<%BANNER%>

Selective Grammatical Impairments across Spoken and Written Modalities in Three Persons with Aphasia

Permanent Link: http://ufdc.ufl.edu/UFE0021984/00001

Material Information

Title: Selective Grammatical Impairments across Spoken and Written Modalities in Three Persons with Aphasia
Physical Description: 1 online resource (78 p.)
Language: english
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: aphasia, impairments, modality, noun, oral, verb, written
Communication Sciences and Disorders -- Dissertations, Academic -- UF
Genre: Communication Sciences and Disorders thesis, M.A.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Grammatical deficits occur in persons with aphasia when the ability to produce one grammatical category, typically nouns or verbs, is more impaired than another category. These deficits are most often examined in oral tasks, such as naming and producing sentences. In the current study, two verbal (i.e., oral naming and oral reading) and two written (i.e., written naming and writing to dictation) tasks were examined to determine if any relationship across or between modalities existed in three persons with nonfluent aphasia. Specifically, the following three questions were asked: (a) Are there significant differences across nouns and verbs in oral naming, oral reading, written naming, and writing to dictation? (b) Does a double-dissociation exist across any of these modalities (i.e., nouns are more impaired in one modality and verbs are more impaired in the other modality)? (c) Based on an error analysis of incorrect responses, are the levels of impairment different for nouns and verbs across modalities? No significant statistical difference across nouns and verbs in any modality was noted. Similarly, no double dissociation across modalities was found in any of the three participants. Trends regarding accuracy data, however, were noted in all three participants. In addition, error analyses served as an appropriate medium to identify linguistic deficits present for nouns and verbs across modalities in each participant.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Thesis: Thesis (M.A.)--University of Florida, 2008.
Local: Adviser: Edmonds, Lisa Anna.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2008-11-30

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0021984:00001

Permanent Link: http://ufdc.ufl.edu/UFE0021984/00001

Material Information

Title: Selective Grammatical Impairments across Spoken and Written Modalities in Three Persons with Aphasia
Physical Description: 1 online resource (78 p.)
Language: english
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: aphasia, impairments, modality, noun, oral, verb, written
Communication Sciences and Disorders -- Dissertations, Academic -- UF
Genre: Communication Sciences and Disorders thesis, M.A.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Grammatical deficits occur in persons with aphasia when the ability to produce one grammatical category, typically nouns or verbs, is more impaired than another category. These deficits are most often examined in oral tasks, such as naming and producing sentences. In the current study, two verbal (i.e., oral naming and oral reading) and two written (i.e., written naming and writing to dictation) tasks were examined to determine if any relationship across or between modalities existed in three persons with nonfluent aphasia. Specifically, the following three questions were asked: (a) Are there significant differences across nouns and verbs in oral naming, oral reading, written naming, and writing to dictation? (b) Does a double-dissociation exist across any of these modalities (i.e., nouns are more impaired in one modality and verbs are more impaired in the other modality)? (c) Based on an error analysis of incorrect responses, are the levels of impairment different for nouns and verbs across modalities? No significant statistical difference across nouns and verbs in any modality was noted. Similarly, no double dissociation across modalities was found in any of the three participants. Trends regarding accuracy data, however, were noted in all three participants. In addition, error analyses served as an appropriate medium to identify linguistic deficits present for nouns and verbs across modalities in each participant.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Thesis: Thesis (M.A.)--University of Florida, 2008.
Local: Adviser: Edmonds, Lisa Anna.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2008-11-30

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0021984:00001


This item has the following downloads:


Full Text

PAGE 1

SELECTIVE GRAMMATICAL IMPAIR MENTS ACROSS SPOKE N AND WRITTEN MODALITIES IN THREE PERSONS WITH APHASIA By MEGAN ELIZABETH FERGUSON A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF ARTS UNIVERSITY OF FLORIDA 2008 1

PAGE 2

2008 Megan Elizabeth Ferguson 2

PAGE 3

To my parents, who enabled me to have the best education and offered unceasing encouragement and prayers throughout the experience 3

PAGE 4

ACKNOWLEDGMENTS I would like to thank Dr. Lisa Edmonds for her stimulating lectures which sparked my interest in completing a masters thesis, as well as, her guidance and patience throughout the research process. I would like to thank Dr. Diane Kendall for he r assistance and support through the writing process as well. 4

PAGE 5

TABLE OF CONTENTS page ACKNOWLEDGMENTS ...............................................................................................................4 LIST OF TABLES ...........................................................................................................................7 LIST OF FIGURES .........................................................................................................................8 ABSTRACT .....................................................................................................................................9 CHAPTER 1 INTRODUCTION................................................................................................................. .11 Brief Overview of Aphasia .....................................................................................................11 Modality Specific Deficits ......................................................................................................12 Noun and Verb Deficits..........................................................................................................12 Double Dissociations across Modality and Grammatical Class .............................................13 Language Models ....................................................................................................................14 Levelt Model ...................................................................................................................15 Rapp and Caramazza Model (2002) ................................................................................16 Ellis and Young Model (1988) ........................................................................................17 Noun and Verb Cortical Locations .........................................................................................19 Evidence for Double Dissociations ........................................................................................21 Questioning the Existence of Double Dissociations ...............................................................22 Research Questions .................................................................................................................22 2 METHODS...................................................................................................................... .......26 Participants .............................................................................................................................26 Materials .................................................................................................................................27 Procedures ...............................................................................................................................28 Accuracy Scoring ....................................................................................................................29 Error Analysis .........................................................................................................................29 Reliability ...............................................................................................................................30 3 RESULTS...................................................................................................................... .........35 Accuracy .................................................................................................................................35 Participant 1 (P1) Accuracy .............................................................................................35 Oral naming ..............................................................................................................35 Oral reading ..............................................................................................................35 Written naming .........................................................................................................35 Writing to dictation ..................................................................................................35 Participant 2 (P2) Accuracy .............................................................................................36 Oral naming ..............................................................................................................36 5

PAGE 6

Oral reading ..............................................................................................................36 Written naming .........................................................................................................36 Writing to dictation ..................................................................................................36 Participant 3 (P3) Accuracy .............................................................................................36 Oral naming ..............................................................................................................36 Oral reading ..............................................................................................................36 Written naming .........................................................................................................37 Writing to dictation ..................................................................................................37 Error Analysis .........................................................................................................................37 Participant 1 Error Analysis ............................................................................................37 Oral naming ..............................................................................................................37 Oral reading ..............................................................................................................38 Written naming .........................................................................................................38 Writing to dictation ..................................................................................................38 Participant 2 Error Analysis ............................................................................................39 Oral naming ..............................................................................................................39 Oral reading ..............................................................................................................39 Written naming .........................................................................................................40 Writing to dictation ..................................................................................................40 Participant 3 Error Analysis ............................................................................................40 Oral naming ..............................................................................................................40 Oral reading ..............................................................................................................41 Written naming .........................................................................................................41 Writing to dictation ..................................................................................................42 4 DISCUSSION................................................................................................................... ......52 Participant 1 ............................................................................................................................53 Participant 2 ............................................................................................................................58 Participant 3 ............................................................................................................................63 5 CONCLUSION................................................................................................................... ....72 LIST OF REFERENCES ...............................................................................................................75 BIOGRAPHICAL SKETCH .........................................................................................................78 6

PAGE 7

LIST OF TABLES Table page 2-1 Descriptive Information Regarding Participants ..............................................................31 2-2 Accuracy Scores for all Participants .................................................................................32 2-3 Lexical Test Scores for all Participants ............................................................................33 2-4 Error Analysis Categories (Targeted Word: Candle) ........................................................34 3-1 P1 Accuracy Data .............................................................................................................43 3-2 P2 Accuracy Data .............................................................................................................43 3-3 P3 Accuracy Data .............................................................................................................43 7

PAGE 8

LIST OF FIGURES Figure page 1-1 Rapp and Caramazza model (2002). .................................................................................24 1-2 Ellis and Young model ......................................................................................................25 3-1 Visual Representation of P1s Accuracy Data. .................................................................44 3-2 Visual Representation of P2s Accuracy Data. .................................................................44 3-3 Visual Representation of P3s Accuracy Data. .................................................................45 3-4 P1s Error Analyses Graphs ..............................................................................................46 3-5 P2s Error Analyses Graphs ..............................................................................................48 3-6 P3s Error Analyses Graphs ..............................................................................................50 4-1 Lexical Deficits Present in Participant 1 Based on Error Analysis ..................................69 4-2 Lexical Deficits Present in Participant 2 Based on Error Analysis ..................................70 4-3 Lexical Deficits Present in Participant 3 Based on Error Analysis ..................................71 8

PAGE 9

Abstract of Thesis Presen ted to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Arts SELECTIVE GRAMMATICAL IMPAIRMENTS ACROSS SPOKE N AND WRITTEN MODALITI ES IN THREE PERSONS WITH APHASIA By Megan Elizabeth Ferguson May 2008 Chair: Lisa Edmonds Major: Communication Sciences and Disorders Grammatical deficits occur in persons with aphasia when the ability to produce one grammatical category, typically noun s or verbs, is more impaired than another category. These deficits are most often examined in oral tasks, such as naming and producing sentences. In the current study, two verbal (i.e., oral naming and oral reading) and two written (i.e., written naming and writing to dictation) ta sks were examined to determine if any relationship across or between modalities existed in th ree persons with nonfluent aphasia. Specifically, the following questions were asked: (a) Are there significan t differences across nouns and verbs in oral naming, oral reading, written naming, and writing to dictation? (b) Does a double-dissociation exist across any of these modaliti es (i.e., nouns are more impaired in one modality and verbs are more impaired in the other modality)? (c) Based on an error analysis of incorrect responses, are the levels of impairment different for nouns and verbs across modaliti es? No significant statistical difference across nouns and verbs in any modality was noted. Similarly, no double dissociation across modalities was found in any of the three participants. Trends regarding accuracy data, however, were noted in all three participants. In a ddition, error analyses served as 9

PAGE 10

an appropriate medium to identify linguistic deficits present fo r nouns and verbs across modalities in each participant. 10

PAGE 11

CHAPTER 1 INTRODUCTION Brief Overview of Aphasia Aphasia is an acquired neurogenic language impa irment that results from damage to the brain most often caused by cerebrovascular accidents (CVAs). There are seven aphasia syndromes that can be subdivided into fluent and nonfluent aphasi as. Nonfluent aphasias include Brocas, transcoritcal motor, and global a phasia. Fluent aphasias include Wernickes, conduction, transcortical sensor y, and anomic aphasia. The patterns of language impairments can vary ac ross and within individuals with aphasia. Deficit patterns in spoken language production and comprehension, as well as, reading and writing can be differentially impaired within th e aphasia types (Chapey, 1994). Impaired naming abilities tend to be the residual deficit following recovery for a ny type of aphasia (Hillis, 2007). Thus, the remaining impairment would then imply that naming deficits can result from damage to various areas of the brain. All aphasia types are characterized by word retrieval difficulties, but differences in production, comprehension, and repetition differentiate the subtypes further. Brocas aphasia, the most common type of nonflu ent aphasia, results from frontal cortical lesions and is characterized by effortful speech wi th impaired grammar (e.g., reduction in use of function words and morphology, prosody), reduced rate of speech and repetition abilities as well as poor spelling and reading; comp rehension is relatively intact. Wernickes Aphasia, the most common type of fluent aphasia, occurs from posteri or cortical lesions whic h result in relatively preserved fluency, poor comprehension, and poor repetition, spelling, and writing abilities. Conduction Aphasia is characterized by anterior-par ietal cortical lesions resulting in fair-to-good fluency, intact comprehension for words and si mple sentences, and poor repetition, spelling, and reading ability. Anomic aphasia is primarily ch aracterized by impaired naming with preserved 11

PAGE 12

fluency, comprehension, repetiti on ability, and variably impair ed spelling and reading. The primary deficit, however, is impaired naming. The location of the lesion in this aphasia type is variable across individuals (Chapey, 1994; Hillis, 2007). Modality Specific Deficits Individuals with aphasia can exhibit patte rns of deficits across modalities (Arevalo, Moineau, Saygin, Ludy, & Bates, 2005; Ber ndt, Mitchum, Haendiges, and Sandson, 1997a, 1997b; Druks, 2002). Traditionally, individuals with Anomic Aphasia are noted to have fluent speech, but with reduced oral naming abilities. Si milarly, individuals with Wernickes Aphasia are noted to have relatively fl uent speech, but with noted reduced abilities with oral naming and comprehension. In Brocas aphasi a, individuals are noted to have nonfluent speech with reduced oral naming abilities and preserved comprehens ion (Chapey, 1994; Druks; Hillis, 2007; Zingeser & Berndt, 1990). Some researchers have found oral naming abilities to be reduced in individuals with fluent aphasias when compared to spoken or written word/picture verification (Hillis & Caramazza, 1991). Arevalo et al. (2005) found th at picture-naming was th e most difficult tasks for individuals with fluent and nonfluent aphasia when compared to word reading and repetition. In a later study with a group of diversely diagnosed persons w ith aphasia (i.e., Brocas, Anomic, and Wernickes Aphasia), Arevalo et al. (2007) again found the par ticipants to be significantly less accurate with picture naming. Noun and Verb Deficits Aware of modality impairment s associated within the aphasias, researchers have more closely examined grammatical deficits made by individuals with aphasia. Traditionally, individuals with Brocas Apha sia were thought to demonstrate more verb deficits, while individuals with Wernickes and Anomic Aphasia tended to demonstrate more noun deficits. Evidence for this assertion, however, has not been consistent across studies ; thus revealing that 12

PAGE 13

specific noun and verb deficits are not solely conf ined to a particular aphasia (Arevalo et al., 2007; Arevalo et al., 2005; Berndt et al., 1997a, 1997b; Druks, 2002; Hillis & Caramazza, 1995; Hillis, Tuffiash, & Caramazza, 2002). Arevalo et al. (2005) found that when cons idering the three word -processing tasks of picture-naming, word reading, and repetition, 21 individuals with aphasia (i.e., Anomic, Brocas, and Wernickes) demonstrated a noun advantage in picture-naming alone. In an individual with primary progressive aphasia, a clearly reduced ability to orally naming verbs was documented, while written naming of verbs and comprehension of verbs remained relatively intact (Hillis et al., 2002). In a single case study, an individual with aphasia dem onstrated a greater difficulty with verbs in reading tasks than with nouns (H illis and Caramazza, 1995). Arevalo et al. (2007) found that in a group of 21 individu als with either Anomic, Brocas, or Wernickes Aphasia, all were significantly less accurate with verbs in picture naming when compared to single word reading and word repetition. Conversely, Berndt et al. (1997a ) found that in a group of 11 persons with aphasia (i.e., Anomic, Brocas, Wern ickes, and transcortical sensory aphasia), five individuals had more trouble w ith nouns, two had more trouble with verbs, and four had no significant difference between nouns and verbs in picture naming ability. The dynamic noun and verb grammatical deficits within modalities noted between the documen ted individuals with aphasia has prompted researchers to more clos ely examine relationships regarding theoretical models of language processing, cortical representation, and noun and verb retrieval. Double Dissociations across Modality and Grammatical Class Double dissociations across modalities and gramma tical deficit can also occur, so that a grammatical class deficit (e.g., noun deficit) can exist within one modality (e.g., oral naming), while a different grammatical class deficit (e.g., ve rbs deficit) can occur in a different modality 13

PAGE 14

(e.g., written naming) (Hillis & Caramazza, 1995; Hillis et al., 2002; Rapp & Caramazza, 2002; Shapiro & Caramazza, 2002). After examining the evidence regarding selective impairments, many researchers have begun to view double-dissociations in individuals with aphasia as a series of interconnected processing subsystems that have been damage d, instead of a single underlying impairment (Berndt et al., 1997a, 1997b; Caramazza & Mice li, 1991; Caramazza & Rapp, 2002; Kay, Lesser, & Coltheart, 1996). Hillis and Caramazza (1995) work ed with an individual with severe anomic aphasia who exhibited increased difficulties prod ucing nouns in oral tasks, and showed greater impairment in verbs when recognizing written forms. Rapp and Caramazza (2002) reported findings from an individual with aphasia who de monstrated increased di fficulties with producing nouns in spoken activities and increased difficulties with verbs in writte n activities (originally reported between May 1994April 1995). The author s posited that the indivi duals deficits arose from problems retrieving/maintaining stable word-forms (i.e., noun word-forms in spoken tasks and verb word-forms in written tasks). The st udy examined the possible effects of abstractness of verbs and selective morphological difficulties within various individuals with aphasia and determined that the deficits were not due to object or action distinc tion, but to grammatical properties associated with the stimuli. The author s concluded that these findings, along with data collected from the studied individual with aphasia, support the ex istence of grammatical category distinctions at post-semantic levels among variou s pathways of each feature set. For further information, see the section title d Rapp and Caramazza model. Language Models Despite a number of cases reporting double-dissociations across modality and grammatical class (Hillis & Caramazza, 1995; Hillis et al.,, 2002; Rapp & Caramazza, 2002; Shapiro & Caramazza, 2002), theories regarding the loci of impairment within the language system remain 14

PAGE 15

controversial. Nonetheless, ev idence regarding double dissociations has been used to further establish the existence of multi-system langua ge models that allow explanation for the occurrence of selective impairments (Berndt et al., 1997a, 1997b; Caramazza & Miceli, 1991; Rapp & Caramazza, 2002; Kay et al., 1996). This section examines some traditional models of lexical access and how they are able or unable to account for the double dissociation findings across modality and grammatical class. Levelt Model Dissociations are generally thought to occur within th e syntactic or semantic areas of the language system. Levelts Tier Speech Production Model (1989) is commonly used to explain the syntactic or semantic basis for the dissociation. The model consists of three levels: the conceptualizer (determines what sounds will be used), formulator (converts conceptual sound structure into a linguistic st ructure), and articulator (phys ical components used for sound production) (Levelt, 1989). In an updated mode l, lexical representati on is comprised of two levels between the formulator and articulator. Th e first level, the lemma, is an abstract lexical representation that matches semantic informati on to the conceptualized information and also plays a role in defining sentence structure. The second level, the lexeme, stores word forms that are specified for oral (i.e., phonologic) and written (i.e., orthographic) output, which are activated and then theoretically inserted into the previously formulated sentence structure (Levelt, 1992). Analyzing the collected data relating to language processing, researchers have made assertions regarding the origin of double dissociations within the Levelt model (1992). A number of researchers have concluded that items are not grouped according to grammatical class, but through semantic/physical feat ures, leaving the dissociation to occur within that system (Berndt et al., 1997a, 1997b). For example, an indi vidual with aphasia may see a picture of tiger, 15

PAGE 16

which would in turn activate other related animal s, such as a lion or cheetah. Berndt et al. (1997a, 1997b) argued that if errors occurred at the lemma level (i.e., semantic level), serious destructive effects would be seen in sentence pr oduction, which is formulated at that level. These researchers argue that if the level of impair ment was at the lower lexical level, the lexeme level (i.e., phonological and/or orthographic output level), senten ce production would subsequently be notably less affected. Convers ely, Caramazza & Hillis ( 1991) argue that their data reveal impairment at the lower lexeme level. They assert that if an individual with aphasia has a noun or verb deficit in one output modality and not in another (e.g., noun impairment with oral naming and not written naming), then the impair ment must exist at the lexeme level. Druks (2002) added that although the evidence for impairment at this level is not equally strong, the majority of investigators infer that the impairments do occur at the lexeme level. Rapp and Caramazza Model (2002) Rapp and Caramazza (2002) created a newer schematic depiction of language processing that would account for the inte rpretation of modality specific grammatical dissociations as modality-specific dissociations between words referring to objects and actions (p. 398). The model is based on the hypothesis that grammati cal category deficits occur because defining features of nouns correspond to perceptual features, while defi ning features of verbs correspond to motor actions. It is assumed that these info rmation categories are lo cated in separate brain areas that communicate with the output lexicons using various pathways. The researchers state that modality and grammatical errors occur from selective damage to perceptual feature and motor feature pathways that correspond to gr ammatical categories, si nce action information plays a significant role in the de finition of many verbs and perceptual features compose a great majority of concrete nouns. Separate pathways exist for each of these feature sets thereby 16

PAGE 17

providing an explanation for noun a nd verb differences in terms of the semantic differences between perceptual objects and actions (Rapp & Caramazza) (Figure 1-1). Ellis and Young Model (1988) The model that will be used to guide the research questions and interp retation of results in this study is the Kay et al. (1996) adapted El lis and Young model (1988). This model depicts language processing as a multi-modal input and output system and consists of three input modalities (i.e., speech, pictures & seen objects and print) and two output modalities (i.e., speech and print) (Figure 1-2). Auditory input is initially processed via auditory phonological anal ysis (i.e., speech) and orthographic information (i.e., print) is processed via the abstract letter identification modality. The stimuli are then processed to the phonologica l input lexicon and orthographic input lexicon, respectively, where a list of appropriately matched phonological and orthographic words is activated. For visual material, such as pictures and objects, the visual object recognition system is activated. Stimuli from a ll three input modalities are then processed through the semantic system where conceptual meaning is retrieved an d assigned to the stimuli. Theoretically, this level is analogous to Levelts lemma level (1992). Next, the identi fied stimuli are transferred to either the phonological output lexicon (i.e., sp eech output) or the orthographic output lexicon (i.e., writing/print) where the appropriate word -form is retrieved. This level corresponds to Levelts lexeme level (1992). Finally, the newly assigned word is temporarily stored in a buffer until it is produced. Using this model, an individual can repeat auditory stimuli (i.e., speech), copy orthographic stimuli (i.e., print) read aloud orthographi c stimuli (e.g., oral reading), and write to auditory stimuli (e.g., writing to dictation) with out utilization of the semantic system. For example, if an individual is to repeat auditory stimuli, the semantic system can be bypassed 17

PAGE 18

completely using the route of auditory phonological analysis phonological input buffer phonological input lexicon phonological output lexicon phonological output buffer. Similarly, an individual can bypass the seman tic system when copying orthographic stimuli using the route of abstract letter identification orthographic input lexicon phonological output lexicon orthographic output lexicon orthographic output buffe r. An individual can bypass the semantic system as well when read ing orthographic stimuli aloud. The route of abstract letter identification orthographic input lexicon phonological output lexicon phonological output buffer would be u tilized in this situation. Fi nally, for writing to auditory stimuli, the route of auditory phonological analysis phonological input buffer phonological input lexicon phonological output lexicon orthographic output lexicon orthographic output buffer would be used. These routes woul d allow an individual w ith aphasia to process and produce words that are regul ar (e.g., dog) and irregular (e .g., leopard) without having to comprehend the meaning of the word (i.e., acces s the semantic system) (adapted Ellis & Young model, 1988; as cited in Kay et al., 1996). The adapted Ellis and Young model (1998; as cited in Kay et al., 1996) also accounts for non-lexical retrieval via acousti c-to-phonological conversion (i.e., auditory stimuli converted into oral speech) and letter-to-s ound rules (i.e., print/r eading stimuli converted into oral speech) pathways. Nonsense words or non-words are pr ocessed through these two pathways; stimuli are not assigned semantic meaning or retrieved from the lexicon be cause they do not exist in the native language. For example, if an individual read th e word tanipe aloud, it would theoretically be processed usi ng the letter-to-sound rules path way (i.e., sounded out) because there is not a lexical item corre sponding to this lette r string in the orthog raphic input lexicon. Another non-lexical route that may be used is the sound-to-let ter rules pathway. This route 18

PAGE 19

enables individuals to transfer non-lexically retrieved words fr om the phonological output buffer (i.e., temporarily stored speech forms of words that have been presented auditorily) to the orthographic output buffer (i.e., tem porarily stored print/ written forms of words) to then produce the word orthographically. This route allows fo r people to write nonwords from dictation. The final non-lexical route is the acoustic-to-phonological c onversion. With spoken input, the stimuli would travel around the semantic and phonological systems to the phonological output buffer. This route enables individuals to repeat nonwords. It is spec ulated that individuals with language processing impairment, such as indivi duals with aphasia, of ten use the non-lexical routes as well as the lexical routes that bypass the semantic system (Kay et al.). The adapted Ellis and Young model (1988; as ci ted in Kay et al., 1996) can be used as a guide to explain single modality deficits in individua ls with aphasia. To that end, Kay et al. have created a psycholinguis tic test battery, The Psycholinguistic Assessmen t of Language Processing in Aphasia (PALPA) to assist clinicians and research ers in localizing deficits in persons with aphasia (1992). Noun and Verb Cortical Locations One theory argues that noun and verb dissociat ion arises along lexical lines, which are processed in separate cortical areas (Caram azza & Hillis, 1991). Hillis & Caramazza (1995) concluded that an implication for this diffe rence involved lexical phonological knowledge for nouns and verbs being independently represented in the brain. The researchers furthered the theory of separate noun and verb locations in various modalities after analyzing an individual with aphasia who was severely impaired in r ecognizing and comprehending written verbs, even though her performance on recognizing and comp rehending spoken verbs and spoken/written nouns was considered within normal limits. Th e participants performance on concrete and abstract nouns was identical; therefore, this situation would further imply that lexical 19

PAGE 20

phonological knowledge and lexical orthograp hic knowledge for nouns and verbs are independently represented in the brain (Hillis & Caramazza, 1995). Recent studies have begun to use imaging techniques (e.g., fMRI and PET) to determine distinct brain regions designated for speech across modality tasks, as well as, noun and verb processing (Paulesu et al., 1997; Perani et al., 1999). No study, however, has offered sufficient evidence to allow researchers to confidently conc lude that nouns and verb s activate separate and discrete cortical areas. Inst ead, the majority of these studie s have shown areas that were activated more by one category than by another or that no areas were activated more by nouns than verbs (Arevalo et al., 2005; Paules u et al., 1997; Pera ni et al., 1999). Other recent studies have assert ed that noun and verb retrieva l differences are not due to separate anatomical locations, but to factors underlying each concep t (Arevalo et al., 2005). In a review of the literatu re, Druks (2002) compiled a list of f actors that objects and actions should possess in order to evaluate word retrieval effect ively. These factors included a high degree of name agreement and matching of frequency, ageof-acquisition, and familiarity (Druks). Arevalo et al. (2007) investigat ed the effect of manipulability on activating and processing nouns and verbs (i.e., objects that can and cannot be ma nipulated and actions th at do and do not involve fine hand movements). The researchers found an ove rall lower verb accuracy across individuals with aphasia and with control participants; howev er, individuals with ap hasia (i.e., Brocas and Wernickes Aphasia) were sign ificantly more accurate with pr ocessing non-manipulable items, while control participants were significantly mo re accurate with processing manipulable items. Further, when analyzed as a group (i.e., Brocas, Wernickes, and Anomic Aphasia), individuals with aphasia showed a significant affect of mani pulability with verbs (ac tions) rather than nouns (objects). This effect of manipulability was seen across all three aphasia types. The researchers 20

PAGE 21

hypothesized this result was due to motor cortex damage in the area were hand movement is processed/analyzed (A revalo et al., 2007). Berndt et al. (1997b) co ncluded that either grammatical cl ass, or imageability, or both features play a role in performance, but the effe cts of the factor or class are independent from one another. Results in this area of resear ch have provided further insight regarding the foundation of nouns and verbs; however, further research is required before a conclusion regarding the origin can be established. Jonkers and Basti aanse (1998) furthered this point by asserting that if these extraneous factors are appropriate ly controlled for, a dissociation between nouns and verbs could prove to be just one-directional (i.e., all pe rsons with aphasia would show greater difficultly with verb retrieval compared to noun retrieval or vice versa). Evidence for Double Dissociations Certain researchers have documented the existence of double dissociations between modalities and grammatical classes in persons with aphasia (Hillis & Caramazza, 1991; Hillis & Caramazza, 1995; Rapp & Caramazza, 2002). Hillis and Caramazza (1991) asserted that the presence of a double-dissociation between two indi viduals using the same stimuli implies that noun and verb differences can be attributed to degree of difficulty between th e two word classes. This was evident when looking at two indivi duals with aphasia with flawless auditory comprehension of both nouns and verbs. One pr esented difficulties w ith oral production of nouns, while the other demonstr ated difficulties with oral production of verbs (Hillis & Caramazza, 1991). In a review of the literature, it was asserted that these noun/verb differences occur at the lexical level. The argument was base d on the fact that the a ffected participants had intact comprehension for the impaired category, wh ich would further suppor t the existence of the double dissociation (e.g., impaired speaking or writing with intact comprehension) (Druks, 2002). Berndt et al. (1997a), however, made the relevant point that res earchers should be aware 21

PAGE 22

that certain comprehension test are not sensitiv e enough to pick up mild semantic deficits, which could affect whether a doubledissociation truly existed. Questioning the Existence of Double Dissociations Although several studies have documented doubl e-dissociations, its existence remains questionable due to the inability to reliably duplic ate the findings (Arevalo et al., 2005; Odell, Hashi, Miller, & McNeil, 1995). Arevalo et al. (2005) mentioned that double dissociation studies often suffer from a small sample size (i .e., single case studies) and are not sufficiently representative of the population. In addition to the limited sample, many of the studies have been designed in such a way that correct/good and incorrect/poor performance was not clearly and statistically defined. Thes e factors contribute to the abi lity to properly duplicate and/or document double-dissociations in various indivi duals, and thus weaken the validity of the findings (Arevalo et al., 2005). Similarly, Odell et al. (1995) found methodological weaknesses in certain studies which asserted the presence of a selective impairment in persons with aphasia. Specifically, test-retest reliability, task difficulty, and task sensitivity were found to be deficient in certain studies. The researchers stated that due to these inequities, th e studies failed to support the existence of a true selective impairment in the individuals. Odell et al. concluded that to meaningfully interpret double-dissociations within an individual, res earchers should first fu lfill the six criterion established by their study. Additional information regarding the criteria can be found in Odell et al. (2005). Research Questions In this study, the verbal and written outputs of three persons with aphasia were examined to determine if a relationshi p, particularly a double dissociat ion between nouns and verbs, existed across modalities. Specifically, the questions asked are as follows: 22

PAGE 23

1. Are there significant differences across nouns and verbs in oral naming, oral reading, written naming, and writing to dictation? 2. Does a double-dissociation exist across any of these modalities (i.e., nouns are more impaired in one modality and verbs are more impaired in the other modality)? 3. Based on an error analysis of incorrect respon ses, are the levels of impairment different for nouns and verbs across modalities? 23

PAGE 24

Figure 1-1. Rapp and Caramazza model (2002). [Reprinted with permission from Elsevier Limited. Rapp, B., & Caramazza, A. (2002). Selective difficulties with spoken nouns and written verbs: A single case study. (Page 397, Figure 4-a). Journal of Neurolinguisitics, 15 373402.] 24

PAGE 25

Figure 1-2. Ellis and Young model (1988; as cited in Kay, Lesser, & Coltheart, 1996). [Reprinted with permission from Psychology Press: Taylor and Francis Group. Kay, J., Lesser, R., & Coltheart, M. (1996). Ps ycholinguistic Assessments of Language Processing in Aphasia (PALPA): An introducti on. (Page 15, Figure 9). Aphasiology, 10, 159-215.] 25

PAGE 26

CHAPTER 2 METHODS Participants Three participants, two females and one male, were recruited from the University of Florida Speech & Hearing Clinic and hospitals in the Central Florida area to participate in this study. Participant 1 (P1) is a 41-year-old, righ t-handed female with 16 years of education. Participant 2 (P2) is a 49-y ear-old, right-handed female w ith 12 years of education, and Participant 3 (P3) is a 57-year -old, right-handed male with 12 years of education. P1 was diagnosed with Brocas Aphasia based on the Western Aphasia Battery (WAB; Kertesz, 1982) criteria. P2 was diagnosed with Conduction Ap hasia based on WAB criteria; however, through clinical judgment and considera tion that the participant demonstrated extremely limited verbal output with relatively intact comp rehension, the diagnosis was determined to be more accurately described as Brocas Aphasia. P3 also wa s diagnosed with Conducti on Aphasia based on WAB criteria; however, through clinical judgment, the diagnosis was determined to be more accurately described as Brocas Aphasia se condary to the occasional strings of fluent and grammatical speech with limited verbal output and relativel y intact comprehension demonstrated by the participant (Table 2-1 and Table 2-2). Inclusion criteria included a) right handedness; b) a diagnosis of aphasia with significant lexical retrieval impairments and relatively int act comprehension; c) no more than moderate cognitive impairments; d) a nega tive history of neurological impai rments (other than stroke), alcoholism, learning disorder, etc. All participants met these cr iteria; however, after completion of testing, a more detailed case history with P3 revealed a possible lear ning disorder, which had not previously been diagnosed. He reported ha ving trouble learning and r eading as a child as well as being held back one y ear in elementary school despit e active parent involvement and 26

PAGE 27

assistance. Two additional people with aphasia were tested and not included in the study due to one subjects over-qualified ability to complete the tasks and another subjects inability to produce words in the oral or written modalities. Pa rticipants who met the above listed criteria completed standardized and non-standardized asse ssments to determine the associated level of language impairments as well as the existence of a selective noun and/or verb deficits. The comprehensive evaluation included (a) Western Aphasia Battery (WAB; Kertesz, 1982), to determine the type and seve rity of the aphasia; (b) Boston Naming Test ( BNT Goodglass, Kaplan, & Weintraub, 1983) to determine oral naming abili ties of nouns across a range of frequencies; (c) word fluency tasks, to look at generative naming abilities; (d) Pyramids and Palm Trees (Howard & Patterson,1992); (e) Kissing and Dancing Test (Bak & Hodges, 2003), to determine the presence of semantic deficits; (f) An Object and Action Naming Test (across multiple modalities) (Druks & Masterson, 2000), to determine deficits in various modalities as well as the possibility of noun or verb deficits. Materials Lexical retrieval abilities as measured by stimuli from An Object & Action Naming Test (Druks & Masterson, 2000) were the prim ary dependent variable for this study. An Object & Action Naming Test consists of two forms (Form A and Fo rm B), which are matched for printed word frequency, rated familiarity of the verbal labels, and rated age-of-acquisition (Druks & Masterson). Form A and Form B each consis t of 81 object line-drawings and 50 action linedrawings. The pictures were us ed for oral naming and written naming. For the oral reading task, the words corresponding to each picture were type d and glued onto three-by-five note cards, in 16-point, Times New Roman font. Action words were typed in the present form rather than the present progressive (i.e., ing) form provided through An Object & Action Naming Test (Druks & Masterson). An example of th is change would include the word Kicking typed as Kick so 27

PAGE 28

that oral reading of verbs was not complicated by additionally requiri ng the processing and reading of morphology. Each participant was assessed using th e 131 words/pictures provided through An Object & Action Naming Test (Druks & Masterson, 2000). Although the test was di vided into object and action sections, it was evident that certain items could be considered both an action and object (i.e., ambiguous) by the participant. For example, the object word tie coul d be considered as a noun (i.e., a garment worn by a male around the neck ) or a verb (i.e., to join or connect two items). Nouns and verbs considered to be am biguous were removed and analyzed separately. Ambiguous nouns and verbs were classified by dete rmining those with a word frequency greater than 25% of the cumulative word frequency. Word frequency was determined using the Computational Analysis of Present-Day American English (Kucera & Fancis, 1964). Statistics regarding word frequency, rated familiarity of the verbal labels, and rated age-of-acquisition were completed to ensure validity of the nonambiguous word lists created for the study. Removing the ambiguous items did not result in a significant difference (p > .05) across any of the above listed characteristics, as determin ed through the use of a Mann-Whitney Test. Procedures An Object & Action Naming Test (Druks & Masterson, 2000) provides noun and verb stimuli in two forms (Form A and B). P1 and P2 received Form A, and P3 received Form B. The battery was administered over several sessi ons, for approximately 2 hours each session. All sessions were videoand/or au diorecorded to ensure the accuracy of scoring and uniform testing. Participants were tested acr oss six modalities; however, this study was completed to examine the relationship between four modalities: oral naming and written naming as well as oral reading and writing-to-dic tation. For oral naming, each part icipant was provided with the 28

PAGE 29

object and action line drawings to name orally. All answers were manually recorded by the examiner. At a later date, the participants were provided with the object and action line drawings in a different order and were as ked to write the name of the obj ect/action using a pen and paper provided. Oral reading was completed using the three-by-five note cards described above. The participant read aloud th e object and actions printed on the ca rds, while the examiner manually recorded his/her reply. Finally, the participants were asked to write object and actions words dictated in a different order by the examiner. This modality was completed with the participant recording his/her answers with a pen and paper pr ovided. Data collected in oral modalities was videoand audio-recorded, while written modalities were video-recorded only. Oral reading and writing-to-dictation action words were auditorily and visually presented to participants in pres ent form rather than the present progr essive form (e.g., crawl instead of crawling). This change was implemented to ensure that participants were solely tested on verb knowledge and not scored or penalized accordin g to morphological changes of a word. Accuracy Scoring Responses were audio recorded or produced orthographically by the participant. All collected data was audioand/or videorecorded to ensure accuracy. Recorded responses were considered accurate if no phonologic (i.e., no sound omission, addition, or substitution), orthographic (i.e., no letter omission, a dditional, or substitution), or se mantic errors were present. During written naming, however, verbs written in present progressive (i.e., using a gerund ing) form or present form by the participant were considered correct. Accuracy results for all participants are present in Ta ble 2-2 and lexical accuracy re sults are present in Table 2-3. Error Analysis An error analysis was completed to determine if any trends emerged through the types of errors made by each participant. Errors were labeled and divided into the following eleven 29

PAGE 30

categories: Semantic Error in the Same Gramma tical Class, Semantic Error in a Different Grammatical Class, Greater than or Equal to 50% of Phonemes/ Graphemes (i.e., sounds/ letters) Produced Correctly, Less than 50% of Phoneme s/ Graphemes Produced Correctly, Mixed Error (i.e., both a phonologic and semantic error), Un related Word within the Same Grammatical Class, Unrelated Word in a Different Grammati cal Class, Perseveration (i.e., repe tition of a previously stated/written word), Jargon (i.e., a nonsense word fabr icated by the participant), I Dont Know (IDK) or No Response (NR) given by the participant, and/or an Unintelligible/Indiscernible answer provided by the participant. Using the targeted word, candle, an example of each error type is provided in Table 2-4. Reliability Inter-rater reliability was done by a trained co mmunication sciences and disorders graduate student familiar with aphasic sp eech on 37% of the naming responses. Two different modalities within different grammatical classe s were evaluated for each particip ant. For example, with Pt 1, reliability was completed for written naming of verbs and oral reading of nouns. A point-topoint evaluation was conducted on each re sponse, resulting in 93% agreeability. 30

PAGE 31

Table 2-1. Descriptive Informa tion Regarding Participants Note: Pt: Participant; M: Male; F: Female; WA B: Western Aphasia Battery; CLQT: Cognitive Quick Linguistic Test. Pt. Sex Education (years) Age Site of Lesion Type of Aphasia Months Post Onset WABAphasia Quotient CLQT 1 F 16 41 Left MCA Brocas 48 55.5 2.0 (moderate) 2 F 12 49 Left MCA Brocas 9 36.8 1.8 (moderate) 3 M 12 57 Left Posterior Watershed Region Brocas 24 55.8 2.8 (mild) 31

PAGE 32

Table 2-2. Accuracy Scores for all Participants Total Possible P1 P2 P3 WAB Spontaneous Speech Information Content 10 8 4 7 Fluency 10 5 2 4 Total 20 13 6 11 Comprehension Yes/No Questions 60 51 58 54 Auditory Word Recognition 60 49 54 54 Sequential Commands 80 51 60 46 Total 200 151 172 154 Repetition Total 100 25 23 46 Naming Object Naming 60 29 11 29 Word Fluency 20 7 3 3 Sentence Completion 10 6 1 6 Responsive Speech 10 5 0 8 Total 100 47 15 46 Aphasia Quotient Total 100 55.5 36.8 55.8 Praxis 60 57 54 60 Reading and Writing Reading 100 54 65 80 Writing 100 46.5 60.5 57.5 Total 200 100.5 125.5 137.5 CLQT Attention Score 215 114(mod) 156(mild) 133(mild) Memory Score 185 83(severe) 70(severe) 134(mod) Executive Function Score 40 23(mod) 15(severe) 25(WNL) Language Score 37 12(severe) 6.5(severe) 19(severe) Visuospatial Skills Score 105 75(mild) 68(mild) 84(WNL) Clock Drawing Score 13 13(WNL) 11(mild) 9(mod) Composite Severity Score 4 2(mod) 1.8(mod) 2.8(mild) Note: WAB: Western Aphasia Battery; CLQT: Cognitive Linguistic Quick Test 32

PAGE 33

Table 2-3. Lexical Test Scor es for all Participants Test Total Possible P1 P2 P3 BNT Raw Score 60 19 1 7 % Correct 31.67 1.67 11.67 Pyramids and Palm Trees Raw Score 52 49 49 52 % Correct 94.23 94.23 100 Kissing and Dancing Raw Score 52 51 47 49 % Correct 98.07 90.38 94.23 Verbal Fluency/ Generative Naming Seconds CLQT CLQT CLQT Animals 0-15 6 1 1 16-30 2 0 2 31-45 1 0 1 46-60 1 0 0 Total animals 10 1 4 M words 0-15 1 1 0 16-30 0 1 1 31-45 0 0 0 46-60 0 0 0 Total M words 1 2 1 Verbs/actions 0-15 1 1 3 16-30 0 0 1 31-45 0 0 0 46-60 0 0 0 Total verbs/ actions words 1 1 4 Note: BNT: Boston Naming Test; CLQT : Cognitive Linguistic Quick Test. 33

PAGE 34

Table 2-4. Error Analysis Cate gories (Targeted Word: Candle) Error Type Example Descriptive Characteristic Waxy Semantic Error in Same Grammatical Class Match Semantic Error in Different Grammatical Class Burn 50% of Phonemes/ Graphemes Correct Canl/ Candl < 50% of Phonemes/ Graphemes Correct Ca/Canl Mixed Phoneme/ Grapheme Error Satch/ Satch Unrelated word within the Same Grammatical Class Table Unrelated word within a Different Grammatical Class See Perseveration Foot* Jargon Cadoo IDK/ NR I dont Know Unintelligible/ Indiscernible C..? Note: *A word previously an swered during the assessment. 34

PAGE 35

CHAPTER 3 RESULTS The accuracy data for each participant will be discussed in turn (Table 3-1 through Table 3-3). Following the accuracy data are the error analysis data (Figure 3-1 through Figure 3-3). Accuracy A Fishers Exact Test was used to analyze oral naming, oral read ing, written naming, and writing to dictation accuracy results to determin e if there was a difference in naming accuracy between nouns and verbs within each modality (Berndt et al., 1997a). Th ere was no significant difference across any modalities for participant 1, 2, or 3. Participant 1 (P1) Accuracy Overall, Participant 1 exhibited higher accuracy on tasks in the oral modality as compared to tasks in the written modality (Table 3-1 and Figure 3-1). Oral naming Although P1 exhibited a higher percent accu racy for noun naming (60.81%) as compared to verb naming (45.16%), this difference was not significant ( p = 0.1956). Oral reading There was no statistical difference between noun oral naming (68.92%) and verb oral naming (67.74%) ( p = 1.000). Written naming There was no statistical difference between noun written naming (9.46%) and verb written naming (3.23%) (p= 0.4314). Writing to dictation There was no statistical difference between noun writing to dictation (10.81%) and verb writing to dictation (6.45%) (p= 0.7196). 35

PAGE 36

Participant 2 (P2) Accuracy Participant 2 demonstrated higher percent accura cy on written tasks as compared to oral tasks (Table 3-2 and Figure 3-2). Oral naming There was no statistical difference between noun oral naming (4.05%) and verb oral naming (0%) (p= 1.437). Oral reading There was no statistical difference between noun oral reading (17.57%) and verb oral reading (16.13%) ( p= 1.0000). Written naming There was no statistical difference betw een noun written naming (66.22%) and verb (51.61%) (p = 0.2789). Writing to dictation There was no statistical difference between noun writing to dictation (74.32%) and verb writing to dictation (83.87%) (p= 0.3225). Participant 3 (P3) Accuracy Participant 3 did not demonstrate greater pe rcent accuracy for oral or written tasks; however, he did reveal a trend toward decreased naming abilities (Table 3-3 and Figure 3-3). Oral naming There was no statistical difference between noun oral naming (30.99%) and verb oral naming (31.03%) ( p= 1.0000). Oral reading Although noun oral reading was lower (45.07%) than verb oral reading (65.52%), the difference was not statistically significant ( p = 0.0793). 36

PAGE 37

Written naming There was no statistical difference betw een noun written naming (24.29%) and verb written naming (27.59%) ( p= 0.8010). Writing to dictation There was no statistical difference between noun writing to dictation (76.06%) and verb writing to dictation (82.76%) (p= 0.1708). Error Analysis The results of the error analysis conducted on al l incorrect responses on An Object and Action Naming Test (Druks & Masterson, 2000) across all m odalities are presented on Figures 34 to 3-6. Participant 1 Error Analysis Oral naming Nouns Semantic errors were the most pre dominant error type for verb naming, with production of semantically related nouns ( 48.28%) more prevalent than production of semantically related verbs ( 3.45%) (as determined by same a nd different grammatical class errors, respectively). The next highest level of errors was phonemic errors (20.69%) with at least 50% of target phonemes produced correctly and mo re than 50% of the target verbs produced incorrectly 3.45% of the time. Unrelated words in the same grammatical class (i.e., nouns) were produced 10.34% of the time, followed by mixed (6.90%) and unrelated di fferent grammatical class errors (6.90%) Verbs Semantic errors were the most predom inant error type for oral naming with verbs, with production of semantically related no uns (i.e., semantic different grammatical class, 29.41%) produced more often than production of related verbs (i.e., semantic same grammatical class, 11.76%). The next highest level of errors was unrelated grammatical class errors (same 37

PAGE 38

grammatical class 5.88%, different grammatical class 29.41%). Only two other error types were observed, mixed errors (17.6%) and jargon (5.88%). Oral reading Nouns The primary error was phonemic (39.13%, 50% phonemes correct and 17.39%, < 50% correct), followed by semantic sa me grammatical class (21.74%), then NR/IDK and Mixed (both = 8.70%), and jargon (4.35%). Verbs The error types for verb reading were ev enly distributed across semantic errors (30%), phonemic errors ( 50% phonemes correct) (30%), and unrelated word errors (30%). NR/IDK errors composed 10% of all errors. Written naming Nouns The predominant error for written naming was NR/IDK responses (50.75%), followed by jargon (14.93%), 50% of graphemes correct (10.45%), semantic same grammatical class (7.46%), and unrelated same grammatical cl ass (7.46%). The remaining 4 error categories were all less than 5% each. Verbs The predominant error for written naming of verbs was NR/IDK responses (83.33%), with the remaining errors falling in to mixed (6.67%), indiscernible (6.67%), and 50% correct graphemes (3.33%) categories. Writing to dictation Nouns The primary error type was NR/IDK responses (68.18%), followed by indiscernible (12.12%) responses, 50% correct graphemes (7.58 %), and jargon (4.55%). The remaining 3 categories were each composed of 3% or le ss of all errors. Verbs The primary error type was NR/IDK responses (75.86%), followed by indiscernible (10.34 %) responses, 50% correct graphemes (6.90%), semantic different grammatical class (3.45%), and unrelated same gramma tical class (3.45%) errors. 38

PAGE 39

Participant 2 Error Analysis Oral naming Nouns Jargon (33.80%) was the main error for oral naming with nouns, followed by NR/IDK errors (23.94%), and phonemic errors (16.90%, 50% phonemes correct and 8.45%, < 50% correct). Unintelligible responses (14.08%), unrelated sa me grammatical class errors (1.41%), and unrelated different grammatical class errors (1.41%) comprised the remaining error types. Verbs NR/IDK errors (54.84%) were the main error for oral naming with verbs, followed by phonemic errors with 50% of phonemes correctly st ated (16.13%), unintelligible responses (12.90%), and unrelated different gra mmatical class errors (6.45%). Semantic different grammatical class erro rs (3.23%), mixed errors (3.23%), and jargon (3.23%) comprised the remaining errors. Oral reading Nouns Jargon (36.07%) was the primary error fo r oral reading with nouns, followed by phonemic errors (24.59%, 50% phonemes correct and 11.48%, < 50% correct), and unintelligible errors (18.03%). Unrelated grammatical class errors (same grammatical class 6.56%, different grammatical class 1.64%) and pe rseveration (1.64%) comprised the remaining errors. Verbs Phonemic errors (38.46 %, 50% phonemes correct and 3.85 %, < 50% correct) were the primary error type for oral readi ng with verbs, followed by unrelated different grammatical class errors (26.92 %), and jargon (19.23%) errors. The remaining errors were comprised of unintelligible response s (7.69%) and persev eration (3.85%). 39

PAGE 40

Written naming Nouns The predominant error for written naming with nouns were grapheme errors (40.00%, 50% graphemes correct a nd 8.00 %, < 50% correct), followed by semantic errors (semantic same grammatical class 16.00%, sema ntic different grammatical class 4.00%), and NR/IDK errors (12.00%). Unrelated grammatical class errors (unrelated same and different grammatical class errors both at 8.00%) and jargon (4.00%) comprised the remaining error types. Verbs Unrelated different grammatical cl ass errors (33.33%) and semantic (same grammatical class 13.33%, differe nt grammatical class 20.00%) were the predominant error with written naming with verbs, followed by <50% of graphemes correctly produced (13.33%) and mixed errors (13.33%), and jargon (6.67%) errors. Writing to dictation Nouns The main error type for writing to dict ation with nouns was grapheme errors (52.63%, 50% graphemes correct a nd 5.26 %, < 50% correct), fo llowed by NR/IDK responses (26.32%), and unrelated same gram matical class errors (10.53%). Verbs The main error type for writing to dictation with verbs was grapheme errors (40.00%, 50% graphemes correct and 20.00%, < 50% correct), which was followed by NR/IDK responses (40.00%). Participant 3 Error Analysis Oral naming Nouns The predominant error for oral naming with nouns was NR/IDK responses (63.27%), followed by unrelated grammatical cl ass errors (same gram matical class 10.20%, different grammatical class 2.04 %), and semantic errors (same grammatical class 8.16%, different grammatical class 2.04 %). Greater than and equal to 50% of phonemes correctly 40

PAGE 41

produced (6.12%) and unintelligible responses (6 .12%), as well as, jargon (2.04%) comprised the remaining error types. Verbs The predominant error for oral naming with verbs was NR/IDK responses (60.00%) as well, followed by semantic errors (same grammatical cl ass 20.00%, different grammatical class 10.00%), and unrelated same grammatical cl ass errors (10.00%). Oral reading Nouns The primary error for oral reading w ith nouns was NR/IDK responses (64.10%), followed by 50% of phonemes correctly produced (17.95% ), and unrelated grammatical class errors (same grammatical class errors 7.69%, di fferent grammatical class errors 2.56%). The remaining errors were unintelligible responses (5.13%) and jargon (2.56%). Verbs The primary error for oral reading with verbs was unrelated grammatical class errors (same grammatical class errors 20.00%, different grammatical class errors 30.00%), followed by phonemic errors ( 50% phonemes correct produced, 20.00% and <50% of phonemes correctly produced, 10.00%), and semantic errors (same grammatical class errors 10.00%, different grammatical class errors 10.00%). Written naming Nouns NR/IDK responses (66.04%) were the main error type for written naming of nouns, followed by grapheme errors ( 50% graphemes correct produced, 15.09% and <50% of graphemes correctly produced, 1.89 %), and jargon (5.66%) and indiscernible responses (5.66%). The remaining errors were comprised of semantic same grammatical class errors (3.77%), and mixed errors (1.89%). Verbs NR/IDK responses (61.90%) were also the main error type for written naming of verbs, followed by semantic errors (same gr ammatical class errors, 9.52% and different 41

PAGE 42

grammatical class errors, 4.76%), 50% of graphemes correctly produced and unrelated same grammatical class errors (both at 9.52%), and jargon errors (4.76%). Writing to dictation Nouns The predominant error for writi ng to dictation with nouns was 50% of graphemes correctly produced (58.82%), fo llowed by jargon (23.53%), and indiscernible responses (17.65%). Verbs The predominant error for writing to dictation with verbs was 50% of graphemes correctly produced (80.00%), which was followed by jargon errors (20.00%). 42

PAGE 43

Table 3-1. P1 Accuracy Data Modality Nouns % Accuracy Verbs % Accuracy Oral Naming 60.81 45.16 Oral Reading 68.92 67.74 Written Naming 9.46 3.23 Writing to Dictation 10.81 6.45 Table 3-2. P2 Accuracy Data Modality Nouns % Accuracy Verbs % Accuracy Oral Naming 4.05 0.00 Oral Reading 17.57 16.13 Written Naming 66.22 51.61 Writing to Dictation 74.32 83.87 Table 3-3. P3 Accuracy Data Modality Nouns % Accuracy Verbs % Accuracy Oral Naming 30.99 31.03 Oral Reading 45.07 65.52 Written Naming 24.29 27.59 Writing to Dictation 76.06 82.76 43

PAGE 44

P1 Percent Accuracy 0 10 20 30 40 50 60 70 80 90 100 Oral Naming Oral Reading Written Naming Writing to DictationPercentage Nouns Verbs Figure 3-1. Visual Representation of P1s Accuracy Data. P2 Percent Accuracy 0 10 20 30 40 50 60 70 80 90 100 Oral Naming Oral Reading Written Naming Writing to DictationPercentage Nouns Verbs Figure 3-2. Visual Representation of P2s Accuracy Data. 44

PAGE 45

45 Figure 3-3. Visual Representa tion of P3s Accuracy Data. P3 Percent Accuracy 0 10 20 30 40 50 60 70 80 90 100 Oral Naming Oral Reading Written Naming Writing to DictationPercentage Verbs Nouns

PAGE 46

Oral NamingP10 10 20 30 40 50 60 70 80 90 100Semantic same grammatical Semantic different grammatical 50% Phonemic < 50% Phonemic Mixed Unrelated same grammatical Unrelated different grammatical Perseveration Jargon NR/IDK Unintelligible Percentage Nouns Verbs A Oral Reading-P10 10 20 30 40 50 60 70 80 90 100Semantic same grammatical Semantic different grammatical 50% Phonemic < 50% Phonemic Mixed Unrelated same grammatical Unrelated different grammatical Perseveration Jargon NR/IDK Unintelligible Percentage Nouns Verbs B 4646 Figure 3-4. P1s Error Analyses Graphs A) Error Analyses Graph of Oral Naming for P1. B) Error Analyses Graph of Oral Reading for P1. C) Error Analyses Graph of Written Naming for P1. D) Error Analyses Graph of Writing to Dictation for P1. 46

PAGE 47

Written NamingP10 10 20 30 40 50 60 70 80 90 100Semantic same grammatical Semantic different grammatical 50% Graphemes < 50% Graphemes Mixed Unrelated same grammatical Unrelated different grammatical Perseveration Jargon NR/IDK Indiscernible Percentage Nouns Verbs C Writing to DictationP10 10 20 30 40 50 60 70 80 90 100Semantic same grammatical Semantic different grammatical 50% Graphemes < 50% Graphemes Mixed Unrelated same grammatical Unrelated different grammatical Perseveration Jargon NR/IDK Indiscernible Percentage Nouns Verbs D 47 Figure 3-4. Continued 47

PAGE 48

Oral NamingP20 10 20 30 40 50 60 70 80 90 100Semantic same grammatical Semantic different grammatical 50% Phonemic < 50% Phonemic Mixed Unrelated same grammatical Unrelated different grammatical Perseveration Jargon NR/IDK Unintelligible Percentage Nouns Verbs A Oral ReadingP20 10 20 30 40 50 60 70 80 90 100Semantic same grammatical Semantic different grammatical 50% Phonemic < 50% Phonemic Mixed Unrelated same grammatical Unrelated different grammatical Perseveration Jargon NR/IDK Unintelligible Percentage Nouns Verbs B 48 Figure 3-5. P2s Error Analyses Graphs A) Error Analyses Graph of Oral Naming for P2. B) Error Analyses Graph of Oral Reading for P2. C) Error Analyses Graph of Written Naming for P2. D) Error Analyses Graph of Writing to Dictation for P2. 48

PAGE 49

Written NamingP20 10 20 30 40 50 60 70 80 90 100Semantic same grammatical Semantic different grammatical 50% Graphemes < 50% Graphemes Mixed Unrelated same grammatical Unrelated different grammatical Perseveration Jargon NR/IDK Indiscernible Percentage Nouns Verbs C Writing to DictationP20 10 20 30 40 50 60 70 80 90 100Semantic same grammatical Semantic different grammatical 50% Graphemes < 50% Graphemes Mixed Unrelated same grammatical Unrelated different grammatical Perseveration Jargon NR/IDK Indiscernible Percentage Nouns Verbs D 49 Figure 3-5. Continued 49

PAGE 50

Oral NamingP30 10 20 30 40 50 60 70 80 90 100Semantic same grammatical Semantic different grammatical 50% Phonemic < 50% Phonemic Mixed Unrelated same grammatical Unrelated different grammatical Perseveration Jargon NR/IDK Unintelligible Percentage Nouns Verbs A Oral Reading-P30 10 20 30 40 50 60 70 80 90 100Semantic same grammatical Semantic different grammatical 50% Phonemic < 50% Phonemic Mixed Unrelated same grammatical Unrelated different grammatical Perseveration Jargon NR/IDK Unintelligible Percentage Nouns Verbs B 50 Figure 3-6. P3s Error Analyses Graphs A) Error Analyses Graph of Oral Naming for P3. B) Error Analyses Graph of Oral Reading for P3. C) Error Analyses Graph of Written Naming for P3. D) Error Analyses Graph of Writing to Dictation for P3. 50

PAGE 51

51 Written Naming AnalysisP30 10 20 30 40 50 60 70 80 90 100Semantic same grammatical Semantic different grammatical 50% Graphemes < 50% Graphemes Mixed Unrelated same grammatical Unrelated different grammatical Perseveration Jargon NR/IDK Indiscernible Percentage Nouns Verbs C Writing to Dictation-P30 10 20 30 40 50 60 70 80 90Semantic same grammatical Semantic different grammatical 50% Graphemes < 50% Graphemes Mixed Unrelated same grammatical Unrelated different grammatical Perseveration Jargon NR/IDK Indiscernible Percentage Nouns Verbs D Figure 3-6. Continued 51

PAGE 52

CHAPTER 4 DISCUSSION The purpose of this study was to examine the existence of a significant difference between nouns or verbs as well as the presence of a double dissociation, in oral naming, oral reading, written naming, and writing to dictation were ex amined for three participants with nonfluent aphasia. Then, using the adapted Ellis and Young model (1988; as c ited in Kay, Lesser, & Coltheart, 1996), the modality and/or pathway defi cit associated with each tested modality was targeted following error analyses in order to ev aluate whether this clos e analysis would reveal differences in noun and verb processing poten tially not seen in accuracy performance. There were no significant di fferences between nouns and verbs in oral naming, oral reading, written naming, and writing to dictati on regarding accuracies in any of the three participants. Additionally, a double dissociation across any of the modalities was not identified in the participants. These results are not unco mmon and appear consistent with findings from other related studies (e.g., Miceli, Silveri, Nocentini, & Caramazza, 1988; Berndt, Mitchum, Haendiges, & Sandson, 1997a). In Miceli et al., only seven out of 25 participants with aphasia were found to demonstrate grammatical category differences. The Ber ndt et al. (1997a) study had similar findings as well. Seven out of the 11 tested individuals with aphasia demonstrated an advantage in either nouns or verbs, while f our individuals showed no grammatical class advantage (Berndt et al., 1997a). Furthermore, selective grammatical impairments and double dissociations in the literature generally pertai n to a limited number of documented individuals (e.g., Hillis & Caramazza, 1991; Hillis & Ca ramazza, 1995; Rapp & Caramazza, 2002). Overall, no significant differen ce in noun or verb retrieval wi thin any modality was noted; thus, no double dissociation within a grammatical deficit across the modalities was present for any participant (i.e., research questions 1 and 2). Therefore, the focus of the discussion will 52

PAGE 53

regard the results of the error analysis and address whether ther e is a difference in impairment level for nouns or verbs in any modality across eac h participant (i.e., resear ch question 3). Since there was no evidence of a double dissociation, the Rapp and Caramazza model (2002) will not be used to examine hypothesized error levels. Rather, the adapted Ellis and Young model (1988; as cited in Kay et al., 1996) will be used because it is well-accepted and appropriate to explain the proposed level of deficit associated with each tested modality. Through the completion of an error analysis for each modalit y, locations of possible deficits for nouns and verbs within the linguistic model were identified for the particip ants. Although no statistical difference between noun and verb accuracies existed, modality impairments and/or noun or verb impairments were noted the participants. These tre nds provide valuable insight rega rding interpretation of the error analyses and possible deficit locations. Participant 1 P1 demonstrated a trend toward greater accuracy with oral tasks (i.e., oral naming and oral reading). P1 also showed a trend toward in creased accuracy with nouns across all modalities (e.g., oral naming nouns, 60.81%; oral naming verbs, 45.16%; written naming nouns, 9.46%; written naming verbs, 6.45%). This trend can al so be seen in results from additional lexical testing. P1s increased ability to name more nouns (i.e., animals) than verbs (i.e., actions) in generative naming tasks further reveals a great er deficit with verb s (Table 2-3). Based on Participant 1s error analysis, the follo wing deficits were rec ognized (Table 4-1). For oral naming, P1s greatest percentage of errors for nouns was semantic, followed by phonemic, and then unrelated same grammatical cl ass errors. Due to th e high percentage of semantic errors, it is likely that the semantic system is impaired; however, considering P1s over all naming accuracy, her Pyramids and Palm Trees (Howard & Patters on,1992) results, and generative naming scores, it is probable that semantic system remains intact to an extent (Table 53

PAGE 54

2-2 and 2-3). A deficit in the visual object recognition system is unlikely for nouns or verbs, considering P1s high scores on the Pyramids and Palm Trees (Howard & Patterson, 1992) and Kissing and Dancing Test (Bak & Hodges, 2003) (94.23% an d 98.07 %, respectively), which utilize the visual object recogniti on naming route. The lack of a deficit in the visual object recognition naming route would be seen in writte n naming as well. Due to the phonemic errors and mixed errors, impairment in the phonological output lexicon and partial impairment in the pathway between the semantic system and phonologica l output lexicon is likely present. It is uncertain, however whether the number of phonemic errors 50% of the targeted phonemes was do to previous semantic system/pathway errors or to impairment of the phonological output system itself. Unrelated same and different gra mmatical class errors made by P1 were often due to naming other objects present in the picture/stimuli or to the participant describing the object (e.g., for a picture of a cherry, the participant sa id yucky), which could also imply that the participant was unable to correctly activate th e targeted word seman tically and therefore described the object or named another item in th e picture that was more easily activated. The majority of errors for oral naming with verbs for P1 was semantic, followed by unrelated different grammatical classes, and then mixed errors. This would also imply that an impairment in the semantic system because ve rbs were not always activated or retrieved appropriately (i.e., words from another grammatical class, such as nouns, were activated instead). Impairment in the pathway leading from the se mantic system to the phonological output lexicon is likely present as well, when considering the nu mber of mixed errors. The number of unrelated grammatical errors was primarily due to P1 na ming nouns rather than verbs, providing more evidence that verbs are more impaired in this part icipant. This type of error would occur in the semantic system as well. 54

PAGE 55

P1 showed the greatest percentage of errors for oral reading with nouns in phonemic errors ( 50% and < 50% phonemes correctly produced), followed by semantic, and then mixed and NR/IDK response errors. This would then imply a primary impairment in the phonological output lexicon. When considering the percentage of semantic same grammatical class and mixed errors, an impairment in the semantic system and a partial impairment in the pathway between the semantic system and the phonological output le xicon is likely present as well. Since P1 was overall more accurate with nouns across modalities, the greate r percentage of 50% phonemes correctly produced and the pres ence of < 50% phonemes correctly produced errors with nouns (as compared to verbs) likely demonstrates a great er attempt of using the semantic system when reading nouns. The greatest percentage of errors pertai ning to oral reading of verbs came from 50% of phonemes correctly produced, which was follo wed by semantic and unrelated different grammatical class errors (i.e., words from other grammatical cl asses were produced, which could suggest semantic system impairment). These er rors would imply a deficit within the semantic system and the phonological output lexicon as well. It is possible, however, that an alternative route ma y have been utilized by P1 for oral reading with nouns and verbs. Through previous testing of P1s ability to copy printed words with the same word list (not re ported in the results section), P1 was able to copy nouns and verbs with 98.77% and 100% accuracy, respec tively. Therefore, it is possibl e that she used the route of abstract letter identification orthographic input lexicon phonological output lexicon phonological output buffer to read the printed word s aloud. Considering P1s primary error type with oral naming of nouns and verbs was semantic (i.e., the semantic system was impaired), it is probable that the participant ut ilized the above route to avoid use of the impaired semantic 55

PAGE 56

system. Partial use of this route to avoid semantic system use would explain the overall increased accuracy within the modality. Thus the primary impairment for nouns and verbs in this modality is primarily within the semantic system with additional involvement in the phonological output lexicon. Impairment in th e orthographic input lexicon is unlikely, considering P1s accuracy with copying words. Use of the non-lexical route: ab stract letter identification letter-to-sound rules phonological output buffer for oral reading is un likely considering P1s copying ability and possible use of the above listed route that bypasses the semantic system. In addition, irregular words, such as weigh, were read aloud corre ctly by the participant, which would not be possible when using th e non-lexical route. In the written modality, greater than 50% of the errors for nouns and verbs in both tasks were NR/IDK responses for P1. In written na ming with nouns P1 displayed the greatest number of errors with NR/IDK responses, followed by jargon, 50% and < 50% graphemes correct, and semantic same grammatical class. This would imply a primary deficit at the orthographic output lexicon rather than the semantic level, since oral naming accuracy was much higher. However, as indicated with the oral modality results, the sema ntic system is likely degraded to some extent, considering the percentage of se mantic and unrelated same gra mmatical class errors. P1 had overall greater accuracy with nouns across all moda lities; therefore, the presence of semantic errors in written naming of nouns is likely due to P1s increased attempts to provide answers. With verbs, however, activation of the semantic system may not have been as strong, considering the greater percentage IDK/NR responses with verbs as compared to nouns. For written naming with verbs the greatest num ber of errors after NR/IDK responses was mixed errors, followed by indiscer nible errors, and errors where 50% graphemes were 56

PAGE 57

correctly written. These errors would imply a deficit in the pathway between the semantic system and orthographic output lexi con. Partial/possible impairment of the semantic system is likely considering the pathway involvement a nd the results from written naming with nouns. For writing to dictation with nouns, P1 had th e primary number of errors with NR/IDK responses. Indiscernible res ponses were next, followed by 50% correct graphemes, and then jargon. These errors would imply a deficit at the orthographic output lexicon for nouns with writing to dictation. For writing to dictation with verbs, th e primary number of errors, after NR/IDK responses, was indiscernible responses, followed by 50% correct graphemes, then semantic different grammatical class and unrelated same grammatical cla ss errors. A deficit in the orthographic output lexicon and a partial/possible deficit in th e semantic system would be consistent with these error types. Impairment in the phonological input lexicon, which is utilized in writing to dictation, is unlikely considering P1s high accuracy with re petition (not reported in results) of nouns (81.48%) and verbs (88.00%) when using the same word lists. During writing to dictation, it is important to consider that the participant may have utilized the route of a uditory phonological analysis phonological input buffer phonological input lexicon phonological output lexicon orthographic output lexicon orthographic output buffer to avoid use of an impaired semantic system. Utilization of this route is possible; however, its use would not increase accuracy due to the impaired orthographic output lexicon. Considering that P1s overall accuracy for writin g to dictation was low, it is likely that the orthographic output lexicon is the primary site of impairment. Possible/partial impairment in the semantic system is likely for nouns and verbs as we ll, considering the seman tic errors made with verbs in writing to dictation and the semantic impairment already documented in previous 57

PAGE 58

modalities. Partial/possible impairment of th e phonological output lexicon (when using the alternative route) is also likely considering the previous deficits present in oral naming and oral reading. Use of the non-lexical route: acoustic-to-phonological conversion phonological output buffer sound-to-letter rules orthographic buffer is unlikely when considering P1s adequate repetition ability and the possible use of the above listed route that bypasses the semantic system. In addition, the semantic system is at least partially in tact in P1 and irregular words, such as knock, were written correctly by the participant, which w ould further prove that the nonlexical route was not utilized. Overall for P1, it is apparent through error anal ysis that the semantic system is primarily impaired for oral tasks, while the orthographic output lexicon, with partial involvement of the semantic system, is primarily impaired for written tasks. The semantic system is impaired for nouns and verbs in the oral modalities, with prob able phonological output lexicon involvement in oral reading of nouns and verbs and oral naming of nouns. Conversely, ther e is an orthographic output lexicon impairment for nouns and verbs in the written modalities, w ith partial semantic system involvement. A pattern regarding noun and verb deficits within specific modalities (i.e., double dissociations) was not apparent for P1 after completing the error analysis. Participant 2 P2 displayed a trend toward increased accuracy in written modalitie s (i.e., written naming and writing to dictation). No tr end of noun or verb impairment wa s present across oral or written modalities. This lack of a trend is further supported by P2s results in lexical testing of noun and verb generative naming, where she obtained th e same score in both tasks (Table 2-3). For oral naming of nouns, P2s greatest pe rcentage of errors was jargon and NR/IDK responses, which were followed by phonemic errors ( 50% and < 50% phonemes correctly 58

PAGE 59

produced). Considering these errors, it is likel y that there is a significant deficit in the phonological output lexicon, which would inhibit P2 from naming several of the pictures (i.e., NR/IDK responses) and/or cause the retrieval of incorrect phonemes (i.e., jargon/phonemic errors) (Table 4-2). Similarly, considering that the majority of errors were with NR/IDK responses and 50% of phonemes correctly produced for oral naming of verbs, it is likely that the phonological output lexicon is impaired for verbs as well. This ex plains P2s inability to correctly retrieved phonemes or name over half of the missed pictures. Less than 10% of the errors for nouns and verb s in oral naming were either semantic or unrelated grammatical class erro rs, which imply partial impairment of the semantic system. However, it is clear that the prim ary deficit location in this modali ty is in the phonological output lexicon. P2s high scores on the lexical tests, Pyramids and Palm Trees (Howard & Patterson, 1992) and Kissing and Dancing Test (Bak & Hodges, 2003) (94.23% and 90.38%, respectively) provide further evidence that th e semantic system remains primarily intact as well as that the visual object recognition systems is functional for this modality and written naming (Table 2-2). Further evidence of phonological ou tput lexicon impairment is seen through P2s repetition accuracy (not present in results section). Fo r repetition with nouns and verbs, P2 achieved 20.99% and 20.00% accuracy, respectively. Since re petition uses a route that bypasses the semantic system, it is clear that the phonological out put lexicon is impaired and it is possible that the phonological input lexicon is impaired as well. The greatest percentage of errors for nouns in oral reading was jargon and phonemic errors ( 50% and < 50% phonemes correctly produced). This would again imply that the phonological output lexicon is greatly impaired, which inhibits P2 from naming several of the pictures and/or 59

PAGE 60

causes the retrieval of incorrect phonemes. Partial impairment of the semantic system is possible considering the small percentage of unr elated same gramma tical class errors. The greatest percentage of errors for verbs with oral reading was phonemic errors ( 50% and < 50% phonemes correctly produced), followe d by unrelated different grammatical class errors, and jargon. These error types would also imply an impairment with the phonological output lexicon. Partial impairment of the semantic system may also be involved when considering the percentage of e rrors made with unrelated different grammatical class (e.g., verbs were produced as related nouns). Similar to P1, it is possible that an alternativ e route may have been utilized by P2 for oral reading with nouns and verbs. Through previous testing of P2s ability to copy printed words using the same word list (not prev iously reported in th e results), P2 was ab le to copy nouns and verbs with 100% accuracy. Therefore, it is possible that she also used the route of abstract letter identification orthographic input lexicon phonological output lexicon phonological output buffer to read the printed words aloud. Cons idering the perfect accuracy P2 showed with copying, it is clear that the orthographic input le xicon is functional. Although partial impairment of the semantic system is evident through the or al naming and oral readi ng analyses, primary use of this route to bypass the semantic system is unlik ely given the similarities in accuracy scores in both modalities (i.e., both with low accuracy). Thus it further supports the conclusion that an impairment in the phonological output lexicon exists. Use of the non-lexical route: ab stract letter identification letter-to-sound rules phonological output buffer is also unlikely consider ing P2s ability to accurately copy. Also, irregular words (e.g., weight) were read aloud co rrectly by the participan t, which would not be possible if using the non-lexical route. 60

PAGE 61

Initially P2 appeared to have greater accura cy with nouns in written naming and greater accuracy with verbs in writing to dictation. Th is prompted further an alysis in the written modalities. The main error type for verbs in wr itten naming was unrelated different grammatical class errors and semantic different grammatical cl ass errors (i.e., semantic system impairment). Considering the unrelated and semantic errors were in different grammatical classes (e.g., verbs nouns), P2s increased accuracy with no uns in this modality is easily understood. Graphemic and mixed errors (both at 13.33%) followed the unrelated grammatical errors/ semantic errors in written naming with verbs, showing partial impairment at the orthographic output lexicon and the pathway between it and the semantic system as well. Graphemic errors were the primary error made for nouns in writ ten naming, followed by semantic and unrelated grammatical class errors. This would imply a deficit at the orthographic output lexicon with partial impairment at the semantic system for nouns. Considering the error results above, a diffe rence in the level of processing for noun and verbs was established in written naming. Over all, a higher level of processing for nouns was apparent through the higher accuracy (66.22%), as we ll as, the fact that responses were closer to the targeted word (i.e., 50% graphemes correctly produced). Increased impairment with verbs in written naming was evident through the slightly decrease accuracy ( 51.61%), as well as, the fact that the primary error t ypes of unrelated grammatical cl ass and semantic class errors involved changing of verbs to another grammati cal class (i.e., nouns). This difference in noun and verb processing is not acc ounted for in the adapted Ellis and Young model (1988); however, it is consistent with the Caramazza model (2002) where nouns and verbs are processed separately at each lexical level. Further investigation of the model regarding the m odality errors would be needed before drawing a conclusion. 61

PAGE 62

Conversely, the error analyses of writing to dictation initiall y showed increased accuracy with verbs. For writing to di ctation with nouns, graphemic erro rs and NR/IDK responses were the primary error, followed by unrelated same grammatical class errors A deficit at the orthographic output lexicon is likel y due to the high number of gra pheme errors and inability to retrieve any correct graphemes (i.e. NR/IDK responses). Partia l/ possible involvement with the semantic system due to the presence of unrelated grammatical class errors is probable as well. For verbs, the predominant error with writing to dictation was graphemic as well, which was followed by NR/IDK responses. A deficit at the orthographic output lexicon is likely when considering the participants inabi lity to correctly retrieve any or all of the accurate graphemes. During writing to dictation, it is important to consider that the participant may have utilized the route of a uditory phonological analysis phonological input buffer phonological input lexicon phonological output lexicon orthographic output lexicon orthographic output buffer to avoid use of the semantic system if it is impaired. Considering P2s poor repetition abilities and phonologica l deficits, it is likely that the phonological input lexicon and/or the phonological output lexicon are impaired and that sh e did not utilized this route for nouns or verbs when writing to dictation. Use of the non-lexical route: acoustic-to-phonological conversion phonological output buffer sound-to-letter rules orthographic buffer is possible, considering P2s poor repetition abilities. P2 may have utilized the non-lexical route on occasion; however, it is evident that the semantic system was also activated at times due to the presences of sema ntic errors with nouns and certain irregular words (e.g., kneel) were wr itten correctly by the participant, which also suggest the use of a lexical route. 62

PAGE 63

Considering the error types and possible routes for writing to dictation, the possibility of increased accuracy with verbs over nouns within the modality is not as evident. For instance, verbs had a higher percentage of IDK/NR responses (i.e., the participant did not attempt to write those words). Also, the increased accuracy with verbs may have been cont ributed to the use of the non-lexical route. Utilization of this route is less likely for nouns considering the presence of semantic errors. Combining the results from the e rror analyses of the two written modalities, it is possible that a different level of processing for nouns and verbs was present in written naming; however, a different level of processing in writing to dictatio n is less likely. Overall for P2, it is apparent through error anal ysis that the primary deficits in the lingual system is at the lexeme (i.e., output lexicon) level. The phonological output lexicon is impaired for nouns and verbs in the oral modalities, with partial semantic system deficits. The orthographic output lexicon is impaired for nouns and verbs in the written modalities, with partial semantic impairment in written naming. A higher level of processing for nouns was present in written naming; however, a pattern re garding noun and verb deficits within specific modalities (i.e., double dissociations) was not found for P2 after completing an error analysis. Participant 3 P3 revealed a trend toward decreased naming abilities (i.e., decreased accuracy with oral and written naming). A trend to ward increased accuracy with verbs across all modalities was also noted except for oral naming. These trends are further reinforced through P3s results from additional lexical testing, wher e he obtained the same score for noun and verb oral/generative naming tasks (Table 2-3). The predominant error for P3 in oral na ming with nouns was NR/IDK responses, followed by unrelated grammatical class errors, semantic errors, and 50% of phonemes correctly produced. It is likely that a defi cit within the semantic system exists for oral naming of nouns. 63

PAGE 64

This type of impairment would account for the unrelated grammatical class words that were activated and the semantic errors. The NR/IDK responses would likely be due the inability of the semantic system to activate/retrieve appropriate words well. Th e small percentage of phonemic errors (6.12%) suggests partial/possible impairment in the phonological output lexicon as well (Table 4-3). A deficit in the visual object recognition system is unlikely for nouns or verbs in oral/written naming, considering the majo rity of the errors were either semantic or phonemic and P3s high scores on the Pyramids and Palm Trees (Howard & Patterson, 1992) and Kissing and Dancing Test (Bak & Hodges, 2003) (100% and 94.23%, respectively), which utilize the visual object recognition naming route. The predominant error for oral naming w ith verbs was NR/IDK responses, followed by semantic errors, then unrelated same grammatical class errors. Similar to nouns in this modality, these errors would likely be due to a deficit in the semantic system. P3s high accuracy with repetition of nouns (97.53%) and verbs (92.00%) (not reported in the results ), which utilizes the route of auditory phonological analysis phonological input buffer phonological input lexicon phonological output lexicon phonological output buffer that bypasses the semantic system, further supports the conclusion that the imp airment is primarily in the semantic system. During oral reading, the primary error for nouns was NR/IDK responses, followed by 50% of phonemes correctly produced, and unrelated grammatical class errors. Without thorough investigation, these errors would initially imply an impair ment at the phonological output lexicon to account for the phonemic errors and the partic ipants inability to co rrectly retrieve any phonemes (i.e., NR/IDK responses). A partial/possible impairment at the semantic system would also be probable, considering the number of unrelated grammatical class errors. 64

PAGE 65

The primary error for oral reading with ve rbs was unrelated grammatical class errors, followed by phonemic errors, and semantic errors. These errors would imply a larger impairment at the semantic level for verbs in this modality, secondary to the increased percentage of unrelated grammatical errors a nd the presence of semantic errors. A notable impairment at the phonological output lexicon is lik ely as well, when considering the percentage of phonemic errors. It is possible, however, that an alternative route ma y have been utilized by P3 for oral reading with nouns and verbs. Through previous testing of P3s ability to copy printed words using the same word list (not prev iously reported in th e results), P3 was ab le to copy nouns and verbs with 92.59 % and 95.92% accuracy, respectively. Therefore, it is possi ble that he used the route of abstract letter identification orthographic input lexicon phonological output lexicon phonological output buffer to read the printed words aloud. Considering P3s primary error type with oral naming of nouns and verbs was semantic (i.e., the semantic system was impaired), it is probable that the participant uti lized the above route to avoid use of the impaired semantic system. Since the majority of errors with nouns were phonemic rather than semantic, it is likely that the partic ipant utilized this route for nouns mo re than verbs (i.e., the semantic system was bypassed more when processing nouns). Use of this route to bypass the semantic system would explain the overall increased accuracy within the modality as well. It is also important to note th at a higher level of processing ex ists for verbs in oral reading. This can be seen through the high percentage of IDK/NR responses and reduced accuracy with nouns. This difference in processing between the two grammatical classes can be further explained using the Caramazza model (2002), whic h has separate locations for noun and verb 65

PAGE 66

processing at the semantic and phonological levels Further investiga tion of the model is required, however, before drawing a conclusion. Considering the above, the primary impairment for nouns and verbs in this modality is within the semantic system with additional involvement in the phonological output lexicon. Impairment in the orthographic input lexicon is unlikely, considering P3s accuracy with copying words. Use of the non-lexical route: ab stract letter identification letter-to-sound rules phonological output buffer is also unlikely cons idering P3s adequate copying ability and possible use of the above listed route that bypasses the semantic system. In addition, irregular words (e.g., cheese) were read al oud correctly by the pa rticipant, which would not be possible through the use of th e non-lexical route. For written naming of nouns, NR/IDK responses were the main error type, which was followed by grapheme errors, and jargon/indiscer nible responses. A small percentage of semantic and mixed errors were noted as well. Considering the high percentage of NR/IDK responses and grapheme errors, it is likely that th e orthographic output lexic on is impaired. It is also possible that partial/possible impairment in the semantic system and the pathway between it and the orthographic output lexi con is present, considering th e semantic and mixed errors. For written naming with verbs, NR/IDK respons es were also the main error type, which was followed by semantic errors, and then 50% graphemes correctly produced and unrelated same grammatical class errors (both at 9.52%). Considering the high percentage of NR/IDK responses as well as the semantic / unrelated same grammatical cla ss errors, it is likely that the semantic system is impaired for verbs in this modality. Orthographic output lexicon impairment would explain the percentage of 50% of graphemes correctly produced errors. 66

PAGE 67

The predominant error for writi ng to dictation with nouns was 50% of graphemes correctly produced, followed by jargon, and indis cernible responses. Similarly, the predominant error for writing to dictation with verbs was 50% of graphemes correctly produced, which was followed by jargon. Phonological input and outpu t lexicon impairment is unlikely considering P3s high accuracy with repetition (not reported in results) of nouns (97.53%) and verbs (92.00%) when using the same word lists. Without thorough investigation, these errors would initially suggest sole impairment at the orthographic output lexicon to expl ain the grapheme errors. It is important to consider, however, that the participant likely utilized th e route of auditory phonological analysis phonological input buffer phonological input lexicon phonological output lexicon orthographic output lexicon orthographic output buffer to avoid use of the impaired semantic system. Use of the non-lexical route: ac oustic-to-phonological conversion phonological output buffer sound-to-letter rules orthographic output buffer is unlikely when considering P3s repetition ability and possible use of the above list ed route that bypasses the semantic system. In addition, irregular words, such as knit, were written correctly by the participant, which would further prove that the non-lexical route was not utilized. P3s overall high accuracy with writing to dictation, the substantial percentage of graphemic errors as well as the absence of semantic errors when the semantic systems has already been established to be impaired, is further evidence that P3 bypassed the semantic system when writing to dictati on. In view of the above, the primary impairment for nouns and verbs would be in the semantic system, follo wed by partial impairment in the orthographic output lexicon for nouns and verbs writing to dictation. 67

PAGE 68

68 Overall for P3, it is apparent through error anal ysis that for both oral and written tasks the semantic system is primarily impaired. The primary semantic system impairment for nouns and verbs in the oral modalities has additional pa rtial phonological output le xicon involvement in oral reading and in oral naming with nouns. Th e semantic system is primarily impaired for nouns and verbs in the written modalities, with additional impairment of the orthographic output lexicon for written naming and partial impairment of the orthographic output lexicon for writing to dictation. A higher level of processing for verbs was present in oral reading; however, a pattern regarding noun and verb deficits within specific modalities (i.e., double dissociations) was not found for P3 after completing the error analyses.

PAGE 69

Table 4-1. Lexical Deficits Present in Participant 1 Based on Error Analysis Modality/Impairment Location Visual Object Recognition System Semantic System PIL POL Pathway Between Semantic and POL OIL OOL Pathway between Semantic and OOL N + + +/Oral Naming V + + N + + +/Oral Reading V + + N + + Written Naming V +/+ + N +/+/+ Writing to Dictation V +/+/+ Note: POL: Phonological Output Lexic on; OOL: Orthographic Output Lexicon; (+): Primary impairment present at that level; (-): No impairment present at that leve l; (+/-): Partial/Possible impairment at that level. 69 69

PAGE 70

Table 4-2. Lexical Deficits Present in Participant 2 Based on Error Analysis Modality/Impairment Location Visual Object Recognition System Semantic System PIL POL Pathway Between Semantic and POL OIL OOL Pathway between Semantic and OOL N +/+ Oral Naming V +/+ N +/+ Oral Reading V +/+ N +/+ Written Naming V + +/+/N +/+ + + Writing to Dictation V + + + Note: POL: Phonological Output Lexic on; OOL: Orthographic Output Lexicon; (+): Primary impairment present at that level; (-): No impairment present at that leve l; (+/-): Partial/Possible impairment at that level. 70 70

PAGE 71

71 Modality/Impairment Location Visual Object Recognition System Semantic System PIL POL Pathway Between Semantic and POL OIL OOL Pathway between Semantic and OOL N + +/Oral Naming V + N + +/Oral Reading V + +/N +/+ +/Written Naming V + + N + +/Writing to Dictation V + +/Note: POL: Phonological Output Lexic on; OOL: Orthographic Output Lexicon; (+): Primary impairment present at that level; (-): No impairment present at that leve l; (+/-): Partial/Possible impairment at that level. Table 4-3. Lexical Deficits Present in Participant 3 Based on Error Analysis 71

PAGE 72

CHAPTER 5 CONCLUSION No significant statistical di fference across nouns and verbs in oral naming, oral reading, written naming, and writing to dictation was noted for participant 1, 2, or 3. Similarly, no double dissociation across the above listed modalities was found in any of th e three participants. Trends regarding accuracy data, however, were noted for all three particip ants. P1 demonstrated greater accuracy within oral tasks as well as increased accuracy with nouns acro ss all modalities. P2 displayed increased accuracy in written tasks. Sh e also initially presented with a possible trend of noun advantage in written naming and verb adva ntage in writing to dictation; however, this apparent trend was proven false after completing error analyses. P3 showed decreased accuracy with oral and written naming tasks, with a tre nd toward increased accuracy with verbs in all modalities except for oral naming (i.e., noun and verb deficits were similar in oral naming). Based on error analyses, hypothesized levels of impairment within the linguistic system were identified. Participant 1 appeared to have impairment of the semantic system for oral tasks and primary impairment of the orthographic ou tput lexicon with partial semantic system involvement for written tasks. The semantic sy stem showed impairment for nouns and verbs in the oral modalities, with probable phonological ou tput lexicon involvement in oral reading of nouns and verbs and oral naming of nouns. C onversely, an orthographic output lexicon impairment, with partial semantic system invol vement, was present for nouns and verbs in the written modalities. P2 displayed a primary deficit at the lexeme (i.e., output lexicons) level for both oral and written modalities. The phonologica l output lexicon was mainly impaired for oral tasks, while the orthographic output lexicon was mainly impa ired for written tasks. The phonological output lexicon appeared impaired for nouns and verbs in the oral modalities, with partial semantic 72

PAGE 73

system deficits. Conversely, the orthographic ou tput lexicon appeared impaired for nouns and verbs in the written modalities, with partial semantic impairment in written naming. A higher level of processing was considered for both writte n modalities in P2. Nouns were noted to be processed at a higher level than verbs in written naming. In writing to dictation, however, it was determined that the greater percentage of accu racy with verbs was due to increased IDK/NR responses and use of an alternative non-lexical route. Therefore, a different level of processing in writing to dicta tion is less likely. P3 demonstrated overall impairment of the sema ntic system for oral a nd written tasks. The primary semantic system impairment for nouns and verbs in the oral modalities showed additional partial phonological output lexicon involvement in oral reading and in oral naming with nouns. The primary semantic system impa irment for nouns and verbs in the written modalities showed additional impairment in th e orthographic output lexicon for written naming and partial impairment in the ort hographic output lexicon for writi ng to dictation. A higher level of processing for verbs was establis hed in oral reading as well. Participant 3s, as well as P2s, apparent higher level of processing for verbs or nouns could not be explained by the adapted Ellis a nd Young model (1988; as cited in Kay, Lesser, & Coltheart, 1996). The Rapp and Caramazza model (2002), however, offers further explanation for the results. More testing is needed to conc lusively determine a higher level of processing and appropriate application of the Rapp and Caramazza model. Although not significant, the impairments inferred from the error analys is provide essential information about each participant. The impor tance of thorough testing regarding oral and written modalities is crucial because it offers valu able information of an individuals strengths. For example, P2 demonstrated high accuracy w ith written naming tasks even though her aphasia 73

PAGE 74

quotient on the Western Aphasia Battery (WAB; Kertesz, 1982) was low (AQ = 36.8). Important conclusions regarding noun/verb deficits, as well as, the presence/absence of a double dissociation can be made using error analyses as well. Although the stimuli from An Object and Action Naming Test (Druks & Masterson, 2000) is controlled for frequency, age of acquisition, and familiarity and it was further controlled for ambiguity in this study, other factors which affect lexical retrieval (e.g., manipulability, imageability, and word length) were not controlled for or examined. Therefore, those features cannot be ruled out as factors affecting the obser ved trends. Nonetheless, the current findings provide a thorough examination of the participants lexical retrie val of nouns and verbs across a variety of tasks, as well as, highlight the impor tance of error analyses in understanding lexical retrieval deficits and interpreti ng them within curren t theoretical models of language processing. 74

PAGE 75

LIST OF REFERENCES Arevalo, A., Moineau, S., Saygin, A., Ludy, C., & Bates, E. (2005). In search of noun-verb dissociations in aphasia acr oss three processing tasks. CRL Newsletter, 17 (1), 3-15. Arevalo, A., Perani, D., Cappa, S. F., Butler, A., Bates, E., & Dronkers N. (2007). Action and object processing in aphasia: From nouns a nd verbs to the effect of manipulability. Brain and Language, 100, 7994. Bak, T. H., & Hodges, J. R. (2003). Kissing and da ncing: A test to distinguish the lexical and conceptual contributions to noun/verb and action/object disso ciation. Preliminary results in patients with frontotemporal dementia. Journal of Neurolinguistics, 16 169-181. Berndt, R. S., Mitchum, C. C., Haendiges, A. N., Sandson, J. (1997a). Verb retrieval in aphasia: 1. Characterizing single word impairments. Brain and Language, 56, 68106. Berndt, R. S., Mitchum, C. C., Haendiges, A. N ., Sandson, J. (1997b). Verb retrieval in aphasia: 2. Relationship to sentence processing. Brain & Language, 56 107-137. Caramazza, A., & Hillis, A. E. (1991). Lexical organization of nouns and verbs in the brain. Nature, 349, 788790. Caramazza, A., & Miceli, G. (1991). Selective im pairment of thematic role assignment in sentence processing. Brain and Language, 41 402436. Chapey, R. (Ed.). (1994). Language intervention strategies in aphasia and related neurogenic communication disorders. Philadelphia, PA: Lippincott, Williams & Wilkins. Druks, J. (2002). Verbs and nouns: A review of the literature. Journal of Neurolinguistics, 15, 289-315. Druks, J., & Masterson, J. (2000). An object & action naming battery. Hove: Psychology Press. Ellis, A., & Young, A. (1988). Human cognitive neuropsychology. Hillsdale, NJ: Lawrence Erlbaum Associates Ltd. Goodglass, H., Kaplan, E., & Weintraub, S. (1983). Boston Naming Test. Philadelphia, PA: Lea and Febiger. Miceli, G., Silveri, M.C., Nocentin i, U., & Caramazza, A. (1988). Patterns of dissociations in comprehension and production of nouns and verbs. Aphasiology, 2, 351-358. Hillis, A. E. (2007). Aphasia: Progress in the last quarter of a century. Neurology, 69 200-213. Hillis, A. E., & Caramazza, A. (1991). Category-specific naming and comprehension impairment: A double dissociation. Brain, 114, 20812094. 75

PAGE 76

Hillis, A. E., & Caramazza, A. (1995). Representation of grammatical categories of words in the brain. Journal of Cognitive Neuroscience, 7 (3), 396407. Hillis, A. E., Tuffiash, E., & Caramazza, A. (2002). Modality-specific de terioration in naming verbs in nonfluent primary progressive aphasia. Journal of Cognitive Neuroscience, 14(7), 10991108. Howard, D., & Patterson, K. (1992). The pyramids and palm trees test. London: Harcourt Assessment. Jonkers, R., & Bastiaanse, R. ( 1998). How selective are selective word class deficits? Tow case studies of action and object naming. Aphasiology, 12(3), 245256. Kay, J., Lesser, R., & Coltheart, M. (1996). Psycholinguistic Assessments of Language Processing in Aphasia (PALPA): An introduction. Aphasiology, 10, 159-215. Kay, J., Lesser, R., & Coltheart, M. (1992). Psycholinguistic Assessments of Language Processing in Aphasia (PALPA) Hove: Psychology Press. Kertesz, A. (1982). Western aphasia battery Austin, TX: Pro-ed. Kucera, H., & Francis, W. N. ( 1967). Computational Analysis of Present-Day American English. Providence: Brown University Press. Levelt, W. J. M. (1989). Speaking: From inte ntion to articulation Cambridge, MA: MIT Press. Levelt, W. J. M. (1992). Accessing words in speech production: Stages, processes and representations. Cognition, 42 1-22. Odell, K. H., Hashi, M., Miller, S. B., McNeil, M. R. (1995). A critical look at the notion of selective impairment. Clinical Aphasiology, 23 1-8. Paulesu, E., Goldacre, B., Scifo, P., Cappa, S. F., G ilardi, M. C., Castiglioni I., Perani, D., Fazio, F. (1997). Neuroreport, 8 20112016. Perani, D., Cappa, S. F., Schnur, T., Tettamanti, M., Collina, S., Rosa, M. M., & Faziol, F. (1999). The neural correlates of noun-verb processing: a PET study. Brain, 122, 23372344. Rapp, B., & Caramazza, A. (2002). Selective difficu lties with spoken nouns and written verbs: A single case study. Journal of Neurolinguisitics, 15 373402. Shapiro, K, & Caramazza, A. (2002). The role and neural representation of grammatical class: A special issue of the Journal of Neurolinguistics, Journal of Neurolinguistics, 15 159170. 76

PAGE 77

Thompson, C. K. (2002). The northwestern verb production battery Unpublished. Zingeser, L. B., & Berndt, S. R. (1990). Retr ieval of nouns and verbs in agrammatism and anomia. Brain and Language, 39 14-32. 77

PAGE 78

BIOGRAPHICAL SKETCH Megan Elizabeth Ferguson is a native Floridian, who was born and raised in Winter Park, Florida. She earned a Bachelor of Arts degree in communication sciences and disorders from the University of Fl orida in December 2005. After graduation she worked as a mentored speech-language therapis t in a Central Florida middle school. She preformed individualized treatment and colla borative therapy with teachers for students with speech, language, and cognitive deficits. Megan later returned to the University of Florida and received a Master of Arts degree in communicati on sciences and disorders in May of 2008. Following her marriage to Richard Ashford Denman III in August of 2008, she will begin her career in the field of speech-language pathology. She will utilize her advanced education as a speech-language therapist in a medical setting. 78