<%BANNER%>

Evaluation of The Boston Naming Test Cueing Hierarchy

Permanent Link: http://ufdc.ufl.edu/UFE0024483/00001

Material Information

Title: Evaluation of The Boston Naming Test Cueing Hierarchy
Physical Description: 1 online resource (37 p.)
Language: english
Creator: Bislick, Lauren
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: access, aphasia, boston, cueing, phonemic, rasch, semantic
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: The Boston Naming Test (BNT; Kaplan, Goodglass, & Weintraub, 1983) is the most commonly used test to assess anomia in aphasia. However, the BNT was standardized on data from individuals with dementia, a neurodegenerative disorder, and not on individuals with aphasia. The semantic impairment researched in these two populations result in diverse language deficits. Thus, how can the BNT, which was only standardized on individuals with storage impairments, be an appropriate assessment battery for an access impaired population? The purpose of the current study is to further elucidate the effect of cueing on the naming abilities of individuals with aphasia. We raise two questions, 1) Can the hypothesized hierarchical structure of the rating scale be empirically validated, and 2) How well is the instrument able to differentiate individuals with aphasia? Item level analysis was applied to the BNT 4-point rating scale scoring system (incorrect, correct with phonemic cue, correct with semantic cue, spontaneously correct) using data from 100 individuals with aphasia. Results show semantic cues only facilitated word retrieval in less > 1% of individuals in this study. Based on the infrequent use of semantic cues we investigated an additional analysis with a 3-point collapsed scale (incorrect, correct with phonemic cue, spontaneously correct). Results indicate semantic cues may not be essential for individuals with aphasia, as the underlying mechanisms of word retrieval deficits are not a result of degraded semantic memory, but impairment in the ability to access word form.
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.
Statement of Responsibility: by Lauren Bislick.
Thesis: Thesis (M.A.)--University of Florida, 2009.
Local: Adviser: Kendall, Diane L.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0024483:00001

Permanent Link: http://ufdc.ufl.edu/UFE0024483/00001

Material Information

Title: Evaluation of The Boston Naming Test Cueing Hierarchy
Physical Description: 1 online resource (37 p.)
Language: english
Creator: Bislick, Lauren
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: access, aphasia, boston, cueing, phonemic, rasch, semantic
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: The Boston Naming Test (BNT; Kaplan, Goodglass, & Weintraub, 1983) is the most commonly used test to assess anomia in aphasia. However, the BNT was standardized on data from individuals with dementia, a neurodegenerative disorder, and not on individuals with aphasia. The semantic impairment researched in these two populations result in diverse language deficits. Thus, how can the BNT, which was only standardized on individuals with storage impairments, be an appropriate assessment battery for an access impaired population? The purpose of the current study is to further elucidate the effect of cueing on the naming abilities of individuals with aphasia. We raise two questions, 1) Can the hypothesized hierarchical structure of the rating scale be empirically validated, and 2) How well is the instrument able to differentiate individuals with aphasia? Item level analysis was applied to the BNT 4-point rating scale scoring system (incorrect, correct with phonemic cue, correct with semantic cue, spontaneously correct) using data from 100 individuals with aphasia. Results show semantic cues only facilitated word retrieval in less > 1% of individuals in this study. Based on the infrequent use of semantic cues we investigated an additional analysis with a 3-point collapsed scale (incorrect, correct with phonemic cue, spontaneously correct). Results indicate semantic cues may not be essential for individuals with aphasia, as the underlying mechanisms of word retrieval deficits are not a result of degraded semantic memory, but impairment in the ability to access word form.
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.
Statement of Responsibility: by Lauren Bislick.
Thesis: Thesis (M.A.)--University of Florida, 2009.
Local: Adviser: Kendall, Diane L.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0024483:00001


This item has the following downloads:


Full Text

PAGE 1

1 EVALUATION OF THE BOSTON NAMING TEST CUEING HIERARCHY By LAUREN BISLICK A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF ARTS UNIVERSITY OF FLORIDA 2009

PAGE 2

2 2009 Lauren Bislick

PAGE 3

3 To those that supported me in my endeavors ; my parents, sisters, boyfriend, and strong support network of peers and professors

PAGE 4

4 ACKNOWLEDGMENTS Ther e are so many people I have to thank who have helped in suppor ting me through the completion of my degree and thesis. I feel truly blessed to have received the many invaluable opportunit ies I have experienced at the University of Florida First off I hav e to thank my mentor and thesis committee for their time, guidance, and knowledge, Dr. Diane Kendall has inspired me as a clinician, researcher, and as a human being. Not only has she provided me with encouragement and installed confidence in my abilitie s as a speech language pathologist, but has bestowed upon me some of the greatest experiences in my career thus far. Dr. Craig Velozo has graciously opened his door to educate me about Rasch analysis. In my first attempts at understanding and applying an item level analysis Dr Velozo patiently provided me with the direction and education required for this project I must also thank Dr. Lisa Edmonds who first initiated my love for aphasia and aphasia research. Dr. Edmonds was the first person to introdu ce me to the concept of aphasia and through her remarkable teachings I developed a tremendous passion for this resilient population. Dr. E d monds has offered her guidance encouragement and numerous opportunities to me throughout my undergraduate and grad uate career. I thank my fellow students, friends and family who have supported my spirit through this project, for this I am greatly appreciative. My graduate and thesis experience would not have been as enjoyable without the companionship of Irene Hue bner, her friendship and leadership provided me with inspiration to make the most of the program and my thesis experience. I must also thank Sarah Mizrahi for all of her encouragement and understanding, and Christina del Toro for her friendship and positi ve attitude, taking the time to answer my many questions, and her

PAGE 5

5 help with Ras c h analysis and my submission to the 2008 American Speech and Hearing Association (ASHA) Conference. Finally, I thank my boyfriend Matthew Wilson who has been my day in and day out support during my graduate and thesis experience and my family especially my mom, whose love and encouragement have always inspired me to do my best

PAGE 6

6 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ ........... 4 LIST OF TABLES ................................ ................................ ................................ ...................... 8 LIST OF FIGURES ................................ ................................ ................................ .................... 9 ABSTRACT ................................ ................................ ................................ ............................. 10 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ ............. 12 Administration of the Boston Naming Test ................................ ................................ ........ 12 Underlying Nature of Anomia in Aphasia ................................ ................................ .......... 13 Anomia in Aphasia Versus Semantic Dementia ................................ ................................ .. 15 Purpose of Study ................................ ................................ ................................ ................ 18 2 METHODS AND MATERIALS ................................ ................................ ....................... 19 Participants ................................ ................................ ................................ ........................ 19 Procedures (data analysis) ................................ ................................ ................................ .. 19 3 RESULTS ................................ ................................ ................................ .......................... 21 4 DISCUSSION ................................ ................................ ................................ .................... 26 Question 1: Can the Hypothesized Hierarchica l Structure of the Rating Scale be Empirically Validated? ................................ ................................ ................................ ... 26 Question 2: How well is the Instrument Able to Differentiate Individuals with Aphasia? ... 28 APPENDIX A TABLES ................................ ................................ ................................ ............................ 30 A 1 Participant demographics ................................ ................................ ......................... 30 A 2 Summary of measured sets for 4 point Scale ................................ ........................... 30 A 3 Summary of measured sets for 3 point scale ................................ ............................ 30 B FIGURES ................................ ................................ ................................ .......................... 31 B 1 Item map of the 60 item BNT using the original 4 point rating scale ....................... 31 B 2 Four point rating scale category probabilities ................................ ........................... 32 B 3 Item map of the 60 item BNT using the 3 point rating scale ................................ .... 33 B 4 Three point rating scale category probabilities ................................ ......................... 34

PAGE 7

7 REFERENCES ................................ ................................ ................................ ......................... 35 BIOGRA PHICAL SKETCH ................................ ................................ ................................ ..... 37

PAGE 8

8 LIST OF TABLES Table page A 1 Participant demographics ................................ ................................ .............................. 30 A 2 Summary of measured sets for 4 point scale ................................ ................................ .. 30 A 3 Summary of measured sets for 3 point scale ................................ ................................ .. 30

PAGE 9

9 LIST OF FIGURES Figure page B 1 Item map of the 60 item Boston Naming Test (BNT) using the original 4 point rating scale ................................ ................................ ................................ .............................. 31 B 2 Four point rating scale category probabilities ................................ ................................ 32 B 3 Item map of the 60 item BNT using the 3 point rating scale ................................ .......... 33 B 4 Three point rating scale category probabilities ................................ ............................... 34

PAGE 10

10 Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of the Arts EVALUATION OF THE BOSTON NAMING TEST CUEING HIERA RHCY By Lauren P. Bislick May 2009 Chair: Diane Kendall Major: Communications Sciences and Disorders The Boston Naming Test (BNT; Kaplan, Goodglass, &Weintraub, 1983) is the most commonly used test to assess anomia in aphasia. However, the BNT was sta ndardized on data from individuals with dementia a neurodegenerative disorder, and not on individuals with aphasia. The semantic impairment researched in these two populations result in diverse language deficits. Thus, how can the BNT, which was only st andardized on individuals with storage impairments, be an appropriate assessment battery for an access impaired population? The purpose of the current study is to further elucidate the effect of cueing on the naming abilities of individuals with aphasia. We raise two questions, 1) Can the hypothesized hierarchical structure of the rating scale be empirically validated, and 2) How well is the instrument able to differentiate individuals with aphasia? Item level analysis was applied to the BNT 4 point ratin g scale scoring system (incorrect, correct with phonemic cue, correct with semantic cue, spontaneously correct ) using data from 100 individuals with aphasia. Results show s emantic cues only facilitated word retrieval in less >1% o f individuals in this stu dy. Based on the infrequent use of semantic cues we investigated a n additional analysis with a 3 point collapsed scale (incorrect, correct with phonemic cue, spontaneously correct) R esults indicate semantic cues may not be essential for individuals with aphasia as the underlying mechanisms of word

PAGE 11

11 retrieval deficits are not a res ult of degraded semantic memory, but impairment in the ability to access word form.

PAGE 12

12 CHAPTER 1 INTRODUCTION Each year 80,000 people acquire aphasia, a common language disorder occurring after left hemisphere cerebral vascular attacks (CVA); currently 1 million people in the U.S. a re living with aphasia (National Institute on Deafness and other Communication Disorders). It is frequently a chronic disorder requiring long term treatment that is rarely funded by insurance for the extent of the patient's needs. In order to effectively treat individuals with aphasia professionals require appropriate and effective assessment tools. The most frequ ent symptom of aphasia is anomia, a language impairment characterized by an inability to name objects, concepts, and people. Anomia can occur in response to stimulus presentation or in spontaneous speech. Naming tasks have the ability to reveal word retrieval or phonological encoding deficits beyond what is evident in conversational speech. Unlike other tasks used to assess language impairment s, the visual input of picture naming tasks do not provide sufficient sublexical cues to activate the phonology of the targeted word ( Dell, Schwartz, Martin, Saffran, & Gagnon, 1997) Currently, t he Boston Naming Test (BNT; Kaplan, Goodglass, &Weintraub, 1983), the focus of this study, is the most commonly used test to assess anomia in aphasia and utilizes picture naming to activate networks involved in lexical semantic retrieval The BNT consists of 60 black and white line drawings of objects ranging in frequency based on the occurrence of the corresponding word in print. Items are presented in order from easiest to hardest; that is most frequent to least frequent. Administration of the Boston Naming Test In the administration of the BNT, a cueing hier archy is implemented to help participants evoke the correct response to the picture item. If the participant is unable to spontaneously produce a correct response within a 20 second time frame he is provided with a stimulus or

PAGE 13

13 semantic cue to help relieve any misperception' the participant may have concerning the picture. For example, if the picture is of a bed but the participant does not produce the correct response or states that he does not know the pictured item a semantic cue, such as "a piece of f urniture", would be provided. If a semantic cue does not evoke the correct response the participant is provided with a phonemic cue (i.e. /b/ for bed ). Phonemic cues assist the individual by providing the opening/initial sound of the target word. The au thors believe this cue has potential to provide some evidence as to the existence of residual phonologic knowledge and the subsequent activation of the word in the lexicon. Lastly, if the participant is still unable to produce the correct response afte r both the phonemic cue and semantic cue are provided then it is assumed there is no ability to access the lexical item or no representation of the lexical item in the lexicon. In this instance, the response is marked incorrect The BNT provides a compos ite score. In particular the evaluator must combine the total number of items produce correctly without a cue and the number of items produced correctly following a stimulus cue. Participants do not receive credit for the number of items correct follow ing a phonemic cue. The cue is simply implemented to provide a possibility of success, relieve the individual of frustration or failure, and help differentiate among the different ability levels of the aphasic' patient. The BNT is particularly useful in the detection of relatively mild word retrieval impairments, such as tho se found in cases of dementia (Kaplan, Goodglass, Weintraub 1983). Underlying N ature of A nomia in A phasia Although this test is frequently used to assess anomia in aphasia it has s ome limitations. One limitation being that it was standardized using data from neurologically healthy individuals and individuals with dementia, a neurodegenerative disorder. Interestingly enough, despite the frequent use with individuals with aphasia, t he BNT was never standardized on data from individuals with aphasia. The degree to which this 60 item test can be applied to individuals

PAGE 14

14 with aphasia is questionable given the differences in underlying mechanism of anomia for these populations: impaired a ccess versus degradation of the semantic system (del Toro, Submitted ) Anomia in aphasia is described as a deficit of word retrieval, resulting from the temporary unavailability of stored representations in the brain. The process of word retrieval has bee n extensively modeled in aphasia (Dell 1986 ; Levelt, 1999 ; Garrett 1984 ; Tyler 2001 ) Dell et al. (1997) simply describes the process of picture naming as a transition of a visual stimulus into a conceptual representation, then into a lemma, which is t he retrieval of the name of the picture, and lastly into the phonological representation or the articulation of that name. In 1997 Dell et al., proposed a two step theory of lexical retrieval developed with information obtained from the results of non imp aired human error studies on picture naming. The first step, lemma access, starts with conceptual representations, which then leads to the activation of lemmas, the semantic and grammatical information pertaining to the targeted word. Step two, phonologi cal access, consists of mapping from the lemma stage to the phonological word form. Lemmas are referred to as the intermediate layer of this two step process, which allows for feedforward and feedback connections. The lexicon is described as consisting of three layers, a semantic, a word, and a phonemic layer. The semantic layer houses units that are connected to word nodes, which are connected to phoneme nodes. All connections are two way connections; at each level a representation of the message to be spoken is formed. Information can be spread from semantic to word to phoneme units, and from phoneme to word to semantic units. As semantic nodes decay phonemic units may activate and will then send activation back up to the word level. The most highl y activated word, that matches the appropriate syntactic category, will be chosen. Speech errors known as semantic paraphasias can occur at this stage of lexical retrieval This error of word meaning is characterized by a n involuntary substitution of a w ord that is

PAGE 15

15 semantically related to the target ed word ( Lambon Ralph, Moriarty, & Sage, 2002 ; Dell, Schwartz, Martin, Saffran, & Gagnon, 1997 ). A semantic cue can be provided to help assist in retrieval of the correct word from the lexicon by providing cue s to the meaning of the word The chosen word will then be put into syntactic frames with categorized slots that define acceptability. Next, meaning to form mapping begins. The most highly activated phoneme nodes are chosen and put into slots in a phono logical frame, which is characterized by word structure, such as consonant vowel consonant ( CVC ) pattern, stress pattern, and number of syllables. Speech errors occurring at the phonological level known as a phonemic paraphasia or sound error results fr om defected encoding for the articulation of the phonological representation of the lemma (Dell et al. 1997) P honemic cue s can be given to help simplify the selection process of the correct word form Semantic and phonemic paraphasias can also occur to gether on the same word; these errors are referred to as mixed paraphasias Anomia in A phasia v ersus S emantic D ementia As reported earlier, the BNT was standardized on individuals with semantic dementia. Taking into consideration the normal process o f lexical retrieval, a breakdown in this process is thought about differently depending upon the effected population or underlying neural mechanisms subserving word retrieval Researchers have intensely investigated the differences between language impair ments caused by stroke aphasia verses those caused by cortical degeneration ( Gotts & Plaut, 2002 ; Jefferies & Lambon Ralph, 2006; Nad uea & Rothi 200 0 ; Patterson, Nestor, & Rogers 2007 ; Warrington & Cipolotti, 1996; Warrtington & McCarthy, 1983). In both aphasia and semantic dementia semantic memory is impaired; however, these contrasting findings between the two populations have led some researchers to believe that there are two distinct forms of semantic impairment, one resulting from damage to semantic access processes and the other resulting from damage directly to semantic representations, respectively

PAGE 16

16 (Gotts &Plaut, 2002). Impai rments in naming ability for aphasia occur from deregulated semantic control and can result in an inability to successfully recover the phonological descriptions of words from their lemmas, and/or lemmas from their semantic descriptions (Nadeau et al., 200 0). In contrast, storage impairments resulting from cortical degeneration, such as semantic dementia, are defined as a stable loss of word knowledge caused by the degradation of semantic representations (Warrington & Cipolotti, 1996). T h is continuous d eterioration found in dementia typically affects proce ssing across selective domains, which may result in d ysfunction of the semantic system or disconnection of the semantic system from the other language processing mec hanisms ( Hodges, Patterson, Oxbury, & Funnell 1992 ). Semantic dementia is not only a language deficit, but includes a continuous loss of semantic memory, which corresponds to general knowledge of objects, word meanings, facts, and people (Patterson Nestor, & Rogers, 2007). It allows us to comprehend a multitude of different stimuli and express knowledg e in a wide variety of domains (Jefferies & Lambon Ralth, 2006). Individuals with semantic dementia show increasing impairments of word retrieval, a uditory and reading comprehension, a nd nonverbal semantic processing. However, abilities such as phonology and syntax remain intact, unlike those found in access impaired individuals. In their 1996 experimental paper Warrington and Cipolotti investigate d characteristics that have been shown to distinguish between access and storage deficits, such as presentation rate, consistency, frequency, and semantic relatedness. Comparative results obtained from individuals with aphasia and individuals with Pick's disease, a neurodegenerative disorder, showed a significant relationship between access disorders and the effects of presentation rate and semantic relatedness. Storage disorders, however, were found to be affected by frequency and yield consistent result s, regardless of modality. In a similar study, storage disorders were

PAGE 17

17 associated with strong frequency effects, highly consistent performance, and no impact of cueing, whereas access deficits were found to produce no frequency effects, inconsistent respon ses, and strong effects of cueing (Jefferies and Lambon Ralph, 2006). Jefferies and Lambon Ralph (2006) propose that semantic cognition is supported by two interacting principal components: 1) a set of amodal representations, which progressively degrade i n semantic dementia, and 2) a set of executive processes that help to direct and control semantic activation in a task appropriate fashion, which are affected in stroke aphasic patients. These authors support the refractory hypothesis to explain why indiv iduals with aphasia have difficulty accessing semantic knowledge. Cipolotti and Warrington (1995) define refractoriness as temporary or even sporadic impairment, which delays the return, following activation, of a set of representations that underpin a wo rd, to a ready' state. In other words, language deficits resulting from aphasia are not the result of lost knowledge like those found in semantic dementia, but a temporary delay of access. The authors found individuals with semantic dementia to be high ly consistent when the same items were probed by different tasks, reinforcing the claim that storage impairments in semantic dementia result from degraded amodal semantic representations. However, this consistency may not be found in access impaired indiv iduals as their impairment is caused by a loss of semantic control due to impaired executive processes, rather than a loss of core amodal semantic knowledge. An access deficit is believed to be multimodal because all tasks, regardless of modality, require semantic control. Thus, if the semantic impairment for these two populations result in such diverse language deficits how can the BNT, which was only standardized on individuals with a storage impairment, be an appropriate assessment battery for an access impaired population?

PAGE 18

18 Purpose of Study The purpose of the current study is to further elucidate the effect of cueing on the naming abilities of individuals with aphasia. We raise two questions, 1. Can the hypothesized hierarchical structure of the rating scale be empirically validated ? 2. How well is the test able to differentiate individuals with aphasia? We predict that the hypothesi zed hierarchical structure does not confirm the actual hierarchy because of the population of which the BNT was originally standardized We believe that semantic cues are not effective in helping to produce a desired result, unlike ph onemic cues. Therefore, the hypothesized hierarchy (incorrect, phonemic cue, semantic cue, spontaneously correct) may not be supported by our data, which was collected from individuals with aphasia. We predict that the cueing hierarchy will be able to differentiate individuals with aphasia in terms of seve re, for those that produce incorrect responses when provided with cues less severe /moderate for those that can produce the correct response when given a phonemic cue and mild/normal, for those that can spontaneously produce a correct response.

PAGE 19

19 CHAPTER 2 METHODS AND MATERIAL S Participants Archival data were used for this study from the Veteran Affairs (VA) Brain Rehabilitation Research Center (BRRC) in Gainesville, F lorida. Individuals who suffered a left hemisphere cerebral vascular accident (CVA) were seen at the BRRC for evaluation to determine eligibility for rehabilita tion studies (IRB # 457 1999). The comprehensive BRRC evaluation included assessment by a beha vioral neurologist, physical therapist and speech language pathologist. The speech and language evaluation included administration of the Western Aphasia B attery (WAB; Kertesz, 1982), BNT non standardized tests of agrammatism, aprosodia, alexia and agrap hia. All data were entered into an Access Database. For the present study, archival data from the BRRC was used to examine performance of 100 individuals diagnosed with aphasia and 30 neurologically healthy adults on the BNT (Kaplan, Goodglass, & Weintr aub, 1983). All individuals with aphasia (6 3 males, 37 females) had a left hemisphere cerebral vascular accident (CVA) at least 6 months prior to examination. A diagnosis of aphasia (WAB AQ <94.7) was based on the Western Aphasia Battery (WAB; Kertesz, 1 982). All 130 participants were right handed, monolingual, English speakers. Characteristics for the individuals with aphasia are displayed in T able A 1. Procedures (da ta analysis ) All participants were given the full 60 item BNT by a licensed speech la nguage pathologist following original directions for administration and scoring, including use of a semantic cue following 20 seconds of an incorrect or no response, then a phonemic cue after another 20 seconds of incorrect or no response. Responses from 30 neurologically healthy controls and 100 individuals with aphasia were analyzed with the WINSTEPS Rasch analysis

PAGE 20

20 computer program (Bond & Fox, 2001; Linacre, 2004, 1994). Linacre (2002) recommends implementing eight essential criteria to optimize the ra ting scale category effectiveness, three of the eight essential criteria were implemented in the present study: 1) frequency of occurrence of each rating must be greater than 10 observations each, 2) average measures advance monotonically with category, an d 3) acceptable outfit mean squares was defined as < to 2.0. Two separate analyses were conducted: 1) 4 point rating scale: 0 = incorrect, 1 = correct with phonemic cue, 2 = correct with semantic cue, 3 = spontaneously correct and 2) 3 point rating sca le, 0 = incorrect, correct with phonologic cue, 3 = spontaneously correct

PAGE 21

21 CHAPTER 3 RESULTS Overview of Rash Analysis 1A: Original Rating Scale Aphasia Group Unidimensionality Unidimensionality is determined by item fit statistics of Rasch model resid uals. Two items misfit based on Infit Mean Square (!1.4) ( abacus, bed ) and one item misfit based on Outfit Mean Square (! 2.0) ranges ( beaver ). Item d ifficulty The range of person ability & item difficulty is an indication of the difficulty of the test. The range of item difficulty and person ability for this form was 3 to 3 logits ( F igure B 1). Participant ability and test difficulty One participant misfit according to I nfit M ean S quare criteria and three participants misfit according to O utfit M ean Square. As shown in Figure B 1, the mean difficulty of the test and the mean ability of the participants are equal. Item h ierarchy Figure B 1 displays the items in the order of easiest to hardest (bottom to top) using the responses from our sample. Th e original order of presentation is represented by the numbers adjacent to each item. Overall, t he items generally maintained their hierarchy according to frequency (Figure B 1. shows the item s ranging in frequency, item 1bed is the most frequent item foll owed by items 4 house, 19 pretzel, 23 volcano, 38 harp 51 latch and 60 abacus the least frequent item ). Visual analysis of the hierarchy reveals several inconsistencies between the proposed hierarchy of lexical frequency and the resulting hierarchy using the responses of our

PAGE 22

22 individual's with aphasia (ex. Items 19 and 24 are placed higher/more difficult in the hierarchy than item 32 and 35). Floor/ c eiling e ffects and i tem r edundancy Analysis was completed on 92 of 100 partic ipants with aphasia. The Rasc h program excluded eight participants because total scores were zero or one, resulting in ins ufficient data for the analysis There were no ceiling effects. So me item redundancy was observed (ex. Seahorse, hammock, stethoscope, muzzle, and unicorn in Fig ure 1). Separation, r atio, and i nternal c onsistency Person separation is the measure of the spread of persons on the variable using a ratio of the person standard deviation and the average measurement error. Based on the real RMSE, the person separation w as 5.34. Person separation is converted to strata, a measure of how many ability levels people are separated into by the test. The calculated strata was 7.45, indicating the 60 item BNT can divide people into approximately 8 ability levels. Person and i tem reliability index are each .97. Essential c riteria Results from the 4 point scale do not meet all three of t he following essential criteria. Criteria 1: F requency of occurrence of each rating must be greater than 10 observations each. The observed count presents the number of categories in each rating, i.e. rating "0", incorrect, has a total 2013 categories or occurrences throughout the 60 item BNT for all (92) individuals. When a category frequency is low the step calibration is ina ccurate and pos sibly unstable. Rating "2", correct with semantic cue, does not have 10 observations in its category (occurred below 10% of the time). In fact, semantic cues occurred least frequently of all ratings and encompassed only 1% of responses (Table A 2 and Fig ure B 3). These findings are also displayed in the category probability rating scale (Figure B 3). The probability curves

PAGE 23

23 act as a close graphic depiction of rating scale observations; rating "2" never moves above 0 and therefore is insignificant. Crit eria 2: A verage measures advance monotonically with category. The observed average measure of a rating scale should increase from negative to positive; in other words higher measure combinations produce observations in higher categori es and lower measure combinations produce observations in lower categories. The average measure by category advance s monotonically up the rating scale from least difficult to most difficult, i.e. category "0", incorrect, corresponds to observe d average measure 1.07, but category "3", spontaneously correct, is the highest measure and corresponds to 1.02 (Table A 2). This same principal is also observed in the category probability rating scale (Figure B 3). As measures increase, or as individu als with incrementally higher measures are observed, each category of the scale is designed to most likely be chosen. The model specifies that respondents with infinitely low (or high) measures must be observed in the lowest (or highest) categories (Linca re, 2002). For example, for a person of relatively low ability (left), the highest probability rating is a "0" which is incorrect with all cues; the next most probable rating is a "1" which is correct with phonemic cues and the least probable response is "3" which is spontaneously correct or "2" which is correct with semantic cues Criteria 3: A cceptable outfit mean squares was defined as < to 2.0 O utfit m ean s quares is the observed variance over the expected variance; it is ideally a ratio of 1 and must be below 2.0. Values greater than 2.0 suggest that there is more unexplained randomness or noise than explained noise, which indicates that there is more misinformation than information in the observation (Linacre 2002). O utfit M ean S quares for this rating scale are all below 2.0, indicating the categories have been used in expected contexts.

PAGE 24

24 1B: New Condensed Rating Scale (3 point scale) Based on the significantly infrequent occurrence of semantic cues, an additional analysis was conducted on a 3 point s cale: incorrect, correct with phonemic cue, spontaneously correct. Rating "2", semantic cues, of the 4 point scale only occurred in 1% of responses. Linacre (2002) suggests rarely observed categories be combined with adjacent categories or their data be om itted Therefore, rating "2" was combined with an adjacent category, rating "1", phonemic cues, to obtain a regular frequency distribution in the resulting 3 point scale (Table A 2) All remaining analyses are based on the 3 p oin t rating scale. In the 3 point scale all three essential criteria were met (Table A 3): 1) frequency of occurrence is greater than 10 observations in each category, 2) observed average measures advanced monotonically with category, and 3) O utfit Mean Squares are less than 2.0. Upon the removal of rating "2" no significant changes in the category probability rating scale occurred (Figure B 4). Unidimensionality The revised 3 point scale showed tha t no items misfit according to I nfit Mean Square statistics and only one item misfit based on O utfit Mean Square statistics Item d ifficulty No significant change in item difficulty was observed (Figure B 3) Participant ability and item difficulty T he mean difficulty of the test and mean ability of the participants remained the same. Item h ierarchy, f loor/ c eiling e ffects and i tem r edundancy Item hierarchy, floor/ceiling effects and item redundancy were unchanged with the new 3 point scale (Figure B 3).

PAGE 25

25 Separation, ratio, and internal c onsistency Based on the real RMSE, the person separation is 5.49, resulting with strata of 7.65. Both the original and condensed rating scales are able to divide people into approximately 8 ability levels. Results for person and item reliability were unchanged.

PAGE 26

26 CHAPTER 4 DISCUSSION The purpose of this study was to apply item level analysis on the 4 point rating scale and investigate the effects of cu e ing on the naming a bilities of 100 individuals with aphasia. Our results show 1) the hypothesized hierarchical structure of the rating scale is not empirically validated and 2) the test is sensit ive enough to differentiate individuals with aphasia into eight ability levels Question 1: Can the H ypothesized H ierarchical Structure of the Rating S cale be E mpirically V alidated ? In this investigation the hypothesized hierarchical structure of the rati ng s cale is not empirically validated According to the 4 point rating scale semantic cues are insignificant; frequency of occurrence is less than 1% (Table A 2 and Figure B 2) This finding is not surprising however, taking into account the underlying nature of word retrieval impairment in individuals with aphasia. While evidence for naming impairment in semantic dementia is largely due to a deficit in semantic knowledge, the naming impairment in aphasia is attributed to inadequate access to the word form. Thus, the participant ma intains the conceptual knowledge but has an impaired ability to access the word without additional information such as those given in cues (Jefferies & Lambon Ralph, 2006). Accordingly, it makes sense that in this group of aphasics semantic cues are not sufficient to activate the stimulus item because semantic knowledge is intact; it is semantic control that is impaired, resulting in an inability to consistently retrieve the correct representation of the word. Category Probability In the creation of th e 3 point scale ratings "2" and "1" are combined to obtain regular distribution of responses, thus optimizing the rating scale category effectiveness However, no significant change is observed in the category probabilit y rating scale upon the removal of rating

PAGE 27

27 "2" (Figure B 4), thus illustrating that semantic cues are inconsequential for this population R esults from the 3 point rating scale lend further support for the idea that semantic cues ar e not necessary or informative with administration of this test to patients with aphasia. Supporting Studies In their 1996 paper Warrington and Cipolotti presented data from four patients with Pick's disease and two patients with global aphasia, which co nfirmed differences in the semantic deficits of these patients based on storage and access. Six years later Je fferies and Lambon Ralph (2006) further investigated the differential effects of access and storage deficits in patients with aphasia and semanti c dementia, respectively. Jefferies and Lambon Ralph (2006) employed a case s eries design and compared aphasia patients to semantic dementia patients on a common batte ry of both verbal and n on verbal semantic tests, including naming tasks. The effect of phonemic cues on picture naming was assessed using t he Boston Naming Test. Participants were asked to name the 60 test items and were provided with a phonemic cue for pictures they could not name. Results showed that the effect of cueing on picture namin g did not allow individuals with semantic dementia to produce object names that they could not recall spontaneously; however, all of the stroke aphasi a patients showed significant improvement with phonemic cueing. Phonemic cues positively influence the n aming abilities of individuals with aphasia because phonemic cues allow for external support, which reduces the need of self generated semantic control which is impaired in individuals with aphasia, to complete the task. In contrast, the n aming abilities of individuals with semantic dementia will not be influenced by phonemic cues because the semantic representation of the word has deteriorated.

PAGE 28

28 Question 2: How well is the Instrument Able to Differentiate Individuals with A phasia? Our results show both th e original hierarchy and condensed rating scales are dividing people into nearly 8 ability levels. This number is much higher than what we predicted of the BNT This finding informs us that the test is extremely sensitive. It is not surprising that both r ating scales are dividing into approximately the same number of strata when considering semantic cues were only used by 1% of the participants; thus combining rating "2" with rating "1" did not affect the sensitivity of the test proving the new 3 point sc ale to be just as sensitive and more appropriate for an access impaired population. Clinical Implications : What D o These Three Points Tell U s? The new condensed 3 point scale provides the evaluator with information about the aphasic individual s word retr ieval ability. If the individual is able to spontaneously provide the correct name to the presented picture it is assumed adequate access is available for word retrieval If the individual requires a phonemic cue to correctly produce the targeted word a ccess to the word appears to be available suggesting that ph onological representations are preserved and semantic abilities are intact (Robson, Marshall, Pring & Chiat, 1998) However access is weakened as a result of reduce d semantic control and assis tance is required for successful word retrieval Phonemic c ues help to reduce the need of self generated semantic control during word retrieval tasks by providing external support (Jefferies & Lambon Ralph, 2006) This deficit in semantic control can res ult in an inability to successfully access the phonological word form from its semantic descr ip tions. If the participant is unable to produce the correct response after a phonemic cue is provided it is assumed the deficit in semantic control is too great an impairment and even with the help of phonemic cues access is insufficient and unable to retrieve the lexical item

PAGE 29

29 The information collected through administration of the 3 point scale may be helpful to the clinician in the construction of treatments. A positive response to phonological cues suggests that therapy should work at the phonological level to improve word retrieval abilities (Robson et al. 1998) In their single case investigation of phonological naming therapy Robson et al. (1998) found a significant relationship between successful affects of phonemic cueing on picture naming tasks and therapy targeted at the phonological level for patient G.F Results show that phonological therapy caused significant improvements in naming, which gene ralized to untreated items and were even maintained at follow up. The results of G.F.'s phonological therapy are consistent with the theoretical interpretation of her naming deficit. Based on the successful result of phonemic cues it is hypothesized tha t access to the lexical item is available, although impaired, and requires assistance to reduce the need of self generated semantic control.

PAGE 30

30 APPENDIX A TABLES Table A 1. Participant d emographics Total 100 Male 63 Female 37 Averages (Std. Dev) Ag e (SD) 62.9 (+/ 12.5) Education (SD) 13.5 (+/ 3.1 ) Months post onset 42.9 (+/ 51.3) Western Aphasia Battery (AQ) 65.7 (+/ 25.2) Anomic 38 Broca's 26 Conduction 13 Unspecified 9 Global 6 Wernicke's 4 TCM/TCS 4 Boston Naming Test 26.1 (+ \ 1 8.5); range = 0 59 Table A 2. Summary of m easured sets for 4 point scale Rating Observed Count (%) Observed Average Infit Mean Square Outfit Mean Square 0 (incorrect) 2013 (36) 1.07 .99 .96 1 (correct with phonemic cue) 940 (17) .11 .98 .76 2 (correct with semantic cue) 55 (1) .33 .90 .49 3 (spontaneously correct) 2512 (46) 1.02 1.08 1.31 Table A 3. Summary of m easured sets for 3 point scale Rating Observed Count (%) Observed Average Infit Mean Square Outfit Mean Squar e 0 (incorrect) 2013 (36) 1.48 1.10 1.13 1 (correct with phonemic cue) 995 (18) .05 .89 .69 3 (spontaneously correct) 2512 (46) 1.66 1.03 1.20

PAGE 31

31 APPENDIX B FIGURES More able persons/Difficult items 3 + | | X | | | T| | 2 + | XXX | X | |T 58PALETTE 59PROTRACTOR 60ABACUS XX | XXX | XXXXX S| 55SPHYNX 56YOKE 1 XXX + XXXXX | 50COMPASS XXXXX |S 47ACCORDION 51LATCH 57TRELLIS XXXXXX | 37ESCALATOR 43PYRAMID 54TONGS XXXXXXXX | 24SEAHORSE 39HA MMOCK 42STETHOSCOPE 44MUZZLE 45UNICORN 49ASPARAGUS 52TRIPOD 53SCROLL | 41PELICAN 48NOOSE XXXXX | 29BEAVER 31RHINOCEROS 40KNOCKER XXXXXX | 27GLOBE 30HARMON ICA 34STILTS 35DOMINOES 0 XX M+M 13OCTOPUS 21RACQUET 28WREATH 33IGLOO 36CACTUS XXX | 19PRETZEL 23VOLCANO 38HARP XXXX | 11HELICOPTER 14MUSHROOM 16WHEELCHAIR 26CANOE 46FUNNEL XXXX | 18MASK 22SNAIL 25DART 32ACORN XX | 10TOOTHBRUSH 15HANGER XX | XXXX |S 17CAMEL 20BENCH 6SCISSORS XXX | 12BRROM 1 X S+ X | 3PENC IL 5WHISTLE 8FLOWER | 2TREE 7COMB 9SAW X | 4HOUSE XXX |T X | | 1BED XXXXX | 2 + X T| X | | | | | | 3 + | | | X | | | | 4 XXXXXXXX + Less able pe rsons/Easier item Figure B 1. Item m ap of the 60 item BNT using the original 4 point rating scale

PAGE 32

32 CATEGORY PROBABILITIES: MODES Step measures at intersections P ++ -------------+ -------------+ -------------+ -------------++ R 1.0 + 3+ O | 333333333 | B |000 3333 | A | 000000 333 | B .8 + 0000 33 + I | 000 3 | L | 00 33 | I | 000 3 | T .6 + 00 33 + Y | 00 3 | .5 + 0 3 + O | 00 3 | F .4 + 00 33 + | | R | 111111111*1** | E | 11111 33 **1 | S .2 + 11111 33 0*111 + P | 111111 33 00 111 | O |1111 333 000 11111 | N | 3333333 000000111111111 | S .0 +*************22222222222222222222222222222222222*************+ E ++ -------------+ -------------+ -------------+ ------------++ 2 1 0 1 2 PERSON [MINUS] ITEM MEASURE Figure B 2. Four point rating scale category probabilities For a person of relatively low ability (left), the highest pro bability rating is a "0" which is incorrect with all cues; the next most probable rating is a "1" which is correct with phonemic cues and the least probable response is "3" which is spontaneously correct or 2 (* means ratings 2 and 3 have equal probability at this point) wish is correct with semantic cues According to this chart semantic cues "2" are not used and therefore are not significant. For a person with high ability (right), the highest probability rating is a "3" spontaneously correct, next most probable is "1" correct with phonemic cue, and the least probable response is "0" incorrect with cues or "2" (* means ratings 0 and 2 have equal probability at this point)

PAGE 33

33 More able persons/Difficult items 5 + | X | | | | 4 + | | T| | | 3 XX + X | X | X | | 60ABACUS XXXX |T 58PALETTE 59PROTRACTOR 2 XX + XXX S| XX | 55SPHYNX XXX | 56YOKE XXXXXX | 50COMPASS 51LATCH XXX |S 47ACCORDION 57TRELLIS 1 XXXXXX + 3 7ESCALATOR 42STETHOSCOPE 43PYRAMID 54TONGS XXXXXX | 24SEAHORSE 39HAMMOCK 49ASPARAGUS 52TRIPOD 53SCROLL X | 41PELICAN 44MUZZLE 45UNICORN 48NOOSE XXXXX | 40KNOCKER XXXX | 29BEAVER 31RHINOCEROS 34STILTS XXX M | 27GLOBE 28WREATH 30HARMONICA 35DOMINOES 0 XX +M 21RACQUET 33IGLOO 36CACTUS XXX | 13OCTOPUS 23VOLCANO 38HARP XXXXX | 11HELICO PTER 14MUSHROOM 16WHEELCHAIR 19PRETZEL 26CANOE XXXXX | 22SNAIL 46FUNNEL | 10TOOTHBRUSH 15HANGER 18MASK 25DART 32ACORN XX | 1 XX + 17CAMEL 20BENCH XX |S 6SCISSORS XX S| 12BRROM XX | X | 3PENCIL 5WHISTLE 8FLOWER | 2 XX + 2TREE 4HOUSE 7COMB 9SAW XX |T X | X X | XX | 1BED | 3 X T+ | | X | | | 4 X + | | | | 5 XXXXX XXX + Less able persons/Easier items Figure B 3.Item m ap of the 60 item BNT sing the 3 point rating scale

PAGE 34

34 CATEGORY PROBABILITIES: MODES Step measures at intersections P ++ -------------+ -------------+ -------------+ -------------++ R 1.0 + + O | | B | | A |00000 33333| B .8 + 0000 3333 + I | 0000 3333 | L | 00 33 | I | 000 333 | T .6 + 00 33 + Y | 000 333 | .5 + 00 33 + O | 00 33 | F .4 + 00 33 + | 0*3 | R | 1 1111111**1**11111111 | E | 11111 333 000 11111 | S .2 + 111111 333 000 111111 + P | 111111 333 000 111111 | O |111 33333 00000 111| N | 3333333333 0000000000 | S .0 +3333 0000+ E ++ -------------+ -------------+ ------------+ -------------++ 2 1 0 1 2 PERSON [MINUS] ITEM MEASURE Figure B 4. Three point r ating s cale c ategory p robabilities This second rating scale (3 p t rating scale) shows that for a person of relatively low ability, the highest probability rating is a "0" which is incorrect with all cues; the next most probable rating is a "1" which is correct with phonemic cues and the least probable response is "3" which is spontan eously correct. Correct with semantic cues "2" no longer exist alone in this new 3 pt rating scale; it has been collapsed into the phonemic cues rating because use of semantic cues was not significant (1% < 10 categories).

PAGE 35

35 REFERENCES American Speech Language Hearing Association. Communication Facts: Special Populations: Stroke 2004 Edition. Ci polotti, L.,Warrington, E.K. ( 1995 ) Towards a unitary account of access dysphagia: a single case study. Memory, 3 309 332. Dell, G.S. (1986). A spreading activation theory of retrieval in sentence production. Psychological Review, 93 (3), 283 321 Dell G.S., Schwartz, M.F., Martin, N., Saffran, E M., & Gagnon, D.A. (1997). Lexical access in aphasic and nonaphasic speakers. Psychological Review, 104 (4), 801 838 del Toro, C.M., Bislick., L.P., Comer, M., Velozo, C., Gonzalez Rothi, L.J., & Kendall., D.L. (2009). Creation of a Short Form Boston Naming Test for Individuals with Aphasia Submitted Garrett, M.F. (1984). The organisation of processing structure for language production: Applications to aphasic speech. In Caplan, D., Leco urs, A.R., and Smith, A. (Eds.), Biological Perspectives on Language. Cambridge, MA: MIT Press. Gotts, S.J. & Plaut, D.C. ( 2002 ) The impact of synaptic depression following brain damage: A connectionist account of "access/refractory" and "degraded store" semantic impairments. Cognitive, Affective, & Behavioral Neuroscience, 2 (3), 187 213. Hodges, J.R., Patterson, K., Oxbury, S., & Funnell, E. (1992). Semantic dementia progressive fluent aphasia with temporal lobe atrophy. Brain : A Journal of Neurology, 115 ( Pt 6) (Pt 6), 1783 1806. Jefferies, E. & Lambon Ralph, M.A. ( 2006 ) Semantic impairment in stroke aphasia versus semantic dementia: a case series comparison. Brain, 129, 2132 2147. Kaplan, E., Goodglass, H., & Weintraub, S. ( 2001 ) Boston naming test Philadelphia, PA: Lea and Febiger. Kertesz, A. ( 1982 ) Western Aphasia Battery. New York: Grune & Stratton. Lambon Ralph M.A. Moriarty L., & Sage, K. ( 2002 ) Anomia is simply a reflection of semantic and phonological impairments: Evidence from a case series study. Aphasiology, 16 (1/2), 56 82. Levelt W.J.M. ( 1999 ) Models of word production. Trends in Cognitive Sciences, 3 (6), 223 232. Linacre, J.M. ( 2002 ) Optimizing rating sale category effectiveness. Journal of applied measurement, 3 (1), 85 106

PAGE 36

36 Nadeau, S.E., Gonzalez Rothi, L.J., Crosson B. ( 2000 ) Aphasia and language : theory to practice New York : Guilford Press National Institute on Deafness and Other Communication Disorders. Aphasia. . Patterson, K., Nestor, P.J., & Rogers T.T. ( 2007 ) Where you know what you know? The representation of semantic knowledge in the human brain. Nature Reviews Neur oscience, 8 976 987. Robson, J., Marshall, J., Pring, T., & Chiat, S. ( 1998 ) Phonological naming therapy in jargon aphasia: Positive but paradoxical effects Journal of the International Neuropsychological Society 4 675 686. Tyler, L.K. & Moss, H.E. ( 2001 ) Towards a distributed account of conceptual knowledge. Trends in Cognitive Sciences, 5 (6), 224 252. Warrington, E.K. & Cipolotti, L. ( 1996 ) Word comprehension: The distinction between refractory and storage impairments. Brain, 119, 611 625. Warr ington, E.K. & McCarthy R. ( 1983 ) Category specific access dysphasia. Brain, 106 859 878.

PAGE 37

37 BIOGRAPHICAL SKETCH Lauren Bislick graduated summa cum laude and received a Bachelor of the Arts in c ommunications s ciences and d isorders from the University of F lorida in May 2007. During her undergraduate experience Lauren completed an undergraduate honors thesis Selective Grammatical Deficits in Aphasia under the mentorship of Dr. Lisa Edmonds In May 2009 she will graduate from the University of Florida w ith a Master of Arts in s peech l anguage p athology The m aster's program at the University of Florida has provided Lauren with numerous opportunities to enrich her development including research opportunities at the Brain Rehabilitation and Research Cente r (BRRC), a poster presentation at the 2008 Amercian Speech and Hearing Association (ASHA) Convention, medical practicum placements at both the Gainesville and Lake City Veteran Affairs Medical Center (VAMC) BRRC, and Shands Rehab, and the most significan t to her graduate experience, the duties of Administrative Coordinator of the first Gainesville Stroke of Hope Stroke Support Group. Lauren has received an invitation to continue her education at the University of Washington in Seattle, where she plans to pursue a Doctor of Philosophy degree focusing on adult neurological disorders under the mentorship of Dr. Diane Kendall Lauren's interests include treatment and diagnostic test development for individuals with aphasia.