Title: Models of stuttering
1Models of stuttering
2Why develop a model of stuttering?
- What does a model of stuttering give us
- In terms of guiding treatment?
- In terms of guiding research?
- For one other good reason to study models of
stuttering, stay awake til the end of this
presentation! - What must a good model of stuttering account for?
3What is a model of stuttering?
- In trying to explain stuttering, we must try to
account for - What initially causes stuttering
- triggers vs. causes?
- What the basis for fluency breakdown is
- How stuttering develops over the lifespan (how it
is maintained and perpetuated) - Some models are not mutually exclusive
- (e.g., genetics and learning)
- Etiology ? model, but can help build one
- A good model must explain
- The phenomenology of stuttering
- When and how stuttering develops
- How fluency breakdown fits within known facts
about speech production
4A historical perspective
- Stuttering is one of the oldest recorded
communication disorders - Theories can be divided into those that view
stuttering as - Physiological
- Psychological
- Environmental
- Multifactorial
5The nature vs. nurture controversy
- Models of stuttering tend to divide in their
emphasis on whether stuttering is primarily due
to physiological bases, environmental influences,
or a mixture of both. - Virtually no models ignore some component of
learning in the DEVLOPMENT of the disorder - Before discussing models, we will examine studies
that have contrasted PWS and normally fluent
speakers on physiological and environmental
features.
6Nature
- Refinement of genetic studies show a clear
evidence of genetic contribution to many cases of
stuttering - such studies weaken environmental accounts when
affected family members cannot serve as models,
or demonstrate patterns of stuttering that differ
from that of the affected child. - Conversely, discordant twins show some impact of
non-genetic factors on the emergence of
stuttering.
7Genetic factors in stuttering.
- At this point, few disagree that stuttering has a
genetic component. This component appears to
predict chronicity and recovery better than it
predicts severity, and the model of the genetic
transmission is still under discussion. Cases in
which monozygotes differ in stutter symptoms are
problematic, although such situations are found
in other disorders having clear genetic roots.
Cases in which adopted children who stutter show
family histories of stuttering in the adopting
family suggest an environmental contribution to
stuttering to some researchers (Bloodstein,
1995). - MISSING DATA Although genetic factors can remove
weight from some hypotheses about the nature of
stuttering, they do not explain the mechanism by
which stuttering physically arises or is best
treated. Moreover, because stuttering is at least
partially predicted by genetic background,
studies of parents of CWS should, but usually
dont, segregate reports of adult parenting
styles by fluency status of parent.
8Nature
- Most studies of physiological function in
stuttering divide into the following topic areas - Reaction time
- Coordination
- Time and visuospatial estimation
- Laterality and cortical organization for language
- Task interference (capacity limitations)
- Most find some, but minor and inconsistent
differences (see book).
9Physiological models
- Perhaps the first models of all models (excluding
religious explanations) Older models tended to
localize stuttering to weaknesses in specific
peripheral systems, particularly the tongue, from
Greco-Roman times through the last century - Other parts of the vocal tract, e.g., the larynx,
were investigated during the 1960s and 1970s. - By far, the most common modern physiological
models focus on cortical representation of, and
control of, speech and language functions.
10Challenges to studies of physiological
limitations in stuttering
- Do such limitations exist at stuttering onset or
do they result from adaptation to stuttering? - Could transient limitations early in development
yield persistent symptoms? - Do observed patterns predict the behavioral
symptoms of stuttering? - Are aberrant physiological patterns amenable to
treatment?
11Lateralizationmodels (Webster, 1996)
- Orton-Travis In normal speech/language
production, a dominant hemisphere takes primary
responsibility for linguistic functions. In
stuttering, dominance fails to occur, causing
competing commands to speech/language functions.
From W. Webster
12Orton-Travis (continued)
- SUPPORTING DATA Differences between
lateralization of language functions in PWS and
fluent speakers, as revealed through numerous
tasks EEG (e.g., Moore Haynes, 1980), dichotic
listening (e.g., Gregory Curry, 1967), PET,
fMRI (current work by teams such as DeNil Kroll
(1995), Fox, Ingham Ingham (1996), Braun, et
al. (1996), more each month. - CONFLICTING DATA No evidence of changed
handedness history in majority of PWS. No
evidence that reverting handedness effects
fluency. - MISSING DATA Whether lateralization differences
are primary or the result of experience with
stuttering lateralization theories are
particularly vulnerable to this challenge, as
therapy (particularly those that deal with speech
planning and pacing) and emotional reactions are
logically presumed to elicit changes in right
hemisphere function.
13Other hemispheric organization models (from
Webster)
- Figure 1C illustrates Websters Two-Factor
Interference Model. Three elements comprise this
model i) we still assume normal left hemisphere
specialization for speech and motor sequencing
ii) we see a left hemisphere system for speech
and motor sequencing that is inefficient and
unusually susceptible to interference
(illustrated by the larger number of "pores"),
not just from activity in the right hemisphere,
but also the left (illustrated with arrows from
both hemispheres impinging on the speech
mechanisms) and iii) a lack of what we call
"left hemisphere activation bias" (illustrated
with equal rather than unequal sized arrows
coming from the mid-and hind-brain activating
system).
- As illustrated in Figure 1B, the interhemispheric
interference model has three major features to
it i) normal left hemisphere lateralization ii)
normal right hemisphere functioning but iii)
interference with the left hemisphere coming from
the right hemisphere through a "slop-over" of
activity (the large arrow). Within this model,
variations in stuttering severity would reflect
variations in the amount of overflow
from the right to left hemispheres.
14Other cerebral lateralization models
- Geschwind-Galaburda model (1985) a variety of
speech-language disturbances (dyslexia,
stuttering, autism) may arise from delays in left
hemisphere development during gestation. Such
delays lead to inefficient right hemisphere
dominance for language functions. Because of the
loading of such disorders in males, the authors
hypothesized a role of fetal exposure to high
levels of testosterone. - SUPPORTING DATA Sex-bias in stuttering
incidence other data supporting Orton-Travis
recent anatomical data from Foundas, et al. - CONFLICTING DATA No obvious conflicts, but
theory leaves stuttering in girls somewhat
unresolved - MISSING DATA Mechanisms by which the different
disorders arise evidence of actual testosterone
influences during development, lack of known
association between stuttering and developmental
conditions characterized by hormonal
abnormalities.
15Data for dominance models
- Handedness (inconclusive)
- Wada test (inconclusive)
- Dichotic listening and CAP batteries (variable)
- Neuroimaging (more promising)
- Structural (e.g., CT, MRI)
- Functional (e.g., fMRI, EEG, EPs)
16More on neuroimaging
- Previous work found few structural differences
between PWS and NS however, work in the past
year by Foundas, et al. (2001) and Sommer, et
al.(2002), have shown anatomical variation
between the groups. - (See a short summary of Foundas work) Sommer
et al. is on your reading list and linked. - Work examining functional differences in
processing has been much more promising - rCBF (Watson, et al., 1994 Ingham, Fox Ingham,
1994) - PET (Wu, et al., 1995 Braun, et al., 1997
DeNil, et al., 2000 Fox et al, 1996
differences found Ingham, et al., 1996 no
support for cerebral asymmetries) - (See Fox et al. data here)
- Some studies suggest that observed differences in
cortical function between PWS and NS may diminish
following intensive fluency therapy (DeNil, et
al., 2000).
17Example Braun, et al., 1997
Fig. 2 Brain map illustrating focal rCBF
activation during the formulation and expression
of language in controls (ad) and stuttering
subjects (eh) during tasks in which stuttering
subjects were dysfluent. Language tasks are
contrasted with the oral motor task as a
baseline, in order to highlight regions involved
in linguistic processing. Data for control
subjects are displayed in the top row (ad), and
for stuttering subjects in the bottom row (eh).
18Effects of fluency therapy on brain activity
during single word reading
Controls
Pre
Post
1 year
Source L. De Nil, Univ. Toronto
19Neuropsycholinguistic models
- Neuroimaging data support but do not constitute a
model of what deficits underlie stuttering - For a terrific review of the problems in
interpreting recent neuroimaging studies, see the
recent presentation by Luc De Nil (be patient, it
has wonderful graphics and can take a while to
load). - Neuropsycholinguistic models are more abstract
and provide a more schematic view of the
underlying deficit in stuttering they tend to be
grounded in psycholinguistic models of the normal
speech/language production process. - Among the more popular are the Covert Repair
Hypothesis (Postma Kolk) and the
Neuropsycholinguistic Model (Perkins, Kent
Curlee).
20Levelts model of speech production and
comprehension
21Covert Repair Hypothesis
- Note that Levelts model of speech production
contains a pre-articulatory monitoring stage, - for which there is considerable evidence.
- The CRH states that people who stutter have a
temporal impairment in phonological encoding (it
is slowed). When the phonetic plan cannot proceed
normally, this impairment is adapted to by covert
repairs, restarts and postponements, which result
in the surface or overt behavior of stuttering.
Thus stuttering is seen as a by-product or
side-effect of self-repairs. - Evidence to support the model is building, but is
primarily being gathered by the models builders. - It is also unclear at which stage phonological
encoding or monitoring is disturbed.
22Covert Repair Hypothesis
23Self-monitoring loops
24Spreading activation and phonological error
- Connectionist models (e.g., Dell, 1988), which
describe speech production and comprehension in
terms of activation of representations and their
spread (spreading activation) may account for
mis-selections or poorly timed phonetic plans.
25Spreading activation the general concept and an
illustration
26A potential model of delayed phonological
encoding from Yaruss Conture, 1996).
27Components of the NPL model
- This is basically a temporal dis-synchrony model
- Disfluencies occur due to a disruption in timing
among various linguistic formulation and motoric
execution demands - Stuttering is perceived as a loss of control the
speaker cannot proceed but does not know why
internally imposed time pressure pushes the
incomplete phonetic plan forward. - For an excellent review of the Covert Repair and
NPL models and their implications, visit Scott
Yaruss's supercourse on this topic.
28NPL (continued)
- SUPPORTING DATA most linguistic and motor
differences noted between PWS and fluent
speakers, including imaging and ERP studies
(e.g., Weber-Fox, 2001) linguistic regularities
in stutter events. - CONFLICTING DATA the theory places stuttering on
a continuum with normal disfluencies, and thus
stutter events should conform to patterns of
normal disfluency, which they do not in advanced
stuttering (Silverman Bernstein Ratner, 1997). - MISSING DATA weak evidence for subjective time
pressure and loss of control (although some data
do exist Ezrati-Vinecour Levin, 2001 Barasch,
et al., 2000).
29Other Higher Order deficit modelsInverse
Modeling Deficit Neilson Neilson
- Some concepts have been lifted into the Demands
and Capacities Model (later), but are based on
cybernetic theory. - Summary There is a deficit in the persons
ability to make and use Inverse internal models
of the speech production system. the child has a
sensory-motor model for speech (it inverts
sensory targets into motor commands) Children
plan an utterance based on what it should sound
like. They generate the motor commands. The motor
commands are sent to the muscles which are stored
in the brain as models. The actual speech is
compared to the model expected. The child has a
weakness in making transformations between what
they want to say and the motor movements required
to say them the sensory to motor and motor
sensory transformation. If the demands are not
great, stuttering doesnt occur because the child
can compensate for the weakness.
30Stuttering as a timing disorder
- Van Riper (1982) favored the view that stuttering
represented mis-coordination of the timing
gestures needed for speech. Elaborated on by
Kent (1984), who also relates earlier theories of
cerebral organization to this theory the left
hemisphere is specialized for processing of
brief, rapidly changing acoustic/motoric events,
and may be impaired in carrying out this function
in PWS. Finally, Kent notes the possible
interference with effective function by
overactivity in the right hemisphere stemming
from emotional arousal. - SUPPORTING DATA Evidence of poorly coordinated
gestures during speech production in PWS (e.g.,
Caruso, Abbs Gracco, 1988) acoustic measures
showing abnormalities in speech timing for VOT
and other parameters of speech (e.g., Zimmerman,
1980 and many others). Slowed RT data are
plentiful for PWS for a variety of targeted
behaviors, both motoric and linguistic (Bosshardt
Frandsen, 1996). - CONFLICTING DATA Few, although mistimed segments
are not well-coordinated with stutter events (see
Smith, 1998) variety of stuttered types not well
accounted for (by any theory). - MISSING DATA more data needed from very young
children, to show that timing problems are not
result of stuttering, or ancillary to stuttering.
31Stuttering as an auditory feedback disorder
- Stromsta (1965) suggested that stutterers have an
built-in DAF-like defect in their auditory
feedback systems Tomatis (1963) believed that
stutterers suffered from a hearing loss in the
dominant ear used to monitor speech. While
changes in feedback (DAF, FAF) does produce
increased fluency for many persons who stutter,
neither theory is highly regarded at this time. -
32Stuttering as learned behavior
- Many of the models in this category can be termed
anticipatory struggle hypotheses. - Some of the earliest models developed in this
century by the first professional speech-language
pathologists fit under this rubric. - The Iowa school of research
33Diagnosogenic theory (Johnson, 1959)
- A popular reduction of this theory into few words
is that stuttering is in the ear of the parent - Parents react poorly to normal disfluencies in
children this reaction is perceived and
responded to with anxiety and avoidance by the
children, setting up a cycle of fluency failure,
anxiety, and more fluency failure. - What types of evidence would be required to
validate this theory? - Still a popular theory with broad implications
for advisement to parents, even though most of
its principal tenets have been disproven.
34Diagnosogenic theory (continued)
- SUPPORTING DATA Early studies showing high
levels of demand in parents of CWS high
sensitivity to normal disfluencies as
pathological. - CONFLICTING DATA Almost all studies done in
recent years. In general, there are few
differences between parents of CWS and their
fluent peers (see Bernstein Ratner Silverman,
2000) it is virtually impossible to tell, when
attitudes differ, which came first the attitude
or the stuttering? - Finally, effectiveness of some operant,
parent-administered programs, such as Lidcombe,
are in direct conflict with predictions flowing
from the diagnosogenic theory.
35Ethics and stuttering research the Monster
Study
- What was the Monster Study?
- Read more about the Monster Study
- Discussion
- What are the ethical obligations of researchers
in communication disorders? - What data are required to support a model of
stuttering?
36Approach-avoidance
- General principles
- Conflict occurs when a person experiences demands
or desires that are incompatible with each other.
In approach-approach conflict we are attracted to
two equally desirable goals. In
avoidance-avoidance conflict we must choose
between two equally undesirable demands. In
approach-avoidance conflict we have one goal that
has positive and negative aspects. And in double
approach-avoidance conflict we experience two or
more goals, both of which have positive and
negative aspects. - How would this apply to speaking and stuttering?
- Read Joe Sheehan's "Message to a stutterer"
37Approach-Avoidance (Sheehan, 1953, 1958) in more
detail
- Stuttering results from the conflict between the
opposed drives to speak and to avoid speaking.
When the desire to speak is stronger than the
desire not to speak, speakers can be fluent. When
the desires to speak and not to speak are equal,
stuttering results. Speech avoidance might result
from 1) reactions to specific words, stemming
from conditioned responses to phonetic factors,
2) reactions to threatening speaking situations,
3) guilt or anxiety about message content, 3)
anxiety about relationship between the speaker
and listener, and 5) the need of the ego to avoid
competitive situations that might result in
failure. This theory is a hybrid of learning
theory and personality factors as primary
components in stuttering development. - Approach-avoidance conflict may have its roots in
generalized emotions about speaking, but quickly
develops strong ties to feelings about fluency
and stuttering itself.
38The approach-avoidance gradient
39The Iceberg of Stuttering
40More recent views of learning and stuttering
- In the 1960s more formal accounts of how
stuttering could be described in terms of
classical (instrumental) and operant (respondant)
conditioning were advanced. - In classical conditioning, the child learns to
associate speaking with an emotional response - In operant conditioning, fluency failures are
shaped by the responses they elicit. - This account explains development better than
onset - Example of such a model Brutten Shoemaker
(1967) two-factor model - Stuttering is the involuntary disruption of
speech resulting from negative emotional
responses that are classically conditioned, while
secondary behaviors are instrumentally
conditioned
41(No Transcript)
42(No Transcript)
43(No Transcript)
44(No Transcript)
45Learning theory (continued)
- SUPPORTING DATA
- Interventions in which the techniques of operant
conditioning are effective suggest some support
for learning theory in stuttering (e.g.,
punishment, negative reinforcement, time-out,
positive reinforcement of fluency, etc.) Ingham
notes that no intervention has managed to
increase stuttering by rewarding it, an imbalance
for a full-fledged model of stuttering as a
conditioned behavior. - CONFLICTING DATA generally low level of success
of such interventions with adults who stutter
genetic factors, motor factors, linguistic
factors that unite many PWS. - MISSING DATA Plots of events that might have
conditioned onsets mechanisms by which operant
techniques appear to achieve results (Bonelli,
Dixon, Ratner Onslow, 1998).
46Related modelBloodstein the continuity
hypothesis
- Stuttering evolves from normal disfluency, and
the childs responses that lead to tension and
fragmentation. - These responses are self-generated and are not
caused by parental reactions.
47Multi-factorial models
- Multi-factorial models attempt to integrate
physiology, learning and the environment in the
etiology and development of stuttering - Because they involve multiple factors, they are
hard to test and validate, although some are more
empirically testable than others (e.g., Smiths
Multifactorial-Dynamic model vs. Demands and
Capacities).
48Demands and Capacities (DCM)
- Model popularized by Starkweather, Gottwald
Halfond, 1990. - Based on earlier proposals by Andrews Neilson
(1981) - Genetically conditioned weaknesses in systems
that support fluency (language, motor, emotional,
cognitive, etc.) interact with environmental
factors (external and internal demands) to
precipitate and maintain fluency failure - Consistent with recent advances in the
genotype/phenotype, nature/nurture controversy - The model has intuitive appeal to many
clinicians, but has been criticized as circular
and untestable (see recent issue of JFD (2000)
devoted entirely to this debate). - However, there is clear evidence of a series of
trading relationships among language, motor
coordination and fluency (Smith colleagues
Ratner colleagues) that provide some level of
support for the model. - These will be explored more fully in a later
lecture.
49DCM (continued)
-
- SUPPORTING DATA Stuttering arises during the
most dynamic period of speech, language and
fluency development, a time in which trade-offs
are readily noted in childrens speech (Crystal,
1987, Ratner, 1997). Some environmental
modifications to reduce parental speech rate and
demand on children appear to facilitate fluency
(but complicated by normal patterns of remission
to be expected in 80 of CWS). - CONFLICTING DATA No evidence that parents of CWS
are more demanding than other parents (Bernstein
Ratner Silverman, 2000 Miles Bernstein
Ratner, 2001) - MISSING DATA The notion that individual children
may have capacity limitations that can be
stressed by normal levels of demand essentially
make the model untestable. Ingham Cordes (1997)
rightfully critique the current model as
circular. Data to suggest that given parental
behaviors affect fluency patterns in a systematic
fashion might address this problem.
50Multifactorial-Dynamic Model
- The symptoms of stuttering are like smoke coming
out of a volcano they tell us little about the
underlying processes that produce the phenomenon. - Stuttering evolves from essentially normal
systems that interact poorly - Neural pathways are dynamic and self-learning
abnormal patterns can self-perpetuate and become
stabile. - Problems integrating systems are apparent in PWS
linguistic and cognitive demands impair their
motor functioning more obviously than in PWDNS. - Similar to DeNil and colleagues
Neurophysiological Model, in which learning of
new skills is seen to change physiological
markers of speech/language processing
51Stuttering as an emotional or psychological
problem
- Still a popular lay account, especially in other
cultures - Stuttering is still classified as a neurosis in
the DSM-IV (Diagnostic and Statistical Manual of
the APA) - Ex repressed need hypothesis stuttering is the
result of a repressed neurotic conflict - Psycho-sexual (Glauber, Freud) or
- Inadequate interpersonal relationships (Barbara,
Wyatt) - Most thoroughly summarized by Travis (1957)
- No evidence to support the theory (e.g., Yairi,
1997) no evidence to suggest that psychotherapy
is effective in managing or curing stuttering
(Andrews, et al., 1983). - Distinguishing cause from effect in psychological
involvement in stuttering - What data would be required to support such
models? What data exist to weaken them?
52Specific theories
- Stuttering as manifestation of repressed need
oral gratification (Coriat (1931) - Repressed desires to remain silent (Fenichel,
1945), or withhold unspeakable thoughts (Barbara,
1954) - Response to childhood stress and trauma (e.g.,
Wyatt, 1958) - SUPPORTING DATA Early studies showing levels of
personality disorder and maladjustment in PWS. - CONFLICTING DATA Most, if not all, recent
research. Genetics, motor data, linguistic data.
Lack of effectiveness of psychotherapy in
treating stuttering symptoms. Obvious difficulty
in determining the directionality of any
differences between personality attributes of
people who stutter and comparison groups. - MISSING DATA Genetic contributions of some
personality traits. - FOR BEST DISCUSSIONS, SEE BLOODSTEINS HANDBOOK,
1995.
53Freuds iceberg (compare to Sheehans)
54Why teach about theories and models?
- Discussion
- How relevant are they to treatment?
- Can good treatment occur in the absence of a good
model or theory? - What will you say when a client or parent asks
(or tells you) what causes stuttering?
55Besides
- It might be on the ASHA exam!
56Evaluating research in stuttering
- Using the internet to do research
- PubMed searches
- http//www.ncbi.nlm.nih.gov/entrez/query.fcgi
- Try the tutorial http//www.nlm.nih.gov/bsd/pubme
d_tutorial/m1001.html - Consider your terms carefully!
- Other good resources EBSCO
57(No Transcript)
58(No Transcript)
59Department research on stuttering
- Work in the department can be used to support
neurospycholinguistic and multifactorial models
of stuttering. Read some recent work, including
our published articles, at - http//www.bsos.umd.edu/hesp/reRatner.html
- Do you have a research question youd like to
ask? Come by, and lets talk!