childhood Apraxia of Speech. American Journal of Speech-Language
Pathology, 23, 486-504.
Murray, E., McCabe, P., & Ballard, K. J. (2015). A randomized controlled trials for children
with childhood apraxia of speech comparing Rapid Syllable Transition treatment and the
Nuffield Dyspraxia Programme–Third Edition. Journal of Speech, Language, and
Hearing Research, 58, 669-686.
Murray, E., McCabe, P., Heard, R., & Ballard, K. J. (2015). Differential diagnosis of children
with suspected childhood apraxia of speech. Journal of Speech, Language, and Hearing
Research, 58, 43-60.
Page 49 of 67 Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
RUNNING HEAD: Bang for your Buck: CAS Tx Amount Distribution
50
Namasivayam, A. K., Pukonen, M., Goshulak, D., Hard, J., Rudzicz, F., … , & van Lieshout, P.
(2015). Treatment intensity and childhood apraxia of speech. International Journal of
Language and Communication Disorders, 50, 529-546.
Preston, J. L., Leece, M. C., & Maas, E. (2016). Intensive treatment with ultrasound visual
feedback for speech sound errors in childhood apraxia. Frontiers in Human
Neuroscience, 10:440.
Preston, J. L., Leece, M. C., McNamara, K., & Maas, E. (2017). Variable practice during
ultrasound visual feedback treatment for childhood apraxia of speech. American Journal
of Speech-Language Pathology, 26, 840-852.
Reynolds, C. R., & Kamphaus, R. W. (2003). Reynolds intellectual assessment scales (RIAS).
Lutz, FL: Psychological Assessment Resources, Inc.
Robbins, J., & Klee, T. (1987). Clinical assessment of oropharyngeal motor development in
young children. Journal of Speech and Hearing Disorders, 52, 271-277.
Robin, D. A., Maas, E., Sandberg, Y., & Schmidt, R. A. (2007). Motor control and learning and
childhood apraxia of speech. In P. Hall, L. Jordan, & D. Robin (Eds.), Developmental
apraxia of speech: Theory and clinical practice (2nd ed.) (pp.67-86). Austin, TX: Pro-Ed.
Rvachew, S., Hodge, M., & Ohberg, A. (2005). Obtaining and interpreting maximum
performance tasks from children: A tutorial. Journal of Speech-Language Pathology and
Audiology, 29, 146-157.
Schmidt, R. A., & Lee, T. D. (2005). Motor control and learning: A behavioral emphasis (4th
ed.). Champaign, IL: Human Kinetics.
Semel, E., Wiig, E. H., & Secord, W. A. (2003). Clinical evaluation of language fundamentals®
(4th ed.). San Antonio, TX: Pearson.
Page 50 of 67Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
RUNNING HEAD: Bang for your Buck: CAS Tx Amount Distribution
51
Shea, C. H., Kohl, R. M., & Indermill, C. (1990). Contextual interference: Contributions of
practice. Acta Psychologica, 73,145–157.
Shea, C. H., Lai, Q., Black, C., & Park, J. H. (2000). Spacing practice sessions across days
benefits the learning of motor skills. Human Movement Science, 19, 737–760.
Shriberg, L. D., Aram, D. M., & Kwiatkowski, J. (1997). Developmental apraxia of speech: I.
Descriptive and theoretical perspectives. Journal of Speech, Language, and Hearing
Research, 40, 273-285.
Skelton, S. L., & Hagopian, A. L. (2014). Using randomized variable practice in the treatment of
childhood apraxia of speech. American Journal of Speech-Language Pathology, 23, 599-
611.
Spielman, J., Ramig, L. O., Mahler, L., Halpern, A., & Gavin, W. J. (2007). Effects of an
extended version of the Lee Silverman Voice Treatment on voice and speech in
Parkinson’s disease. American Journal of Speech-Language Pathology,16, 95-107.
Størvold, G. V., Jahnsen, R. B., Evensen, K. A. I., & Bratberg, G. H. (in press). Is more frequent
physical therapy associated with increased gross motor improvement in children with
cerebral palsy? A national prospective cohort study. Disability & Rehabilitation.
Strand, E. A., Stoeckel, R., & Baas, B. (2006). Treatment of severe childhood apraxia of speech:
A treatment efficacy study. Journal of Medical Speech-Language Pathology, 14, 297-
307.
Terband, H., & Maassen, B. (2010). Speech motor development in childhood apraxia of speech
(CAS): Generating testable hypotheses by neurocomputational modeling. Folia
Phoniatrica et Logopaedica, 62, 134-142.
Terband, H., Maassen, B., Guenther, F. H., & Brumberg, J. (2009). Computational neural
Page 51 of 67 Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
RUNNING HEAD: Bang for your Buck: CAS Tx Amount Distribution
52
modeling of childhood apraxia of speech (CAS). Journal of Speech, Language and
Hearing Research, 52, 1595-1609.
Terband, H., Maassen, B., Guenther, F. H., & Brumberg, J. (2014). Auditory-motor interactions
in pediatric motor speech disorders: Neurocomputational modeling of disordered
development. Journal of Communication Disorders, 47, 17-33.
Thomas, D. C., McCabe, P., & Ballard, K. J. (2014). Rapid Syllable Transitions (ReST)
treatment for childhood apraxia of speech: The effect of lower dose-frequency. Journal of
Communication Disorders, 51, 29-42.
Thoonen, G., Maassen, B., Gabreëls, F., & Schreuder, R. (1999). Validity of maximum
performance tasks to diagnose motor speech disorders in children. Clinical Linguistics &
Phonetics, 13, 1-23.
Thoonen, G., Maassen, B., Wit, J., Gabreëls, F., & Schreuder, R. (1996). The integrated use of
maximum performance tasks in differential diagnostic evaluations among children with
motor speech disorders. Clinical Linguistics and Phonetics, 10, 311-336.
Wagner, R. K., Torgesen, J. K., & Rashotte, C. A. (1999). Comprehensive test of phonological
processing. Austin, TX: Pro-Ed.
Wambaugh, J. L. , Nessler, C., Cameron, R., & Mauszycki, S. C. (2013). Treatment for acquired
apraxia of speech: Examination of treatment intensity and practice schedule. American
Journal of Speech-Language Pathology, 22, 84-102.
Warren, S. F., Fey, M. E., & Yoder, P. J. (2007). Differential treatment intensity research: A
missing link to creating optimally effective communication interventions. Mental
Retardation and Developmental Disabilities Research Reviews, 13, 70-77.
Williams, K. (2007). Expressive vocabulary test (2nd ed.). Circle Pines, MN: American
Page 52 of 67Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
RUNNING HEAD: Bang for your Buck: CAS Tx Amount Distribution
53
Guidance Services.
Wilson, A., & Gildersleeve-Neumann, C. E. (2014). Functional communication parent
questionnaire. Unpublished instrument, Portland State University.
Wolpert, D. M., Ghahramani, Z., & Flanagan, J. R. (2001). Perspectives and problems in motor
learning. Trends in Cognitive Sciences, 5 (11), 487-494.
Wright, D. L., Black, C. B., Immink, M. A., Brueckner, S., & Magnuson, C. (2004). Long-term
motor programming improvements occur via concatenating movement sequences during
random but not blocked practice. Journal of Motor Behavior, 36, 39-50.
Page 53 of 67 Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
RUNNING HEAD: Bang for your Buck: CAS Tx Amount Distribution
54
Figure 1. Overview of study design. BL = initial baseline phase; Tx1 = treatment phase 1; M1 =
maintenance phase 1; Tx2 = treatment phase 2; M2 = maintenance phase 2; FU = follow-up.
Figure 2. Data plots for children 001, 002, and 003. Shading indicates treatment phases. Note
that due to scheduling constraints, the number of probes during maintenance phase 1 varied
between children.
Figure 3. Data plots for children 004, 005, and 008. Shading indicates treatment phases. Note
that due to scheduling constraints, the number of probes during maintenance phase 1 varied
between children.
Supplemental File S1. Detailed information about the items sets for each child, including the
items and the condition matching information.
Page 54 of 67Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
Table 1. Participant information.
ID
Sex
Age
GFTAa
DEAPb
EVTc
PPVTd
CELFe
RIASf
MaxPT Dysg
MaxPT CASh
CAS Dxi
001
M
11;3
<40
68%
73
74
61
48
2
2
1.50 (0.58)
002
M
7;11
55
80%
88
89
105
42
0
2
1.75 (0.50)
003
M
5;11
47
52%
88
92
98
43
0
2
1.75 (0.50)
004
M
5;11
41
72%
84
82
n/c
48
0
2
1.50 (0.58)
005
M
6;0
45
64%
60
75
n/c
45
2
2
1.75 (0.50)
008
M
4;7
75
52%
96
116
n/a
61
0
2
1.25 (0.50)
a Goldman-Fristoe Test of Articulation, 2nd ed. (Goldman & Fristoe, 2000) standard score; b Diagnostic Evaluation of Articulation and Phonology (Dodd et al.,
2006) Word Inconsistency Subtest percent inconsistent; c Expressive Vocabulary Test, 2nd ed. (Williams, 2007) standard score; d Peabody Picture Vocabulary
Test, 4th ed. (Dunn & Dunn, 2007) standard score; e Clinical Evaluation of Language Fundamentals, 4th ed. (Semel et al., 2003) Receptive Language Index
standard score; f Reynolds Intellectual Assessment Scales (Reynolds & Kamphaus, 2003) nonverbal cognition composite T-score; g Maximum Performance Test
Protocol (Rvachew et al., 2005; Thoonen et al., 1999) dysarthria score; h Maximum Performance Test Protocol (Rvachew et al., 2005; Thoonen et al., 1999)
apraxia of speech score; i Mean (standard deviation) of four expert SLP ratings (0 = no CAS, 1 = possible CAS, 2 = CAS). n/c = not computed (see text); n/a =
not administered.
Page 55 of 67 Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
Table 2. Effect sizes (d) and percent change by condition for Amount (treatment phase 1).
Child
High amount
Low amount
Control
d
%change
d
%change
d
%change
001
1.09
16.7%
2.31
20.0%
0.57
10.0%
002
1.83
17.5%
-1.79
-11.3%
-2.98
-18.8%
003
2.35
22.5%
1.13
5.4%
1.12
9.6%
004
2.83
14.2%
-0.41
-1.7%
-0.60
-3.8%
005
3.18
28.6%
2.47
13.2%
1.39
15.0%
008
0.21
1.8%
0.89
9.4%
1.38
9.0%
Mean
1.45
13.5%
0.69
6.6%
0.26
5.3%
SD
1.22
9.9%
1.45
11.4%
1.77
14.1%
Page 56 of 67Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
Table 3. Effect sizes (d) and percent change by condition for Distribution. M2 = maintenance
phase 2 (probes M2-1 and M2-2); FU = follow-up period (4 & 8 week follow-up combined).
Period
Child
Massed
Distributed
Control
d
%change
d
%change
d
%change
M2
001
2.74
15.8%
2.31
20.0%
0.24
4.2%
002
2.53
11.3%
0.60
3.8%
0.60
3.7%
003
2.58
16.3%
1.83
8.8%
0.73
6.3%
004
3.50
17.5%
1.84
7.5%
0.20
1.3%
005
2.32
18.6%
1.07
5.7%
-0.23
-2.5%
008
0.48
4.8%
2.06
21.9%
0.61
4.0%
Mean
2.36
14.1%
1.62
11.3%
0.36
2.8%
SD
1.01
5.2%
0.65
7.7%
0.36
3.1%
FU
001
2.74
15.8%
3.46
30.0%
0.52
9.2%
002
2.87
13.8%
0.99
6.3%
-0.20
-1.3%
003
2.19
13.8%
2.35
11.3%
1.02
8.8%
004
3.50
17.5%
0.61
2.5%
0.20
1.3%
005
2.63
21.1%
2.00
10.7%
0.69
7.5%
008
1.47
14.8%
0.89
9.4%
1.00
6.5%
Mean
2.57
16.1%
1.72
11.7%
0.54
5.3%
SD
0.68
2.8%
1.09
9.5%
0.47
4.3%
Page 57 of 67 Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
phase: BL Tx1 M1 Tx2 M2 FU1 FU2
weeks: 1-6 7-10 11-12 13-16 17-18 20 24
2 sessions/week
50-min. sessions
25 min./condition
(total: 200 min./condition)
Massed Subset 1
5 targets
Distributed Set
10 targets
2 sessions/week
50-min. sessions
25 min./condition
(total: 200 min.condition)
Massed Subset 2
5 targets
Distributed Set
10 targets
Page 58 of 67Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
BL1 BL2 BL3 Pr1-1 Pr1-2 Pr1-3 M1-1 M1-2 M1-3 Pr2-1 Pr2-2 Pr2-3 M2-1 M2-2 FU-1 FU-2
Accuracy (%)
001 Distributed
Massed_1
Massed_2
Control
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
BL1 BL2 BL3 BL4 Pr1-1 Pr1-2 Pr1-3 M1-1 M1-2 Pr2-1 Pr2-2 Pr2-3 Pr2-4 M2-1 M2-2 FU-1 FU-2
Accuracy (%)
002 Distributed
Massed_1
Massed_2
Control
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
BL1 BL2 BL3 BL4 Pr1-1 Pr1-2 Pr1-3 M1-1 M1-2 M1-3 Pr2-1 Pr2-2 Pr2-3 M2-1 M2-2 FU-1 FU-2
Accuracy (%)
003 Distributed
Massed_1
Massed_2
Control
Page 59 of 67 Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
BL1 BL2 BL3 BL4 Pr1-1 Pr1-2 Pr1-3 M1-1 M1-2 M1-3 Pr2-1 Pr2-2 Pr2-3 M2-1 M2-2 FU-1 FU-2
Accuracy (%)
004 Distributed
Massed_1
Massed_2
Control
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
BL1 BL2 BL3 BL4 BL5 BL6 BL7 Pr1-1Pr1-2Pr1-3M1-1M1-2 Pr2-1Pr2-2Pr2-3 Pr2-4M2-1M2-2 FU-1 FU-2
Accuracy (%)
005 Distributed
Massed_1
Massed_2
Control
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
BL1 BL2 BL3 BL4 BL5 Pr1-1Pr1-2Pr1-3M1-1M1-2M1-3Pr2-1Pr2-2Pr2-3M2-1M2-2 FU-1 FU-2
Accuracy (%)
008 Distributed
Massed_1
Massed_2
Control
Page 60 of 67Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
APPENDIX A: TREATMENT PROTOCOL
Steps for each teaching episode are outlined in the Protocol Flowchart below.
In each treatment phase, there are 5 targets for the massed condition and 10 targets for the
distributed condition. A card deck contains 2 stimulus cards for each target per condition.
Teaching episodes are presented in random order for both conditions. Random order of teaching
episodes is implemented by shuffling the cards before the treatment session and reshuffling the
deck as needed during the session once every card has been used.
Feedback is given with decreasing frequency during each condition. Feedback is provided for
100% (10) of the first 10 teaching episodes, 90% (9) of the next 10 teaching episodes, and so on
until feedback is given on 10% (1) of the last 10 teaching episodes. To facilitate keeping track of
feedback schedules, the SLP has a feedback tracking sheet with 100 slots with the fading
feedback schedule marked (Ballard et al., 2007).
Start each condition in each session with immediate imitation at normal rate. The criterion to
increase difficulty level of elicitation condition is 2/2 consecutive correct Attempts #1. That is, if
the child produces a correct response on Attempt #1 on two successive teaching episodes
(whether these are the same target or not), then the third teaching episode begins with the next
level elicitation method. Similarly, if the child produces 2/2 consecutive incorrect Attempts #1
with an elicitation method, the next teaching episode reverts to the previous difficulty level of
elicitation method (or stay in immediate imitation at slow rate).
Continue until session time has elapsed (determined by an egg-timer set to 25 minutes per
condition). After a brief break, begin the second treatment condition and continue until session
time is up.
Page 61 of 67 Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
SLP elicits attempt #1 using one of following elicitation methods:
Immediate imitation, slowed rate (“watch me, listen carefully, and say .”)
Immediate imitation, normal rate (“watch me, listen carefully, and say .”)
Delayed imitation, normal rate (“watch me, listen carefully, and say .” Points to child after 3 seconds to cue delayed imitation)
Spontaneous production (Shows picture and says “What is this?”)
Child attempt #1
Correct
Incorrect
No FB trial
SLP provides no feedback
but waits 2-3 seconds
FB trial:
SLP provides feedback
after 2-3 seconds
(e.g., “good one!”)
No FB trial
SLP provides no feedback
but waits 2-3 sec.
FB trial:
SLP provides feedback after 2-3 sec.
(e.g., “good try, but not quite”, “your
lips were not rounded enough”)
SLP elicits attempt #2 (e.g., “can you say that again?”)
Child attempt #2
Correct
Incorrect
No FB trial
SLP provides
no feedback
but waits 2-3
seconds
FB trial:
SLP provides
feedback after
2-3 seconds
(e.g., “good!”)
SLP provides support (e.g., simultaneous production at slow rate,
tactile cue) & elicits attempt #2 using same method as for attempt #1
SLP starts next teaching episode (e.g., “let’s do another one”)
Child attempt #2
Correct
Incorrect
FB trial:
SLP provides
feedback after
2-3 seconds
(e.g., “close!”)
FB trial:
SLP provides
feedback after
2-3 seconds
(e.g., “good!”)
FB trial:
SLP provides
feedback after
2-3 seconds
(e.g., “close!”)
No FB trial
SLP provides
no feedback
but waits 2-3
seconds
No FB trial
SLP
provides no
feedback but
waits 2-3
seconds
No FB trial
SLP provides
no feedback
but waits 2-3
seconds
Page 62 of 67Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
APPENDIX B. TREATMENT FIDELITY DATA.
Table B. Fidelity measures. M = massed; D = distributed.
Child
001
002
003
004
005
008
Fidelity Aspect
M
D
M
D
M
D
M
D
M
D
M
D
# & % sessions reviewed
4/16 (25%)
3/16 (19%)
2/16 (12.5%)
2/16 (12.5%)
3/16 (19%)
2/16 (12.5%)
Duration (minutes)a
23
24
23
23
24
23
21
21
22
23
23
22
Number of TEsb
26
28
25
26
27
28
16
15
34
31
22
27
Total whole-word attemptsc
64
78
63
75
67
70
41
38
108
92
74
78
Whole-word attempts / TEd
2.48
2.72
2.53
2.88
2.49
2.48
2.53
2.58
3.30
2.96
3.40
2.89
FB frequencye
74%
64%
75%
58%
63%
53%
73%
72%
74%
80%
84%
58%
FB on 1st & last attemptf
58%
43%
59%
34%
42%
28%
53%
54%
59%
67%
70%
43%
FB delayg
< 1 sec.
68%
71%
70%
82%
87%
80%
59%
84%
79%
76%
89%
86%
1-3 sec.
29%
27%
27%
16%
12%
15%
28%
14%
19%
23%
9%
14%
4-6 sec.
3%
1%
3%
3%
1%
5%
12%
0%
2%
1%
2%
0%
> 6 sec.
0%
1%
1%
0%
0%
0%
2%
3%
0%
0%
0%
0%
Elicitation per protocolh
96%
99%
97%
98%
96%
100%
100%
96%
85%
84%
87%
89%
Same elicitation 1st & lasti
92%
88%
88%
86%
92%
97%
98%
91%
90%
84%
78%
74%
a Treatment duration per session (time from start of first teaching episode to start of last teaching episode; rounded to nearest minute).
b Number of teaching episodes per session.
c Number of whole-word target attempts per session.
d Number of whole-word target attempts per teaching episode (including first and last attempt).
e Proportion of total number of attempts (initial and final) with FB (# initial & final attempts with FB / total # of initial & final attempts).
f Proportion of teaching episodes with same presence/absence of FB on initial and final attempt.
g Proportion of initial and final attempts with feedback delays in each given range.
h Proportion of teaching episodes in which the initial elicitation method adhered to protocol.
i Proportion of teaching episodes with the same elicitation method for initial and final attempt (based only on TEs with incorrect initial attempt).
Page 63 of 67 Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
APPENDIX C. RESULTS FROM NONPARAMETRIC STATISTICAL GROUP ANALYSES.
Table C. Results of nonparametric group analyses for Research Question 1 (practice amount). CMH = Cochran-Mantel-Haenszel test;
Wilcoxon = Signed rank test. Significant effects (p < 0.05) are bolded, and direction of effect is indicated (Advantage column).
Dependent variable
Test
Comparison
Statistic
p-value
Advantage
CMH
High vs. Low vs. Control
χ2 = 4.33
0.115
High vs. Low
S = 6
0.250
-
High vs. Control
S = 8.5
0.094
-
d
Wilcoxon
Low vs. Control
S = 7.5
0.156
-
CMH
High vs. Low vs. Control
χ2 = 3.00
0.223
High vs. Low
S = 7.5
0.156
-
High vs. Control
S = 8.5
0.094
-
% change
Wilcoxon
Low vs. Control
S = 4.5
0.4375
-
Table D. Results of nonparametric group analyses for Research Question 2 (practice distribution). M2 = maintenance phase 2; FU =
follow-up; CMH = Cochran-Mantel-Haenszel test; Wilcoxon = Signed rank test. Significant effects (p < 0.05) are bolded, and
direction of effect is indicated (Advantage column).
Phase
Dependent variable
Test
Comparison
Statistic
p-value
Advantage
CMH
Distributed vs. Massed vs. Control
χ2 = 6.35
0.042
Distributed vs. Massed
S = 6.5
0.219
-
Distributed vs. Control
S = 7.5
0.063
-
d
Wilcoxon
Massed vs. Control
S = 9.5
0.063
-
CMH
Distributed vs. Massed vs. Control
χ2 = 9.33
0.009
Distributed vs. Massed
S = 3.5
0.500
-
Distributed vs. Control
S = 10.5
0.031
Distributed > Control
M2
% change
Wilcoxon
Massed vs. Control
S = 10.5
0.031
Massed > Control
CMH
Distributed vs. Massed vs. Control
χ2 = 7.00
0.030
Distributed vs. Massed
S = 5.5
0.313
-
Distributed vs. Control
S = 9.5
0.063
-
d
Wilcoxon
Massed vs. Control
S = 10.5
0.031
Massed > Control
CMH
Distributed vs. Massed vs. Control
χ2 = 10.33
0.006
Distributed vs. Massed
S = 5.5
0.313
-
Distributed vs. Control
S = 10.5
0.031
Distributed > Control
FU
% change
Wilcoxon
Massed vs. Control
S = 10.5
0.031
Massed > Control
Page 64 of 67Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
SUPPLEMENT S1.
ITEM SET INFORMATION.
PART 1: ITEM SETS.
Notes:
Identifying information is redacted and replaced with X (each X represents a syllable)
Numbers following items in Massed sets indicate treatment phase (Massed subset 1 or 2)
CAS 001
DISTRIBUTED
MASSED
CONTROL
I love cheeseburgers
I'm hungry (1)
Let's have breakfast
Exterminate
I know karate (1)
Tardis
I'm taking drum lessons
Can I have the iPad please? (1)
Let's do something else
Do you want to play Xbox?
What do you want to drink? (1)
Are you ready to go home?
Where is the bathroom?
XX is my sister (1)
Who is the Doctor?
Please don't do that
Thank you very much (2)
XXX is my friend
I play basketball
We have eight cats (2)
I like Target
XXX (last name)
Minecraft (2)
XX X (hometown)
Gingercat
My name is X (2)
Alfie is one of our dogs
Mrs. XX is my speech teacher
Time and Relative Dimension in
Space (2)
Let's watch YouTube videos
CAS 002
DISTRIBUTED
MASSED
CONTROL
I want a cereal bar
breakfast (1)
I like chocolate milk
XXX is my speech teacher
My dad's name is XX (1)
X is my brother
XXX (friend’s name)
When are we going? (1)
it's almost 6 o'clock
I like playdough
5 minutes later (1)
XXXX XXX Lane (street address)
I live in XXX
When is the movie? (1)
Where's the bathroom?
What time is it?
Minecraft (2)
Do you want to play?
memory
XX is my friend (2)
I'm sorry
Talk to you later
My name is XX XXX (2)
kitchen
Beauty and the Beast
Can you help me please? (2)
after 2 hours
What's your name?
Have a nice day (2)
Where are my shoes?
CAS 003
DISTRIBUTED
MASSED
CONTROL
My name is XXX
XX is my cousin (1)
X is my twin brother
Spiderman
Wolverine (1)
Captain America
Can you help me please?
Thank you (1)
I'm hungry
Batman
Excuse me (1)
Iron Man
Mrs. XX (SLP)
Charlie is our dog (1)
Paw Patrol
How are you today?
I wanna play outside (2)
Good morning mom
See you later
cereal (2)
I love you
pepperoni
strawberry milk (2)
chicken nuggets
breakfast
Mrs. X (teacher) (2)
Burger King
Where is the bathroom?
Who is that? (2)
What's for dinner?
Page 65 of 67 Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
CAS 004
DISTRIBUTED
MASSED
CONTROL
My name is X
What's on TV? (1)
XX is my cousin
Spiderman
superhero (1)
Wolverine
Can you help me please?
I'm thirsty (1)
I wanna play outside
Polar Bear
Batmobile (1)
Paw Patrol
Mrs. XX (SLP)
Good morning dad (1)
Charlie is our dog
How are you today?
I love you (2)
Thank you
See you later
Mrs. S (2)
Excuse me
pepperoni
chocolate milk (2)
cheese-its
Where are my shoes?
cereal (2)
Burger King
Where is the bathroom?
X is my twin brother (2)
Who is that?
CAS 005
DISTRIBUTED
MASSED
CONTROL
XX XX (name)
excuse me (1)
Pepper (dog)
thank you
yesterday (1)
hello
teacher
Mickey (1)
towel
doing
What's your name? (1)
tomorrow
Disney World
sister (1)
Phanatic
XX (friend)
XXX (sister) (2)
XX (friend)
please
XX (friend) (2)
see you later
hamburger
good morning (2)
guacamole
Where is the bathroom?
pizza (2)
How are you?
I'm tired
I'm hungry (2)
I'm thirsty
CAS 008
DISTRIBUTED
MASSED
CONTROL
XX XXX (name)
lollipops (1)
XX is my cousin
Petey the fish
museum (1)
Tony Box Turtle
XX (friend)
See you later (1)
Train station
Miss XX (SLP)
I want my Teddy (1)
Stewart (character)
Amelia Bedelia
Let's go (1)
Sister X (family friend)
Excuse me
Oreo (2)
No thank you
Please stop
Uncle XX (2)
Juice
What's your name?
breakfast (2)
How are you?
Play-doh
Where's the bathroom? (2)
Yesterday
I'm hungry
Angry Birds (2)
Cat in the Hat
Page 66 of 67Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Peer Review
PART 2: MATCHING INFORMATION.
Descriptive statistics for item set matching for length and baseline accuracy (SD for baseline
accuracy based on averages of item scores, not based on set scores per probe). IPC = Index of
Phonetic Complexity (Jakielski, 2002, 2016); IPC was calculated by treating each target as a
single word. Within each child, there were no differences between any item sets (ps > 0.05).
Participant
Set
# syllables (SD)
Baseline accuracy (SD)
IPC
001
Distributed
5.10 (1.85)
0.10 (0.11)
12.0 (4.2)
Massed
5.20 (2.25)
0.11 (0.11)
10.4 (3.6)
Massed 1
5.40 (1.52)
0.10 (0.14)
10.0 (3.9)
Massed 2
5.00 (3.00)
0.12 (0.08)
10.8 (3.6)
Control
5.00 (1.76)
0.15 (0.23)
11.4 (4.4)
002
Distributed
4.80 (1.75)
0.18 (0.21)
7.8 (4.2)
Massed
4.70 (1.77)
0.27 (0.18)
10.1 (3.5)
Massed 1
4.60 (1.52)
0.27 (0.19)
8.6 (1.1)
Massed 2
4.80 (2.17)
0.27 (0.19)
11.6 (4.5)
Control
4.70 (1.64)
0.22 (0.22)
10.9 (5.1)
003
Distributed
4.00 (1.33)
0.17 (0.24)
8.2 (3.8)
Massed
3.80 (1.40)
0.15 (0.24)
9.0 (2.5)
Massed 1
3.80 (1.64)
0.10 (0.22)
8.4 (2.4)
Massed 2
3.80 (1.30)
0.20 (0.27)
9.6 (2.7)
Control
3.90 (1.20)
0.10 (0.14)
7.9 (2.3)
004
Distributed
4.10 (0.74)
0.20 (0.20)
7.9 (2.9)
Massed
3.60 (0.97)
0.17 (0.31)
8.1 (3.7)
Massed 1
3.60 (0.55)
0.20 (0.36)
7.2 (1.6)
Massed 2
3.60 (1.34)
0.13 (0.30)
9.0 (5.1)
Control
3.60 (1.51)
0.10 (0.18)
8.7 (2.2)
005
Distributed
2.70 (1.16)
0.11 (0.23)
5.2 (2.7)
Massed
2.70 (0.67)
0.11 (0.16)
5.4 (2.9)
Massed 1
2.60 (0.55)
0.06 (0.13)
5.4 (3.0)
Massed 2
2.80 (0.84)
0.16 (0.36)
5.4 (3.1)
Control
2.70 (0.67)
0.12 (0.20)
3.7 (2.2)
008
Distributed
3.30 (1.34)
0.05 (0.12)
7.1 (2.3)
Massed
3.30 (0.95)
0.08 (0.12)
6.6 (3.4)
Massed 1
3.40 (1.14)
0.10 (0.16)
5.0 (1.6)
Massed 2
3.20 (0.84)
0.05 (0.07)
8.2 (4.2)
Control
3.30 (1.42)
0.07 (0.17)
6.5 (2.5)
Page 67 of 67 Journal of Speech, Language, and Hearing Research
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ResearchGate has not been able to resolve any citations for this publication.
Article
Full-text available
Purpose: To investigate the association between physical therapy frequency and gross motor improvement in children with cerebral palsy (CP). Materials and methods: This is a prospective cohort study of 442 children aged 2-12 years, Gross Motor Function Classification System levels I-V, from the Cerebral Palsy Follow-up Program and the Cerebral Palsy Register of Norway. Outcome was change in reference percentiles for the Gross Motor Function Measure (GMFM-66) between two subsequent assessments (N = 1056) analyzed in a linear mixed model. Results: It was a dose response association between physical therapy frequency and gross motor improvement. Mean change was 4.2 (95% CI: 1.4-7.1) percentiles larger for physical therapy 1-2 times per week and 7.1 (95% CI: 2.6-11.6) percentiles larger for physical therapy >2 times per week, compared to less frequent physical therapy when analyzed in a multivariable model including multiple child and intervention factors. The only statistically significant confounder was number of contractures which was negatively associated with gross motor improvement. Conclusions: When gross motor improvement is a goal for children with CP, more frequent physical therapy should be considered. Implications for rehabilitation In general, the gross motor development of Norwegian children with cerebral palsy was as expected according to the reference percentiles for the GMFM-66. When gross motor improvement is a goal for children with cerebral palsy, high-frequency physical therapy should be considered. Contractures should be addressed in order to optimize gross motor improvement for children with cerebral palsy.
Article
Full-text available
Purpose: The purpose of this study was to evaluate the role of practice variability, through prosodic variation during speech sound training, in biofeedback treatment for children with childhood apraxia of speech. It was hypothesized that variable practice would facilitate speech sound learning. Method: Six children ages 8-16 years with persisting speech sound errors due to childhood apraxia of speech participated in a single-subject experimental design. For each participant, 2 speech sound targets were treated with ultrasound visual feedback training: one with prosodic variation (i.e., practicing sound targets in words and phrases spoken fast, slow, loud, as a question, command, and declarative), and one without prosodic variation. Each target was treated for half of the 1-hr session for 14 treatment sessions. Results: As measured by standardized effect sizes, all participants showed greater change on generalization probes for sound targets treated under the prosodic variation condition with mean effect sizes (d2) of 14.5 for targets treated with prosodic variation and 8.3 for targets treated without prosodic variation. The average increase in generalization scores was 38% in the prosodic variation condition compared to 31% without. Conclusions: Ultrasound visual feedback may facilitate speech sound learning and learning may be enhanced by treating speech sounds with explicit prosodic variation. Supplemental materials: https://doi.org/10.23641/asha.5150119.
Article
Full-text available
Ultrasound imaging is an adjunct to traditional speech therapy that has shown to be beneficial in the remediation of speech sound errors. Ultrasound biofeedback can be utilized during therapy to provide clients with additional knowledge about their tongue shapes when attempting to produce sounds that are erroneous. The additional feedback may assist children with childhood apraxia of speech (CAS) in stabilizing motor patterns, thereby facilitating more consistent and accurate productions of sounds and syllables. However, due to its specialized nature, ultrasound visual feedback is a technology that is not widely available to clients. Short-term intensive treatment programs are one option that can be utilized to expand access to ultrasound biofeedback. Schema-based motor learning theory suggests that short-term intensive treatment programs (massed practice) may assist children in acquiring more accurate motor patterns. In this case series, three participants ages 10–14 years diagnosed with CAS attended 16 h of speech therapy over a 2-week period to address residual speech sound errors. Two participants had distortions on rhotic sounds, while the third participant demonstrated lateralization of sibilant sounds. During therapy, cues were provided to assist participants in obtaining a tongue shape that facilitated a correct production of the erred sound. Additional practice without ultrasound was also included. Results suggested that all participants showed signs of acquisition of sounds in error. Generalization and retention results were mixed. One participant showed generalization and retention of sounds that were treated; one showed generalization but limited retention; and the third showed no evidence of generalization or retention. Individual characteristics that may facilitate generalization are discussed. Short-term intensive treatment programs using ultrasound biofeedback may result in the acquisition of more accurate motor patterns and improved articulation of sounds previously in error, with varying levels of generalization and retention.
Article
Full-text available
One challenging area of speech-language pathology is evaluating treatment change in children with speech sound disorder (SSD) with a motor basis. A clinician’s knowledge and use of outcome measures following treatment are central to evidence-based practice. This narrative review evaluates the use of outcome measures to assess treatment change in motor-based SSDs. Seven databases were searched to identify studies reporting outcomes of treatment in SSDs between 1985 and 2014. Sixty-six studies were identified for analysis, and reported outcome measures were categorized within the International Classification of Functioning, Disability and Health framework (ICF-CY). The majority of studies used perceptual methods (despite their limitations) to evaluate change at the impairment level of the ICF-CY and only three studies examined participation level factors. Accurate outcome measures that reflect the underlying deficit of the SSD as well as activity/participation level factors need to be implemented to document intervention success in this population.
Book
An intervention approach for children with moderate to severe speech sound disorder. The BSQC kit contains a manual explaining the phonetics-based intervention approach and the Index of Phonetic Complexity measure for quantifying speech skills, in addition to over 800 stimulus cards.
Article
Purpose: Treatment intensity (sometimes referred to as ‘‘practice amount’’) has been well-investigated in learning non- speech tasks, but its role in treating speech disorders has not been largely analysed. This study reviewed the literature regarding treatment intensity in speech disorders. Method: A systematic search was conducted in four databases using appropriate search terms. Seven articles from a total of 580 met the inclusion criteria. The speech disorders investigated included speech sound disorders, dysarthria, acquired apraxia of speech and childhood apraxia of speech. All seven studies were evaluated for their methodological quality, research phase and evidence level. Result: Evidence level of reviewed studies ranged from moderate to strong. With regard to the research phase, only one study was considered to be phase III research, which corresponds to the controlled trial phase. The remaining studies were considered to be phase II research, which corresponds to the phase where magnitude of therapeutic effect is assessed. Results suggested that higher treatment intensity was favourable over lower treatment intensity of specific treatment technique(s) for treating childhood apraxia of speech and speech sound (phonological) disorders. Conclusion: Future research should incorporate randomised-controlled designs to establish optimal treatment intensity that is specific to each of the speech disorders.
Article
Three previous letters written to the parent(s) of children with developmental apraxia of speech (DAS) described the speech characteristics of DAS, the nature and causes of the disorder, and problems that often co-occur with DAS. This final letter discusses the treatment DAS and includes an appendix or publications for a are complete review of this aspect of the problem.
Article
This study employed a single subject, multiple baseline across behaviors design to examine the effects of dynamic temporal and tactile cueing (DTTC). The participant is a 12-year-old male with CHARGE association, a complex genetic disorder. He exhibited moderate mental retardation and severe childhood apraxia of speech (CAS). He was nonverbal and exhibited communicative vocalizations that drew negative attention. Treatment was implemented to extinguish aberrant vocalizations, improve his ability to produce accurate movement gestures for a core set of functional words and phrases, reduce rate, and improve comprehensibility. Treatment was provided over three phases. Results are reported for the effects of treatment focused on establishing a core functional vocabulary as well as for pre- and post changes in the number of aberrant vocalizations and speech rate. Changes in his ability to be understood in context are also reported. Our data provide evidence that improvement in functional verbal communication can occur even in children with severe apraxia who are over 12 years of age. Discussion focuses on the need for children with severe apraxia to continue in treatment, sometimes for long periods of time, before the full effects of therapy are seen. Issues related to measures of disability and societal limitation and the need for more treatment efficacy research in children with severe speech disorders are addressed.
Article
The abstract for this document is available on CSA Illumina.To view the Abstract, click the Abstract button above the document title.