Transcript Language Processing in Children and Adults: Producing Words

Language and Cognition
Colombo June 2011
Day 5
Psycholinguistic Assessments of
Language Processing in Aphasia
Producing Words
Acknowledgement to Jane Marshall
Overview
• This lecture aims to give you an
understanding of:
– A model of word retrieval
– Patterns of word retrieval deficit
– Methods of investigation
– Therapy approaches
– Therapy outcomes
Aphasic word retrieval difficulties
Target
Production
anchor
For holding the sleep steady
when its sunk in water
iron
hoover
bus
sarabang
dart
dark
table
Kurzle, kazle, tazle, tayzle,
table
jacket
helicopter
onion
thustle
Classification
Aphasic word retrieval difficulties
Target
Production
Classification
anchor
For holding the sleep steady circumlocution
when its sunk in water
iron
hoover
Semantic error
bus
sarabang
Mixed error
dart
dark
Phonological error
table
Kurzle, kazle, tazle, tayzle,
table
Conduite
d’approche
jacket
helicopter
Verbal paraphasia
onion
thustle
neologism
To name a seen object involves:
Spoken word
Written word
Auditory
Analysis
Visual
Analysis
Object
Recognition
AIL
VIL
Semantics
POL
Buffer
Motor speech
production
Grapheme to
Phoneme
Conversion
Failure in Object/Picture Recognition
Visual Agnosia
‘A continuous surface infolded on itself. It appears to
have .. five outpouchings’ (re glove)
Item later recognised through touch: ‘it’s a glove’
Intact
Impaired
Vision
Object recognition
Semantic knowledge
Language
Spoken word
Written word
Auditory
Analysis
Visual
Analysis
Object
Recognition
AIL
VIL
Semantics
POL
Buffer
speech
Grapheme to
Phoneme
Conversion
Effects of a Semantic Deficit
• Poor understanding of both written and spoken words
(semantic errors in testing)
• Impaired word production, with semantic errors
• Production affected by semantic factors, like imageability
• Poor performance on non-verbal semantic tasks, like the
Pyramids and Palm Trees Test
Spoken word
Written word
Auditory
Analysis
Visual
Analysis
Object
Recognition
AIL
VIL
Semantics
POL
Buffer
speech
Grapheme to
Phoneme
Conversion
Effects of Deficit at POL
• Knowledge of word meaning is retained, may
circumlocute
• May produce a phonologically related word/non word
• High frequency words may be named better than low
frequency words
• May produce a semantically related word, but should
know that this is not the target
Loss of Entries in POL or Impaired
Access?
Evidence:
• Naming is often inconsistent
• Cues assist naming, such as providing the first
phoneme
Argues for access problem
Phonological Assembly
Whole word phonology has to be ‘assembled’ prior to
speech
Assembly failure:
- Phoneme omissions, substitutions and exchanges
- Conduite d’approche – sequence of phonological errors
getting closer to the target
- Length effect (short words easier than long)
Factors Affecting Naming
•
•
•
•
Frequency/familiarity
Age of acquisition
Word class (noun>verbs)
Word category (animate > inanimate)
• Note these variables can interact e.g.
Common words may be acquired early
Investigating a Word Retrieval
Impairment
Questions to Ask
•
•
•
•
•
Is word retrieval failing?
What is the extent of the problem?
Does it impede everyday communication?
Is it a therapy priority for the aphasic person?
Why is word retrieval failing: what is the level of
breakdown?
• What helps?
• Is the person using any strategies?
RS (Marshall et al 1990)
J
Can you tell me how far you have got with selling your
business R?
R
er ………… Mr N
Mrs S: Your accountant
R
Yes … I’ve left it to him
J
And how far has he got with it?
R
er … one chap has come up with a er …………… fee
… but there’s three more coming
J
That’s quite good isn’t it? Are they offers that you can
accept?
R
Not really
J
So you want slightly more?
R
Yes
J
How quickly do you want to sell it?
R
As soon as possible … just for me to …. Call it a day
… but it could take as long as three months
J
What will you do with the capital?
R
Put it into the …… one in the …. What’s name ……
bank
J
What’s happening to the staff?
R
er …….. (waves) goodbye … goodbye
Assessment Plan
• Aims to find out:
• The extent of R’s naming problem
• Whether he can be cued
• Where word retrieval is breaking down
– Semantics
– Phonology
Semantic Tests:
Pyramids and Palm Trees
PALPA Spoken word to picture
matching
Semantic Tests:
• Synonym Judgement (PALPA)
–
–
–
–
Shovel
Shovel
Menace
Menace
Spade
Tale
Threat
Discovery
• Judging Picture Names (Informal Task)
Is this a banana?
Is this an apple?
Results
•
•
•
•
Pyramids and Palm Trees
Spoken word to picture matching
Synonym judgements (concrete)
Judging picture names
3 errors
98%
95%
100%
Naming Test
Say the names of 30 pictures
• If cannot is given phonological cue
– Target (/t/ for tiger)
– Miscue (/l/ for tiger)
Results
• 10 pictures named
• Correct cues elicit correct names
• Miscues elicit no response
Interpretation
Word retrieval is not failing at semantics:
• R has good understanding of concrete words
and pictures
• He can understand words that he cannot name
• He cannot be miscued into making semantic
errors
Phonological Representations are retained
• R responds to target phonological cues
More Evidence re Phonology
Reading aloud 100%:
• Concrete words
• Abstract words
• Spell regular words (rabbit)
• Spell irregular words (yacht)
• The phonologies of words are available and can
be accessed from the written word
Problem is connection between semantics and phonology:
Spoken word
Written word
Auditory
Analysis
Visual
Analysis
Object
Recognition
AIL
VIL
Semantics
POL
Buffer
speech
Grapheme to
Phoneme
Conversion
Therapy
• Treatment needs to help RS reconnect
semantic with phonological representations
• Treatment exploits two strengths: reading aloud
and semantic discrimination
• 3 hours therapy using word to picture matching
Therapy Example
Radio
Television
Hi Fi
Computer
Camera
Task: Find the correct word and read it aloud
Discuss differences between target and foils
Results: % correct in picture naming
100
90
80
70
60
50
40
30
20
10
0
pre
post
f up
treated words
untreated words
foils
Conclusions from RS
• Therapy improves naming of treated words
• Effects are very well maintained (still evident one
year later)
• There may be some generalisation to related
words that appeared in therapy as foils
GF (Robson et al 1998)
• L CVA
• Jargon Speaker
• Very incomprehensible speech
I was quite erm that’s why I can’t get weyerd keep ... erm makes
me very um here up here makes him all /s/ all mingsing but these
come and I can’t it might be because I had another setoid no sort of
um I mean but when you cough you different but when you right you
lie to her ...’ (replying to a question about her holiday)
Input Tests
•
•
•
•
Pyramids and Palm Trees
Auditory lexical decision
Spoken word to picture match
Spoken synonym judgements
4 errors
100%
98%
87%
Conclusions: GF can access semantic
representations of pictures and concrete words
Production
Naming pictures
1/40
– Some help from phonological cues (5/15)
– Not helped by semantic cues (irritated by them!)
Her comments:
‘I had it there and then it went’
• Reading aloud 10/40 (regular and irregular
words equal success)
Conclusions
• Naming is not failing at the semantic level (input
tests)
• Entries are retained in POL, since GF can read
some words
• Words/non-words same: not reliant on GPC
• The problem is mainly between semantics and
phonology
Objects/pictures
Spoken word
Auditory
Analysis
AIL
Written word
Object/
Picture
Recog
Visual
Analysis
VIL
Semantics
GPC
POL
Buffer
speech
Therapy Decisions
• Help GF access POL
• Use good input skills/monitoring
• Help GF use any partial phonological knowledge
about words
• Aim to develop a phonological self cueing
strategy
• Tasks: making phonological judgements about
target words
Therapy Example 1
Task: How many syllables (‘beats’)
i)
ii)
‘Carrot’
1
2
1
2
Therapy Example 2
Task: Initial phoneme judgement:
i)
‘carrot’
/k/
/m/
ii)
‘carrot’
/b/
/k/
iii)
/k/
/m/
iv)
/m/
/l/
/k/
/b/
Therapy Example 3
i) Indicate number of syllables
ii) Indicate first phoneme
iii) Produce first phoneme
iv) Attempt to name picture
Results: % correct in picture naming
80
70
60
50
pre
post
f up
40
30
20
10
0
treated words
phon related
controls
unrelated
controls
Conclusions from GF
• Phonological therapy has improved naming of treated
words
• Good generalisation to untreated words
• Stable baseline and no change in unrelated tasks
(therefore effects can be attributed to therapy)
Why is generalisation achieved?
• Self cueing strategy (but little evidence of this in testing)
• Generalised recovery of access to POL
Conclusions
• Using careful assessment we can try to pin point
why word retrieval is failing
• This can inform our therapy with clients
• A number of studies show that therapy can
improve picture naming, with variable
generalisation to untreated words
• Hickin et al (2007) suggest that therapy may
also improve everyday speech, but only when
effects generalise beyond treated words
References
• Hickin J, Herbert R, Best W, Howard D, and Osborne F (2007)
Efficacy of treatment: effects on word retrieval and conversation. In
S Byng, K Swinburn and C Pound (eds) The Aphasia Therapy File,
Psychology Press.
• Marshall J, Pound C, White-Thomson M, Pring T (1990) The use of
picture/word matching tasks to assist word retrieval in aphasic
patients. Aphasiology 4, 167 - 184.
• Robson J, Marshall J, Pring T and Chiat S (1998) Phonological
naming therapy in jargon aphasia: Positive but paradoxical effects.
Journal of the International Neuropsychological Society, 4, 675 686. (available from Jane)