Treatment of Word retrieval impairments: Do we know which tasks work for whom and why? Continued! Associate Professor Lyndsey Nickels National Health.

Treatment of Word retrieval impairments:

Do we know which tasks work

for whom and why?

Continued!

Associate Professor Lyndsey Nickels

National Health and Medical Research Council

Senior Research Fellow,

Macquarie Centre for Cognitive Science (MACCS),

Macquarie University, Sydney, Australia

How do we decide which treatment?

Each different level of breakdown in word production will be best remediated by a different type of treatment

(e.g. Hillis & Caramazza, 1994; Nettleton & Lesser, 1991)

• impaired word meaning (semantics)

→ treatment focusing on meaning

• impaired retrieval of the phonological form from semantics

→ treatment focusing on providing/accessing the phonological form

• impaired phoneme level/phonological encoding→ treatment focusing on phonemes

Remember…. Therapy for word

retrieval DOES NOT have to include word retrieval

How do we decide which treatment?

Each different level of breakdown in word production will be best remediated by a different type of treatment

(e.g. Hillis & Caramazza, 1994; Nettleton & Lesser, 1991)

• impaired word meaning (semantics)

→ treatment focusing on meaning

• impaired retrieval of the phonological form from semantics

→ treatment focusing on providing/accessing the phonological form

• impaired phoneme level/phonological encoding→ treatment focusing on phonemes

What treatment is appropriate?

Semantic impairments

The most successful therapy seems to involve exploring the semantic attributes of a stimulus.

e.g. Boyle & Coelho, 1995.

Coelho, McHugh & Boyle, 2000.

Hillis, 1991, 1998.

Nickels & Best, 1996.

Phonological Output Buffer

Speech

Phonological Output Lexicon

Lexical Semantics

Therapy for semantic impairments

e.g. Semantic feature analysis (SFA) (Boyle & Coelho, 1995; Coelho, McHugh & Boyle, 2000)

• name a picture• generate features relating to …..

Exploring semantic attributes of a stimulus

GROUP: is an ANIMAL

USE: is used for protection

ACTION: does what? Barks

LOCATION: is found at home

ASSOCIATION: reminds me of a kennel

Can we be sure this is effective?

Not entirely … they use very small numbers of items in

the sets, and examine changes in performance

using visual inspection with no statistics

Therapy for semantic impairments (cont)

e.g. Nickels & Best (1996) AER

“Relatedness judgements” (with feedback)

Statistically significantly improved naming of treated and untreated stimuli

Nickels & Best argue that feedback is

critical to obtaining generalisation to untreated items


e.g. Hillis (1991, 1998) HG.

Red vs yellowRound vs oval

Thin skin vs thick skinSweet vs sour

Lemon


e.g. Hillis (1991, 1998) HG. • name a picture, • if and semantic error was produced, • → drawing of error • compare drawing to the target picture • discuss semantic distinctions

• Significant improvement in written naming (treated)• also in spoken naming and word comprehension • Improvement of treated items and untreated items in the

same semantic category(but this may just be that untreated items appeared in therapy)

How does the treatment for

semantic impairment work?

Doesn’t it improve the

semantic impairment?

If so what would the effects be?


Speech output


Lexical Semantics

Writing

Heard SpeechPrint

Idea, Picture, or seen object

Lexical Semantics


Speech output


Lexical Semantics

Writing

Heard SpeechPrint


Lexical Semantics


Speech output


Lexical Semantics

Writing

Heard SpeechPrint


Lexical Semantics


Speech output


Lexical Semantics

Writing

Heard SpeechPrint


Lexical Semantics

Semantic therapy for semantic

impairments should improve

all modalities

But although this was true for

Hillis’ study with HG, it was not

for AER.Improvement in all

modalities only occurs if treatment improves semantic

processing

How did the authors think these treatments for semantic impairments were having their effects?

Hillis (1998) “improvement at the semantic level itself – perhaps through increased specificity of semantic representations of trained items”

Coelho et al (2000) - a strategy for word retrieval“by activating the semantic network surrounding the target word, that word may be activated above its threshold, thereby facilitating retrieval”

Nickels and Best (1996)

“the use of a strategy (which may be unconscious) of exploring the semantics of an item … which facilitates retrieval of that item (perhaps by increasing the semantic information addressing the output lexicon…)”

Consistent with improvement

across modalities

Consistent with improvement

restricted to speech production

• Semantic tasks– Those tasks which involve focusing on word meanings

• Semantic tasks may be used to treat semantic impairments– Those tasks which seem to most often improve semantic

impairments involve reflecting on semantic features

– However even then they may NOT improve semantics itself but may instead provide a strategy for improving word retrieval.

Didn’t we contrast semantic impairments

and post-semantic word retrieval impairments?

Semantic tasks and semantic impairmentsHold on! How can

therapy for semantic impairments improve word retrieval without improving semantics

Yes, but semantic impairments cause word

retrieval impairments …

Semantics

Object, picture or idea

purrs

cat dog rabbit fish

barksfur pet4-legs scales

houserobin

Phonological Lexicon

PhonologicalBuffer/

Phonemesæ td o gk

Post semantic word retrieval impairment

Semantic impairment

Sem impairment

Semantics


purrs

cat dog rabbit fish


houserobin


PhonologicalBuffer/

Phonemesæ td o gk

Semantic impairment

Semantics


purrs

cat dog rabbit fish


houserobin


PhonologicalBuffer/

Phonemesæ td o gk

Semantic impairment also leads to a

reduction in activation at the word form level

Semantic impairment

Although, as we will see, this is usually

only the case in people with less severe semantic

impairments.So tasks can be

effective in improving word

retrieval WITHOUT improving the

semantic impairment

SUMMARY: What treatment is appropriate?

Semantic impairments

The most successful therapy seems to involve exploring the semantic attributes of a stimulus.

e.g. Boyle & Coelho, 1995.

Coelho, McHugh & Boyle, 2000.

Hillis, 1991, 1998.

Nickels & Best, 1996.


Speech


Lexical Semantics

Any questions?

Therapy for semantic impairments or semantic tasks as therapy?

• Virtually all tasks involve semantic processing• Some focus on semantic processing to a greater extent

(we tend to refer to these as semantic tasks)

e.g.odd one out

(spoken or written) word to picture matching

These tasks are widely used in the remediation of word-retrieval impairments, however, their use is not restricted to those individuals with semantic impairments.

word-picture verification

Semantic tasks for improving word retrieval

The use of semantic tasks is not restricted to individuals with semantic impairments…...

semantic tasks can improve word production even for those individuals with good semantic processing and even when the tasks are performed accurately

(e.g. Nickels & Best, 1996).

…. In fact they are probably more likely to improve word production for those with relatively less semantic impairment – those with ‘post-semantic lexical retrieval impairments’.


Word retrieval impairments

Tasks focusing on semantics and phonology

- improve word retrieval

e.g. Howard et al 1985

Nickels & Best 1996 Phonological Output Buffer

Speech


Lexical Semantics



Tasks involve activation of both semantics and phonology

But may focus more on semantics….


Speech


Lexical Semantics



All the tasks involve activation of both semantics and phonology

But may focus more on semantics

or phonology


Speech


Lexical Semantics

Repeat “kangaroo”

It starts with /k/



All the tasks involve activation of both semantics and phonology

They produce long lasting, item specific effects in the majority of individuals with impaired activation of the correct target in the phonological lexicon

Improves likelihood of the target being sufficiently activated to be retrieved successfully.


Speech


Lexical Semantics

Let’s explore these tasks that improve word retrieval in

more detail

“Semantic” tasks as therapy

Word-picture matching

Word-picture verification

Widely effective (Howard et al, 1985; Marshall et al, 1989)

Long lasting effects (Pring et al., 1990)

Generally, lasting effects are item specific

These tasks typically don’t require word

production but improve naming at a later point

nevertheless

The tasks work best with people that can do them easily (people with

less of a semantic problem)

Howard et al (2006)

Compared the efficacy of word-picture matching with related and unrelated distractors

(related distractors = deeper processing = more effective?)

How important is the emphasis on

semantic processing?

Howard et al (2006)


All incorrect before word-picture matching

After word-picture matching

Related distractors: 45% correct

Unrelated distractors: 50% correct

Controls: 30% correct



No evidence that deeper semantic processing

improves efficacy of word-picture matching

Not even for those with more severe

semantic impairments?

Howard et al (2006)


All incorrect before word-picture matching

After word-picture matching

Less sem imp. More sem imp.

Related distractors: 50% 40%

Unrelated distractors: 60% 40%

Controls: 40% 23%



No evidence that deeper semantic processing improves efficacy of

therapy: not even for those with more severe semantic impairments

Martin, Laine et al (2000, 2003, 2004, 2005) ‘Contextual Priming’

Compared the efficacy of naming/repetition in the context of semantically related distractors and unrelated distractors.



Martin, Laine et al (2000, 2003, 2004, 2005) ‘Contextual Priming’



Martin, Laine et al (2000, 2003, 2004, 2005) ‘contextual priming’

Compared the efficacy of naming/repetition in the context of semantically related distractors and unrelated distractors.

- Naming is worse at the time with semantically related distractors.

- Naming is improved no more (at a later point) with semantically related distractors.



No evidence that a semantic context

improves efficacy of therapy



Not as important as we thought!

• Tasks which have traditionally been labelled ‘semantic’, such as word-picture matching are equally effective with unrelated items.

• Adding a semantic context to a (naming/repetition) task does not make it more effective.

Word form is usually provided.

e.g. Le Dorze et al (1994)

“Show me the octopus”

“Show me the mollusc with long legs”

How important is the phonological form in the semantic task?

Most probably critical to its effectiveness

Therapy: The application of a

technique several times, over days, weeks

or months

Immediate Cueing effects: The effect of a technique at the time of

its application

Nickels, McDonald, Makin, Moses & Taylor –Facilitation Experiments

What is a facilitation

experiment?

It looks at the effect of a task performed ONCE on

another task at a later point and contrasts with…

Cueing, Facilitation, Therapy

“k”

kangaroo


“k”

kangaroo

….later


“k”“k”

“k”“k”

“k”“k”

“k”“k”

“k”

kangaroo


Phonological

Repetition

“kangaroo”

Semantic without the word form

Feature verification

Does it hop?


Name 300 pictures until 95 failed (10 second limit)

Sort the failed items into 3 (frequency & length)matched sets

Facilitation Task 1 Facilitation Task 2 Control

Facilitation

Name all 3 sets of pictures+ 30 easy fillers

1 week later

10 mins later

Comparing repetition and a semantic task with no word form

-0.20

-0.10

0.00

0.10

0.20

0.30

0.40

0.50F

ME

JNI

KC

C

PS

J

DR

S

JUE

SJS

JMM

DIP

MC

B

WV

K

FA

B

JOW

MF

T

SM

G

Gro

up

participant

ben

efit

co

mp

ared

to

co

ntr

ols semantic

phonological


-0.20

-0.10

0.00

0.10

0.20

0.30

0.40

0.50F

ME

JNI

KC

C

PS

J

DR

S

JUE

SJS

JMM

DIP

MC

B

WV

K

FA

B

JOW

MF

T

SM

G

Gro

up

participant

ben

efit

co

mp

ared

to

co

ntr

ols semantic

phonological


-0.20

-0.10

0.00

0.10

0.20

0.30

0.40

0.50F

ME

JNI

KC

C

PS

J

DR

S

JUE

SJS

JMM

DIP

MC

B

WV

K

FA

B

JOW

MF

T

SM

G

Gro

up

participant

be

ne

fit

co

mp

are

d t

o c

on

tro

l pic

ture

s

semanticphonological

Very little benefit from a semantic task without the word form for most

individuals

Semantic tasks without the word form

We also looked at the same tasks used as therapy with 4 individuals with aphasia

… once again, semantic tasks without the word form had little benefit

…. for example….

DRS

8 sessions of semantic therapy (feature verification)

over 2-3 weeks

Before each therapy session named all those items that were to be treated, and a set of control pictures (50 items in each set).

DRS - Semantic Therapy

0

5

10

15

20

25

30

35

40

1 2 3 4 5 6 7 8

daily pretests

nu

mb

er c

orrect

( /5

0)

no. correct THERAPY

no. correct CONTROL

Linear no. correct THERAPY

Linear no. correct CONTROL

McNemar Day 1 vs Day 8

Wilcoxon 1 sample

Sem Treat ns 0.028773653

Sem control p= 0.0117 0.04377147

Sem treat vs sem control Wilcoxon 2 sample NS

No extra benefit from the

semantic task

But the untreated items seem to improve.....

Spontaneous recovery?Generalisation?

Benefits from repeated naming attempts?

0

5

10

15

20

25

30

35

40

Pre-test 1 Pre-test 2 Post sem 1 Post sem 2

Treated (sem)

Naming control

Unseen

0

5

10

15

20

25

30

35

40


Treated (sem)

Naming control

Unseen

Treated items and repeatedly named items

improve to the same degree – so there is no specific benefit from the

semantic task

0

5

10

15

20

25

30

35

40


Treated (sem)

Naming control

Unseen

Treated items and repeatedly named items

improve to the same degree – so there is no specific benefit from the

semantic task

The improvement from naming controls isn’t

generalisation or spontaneous recovery (unseen controls don’t

improve) – but benefit from repeated attempts at

naming

Summary: semantic tasks in word retrieval

• Degree/depth of semantic processing appears not to be critical (Howard et al)

• Presenting tasks in a semantic context can interfere short-term and mostly has no long term advantage over presentation in unrelated contexts (Martin, Laine et al)

• Presence of phonological form appears critical in most cases (Le Dorze et al; Nickels et al: facilitation studies, DRS therapy study)

…… and so to Phonological tasks

Any questions?

“Phonological” tasks

• Repetition of target• Reading aloud• Phonological & orthographic cueing• Rhyme judgements • Syllable and phoneme counting• Phoneme segmentation• Anagrams

…….but nearly always in the presence of the picture.

Again we need to remember the distinction between the nature of

the task and the nature of the impairment…. Here we are talking about phonological tasks but NOT impairments to the phoneme level.

Phonological tasks and word retrieval

• Widely argued to be the most appropriate for impairments in retrieval of (or damage to) the phonological form from the phonological output lexicon (e.g. Hillis & Caramazza, 1994; Miceli,et al, 1996; Nettleton & Lesser, 1991)

Miceli et al (1996): as these tasks focus at the level of activation of individual entries in the phonological output lexicon, their effects should be item specific

– a result of ‘priming’ retrieval of the phonological form.

For many years ‘semantic’ tasks were thought to be

more effective than ‘phonological’ tasks in

improving word retrieval

How true is this?


-0.20

-0.10

0.00

0.10

0.20

0.30

0.40

0.50F

ME

JNI

KC

C

PS

J

DR

S

JUE

SJS

JMM

DIP

MC

B

WV

K

FA

B

JOW

MF

T

SM

G

Gro

up

participant

ben

efit

co

mp

ared

to

co

ntr

ols semantic

phonological


-0.20

-0.10

0.00

0.10

0.20

0.30

0.40

0.50F

ME

JNI

KC

C

PS

J

DR

S

JUE

SJS

JMM

DIP

MC

B

WV

K

FA

B

JOW

MF

T

SM

G

Gro

up

participant

be

ne

fit

co

mp

are

d t

o c

on

tro

l pic

ture

s

semanticphonological

Very little benefit from a semantic task without

the word form … repetition is almost

always more beneficial

DRS - Phonological Therapy

0

5

10

15

20

25

30

35

40

45

50

1 2 3 4 5 6 7 8

daily pretests

nu

mb

er c

orrect (

/50)

No. correctTHERAPY

no. correctCONTROL

Linear no.correctTHERAPY

Linear no.correctCONTROL

McNemar Day 1 vs Day 8

Wilcoxon 1 sample

Phon Treat p= 0.0001 < .0001Phon control p= 0.0042 0.002304428

Phon treat vs phon control

Wilcoxon 2 sample

p(1 tail) =.048

Significant extra benefit from the repetition task

For many years ‘semantic’ tasks were thought to be

more effective than ‘phonological’ tasks in

improving word retrieval

This seems NOT to be true in general, particularly if

the semantic task does not include the word form

Semantic & phonological treatments: an overstated distinction?

• BOTH have semantic and phonological components– Semantic tasks without the phonological form are generally

not effective

• Howard (2000)

– tasks having the same effects in the same way

• Strengthening the connections between semantics and phonological form when both are simultaneously active

• equivalent effects for semantic and phonological tasks for most individuals

Semantics


purrs

cat dog rabbit fish


houserobin


PhonologicalBuffer/

Phonemesæ td o gk

Semantics


purrs

cat dog rabbit fish


houserobin


PhonologicalBuffer/

Phonemesæ td o gk

Semantics


purrs

cat dog rabbit fish


houserobin


PhonologicalBuffer/

Phonemesæ td o gk

‘Semantic’ and ‘phonological’ tasks are working in similar ways

By strengthening the connections between

semantics and word form by activating both

semantics and phonology

The role of effort and error

• Effortful >benefit than automatic? – Greater semantic processing

• Errorless >benefit than errorful?

How does amount of ‘effort’ or accuracy in

a task affect the benefit for naming?

Comparing Phonological Tasks

“kangaroo”

Repetition

Effortless, ‘shallow’ processing,

error reducing

“k”

Phonemic cueing

Effortful, deeper processing,

errorful

Method

Name 300 pictures until 95 failed (10 second limit)

Sort the failed items into 3 (frequency & length)matched sets

Facilitation Task 1 Facilitation Task 2 Control

Facilitation

Name all 3 sets of pictures+ 30 easy fillers

1 week later

10 mins later

-0.10

0.00

0.10

0.20

0.30

0.40

0.50

FM

E

JNI

KC

C

PS

J

DR

S

JUE

SJS

JMM

DIP

MC

B

WV

K

FA

B

JOW

MF

T

SM

G

Mea

n

Phon Cues

Repetition

*

*

*

*

*

**

*

Facilitation effect compared to control pictures:Errorless/Automatic (Repetition) vs Errorful /Effortful (Phonemic cues)

-0.10

0.00

0.10

0.20

0.30

0.40

0.50

FM

E

JNI

KC

C

PS

J

DR

S

JUE

SJS

JMM

DIP

MC

B

WV

K

FA

B

JOW

MF

T

SM

G

Mea

n

Phon Cues

Repetition

Facilitation effect compared to control pictures:Errorless/Automatic (Repetition) vs Errorful /Effortful (Phonemic cues)

*

*

*

*

*

**

*

Comparing Phonological Tasks

“kangaroo”

Repetition

Effortless, ‘shallow’ processing,

error reducing

“k”

Phonemic cueing

Effortful, deeper processing,

errorful

> benefit than

No need for it to be effortful to be

helpful – FOR MOST PEOPLE

Well, repetition is better and it does improve more BUT

the proper comparison is within

a task ….

Is errorless (or error reducing) better than errorful?

Fewer errors means greater benefit?

Repetition is more error-free than cueing… but the proper comparison is within a task

Cueing

Do the people who are more accurate with cueing (at the time) have more benefit for naming (later)?

-0.2

-0.1

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

MFT

DRSFAB

FME

KCCSJS DIP JN

IJM

MM

CBPSJ

JUE

WVK

SMG

JOW

Phon Cue - Facilitationaccuracy

Phon Cue naming benefit

Errorless items benefit more?

Repetition is more error-free than cueing… but the proper comparison is within a task

Cueing• no significant correlation between success of

the cue during facilitation and benefit of cueing for subsequent naming.

• i.e. errorless is not better than errorful for cueing.Why did we think errorless might be better in the first

place?

Errorless learningHypothesis• Remediation is more effective if errors are prevented

- the act of producing an error may strengthen the incorrect association, and make the correct response less likely to occur.

History• Animal learning (Terrace, 1963)

• Children with developmental learning difficulties (e.g. Sidman & Stoddard, 1967).

• Acquired memory impairments (amnesia) (e.g. Baddeley & Wilson, 1994)

Errorless learning and anomia treatment

• Fillingham et al (2003)“error reducing techniques do have positive effects for patients with

word finding difficulties. As yet there is limited information on which to judge whether this technique is significantly advantageous over errorful approaches” (p358).

• Fillingham et al (2005, 2006)– No significant difference between errorless (repetition) and errorful

(cueing) tasks.

• Abel et al (2005)– No significant difference between increasing and vanishing cues

methods (errorful vs error reducing).

BUT nonetheless repetition may be more

beneficial than phonemic cueing for

many people

There appears to be no difference between

errorful and errorless techniques in aphasia

Remember that this will all depend on the processing strengths and weaknesses of

the aphasic individual

Any questions?

•How semantic & phonological tasks

work

•Error & Effort

Generalisation in the treatment of word retrieval.

• the most successful treatment is one which effects improvement not only for the items used in therapy but also for any other item, in any other context.

• many treatments produce clear long lasting effects on the treated items, generalisation to untreated items is

less common, and when obtained, often less robust (e.g. Nickels and Best, 1996b)

When perceived generalisation is actually an effect of (assessment during) treatment!

DRS

0

5

10

15

20

25

30

35

40


Treated (sem)

Naming control


DRS

0

5

10

15

20

25

30

35

40


Treated (sem)

Naming control

Unseen


Howard et al (1985) “Naming controls”• naming daily but NOT treated• Improved significantly

• Widely interpreted as generalisation of treatment effects– more probably the result of repeated attempts at naming (Howard, 2000)

Nickels (2002) JAW- Repeatedly attempted to name a set of pictures (daily for a

week)- Significantly improved naming

Practice makes (closer to) perfect:trying to name helps naming!

HOW?- some names will be successfully retrieved on one occasion but not on

another.

- When, by chance, an individual produces a picture name successfully, both the semantic representation and the phonological representation for that item are simultaneously active.

- This will then strengthen the mapping for that item, making it more likely that the word will be produced correctly on a subsequent occasion (which will again strengthen the mapping).

- Over time, more items will have strong enough mappings to be produced accurately on every occasion

Or perhaps just partially activating the phonological

representation when trying to name is enough to strengthen

the mappings

More work needs to be done to determine who

this works for, how close together naming

attempts have to be etc etc

But it could have important

implications-Use it or lose it- (Attempts at)

conversation might improve word retrieval- (Trying to) talk about the same topics may be

better

When generalisation is a reflection of (strategic?) changes in processing

• Semantic tasks (with semantic disorders) Techniques promoting reflection on semantic properties may be better

viewed as teaching a strategy, albeit an unconscious strategy.

• Phonological tasks– Best (2006) examined who showed generalisation using a cueing

treatment – it was those individuals with relatively more of a phonological impairment (and less semantic impairment)

• Self-cueing– E.g. phonologically mediated orthographic cueing

Mostly generalisation is limited…

Therefore the emphasis is on the clinician to ensure that all items used in therapy are functionally relevant and chosen in collaboration with the person with aphasia.

Nickels et al – in progress

‘Home Programme’

• Using the most successful task (repetition)

• 60-100 items chosen by the person with aphasia as being words they wanted to be able to say.

• Therapy presented by computer in powerpoint

• ‘Structured’ conversation used to evaluate carry-over from naming to conversation

Treatments aimed at remediating word production impairments

Tasks can be • effective (e.g. Hillis & Caramazza, 1994; Howard et al, 1985b),

• produce durable effects (e.g. Pring et al, 1990),

• be administered by clinician &/or computer (e.g. Fink et al, 2002)

• be obtained on verbs & nouns (e.g. Murray & Karcher, 2000; Raymer and Ellsworth, 2002).

• carry-over into connected speech & conversation (e.g. Hickin et al, 2002),

BUT

treatment tasks are not invariably effective

Do impairment-based treatments impact on functional skills and increase activity &

participation?• word retrieval/production underlies every attempt at verbal

communication hence improving word retrieval impairments is of course functionally relevant

Simplistic and/or fundamentally misguided

BUT

the importance of an increased ability to name treated items following therapy should not be underestimated

Hillis (1998) HG

- small change in impairment – significant gains in activity & participation with the use of functionally relevant personal items (Bacardi & Coke!)

Do impairment-based treatments impact on functional skills and increase activity &

participation?• Yes …. (functional gains can be achieved)• but…. (they are by no means guaranteed)

• few studies that have formally evaluated the effects on disability of impairment-based remediation.

• It seems particularly important given the prevalence of item specific effects that remediation should include personally relevant stimuli.

Any questions?

•Generalisation•Effects of repeated naming•Functional relevance

So far we’ve talked about tasks that aim to improve the underlying

word retrieval impairment – some

tasks may work in other ways too

Lets take another look

at cueing

?

?

/t/

tomato

/t/

?

T

tomato

TLetter cueing has been found to have the same effects as phonemic cues

– both immediate and longer term benefits for naming for some individuals

(Best et al. 2002)

?

Some individuals can generate their own orthographic

cues

?

TOMATO

tomato

TOMATO

How does this work?

?

T -> /t/

How does this work?

Phonologically-mediated self-cueing (e.g. Bachy-Langedock & DePartz, 1986; Nickels, 1992)

tomato

T -> /t/


Requires phonological

cueability

Requires access to the written

form when the spoken form is

unavailable Requires an ability to convert

letters into sounds


Requires phonological

cueability

Requires access to the written

form when the spoken form is

unavailable Requires an ability to convert

letters into sounds

Can use a computer

cueing aid to do the conversion

Can be retaught

Generating phonemic cues from the initial letter

1. Spoken naming

2. Written naming

3. Convert letters to sounds

dog

Nickels (1992) TC

Generating phonemic cues from the initial letter

1. Spoken naming

2. Visualise written word

3. Sound out initial letter

& cue word production

Nickels (1992)

retaught letter-sound correspondences

ddog

dog

• This improved TCs spoken naming to almost the same level as his written naming. • He used this spontaneously in conversation.• Could be (and was) used for any word he was trying to retrieve (only fails for words with irregular initial letters e.g. onion, Cinderella)

Using a computer to generate phonemic cues from the initial letter

e.g. Bruce & Howard (1987); Best et al. (1997)

/d/

dog

d

3. Press letter 2. Visualise first letter

4. Computer produces phoneme

5. Cue word production

1. Spoken naming

Generating phonemic cues

When spoken naming

IF individuals can identify initial letter

convert letters to sounds

and are phonemically cueable

Then they can generate their own cues.

If unable to convert letters to sounds

- Can be retaught

- Can use computer generated cues

- may be able to use the letters without converting to sounds – Direct orthographic cueing.

Direct Orthographic Cueing

SD: Howard & Harding (1998)

• impaired access to the phonological output lexicon (good semantics)• Unable to form letters

• Given an alphabet board – naming improved dramatically

• Was not using phonological mediation• unable to convert letters to sounds• unaffected by initial letter regularity (e.g. onion; eye)

Self-generated orthographic cues

When spoken naming

IF individuals can identify initial letter

They may be able to use direct orthographic cueing

• If they are phonemically cueable

They may be able to use a computer to generate the cues

• If they can (or be taught to) convert letters to sounds

They may be able to generate their own phonological cues.

Any questions?

•Self-generated cues

When things don’t go the way you

planned!

Using a cueing aid as therapy – who benefits?

Best, Howard, Bruce & Gatehouse (1997)• 13 individuals treated with the cueing aid

• predicted benefit only for those who when spoken naming

can identify initial letter & are phonemically cueable

•They should be better at naming with the aid than without it (this is providing the cues).

• Only two individuals had both skills

After treatment:

Only 1 individual had better naming with the aid than without.

BUT 12 showed significant improvement in naming.

WHY???

Why does treatment with a cueing aid help individuals who are not predicted to benefit?

Treatment tasks can work in different ways for different individuals

Multicomponent treatment • orthography • phonology • semantics (presentation of the picture and

attempted naming).Only one individual was

helped by the aid providing the cue - for many others this was a standard word

retrieval therapy

Phonemic cueingRepetition in the presence of the

picture

Cueing aid reorganising the naming system: JOW (Best et al, 1997)

• Substantial and long-lasting effects of treatment• Improvement in treated and untreated items

• treatment drew attention to the relationship between orthography and phonology

• altered automatic (not strategic) processes in his word retrieval

• A direct orthographic cueing mechanism

Why does treatment with a cueing aid help individuals who are not predicted to benefit?

Treatment tasks can work in different ways for different individuals

…. especially when they are multicomponentHow tasks actually work may not be

how you planned them

to!And sometimes they don’t work

even though you would

predict they should!

RECAP

BEFORE THE BREAK• Assessment of level of breakdown in word production• Methodological issues in evaluating treatment

RECAP (continued)

AFTER THE BREAK• Improving naming when there are semantic impairments

– Exploring semantic features – It may work by providing a strategy to improve naming rather than

improving semantics itself.

• Word retrieval impairments– Semantic vs phonological tasks– Semantic tasks require the word form, depth of semantic

processing is not important– Phonological tasks are also effective– Tasks are working in the same way

• Effort & error– Tasks do not need to be effortful– There is no additional benefit from errorless performance

RECAP (continued)• Generalisation

– Treatment for Word retrieval impairments usually result in item specific improvements

– but may generalise to conversation, and can produce effects on participation/quality of life

– Use of functionally relevant items is a priority– Repeated attempts at naming may improve word retrieval for some

individuals– Generalisation can result from strategic/compensatory approaches

• Self-generated Cues– Phonologically mediated/computer generated/direct orthographic

• Tasks may not always work the way you expect

• ALL OF THE ABOVE ARE GENERALISATIONS

– every person with aphasia is different and could be the exception to these general principles

Relationship between task, impairment and efficacy

Therapist’s dream = Prescription guide

“to unambiguously pair a particular functional impairment with a treatment task that has guaranteed success for that impairment”

Clinicians need to know what to do NOW!!!

• choose a therapy that has the best chance of succeeding

• Choose a task that has successfully improved naming for a wide variety of people with spoken word retrieval impairments

- multicomponent approaches (e.g. cueing hierarchies or multimodality/multi-task therapies – see e.g. Best et al, 1997; Hickin et al, 2002)

• Use pilot studies using different tasks for short periods (e.g. 1 week) to establish which tasks seem effective and which do not prior to continuing with the most effective therapies over longer period.

Prove that it works!

• ensure that it is possible to establish whether or not a particular therapy is effective by using a sound methodology.

We have a duty to be able to demonstrate that a therapy is effective to the individuals with aphasia, their families/friends/carers, those funding the treatment and not least to ourselves.

Thank you!

Any questions

or comments?

Treatment of Word retrieval impairments: Do we know which tasks work for whom and why? Continued! Associate Professor Lyndsey Nickels National Health.

Documents

semantic level

semantic processing

semantics treatment

semantic error

semantic attributes

semantic features

semantic information

semantic network