Treatment of Abstract and Concrete Words for Lexical Retrieval in Aphasia

1
Treatment of Abstract and Concrete Words for
Lexical Retrieval in Aphasia Chaleece Sandberg,
Karen Abbot, Swathi Kiran Aphasia Research
Laboratory, Department of Communication Sciences
and Disorders The University of Texas at
Austin This research was supported by NIDCD
DC006359 to the third author
Background
Methods
P4
Effect Size (CHURCH) Effect Size (CHURCH) Effect Size (CHURCH) Effect Size (CHURCH)
Abstract (untrained) Concrete (trained) Other Abstract Other Concrete
-0.20 5.34 -0.37 0.09
Task during treatment and probes Generative
naming of category items

• Behavioral data from both normal and
  brain-injured subjects suggests that abstract
  words and concrete words are processed
  differently.
• Theories that explain this difference in
  processing
• Dual Coding Theory (Paivio, 1991)
• Context Availability Theory (Schwanenflugel,
  Harnishfeger, Stowe, 1988)
• NICE (Normally Isolated Centrally Expressed
  semantics) Model (Newton Barry, 1997)
• Abstract words appear to be linked by
  associations, whereas concrete words are linked
  by category (Crutch Warrington, 2005).
• One study has examined the use of low or high
  imageable stimuli in the treatment of reading
  deficits in patients with deep dyslexia (Kim
  Beaudoin-Parsons, 2007).
• Evidence for the effect of concrete or abstract
  stimuli on lexical retrieval deficits subsequent
  to treatment for patients with aphasia is limited.

Sample training session
Most frequent other and incorrect responses (CHURCH) Most frequent other and incorrect responses (CHURCH) Most frequent other and incorrect responses (CHURCH)
other abstract other concrete incorrect
God (50.0) funerals (16.2) preaching (28.6)
Baptists (14.7) attendees (14.7) death services (28.6)
Catholics (14.7) marriage (13.2) spear (14.3)
Methodists (5.9) members (11.8) infant services (14.3)
ordained (2.9) churchgoers (5.9) church house (14.3)
Effect Size (HOSPITAL) Effect Size (HOSPITAL) Effect Size (HOSPITAL) Effect Size (HOSPITAL)
Abstract (trained) Concrete (untrained) Other Abstract Other Concrete
2.56 2.46 0.00 -1.47
Most frequent other and incorrect responses (HOSPITAL) Most frequent other and incorrect responses (HOSPITAL) Most frequent other and incorrect responses (HOSPITAL)
other abstract other concrete incorrect
death (35.3) nurse (23.4) city hospital (9.1)
cleanliness (11.8) visitors (14.0) older hospital (9.1)
healing (5.9) rooms (4.7) medical hospital (9.1)
wellness (5.9) surgeons (4.7) presbyterian hospital (9.1)
diagnosis (5.9) paramedic (2.8) federal hospital (9.1)
Analysis of responses e.g., CHURCH
Target Concrete N 10 (e.g., steeple, candle) Target Abstract N 10 (e.g., penance, belief) Normed and balanced for stimulus variables
Other concrete words Up to 15 words (e.g., choir, cross) Other abstract words Up to 15 words (e.g., god, ordained) Set of responses different for each patient - not normed
Pre/Post Testing
Present Study
Results

• Examine the effect of training generative naming
  of abstract versus concrete items within a
  category.
• Underlying basis Previous work on training
  complex atypical examples results in
  generalization to untrained typical examples but
  not vice versa (Kiran Thompson, 2003 Kiran
  Johnson, submitted Kiran, submitted Kiran et
  al., in preparation).

P1
Effect Size (CHURCH) Effect Size (CHURCH) Effect Size (CHURCH) Effect Size (CHURCH)
Abstract (trained) Concrete (untrained) Other Abstract Other Concrete
1.50 0.64 0.29 0.46
Most frequent other and incorrect responses (CHURCH) Most frequent other and incorrect responses (CHURCH) Most frequent other and incorrect responses (CHURCH)
other abstract other concrete incorrect
God (33.3) pew (44.4) forgotton (100)
religion (23.8) cross (22.2)  
Christ (14.3) preacher (11.1)  
heaven (9.5) choir (11.1)  
unfaithful (4.8) grail (11.1)  
Theoretical Assumption

Effect Size (HOSPITAL) Effect Size (HOSPITAL) Effect Size (HOSPITAL) Effect Size (HOSPITAL)
Abstract (untrained) Concrete (trained) Other Abstract Other Concrete
0.41 1.29 1.32 1.90
Most frequent other and incorrect responses (HOSPITAL) Most frequent other and incorrect responses (HOSPITAL) Most frequent other and incorrect responses (HOSPITAL)
other abstract other concrete incorrect
sick (33.3) nurse (60.0) hearing (100)
aid (16.7) technician (10.0)  
well (16.7) nurse's aid (10.0)  
helpful (8.3) sling (10.0)  
cure (8.3) stroke (10.0)  
P2
Hypothesis and Predictions
Effect Size (CHURCH) Effect Size (CHURCH) Effect Size (CHURCH) Effect Size (CHURCH)
Abstract (trained) Concrete (untrained) Other Abstract Other Concrete
3.46 1.59 2.02 0.55
Conclusion
Most frequent other and incorrect responses (CHURCH) Most frequent other and incorrect responses (CHURCH) Most frequent other and incorrect responses (CHURCH)
other abstract other concrete incorrect
communion (40.0) choir (16.4)
God (13.3) piano (16.4)
vows (13.3) pew(s) (14.5)
Presbyterian (6.7) stained glass (12.7)
love (6.7) cross (7.3)

• Results are consistent with our previous work
  examining selective generalization patterns
  (Kiran, 2007) using typicality of category
  exemplars.
• Generalization from abstract to concrete, but not
  vice versa.
• Which mechanisms underlying treatment induce
  generative naming recovery remains unclear.
• No improvement on other concrete or abstract
  words for all but one participant (based on
  effect size).
• However, from a clinical standpoint, the total
  number of items each patient produced increased
  as a function of treatment.
• The influence of working memory and other
  cognitive-linguistic factors on the task is not
  well understood.
• Interaction of variables such as AoA and
  frequency of co-occurrence (both normal and
  pathological) need to be examined.

P3
Effect Size (HOSPITAL) Effect Size (HOSPITAL) Effect Size (HOSPITAL) Effect Size (HOSPITAL)
Abstract (trained) Concrete (untrained) Other Abstract Other Concrete
16.88 4.02 -0.11 -0.33
Most frequent other and incorrect responses (HOSPITAL) Most frequent other and incorrect responses (HOSPITAL) Most frequent other and incorrect responses (HOSPITAL)
other abstract other concrete incorrect
sleeping (33.3) nurse (10.2) socks (16.7)
death (33.3) MRI (8.8) pictures (8.3)
reading (16.7) babies (8.0) white stuff (8.3)
emotional (16.7) bed (7.3) neo (8.3)
  stethoscope (5.8) french fries (8.3)
Participants
P1 P2 P3 P4
Age (yrs) 54 57 39 77
Gender male female female male
Handedness right right right right
Occupation retired auto body technician retired office manager software tester retired lawyer
Etiology left MCA left MCA left temporal intra-parenchymal hemorrhage left thalamic hemorrhage
MPO 19 43 8 32
Aphasia DX Anomic Aphasia Anomic Aphasia Anomic Aphasia Anomic Aphasia
References
Effect Size (COURTHOUSE) Effect Size (COURTHOUSE) Effect Size (COURTHOUSE) Effect Size (COURTHOUSE)
Abstract (untrained) Concrete (trained) Other Abstract Other Concrete
2.35 6.66 2.10 -1.07
Crutch, S. J. Warrington, E. K. (2005).
Abstract and concrete concepts have structurally
different representational frameworks. Brain,
128(3), 615- 627. Kim, M. Beaudoin-Parsons, D.
(2007). Training phonological reading in deep
alexia Does it improve reading words with low
imageability? Clinical Linguistics Phonetics,
21(5), 321-351. Kiran, S. Thompson, C. K.
(2003). The role of semantic complexity in the
treatment of naming deficits Training semantic
categories in fluent aphasia by controlling
exemplar typicality. Journal of Speech, Language,
and Hearing Research, 46, 773787. Kiran, S.
(2007). Complexity in the treatment of naming
deficits. American Journal of Speech Language
Pathology, 16(1), 18-29. Newton, P. K. Barry,
C. (1997). Concreteness effects in word
production but not word comprehension in deep
dyslexia. Cognitive Neuropsychology, 14,
481-509. Paivio, A. (1991). Dual coding theory
Retrospect and current status. Canadian Journal
of Psychology, 45(3), 255-287. Schwanenflugel, P.
J., Harnishfeger, K. K., Stowe, R. W. (1988).
Context availability and lexical decisions for
abstract and concrete words. Journal of Memory
and Language, 27(5), 499-520.
Most frequent other and incorrect responses (COURTHOUSE) Most frequent other and incorrect responses (COURTHOUSE) Most frequent other and incorrect responses (COURTHOUSE)
other abstract other concrete incorrect
law (23.3) judge (11.1) court (30.8)
appeal (13.3) witness (9.5) courthouse (15.4)
innocence (10) criminals (7.9) 12 people (7.7)
defense (10) police (7.9) hammer (7.7)
verdict (10) defendants (7.9) offendant (7.7)