Docstoc

Treatment Termination in Chronic Aphasia Deficit

Document Sample
Treatment Termination in Chronic Aphasia Deficit Powered By Docstoc
					Noun-verb dissociations-evidence from
acquisition and developmental and acquired
impairments

      Maria Kambanaros PhD CSP MSPA
      Speech Pathologist
      Assistant Professor of Speech-Language Therapy
      ATEI Patras
Key issue


   Grammatical class seems to be a universal
    linguistic feature, since the difference
    between nouns and verbs is present in all
    languages.
Introduction

   People with language impairments often
    show word class effects or differential
    impairments for nouns and verbs.

   In the case of children, verbs are typically the
    impaired class (Rice & Bode, 1984, Chiat,
    2001, Schultz et al., 2001, Black & Chiat,
    2002)
SLI children

   For example, children with SLI had reduced
    verb (but not noun) production and made
    heavy use of uninflected verbs.

   That were highly input dependent in their use
    of verbs, in that verbs, which featured in their
    speech had often been previously used by
    their parents in the same session.
   Conti-Ramsden & Jones, 1997
   Verb deficits are also common in acquired
    language disorders

   Several reports of 3 different patterns
   Verbs < Nouns
   Verbs > Nouns
   Verbs = Nouns
What’s in a word?

   Phonological information: the sound pattern
    of the word

   Semantic information:the meaning to which
    the sound pattern can refer

   Syntactic information: the position which the
    word may occupy in relation to other words
Mapping

To acquire words: Need to segment

   The stream of speech separating out the sound pattern of each
    word

   The scenes in which the speech occurs, separating out the
    aspects of the scene which are picked out by the words

   Map the one onto the other, making connections between the
    sound patterns they have segmented and the aspects of scenes
    they have segmented
Evidence from normal
acquisition
Work by Black and Chiat, 2002

   Verbs: world to word mappings-the child has
    to learn which features of an event map onto
    the specific meaning of the verb.

   Many features of verb meaning cannot be
    grasped at all from the perceived event, such
    as the difference between “buy” and “sell”.
Buy vs Sell

   The contrastive meanings cannot be derived simply
    by observing buying and selling events

   The structures in which they occur do offer clues
    about their particular focus
   Buy: describes the event from the point of view of
    the purchaser maps goal onto subject
   Sell: focuses on the seller, maps source onto
    subject
   Hearing many examples of these verbs in sentences
    can help the child derive this opaque information
Work by Black and Chiat, (2002) for
English
   Verbs have difficult phonological properties.

   In bisyllabic words, nouns normally stress the first syllable, while
    verbs often stress the second e.g. honey vs attract

   A listener assuming the typical pattern in which stressed
    syllables mark word onsets, may confront particular difficulties in
    segmenting verbs from the speech stream

   Verbs typically occupy sentence medial, rather than final or initial
    positions, which again makes them less salient than nouns.
“Motherese” or infant directed
speech(Aslin et al., 1996)
   English speaking mothers were asked to teach their 12 month children
    three verbs and three nouns.

   The nouns and verbs had very different distributions in the mothers’
    speech.

   Target nouns-overwhelmingly located at the end of utterances,
    suggesting that mothers knew intuitively that this position favoured
    acquisition.

   Target verbs-more variable locations, but most were used in sentence
    medial positions.

   It seemed that the mothers were bound by the constraints of English,
    even when they were consciously teaching vocabulary to their infants.
Conclude

   Verbs entail semantic and phonological
    complexities, which are not easily compensated for
    in motherese

   Verbs have an argument structure while nouns,
    especially those referring to concrete entities, often
    do not

   Given these contrasts, we would expect verbs to be
    acquired later than nouns in normal acquisition.

   There is substantial evidence that this is the case
Exceptions to the typical noun bias

   Children learning Korean (Choi & Gopnick, 1995)
    and Mandarin (Tardiff, 1996) have shown earlier
    verb acquisition than their English speaking peers

   Explanations: ?different patterns of parental input
    (Tardiff et al., 1997)

   Common view: the noun bias is moderated in
    languages which allows verbs to appear more in
    isolation or sentence final positions = greater
    phonological prominence
Evidence from German
(De Bleser & Kauschke, 2002)
   Q: Does the sentence position adopted by verbs
    influence order of access?
   German = a verb final language
   Children ranging from 2;6 to 4;11 years were tested
    in naming nouns and transitive and intransitive verbs
   Despite the characteristics of the language, the
    children showed a consistent noun advantage and
    across all age groups
   A: Not always the case
What affects acquisition?


   Nouns and Verbs have         Require different
    different semantic and        learning styles
    phonological properties
Evidence (Goldfield, 2000)

   Source: Parental input
   44 mother child interactions at 1;8y.o
   Parental input aiming to elicit production typically
    targeted nouns
   While verbs typically occurred in utterances aiming
    to elicit behaviours
   Children have to perform different functions with
    nouns and verbs
   Asked to produce nouns, but comprehend verbs
Implications

   1. studies focussing purely on production may
    underestmate children’s verbal skills in
    comprehension - children in the study demonstrated
    good verb understanding in carrying out mothers’
    requests

   2 to the encoding of lexical information: early
    learning of verbs is strongly associated with non-
    verbal action while nouns are associated with
    contexts in which an utterance accompanies
    attention to an object
   Learning of nouns and verbs may be
    qualitatively different
   Verb meanings are linked to
   Knowledge about action and function
   Noun meanings are linked to
   Sensory inputs e.g. the shape or appearance
    of the item
Verb learning
   Through the linguistic structures in which they appear

   Supported by richer linguistic input

   Child provided with many exemplars of each item

   May compensate for the opaque mappings between verbs and
    their real world referents in that the repeated linguistic context
    provides children with evidence about verb meaning

   Example: buy and sell
Lexical-semantic complexity

   The complexity of the semantic representations
    of verbs also influences retrieval accuracy
   Verb meanings =
     √ describe relational concepts
     √ structured entities that vary in type and number of
     constituents
    √Linked to thematic role assignment and argument
     structure
    √ Organised in matrices
    √ Multiple meanings (e.g. break a glass, break the news,
     break up a marriage)
Evidence from impaired
acquisition
Definition

   Children can be described as having
    language disorders if they have a significant
    deficit in learning to talk, understand, or use
    any aspect of language appropriately, relative
    to both environmental and norm-referenced
    expectations of similar developmental level
    (Paul, 2003)
Are Verbs more vulnerable in cases of
developmental language disorder?
   Because of semantic and phonological complexities
    which delay normal verb learning

   If language is one route to verb meaning, a vicious
    circle may be established: a child who cannot
    segment phonology or retain word order will be
    unable to identify verb arguments or their sentence
    order

   = lack the information needed to unravel the
    meaning of many verbs..without verbs..unable to
    build linguistic contexts
Evidence from SLI (English)

   Dockrell et al, (2001)

   31 children with word finding difficulties (WFD) in
    their naming of verbs and nouns and compared
    performance to chronological and language age
    matched controls

   WFD children did not show a particular
    disadvantage with verbs in naming accuracy or
    latency
Dockrell et al, (2001)

   WFD children were worse than chronological
    controls with nouns, but not with verbs
    (accuracy).

   Like the control groups children with WFD
    made a higher proportion of “other” errors
    with verbs e.g. Inappropriate verbs
Dockrell et al, (2001)

   Explanations:
   1. Group results obscure individual children who
    were showing such an effect
   2. WFD are not the group in which to expect verb
    deficits i.e. showed better comprehension than
    production suggesting that semantic knowledge was
    intact and this may minimise the likelihood of a verb
    deficit
   3. Picturable actions: being concrete and with
    observable world to word mapping are relatively
    easy for language impaired children
Evidence from normal acquisition and SLI
(Cypriot-Greek)
   Kambanaros & Grohmann (work in progress)

   30 6y.o monolingual Cypriot-Greek Grade 1
    children in naming action and object pictures

   2 chidren with SLI
   4;1 y.o boy = 33 action names incorrect (/42), 27
    object names incorrect (/42)

   9 y.o girl = 13 action names incorrect (/42) and 11
    object names incorrect (/42)
Evidence from SLI (Greek)
   Stavrakaki, S. (1996)

   Verb production

   4 SLI children: 2 boys and 2 girls MA 6;1 years, Mean RA 3;7 years,
    Mean EA 2;8 years

   12 controls: 6 boys and 6 girls MA 6;1 years

   SLI children have difficulties in verb production, whereas AM controls
    exhibit high level of correct performance An independent-sample T test
    carried out indicated that SLI children perform significantly below AM
    controls on the verb production task [t (3.057)=-8.083 p=.004]
Evidence from SLI (Greek)

   Kambanaros et al., (work in progress)
   8 children with SLI: 3 boys and 5 girls MA 5;6 years
    and LA 3;0yrs

   8 normal language developing children : 3 boys and
    5 girls who were matched on chronological age with
    the SLI group (MA 5;06).

   8 normal language developing children 3 boys and
    5 girls who were matched on language age with the
    SLI group (MA 3;06).
Object name
Action name
Γενικός Μέζος Όρος Λαθών

Παπακάηω αναθέπεηαι ο ζςνολικόρ μέζορ όπορ ηων λαθών ηόζο
ζηην καηηγοπία ηων οςζιαζηικών όζο και ζε αςηή ηων πημάηων



   ΠΙΝΑΚΑΣ 1:   Παροσσίαση τοσ Μέσοσ Όροσ των Λαθών στις σποκατηγορίες τοσ τεστ:
                              Οσσιαστικά και Ρήματα.


                            ΜΕΣΟΣ ΟΡΟΣ ΛΑΘΩΝ

 ΚΑΤΗΓΟΡΙΑ                          SLI           NL CA               NL LA

 Ουσιαστικά                            53%                30%                 49%

 Ρήματα                                54%                41%                 54%
Μέσος όρος λαθών ανά κατηγορία



       Σηο ζύνολο ηος ηεζη ηα παιδιά με SLI έσοςν ηον ίδιο μέζο όπο
       λαθών με ηα παιδιά θςζιολογικήρ γλωζζικήρ ανάπηςξηρ ίδιαρ
       γλωζζικήρ ηλικίαρ με ηα SLI.


                   ΜΕΣΟΣ ΟΡΟΣ ΛΑΘΩΝ ΣΤΟ ΣΥΝΟΛΟ ΤΟΥ ΤΕΣΤ

 60%        53%                                    52%
 50%

 40%                            35%

 30%                                                             ΢υνολικό τεστ

 20%

 10%

  0%
             SLI               NL CA              NL LA
Είδος λαθών
    Η πιο ζςσνή καηηγοπία λαθών ζε παιδιά με SLI είναι ηο καμία
    απάνηηζη με μέζο όπο 43% και ηςπική απόκλιζη 0,15

 Διάγραμμα: Σστνόηηηα λαθών ζε παιδιά με SLI


      50%                                           43%
      45%
      40%
      35%
      30%                23%
      25%                                                                                                        ΢ειρά1
      20%                               12%
      15%                                                      10%
      10%     5%                                                              6%
       5%                                                                                  0%       1%
       0%
            ΢Τ΢ΥΕΣΙ΢Η




                                                               ΜΗ ΢ΥΕΣΙΚΟ




                                                                            ΟΤ΢ΙΑ΢ΣΙΚΟΤ
                                        ΠΕΡΙΓΡΑΦΗ
                        ΢ΗΜΑ΢ΙΟΛΟΓΙΚΗ




                                                                                                    ΑΚΑΣΑΛΗΠΣΟ
                                                    ΑΠΑΝΣΗ΢Η




                                                                                          ΡΗΜΑΣΟ΢
                                                                                           ΥΡΗ΢Η
             ΟΠΣΙΚΗ




                                                      ΚΑΜΙΑ
                          ΢Τ΢ΥΕΣΙ΢Η




                                                                              ΥΡΗ΢Η
Τα παιδιά φυσιολογικής γλωσσικής ανάπτυξης που ταιριάζουν ως προς τη
χρονολογική ηλικία (CA) με τα παιδιά με SLI παρουσιάζουν κυρίως περιγραφή με
μέσο όρο 39% και τυπική απόκλιση 0,09.
          Διάγραμμα: Συχνότητα λαθών σε παιδιά NL CA
        45%                               39%
        40%
        35%
        30%
        25%                                           21%
                           18%                                                                                     ΢ειρά1
        20%                                                      14%
        15%    7%
        10%
         5%                                                                     0%           1%       0%
         0%
              ΢Τ΢ΥΕΣΙ΢Η




                                                                 ΜΗ ΢ΥΕΣΙΚΟ




                                                                              ΟΤ΢ΙΑ΢ΣΙΚΟΤ
                                          ΠΕΡΙΓΡΑΦΗ
                          ΢ΗΜΑ΢ΙΟΛΟΓΙΚΗ




                                                                                                      ΑΚΑΣΑΛΗΠΣΟ
                                                      ΑΠΑΝΣΗ΢Η




                                                                                            ΡΗΜΑΣΟ΢
                                                                                             ΥΡΗ΢Η
               ΟΠΣΙΚΗ




                                                        ΚΑΜΙΑ
                            ΢Τ΢ΥΕΣΙ΢Η




                                                                                ΥΡΗ΢Η
Η συχνότητα των λαθών των παιδιών φυσιολογικής γλωσσικής ανάπτυξης που
ταιριάζουν ως προς τη γλωσσική ηλικία με τα παιδιά με SLI (NL LA) και
παρατηρούμε ότι εμφανίζονται με την ίδια συχνότητα η μη σχετική απάντηση και
το σημασιολογικά συσχετιζόμενο λάθος

Συχνότητα λαθών σε παιδιά NL LA


 30%                27%                                   27%
 25%                               20%
 20%
 15%                                           13%                                                          ΢ειρά1
        10%
 10%
  5%                                                                     0%           1%       1%
  0%
       ΢Τ΢ΥΕΣΙ΢Η




                                                          ΜΗ ΢ΥΕΣΙΚΟ




                                                                       ΟΤ΢ΙΑ΢ΣΙΚΟΤ
                                   ΠΕΡΙΓΡΑΦΗ
                   ΢ΗΜΑ΢ΙΟΛΟΓΙΚΗ




                                                                                               ΑΚΑΣΑΛΗΠΣΟ
                                               ΑΠΑΝΣΗ΢Η




                                                                                     ΡΗΜΑΣΟ΢
                                                                                      ΥΡΗ΢Η
        ΟΠΣΙΚΗ




                                                 ΚΑΜΙΑ
                     ΢Τ΢ΥΕΣΙ΢Η




                                                                         ΥΡΗ΢Η
Evidence from acquired
language disorders
Acquired Language disorders as a result of

   Stroke-Aphasia Syndromes
   Right Hemisphere Brain Injury
   Traumatic Brain Injury
   Dementia of the Alzheimer Type
   Schizophrenia
Aphasia
   A disorder of language due to
    brain damage, most commonly
    stroke. It often affects
    language production and
    comprehension, compromising
    both spoken and written
    language. Problems can arise
    at linguistic levels such as
    semantics, syntax, morphology
    and phonology. Aphasia may
    range from mild to severe, but
    it almost universally affects the
    ability to find words, nouns and
    verbs in particular
What aphasia is

   Neurogenic
   Acquired
   Language Specific
   Not modality bound
   Not a speech problem
   Not a problem of sensation or intellect
   Affects psycholinguistic processes
   Treating patients with aphasia involves more
    than remediating the language disorder
Grammatical word class differences

                                     were the first to
    Goodglass, Klein, Carey, and Jones (1966)
    report differential noun and verb production in
    aphasia: individuals with fluent aphasia showed
    more difficulty naming objects (nouns) whereas
    individuals with non-fluent (Broca’s) aphasia more
    impaired in naming action pictures (verbs)

    The (double) dissociation between noun and verb
    processing in aphasic individuals has been studied
    in a number of different languages
Previous research:Double dissociation


   English: (Berndt et al., 1997; Breedin et al., 1998; Breen & Warrington,
    1994; Kim & Thompson, 2000; Manning & Warrington, 1996; McCarthy
    & Warrington, 1985; Marshall, Pring & Chiat, 1998; Shapiro et al., 2000;
    Zingeser & Berndt, 1990).
   Chinese (Bates, Chen, Tzeng, Li & Opie, 1991), Danish (Jensen, 2000),
   Dutch (Bastiaanse, 1991; Jonkers, 1998; Jonkers & Bastiaanse, 1996),
   Finnish (Laine, Kujala, Niemi, & Uusipaikka, 1992),
   German (De Bleser & Kauschke, 2002),
   Hungarian (Osman-Sagi, 1987),
   Italian (Daniele, Guistolisi, Silveri, Colosimo & Gainotti, 1994; Luzzati,
    Raggi, Zonco, Pistarini, Contardi et al., 2002; Miceli, Silveri, Villa &
    Caramazza, 1984; Miceli, Silveri, Noncentini & Caramazza, 1988;
    Silveri & Di Betta, 1997)
   Greek (Tsapkini, Jarema & Kehayia, 2002).
Previous research: nouns<verbs in fluent aphasia


   Individuals with posterior lesions of the left hemisphere, fluent
    aphasia, that of, impaired noun retrieval with relatively spared
    verb naming. Again this has been demonstrated in a number of
    languages:


    English (Berndt et al., 1997; Berndt, Haendiges & Wozniak, 1997;
    Breen & Warrington, 1994; Caramazza & Hillis, 1991; Shapiro, Shelton
    & Caramazza, 2000; Zingeser & Berndt, 1990),
    Chinese (Bates, Chen, Tzeng, Li & Opie, 1991; Chen & Bates, 1998),
   French (Bachoud-Levi & Dupoux, 2003),
   Italian (Daniele et al., 1994; Luzzatti et al., 2002; Miceli et al., 1984;
    Miozzo, et al., 1994) and
   Hungarian (Osman-Sagi, 1987).
Action and Object Naming in anomia

   Verbs more difficult         Nouns more difficult
   Berndt et al, 1997           Basso et al., 1990
   Kohn et al, 1989             Marshall et al., 1996
   Jonkers, 1998                Zingeser & Berndt, 1990
   Jonkers & Bastiaanse,        Miceli et al, 1984, 1988
    1996                         Luzzatti et al., 2002
   Osman-Sagi, 1987
   Williams & Canter, 1987
   Berndt et al., 2002
                                 Equal Difficulties
                                 Luzzatti et al., 2002
   Luzzatti et al., 2002
Previous research: nouns>verbs in fluent aphasia



   Findings suggest that selective verb impairments
    are not exclusively found in non-fluent, Broca’s
    patients, but can also be observed in fluent patients
    with Wernicke’s aphasia (Berndt, Mitchum, Haendiges &
    Sanderson, 1997) and anomia (Bastiaanse, 1991; Berndt et
    al., 1997; Breedin & Martin, 1996; Jonkers & Bastiaanse, 1998;
    Luzzatti et al., 2002).
Grammatical word class deficits
in fluent aphasia

   Impaired noun retrieval with fluent
    aphasia.
   Impaired verb retrieval with fluent
    aphasia.
   No differential impairment for verbs and
    nouns in aphasia.
Verb retrieval in fluent aphasia
   Edwards (2002) has claimed that access to grammar of
    the language and the implementation of grammatical
    rules and processes could be faulty in fluent aphasia or
    anomia leading to a deficit in processing/retrieving verbs.

   Contrastingly, other aphasic subjects with anomia have
    demonstrated impaired knowledge of verb meanings
    (semantics) rather than syntax (Breedin & Martin, 1994, 1996;
    Breedin, Saffran & Schwartz, 1998; Manning & Warrington, 1996).

   For example, some anomic subjects have shown
    difficulties processing core meanings of verbs despite
    preserved syntactic knowledge of how the verb is used
    in a sentence (Breedin & Martin, 1994).
Bilingual aphasia research 1980-2004

   Nouns                    Verbs and Nouns
   Kremin & De Agostini     Kremin & De Agostini

    (1995)                    (1995)
                             Sasanuma & Park
   Junque et al., (1989)
                              (1995)
   Stadie et al., (1995)   No study investigating
                              verb/ noun processing
                              in L1 and L2
Recent Bilingual aphasia research

   Hernandez, Costa, Sebastian-Galles,
    Juncadella, & Rene (2006)
   1 female anomic (BDAE-Cookie theft picture)
   Alzheimer’s disease
   Catalan (L1)Spanish (L2)
   Early bilingual
   Highly proficient
   Nouns < Verbs in picture naming
Recent Bilingual aphasia research

   Hernandez, Cano, Costa, Sebastian-Galles,
    Juncadella & Gascon-Bayarri, (2008)
   1 male non-fluent agrammatic (BDAE-Cookie
    theft picture)
   Primary Progressive Aphasia
   Spanish (L1) - Catalan (L2)
   Late bilingual
   Highly proficient
   Nouns > Verbs in picture naming: verbal and
    written
Recent Bilingual aphasia research

   Kambanaros & van Steenbrugge (2006)
   12 bilingual aphasic participants: 8 males and
    4 females
   anomic aphasia (Left CVAs)
   Greek (L1) - English (L2)
   Late bilinguals
   Highly proficient
   Results: ??
Research Question



   Do bilingual aphasic subjects have difficulties
    comprehending and or retrieving verbs and/
    nouns ?
Why is this study important ?

   Theory: one lesion-two languages
   Models of bilingual language organization can be
    developed further (Lalor & Kirsner, 2001; Roberts,
    1998)
   Previous research-descriptive case-studies with little
    hypothesis testing, Studies fail to include a non-
    aphasic bilingual control group
   Little research so far
Modern Greek vs English

Modern Greek             English

Highly inflected         Minimally inflected

Stems-representational   Stems-real words
   units                 e.g sweep (s, ing)
e.g. a) /skoup/          broom (s)
Inflected –a
Inflected-izo
Subjects

   Anomic aphasics (N=12)        BDAE profiles-
   One unilateral lesion in
    the left hemisphere            anomic in both
   Bilingual pre-morbidly-        languages
    Greek native language
   No reported hearing or        8 males, 4 females
    visual problems               Mean age: 60y.o
   Right- handed




                                      Controls (N=12)
Method

Research Materials
 The Greek Object and Action Test (GOAT)
  developed for this study.
 110 coloured photographs (10cms x 15cms)

   55 objects

   55 actions

   Verb comprehension subtest ( 55 items)

   Noun comprehension subtest (55 items)

 All Greek tests = English tests

 Language/Subtests were counterbalanced
Language Tasks


   Action/verb comprehension/naming

   Object/noun comprehension/naming

       Analysis
         Repeated measures analysis of variance
         Within subject factors
                 (L1 and L2 verbs/ nouns)
Question 1:
Verb/Noun comprehension


   Do bilingual anomic aphasic subjects
    have more difficulties comprehending
    verbs and /or nouns in a comprehension
    task ?
Comprehension: verbs versus nouns
    100

    95

    90

    85

    80

    75

    70
          COMPVALL     COMPNALL
Comprehension:
verbs and nouns in L1 and L2

100


 96


 92


 88


 84


 80
      COMPVAGR   COMPNAGR   COMPVENG   COMPNENG
Question 2:
Verb/Noun production


   Do bilingual anomic aphasic subjects
    have more difficulties producing verbs
    and /or nouns in a naming task ?
Production: verbs versus nouns

     80




     70




     60




     50
          TOTVEBAL     TOTNNALL
Verbs and nouns within languages
  70

  65

  60

  55

  50

  45

  40
       TOTVEGRE   TOTNNGRE   TOTVEENG   TOTNNENG
Verb Errors



                1.4   0.5
          5.5
                                 Verb Error
                                 semantic
                                 other
                                 c/s
                            23
                                 gram.class
   14.5                          phonemic
Noun Errors



             0.5   0.5
                                noun
      9.4
                         15.3   other
                                semantic
                                c/s
                                gram.clas
            11                  phonemic
                       Pictures

Level of breakdown
                         Visual Object.
                         recognition
                         system



                          Semantic        LEMMA
                          System

   LEXEME
             Phono
             Output
             Lexicon
    Phono
    Output
    Buffer
Results-summary


      Verbs were more difficult to retrieve
       compared to nouns.
Conclusions:

   Verbs were more difficult to retrieve than nouns
    between and within languages

    language specific
    Comprehension problems
   = retrieving the (morpho) phonological
    representation      (lexeme)
Semantic accounts of word
class effects
   A number of individuals exhibit verb difficulties which
    are attributed to a semantic problem either in the
    context of aphasia, dementia etc

   Verbs are suggested to be particularly vulnerable to
    semantic damage, e.g., because they are less
    imageable than nouns.

   Semantic disorders affecting nouns but sparing
    verbs has also been described.
   Such cases suggest that nouns and verbs
    are distinguished by different distributions of
    semantic features. For example, verbs may
    be defined predominantly by functional,
    thematic or action related features and nouns
    more by sensory/perceptual features.

   = word class effects can arise from
    impairments to these different domains of
    meaning
Conclusion

   Semantic Hypothesis provides a convincing
    explanation for some individuals.

   Whether it can explain all word class effects
    is much more disputed
Modality specific deficits and
non-semantic accounts
   A number of people have been described with impaired verb
    production but intact comprehension.

   Kim & Thompson (2000) tested 7 people with agrammatism on:

   Noun/verb naming
   Noun/verb comprehension
   Grammatical judgements (with anomalous sentences violating the
    argument properties of verbs)
   Noun/verb categorisations
Results (Kim and Thompson, 2000)

   Grammaticality           93.6%
    judgement
   Noun comprehension       99.8%
   Verb comprehension       97.5%
   Noun naming              92.6%
   Verb naming              71.2%
   Noun categorisation      99.4%
   Verb categorisation      66.1%
   Subjects had good input skills with verbs i.e. They
    could comprehend them and perform grammaticality
    judgements.

   In contrast poor production.

   There was a clear hierarchy of difficulty which was
    determined by the number of arguments required by
    the verb (i.e. One place verbs were produced more
    successfully than two place and two place more
    successfully than three).
Striking examples-problems confined to
one modality of output.
   Individuals that show focal impairments
   Caramazza & Hillis, (1991)
   2 individuals with excellent comprehension
    for verbs
   1 had difficulties with verbs in speech only
    and for the other only in writing.
Striking examples-problems confined to
one modality of output.
   Marshall et al., (1998)
   Patient EM- had agrammatic speech which lacked
    verbs
   Naming: significant noun advantage but only in
    speech
   Written naming: nouns=verbs
   Comprehension: good for verbs and nouns
   Could access verbs; thematic properties to
    distinguish reverse role verbs like buy and sell
Striking examples-problems confined to
one modality of output.
   Individuals who have word class effects in
    one direction in production, and another in
    comprehension
   Shapiro et al., (2000)
   Patient JR
   Named verbs significantly better than nouns
   In comprehension verbs were significantly
    more impaired than nouns
Explanations

   Reflect a lexical rather than semantic deficit

   Nature of the deficit is disputed-different
    theories of lexical access
2-stage theory of lexical access (Levelt and
colleagues)

   Stage 1
   Lemmas: amodal representation carrying
    grammatical information e.g. Relating to
    grammatical class, subcategorisation, gender

   Stage 2
   Lexeme: retrieving the
    phonological/orthographic form
Greek Verb lemma
   KEY
   Vt = Transitive Verb
   pres = Present Tense
   prog = Progressive
   MG = Modern Greek
   x,y = 2 argument position




      Vt              Lexical
                                to hammer
      x,y           Category
                                karfoni      pres




                                            prog
Serial model of word processing (Levelt et al., 1999).
   Kim & Thompson (2000) attribute their results
    to a breakdown at the level of the lemma.

   EM: Lexeme level of breakdown

   JR: input and output lexicons are marked for
    word class and that he has a deficit here for
    verbs.
                          Pictures




                     Visual Object.
                     recognition
                     system



                      Semantic         LEMMA
                      System
                                      agrammatism
LEXEME
           Phonol
Anomia     Output
           Lexicon

 Phonol.
 Output
 Buffer
The relationship between lexical
and grammatical knowledge
   Lemma model argues that lexical access
    entails access to lexical syntactic properties

   ? Relationship between lexical and syntactic
    deficits
   Deficits in verb access = profound
    consequences for sentence building i.e.
    Lemma level breakdown

   Marshall et al., (1998) suggest that even
    lexeme level verb deficits can impair
    sentence generation
   EM: had agrammatic speech coupled with
    impaired verb production. Problem specific to
    phonological retrieval in that comprehension and
    written production were intact. Despite this it
    seemed a major contributor to her sentence
    deficit.

   Sentence production improved when verbs were
    provided and lexical therapy with verbs
    substantially improved sentence production.
   The relationship between verb and syntactic
    impairments is far from clear cut.

   One problem is that not all individuals with
    verb deficits have syntactically impoverished
    speech.
   Jensen, (2000)

   IB: impaired in both producing and understanding verbs = lemma
    level deficit.

   BUT able to produce canonical sentence structures, either with a
    realised verb or with an empty verb slot

   Could also map agent and theme nouns appropriately onto these
    structures (even with reversible sentences)

   It seemed that he had a basic sentence schema, and its
    associated mapping routines, in the absence of verb specific
    information.
   Shapiro and Caramazza, (2002)

   HG: showed impaired verb access in naming and
    description

   Yet morpho-syntactic processes involved with verbs
    were intact i.e. He could inflect verbs to mark tense
    or number even pseudo verbs.

   She could also draw upon her morphological skills to
    generate novel verbs (like laddering) from
    instrument names.
   These results suggest that morpho-syntactic
    processes are functionally separable from
    lexical access.
   HG: dissociations with verbs in an odd one out comprehension
    task.

Did ok on:
 Manner information (e.g. To mutter, to whimper, to mumble)

 Non-relational semantic oppositions (e.g. To open, to close, to
   shut)

BUT not ok on:
 Relational oppositions (e.g. To eject-to exit-to expel) = could not
  access verbs’ thematic/argument structures
   It seems that morpho-syntactic information is
    not only seperable from word forms, but also
    from thematic/argument information.
   Shapiro & Caramazza suggest that
    thematic/argument information is stored at
    the semantic level, and separately from more
    perceptual features of verb meaning.

   An impairment in accessing such information
    gives rise to a word class effect, given that
    verbs are much more dependent than nouns
    on thematic properties
   Both HG and PB indicate that thematic information
    is dissociable from morpho-syntactic properties.

   Shapiro & Caramazza suggest the latter are stored
    in an independent syntactic network. Impairment to
    the network generates a morpho-syntactic deficit,
    which can be specific to nouns or verbs.

   In contrast, preservation of the network allows for
    morpho-syntactic skills, even in the face of impaired
    word retrieval.
Some concluding comments
 Why are verbs                                  Picture

more difficult ?

   Linguistically more complex            Visual Object
                                           recognition
                                           system




                                            Semantic       LEMMA
                                            System

           LEXEME
                            Phonological
                            Output
                            Lexicon


           Phonological

           Output Buffer
   VERBS
   Less imageable
   More complex world mappings
   More complex syntactic-morphological forms
   Less phonologically salient

   Unlikely to show an advantage in
    developmental disorders and are more
    vulnerable in aphasia
   Dissociations can be modality specific i.e.
    confined to output, or to speech or writing.

   Dissociations can also occur within the
    classes (noun-verbs), in that strengths or
    weaknesses or nothing at all
Conclusion

   According to Marshall (2002: 77),
   [t]hese different accounts need not be
    mutually exclusive. Rather, it seems that
    individuals show word class effects for
    different reasons. In other words, despite
    some common symptoms, people with word
    class effects are probably a very disparate
    group, a situation which is not at all unfamiliar
    in aphasia.
Thankyou for your attention
Questions??

				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:7
posted:12/30/2010
language:English
pages:104