Content uploaded by Yvonne van Zaalen-Op't Hof
Author content
All content in this area was uploaded by Yvonne van Zaalen-Op't Hof on Jan 30, 2014
Content may be subject to copyright.
1
Yu shu zhang ai Tartajeo
Tartagliare Taquifemia
Bredouillement Клáттеринг
Hadard Bruddeln Kleteringas
Løpsk tale إما ج Broddelen
Poltern
พูดรัวเร็ว
Brebtavost
Cluttering Groyok
Sokellus
Cluttering identified
Differential diagnostics between cluttering, stuttering
and speech impairment related to learning disability
Yvonne van Zaalen – op ’t Hof
Cluttering identified
Differential diagnostics between cluttering, stuttering
and speech impairment related to learning disability
ISBN:
978-90-76912-96-7
Copyright © 2009 Yvonne van Zaalen-op ’t Hof
Cluttering identified: differential diagnostics between cluttering, stuttering and speech
impairment related to learning disability.
All rights reserved. No part of this publication may be reproduced in any form or by any
means, electronically, mechanically, by print or otherwise without written permission of the
copyright owner.
Cluttering identified
Differential diagnostics between cluttering, stuttering and speech
impairment related to learning disability
Broddelen geïdentificeerd
Differentiaal diagnostiek tussen broddelen, stotteren en
spreekstoornissen gerelateerd aan leermoeilijkheden
(met een samenvatting in het Nederlands)
PROEFSCHRIFT
ter verkrijging van de graad van doctor aan de Universiteit Utrecht
op gezag van de rector magnificus,
prof. dr. J.C. Stoof,
ingevolge het besluit van het college voor promoties
in het openbaar te verdedigen op
dinsdag 17 november 2009
des ochtends te 10.30 uur
door Yvonne op ‘t Hof
geboren op 25 februari 1966
te Haarlem
Promotoren: Prof. dr. Ph. Dejonckere
Prof. dr. F.N.K. Wijnen
Aan Bertram
Voor Desirée, Anouschka, Lloyd en Giorgio
7
CONTENTS
CHAPTER 1 Introduction and thesis outline 9
PART one
Differential diagnostic characteristics
21
CHAPTER 2 Differential diagnostics between cluttering and stuttering, part one.
Speech characteristics of persons who clutter, clutter-stutter or
stutter.
23
CHAPTER 3 Differential diagnostics between cluttering and stuttering, part two.
Validation of the Predictive Cluttering Inventory.
43
CHAPTER 4 A test on speech motor control on word level productions:
The SPA Test (Dutch: Screening Pittige Articulatie).
57
CHAPTER 5 Language planning disturbances in children who clutter or
have speech impairment related to learning disability.
71
PART two
Underlying neurolinguistic processes
97
CHAPTER 6 Cluttering and stuttering: different disorders.
A neuroimaging study.
99
CHAPTER 7 General discussion
125
Summary 151
Samenvatting (summary in Dutch)
List of abbreviations
Dankwoord (acknowledgements in Dutch)
Publications
Curriculum vitae
8
9
CHAPTER One
Introduction and thesis outline
10
Introduction
In spontaneous speech no one is perfectly fluent. Even the most eloquent speaker suffers
from speech failures every now and then. Probably, most of us make these mistakes more
often than we actually want to. Different kinds of speech failures exist. For instance, we can
insert pauses, or add interjections (‘well’) or meaningless sounds (‘uh’) to gain time. It is
also possible to restart a sentence when we notice that it does not properly express what
we intend to communicate. Word repetitions and “stumbling over one’s words”, or,
technically, difficulties in realizing a word form, are also quite common. In response to
failures like this, some people may say: “Oh, damn, I am stuttering again”. But such
hesitations and sentence or word structure errors are not stuttering. Only when a person
very frequently produces very many of these hesitations and slips of the tongue, in all kinds
of different speaking situations, this may be indicative of a disorder, not stuttering but,
arguably, cluttering.
For a long time, cluttering was the orphan of speech- and language pathology. After the
German Kussmaul (1877) and the Austrian Weiss (1964) drew attention to this remarkable
phenomenon, it was – in Europe in particular – recognized as a specific disorder. Cluttering
did not fit in any other nosological class defined until then and remained poorly understood
until the end of the last century. In the last century a diversity of symptoms were associated
to cluttering. In the United States of America cluttering was not recognized as a disorder
separate from stuttering, till, in the ‘90’s a handful of publications clarified the difference
between stuttering and cluttering (e.g. St. Louis, Raphael, Myers, and Bakker 2003, 2007).
Cluttering is now generally characterized by three main features: (1) a rapid and/or irregular
articulatory rate (Daly, 1993; Damsté, 1984; Dinger, Smit & Winkelman, 2008; St. Louis,
1992; St. Louis, Myers, Cassidy, Michael, Penrod, Litton et al., 1996; St. Louis, Raphael,
Myers & Bakker, 2003; Weiss, 1964); (2) a higher than average frequency of normal (non-
stutter-like) disfluencies, (Myers & Bradley, 1992; St. Louis, 1992,1996; St. Louis et al.,
2003) and (3) reduced intelligibility due to exaggerated coarticulation (deletion of syllables
or sounds in multi-syllabic words) and imprecise articulation (Daly & Burnett, 1999; Damsté,
1984; Dinger et al., 2008; Gutzman, 1893; Mensink-Ypma, 1990; St. Louis et al., 2003; St.
Louis, Raphael, Myers & Bakker, 2007; Van Zaalen & Winkelman, 2009; Voelker, 1935;
Ward, 2006; Weiss, 1964). A remarkable characteristic of persons who clutter (PWC),
already described by Kussmaul in 1887, is that their speech production problems diminish
11
when they pay attention to their speech production or when they slow down their speaking
rate.
Aetiology
Weiss's (1964) often quoted definition considers cluttering to be the verbal manifestation of
an underlying central language imbalance. As ‘language’ as a cognitive faculty involves
many components and processes, one may ask what an ‘imbalance’ in this system would
entail. According to Myers (1992), the imbalance is to be viewed as a problem in
synchronizing various components of the language system during utterance production:
“The propositions, pragmatic intent, meaning, sequencing, phrasing, timing, articulation,
and orchestration….” of speech and language functions “…require synergy and synchrony
of function and form.” (Myers, 1992, p.175). When one or more of the concomitant
functions go awry (e.g., speaking at a rate that is faster than the individual can handle,
producing an intent or meaning that cannot be readily coded), symptoms of cluttering
surface. According to Myers (1992), cluttering may ultimately be looked upon as a disorder
of timing both for the production of speech and language units. It has been suggested by
many researchers that this disorder is – ultimately – based in a neurological deficit. Alm
(2007) proposes that the problem in adjusting speaking rate in PWC is due to an inhibition
problem in the basal ganglia system. Further research on underlying neurolinguistic
processes like the role of the basal ganglia circuits in cluttering is needed to confirm this
hypothesis.
Prevalence
Pure cluttering is supposed to occur in 5-16% of disfluent speakers (Bakker et al., 2005;
St.Louis & Mc Caffrey, 2005) and 21-67% of PWS also show cluttering characteristics
(Preus, 1992). In the adult population 1-2 % is considered to be a disfluent speaker,
whereas 5% of the children (aged 2-9 years) are diagnosed as disfluent. Cluttering may be
more prevalent than the literature suggests, with cluttering and stuttering-cluttering almost
as prevalent as stuttering (St. Louis & McCaffrey, 2005). The co-morbidity of stuttering and
cluttering is high. Weiss (1964) was even doubtful of the existence of pure stuttering and
argued that all stuttering is based on cluttering. Published prevalence and incidence rates
for cluttering were not based on the current working definition of cluttering (St. Louis et al.,
2007) and should therefore be used with caution.
12
Treatment
Systematic studies on the efficacy of therapy for PWC are virtually nonexistent in the
literature. Possibly as a side effect of the diagnostic problems, there is little consensus
regarding appropriate intervention techniques. The focus of therapy for people who clutter
is on strategies to improve rate control and intelligibility, language production and
monitoring skills (Daly, 1996; Mensink, 1990; Myers, 1996; Van Zaalen & Winkelman,
2009; Weiss, 1964; Ward, 2006). Less often cognitive restructuring and social changes (for
instance, change of profession) are mentioned as part of intervention programs in cluttering
(Daly, 1996; Winkelman, 1990; Van Zaalen & Winkelman, 2009). Changing behaviour,
especially speech behaviour is very difficult. Therefore, treatment sessions with a cluttering
client should be held frequently and at short intervals in order to be effective (Winkelman,
1990). Issues concerning treatment will not be addressed in this study, the focus will be on
(differential) diagnosis and underlying neurocognitive processes.
Diagnosing cluttering
“One of the problems in diagnosing and treating cluttering is that it often occurs in
conjunction with other disorders, some of which are speech-language based and others
that are not” (Ward, 2006, p. 359). Differential diagnostics between cluttering and stuttering
is difficult because these disorders have similar characteristics and often occur in
conjunction with each other. Differential diagnostics in cluttering and stuttering has been
based predominantly on subjective clinical judgments. Clinical judgment in the assessment
of cluttering should be based on different aspects of communication and cognition, such as
speech rate, intelligibility or fluency (Myers, 1996; Sick, 2004; St. Louis et al., 2003, 2007;
Van Zaalen, Myers, Bennett and Ward, 2007; Ward, 2006).
Objective norms for speech and language characteristics and results on
questionnaires developed especially for cluttering are needed to complement and support
subjective clinical assessments. In addition to improving diagnostic reliability, it is assumed
that by formulating objective criteria more light can be shed on neurolinguistic processes
underlying different forms of speech disfluency.
Cluttering and language
The widely accepted working definition of cluttering by St. Louis, Raphael, Myers and
Bakker (2007) describes cluttering as a fluency disorder characterized by rate
abnormalities, but does not refer to language impairments in PWC. However, hypotheses
13
stressing a central role of (high level) linguistic processes in cluttering have a long history.
Weiss (1964, p.1) assumed that cluttering was the manifestation in speaking of a ‘Central
Language Imbalance’, a disorder affecting all channels of verbal communication as well as
some other, non-verbal skills. Earlier still, Freund (1952) and Luchsinger (1963) identified
linguistic components in cluttering when they characterized the disorder as a ‘dysphasia-
like’ disability. Grewel (1970) observed that cluttering is often found in children with a
delayed speech and language development. The linguistic attributes of cluttering were also
noted by Van Riper (1982) when he included linguistic anomalies next to articulatory rate
variations in his “track II stuttering” (stuttering with a strong cluttering component). Damsté
(1984) had a similar approach, when he described ‘dysphasia-like cluttering’. However,
despite the fact that these authors and more recent studies (St. Louis, 1992; Daly, 1992;
Ward, 2004) also pointed to the importance of language difficulties in cluttering, research
on the language skills of PWC has thus far not yielded more than vague and broad
descriptions, such as “problems in retelling a story,” (Mensink, 1990) or “a limitation in
language formulation” (St. Louis, 1992) and “disorganized language formulation” (Daly,
1996).
It is assumed here that more knowledge on the language component in cluttering
can be gained by studying the formulation skills of PWC at different speaking rates and
underlying lexical complexity conditions. It is hypothesized that PWC do not exhibit a
language disorder, but do exhibit (transient) language formulation difficulties that are
induced and/or exacerbated by an abnormally high or a highly variable speaking rate
(Myers, 1992). A language disorder can be defined as a disorder that affects all kinds of
linguistic information processing, particularly both receptive and expressive tasks. By
contrast, language formulation disturbances affect production only, and are reflected by
specific types of disfluency: hesitations, interjections and sentence and word revisions. A
detailed qualitative and quantitative comparative analysis of language formulating
difficulties in PWC and unaffected controls is assumed to assist in clarifying the nature of
the deficit underlying cluttering.
Cluttering and language/speech impairment related to children with learning disability
The disorder of cluttering provides us with an example of how much speech/language
disorders and learning disabilities can have in common (Gregory, 1995). For Preus (1996),
cluttering has more in common with learning disabilities (LD) than with stuttering. Daly
(1986) claimed that decreased expressive language skills are common characteristics of
14
children who clutter (CWC) or have learning disabilities. Many researchers contend that the
overlap between cluttering and learning disability exists mainly with regard to problems in
expression, reading aloud and writing (Daly & St. Louis, 1986; Mensink-Ypma, 1990; St.
Louis, 1992; St. Louis, Myers, Raphael & Bakker, 2007 in Curlee & Conture, 2007; Tiger,
Irvine & Reis, 1980; Ward, 2006; Weiss, 1964, 1968). Nevertheless, clear descriptions of
the commonalities and differences of disturbances in language production in CWC and LD-
children are lacking in the scientific literature. Research on specific aspects of language
abilities in cluttering and speech impairment related to learning disability has been limited
to merely mentioning problems in language production, as reflected predominantly in a high
occurrence of disfluencies.
Brain imaging
Differential diagnosis of cluttering and stuttering is also important when disfluent
participants are included in research projects. In brain imaging studies participants are
often included on the basis of a stuttering diagnosis based on the percentage stuttered
syllables or a Stuttering Severity Instrument-III score (Riley, 1994), (Blomgren, Nagarajan,
Lee, Li & Alvord, 2003; Giraud, Neumann, Bachoud-Levi, Wolff von Gudenberg, Euler,
Lanfermann & Preibisch, 2008; Smits-Bandstra & de Nil, 2007). In some studies the
presence of other speech language disorders is used as an exclusion criterion (Neumann,
Euler, Wolff von Gudenberg, Giraud, Lanfermann, Gall & Preibisch, 2003). In none of the
brain imaging studies cluttering components are described, neither as inclusion nor as
exclusion criteria. According to Preus (1992) 21-67% of the disfluent population displays
cluttering symptoms. When a cluttering component in stuttering participants of brain
imaging projects is neglected, it is possible that in comparing brain activation of stuttering
participants to controls inconclusive results will be found.
Personal characteristics
Initiated by Weiss (1964) also various cognitive weaknesses, as well as personality traits
are related to cluttering (e.g. poor concentration and attention span, reading disorders,
writing disorders, unawareness of symptoms, restlessness and hyperactivity, impatience,
superficiality, casual acceptance of life, lack of consideration of the consequences of a
given act or for other people, and a short temper that is easily placated). Weiss’ findings
were based on clinical observations and not studied thoroughly. At present it is uncertain if
these characteristics indeed are related to cluttering. It is possible that the (speaking) rate
15
abnormalities in cluttering underly most of the above mentioned personal traits. Although I
am aware that these characteristics are ascribed to cluttering, issues concerning non
speech related behavioural aspects will not be addressed in this thesis.
Underlying neurolinguistic processes
When cluttering and stuttering can be differentiated (by for instance differentiating
symptoms or different responses to intervention) it is reasonable to assume that underlying,
neurolinguistic processes/deficits are also different. It is a challenge to describe the
underlying processes/deficits as precisely as possible and in such a way that the
observations (different symptoms, different responses to intervention) can be explained
adequately and new testable predictions can be derived. Description of underlying
processes/deficits should fit in knowledge of and models for normal production of spoken
language.
Thesis outline
This study has two following objectives: (1) clinical, diagnostic classification of the
syndrome of cluttering and particularly the difference in symptomatology between cluttering
and stuttering and between cluttering and speech problems related to learning disability; (2)
to contribute to a (neurolinguistic) model of cluttering that provides a coherent explanation
for the observed symptomatology; and elucidates the difference between cluttering and
stuttering.
(1) For that purpose diagnostic instruments used in stuttering assessment are
adapted to cluttering and new assessment instruments to identify cluttering are designed
and validated in a large group of disfluent speakers in the age range 6;6 – 50 years. Since
in current clinical practice cluttering is diagnosed and differentiated from stuttering on the
basis of subjective interpretation of articulatory rate (variations), type of speech errors and
type and frequency of normal disfluencies, I will take these symptoms as a starting point.
Consequently, it is hypothesised that objective measurements of articulatory rate,
articulatory rate variation; type and frequency of disfluencies and errors in word or sentence
structure will differ between persons who clutter (PWC) and persons who stutter (PWS). In
determining norms for speech and language characteristics, a deeper understanding of
some of the variables underlying different neurolinguistic processes of fluency disorders will
be acquired.
16
Due to a lack of differential diagnostic criteria between cluttering, stuttering and
language/speech impairment related to a learning disability, cluttering is often detected
later in life, or not at all. This has the undesired result that therapy results are very limited
and communicative skills of affected persons remain poor. It is assumed that in cluttering
rate abnormalities manifest in language formulation disturbances and that these
disturbances diminish when speaking rate is reduced or linguistic demands (complexity)
decrease(s). By contrast, I hypothesize that the language production errors in LD-children
will not be affected by rate or linguistic task in the same way.
(2) Exact objective diagnostic criteria for cluttering will guide formulating the
characteristics of the underlying deficits. A functional MRI study showing differences
between PWC and PWS will assist in corroborating the differences. A theoretical analysis
of language production in cluttering will be given using Levelt’s model of language
production (Levelt, 1989).
Consequently, this thesis is divided in two parts. Part I addresses differential diagnostic
characteristics of cluttering, stuttering and speech impairment related to a learning
disability. Part II addresses underlying neurolinguistic processes in these disorders.
The first chapter of part I (Chapter 2) describes an empirical study aiming to set
objective norms for differential diagnostic assessment of cluttering and stuttering
symptoms, based on the three main characteristics of cluttering proposed by St. Louis,
Raphael, Myers & Bakker (2003): (1) fast and/or irregular articulatory rate together with (2)
errors in syllable, word or sentence structure and/or (3) a high frequency of normal
disfluencies (not being stuttering). Objective measures are compared to the subjective
clinical judgment made by expert fluency therapists. As a result of this work an assessment
protocol differentiating between cluttering and stuttering was developed for use in further
research on cluttering. Chapter 3 describes results of the Predictive Cluttering Inventory
(Daly and Cantrell, 2006) of persons with fluency disorders (participants were children who
clutter, stutter or had speech impairment due to a learning disability) in relation to the
subjective and objective measurements described in Chapter 2. A revised version of the
Predictive Cluttering Inventory checklist is validated to detect cluttering symptoms.
PWC experience difficulties in making themselves understood in conversations, but
many are able to produce correct syllable and word structures in restricted situations
(Weiss, 1964; Damsté, 1984; Bezemer et al., 2006; Ward, 2006; St. Louis et al., 2007; Van
Zaalen & Winkelman, 2009). To produce intelligible syllable or word structures, the speaker
17
must exercise appropriate levels of speech motor control. Results of the study described in
Chapter 2 made clear that in order to differentiate cluttering from stuttering a validated test
on speech motor control in stuttering and cluttering at word level was needed. In Chapter 4
the validity of an assessment instrument specifically designed to assess speech motor
control at the word level, was tested. Such an instrument may enable the speech language
pathologist to differentially diagnose the speech characteristics between PWC and PWS.
PWC differentiate themselves from PWS on speech motor control at word level. The
question arose whether PWC exhibit language production disturbances comparable to
persons with speech impairment associated with a learning disability. Chapter 5 describes
to what extent disturbances in the fluency of language production of children who clutter
might be comparable to, or differ from, those observed in LD-children. A tentative
connection is made with the underlying processes of language formulation. It is
hypothesized that an increase in normal disfluencies and sentence revisions in children
who clutter reflects a different neurolinguistic deficit than that in LD-children.
In Part II the underlying neurolinguistic processes in cluttering and other disorders of
fluency are described and placed into a model of speech and language production.
Chapter 6 describes an fMRI study in which the findings of studies described in Chapter 2-5
will be confronted with brain activation data in persons diagnosed with either pure stuttering
or pure cluttering while producing strings of multisyllabic words. In this fMRI study the
question will be addressed whether PWC and PWS display different neurocognitive
processes when performing speech tasks that call upon increasing demands on speech
motor and linguistic skills.
Finally, in Chapter 7 cluttering is discussed within the framework of Levelt’s (1989)
language production model. Underlying neurolinguistic processes are described in relation
to articulatory rate. In the final discussion an answer is provided to the question if cluttering
is a language based fluency disorder. The thesis ends with a general summary.
18
References
Alm, P. (2007). On the causal mechanisms of stuttering. PhD thesis, University of Lund,
Sweden.
Bakker, K., St. Louis, K. O., Myers, F., & Raphael, L. (2005). A freeware software tool for
determining aspects of cluttering severity. Annual National Convention of the
American Speech Language and Hearing Association, San Diego, CA.
Bezemer, B.W., Bouwen, J., & Winkelman, C. (2006). Stotteren van theorie naar therapie.
Bussum: Coutinho.
Blomgren, M., Nagarajan, S., Lee, J., Li, T. & Alvord, L. (2003). Preliminary results of a
functional MRI study of brain activation patterns in stuttering and nonstuttering
speakers during a lexical access task. Journal of Fluency Disorders, 28, 337-356.
Bloodstein, O. (1975). A handbook on stuttering. Chicago: National Easter Seal Society for
Crippled Children and Adults.
Daly, D. (1986). The Clutterer. In K. St. Louis (ed) The Atypical Stutterer: Principles and
Practice of Rehabilitation. New York: Academic Press, 155-192.
Daly, D. (1992). Helping the clutterer: Therapy considerations, in F. Myers and K St. Louis
(eds) Cluttering: A Clinical Perspective, Leicester, England: FAR Communications,
107-124.
Daly, D.A. (1993). Cluttering: Another Fluency Syndrome. In R. Curlee (ed) Stuttering and
Related Disorders of Fluency, New York: Thieme Medical Publishers, Inc.
Daly, D. (1996). The source for stuttering and cluttering. East Moline, IL: LinguiSystems.
Daly, D. A., & Burnett, M. L., (1999). Cluttering: Traditional views and new perspectives. In
R. F. Curlee (Ed). Stuttering and Disorders of Fluency, 222-254. 2nd ed. Thieme,
New York.
Daly, D. A. & Cantrell, R.P. (2006). Cluttering characteristics identified as diagnostically
significant by 60 fluency experts. Second World Congress on Fluency Disorders.
Proceedings.
Damste, P.H. (1984). Stotteren. Utrecht: Bohn, Scheltema & Holkema.
Dinger, T., Smit, M. and Winkelman, C.(2008). Expressiever en gemakkelijker spreken,
Uitgeverij Coutinho, Bussum, the Netherlands.
Freund, H. (1952). Studies in the interrelationship between stuttering and cluttering. Folia
Phoniatrica, 4, 146-168.
Giraud, A.L., Neumann, K., Bachoud-Levi, A.C., Wolff von Gudenberg, A., Euler, H.A.,
Lanfermann, H. & Preibisch, C. (2008). Severity of dysfluency correlates with basal
ganglia activity in persistent developmental stuttering. Brain and Language, 104,
190–199.
Gregory, H. (1995). Analysis and commentary. Language, Speech, and Hearing Services in
Schools, 26, 19-25.
Grewel, F. (1970). Cluttering and its problems. Folia Phoniatrica 22, 301-310.
Gutzmann, H. (1893). Vorlesungen über die Störungen der Sprache und ihre Heilung.
Berlijn.
Kussmaul, A. (1887). Speech disorders, Encyclopedia of the practice of medicine, XIV.
New York: William Wood en Co.
Levelt. W.J.M. (1989). Speaking: from intention to articulation. Cambridge, MA: MIT Press.
Luchsinger, R. (1963). Poltern. Berlin-Charlottenburg: Manhold Verlag.
Mensink – Ypma, M. (1990). Broddelen en leerstoornissen, Bohn Stafleu van Loghum,
Houten/Antwerpen.
Myers, F. (1992). In: Myers and St. Louis: Cluttering: A Clinical Perspective, Leicester,
England: Far Communications, 85-105.
19
Myers, F. (1996). Cluttering: A matter of perspective. Journal of Fluency Disorders, 21,
175-186.
Myers, F. L., & Bradley, C. L. (1992). Clinical management of cluttering from a synergistic
framework. In F. L. Myers & K. O. St. Louis (Eds.). Cluttering: A clinical perspective,
85-105. Kibworth, Great Britain: Far Communications.
Neumann, K., Euler, H.A., Wolff von Gudenberg, A., Giraud A.L., Lanfermann H., Gall, V. &
Preibisch, C. (2003). The nature and treatment of stuttering as revealed by fMRI. A
within- and between-group comparison. Journal of Fluency Disorders, 28, 381–410
Preus, A. (1992). Cluttering or stuttering: Related, different or antagonistic disorders. In F.L.
Myers & K.O. St. Louis (Eds.), Cluttering: A clinical perspective. Kibworth: Far
communications.
Preus, A. (1996). Cluttering upgraded. Journal of Fluency Disorders, 21, 349-358.
Riley, G. D. (1994). SSI-3Stuttering severity instrument for children and adults (3
rd
ed.).
Austin, TX: Pro-ed.
Sick, U. (2004). Poltern, Theoretische Grundlagen, Diagnostik, Therapie. Stuttgart
(Thieme).
Smits-Bandstra, S., De Nil, L.F. (2007). Sequence skill learning in persons who stutter:
Implications for cortico-striato-thalamo-cortical dysfunction, Journal of Fluency
Disorders, 251-278.
St. Louis, K. O. (1992). On defining cluttering. In F. L. Myers & K. O. St Louis (Eds.).
Cluttering: A clinical perspective (pp. 37-53). Kibworth, Great Britain: Far
Communications.
St.Louis, K.O., McCaffrey, E., (2005). Public Awareness of Cluttering and Stuttering:
Preliminary Results; Poster Presented at the 2005 ASHA Convention, San Diego,
CA, November 18
St.Louis, K.O. and Myers, F.L. Cassidy, L.J., Michael, A.J., Penrod, S.M., Litton, B.A., et al
(1996). Efficacy of delayed auditory feedback for treating cluttering: Two case
studies. Journal of Fluency Disorders, 21, 305-314.
St. Louis, K.O., Raphael, L. J., Myers, F.L., & Bakker, K. (2003). Cluttering Updated. The
ASHA Leader, ASHA, 4-5, 20-22.
St. Louis, K.O., Myers, F. L., Bakker, K. & Raphael, L. J. (2007). In Conture, E. and Curlee,
R. (Eds.). (2007). Stuttering and Other Fluency Disorders, (3
rd
Ed.). Philadelphia,
PA: Thieme Medical.
Tiger, R.J., Irvine, T.L., & Reis, R.P. (1980). Cluttering as a complex of learning disabilities,
Language, Speech and Hearing Services in Schools, 11, 3-14.
Van Riper, C. (1982). The nature of stuttering. Englewood Cliffs, NJ: Prentice Hall.
Van Zaalen, Y., Myers, F., Ward, D., & Bennet, E. (2008). The cluttering assessment
protocol. Retrieved August 25
th
, 2007 from http://associations.missouristate.edu/ICA/
Van Zaalen, Y., & Winkelman, C. (2009). Broddelen, een (on) begrepen stoornis. Bussum:
Coutinho.
Voelker, C.H. (1935). The Prevention of Cluttering, The English Journal, 24, 808-810, 10,
December.
Ward, D. (2006). Stuttering and cluttering, Frameworks for understanding and treatment,
East Sussex: Psychology Press.
Weiss, D.A. (1964). Cluttering. Englewood Cliffs, NJ: Prentice-Hall.
Weiss, D.A. (1968). Cluttering: Central language imbalance. Pediatric Clinics of North
America, 15, 705-720.
Wigg, E.H. & Semel, E.M. (1984). Language Assessment and Intervention for the Learning
Disabled. Second Edition, Columbus, OH: Charles E. Merrill.
20
Winkelman, C.L. (1990). Broddelen, in Mensink: Broddelen en Leerstoornissen, Bohn,
Scheltema & Holkema. Utrecht.
21
PART one
Differential diagnostic characteristics
22
23
Chapter Two
Differential diagnostic characteristics between cluttering and stuttering – part one:
Speech characteristics of persons who clutter, clutter-stutter or stutter.
A slightly adapted version of:
Y. van Zaalen – op ’t Hof, F. Wijnen and P.H. DeJonckere
(2009c)
Journal of Fluency Disorders, in press, DOI: 10.1016/j.jfludis.2009.07.001.
24
Abstract
Speech-language pathologists generally agree that cluttering and stuttering represent two
different fluency disorders. Differential diagnostics between cluttering and stuttering is
difficult because these disorders have similar characteristics and often occur in conjunction
with each other. This paper presents an analysis of the differential diagnostic
characteristics of the two disorders, and a proposal for distinguishing between the two in
clinical settings.
The main goal of this article is to set objective norms for differential diagnostic
assessment of cluttering and stuttering symptoms, based on the three main characteristics
of cluttering indicated/identified by St. Louis, Raphael, Myers & Bakker [St. Louis, K.O.,
Raphael, L.J., Myers, F., & Bakker, K. (2003). Cluttering updated, The ASHA leader,
ASHA, 4-5, 20-22]: a fast and/or irregular articulatory rate together with errors in syllable,
word or sentence structure and or a high frequency of normal disfluencies (not being
stuttering). In this article objective measures are compared to the subjective clinical
judgment made by fluency experts. In other words, which characteristics can be found in
the speech profiles of persons who were diagnosed as people who clutter or stutter?
1. Introduction
Speech-language pathologists generally agree that cluttering and stuttering represent two
different fluency disorders. Whereas research into stuttering has increased markedly in the
past century, studies on cluttering remain scarce. “One of the problems in diagnosing and
treating cluttering is that it often occurs in conjunction with other disorders, some of which
are speech/language based and others that are not” (Ward, 2006, p. 359). Differentially
diagnosing between cluttering and stuttering is difficult because these disorders have
similar characteristics and often occur in conjunction with each other. For example, Van
Borsel and Tetnowski (2007) reviewed stuttering patterns in clients with mental retardation
who showed evidence of disfluency patterns, concluding that not all would be considered
stuttering. This paper presents an analysis of the differential diagnostic characteristics of
the two disorders, and a proposal for distinguishing between the two in clinical settings.
Stuttering is a disorder characterized by a high frequency of involuntary interruptions
of the forward flow of speech, regarded by the person who stutters (PWS) as “stutters”,
which are often accompanied by a feeling of loss of control (Curlee & Conture, 2007;
Guitar, 2006; Quesal, 2004; Shapiro, 1999; Van Borsel & Tetnowski, 2007; Ward, 2006).
These interruptions usually take the form of (1) repetitions of sounds, syllables or one
25
syllable words; (2) prolongations of sounds; (3) blocks of airflow or voicing in speech. The
results of an exploratory study on speech motor practice and learning by Namasivayam
and Van Lieshout (2008) indicated that PWS and persons who do not stutter may resemble
each other on a number of performance variables (such as movement amplitude and
duration), but they differ in terms of practice and learning on variables that relate to
movement stability and strength of coordination patterns.
In the last decades of the previous century research on cluttering has addressed
overt as well as covert symptomatology (Daly, 1996; Daly & Burnett, 1999; Myers &
Bradley, 1992; St. Louis, 1992; Weiss, 1968). Weiss (1964) described cluttering as a
disorder in the fluent flow of communication. According to experts, cluttering is
characterized by three main features: (1) a rapid and/or irregular articulatory rate (Daly,
1993; Damsté, 1984; Dinger, Smit & Winkelman, 2008; St. Louis, 1992; St. Louis, Myers,
Cassidy, Michael, Penrod, Litton et al., 1996; St. Louis, Raphael, Myers & Bakker, 2003;
Weiss, 1964); (2) a higher than average frequency of disfluencies, dissimilar to those seen
in stuttering [see section on stuttering characteristics above], (Myers & Bradley, 1992; St.
Louis, 1992,1996; St. Louis et al., 2003) and (3) reduced intelligibility due to exaggerated
coarticulation (deletion of syllables or sounds in multi-syllabic words) and indistinct
articulation (Daly & Burnett, 1999; Damsté, 1984; Dinger et al., 2008; Gutzman, 1893;
Mensink-Ypma, 1990; St. Louis et al., 2003; St. Louis, Myers, Bakker & Raphael, 2007; Van
Zaalen & Winkelman, 2009; Voelker, 1935; Ward, 2006; Weiss, 1964).
1.1 Rapid and/or irregular articulatory rate
According to the St. Louis et al., (2003) working definition of cluttering a high and/or
irregular articulatory rate is a main characteristic in differential diagnostics between
cluttering and stuttering, however, agreement on what defines abnormally fast and
abnormally irregular articulatory rate is needed. It is hypothesized that there are persons
who clutter who maintain a high articulatory rate in a more demanding speaking situation,
and their speech-language system can not handle such fast speed. Due to speech motor or
language planning problems in a high articulatory rate, intelligibility problems or disfluencies
occur (Daly, 1992).
1.2 Intelligibility and articulatory accuracy
Many researchers and clinicians (Bezemer, Bouwen, & Winkelman, 2006; Ward, 2006)
report that people who clutter experience intelligibility problems due to exaggerated
26
coarticulation (deletion of sounds or syllables in multi-syllabic words) and indistinct
articulation (substitution of sounds and/or syllables). Several researchers discuss the fact
that although persons who clutter experience intelligibility problems in running speech, they
are able to produce correct syllable and word structures in controlled situations (Damsté,
1984; Van Zaalen & Winkelman, 2009; Ward, 2006; Weiss, 1964). The findings of
Hennessey, Nang, and Beilby (2008) suggest that contrary to persons who clutter (PWC),
PWS were not deficient in the time course of lexical activation and selection, phonological
encoding, and phonetic encoding. In order to be able to produce correct syllable or word
structures, speech motor control should be appropriate. Riley and Riley (1985) defined
speech motor control as the ability to time laryngeal, articulatory and respiratory
movements, that lead to fast and accurate syllable production. It is hypothesized that
cluttering is a fluency disorder in which speech motor control at word level is disturbed in
high speech rate, resulting in errors in word structure.
1.3 Frequency and type of disfluencies
Relying upon clinical experience, St. Louis, Hinzman, and Hull (1985) and St. Louis et al.,
1996 differentiated the fluency disorders of disfluent people and concluded that PWC had a
high frequency of normal disfluencies (e.g. revisions, interjections, phrase- and syllable
repetitions) and a low frequency of disfluencies typical for stuttering. A higher than average
frequency of disfluencies, dissimilar to those seen in stuttering is considered to be a
characteristic of cluttering.
1.4 Subjective clinical judgment
Differential diagnostics in cluttering and stuttering has up till now mainly been based on the
subjective clinical judgment of the speech-language therapist. Clinical judgment in the
assessment of cluttering and stuttering should be based on different aspects of
communication and cognition, for instance oral reading aloud, spontaneous speech,
retelling a memorized story, and questionnaires (Sick, 2004; St. Louis et al., 2003, 2007;
Ward, 2006). It would appear important to develop a more objective assessment method for
the above mentioned aspects besides the subjective clinical judgment.
The main goal of this article is to set objective norms for differential diagnostic
assessment of cluttering and stuttering symptoms, based on the three main characteristics
of cluttering indicated/identified by St. Louis et al., (2003): a fast and/or irregular articulatory
rate together with errors in syllable, word or sentence structure and or a high frequency of
27
normal disfluencies (not being stuttering). In this article, objective measures are compared
to the subjective clinical judgment made by fluency experts. In other words, which
differentiating characteristics can be found in the speech profiles of persons who were
diagnosed as people who clutter or stutter?
2. Method
2.1 Participants
All persons participating in this study had been referred to centres for stuttering therapy in
the centre of the Netherlands (between January 2006 and January 2007) with self-reported
fluency problems. Participants were 41 males (mean age 10.2; range 6–39 years) and 13
females (mean age 12.9; range 6-47 years). Controls for disfluent adolescents and adults
were 17 males and 8 females (see Table 1.). A control group was included in order to
obtain normative values for articulatory rate in retelling a memorized story and scores on
speech motor control on word level in Screening Pittige Articulatie (SPA, van Zaalen
Wijnen & Dejonckere, 2009a). For the tests used in the younger children no controls were
needed, due to existing normative data. For the older participants, groups were matched for
age and gender.
None of the participants (including controls) reported any neurological or hearing
disorders and all were Dutch speaking mono- or bilinguals with an intermediate to high
educational level. Subjects were tested in the first assessment session prior to therapy or in
case of subjects that were already in the course of treatment (at the most 3 months) prior to
any therapy session.
Males Females Total
N Mean age Range N Mean
age
Range N
Disfluent
41 10.2 6.0 –
39.4
13 12.9 6.3 – 47.2 54
Controls 17 24.3 12.6 –
47.3
8 25.2 12.4 – 52.1
25
Total 58 21 79
Table 1.: Participants divided in gender, mean age and age range
2.2 Diagnostic decision making
Participants were diagnosed based on subjective clinical judgment on audio recordings of
three different speech tasks: spontaneous speech, reading aloud and retelling a story.
28
Diagnostic decisions were separately determined by two speech language pathologists
specialized in fluency disorders (both cluttering and stuttering). Data was blinded and
labelled in code. An independent researcher was in control of the coding system. Both
speech language pathologists (SLPs) were aware of the age and gender of the
participants. SLPs could choose between three diagnostic codes: cluttering (C), stuttering
(S) and cluttering-stuttering (CS). A participant was appointed to a diagnostic group based
on the diagnoses of both SLPs. When SLPs both diagnosed a person as cluttering, the
participant was placed in the PWC group. When SLPs both diagnosed a person as
stuttering, the participant was placed in the PWS group. When a participant was diagnosed
as both cluttering and stuttering, he/she was placed in the PWCS group. When SLPs
disagreed the participant was placed in the undecided group. After analysing the objective
measurements of the PWS, PWCS and PWC group, participants in the undecided group
were diagnosed based on the objective measurement in the three diagnostic groups.
2.3 Speech tasks
Data was gathered on: articulatory rate; articulatory accuracy and smooth-flow frequency
and type of normal disfluencies. The test sequence for all participants
was: (1) monologue;
(2) reading aloud; (3) story retelling; and (4) speech motor coordination.
2.3.1. Task 1: Monologue
Participants were asked to recount an event in the recent past of their own choosing
without intervention of the speech pathologist. Recordings lasted 3 - 5 min.
2.3.2. Task 2: Reading aloud task
This task and the next were adapted to the age / reading skills of the participants. Children
read a standardised story, two levels below their reading level (as assessed in school). This
was assumed to give the investigator a reasonable degree of certainty that not the reading
skills but the speech skills were tested. The adults read a text above childhood reading
level in order to examine complex sentences, multi-syllabic words and the appearance of
more than one person in the story.
2.3.3. Task 3: Story retelling: the bus-story and the wallet-story
For the children under the age 12 we used “the Bus story” (Renfrew, 1997, Dutch version
by Jansonius & Roelofs, 2006). This task is designed for use with children and has been
29
used in several studies. Scoring norms are available. The researcher first tells the child the
story while showing the child a colour book with 12 pictures of the story and accordingly the
child is asked to retell the story. For the adolescents and the adults, we used the Wallet
story”, which was adapted by Van Zaalen and Bochane (2007) from the ABCD test
(Arizona Battery for Communication Disorders of Dementia).
In this study, an analysis of articulatory rate and disfluency measurements was
undertaken. Due to lack of norms on this test, fluent age matched controls were also tested
and means were compared.
2.3.4 Task 4a: Speech motor control on syllable level
Speech motor control can be measured with the Oral Motor Assessment Scale (OMAS,
Riley & Riley, 1985). In a stable elicitation procedure ten repetitions of /puh/, /tuhkuh/ and
/puhtuhkuh/ are obtained. These repetitions are judged on articulatory accuracy, smooth-
flow (co-articulation, flow and sequencing) and rate.
2.3.5. Task 4b: Speech motor control on word level: the SPA test (Van Zaalen et al.,
2009a)
Skills in oral motor coordination in multi-syllabic words were tested with the SPA test
(Dutch: Screening Pittige Articulatie, Van Zaalen et al., 2009a). SPA, designed by the first
author, is a specially designed speech task to provide information on speech motor control
at word level in a fast speech rate. SPA can provide information on retaining correct word
structure and intelligibility and was therefore included in our assessment procedure. In an
elicitation procedure, three repetitions of ten complex multi-syllabic words at a fast speech
rate were obtained. In order to allow a comparison with OMAS results, the SPA elicits three
words containing (a) mostly bilabial onset consonants (similar to [pə]; e.g. Dutch:
[
Ǥ
pərserəmờnimestər], or (b) mostly alveolar and velar onsets (as in [təkə], e.g. Dutch:
[vərǡndərəndə ǽevəns
Ǥ
mstǡndǺxhedən]; (c) a combination of bilabial, alveolar and velar
consonants (as in [pətəkə], e.g. Dutch: [Ǥnœytsprekələk vǫrveləndə vǫrhǡndəǽǺȃən]).
These repetitions are judged on articulatory accuracy, smooth-flow (co-articulation, flow
and sequencing) and rate.
30
2.4. Rate measurements
In order to obtain articulatory rate norms for fast rate, we needed an answer to the
question: ‘What is a fast articulatory rate?’ To our knowledge, there is no clinical or
scientific consensus on this issue. In disfluent speech, speech rate variation can be
influenced by extra or extended pauses. Pindzola, Jenkins, and Lokken (1989) and Hall,
Amir, and Yairi (1999) stated that articulatory rate measures are intended to reflect how
quickly sound segments are produced in stretches of speech that have no pauses. We
decided to base our judgment on Mean Articulatory Rate (MAR), which was defined as the
mean of five rate measures in minimally 10 to maximally 20 consecutive syllables in
perceptually fluent speech without pauses. We defined fast articulatory rate in
spontaneous speech for disfluent speakers as a rate ≥ 1.0 SD above the MAR of disfluent
speakers. It is well known that the fluent speech of PWS is slower compared to age and
gender matched controls. The disfluent people are chosen as a reference group because it
is known that differences in MAR between fluent and disfluent people exist.
Articulatory rate, in stuttering research, is usually calculated in syllables per second
or phonemes per second by analysing only perceptually fluent utterances. Perceptually
fluent utterances are defined as those utterances that exclude “within- or between-word
disfluencies, hesitations, or pauses greater than 250ms” (Yaruss, Logan & Conture, 1994,
p. 221).
In counting syllables one has to decide between the linguistic word form or the
speech motor output. There are two reasons for choosing the linguistic word form in
counting syllables of disfluent persons: (1) as cluttering is often considered to be a disorder
of speech planning, it is important to know how much time the person planned to produce
the word (Verhoeven, de Pauw & Kloots, 2004); (2) PWC sometimes produce unintelligible
speech in which it is difficult to objectively determine how many and which syllables and
phonemes have been realized. To avoid overly subjective assessment, the articulatory rate
was calculated on the basis of the number of syllables that should have been realized [for
the citations forms of words].
In the St. Louis et al., (2003) definition of cluttering deviant rate variability is
considered to be a key symptom. For each subject, articulatory rate variability was
computed as the standard deviation around the mean for the rate measurements. There
are no normative values for MAR-variation (MAR-v). In our study, a deviant-MAR-v was
defined as a variation in articulatory rate ≥ 1.0 SD above the mean articulatory rate
variation: an indicator of cluttering.
31
2.5. Ratio Disfluencies
Evaluation of disfluencies, as a supplement to the diagnostic criteria based on articulatory
rate and articulatory accuracy was done using Campbell and Hill’s Systematic Disfluency
Analysis procedure (Campbell & Hill, 1994). All disfluencies in samples of spontaneous
speech and speech in story retelling were counted. The percentage of stutter-like
disfluencies and normal disfluencies was calculated. The ratio disfluencies was obtained by
dividing the percentage non-stutter disfluencies by the percentage stutter disfluencies, (for
instance a participant with 20% non-stutter disfluencies and 2% stutter disfluencies had a
ratio of 20/2=10) (Van Zaalen & Winkelman, 2009). It is expected that PWC will have a
higher frequency of non-stutter disfluencies and therefore their ratio will be above one,
while the person who stutters will experience more stutter disfluencies and their ratio will be
below one. It is also expected that the mix group (PWCS) ratios will be either below or
above one.
2.6. Articulatory accuracy and coarticulation measurements
In order to classify errors in sound, syllable or word structure, the Oral Motor Assessment
Scale (OMAS; syllable level) and the SPA (Dutch: Screening Pittige Articulatie; word level)
were applied. Scoring was done according to Riley’s Oral Motor Assessment Screening
protocol. Accuracy, smooth flow and rate were scored. Problems in sound accuracy
(distortions and substitutions of voicing and devoicing) were scored. Accuracy scores on
the OMAS or SPA more than 1.5 SD below the group mean were considered as an
indication of severe problems in adjusting voicing to articulatory movement: an indicator of
severe problems in realizing adequate voice-onset time. Problems in smooth flow were
divided into three categories: coarticulation, flow and sequencing. Coarticulation is the
gradual transfer from one speech movement to the other; errors in coarticulation include
telescoping syllables and within sequence pausing. Flow is the gradual stressing and
rhythm of the sequence; errors are for instance changes in stress pattern of the sequence.
Sequencing errors are scored when a person makes errors in sound or syllable order.
Smooth flow scores on the OMAS or SPA more than 1.5 SD below the group mean, are
considered to be an indication of severe problems in oral motor coordination. Rate was
determined by counting the mean time in seconds needed to produce three sequences of
target syllables.
32
A reference test on oral motor coordination on syllable and word level is not
available for ages over 12/13 years. In order to compare the results of the non-fluent
speakers with fluent speakers, results of both adolescents and adults were compared to
results of gender and age matched controls.
2.7. Diagnostic decision based on subjective and objective clinical judgment
The data of the participants with an undecided diagnosis were reanalysed, in order to
determine the value of adding objective values to a subjective clinical judgment. Objective
values obtained from the PWS, PWCS and PWC group, were used as an additional
diagnostic component to diagnose the participants in the undecided group. For instance,
when a participant was diagnosed as PWS by one SLP and PWCS by the other SLP and
this participant scored according to the mean of the cluttering group (on ratio disfluencies or
sequencing scores), the participant was diagnosed as a PWCS. In doing this, we could
objectify the percentage of clients in which the cluttering component was missed by one
SLP.
2.8. Data analysis
Recordings were made in a sound protected room. Digital audio- and video tape recordings
were made of all speaking tasks using a Sony digital video camera, a Trust digital head
microphone, and a GoldWave Digital Audio Editor v5.18.
Articulatory rate was determined using a speech analysis program, PRAAT
(Boersma & Weenink, 2007), (see Figure 2). All of the fluent speech utterances produced
by the individual subjects were recorded through a Trust head microphone into a high
quality sound card using a HP Pavilion zv6000 laptop, sample frequency 22.050 Hz.
Digital audiotakes were edited, replayed and blindly analysed with PRAAT. The
Mean Articulatory Rate (MAR) of five consecutive syllable strings was measured by
counting the number of syllables per second in a string of at least 10 and maximum 20
consecutive syllables spoken, excluding pauses. Durational measures were made in ms by
placing a cursor at the onset and another cursor at the offset. Onset was visually defined as
the first peak (maximum amplitude in millivolts) that corresponded with a burst of spectral
acoustic energy in the corresponding microphone signal or oscillogram. Offset was defined
as the last consecutive peak in the waveform that was followed by a non-speech signal and
also corresponded to the termination of spectral energy. Onset and offset of the utterance
detected in the oscillogram and corresponding spectrographic display of each utterance
33
were verified through playback of the auditory signal. Onsets and offsets of voiceless
consonants that could not be clearly identified were excluded from the analyses, such as in
the Hall, Amir and Yairi (1999) study. Next the duration of each utterance was calculated by
subtracting the onset time from the offset time. Finally, the number of syllables for each
utterance was divided by the duration of the utterance to provide a measure of articulation
rate: syllables per second (SPS).
2.9. Reliability
A random sample of 8 participants was re-diagnosed to evaluate intra-judge reliability. A
paired sample t-test between the two diagnoses was done (p <.05). In addition, a different
random sample of 8 participants was re-analysed by the second experimenter, for
evaluating inter-judge reliability on articulatory rate, articulatory accuracy and smooth flow
measurements. In a paired sample t-test results obtained by the two experimenters were
compared (p < .05).
Diagnosis
PWC PWCS PWS Undecided
Total
Subjective judgment
9
(17%)
10
(18%)
8
(15%)
27
(50%)
54
(100%)
Objective and
subjective judgment
10
(18%)
23
(43%)
9
(17%)
12
(22%)
54
(100%)
Table 2.: Agreement on diagnostic decisions between speech-language pathologists based on
subjective judgment or subjective judgment added with objective measurements.
3. Results
3.1. Diagnostic decision making
Pearson’s correlation between SLP diagnoses was low (r = .638). Of the 54 male and
female disfluent speakers, only 27 (50%) were agreed upon by the SLPs. Of the 54 male
and female disfluent speakers, 7 (13%) were diagnosed as PWC by one researcher and
PWS by the other researcher; and 20 (37%) were diagnosed as PWC or PWS by one
researcher and PWCS by the other researcher. Analyses presented here were carried out
only on the 27 subjects that were agreed upon by the judges (PWC: N=9; PWCS: N=10;
PWS: N=8) (see Table 2.).
34
3.2. Articulatory rate
Table 3 presents mean articulatory rate (MAR) in syllables per second (SPS) on three
speech tasks (spontaneous speech, reading aloud and retelling) presented for each
diagnostic group. Individual subject means for each speech task were calculated. In the
case of two participants no individual mean articulatory rates could be calculated, as they
were not able to produce at least 10 consecutive syllables.
No sex differences were found for MAR in monologue or retelling a memorized story,
however, females were superior to males for reading aloud, [F(1,22)=4.662, p= .047].
Group differences were found for MAR in retelling a memorized story [F(1,22)=8.489, p=
.002]. MAR of PWS was slower (SPS: M=3.7, SD=1.5) compared to PWC (SPS: M=4.9,
SD=0.9) and controls (SPS: M=5.9, SD=0.5) (see Table 3.).
3.3 Fast articulatory rate
Fast articulatory rate (more than 1.0 SD above the MAR) was set at 5.5 SPS in monologue;
5.8 SPS in retelling a memorized story and 5.7 SPS in reading aloud. The majority of PWC
(56%) met the description of fast articulatory rate in spontaneous speech, where PWS did
not. No group differences were found for fast articulatory rate in reading aloud and retelling
a memorized story.
PWC
(N=9)
PWCS
(N=30)
PWS
(N=14)
Controls
(N=25)
M SD M SD M SD M SD F Sig.
MAR monologue 5.3 0.7 4.7 0.8 3.7 0.8
MAR reading 4.5 1.7 4.6 0.7 4.0 1.0
MAR retelling 4.9 0.9 4.8 1.1 3.7 1.5 5.9 0.5
RD Monologue 6.4 3.9 3.2 5.2 0.4 0.5 8.7 .001**
RD Retelling 7.6 4.4 1.7 2.9 2.0 2.0 8.5 .001**
Accuracy SPA 2.1 1.7 1.6 1.3 0.2 0.8 0.1 0.2 11.4 .001**
Smooth flow SPA 8.7 2.0 6.8 2.2 1.4 3.2 0.8 0.9 38.4 .001**
Rate SPA 4.3 0.8 4.8 1.6 8.7 6.1 5.2 0.8
MAR-v 2.5 2.2 2.4 0.1 .884
Table 3.: Speech characteristics and between group analyses of variance, corrected by Tukey’s b for unequal group size, on mean
articulatory rate (MAR), ratio disfluencies (RD) and error scores on Screening Pittige Articulatie (SPA).
35
3.4. Articulatory Rate variability
No group differences were found for Mean Articulatory Rate –variation (MAR-v) (see Table
2). A deviant MAR-v was ≥ 3.99 SPS. A small number of persons who stutter (PWCS:23%;
PWS:15%) fit the description of deviant MAR-v, while PWC did not.
3.4.1. Ratio disfluencies
Between group differences were found in ratio disfluencies for spontaneous speech
[F(1,21)= 34.787, p < .001] and retelling a story [F(1,21)= 16.874, p= .001], but not in
reading aloud [F(1,17)=3.171, p= .094]. PWC produced 6.4 times more normal disfluencies
compared to stutter disfluencies in spontaneous speech and 7.6 in retelling a memorized
story. The Ratio Disfluencies for PWS was 0.4 for spontaneous speech and 1.2 for retelling
a memorized story (see Table 3.).
A ratio of disfluencies above 2.9 is considered to be a cluttering symptom; whilst a
ratio normal disfluencies below 0.9 is considered to be a stuttering symptom. In the PWC
group 75% met the ratio disfluencies criteria for cluttering in both spontaneous speech and
retelling a memorized story. In the PWS group 85.7% met the ratio disfluencies criteria for
stuttering in spontaneous speech (see Fig. 1).
Fig. 1: Ratio disfluencies in the cluttering and stuttering group in
retelling a story, spontaneous speech and reading aloud.
Fig. 2. Smooth flow and accuracy errors in Screening Pittige
Articulatie (SPA)
36
3.5. Articulatory accuracy and smooth flow
PWC produced significantly more (M= 2.1) accuracy errors compared to controls (M= 0.19)
and PWS (M= 0.21) in repeating multi syllabic word strings [F(1,21)=11.386, p< .0001].
Controls produced less smooth flow errors compared to PWS; PWS produced less smooth
flow errors compared to PWC, [F(1,54) =38.413, p< .0001]. Smooth flow scores of PWCS
were between PWS and PWC scores (see Fig. 2).
3.6. Adding objective measures to the subjective diagnostic decision making
When a ratio disfluencies > 2.87, reflecting a cluttering symptom, was added to the
subjective clinical judgment 11 cases out of 54 could be added to the 27 cases SLPs
agreed on diagnosis, with 29.6% of the disfluent cases still remain undecided. In adding
accuracy problems > 2.1 (a cluttering component), to the subjective clinical judgment, the
diagnosis of an extra 9 participants could be confirmed. In adding measures of both ratio
disfluencies and accuracy error scores to the undecided diagnosis made by the subjective
clinical judgment of the SLPs an agreement of 42 out of 54 (77,8%) diagnosis were agreed
upon (see Table 2.).
3.7. Reliability
Intra-judge correlation coefficients on all metrics ranged between .993 < r < .999. Inter-
judge reliability on articulatory rate, articulatory accuracy and smooth flow measurements
ranged between .675 < r < .868.
4. Discussion
The main goal of this research was to compare the subjective clinical judgments made by
experts in fluency disorders to results obtained by objective measurements. Findings
indicate that a differential diagnosis based on a subjective clinical judgment of a speech-
language pathologist specialized in fluency disorders appeared to correspond with the
subjective clinical judgment of another SLP specialised in fluency disorders in only 50% of
all disfluent cases. In 37.0% of all disfluent cases a client was diagnosed as PWC or PWS
by one researcher and PWCS by the other researcher, in other words, one of the SLPs did
not add the cluttering component to the diagnosis. Adding ratio of disfluencies and
accuracy and speech flow error scores on word level to the subjective clinical judgment
37
appeared to be of substantial diagnostic value, especially in locating a cluttering
component. Of all cases, 15.2 % of diagnoses remained undecided after adding
differentiating objective measures to the subjective diagnostic decision.
Overall, the rate data presented in our study are higher than those reported for (American)
English-speakers. This finding corroborates data from earlier research where substantial
rate differences among languages have been observed (Carlo, 2006; Grinfeld & Amir,
2006; Verhoeven et al., 2004). Articulation rate could be affected by linguistic as well as
cultural aspects, thus the establishment of normative bases for mean articulatory rate for
each language is essential. Fast articulatory rate for disfluent speakers, as defined as a
rate ≥ 1.0 SD above the MAR of disfluent speakers, appeared to be of no differentiating
value. One possible explanation of this result might reflect the decision to analyse only
perceptually fluent or intelligible utterances.
In cluttering, a fast articulatory rate is mostly found in those utterances that are not fluent (in
calculating SPS only fluent utterances were included) or in intelligible “spurts”. It would be
important to consider these factors in future studies and find out a way to objectively
measure such accelerated bursts of speech.
Sawyer, Chon, and Ambrose (2008) concluded, based on a single-speech sample in
preschool children who stutter, that influences of rate, length, and complexity were not
significantly correlated to stutter-like disfluencies. Contrary to that, a high amount of normal
disfluencies in combination with a high level of syllable structure errors, can have a
negative influence on the naturalness and intelligibility of speech (Levelt, 1989). Thus,
cluttered speech that is perceived to be fast, may well be within normal limits when
measured objectively.
Ratio disfluencies offer additional diagnostic criteria in retelling a memorized story. Based
on the results of Boey, Wuyts, Van de Heyning, De Bodt, and Heylen (2007) we assume
that results on the ratio of disfluencies can reasonably be used on both Dutch and English
data. It is hypothesized that in retelling a memorized story a person who clutters does not
adjust speech rate to the more complex language level resulting in a high level of normal
disfluencies (word and phrase repetitions and interjections) and sentence structure errors
(Van Zaalen & Winkelman, 2009; Van Zaalen, Wijnen & Dejonckere, 2009b).
38
Errors in speech motor control on word level and a ratio disfluencies above three, appear to
be indicators for cluttering behaviour. Speech situations where a client is not focused on
his/her speech or talks at a fast articulatory rate are sensitive to errors in speech motor
control and disfluencies. In case of cluttering this will mainly be outside the clinic and in
unstructured speech situations. Based on our results we advise SLPs to obtain data both
outside and inside the clinic (both with the conscious knowledge of the client and not).
Although the present research provides ideas of setting normative data and procedures for
differential diagnosis between cluttering and stuttering, the objective measurement values
in this research are based on a small group of disfluent participants that both SLPs agreed
upon. It is recommended that future studies on cluttering and stuttering include multiple
factors or domains in the data collection process, especially with young children during the
formative years of the disorder, when substantial overlap in the development of several
speech/language domains occurs (Yairi, 2007), in order to better understand the intriguing
disorder of cluttering.
39
References
Bezemer, B. W., Bouwen, J., & Winkelman, C (2006). Stotteren van theorie naar therapie.
Bussum, Uitgeverij Coutinho.
Boersma, P. & Weenink, D. (2007). Praat: doing phonetics by computer (Version 4.4.26)
[Computer program]. Retrieved July 24, 2006, from [http://www.praat.org/]
Boey, R.A., Wuyts, F.L., Van de Heyning, P.H., De Bodt, M.S., & Heylen, L. (2007).
Characteristics of stuttering-like disfluencies in Dutch-speaking children, Journal of
Fluency Disorders, 32, 310-329.
Campbell, J. G., & Hill, D. G. (1994). Systematic Disfluency Analysis. Evanston, IL:
Northwestern University.
Carlo, E.J. (2006). Speech rate of non-stuttering Spanish speaking adults, Second World
Congress on Fluency Disorders. Proceedings,111-117.
Curlee, R.F. & Conture, E.G. (2007). Stuttering and related disorders of fluency, (3
e
ED),
Thieme, New York, Stuttgart.
Daly, D. (1992). Helping the clutterer: Therapy considerations, in F. Myers and K St. Louis
(Eds) Cluttering: A Clinical Perspective, Leicester, England: FAR Communications,
107-124.
Daly, D. (1993). Cluttering, another fluency syndrome in Curlee, R.F. Stuttering and related
disorders of fluency (pp. 179-204). Thieme Medical publishers, Inc, New York.
Daly, D. (1996). The source for stuttering and cluttering. East Moline, IL: LinguiSystems.
Daly, D. A., & Burnett, M. L., (1999) Cluttering: Traditional views and new perspectives. In
R. F. Curlee (Ed). Stuttering and Disorders of Fluency (pp. 222-254). 2nd ed.
Thieme, New York.
Damste, P.H. (1984). Stotteren. Bohn, Scheltema & Holkema. Utrecht.
Dinger, T., Smit, M. & Winkelman, C.(2004). Expressiever en gemakkelijker spreken,
Uitgeverij Coutinho, Bussum, the Netherlands.
Grinfeld, D., & Amir, O. (2006). Articulation rate in children and adolescents: Hebrew
speakers. Second World Congress on Fluency Disorders. Proceedings. (pp 125-
129).
Guitar, B., (2006). Stuttering an integrated approach to its nature and treatment, 3rd
edition, Lippincott Williams & Wilkins.
Gutzmann, H. (1893). Vorlesungen über die Störungen der Sprache und ihre Heilung.
Berlijn.
Hall, K. D., Amir, O., & Yairi, E. (1999). A longitudinal investigation of speaking rate in
preschool children who stutter. Journal of Speech Language and Hearing Research,
42, 1367-1377.
Hennessey, N.W., Nang, C.Y., & Beilby, J.M. (2008). Speeded verbal responding in adults
who stutter: Are there deficits in linguistic encoding? Journal of Fluency Disorders,
33, Issue 3, 180-202.
Levelt. W.J.M. (1989) Speaking: from intention to articulation. Cambridge, MA: MIT Press.
Mensink – Ypma, M. (1990). Broddelen en leerstoornissen, Bohn Stafleu van Loghum,
Houten/Antwerpen.
Myers, F. L., & Bradley, C. L. (1992). Clinical management of cluttering from a synergistic
framework. In Cluttering: A clinical perspective (pp. 85-105). K. St. Louis. Kibworth,
Great Britain, Far Communications. (Reissued in 1996 by Singular, San Diego, CA.).
Namasivayam, A.K., & Van Lieshout, P. (2008). Investigating speech motor practice and
learning in people who stutter. Journal of Fluency Disorders, 33, 32-51.
Pindzola, R.H., Jenkins, M.M., & Lokken, K.J., (1989). Speaking Rates of Young Children.
Language, Speech, and Hearing Services in Schools, 20, 133-138 April.
40
Quesal, B. (2004). Fluency and Fluency Disorders. Stuttering course. Retrieved June 1,
2007 from: http://www.mnsu.edu/comdis/kuster/StutteringCourseSyllabi/Quesal.html.
Renfrew, C. (1997). The Renfrew language scales: Bus story test, a test of narrative
speech, Speechmark.
Riley, G.D. and Riley, J. (1985). Oral Motor Assessment and treatment: improving syllable
production. Austin: Pro-Ed.
Sawyer, J., Chon, H., & Ambrose, N.G. (2008). Influences of rate, length, and complexity
on speech disfluency in a single-speech sample in preschool children who stutter,
Journal of Fluency Disorders, 33, 220-240.
Shapiro, D. (1999). Stuttering Intervention: A Collaborative Journey to Fluency Freedom,
Austin, TX: Pro Ed.
Sick, U. (2004). Poltern, Theoretische Grundlagen, Diagnostik, Therapie. Stuttgart
(Thieme).
St. Louis, K. O. (1992). On defining cluttering. In F. L. Myers & K. O. Louis (Eds.),
Cluttering: A clinical perspective. Kibworth, Great Britain: Far Communications, Ltd.
St. Louis, K.O. (1996a). A tabular summary of cluttering subjects in the special edition,
Journal of Fluency Disorders, 21, 337-344.
St. Louis, K.O., Hinzman, A.R., & Hull, F.M. (1985). Studies of cluttering: Disfluency and
language measures in young possible clutterers and stutterers. Journal of Fluency
Disorders, 10, 151-172.
St.Louis, K.O., Myers, F.L. Cassidy, L.J., Michael, A.J., Penrod, S.M., Litton, B.A., et al
(1996). Efficacy of delayed auditory feedback for treating cluttering: Two case
studies. Journal of Fluency Disorders, 21, 305-314.
St. Louis, K. O., Raphael, L. J., Myers, F. L., & Bakker, K. (2003). Cluttering Updated. The
ASHA Leader, ASHA, 4-5, 20-22.
St. Louis, K.O., Myers, F. L., Bakker, K. & Raphael, L. J. (2007). In Conture, E. and Curlee,
R. (Eds.). (2007). Stuttering and Other Fluency Disorders, (3
rd
Ed.). Philadelphia,
PA: Thieme Medical.
Van Borsel, J., & Tetnowski, J.A. (2007). Fluency disorders in genetic syndromes, Journal
of Fluency Disorders, 32, 279-296.
Van Zaalen, Y. & Bochane, M. (2007). “The Wallet story”, 27
th
World congress of the
International Association of Logopedics and Phoniatrics, Proceedings, 85.
Van Zaalen, Y., Wijnen, F., Dejonckere, P.H. (2009a). A test on speech motor control on
word level, the SPA test. International Journal of Speech and Language Pathology,
11:1, 26-33.
Van Zaalen, Y., Wijnen, F., Dejonckere, P.H. (2009b). Language planning disturbances in
children who clutter or have learning disabilities. International Journal of Speech and
Language Pathology, in press.
Van Zaalen, Y., & Winkelman, C. (2009). Broddelen, een (on) begrepen stoornis. Bussum:
Coutinho.
Verhoeven, J., Pauw, G., & Kloots, H. (2004). Speech rate in a pluricentric language: A
comparison between Dutch in Belgium and the Netherlands. Language and Speech,
47, 297-308.
Voelker, C.H. (1935). The Prevention of Cluttering, The English Journal, 24, pp. 808-810,
No. 10, December.
Ward, D. (2006). Stuttering and cluttering, Frameworks for understanding and treatment,
Psychology Press, East Sussex.
Weiss, D.A. (1964). Cluttering. Englewood Cliffs, NJ: Prentice-Hall.
Weiss, D. (1968). Cluttering: Central language imbalance. Pediatric Clinics of North
America, 15, 705-720.
41
Yairi, E. (2007). Subtyping stuttering I: A review, Journal of Fluency Disorders, 32, 165-196.
Yaruss, S., Logan, K., & Conture, E. (1994). Speaking rate and diadochokinetic abilities of
children who stutter. Journal of Fluency Disorders, 19, 221-222, September.
42
43
Chapter Three
Differential diagnostic characteristics between cluttering and stuttering – part two:
Validation of the revised Predictive Cluttering Inventory.
A slightly adapted version of:
Y. van Zaalen – op ’t Hof, F. Wijnen and P.H. Dejonckere
(2009d)
Journal of Fluency Disorders, in press, DOI: 10.1016/j.jfludis.2009.07.001.
44
Abstract
Speech-language pathologists generally agree that cluttering and stuttering represent two
different fluency disorders. Differential diagnostics between cluttering and stuttering is
difficult because these disorders have similar characteristics and often occur in conjunction
with each other. The main goal of this second part of a two-part article is to discuss results
on the Predictive Cluttering Inventory (Daly & Cantrell, 2006). Cluttering characteristics
identified as diagnostically significant by 60 fluency experts. Proceedings of second world
congress on fluency disorders are discussed in relationship to the subjective and objective
measurements studied in the first half of the article and validate a revised version of the
Predictive Cluttering Inventory.
1. Introduction
Van Zaalen, Wijnen, and Dejonckere (2009c) compared objective measures on articulatory
rate, rate variation, type and frequency of disfluencies and scores on speech motor
coordination tasks to the subjective clinical judgment made by fluency experts. This
enabled the identification of speech characteristics that can be found in the speech profiles
of persons who were diagnosed as people who clutter or stutter. In this article, results on
the Predictive Cluttering Inventory (Daly & Cantrell, 2006) are discussed in relationship to
the subjective and objective measurements studied in the first part of the article.
The Predictive Cluttering Inventory (PCI), (Daly, 1996; Daly and Cantrell, 2006) is a
frequently used assessment tool. Daly and Cantrell (2006) called on 60 expert researchers
and clinicians in cluttering worldwide to respond to a questionnaire containing a number of
statements about the disorder. After analysing the data, the checklist contained 33
symptoms associated with cluttering, in four domains (pragmatics, speech motor, language
and cognition & motor coordination and writing problems). Every symptom can be ranked
with a score on a seven-point scale (0 = not present, 6 = always present) in order to predict
possible cluttering. The PCI is produced without a norm for possible cluttering. The aim of
the present study is to correlate PCI data with the characteristics of spontaneous speech
production in disfluent and fluent speakers and validate the PCI as a cluttering detection
instrument. For use in the Netherlands the PCI was translated into Dutch and back
translated into English.
45
It is important to consider that the PCI is based on an email inventory developed by Daly
and Cantrell (2006), where stuttering therapists around the world had to identify
characteristics they assumed to be symptoms of cluttering. The 33-item PCI is the result of
a factor analysis performed on all the characteristics named by the fluency experts. The
PCI is produced without a discriminative norm, so it can be used as an inventory or
checklist rather than a decisive instrument. Adding a discriminative norm could help SLPs
in detecting cluttering more easily.
The PCI was developed to detect cluttering symptoms in running speech. A high
sensitivity level is needed in order to accurately detect cluttering symptoms. A high level of
sensitivity (> 75%) and a high level of specificity (> 75%) are needed in order to make the
PCI a useful differentiating screening instrument. Children with learning disability (LD) were
included, but while they have language disturbances similar to cluttering; they also produce
a high frequency of normal (i.e., non-stutter-like) disfluencies, such as interjections, fillers,
pauses, word- and phrase repetitions (Wigg & Semel, 1984). Fluent children with LD were
hypothesized to score low on the PCI. The main purpose of this research was to investigate
the validity of the PCI in a fluent and disfluent population. In other words is the PCI
sensitive enough to detect cluttering in persons with cluttering symptoms and specific
enough to reject people that do not clutter?
2. Method
2.1. Participants
In this study 137 Dutch speaking participants ranging in age between 10.6 and 12.11
years, were divided into five groups and examined by eight different SLPs. The groups
were based on SLPs diagnosis as described in Van Zaalen et al., (2009c). Age range in all
groups was restricted to exclude bias due to developmental issues. Group 1 consisted of
cluttering children (N=17, M=11.5 years); Group 2 consisted of cluttering-stuttering children
(N=25, M=11.6 years); Group 3 consisted of stuttering children (N=15, M=11.6 years);
Group 4 consisted of children with learning difficulties (N=29, M=11.5 years); and Group 5
consisted of controls (N=51, M=11.2 years). None of the participants had known hearing or
neurological problems. Disfluent children were recruited in two centres for stuttering
therapy in the central and eastern parts of the Netherlands. Children that satisfied the
inclusion criteria for age (from 10.6 to 12.11 years) at the time of the assessment in the
therapy centres were included in the study. Children with learning disability and controls
were recruited from a total of 21 (15 normal education; 6 special education) primary
46
schools in Brabant (part of Holland). All schools have numbered class lists. In every
participating school, six children were randomly selected on the basis of the student
number on the class list. The parents received a consent form with explanation of the
project prior to the assessment. Parents were asked for permission for participation of their
children in the research project (informed consent). All parents gave permission for
inclusion in the study.
2.1.1. Checklist
The Dutch translation of Predictive Cluttering Inventory, “Checklist Broddel Kenmerken”
(CBK, Van Zaalen & Winkelman, 2009) was completed by a group of seven research
assistants on the basis of observation of spontaneous speech, retelling a memorized story,
reading aloud and parental information. All data was coded by an independent researcher
who had access to the coding system. Researchers were SLPs with less than one year
working experience. Completing the checklist was done by the researchers blinded (for
SLP diagnosis). Groups were equally divided between researchers. In order to come
closest to the use by SLPs around the world who download the checklist from the internet
(the available source of the list), researchers were not informed on how to use the checklist
other than by the notification made by Daly and Cantrell (2006) on top of the form.
2.1.2. Norm hypothesized by Daly (2008)
Daly (in press) hypothesized that a norm of 120 points in a 7-point scale (≥3 per item)
would be sufficiently able to detect possible cluttering components in speech, based on
clinical experience with cluttering clients. These hypothetical norms are based on
subjective clinical observation by Daly and research is needed for their validation.
2.1.3 Reference test
While a ‘gold standard’ in cluttering assessment is lacking, we chose to use the subjective
clinical judgment of two highly experienced SLPs specialized in fluency disorders,
combined with objective measures (as described in part one of the article) as the reference
test.
2.1.4. Analyses
Pearson correlations were used to determine relationships between subjective and
objective clinical judgment to the checklist norm studied. Significant differentiating items
47
were analysed. Within group and between group differences were studied, using ANOVA.
The sensitivity level, the proportion of actual positives correctly identified, and the
specificity, the proportion of negatives which are correctly identified, were also determined.
A factor analysis was conducted to determine factors that together may explain the
variance present in the basic variables. Results of the factor analysis were compared to
results from a cluster analysis. The factor analysis grouping was based on the SLPs
diagnosis (cluttering – non-cluttering), while in the cluster analysis, clustering of the items
was conducted based on the individual item scores. In the hierarchical cluster analysis,
between group linkages were displayed with squared Euclidean distance. It is
hypothesized that items detected by both factor and cluster analysis present the best
predicting cluttering items.
Items with significant differentiating value between diagnostic groups were found in
the 33-item list by using an ANOVA with post hoc Tukey’s-b correction for different group
size and a significance level of p < .05. It was hypothesized that significant different items
(for cluttering) with a mean value equal or above three (according to Daly’s hypothesis) can
serve as a predictor of possible cluttering. Supplementary to the total score of
discriminating items, scores on factor one and cluster one items were studied. We
hypothesized that a mean score above 3 (often – always) on all of these items provides a
norm for cluttering.
Before using the checklist, items were not discussed between researchers to make
sure that the use was similar to those therapists who download the file from the Internet.
After completion of all checklists, all items were evaluated between researchers for the
level of clarity. Item clarity analyses was done, by establishing inter judge correlation
scores. Inter judge agreement levels were computed for two items on the checklist. Inter
judge reliability level of >.70 was believed to be acceptable, meaning that both the item and
the scoring system were interpreted similarly by different researchers.
After completion of all checklists, researchers could make comments on items they
thought needed further explanation or clarification. Comments were taken into account in
describing possible alterations of the checklist.
48
3. Results
3.1. Checklist
Based on the proposed norm score for cluttering by Daly (in press), only two participants
from the cluttering or cluttering-stuttering group were detected by the checklist as having
cluttering components (see Table 1.). That is, only 4.44% (2/45) of persons with cluttering
symptoms were tested positive on the index test. This sensitivity level is considered to be
extremely low. The specificity of the checklist is 94.7% (89/94), meaning, that 94.7% of a
group of persons without cluttering indeed tested negative on the index test. The
percentage of false positives, i.e., the percentage of positive tested people that did not
clutter was high: 71.4%. The percentage of false negatives was high: 67.4%.
Reference
norm +
Reference
norm -
Total
Reference
norm +
Reference
norm -
Total
Index test PCI + 2 5 7 Index test
PCI-r +
22 9 31
Index test PCI - 43 89 132 Index test
PCI-r -
10 86 96
Total 45 94 139 Total 32 95 127
Sensitivity 2/45= 4% Sensitivity
22/32= 69%
Specificity 89/94= 95% Specificity
86/95= 91%
False negatives 43/132= 33% False
negatives
10/96= 10%
Table 1. Specificity and sensitivity of total scores on PCI and PCI-revised (PCI-r).
A factor analysis with a Varimax rotation on the cluttering item identified two factors with an
eigenvalue above 1 explaining 85.8% of variance sums of squared loadings: Factor 1
included seven speech planning related items (loading: .61 - .86); factor 2 included six
language structure items (loading: .61 - 83) (see Table 2.).
The cluster analysis with all 33 items on the list entered as variables identified four
major clusters of coherent variables (see Table 3.). Cluster one contained those items
involved in speech planning; cluster two contained items involved in language structure;
cluster 3 contained items involving attentiveness and cluster 4 contained more common
communicative skills.
49
Factor 1 Factor 2
Irregular speech rate; speaks in spurts or
bursts
Language is disorganized; confused
wording; word-finding problems
Rapid rate (tachylalia) Disorganized language increases as topic
becomes more complex
Telescopes or condenses words Poor language formulation; poor story-
telling; sequencing problems
Co-existence of excessive disfluencies and
stuttering
Inappropriate topic introduction,
maintenance, or termination
Initial loud voice trailing off to unintelligible
murmur
Seems to verbalize before adequate thought
formulation.
Little or no anxiety regarding speaking;
unconcerned
Oral diadochokinetic coordination below
expected levels.
Table 2. Items adjusted to factors 1 and 2 according to a factor analyses of the Predictive Cluttering
Inventory (Daly & Burnett, 2006).
An analysis of variance between groups on total cluster scores corrected for uneven group
size by Tukey’s-b revealed significant differences (p= .01) between disfluent (cluttering and
stuttering) and fluent groups (controls and LD) on cluster one: [F(4,134)= 22.975, p<
.0001]; cluster two [F(4,134)= 5.806, p< .0001] and between controls and other diagnostic
groups on cluster three [F(4,134)= 12.961, p< .0001]. There were no significant between
group differences on cluster four.
A closer examination of these results in an analysis of variance on the clusters on
mean cluster scores between disfluent speakers revealed no significant difference on the
three clusters. While differences on total cluster score failed within disfluent speakers, both
cluster two and cluster three showed significant differences between fluent speaking LD
and controls (cluster two: [F(1,78)= 12.146, p< .001]; cluster three: [F(1,78)= 25.230, p<
.0001]).
An analyses of variance between different diagnostic groups corrected for uneven
group size by Tukey’s-b procedure (p=.01) revealed six significant item scores on
cluttering: irregular speech rate, speaks in spurts or bursts; rapid rate (tachylalia); initial
loud voice trailing off to unintelligible murmur; little or no anxiety regarding speaking,
unconcerned; co-existence of excessive disfluencies and stuttering; disorganized language
increases as topic becomes more complex. The first five items mentioned above also
50
appeared in factor 1 and cluster 1. The item “co-existence of excessive disfluencies”
appeared in factor two and cluster two.
Table 3.: Distribution of items in cluster analysis.
3.2. Item clarity
Researchers were asked to comment on the comprehensibility of the checklist. Subjects
and researchers should be able to comprehend the behaviors required to secure accurate
and valid measures. One main concern (noted by all researchers) was uncertainty of how
to interpret the scoring system. For instance, the word “often” can be interpreted as ‘in
almost every speaking situation’ or ‘in almost every sentence within a speaking situation’.
Researchers interpreted this differently. This may have affected final scores.
Some items (item 3, item 6, item 21 and item 22) were considered difficult to score
while a couple of conflicting symptoms were within one item description together. For
instance, item 3 was difficult to score in case of conflicting scores on one part of the item
description. Item 31 was difficult to score because oral diadochokinetic norms for
adolescents and adults are not available and because it was not clear whether coordination
on syllable, word level or during conversation was considered.
A checklist for possible cluttering should be highly sensitive on detecting cluttering.
Therefore an adaptation of the PCI, based on the results of the present study was
Cluster 1.:
Speech planning
Cluster 2.:
Disorganized
language
Cluster 3.:
Attentiveness
Lack of pauses
between words and
phrases; repetition of
multi-syllabic words
and phrases; irregular
speech rate; speaks in
spurts or bursts;
telescopes or
condenses words;
initial loud voice trailing
off to unintelligible
murmur; oral
diadochokinetic
coordination below
expected norm levels;
co-existence of
excessive disfluencies
and stuttering; speech
rate progressively
increases (festinating).
Disorganized language
increases as topic
becomes more
complex; poor
language formulation;
poor story-telling;
sequencing problems
many revisions;
interjections; filler
words; language is
disorganized; confused
wording; word-finding
problems ;
inappropriate topic
introduction,
maintenance, or
termination; Improper
linguistic structure;
poor grammar; syntax
errors
Does not recognize or
respond to listener’s
visual or verbal
feedback; does not
repair or correct
communication
breakdowns; lack of
awareness of own
communication errors
or problems; speech
better under pressure;
lack of effective self-
monitoring skills;
distractible; poor
concentration; attention
span problems
51
analysed. This involved selecting all the items that significantly differentiated cluttering from
stuttering and controls (see Appendix A). A shortened version of the PCI was compiled on
these eight differentiating items. A score of ≥ 3 on (24 points in total) all items of the
shortened PCI resulted in an increase in sensitivity level to 69%. Specificity of the
shortened PCI was very high at 91 % and the percentage false negatives was lowered to
10% (see Table 5.).
4. Discussion
The main purpose of this research was to investigate the validity of the PCI in a fluent and
disfluent population. PCI/CBK in its original form is not sensitive enough to detect cluttering
in persons with cluttering symptoms, but is specific enough to reject people that do not
clutter.
The low sensitivity level of the PCI is equal to frequent type I errors, that is, a
difference was observed when there was none. This may be explained by the fact that a lot
of symptoms on the list are symptoms common to all disfluent speakers. For instance,
every fluent speaker will produce more normal disfluencies compared to cluttering
disfluencies; and weak time planning skills occur in people that do not clutter. Lack of
adequate pausing is common in adolescent speech, but only considered a cluttering
symptom when influencing speech intelligibility. Besides that, some items were expected to
score high in the LD group. For instance, almost every child with learning disability will
score high on attentional focusing problems and weak social skills. A few type II errors
occurred, where there is a failure to observe a difference when in truth there is one. In case
of cluttering, a type II error occurs if the test reports false when the person, in fact, clutters.
The explanation of these errors is that some items are formulated in a way to differentiate
cluttering from stuttering (items 8, 9, 19 and 20). It is expected that persons who only
stutter will score low on this items but the difference between fluent speakers and PWC can
be too small to be statistically significant. In solving these type I and type II errors,
individual item scores were combined to cluster scores on a revised PCI (see Appendix A).
The interpretation of item scores in the revised PCI heightened the sensitivity to a low but
acceptable score of 69%. According to Pollit and Beck (2003), for group-level comparisons,
coefficients in the vicinity of .70 are usually adequate. A sensitivity score of 70% is
acceptable but due to the fact that some items and issues concerning the scoring system
are not absolutely clear to all SLPs a short manual with clarification of items content and
scoring system could further heighten the sensitivity of the PCI. The supporting symptoms
52
are still in the revised PCI as these symptoms can be of great importance in therapy
planning for an individual client (Bezemer et al., 2006; Daly & Cantrell, 2006; Ward, 2006).
In comparing the results of factor and cluster analysis, it was noticed that both
analysing techniques came up with the same two main clusters: speech planning and
language structuring. Another quality to consider in assessing this quantitative instrument is
speed. Researchers were able to complete the PCI within 20-30 minutes. In a relatively
short-time period the SLP is able to collect a substantial important data.
“One significant problem in trying succinctly to identify the characteristics of a clutter
lies in the fact that there may be two basic strands to the disorder; a language component
and a motor one” (Ward, 2006, p.141). The fact that it is common for cluttering to present
more as a language problem than a motoric one, was supported by both factor and cluster
analysis which proposed two major clusters of variables: a speech motor and a language
component. “In case of linguistic cluttering speech output is more likely to show a lack of
linguistic fluency, characterized by poorly constructed language rather than as an output
which is motorically disrupted” (Ward, 2006, p.141), or as Daly described: “in cluttering
accelerated speech is not always present, but an impairment of language formulation
always is” (Daly, 1992, p.107). In cases of motoric cluttering speech output is more likely to
show a lack of speech flow fluency characterized by excessive coarticulation, lack of
speech rhythm, fast bursts of speech interspersed with short inappropriate pauses
(Bezemer et al., 2006; Daly, 1996; Damsté, 1984; Dinger et al., 2008; St. Louis, 1992; St.
Louis et al., 2003, 2007; Ward, 2006; Winkelman, 1990).
As described in the results section of item clarity, researchers reported that the scoring
system was multi-interpretable. It is not clear whether scoring is done on one particular
speech moment (for instance spontaneous speech) or concerns all speaking situations of a
day (both focused and unattended speech) and what a particular score means (‘always’ =
every day, every speaking situation or every word). It is known that speech and language
disturbances of fluent speakers can differ between speaking situations and themes. In
PWC the differences between speaking situations can be very large; especially the
difference in speaking situations when attention to speech is given and those when a PWC
is not alert to his/her speaking performance (Bezemer et al., 2006; Daly, 1992; Damsté,
1984; Dinger et al., 2008; Mensink, 1990; St. Louis et al., 2003; Ward, 2006; Weiss, 1964,
1968). This difference can partially be scored in item ten: Speech better under pressure
53
(improves short-term with concentration), but sensitivity of the PCI would benefit from a
more prominent place of this important symptom in cluttering.
Some item formulation more than likely needs to be reconsidered as many contain
conflicting elements or researchers had problems in dealing with the negative statement in
relation to the scoring system. However, using this revised scoring interpretation system
produced improved sensitivity and specificity. While a number of type I errors still occurred
in the revised PCI, it is not meant to be a diagnostic instrument on its own, though it has the
potential to make contribution to the SLP working with disfluent persons.
In conclusion, the PCI in its current state does not serve as a valid diagnostic
instrument for cluttering, but it serves as a valid screening instrument for possible cluttering
symptoms. In its current state it does differentiate between fluent and disfluent speakers,
but a differentiation between different fluency problems can not be based on total PCI
scores only. Although the revised PCI subtotal score did not differentiate between cluttering
and stuttering, the revised PCI could be of value in the prediction of cluttering components
in speech. Further research on defining items is required in order to make this screening
instrument a valid diagnostic tool that can be used by SLPs.
54
Appendix A
PREDICTIVE CLUTTERING INVENTORY (PCI)-revised
Original by Daly and Cantrell (2006); revised version by Van Zaalen et al.,
(2009)
INSTRUCTIONS to SLP: Please respond to each description section below. Circle the number you believe is the common
most descriptive of this person's cluttering during the day. Count the scores of the itilized items in each section.
5.Always
4.Almost Always
3.Frequently
2.Sometimes
1.Almost Never
0.Never
Section 1: Speech motor
1
Lack of pauses between words and phrases
2
Repetition of multi-syllablic words and phrases
3
Irregular speech rate; speaks in spurts or bursts
4
Telescopes or condenses words
5
Initial loud voice trailing off to unintelligible murmur
6
Oral diadochokinetic coordination below expected normed levels
7
Rapid rate (tachylalia)
8
Co-existence of excessive disfluencies and stuttering
9
Speech rate progressively increases (festinating)
10
Poor planning skills; misjudges effective use of time
11
Little or no excessive effort observed during disfluencies
12
Poor planning skills; mis-judges effective use of time
13
Articulation errors
Section 2: Language planning
14
Disorganized language increases as topic becomes more complex
15
Poor language formulation; poor story-telling; sequencing problems
16
Language is disorganized; confused wording; word-finding problems
17
Many revisions; interjections; filler words
18
Inappropriate topic introduction, maintenance, or termination
19
Improper linguistic structure; poor grammar; syntax errors
20
Variable prosody; irregular melody or stress pattern
Section 3: Attentiveness
21
Does not recognize or respond to listener’s visual or verbal feedback
22
Does not repair or correct communication breakdowns
23
Lack of awareness of own communication errors or problems
24
Speech better under pressure (improves short-term with concentration)
25
Distractible; poor concentration; attention span problems
26
Attention span problems;
27
Seems to verbalize before adequate thought formulation
28
Little or no anxiety regarding speaking; unconcerned
Section 4: Motor and planning (describe these symptoms compared to
age level norms)
29
Clumsy and uncoordinated; motor activities accelerated or impulsive
30
Writing includes omission or transposition of letters, syllables, or words
31
Poor motor control for writing (messy)
32
Compulsive talker; verbose; tangential; word-finding problems
33
Poor social communication skills; inappropriate turn-taking; interruptions
Section one: > 24 points in itilized items => possible cluttering
Section two: itilized items provide supporting information on linguistic component in cluttering
Section three and four provide additional information on personal communicative skills
55
References
Bezemer, B.W., Bouwen, J., & Winkelman, C. (2006). Stotteren van theorie naar therapie.
Bussum: Coutinho.
Daly, D. (1996). The source for stuttering and cluttering. East Moline, IL: LinguiSystems.
Daly, D. (2008). Strategies for identifying and working with difficult to threat cluttering
clients, Presented July 5
th
, Proceedings of the Oxford Dysfluency conference, in
press.
Daly, D. A. & Cantrell, R.P. (2006). Cluttering characteristics identified as diagnostically
significant by 60 fluency experts. Second World Congress on Fluency Disorders.
Proceedings.
Damste, P.H. (1984). Stotteren. Utrecht: Bohn, Scheltema & Holkema.
Dinger, T., Smit, M., & Winkelman, C. (2008). Expressiever en gemakkelijker spreken,
Bussum, the Netherlands: Coutinho.
Mensink – Ypma, M. (1990). Broddelen en leerstoornissen, Houten/Antwerpen: Bohn
Stafleu van Loghum.
Pollit, D.F., & Beck, C.,T. (2003). Nursing research. Principles and methods,7
th
ed,
Baltimore: Lippincott Williams & Wilkens.
St. Louis, K.O. (1992). On defining cluttering. In F.L. Myers & K.O. St. Louis (Eds.),
Cluttering: A clinical perspective. Kibworth, Great Britain: Far Communications, Ltd.
St. Louis, K.O., Raphael, L. J., Myers, F.L., & Bakker, K. (2003). Cluttering Updated. The
ASHA Leader, ASHA, 4-5, 20-22.
St. Louis, K.O., Myers, F. L., Bakker, K. & Raphael, L. J. (2007). In Conture, E. and Curlee,
R. (Eds.). (2007). Stuttering and Other Fluency Disorders, (3
rd
Ed.). Philadelphia,
PA: Thieme Medical.
Van Zaalen, Y., Wijnen, F., Dejonckere, Ph., (2009c). Differential diagnostics between
cluttering and stuttering, part one. Journal of Fluency Disorders, in press.
Van Zaalen, Y., & Winkelman, C. (2009). Broddelen, een (on) begrepen stoornis. Bussum:
Coutinho.
Ward, D. (2006). Stuttering and cluttering, Frameworks for understanding and treatment,
East Sussex: Psychology Press.
Weiss, D.A. (1964). Cluttering. Englewood Cliffs, NJ: Prentice-Hall.
Weiss, D.A. (1968). Cluttering: Central language imbalance. Pediatric Clinics of North
America, 15, 705-720.
Wigg, E.H., & Semel, E.M. (1984). Language Assessment and Intervention for the Learning
Disabled. 2nd Ed., Columbus, OH: Charles E. Merrill.
Winkelman, C.L. (1990). Broddelen. In W. Mensink, Broddelen en Leerstoornissen, Utrecht:
Bohn, Scheltema & Holkema.
56
57
Chapter Four
A test on speech motor control on word level productions: The SPA Test
(Dutch: Screening Pittige Articulatie)
Y. van Zaalen – op ’t Hof, F. Wijnen and P.H. Dejonckere (2009a)
International Journal of Speech and Language Pathology, 11:1, 26-33.
58
Abstract
The primary objective of this article is to study whether an assessment instrument
specifically designed to assess speech motor control on word level productions would be
able to add differential diagnostic speech characteristics between people who clutter and
people who stutter. It was hypothesized that cluttering is a fluency disorder in which speech
motor control on word level is disturbed in high speech rate, resulting in errors in flow of
speech and sequencing. An assessment instrument on speech motor coordination on word
level was developed and validated. In an elicitation procedure, repetitions of complex multi-
syllabic words at a fast speech rate were obtained from 47 dysfluent participants (mean
age 24.3; SD 10.25, range 14.2–47.4 yrs) and 327 controls (mean age 25.56 yrs; SD 8.49;
age range 14.3–50.1). Speech production was judged on articulatory accuracy, smooth-
flow (coarticulation, flow and sequencing) and articulatory rate. Results from people who
clutter (PWC) and people who stutter (PWS) were compared to normative data based on
control group data. PWC produced significantly more flow and sequencing errors compared
to PWS. Further research is needed in order to study speech motor control in spontaneous
speech of people who clutter.
1. Introduction
Successful communication requires the active combination of a range of cognitive and
linguistic skills. On the one hand, a speaker must coordinate a range of language-based
faculties, including those required to competently formulate and structure sentences.
Equally, speech planning and speech production processes are utilized to ensure that
language-based elements are produced in an intelligible and coherent manner. Intelligibility
itself is related to a number of factors, such as accurate sound production including speech
rhythm, stress patterning, and articulatory rate.
It is a well known fact that persons with cluttered speech (PWC) experience
problems in speech production resulting in unintelligible speech (Daly, 1992; St. Louis,
Myers, Raphael and Ward, 2003; Weiss, 1964; Ward, 2006). It is hypothesized that
cluttered speech occurs when speech rate is too fast for the speech system to handle, or
when the person with cluttered speech does not give enough attention to the task of
speech production.
Many researchers and clinicians report that PWC experience intelligibility problems
due to exaggerated coarticulation (deletion of sounds or syllables in multi-syllabic words),
indistinct articulation (substitution of sounds and/or syllables), and problems in accurate
59
pausing (Bezemer, Bouwen & Winkelman, 2006; Daly & Cantrell, 2006; St. Louis, Myers,
Bakker & Raphael, 2007; Ward, 2006). Several researchers discuss the fact that although
PWC experience intelligibility problems in running speech, many are able to produce
correct syllable and word structures in controlled situations (Bezemer et al., 2006; Damsté,
1984; St. Louis et al., 2007; Ward, 2006; Weiss, 1964). To produce intelligible syllable or
word structures, the speaker must exercise appropriate levels of control over speech motor
processes. Riley and Riley (1985) defined speech motor control as the ability to time
laryngeal, articulatory, and respiratory movements that lead to fast and accurate syllable
production. This ability is implicit in the widely accepted working definition of cluttering by
St. Louis et al., (2007) which describes cluttering as:
A fluency disorder characterized by a rate that is perceived to be abnormally
rapid, irregular or both for the speaker (although measured syllable rates may not
exceed normal limits). These rate abnormalities further are manifest in one or more of
the following symptoms: (a) an excessive number of disfluencies, the majority of which
are not typical of people who stutter; (b) the frequent placement of pauses and use of
prosodic patterns that do not conform to syntactic and semantic constraints; and (c)
inappropriate (usually excessive) degrees of coarticulation among sounds, especially in
multi-syllabic words. (p. 299).
In 1985, Riley and Riley published the Oral Motor Assessment Scale (OMAS). This
instrument tests the ability of a speaker to produce intelligible syllable strings at a fast rate.
Recent research (van Zaalen, Wijnen & Dejonckere, 2009c) revealed that adult and
adolescent PWC performance on the OMAS cannot be differentiated from those of persons
who stutter (PWS) or controls. PWC experienced no significant difficulties in oral motor
coordination at the syllable level. Based on clinical observations in working with PWC, it is
hypothesized that cluttering is a fluency disorder in which speech motor control at the word
level is disturbed when speaking at a fast speech rate, resulting in errors in the flow and
sequencing of speech.
The main purpose of this study is to test whether an assessment instrument
specifically designed to assess speech motor control at the word level would be able to
differentially diagnose the speech characteristics between PWC and PWS.
60
2. METHOD
2.1. Participants
All subjects in the PWC and PWS groups were referred to a centre for fluency therapy
(between January 2006 and May 2008) in the centre of the Netherlands with self-reported
fluency problems. Participants were 47 disfluent persons including 33 males (age: M=24.7,
SD=9.8, range 14.4-49.3 years) and 14 females (age: M=24.3, SD=10.3, range 14.2-47.4
years) and 327 controls including 271 males (age: M=25.5, SD=7.9, range 14.1 – 54.3 yrs)
and 56 females (age M=28.8, SD=8.5, range 14.3-50.1). Participants were divided in three
diagnostic groups (PWC, PWS and controls). Diagnostic decision making was based on
the objective results of the measurements on articulatory rate, ratio disfluencies,
intelligibility, and the score on the Stuttering Severity Instrument (Riley, 1994). Diagnostic
decision making procedures are described in detail in van Zaalen et al., (2009c). Controls
were included in order to obtain normative values on speech motor control at the word level
on the SPA test. None of the participants (including controls) reported any neurological or
hearing disorders and all were Dutch speaking mono- or bilinguals with an intermediate to
high educational level. Subjects were tested in the first assessment session prior to therapy
or in case of subjects that were already in the course of treatment, (at the most 3 months)
before the therapy session began. Adolescent and adult control subjects were selected at
random from a database of volunteers originating in different parts of the country and who
participated in the study. A total of 374 subjects participated in the study.
2.2. Speech motor control at the word level: The SPA Test
In order to examine the speech motor control on the word level, an assessment instrument
was developed. The SPA (Dutch: Screening Pittige Articulatie), designed by the first author,
is a specially created speech task that provides information on speech motor control and
word structure productions at the word level when speaking at fast rates. In an elicitation
procedure, three repetitions of ten multi-syllabic words at a fast speech rate were obtained.
Stimuli were similar to those on the OMAS in that the SPA elicited three words containing
(a) mostly bilabial onset consonants (similar to [pə]; e.g. Dutch: [Ǥpərseremonimestər]), (b)
mostly alveolar and velar onsets (as in [təkə], e.g. Dutch: [vərǡndərəndə
ǽevənsǤmstǡndǺxhedən]); or (c) a combination of bilabial, alveolar and velar consonants
(as in [pətəkə], e.g. Dutch: [Ǥnœytsprekələk vǫrveləndə vǫrhǡndəǽǺȃən]). These repetitions
61
were judged on articulatory accuracy, smooth-flow (coarticulation, flow and sequencing),
and articulatory rate.
Articulatory accuracy and smooth flow measurements
In order to classify errors in word structure, the SPA (Dutch: Screening Pittige Articulatie)
was developed (see Appendix 1). Scoring was devised to be consistent with Riley and
Riley’s (1985) Oral Motor Assessment Screening (OMAS) protocol. In this study, errors
were defined within three different categories: (a) Accuracy, (b) Smooth Flow, and (c) Rate.
Judgment of errors in sound or syllable production was based on a three point scale: Zero
errors = 0 points, one to two errors = 1point, and three+ errors = 3 points. The more errors
one produces, the higher the score.
2.2.1. Accuracy
Problems in sound accuracy (distortions or substitutions of voicing and devoicing) were
scored. In the Dutch language, substitution of a target sound and the error may not have
the same voicing category, as in English (i.e., [th] -> [s]). Accuracy scores for both PWS
and PWC group that fell more than 1.5 SD above the mean score for the controls were
considered to be an indication of problems in adjusting voicing to articulatory movement: an
indicator of difficulty realizing adequate voice-onset time.
2.2.2. Smooth flow
Problems in smooth flow were subdivided into three categories: Coarticulation, flow, and
sequencing. Coarticulation is the gradual transfer from one speech movement to the next.
Errors in coarticulation were: telescoping syllables (i.e., Dutch: [Ǥpsermonmestər] in stead
of [Ǥpərseremonimestər]), and within sequence pausing (i.e., Dutch: [Ǥpərsere ….
monimestər]). Flow is the gradual stressing and rhythm of the sequence. Errors, for
example, may include changes in the stress pattern for the sequence (i.e., Dutch:
[Ǥpərseremonimestər] instead of [Ǥpərseremonimestər]). Sequencing errors were scored
when a person makes sound order errors between or within syllables (i.e., Dutch:
[Ǥpərmoniseremestər] instead of [Ǥpərseremonimestər]). Total smooth flow scores that are
more than 1.5 SD above the mean score of the controlgroup were considered to be an
indication of problems in speech motor control at the word level.
62
2.2.3. Rate
Rate was determined in syllables per second (SPS) by counting the mean time in seconds
needed to produce a sequence of three target syllables. Normative comparison data was
derived from the controls in this study. A reference test of speech motor control at the word
level was not available. In order to get normative comparison data on accuracy, smooth
flow, and rate, mean results of the non-fluent speakers were determined in z-scores.
3. RESULTS
3.1.1. Accuracy
A univariate analysis of variance between diagnostic groups corrected with Tukey’s-b
procedure for unequal group size revealed a significant group difference in Accuracy
scores [F(2,373)=66.675, p< .0001]. Controls produced a mean of .23 (SD .53) accuracy
errors. Controls produced significantly (p < .0001) fewer accuracy errors compared to
PWC. PWC produced significantly [F(1,46)=5.600, p= .022] more accuracy errors
compared to PWS (PWC: z-score M=1.65, SD=1.46; PWS: z-score M= .67, SD=1.23).
(see table 1-3 and figure 1).
Figure 1.: Mean scores on accuracy, smooth flow and rate for PWC
and PWS group and additional normative data and reference lines
63
3.1.2. Smooth flow
An analysis of variance between diagnostic groups on Smooth Flow errors revealed a
significant group difference between controls and diagnostic groups [F(2,373) =116.460, p<
.0001]. Controls produced a mean of 2.04, SD 1.91 smooth flow errors. Controls produced
significantly (p < .0001) fewer smooth flow errors compared to PWC (z-score M=1.61, SD=
1.11) and PWS (z-score M=1.68, SD=1.20). Z-scores of PWS and PWC did not differ
significantly [F(1,46)= .039, p= .844], (see Table 1-2). A closer examination of the smooth
flow scores for the three different categories of coarticulation, flow, and sequencing was
done.
3.1.3. Coarticulation
An analysis of variance between diagnostic groups on Coarticulation errors revealed a
significant group difference [F(2,373) =88.959, p< .0001]. Both PWC and PWS produced
significantly (p < .0001) more coarticulation errors compared to controls. (Coarticulation
errors: Controls: M= .617, SD .97; PWC: z-score M= .63, SD= .68; PWS: z-score M=2.34,
SD=2.22). PWS had significant higher Z-scores compared to PWC [F(1,46)=15.109, p <
.0001] (see table 1-3 and figure 2).
N Minimum
Maximum
Mean
Std.
Deviation
F
Sign.
accuracyABC 327 .00 3 .23 .53
66.675
0.00
smoothABC 327 .00 10 2.04 1.91
116.46
0.00
Coarticscore 327 .00 5 .62 .97
88.959
0.00
Flowscore 327 .00 3 .62 .72
58.428
0.00
Sequencescore
327 .00 5 .81 1.12
42.121
0.00
rateABC 327 3.1 8.5 5.2 .92
81.288
0.00
Passcore 327 3.8 16.5 7.5 2.35
194.267
0.00
Valid N
(listwise) 327
Table 1.: Normative values for Screening Pittige Articulatie
3.1.4. Flow
An analysis of variance between diagnostic groups on Flow scores revealed significant
group differences [F(2,373)=58.428, p< .0001]. (Flow errors: Controls: M= .62, SD= .72;
PWS: z-score M= .66, SD=1.33; PWC: z-score M=1.57, SD= .85). Both PWS and PWC
64
produced significantly (p < .0001) more flow errors compared to Controls. PWC had
significant higher Z-scores compared to PWS [F(1,46)=8.079, p = .007] (see Table 1-2 and
Figure 2).
3.1.5. Sequencing
An analysis of variance between diagnostic groups on Sequencing errors revealed a
significant group differences [F(2,373)=42.121, p< .0001]. Controls produced a mean of
.807 (SD=1.12) sequencing errors (see Table 1.). Z-scores of PWS and PWC were
significantly different, [F(1,46)=4.782, p = .034]. PWS (z-score M.41, SD 1.07) scored
according to the controls. PWC (z-score M=1.42, SD=1.77) produced significantly more
sequencing errors compared to controls. (see Table 1-2, Figure 2).
Group N Minimum Maximum Mean Std. Deviation
Accuracy 29 -,52 3,78 1,6541 1,46190
Smooth flow 29 -,25 3,90 1,6096 1,11368
coarticulation 29 -,62 1,51 ,6266 ,67576
Flow 29 -,90 2,62 1,5682 ,85013
Sequencing 29 -,71 4,16 1,4234 1,77168
Rate ABC 29 -1,02 9,46 2,8257 3,18358
SPA score 29 1,40 18,96 9,3678 4,68723
PWC
Valid N (listwise) 29
Accuracy 18 -,52 2,34 ,6757 1,22667
Smooth flow 18 -1,01 3,15 1,6775 1,19957
coarticulation 18 -,62 5,77 2,3382 2,22438
Flow 18 -,90 2,62 ,6643 1,33520
sequencing 18 -,71 2,07 ,4100 1,07005
Rate ABC 18 -,47 6,38 2,4800 2,09718
SPA score 18 -1,97 16,97 8,2626 5,54493
PWS
Valid N (listwise) 18
Table 2.: Z-scores on subcategories (total error score on accuracy and smooth flow (coarticulation, flow and
sequencing) and total rate in PWC and PWS groups).
65
Figure 2. : Z-scores and normative reference lines on coarticulation,
flow and sequencing
3.1.6. Rate
An analysis of variance between diagnostic groups on Rate scores revealed a significant
group difference [F(2,373)=81.288, p< .0001]. Controls had a mean rate of 5.24, SD .92
seconds for the three words (see Table 1.). Both PWC and PWS had a significantly slower
rate compared to controls PWC: z-score M 2.83,
SD 3.18 and PWS z-score M 2.48, SD 2.10. Z-
scores of PWS and PWC were not significantly
different, [F(1,46)=. 167, p = .685] (see Table 1-
2).
3.1.7. Total SPA z-score
An analysis of variance on Total SPA z-scores
between diagnostic groups showed a significant
group difference. Controls had a significantly
lower total score compared to PWC and PWS
[F(2,371= 194.267, p < .0001]. Total SPA scores
Fig. 3.: Total SPA z-scores for PWC, PWS and controls
of PWC (z-score: M=9.4; SD=4.7) did not differ
and normative reference line.
significantly from PWS (z-score M=8.3, SD=5.5), [F(1,46)= .536, p= .468]. (see Table 1-2
and Figure 3).
66
4. DISCUSSION
Many researchers have reported intelligibility problems in persons with cluttered speech
(Bezemer, et al., 2006; Daly, 1996; Dinger, Smit & Winkelman, 2008; St. Louis et al., 2003;
St. Louis, 2007; Ward, 2006). Until now, an assessment instrument for speech motor
control at the word-level has not been validated for the disfluent population. The main
purpose of this study was to test whether an assessment instrument specifically designed
to evaluate speech motor control at the word level would be able to help differentially
diagnose the speech characteristics of persons with cluttered speech and PWS. Results
indicate that high scores on accuracy, flow and sequencing can differentiate PWC from
PWS, but the total score on the Screening Pittige Articulatie (SPA) did not differentiate
between PWC and PWS.
Accuracy scores
Accuracy scores on the SPA for the PWS and PWC groups with more than 1.5 SD above
the mean of controls were considered to be indicative of problems in adjusting articulatory
movement necessary to realize adequate voice-onset time. In this study, PWC had a mean
z-score of 2.40 (SD=1.91) and met the criterion of severe problems in adjusting voicing to
articulatory movement, while PWS did not. Problems in adjusting voicing can be seen when
the timing demands for planning and execution processes can not meet (Howell, 2004).
Target words used in the SPA, with complex phonetic and phonological properties that
carry lexical stress, may cause more difficulty than those that do not have these properties
(Howell & Dworzynski, 2005). This was evident for persons with cluttered speech.
Smooth Flow scores
Total Smooth Flow scores falling more than 1.5 SD above the group mean of the controls
were considered to be an indication of severe problems in speech motor control at the
word level. Total smooth flow scores appeared to be of no significant value in differentiating
between PWC and PWS.
Coarticulation
In developing the assessment protocol, we replicated the major judgments categories used
in the OMAS to the SPA. In the Oral Motor Assessment Scale (Riley & Riley, 1985)
protocol, a major category named “coarticulation” contains both telescoping of syllables
and extra pausing. Because PWS produced frequent extra pauses between and within
67
words, their score on coarticulation was high. On the other hand, persons with cluttered
speech produced frequent telescoping errors, while PWS rarely telescoped syllables. In
combining pausing and telescoping data within the coarticulation category, potential
difference might have been masked in the overall analysis. In future versions of the SPA, it
is recommended that telescoping and pausing errors be split into two categories.
Flow and Sequencing scores
Scores on flow and sequencing errors differentiated the PWC from controls and PWS. Flow
and sequencing abilities can be disturbed when speech planning has to be performed
within small time limits. In the SPA, participants had to repeat test words at a fast speech
rate. Goberman and Blomgren (2008) reported that stuttering speakers exhibited
significantly more stuttering on variable rate tasks than on habitual rate tasks. Rieber,
Breskin & Jaskin (1971) reported that PWS tend to have greater mean pause times and
lower mean phonation times than persons with cluttered speech. In a fast rate, pause time
is reduced, resulting in a higher frequency of flow and sequencing errors in the PWC group
and a longer phonation time in the PWS group. PWS produced extra pauses that disturbed
their flow of speech, but the syllable order was not disturbed.
PWC produced a high number of errors in a rate perceived to be fast, but not
statistically different from the PWS. It is suggested that the accuracy and smooth flow
problems in PWC negatively influenced their intelligibility. It is further suggested that
articulatory rate in the PWC, although measured within normal limits, is perceived to be
abnormally fast as a side-effect of other issues relating to problems in speech motor
planning and speech motor execution (St. Louis et al., 2007; VanZaalen et al.,2009c). As
Ward (2006) described, PWS seem to have problems producing what is already coded,
while PWC experience problems in coding speech during conversation.
Total SPA score
The Total SPA Score was calculated by the summation of accuracy errors, smooth flow
errors, and rate scores. The total SPA score for PWC and PWS was negatively influenced
by speech rate. The total SPA score appeared to be of no value in the differential diagnosis
within the disfluent population. This can be explained by the high scores for PWS in some
categories and high scores for PWC in other categories. Both PWC and PWS experienced
difficulties with speech motor skills.
68
Although this study has presented some new insights in the speech motor control in
persons with cluttered speech or PWS, further research is needed in areas of spontaneous
speech and other diagnostics groups that are related to speech motor planning or
execution problems.
Speech motor control on word level
“Speech will be fluent if execution time for the segment currently being produced is
sufficiently long for the plan for the following segment to be ready, after the current
segment has been executed”(Howell & Dworzynski, 2005, p.352). Fluent speech needs
separately planning and execution components (Levelt, 1989). Speakers can start an
utterance (execution) before they have the complete plan (Kolk & Postma, 1997). When
execution is getting ahead of planning, fluency problems arise. In testing speech motor on
word level (multi-syllabic words) at a fast rate planning time was shortened. While PWC
experienced accuracy, flow and sequencing errors as a result of that, it can be assumed
that planning problems underlie on PWC production problems. While PWS experienced
mainly coarticulation problems, it can be assumed that execution problems underlie on
PWS production problems. Further research is needed to confirm this finding.
CONCLUSION
The main purpose of this study was to answer the question whether an assessment
instrument especially designed to assess speech motor control at the word level would be
able to differentially diagnose speech characteristics of PWC and PWS. Results show that
smooth flow scores differentiate PWC from PWS. PWC produced significantly more flow
and sequencing errors compared to the PWS. In addition, PWS produced significantly
more errors on coarticulation compared to PWC and controls. Overall, the total score on
the SPA test served to differentiate between fluent and disfluent participants. The SPA test
on speech motor control on word level productions differentiated PWC from PWS.
69
References
Bezemer, B.W., Bouwen, J., & Winkelman, C. (2006). Stotteren van theorie naar therapie,
De broddelcomponent in stotteren, p. 277-290. Bussum: Coutinho.
Daly, D. (1996). The source for stuttering and cluttering. East Moline, IL: LinguiSystems.
Daly, D. A., & Cantrelle, R. P. (2006). Cluttering characteristics identified as diagnostically
significant by 60 fluency experts. Paper presented at the 5th World Congress on
Fluency Disorders: Proceedings.
Damste, P.H. (1984). Stotteren. Utrecht: Bohn, Scheltema & Holkema.
Dinger, T., Smit, M., & Winkelman, C. (2008). Broddelen. In Expressiever en gemakkelijker
spreken, 100-104. Bussum, MA: Coutinho.
Goberman, A., & Blomgren, M. (2008). Fundamental Frequency Change During Offset and
Onset of Voicing in Individuals with Parkinson Disease, Journal of Voice, Volume
22, Issue 2, 178-191.
Howell, P. (2004a). Assessment of some contemporary theories of stuttering that apply to
spontaneous speech, Contemporary Issues in Communicative Sciences and
Disorders, 39,122–139.
Howell, P., & Dworzynski, K. (2005). Planning and execution processes in speech control
by fluent speakers and speakers who stutter, Journal of Fluency Disorders, 30 (4),
343–354.
Levelt, W. J. M. (1989). Speaking: from intention to articulation, Cambridge, MA: MIT
Press.
Kolk, H., & Postma, A. (1997). Stuttering as a covert repair phenomenon. In: Curlee, RF.,
Siegel, G., M., (Eds.), Nature and treatments of stuttering: New directions, Needham
Heights, MA: Allyn & Bacon, 182-203.
Rieber, R., W., Breskin, S., & Jaffe, J. (1972). Pause Time and Phonation Time in
Stuttering and Cluttering, Journal of Psycholinguistic Research, Vol. 1, No. 2, 149-
153.
Riley G. D., & Riley, J. (1985). A component model for treating stuttering in children,
Peins, M., Contemporary Approaches in Stuttering Therapy, Boston/Toronto,
MA:Little, Brown &Company: Boston/Toronto.
Riley, G. D. (1994). Stuttering severity instrument for children and adults (3
rd
ed.). Austin,
TX:Pro-Ed.
St. Louis, K. O., Raphael, L. J., Myers, F. L., & Bakker, K. (2003). Cluttering Updated. The
ASHA Leader, ASHA, 4-5, 20-22.
St. Louis, K.O., Myers, F. L., Bakker, K. & Raphael, L. J. (2007). In Conture, E. and Curlee,
R. (Eds.). (2007). Stuttering and Other Fluency Disorders, (3
rd
Ed.). Philadelphia,
PA: Thieme Medical.
Van Zaalen, Y., Wijnen, F., & Dejonckere, P.H. (2009c). Differential diagnostic
characteristics between cluttering and stuttering, Journal of Fluency Disorders, in
press.
Ward, D. (2006). Stuttering and cluttering, Frameworks for understanding and treatment,
Psychology Press, East Sussex.
Weiss, D. A. (1964). Cluttering, Englewood Cliffs, NJ. Prentice-Hall, Inc.
70
Appendix 1: Test protocol
Screening Pittige Articulatie (SPA)
van Zaalen, 2009
Purpose:
This assessment can produce insight in speech motor skills on word level within test
circumstances. Of ten test words, only the three bold printed words are analysed on
accuracy, smooth flow and articulatory rate.
Instruction:
The client can look at the word to produce during 5 seconds maximum. The word is
covered and subsequently the client may repeat the words 3 times in consecutive syllable
strings, in a fast but still intelligible way and without pauses.
Words (Dutch, bold words are test words):
periodieke uitkeringen = letterlijk en figuurlijk = veranderlijke wind uit westelijke richtingen
= onuitsprekelijk vervelende verhandelingen = woordelijke aanhalingen = geldelijke
tegemoetkoming = opperceremoniemeester = onverantwoordelijke elementen =
veranderende levensomstandigheden = maatschappelijke verhoudingen
Name:
Mean number of errors in three attempts is determined and pointed out. Error score is pointed out in the row
directly below.
Accuracy Smooth Flow
Word set Distorsion Voicing Coarticulatio
n Flow Sequencing
Rate in SPS
Opperceremonie-
meester 0 1-2 3+ 0 1-2 3+ 0 1-2
3+
yes
no 0 1-2 3+
Error score 0 1 3 0 1 3 0 1 3 0 1 0 1 3
Sec.
A
Veranderende
levensomstandigh
eden
0 1-2 3+ 0 1-2 3+ 0 1-2
3+
yes
no 0 1-2 3+
Error score 0 1 3 0 1 3 0 1 3 0 1 0 1 3
Sec.
B
Onuitsprekelijk
vervelende
verhandelingen
0 1-2 3+ 0 1-2 3+ 0 1-2
3+
yes
no 0 1-2 3+
Error score 0 1 3 0 1 3 0 1 3 0 1 0 1 3
Sec.
C
Totaalscore
A+B+C =
sec
Z-score
Overallscore
and interpretation
71
Chapter Five
Language planning disturbances in children who clutter or have speech
impairment related to learning disability
Y. van Zaalen – op ’t Hof, F. Wijnen and P.H. Dejonckere (2009b)
International Journal of Speech and Language Pathology, in press,
DOI: 10.1080/17549500903137249.
72
.
Abstract
The primary objective of this paper is to determine to what extent disturbances in the
fluency of language production of children who clutter (CWC) might be related to, or differ
from difficulties in the same underlying processes of language formulation seen in children
with learning disability (LD). It is hypothesized that an increase in normal disfluencies and
sentence revisions in CWC reflect different neurolinguistic process to those of LD-children.
To test this idea, 150 Dutch speaking children, age 10;6 - 12;11 years, were divided in
three groups (cluttering, learning difficulties and controls), and a range of speech and
language variables were analysed. Results indicate differences in the underlying processes
of language disturbances between children with cluttered speech and those with learning
disability. Specifically, language production of LD-children was disturbed at the
conceptualiser and formulator stages of Levelt's (1993) language processing model, whilst
language planning disturbances in CWC were considered to arise due to insufficient time to
complete the editing phase of sentence structuring. These findings indicate that CWC can
be differentiated from LD-children by both the number of main and secondary plot elements
and by the percentage correct sentence structures.
1. Introduction
The disorder of cluttering provides us with an obvious example of how much speech
language disorders and