SPEECH MOTOR CONTROL IN FLUENCY DISORDERS.pdf /KUNNAMPALLIL GEJO

SPEECH MOTOR CONTROL

IN

FLUENCY DISORDERS

KUNNAMPALLIL GEJO

JOHN,MASLP KUNNAMPALLIL GEJO JOHN

• Stuttering can be conceived of as a disorder of speech

motor control.

• There is a motor disturbance which is evident in the

abnormal types and amount of speech dysfluencies.

• There is a complex set of relations between this motor

disturbance, the emotional accompaniments of the

disorders and further alterations in speech behaviors.

KUNNAMPALLIL GEJO JOHN

• Vocal reaction times and manual reaction times of stutterers

was examined and compared with those of nonstutterers

and found a greater difference b/w them (Starkweather,

Franklin and Smigo, 1984).

• This indicated that stutterers had a general motoric deficit

which slowed their reaction time, in addition they had

learned habits of tension and struggle that were specific to

the speech mechanism.

• The presence of this additional muscular tension acted as a

drag on the coordinative structures of speech and slowed the

vocal reaction time even further. KUNNAMPALLIL GEJO JOHN

• Kent 1983, examined the variability of speech segment

durations in stutterers and nonstutterers and found that the

stutterers produced segments that were not more variable in

their duration.

• The presence of a generally raised level of muscle activity

could also be used to explain such a result since the temporal

location of a gesture would be less sure if movements

initiating and terminating the segment were made less

smoothly.


• Zimmerman 1980 put forth the idea that stuttering

should be regarded as a disorder of movement.

• He suggested that at the level of the motor neuron a

number of inputs from diverse sources are integrated,

and the sum of these inputs determines the

background tonus and triggering thresholds for

coordinated structures


Characteristics of stuttering as

a motor control disorder

• Speech initiation problems both in conversational

speech and in formal reaction time testing.

• Initiation difficulties are not due to slowed neural

conduction time but due to problems with

coordinating and timing movement (Harbison,

Porter and Tobey, 1989) which is typical of many

motor control disorders (Ludlow and Connor,

1987).


Involuntary, abnormal or excessive muscle activation

patterns occur during speech oscillations in

stuttering.

These can be rhythmic oscillations involving several

muscles with a frequency range of 5 – 12 Hz

(McClean, Goldsmith and Cerf, 1984) heightened

muscle activity (Freeman and Ushijama 1978) or

absence of the expected muscle activity (McClean

1984).


• Stuttering is a task specific disorder; these patients

have no clinical evidence of difficulties with other

vocal tract actions such as chewing, swallowing,

eating or singing

• Stuttering seems to appear only when the speaker

becomes a more fluent speaker of a language

• Stuttering is exacerbated by stress. Particular situations

and listeners seem to trigger stress, affecting stuttering

severity.


• Stuttering is a focal disorder, affecting only

certain structures or movements of vocal tract

like in dysarthria and dystonias.

• Stutterers speech dysfluencies increases with

increased constraints on performance, which is

typical of most movement disorders

(Parkinsonism).

• Stuttering improves with practice like other

motor disorders.


MODELS OF SMC IN FLUENCY

DISORDERS

1. MacKay‟s Model

BUFFER DISPLAY

INDIVIDUAL PHONEME

LEVEL

CONTEXTUAL INTEGRATION

MOTOR UNITS


• The model is composed of a buffer system which displays

phonetic units in abstract forms but in correct serial

order. This may be equated to conceptual level.

• This is fed into an individual phoneme level partially

activating or priming a set of singly represented

phonemic units which are unordered.

• The buffer also generates a set of programs for modifying

the phonemes according to contextual constraints. This

may be equated to encoding process.


• These levels are then fed into a motor unit level, where the

contextual variants are coded

• A “scanner” which sweeps over the motor variants in the

motor unit level in a unidirectional manner at a voluntarily

determined rate.

• When a partially primed unit is passed by the scanner, it

receives an added boost of excitation, bringing it to

threshold for a series of motor commands to be sent to the

speech musculature.

• This is then carried out, which becomes the speech output. KUNNAMPALLIL GEJO JOHN


MODEL 1(NS) MODEL 2(S)

MODEL 3(S) MODEL 4(S) KUNNAMPALLIL GEJO JOHN

• For the NS, hyperexcitability of motor units following their

activation is not sufficient to exceed the motor unit

threshold for stuttering (MODEL 1) yet, stressed units

come closer to threshold than unstressed units

• In stutterers, (a) the same preprimed levels for stressed and

unstressed units as normal, but a lowered motor unit

threshold (MODEL 3)

(b) Greater levels of hyperexcitability than normal (M 2)

(c) Greater prepriming for stressed units, but normal thresholds

and excitability boosts (M 4)


PROLONGATION:

• Individual features may interact in a mutually inhibitory

way e.g. the lip movements for /p/ may be stuttered silently

with neither airflow nor phonation, as if the motor units for

phonation were inhibited, while those for lip movement

were in a state of oscillation

• If oscillations between contradictory elements were rapid

enough, fusion might appear, thus lengthening the speech

sound (i.e.) If the time of triggering is longer, then the

speech sounds may be prolonged.


MASKED OR OMITTED:

– It depends on the time required for a unit to reach threshold and the rate of scanning.

• Let us assume that 2 units, the 1st unstressed and the 2nd stressed, are passed over by the scanner. The 1st will be omitted if the time required for it to reach threshold is greater than the time required for the scanner to activate the 2nd unit and for it to reach threshold. But sweeping rate of the scanner is voluntary and determines the speaking rate.

• If swept rapidly, omission is likely to occur, otherwise there is no omission.


• Blocks:

– When there are omissions, stutterers generally

closely monitor their speech, so to overcome this,

he simply stops when a planned phoneme does

not materialize.


2. THE NEUROSCIENCE MODEL:

(Nudelman et al, 1992)

• The neuroscience model proposes that stuttering is caused by

instability in speech motor control.

• Specifically, a stuttered event is viewed as consisting of two

components :

(1) a momentary instability (in the control theory sense) in the

speech motor control system.

(2) the system‟s response (including its corrections) to this

instability. KUNNAMPALLIL GEJO JOHN

• According to this model, the instability of the

speech control system that leads to stuttering

depends on the interaction of the two nested

functional loops:

1. an outer cognitive loop that provides the reasoning

behind and choice of the words being said

2. an inner production loop that programs and

monitors the sounds being made.


• This model provides two possible reasons why the speech

systems of people who stutter are more likely to become

unstable than those of normally fluent speakers:

1. more processing time may be needed by the outer loop,

2. phase margins in the inner loop may be smaller.

The phase margin is, in effect, the margin of error.

In general, stuttering occurs when there is mismatch between

the selection and programming of speech sounds and the

production of these sounds. It is a causal theory


Background and Development:

• The model draws on a reductionist or “top – down” model of motor

control, which is described in terms of “Functional loops”

(Nudelman et al.,1989).

• Functional loops underpin the four stages of speech production:

Ideation, linguistic programming, motor programming and motor

output.

• They presumably involve the “ temporally overlapping, parallel

execution of the stages with feedback”.

• Functional control loops perform functions that hypothetically must

be accomplished before the desired motor behavior occurs.


• vocal tracking tasks in which adult who stutter and control

subjects tracking computer – generated frequency –

modulated sound waves by humming.

• The stutters responded as quickly as the control subjects to

changes in frequency, but needed more time for processing

the change in tracking frequency.

• The stutters were more variable in terms of phase shift and

so, it was concluded, were more likely to develop momentary

instabilities


• The parameters of the functional loops fluctuate widely and continuously: consequently mistiming occurs intermittently.

• This explains the natural variability of stuttering within individuals; namely why people do not stutter on every word or syllable and why stuttering varies over time.

• Factors that may influence processing time in the outer loop may be psychological, linguistic or sociologic.

• Factors in the inner loop that can increase the risk of breakdown are a function of “ motor context”, namely the position and tension of articulators and the complexity of articulatory movements.


• The model predicts that fluency can be enhanced into two ways:

1. By increasing the phase margin in the outer loops. This can be done directly by “ making a therapeutic strategy automatic”, and indirectly by the person adopting a speech pattern that “slows, rounds, or smoothly shapes the movements”.

2. By decreasing the amount of processing time required by the inner loop, for example by practice and these are the sorts of strategies used by SLP‟s in fluency – shaping treatments.


• Foundas et al, 2001 suggested that their findings of

structure abnormalities in the brains of adults who stutter

lend support to the neuroscience model.

• They suggested that the anomalies they identified, which

were in the perisylvian speech – language areas, could cause

stuttering by reducing the efficiency of neural processing in

the outer loop referred to in the neuroscience model.


EXPLAN MODEL:

Howel 2004

• EXPLAN is an autonomous model of the production of the spontaneous speech that applies to speakers who stutter and fluent speakers.

• Planning (PLAN) and execution (EX) are independent process that reflects the linguistic and motor levels, respectively.

• Failures in the normal mode of interaction between the PLAN and EX processes can lead to fluency failures when plans are too late in being supplied to the motor systems


• The EXPLAN model (Howell, 2004) argues that speech

production involves independent planning and execution

processes.

• Fluency failures such as repetition of prior words, pausing,

prolongation and repetition of parts of the current word

occur when the word to be produced is not ready (the

planning process is not complete) by the time the execution

of the previous word is concluded.

• EXPLAN can explain the behaviour of both fluent speakers

and also the speech of speakers who stutter (SWS).


• The independence of planning and execution systems

allows the plan for a future segment to be generated

during the time the plan for current segment is being

executed.

• If execution time is long enough, the plan for the

following word will be ready after execution of

present word has been completed; and is executed in

sequence


• Fluency fails when a speaker has finished execution

of one plan and next one is not ready for execution.

There are 2 Main reasons why this arises.

1. The inherent properties of linguistic segments make

their planning slow (difficult)

2. Speech is executed at a high rate. The role of each of

the factors can be appreciated by examining the fig.


Plan (n) Plan (n*1) Plan (n*2)

Ex (n) Ex (n*1)

fig 1 fig 2

Difficulty increases the planning time of following word beyond time needed to execute the current word.

Increasing execution rate has effect of shortening the planning time allowed for following word.


Adaptive model theory (AMT)

Megan D Neilson & Peter D Neilson

• The AMT of motor control is a general formulation of the interactions between the sensory and motor processes underlying purposive control of movement.

• Skilled movement is generated with reference to an internal store of information established by integration of sensory feedback with the motor activity which produced it.

• AMT addresses the mechanisms by which sensorimotor relationships might feasibly be established, maintained, refined and whenever necessary, adaptively modified.


• According to AMT any system responsible for voluntary motor

behavior can be conceptualized as a controlled dynamic system

driven by an adaptive controller.

• The controller has the task of transforming a preplanned desired

sensory trajectory into set of motor commands, which when passed

through the controlled system will produce the required sensory event

• Establishing and adaptively maintaining a multivariable, nonlinear,

time varying sensorimotor model requires many network modules

operating as a distributed parallel processor.

• It follows that the greater the modular resources available for the

task, the faster and/or more precisely a model can be determined


• The implications of AMT for stuttering stem from this

neural resources conceptualization.

• That stutterers lack response for efficient sensorimotor

modeling is strongly suggested by the finding that stutterers

perform poorly on auditory tracking tasks, given that the

problem of transforming a body movement into the variation

of an auditory tracking marker closely parallels the problem

of transforming a changing VT configuration into a

changing speech signal


• The authors therefore, suggest that stutterers are deficient

in the processing resources normally responsible for

determining and adaptively maintaining the adaptive

models which sub serve speech production.

• As a consequence of this deficiency, the stutterer must spend

longer in evaluating the sensorimotor relationships involved

in speech, evaluate them less precisely, or deploy additional

resources at the expense of other concurrent functions


• Because the availability and efficiency of resources

are affected by factors such as task familiarity,

competing cognitive workload, motivation and

arousal, the proposed framework predicts variation of

fluency with changing circumstances of capacity and

demand.


Assessment of stuttering:

Aspects of stuttering to be assessed:

1. The effects of sensory set, different speaking situations on eliciting stuttering.

2. The effects of motor set, speaking tasks on eliciting stuttering.

3. The effects of motor response patterns, speech targets on eliciting stuttering.

4. Changes in psychophysioloigcal set during episodic periods of marked fluency and non fluency

5. The effects of feedback elimination or enhancement of eliciting stuttering.


How do we assess these?

To assess these we should first of all have knowledge about:

1. Severity of stuttering: consider secondaries and speech

behaviors

2. Effects of speaking situation of moments of stuttering

3. Effects of speech tasks or loci of stuttering.

Knowing these, we have to assess the effects of speech targets (i.e.)

to determine which motor response patterns are most affected

by stuttering, an analysis is needed of disfluent behavior

during particular speech targets.


• When measuring, the percentage time disfluent in various

speech situations and tasks, an inventory of the gestures

observed during stuttering should determine, which

structures and muscles should be studied during physiologic

and acoustic analysis of stuttering.

Methods to study:

1.Movement transduction:

• This needed to study movement abnormality. i.e. either

tremor, an absence of movement or incorrect or abnormally

slow or fast movement.


• Transducers useful for clinical and experimental

measurement of movement by vocal tract structure are

shown in the table below:

• Structure Clinical Experimental

1. Abdomen and Magneto Respitrace

chest wall meter

2. Larynx Pizeoelectric Fiberoptic video

sensor, EGG palatography

3. Lips and jaw Pizeoelectric strain guage

sensor

4. Tongue Articulograph Ultra sound


Assessment of:

a) Respiration: Here we do not assess the amount of air volume

the subject uses but rather the timing of the chest wall and

abdominal movement and their coordination during

breathing.

b) Laryngeal movement: EGG can be used to identify VF

adduction and glottal vibration. But it cannot identify the

tremor when VF are in abducted position or during whisper.

So piezoelectric accelerometer can be used


• Lips and Jaw: Here the intention is to measure the displacement degree, tremor, identifying irregularity or absence of movement and to measure movement time (from the onset to the completion of a gesture) Pizeoelectric sensor can be used

• Tongue movement: Here changes in tongue shape is measured using ultrasound palatography. It involves placing of electrodes over the palate which has sensors to detect tongue contact. This method alters sensation in the mouth and changes oral shape, both of which are unacceptable for the study of SMC.


From all the above, we can measure:

• Number of tremor oscillations / time

• the time of initiation and completion of movement

• Coordination between different movements and

structures and

• the movement duration


2. Muscle activation:

Which muscles are involved in dysfluent behaviors?

• EMG can be used to study muscle activation for

examining which muscles are involved in disfluent

behaviors. This is to determine whether biofeedback

training might be useful or not.


Muscle activation during disfluency may vary from

hypertonicity, abnormal muscle activity patterning or an

absence of activity. This may be helpful in determining

whether reducing activation in particular group of muscles

will be beneficial for that speaker.

EMG measures: It can be done using a surface electrode or

hooked electrode (indwelling wired electrodes).

• Measures of muscle activation must be relative to the mean

amount of muscle activity during similar movement patterns

which do not involve disfluent speech.


Compare muscle activity with same movement but with disfluent speech.

– Then increases in muscle activation 200% greater than mean – muscle is abnormally active

– A decrease in activation in 50% below the mean – muscle is abnormally quiet.

– Regular repetitive bursts during disfluency – tremor

– Prolonged burst – dystonic posturing

• Experience and familiarity with normal patterns during speech are needed to identify abnormal patterns


ACOUSTIC MEASURES:

Spectrography is used to measure:

• Speech timing: execution time for syllables, the pause time

between syllables, syllable repetition rate – describe timing

abnormalities

• Duration of disfluency, rate of syllable or sound repetition,

breaks in phonation

• F0 and 1st harmonics to examine voicing changes

• Formant frequencies, transitions, fluctuations, vowel durations


• Recent acoustic studies have continued to identify

unusual vocal behavior characteristics among stutterers.

• On interesting pattern of finding concerns the

fundamental frequency of their pitch.

• Two studies (Adams, sears and Raming, 1982; Ramig

and Adams, 1981) showed the restricting fundamental

range by either monotone or high and low pitched oral

reading may reduce stuttering.


• Even without such manipulations there is evidence that

adult stutterers probably use a relatively restricted

fundamental frequency range (Healey, 1982)..

• Also the formant frequencies of adult stutterers show

relatively little variation, even during changes in their

frequencies of „dysfluencies‟, oral reading rate and vowel

durations under fluency inducing conditions


• It has been found that voice onset & offset actions in

oral reading may partially influence stuttering ( Adams

and Reis,1971,1974), but these findings was not

replicated when young stutterers performed similar oral

readings (McGee, Hutchinson and Depty,1981).

• Also findings of slow voice initiation and termination

among stutterers (Adam and hyden, 1976) appear to be

conditioned by subject age and the severity of stuttering

factors.


• Nevertheless a careful analysis of the speech of the young

stutterers (Wall, Stark weather, and Harris,1981) has

shown that majority of their stutterings occur when a

voice onset is required after a pause; regardless of its

location in a sentence.

• The simplest interpretation of these findings is that

voice timing problems may initially surround stuttering

events, but, with time, they spread to the rest of a

stutterer‟s speech.


• Acoustic analysis have been widely used to indirectly

assess both the temporal (e.g., Healy and Gutkin, 1984;

stark weather, Hirschman and Tannentaun, 1976) and

spatial (e.g. klich and May 1982; Pindzola, 1986)

characteristic of stutterers fluent speech.

• Recent acoustic studies lend support to the findings of

aberrant temporal characteristics without similar

aberration in spatial characteristics during stutterer‟s

fluent speech movements.


Acoustic and articulatory research

• Direct articulatory measurements cause so much discomfort to

patients and are so difficult to perform the alternative indirect

measurements based on the acoustic signal would be very

attractive, if they would give equivalent information.

• The success of the acoustic theory of speech production in

predicting acoustic patterns from articulatory shapes suggest

that the inverse operation, where articulatory movements are

estimated from acoustic measurements should be viable.


• However, the relation between acoustic patterns and articulatory

configurations proves to be far from being one to one (because in

general the dimensionality of the articulatory space is much

larger than that in the acoustic domain).

• To some extent this can be obtained by the dynamic programming

procedures applied to dynamic articulatory gestures.

• Thus sophisticated modeling techniques are necessary in order to

make reliable inferences about articulation from acoustic

measurements.



Babol

Basu

(1979)

5 persons

with &

without

stuttering

( age and

sex

matched )

age 15 to

20 yrs

A passage in

kannada which

had words with

plosives /p/, /t/,

/k/, /b/,/t/ and

/g/ in word initial

position. CV

syllables where „k‟

was a plosive and

„u‟ was the vowel

a wide band

spectrographic

analysis was done

CV syllables revealed

longer VOT,s in persons

with stuttering.

Voicing lag for unvoiced

plosives and voicing lead

for voiced plosives in

both groups.

In normals VOT

increased as the

articulator moved

backwards in the oral

cavity which was not

observed in persons with

stuttering. KUNNAMPALLIL GEJO JOHN

RC perumal

(2001

15 normal and

stuttering

children (6 to

9yrs)

3 age groups

were included(

1yr gap)

Task:

phonation and

termination of

/a/ on hearing

auditory

stimuli

No significant

difference

between voice

initiation time

(VIT) and

voice

termination

time (VTT)

within each

group.


MV

Rames

h

(1983)

4

kannada

speaking

stutterers

and non-

stutterers

( age, sex

and

reading

proficienc

y matched

Material; 2 non

emotional and

meaningful kannada

passage, with equal

number of syllables.

Conditions (1) DAF

(2)NAF (normal0-

duration of delay of

200msec

Measures VOT

vowel duration,

vocal intensity and

Fo, rate of speech

and syllables

Normals manifested

stuttering like behavior under

DAF whereas stutterers

showed a reduction in

dysfluencies

Rate of speech decreased in

both groups under DAF Fo

increased in normals but no

change in stutterers

VOT decreased in stutterers

Vocal intensity increased in

normals KUNNAMPALLIL GEJO JOHN

Suchitra

Sridhar

(1991)

Swapna

Sebastia

n (1997)

10 normals

and 10

stutterers (

age 15 to

20)

5 normals

and

stuttering

adults ( 15

to 25yrs)

Exp 1: task

reading

„rainbow‟

passage,

Exp 2: read „pa‟

„ta‟ „ka‟

Read an „all

voiced‟ and a

combined

passage‟

Speech and

laryngeal output

were recorded

Reduction in VOT

frequency of

stutteringSIT,STT in post

therapy reduced.

Differences in Fo, voice

onset time, word duration

and vowel duration

No differences in formant

frequencies


JFD 28, 2003 pg: 273 – 295

STUTTERING : A DYNAMIC MOTOR CONTROL DISORDER

Christy L. Ludlow, Torrey Loucks

• The purpose of this review article is to determine what neural mechanism may be dysfunctional in stuttering

• Three sources of evidence are reviewed:

1. Studies of dynamic inter – relationships among brain regions during normal speech and in persons who stutter suggest that the timing of neural activity in different regions may be abnormal


2. The brain lesions associated with acquired stuttering are

reviewed. These indicate that in a high percentage of cases, the

primary speech and language regions are not affected but lesions

involve other structures, which may modulate the primary sp & lg regions

3. To characterize the motor control disorder in stuttering,

similiraties and differences from focal dystonias such as

spasmodic dysphonia and Tourette‟s syndrome are viewed.

These indicate that central control abnormalities in stg are not due

to disturbance in one particular brain region but rather a system

dysfunction that interferes with rapid dynamic speech processing

for production


JFD 20, 1995 Pg 157 – 170

SPEECH MOTOR AND LINGUISTIC SKILLS OF YOUNG

STUTTERERS PRIOR TO ONSET

Kloth & Jansen

• Theorist have increasingly suggested that both speech motor and

linguistic factors are involved in the etiology of stuttering.

• This contention has been supported by findings that tend to

indicate that youngsters who stutter have a slower speech rate

and are less linguistically skilled than nonstutterers

• However, no inferences can be drawn from these findings as to

the nature or the causation of this disorder


• This is because the a fore mentioned findings might be a result tather than a cause of the disorder.

• In order to clarify the directionality issue, a multi – year prospective study was undertaken that involved 93 preschool children with parental history of stuttering

• At the initial session, none of the high risk children sample was regarded as having a stuttering problem.

• One year later, 26 children were classified as stutterers. Statistical analysis revealed that prior to the onset of stuttering these children did not differ from the other youngsters studied with respect to either their receptive or expressive language abilities

• However, their rate of articulation was significantly faster


• The latter finding is taken to mean that the children who developed stuttering were not limited in speech motor ability.

• Rather, their fluency failures are seen as a result of a relatively high articulation rate.

• It is noteworthy, in this regard, that the rate of the high – risk children who continued to be viewed as non stutterers was slower than that previously reported for youngsters of their age.

• This suggests that the slower rate served as a buffer against fluency breakdown


JFD 5, 1980 359 – 372

MOTOR PERSEVERATIVE BEHAVIOR IN AFULT SUTTERERS AND NONSTUTTERERS

CHERYL LYNN & COOPER

• The performances of 22 adult stutterers and nonstutterers on motor task requiring the graphic reproduction of a sequence of alternating figures following the production of a series of non alternating figures were studied

• The finding that there were no differences in the performance of the two groups on the alternating motor tasks are interpreted as challenging the validity of previous observations that motor perseverative behavior exists in stutterers and is indicative of a central neurologic deficit in that population


JSHD 57, 1993 Pg 504 – 518

MOTORIC AND LINGUISTIC VARIABLES AMONG CHILDREN WHO STUTTER: A FACTOR ANALYSIS

GLYDON D. RILEY & JEANNA RILLEY

• 76 children who stuttered, aged 5 – 12 were administered tests of motor coordination, psycholinguistic abilities, and stuttering severity

• A factor analysis of 19 selected variables yielded four statistically useful factors that implicated linguistic integration, oral motor ability, and auditory processing as underlying components among the same population

• These components are assumed to affect a child‟s severity was not related to any of the four factors.


JCD 33, 2000, 391 – 428

RESEARCH ON SPEECH MOTOR CONTROL AND ITS

DISORDERS: A REVIEW AND PROSPECTIVE

RAY D. KENT

• This paper reviews issues in speech motor control and a class of

communication disorders known as motor speech disorders

• Speech motor control refers to the system and strategies that

regulate the production of speech, including the planning and

preparation of movements and the execution of movement plans

to result in muscle contractions and structural displacements


• Traditionally, SMC is distinguished form phonologic

operations but in some recent phonologic theories, there is

a deliberated blurring of the boundaries between

phonologic representation and motor functions.

• Moreover, there is continuing discussion in the literature

as to whither a given motor speech disorder, especially

apraxia and stuttering, should be understood at

phonologic level, the motor level or both of these.


THANK YOU


Documents

SPEECH MOTOR CONTROL IN FLUENCY DISORDERS.pdf /KUNNAMPALLIL GEJO