<%BANNER%>

The Effect of Computerized Verb Network Strengthening Treatment on Lexical Retrieval in Aphasia

Permanent Link: http://ufdc.ufl.edu/UFE0045090/00001

Material Information

Title: The Effect of Computerized Verb Network Strengthening Treatment on Lexical Retrieval in Aphasia
Physical Description: 1 online resource (85 p.)
Language: english
Creator: Furnas, Daniel Wade
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2012

Subjects

Subjects / Keywords: aphasia -- computer -- language -- semantics -- stroke -- teletherapy -- therapy -- treatment
Speech, Language and Hearing Sciences -- 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: Aphasia, an acquired disorder that affects both production and comprehension of language, is a common sequella following stroke. There have been many computerized treatments developed for chronic aphasia, including some with promising results, but few of these treatments have been implemented over the Internet. The current study sought to create a computerized form of the previously established Verb Network Strengthening Treatment (VNeST)(Edmonds, Nadeau, & Kiran, 2009). This Computerized VNeST (VNeST-C) was then used to perform telerehabilitation over the Internet. This study was conducted in three stages. In the first stage, neurotypical young adults were trained on the VNeST-C program and reported on its usability. In the second stage, neurotypical older adults also tested the program and reported on its usability. During the third stage, in which three people with aphasia were treated using the VNeST-C program, a single-subject, multiple baseline design was used. The treatment materials were originally created for previous VNeST studies and adapted for use on the computer. Treatment was performed using Adobe® Connect™ software to facilitate communication between the researcher and participants. The results of this study suggest that two of the three participants showed modest gains as a result of the VNeST-C treatment. The third participant showed minimal improvement as a result of the treatment, but his lack of improvement may be a result of several complicating factors. The overall results of this study show that the VNeST-C program and protocol have promise and further investigation into the effects of the treatment are warranted.
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 Daniel Wade Furnas.
Thesis: Thesis (M.A.)--University of Florida, 2012.
Local: Adviser: Edmonds, Lisa Anna.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2014-12-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2012
System ID: UFE0045090:00001

Permanent Link: http://ufdc.ufl.edu/UFE0045090/00001

Material Information

Title: The Effect of Computerized Verb Network Strengthening Treatment on Lexical Retrieval in Aphasia
Physical Description: 1 online resource (85 p.)
Language: english
Creator: Furnas, Daniel Wade
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2012

Subjects

Subjects / Keywords: aphasia -- computer -- language -- semantics -- stroke -- teletherapy -- therapy -- treatment
Speech, Language and Hearing Sciences -- 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: Aphasia, an acquired disorder that affects both production and comprehension of language, is a common sequella following stroke. There have been many computerized treatments developed for chronic aphasia, including some with promising results, but few of these treatments have been implemented over the Internet. The current study sought to create a computerized form of the previously established Verb Network Strengthening Treatment (VNeST)(Edmonds, Nadeau, & Kiran, 2009). This Computerized VNeST (VNeST-C) was then used to perform telerehabilitation over the Internet. This study was conducted in three stages. In the first stage, neurotypical young adults were trained on the VNeST-C program and reported on its usability. In the second stage, neurotypical older adults also tested the program and reported on its usability. During the third stage, in which three people with aphasia were treated using the VNeST-C program, a single-subject, multiple baseline design was used. The treatment materials were originally created for previous VNeST studies and adapted for use on the computer. Treatment was performed using Adobe® Connect™ software to facilitate communication between the researcher and participants. The results of this study suggest that two of the three participants showed modest gains as a result of the VNeST-C treatment. The third participant showed minimal improvement as a result of the treatment, but his lack of improvement may be a result of several complicating factors. The overall results of this study show that the VNeST-C program and protocol have promise and further investigation into the effects of the treatment are warranted.
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 Daniel Wade Furnas.
Thesis: Thesis (M.A.)--University of Florida, 2012.
Local: Adviser: Edmonds, Lisa Anna.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2014-12-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2012
System ID: UFE0045090:00001


This item has the following downloads:


Full Text

PAGE 1

1 THE EFFECT OF COMPUTERIZED VERB NETWORK STRENGTHENING TREATMENT ON LEXICAL RETRIEVAL IN APHASIA By DANIEL FURNAS 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 2012

PAGE 2

2 2012 Daniel Furnas

PAGE 3

3 To my sister Emily who is always there to remind me that there is more than one way to tackle a problem.

PAGE 4

4 ACKNOWLEDGMENTS First and foremost, I thank my advisor Dr. Lisa Edmonds for support and guidance throughout the research process, without whose assistance, this project would never have even started. I thank all the members of the UF Aphasia and Bilinguali sm Lab for support and feedback. In particular, I thank Sachit Mishra for assistance in developing the VNeST C computer program His insights and programming skills made the VNeST C program what it is today. Additionally, I thank Angela LaGambina for assistance in pr eparing the data for this study, without whom there would not be results to be discussed and Victoria Lastinger who assisted me with the testing of the older adult patients Finally, I am deeply grateful to Iomi Patten for all her insight into the thesis process and her support and advice along the way.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS .................................................................................................. 4 LIST OF TABLES ............................................................................................................ 8 LIST OF FIGURE S .......................................................................................................... 9 LIST OF ABBREVIATIONS ........................................................................................... 10 ABSTRACT ................................................................................................................... 11 CHAPTER 1 BACKGROUND AND INTRODUCTION ................................................................. 13 Aphasia ................................................................................................................... 13 Verb Network Strengt hening Treatment .................................................................. 14 In Clinic Computerized Aphasia Therapy ................................................................ 17 Computerized Teletherapy ...................................................................................... 19 The Present Study .................................................................................................. 21 Research Q uestions ............................................................................................... 24 Research Question 1 ........................................................................................ 24 Research Question 2 ........................................................................................ 24 Research Question 3 ........................................................................................ 24 Research Question 4 ........................................................................................ 25 Research Question 5 ........................................................................................ 25 2 PROCEDURE ......................................................................................................... 27 Participants ............................................................................................................. 27 Younger Adults ................................................................................................. 27 Older Adults ...................................................................................................... 27 Treatment Participants ..................................................................................... 28 Study Design .......................................................................................................... 30 Stimuli development ............................................................................................... 31 Sentence Probe Stimuli .................................................................................... 31 Treatm ent Stimuli ............................................................................................. 31 VNeST C Computer Program ........................................................................... 31 Computer Questionnaire .................................................................................. 32 Procedure ............................................................................................................... 33 Young and Older Adults ................................................................................... 33 Treatm ent Participants ..................................................................................... 33 Pre treatment testing ................................................................................. 33 Baseline testing .......................................................................................... 34 Treatment ................................................................................................... 36

PAGE 6

6 Step 1: agent verb patient triad creation .................................................... 37 Step 2: whquestions ................................................................................ 39 Step 3: semantic judgments ....................................................................... 40 Step 4: recall and review ............................................................................ 40 Treatment Probes ............................................................................................. 41 Post treatment Testing ..................................................................................... 41 Treatment Reliability ......................................................................................... 42 Scoring Reliability ............................................................................................. 42 3 RESULTS ............................................................................................................... 43 Younger and Older Adults ....................................................................................... 43 Research Question 1 ........................................................................................ 43 Research Question 2 ........................................................................................ 43 Participant 1 ............................................................................................................ 44 Research Question 3 ........................................................................................ 44 Research Question 4 ........................................................................................ 44 Research Question 5 ........................................................................................ 45 Participant 2 ............................................................................................................ 46 Research Question 3 ........................................................................................ 46 Research Question 4 ........................................................................................ 46 Research Question 5 ........................................................................................ 46 Participant 3 ............................................................................................................ 48 Research Question 3 ........................................................................................ 48 Research Question 4 ........................................................................................ 48 Research Question 5 ........................................................................................ 48 Figures and Tables ................................................................................................. 50 4 DISCUSSION ......................................................................................................... 64 Discussion of Research Questions ......................................................................... 64 Research Question 1 ........................................................................................ 64 Research Question 2 ........................................................................................ 65 Research Question 3 ........................................................................................ 65 Research Question 4 ........................................................................................ 66 Research Question 5 ........................................................................................ 66 General Discussion ................................................................................................. 68 Future Directions of Study ...................................................................................... 72 APPENDIX A EXAMPLE PICTURES FROM THE SENTENCE PROBE TASK ............................ 73 B PICTURES OF THE VNeST C COMPUTER PROGRAM INTERFACE AND OUTPUT ................................................................................................................. 75 C COMPUTER QUESTIONNAIRE ............................................................................. 78

PAGE 7

7 D LIST OF SELECTED WORDS FROM THE JOHNS HOPKINS DYSLEXIA AND DYSGRAPHIA BATTERY ....................................................................................... 80 LIST OF REFERENCES ............................................................................................... 81 BIOGRA PHICAL SKETCH ............................................................................................ 85

PAGE 8

8 LIST OF TABLES Table page 3 1 P1s spoken agent, verb and patient production during sentence probes. ......... 59 3 2 P1s typed agent, verb and patient production during sentence probes. ............ 59 3 3 List of results from Participant 1s preand post treatment testing scores. ........ 60 3 4 P2s spoken agent, verb and patient production during sentence probes. ......... 60 3 5 P2s typed agent, verb and patient production during sentence probes. ............ 61 3 6 Li st of results from Participant 2s preand post treatment testing scores. ........ 61 3 7 P3s spoken agent, verb and patient production during sentence probes. ......... 62 3 8 P3s typed agent, verb and patient production during sentence probes. ............ 62 3 9 List of results from Participant 3s preand post treatment testing scores. ........ 63

PAGE 9

9 LIST OF FIGURES Figure page 3 1 P1s cue distribution for both A) agents and B) patients during treatment. ......... 50 3 2 P1s typing errors of target A) agents and B) patients during treatment. ............ 51 3 3 Participant 1s spoken and typed sentence probe results. .................................. 52 3 4 Participant 1 s results on a control task PALPA Subtest 13. ............................ 52 3 5 P2s cue distribution for A) agents and B) patients during treatment. ................. 53 3 6 P2s errors of target A) agents and B) patients during treatment. ....................... 54 3 7 Participant 2s spoken and types sentence probe results. .................................. 55 3 8 Participant 2s results on a control task PALPA Subtest 8. .............................. 55 3 9 P3s cue distribution for A) agents and B) patients during treatment. ................. 56 3 10 P3s errors of target A) agents and B) patients during treatment ........................ 57 3 11 Participant 3s spoken and types sentence probe results. .................................. 58 3 12 Participant 3s results on a control task PALPA Subtest 13. ............................ 58 A 1 The picture for The mailman is delivering the letter. ......................................... 73 A 2 The picture for The fireman is holding the cat. ................................................. 74 A 3 The picture for The snake is biting the hand. ................................................... 74 B 1 The triad creation step of VNeST C. ................................................................... 75 B 2 The Whquestion step of VNeST C. .................................................................. 75 B 3 The semantic judgement step of VNeST C. ....................................................... 76 B 4 The review step of VNeST C. ............................................................................. 76 B 5 Sample output data from the Triad, Whquestion and Recall steps ................... 77 B 6 Sample output data from the semantic judgment step ........................................ 77

PAGE 10

10 LIST OF ABBREVIATIONS AHA American Heart Association AOS Apraxia of speech AQ Aphasia Quotient CETI Communication Effectiveness Index CLQT Cognitive Linguistic Quick Test CVA Cerebrov ascular accident; stroke NAVS Northwestern Assessment of Verbs and Sentences O&A An Object & Action Naming Battery PALPA Psycholinguistic Assessment of Language Processing in Aphasia PICA Porch Index of Communicative Abilities TEA Test of Everyday Attention Tx Treatment VNeST Verb Network Strengthening Treatment VNeST C Computerized Verb Network Strengthening Treatment WAB Western Aphasia Battery WTD Writing to dictation

PAGE 11

11 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 Arts THE EFFECT OF COMPUTERIZED VERB NETWORK STRENGTHENING TREATMENT ON LEXICAL RETRIEVAL IN APHASIA By Daniel Furnas December 2012 Chair: Lisa A. Edmonds Major: Comm unication Sciences and Disorders Aphasia, an acquired disorder that affects both production and comprehension of language, is a common sequella following stroke. There have been many computerized treatments developed for chronic aphasia, including some wi th promising results, but few of these treatments have been implemented over the Internet. The current study sought to create a computerized form of the previously established Verb Network Strengthening Treatment (VNeST ) (Edmonds, Nadeau, & Kiran, 2009) T his Computerized VNeST (VNeST C) was then used to perform telerehabilitation over the Internet. This study was conducted in three stages. In the first stage, neurotypical young adults were trained on the VNeST C program and reported on its usability. In the second stage, neurotypical older adults also tested the program and reported on its usability. During the third stage, in which three people with aphasia were treated using the VNeST C program, a singlesubject, multiple baseline design was used. The tre atment materials were originally created for previous VNeST studies and adapted for use on the computer. Treatment was performed using Adobe Connect software to facilitate communication between the researcher and participants.

PAGE 12

12 The results of this study suggest that two of the three participants showed modest gains as a result of the VNeST C treatment. The third participant showed minimal improvement as a result of the treatment but his lack of improvement may be a result of several complicating factors. The overall results of this study show that the VNeST C program and protocol have promise and further investigation into the effects of the treatment are warranted.

PAGE 13

13 CHAPTER 1 BACKGROUND AND INTRODUCTION A phas ia Each year, approximately 795,000 people suffer a stroke in the United States, 185,000 of which are recurrent attacks (AHA, 2011) One of the most common sequella following a cerebrovascular accident (CVA) is aphasia, occurring in 27% of all patients who incur some form of brain injury (Duffy, 2005). While some people with CVAs may recover spontaneously from aphasia, 19% of ischemic stroke survivors still had aphasia 6 months post onset. (AHA, 2011). Aphasia is an acquired disorder that limits a person's expression and comprehension of language. For many people with aphasia, treatment of this condition is a long term proposition, with effects from aphasia continuing after the expi ration of outpatient treatment options. Aphasia management, often already limited by the availability of insurance and reimbursement options, can be further complicated by a lack of access to local treatment facilities (Theodoros Hill, Russell, Ward, & Wo otton 200 8 ). Some people in these situations are able to receive home healthcare, however the frequency and duration of these sessions may not be sufficient, and these services are not readily available everywhere. Considering the fact that intensive, long term aphasia rehabilitation has been correlated to increased quality of life after a stroke (Sarno, 1997), having a variety of rehabilitation options is in the best interest of the patient. Advancements in computer and Internet based aphasia therapy all ow for increased access for people with chronic aphasia to receive therapy in situations that they might not otherwise have. As a result, people with aphasia with inadequate access

PAGE 14

14 to therapy for transportation or logistical reasons now have options to rec eive therapy without leaving their home. While internet based teletherapy is still a relatively young field in speechlanguage pathology, there have been some studies with promising results, and as more research is compiled, more avenues of treatment will become available. Verb Network Strengthening Treatment Verb Network Strengthening Treatment (VNeST) is a semantic treatment originally designed as a faceto face clinical program for improving lexical retrieval in people with aphasia. The VNeST protocol s hares a theoretical basis with the verb as core treatment system explored by Loverso and colleagues ( 1982). Loverso and colleagues were able to show that verbcentered semantic therapy has potential to improve lexical retrieval in people with aphasia. Ac ross a number of studies, Loverso and colleagues found improvement in the ability to independently generate agents and patients given a verb (Loverso, Prescott, Selinger, Wheeler, & Smith, 1982; Loverso, Prescott, & Selinger, 1988) as well as improvement on the Porch Index of Communicative Abilities (Prescott, Selinger, & Loverso, 1982). Thus, their treatment approach showed promise, but their investigation to generalization was limited to trained items and a comprehensive aphasia battery. VNeST builds up on the idea of verbcentered treatment but expands on the production and comprehension required from Loversos approach with the goal of increasing lexical retrieval to a hierarchy of lexical retrieval contexts. VNeST targets the semantic associations betw een agent, verb and patient in order to systematically improve lexical access. During subsequent studies, Loverso and colleagues found improvement on ability to independently generate agents and patients given a verb

PAGE 15

15 (Loverso, et al., 1982) as well as impr ovement on the Porch Index of Communicative Abilities (Prescott, et al 1982). The interrelatedness of the semantic concepts of an agent (who performs an action), verb and patient (on which an action is performed) has been investigated in semantic priming studies. A number of these studies indicate priming between a verb and related nouns serving as various thematic roles (e.g., agent, patient and instrument) (Ferretti, McRae, & Hatherell 2001). Further study by McRae and colleagues found that these priming results may be bidirectional, allowing for priming from noun to verb as well as verb to noun. The implications of these studies is that through experience and knowledge event schemas o f verbs and related thematic roles (e.g., A teacher erases a chalkboard in a classroom with an eraser) are developed and coactivated neutrally, implying a network of semantic connectivity. Based on these theoretical underpinnings, VNeST provides the par ticipant with a verb and asks them to provide appropriate agents and patients. Some of these agent verb patient triads produced are then expanded upon to increase the semantic salience and breadth of the association by requiring the participant to provide when, where and why the action would take place. Additionally as part of the protocol, participants are asked to make semantic judgments about a variety of syntactically balanced and correct sentences (e.g., A pilot flies a car.). A more detailed explanati on of the traditional VNeST protocol may be found in Edmonds & Babb (2011). In the initial VNeST study (Edmonds, Nadeau, & Kiran, 200 9 ) the four participants with moderate aphasia showed marked improvement in the ability to generate specific agents, verbs and patients during a picturebased sentence elicitation

PAGE 16

16 task. The pictures developed for and used in this study targeted the semantic ally related pairs of verbs, one of which was treated during the study and the other remained untreated (e.g., measure and weigh). The improvement on this task is indicative of generalization, as the pictures themselves were not used during the course of t herapy. The participants only had access to the pictures seen during testing. Additionally, the participants showed indications of improvement in untrained singleword naming and sentence elicitation tasks, and three of the four participants significantly improved on the number of complete utterances (Subject Verb Object) they produced in a discourse task. In the subsequent study (Edmonds & Babb, 2011), two participants with more severe aphasia were enrolled. The testing protocol was modified from the ori ginal study for one of the participants to allow for writing of responses (after attempts at spoken production) during treatment due to a preponderance of neologistic responses. While these participants did not show as large an improvement as the initial s tudy, due t o more severe aphasia, there were still statistically significant improvements in sentence production of trained and untrained pictures and singleword naming as measured by the Boston Naming Test (Kaplan, Goodglass, & Weintraub, 1983) and An Object and Action Naming Batter y ( Druks & Masterson, 2000). The participant whose protocol was modified to allow for writing showed improvements in both verbal and written production. While the treatment did not directly target written production, the increase seen in the written modality is likely due to generalization of improved lexical access as a result of the semantic treatment since the semantic system supports both written and spoken production.

PAGE 17

17 In Clinic Computerized Aphasia Therapy Most of the stu dies using computer based treatment have been done inclinic and still require travel by the participant to receive therapy. Computer based treatment of aphasia first started with Seron and colleagues (1980) who used a set of typing tasks to improve the spelling of 50 trained words in 5 people with aphasia. Four of the participants were 16 months or more post onset, while one was only 3 months post onset. The task required a clinician to dictate a word, then the participant typed it with either the visual c ue of a box for each letter or onto a single box for the whole word. The results of this study showed mild gains on trained items with some of the participants. While Seron focused on typing production, other researchers have sought to use computers to adv ance rehabilitation in other ways. Computerized reading therapies are one of the most frequently explored avenues of rehabilitation (Katz & Nagy, 1982; Katz & Wertz, 1997; Katz, 2010). In each of the studies, the participants had limited researcher interv ention. The researcher introduced the procedure for the task and remained onhand to answer questions as necessary, but the participant primarily operated the program independently. In the original study, Katz and Nagy (1982) employed a series of computeri zed reading tasks with no specific treatment objective, achieving limited effects on reading comprehension. The Katz & Wertz's (1997) study used new software with a mixture of matching and reading comprehension activities, designed specifically for the stu dy. Patients with mild to moderate aphasia again had limited interaction with the clinician, but did manage some significant improvement on the Porch Index of Communicative Ability (PICA : Porch, 1981). The phase 2 study by Katz and Wertz, employing the sam e methods as the previous study again showed gains on the PICA and Western Aphasia Battery

PAGE 18

18 (Kertesz, 1982) as a result of reading stimulation therapy in people with chronic aphasia (Katz, 2010). Other research has taken advantage of the computers abil ity to store and play media files. One such early study was a case study on a person with chronic aphasia. The treatment course covered 13 months of listening therapy on the computer. In part one of the treatment, the participant is given a set of pictures on the screen and then the program plays an audio file with the directions for an auditory comprehension task. In this study, significant gains were made on subtests VI and X of the PICA, which test auditory comprehension (Mills, 1982). In a study of the Computer assisted Anomia Rehabilitation Program (CARP 2), Adrian and colleagues used a multicue naming program to treat naming of 200 picture stimuli. As a result of the treatment, a group analysis of 15 people with chronic aphasia showed statistically sig nificant improvement in their ability to perform the treated singleword naming task after treatment. The study also showed improvement in similar naming tasks of untrained stimuli but it did not investigate generalization outside of singleword picture naming. As a result, it is unclear if CARP 2 resulted in a notable improvement in daily functional communication (Adrian Gonzalez, Buiza, & Sage, 2011). A computer therapy program entitled AphasiaMate has been investigated for improvement of language ski lls in people with aphasia. The AphasiaMate program has eight different facets of treatment: auditory processing, visual matching, reading comprehension, spelling, semantics, sentence processing, time and money. Eight people with aphasia were treated with the intervention program, which was not clinicianguided. Rather, the participants were able to run the program at their own speed. After

PAGE 19

19 15 weeks of treatment, the participants showed mixed results on the Western Aphasia Battery's Aphasia Quotient (WAB AQ ) and the Communicative Effectiveness Index (CETI), a questionnaire which participants indicate their communicative abilities (relative to premorbid abilities) on 16 functional communication scenarios. While six participants showed improvement on the post treatment measures, two participants showed decreases in their WAB AQ. On the CETI, the aggregate score of 6 participants improved while one stayed the same and one decreased. Based on the increase seen in CETI scores, the participants expressed that their functional communication abilities had improved; however, functional communication abiltiies were not directly tested in the study (Archibald, Orange, & Jamieson, 2009). While all of these treatment methods use the computer as a method of treatment deli very, they still require the participant to travel to the clinic. More recently, computer based aphasia therapy has been combined with teletherapy in part due to improved security protocols over the Internet. Computerized Teletherapy As Internet speed and availability have increased, so too has the viability of providing remote treatment. Using secure videoconferencing software, in conjunction with high speed Internet, speechlanguage pathologists can provide services for people who could not otherwise rec eive them due to accessibility issues. Remote therapy alleviates difficulties with transportation and also reduces secondary costs associated with receiving healthcare such as public transportation fares, gasoline costs or leave from work for caregivers who may have to drive patients to the clinic. Initially, video based telerehabilitation was explored for the purpose of performing evaluations remotely. Duffy and colleagues (1997) successfully performed oral motor,

PAGE 20

20 motor speech and language evaluations usi ng real time audio and visual feeds. Evaluations performed remotely were found to be as accurate as those performed faceto face with the patients. Subsequently, Brennan and colleagues (2002) investigated whether there were significant differences in the performance of patients with brain injury during remote evaluations as compared to faceto face evaluations. Again, their study found no significant difference in accuracy of evaluation or participant performance between faceto face and teletherapy evaluations. More recently, remote evaluation of patients with aphasia was investigated using the short form of the Boston Diagnostic Aphasia Examination (Goodglass, Kaplan, & Barresi, 2000) and Boston Naming Test (Kaplan, et al ., 1983) In this study, it was fo und that standardized assessment tools could be used to accurately evaluate people with aphasia over the Internet. These studies all proved that evaluations could be performed remotely on different groups of people with a variety of conditions. While the a bility to conduct evaluations remotely is important, evaluations are only one part of the service provided by speechlanguage pathologists. For many patients, therapy makes up a much larger portion of their visits, so simply evaluating patients online does not eliminate access and transportation problems. While research into online treatment has been relatively limited as compared to evaluations, the results have been generally positive. A case study done by Kully (2000) showed that remote, follow up treat ment for people who stutter was an effective means of maintaining their gains from therapy. However, in this study, initial treatment was first provided faceto face. Theodoros and colleagues (2006) successfully implemented a complete course of Lee Silverm an Voice Therapy to participants with Parkinson s

PAGE 21

21 disease over the Internet. A comparison of pretreatment and post treatment measures found that participants showed statistically significant improvement in pitch range and sound pressure level during sustained phonation, reading and conversational speech. Aphasia management via telerehabilitation is a relatively new yet potentially promising method of treatment delivery. One study of a singleword multimodality naming treatment conducted over the Internet found some gains after 12 1hour sessions, and participants in the study indicated a high degree of satisfaction ( Dechene, Tousignant, Boissy, Macoir, Heroux, Hamel, Briere, & Page, 2011). Another study, using a 2week course of Constraint Induced Language Therapy, found gains on the WAB and an increased number of propositions in the telling of the Cinderella Story. The study used an intensive course of therapy 4 days per week for 3 hours per day, all provided remotely ( Maher Kendall, Swearengin, Rodriguez, Leon, Pingel, Holland, & Rothi, 2006). These st udies show that there is promise in telerehabilitation for people with aphasia. However, the fact remains that remote aphasia management is an under investigated field, and more work is necessary to determine the best methods of remote treatment delivery. The Present Study The current study shares some aspects of both Phase I and II treatment studies (Robey, 2004), as it takes a preexisting treatment and changes the format to accommodate for computer based delivery. The singlesubject design study explor es whether a computerized version of the VNeST protocol (VNeST C) still promotes improvement in lexical access for people with aphasia. With advancement of personal computer, tablet and cellular phone technology, the ability to accurately and efficiently t ype has become a matter of daily life. Now, communication frequently occurs via the

PAGE 22

22 Internet in the form of email, shopping, banking, or social networking websites, and information is frequently gathered via Internet searches. These tasks all require some proficiency and ability to type. In the present study, three participants were enrolled in VNeST modified for presentation via computer. The Computerized Verb Network Strengthening Treatment (VNeST C) protocol maintains the basic theoretical foundation of the previous VNeST studies with the addition of a typing modality and a more complex semantic sentence judgment task. Given the fact that Edmonds & Babb (2011) integrated written production during their previous study with success, the addition of the typed modality is a natural extension of the treatment. Based on the improvements on both spoken and written responses seen in Edmonds & Babb (2011), a combination of spoken and typed VNeST treatment has a sound theoretical basis for an improvement. In Edmond s & Babbs study, VNeST was administered only using verbal responses; however, one participant showed significant improvement in both spoken and written modalities Based on the promise seen in the previous studies results, this study seeks to investigate whether the addition of a typing modality will act as an adjuvant to further increase production in modalities other than speech. While the basic tenets of the treatment have been previously studied, the addition of a second modality effectively renders t his a pilot study in its own right. Even with the addition of the typing modality, VNeST C continues to target the same lexical semantic systems as previous VNeST studies. As previously stated, the VNeST system of treatment seeks to strengthen the semantic connections between associated words. However, in order for a person to express these semantic concepts,

PAGE 23

23 they must then be expressed either through the phonological or orthographic lexicon and output buffers (which are both activated via a shared semantic system) (e.g., Kay, Lesser, & Coltheart, 1992) T he dominant pathway for organizing the spelling for familiar words is the lexical semantic pathway ( Rapcsak & Beeson, 2002). This is the specific access system that VNeST targets during treatment. Previous VNeST studies, which only trained participants using verbal responses primarily targeted the path that traces from the semantic system through the phonological output lexicon and buffer. However, since the same basic lexical semantic connections are being made for both spoken and written responses, Edmonds and Babb were able to show some improvement in writing without targeted treatment (2011). Expanding on the typing output, the model of orthographic output put forth by Cameron, Cubelli and Della Sala (2002) suggests that while handwriting and typing have separate final outputs (a graphememotor buffer for writing and a letter detection buffer for typing), they share the same orthographic and allographic buffers. As such, trained improvement in typing coul d also lead to additional improvement in writing, but not necessarily accuracy of writing specific motor movement since the graphememotor system is not being directly exercised and as such implicit motor learning is not being experienced. Improvements seen in written output during a previous VNeST study were theoretically due to improved lexical access resulting from semantic treatment and a course of treatment that targets writing may have further improved the graphememotor buffer outputs and strengt hened written response accuracy. Based on these theories and previous results this study seeks to improve the accuracy of lexical retrieval and subsequent output through both the phonological and orthographic output systems through the combination of spok en and typed treatment

PAGE 24

24 Research Questions Research Q uestions 1 and 2 of this study were performed with neurotypical older and younger adults prior to the start of treatment. These questions were investigated during the piloting and testing stages of crea ting the VNeST C computer program. Research Question 1 Does the VNeST C computer program demonstrate high usability as measured by responses from neurotypical young adults? Usability will be measured via a questionnaire given to participants after being t rained on the VNeST C program. The questionnaire allows the participant to rate usability in one of five categories: no difficulty, little difficulty, mild difficulty, moderate difficulty, and high difficulty. The criterion for usability is met when ninet y percent of participants indicate little or no difficulty using the program following a simulated treatment session. Research Question 2 Does the VNeST C computer program demonstrate high usability as measured by responses from neurotypical older adults? As with younger adults, usability will be measured via the same questionnaire given to participants after being trained on the VNeST C program. The criterion for usability is met when ninety percent of participants indicate little or no difficulty using t he program. Research Question 3 Does treatment with VNeST C result in improved lexical access during treatment sessions as measured by level of cueing required?

PAGE 25

25 This will be measured by evaluating the level of cueing (e.g., independent, minimum or maximum ) required to elicit the first two nontarget agent and patient pairs. Target agent and patient pairs, those tested during the sentence probe task, are excluded from this task as they are required during each session and, as a result, are frequently cued t o ensure inclusion in treatment (see Method section for details). Responses were grouped into cycles where each of the 10 trained verbs w as practiced once. Research Question 4 Does treatment with VNeST C result in improved typing accuracy of agents and patients during treatment sessions? T yp ed responses for non target agent s and patient s for a given verb will be analyzed for number of typographical errors. Responses will be grouped into correct attempts, single grapheme errors and multiple grapheme errors. Research Question 5 Does treatment with VNeST C generalize to increased lexical retrieval as measured by a hierarchical series of production tasks? As this study is only directly providing treatment within the confines of the VNeST C computer program, any gains made to other clinical measures or functional communication would require generalization. This would include generalization to single word naming, sentencelevel production, and discourse. Based on previous VNeST studies (Edmonds, et al., 2009; Edmonds & Babb, 2011), we predict some improvement on picture description using sentences with trained and untrained verbs. This is due to the fact that this treatment seeks to improve the independent production of agents and patients based around a series of verbs. Should this treatment result in

PAGE 26

26 improved lexical access, the participants should show improvements in the ability to produce the content words required for picture description. Additionally, we will compare pre treatment and post treatment singl e word naming, discourse samples and standardized test scores for further indication of improvement.

PAGE 27

27 CHAPTER 2 PROCEDURE Participants Two groups of participants were recruited to test the usability of the VNeST C program and refine the interface. Five cognitively normal younger adults and five cognitively normal older adults were recruited. After the completion of program testing, four persons with aphasia were enrolled in this study for treatment purposes. All participants completed appropriate informed consent paperwork from the University of Florida prior to testing. Younger Adults Five younger adults with a mean age of 26.8 years (range = 20 to 32 years) were enrolled to beta test the VNeST C computer program. All of the younger adult participants wer e monolingual English speakers with no history of learning disability, neurological disorder or traumatic brain injury as per self report. They had a mean of 16 years of education with a range of 14 to 17 years. These participants were recruited via flyers located around the University of Florida campus. Older Adults Subsequent to the younger adult testing, three older adults, mean age of 64.3 years (range = 56 to 71 years) were enrolled to beta test the VNeST C computer program. All of the older adult part icipants were monolingual English speakers with no history of learning disability, neurological disorder or traumatic brain injury as per self report. These participants have a mean education of 17.3 years with a range of 15 to 21 years. These participants were recruited via flyers located around the University of

PAGE 28

28 Florida campus and the use of a participant recruitment service at the University of Florida. One participant was consented for this study but her data was not used due to signs of mild cognitive impairment as evidenced by scores on the Cognitive Linguistic Quick Test (CLQT: Helm Estabrooks, 2001). Treatment Participants Participants were asked about their stroke history as all participants had to be more than 9 months post onset. Additionally, al l participants had to be functionally monolingual. They may have studied languages previously but could not use them on a regular basis around the time of or subsequent to the stroke. The self reported stroke data will be confirmed by medical records as they are acquired. After being consented into the study, treatment participants were screened using the WAB to determine aphasia type and severity and the ABA 2 to determine existence and severity of apraxia of speech. Participant 1 (P1) is a 73 year old, monolingual, left handed Caucasian male. He suffered an ischemic stroke with hemorrhagic conversion, but had no other history of neurological disorder. Prior to retirement, P1 worked for the sheriffs department. Once he retired from the sheriffs department, he became heavily active in local politics. Since his stroke, he continues to assist in local politics but in a reduced capacity and also attends a book club for people with aphasia. He currently lives with his girlfriend who assists him with t ransportation and other activities as necessary. He received speechlanguage therapy until two weeks prior to the beginning of testing. During our pretreatment testing, P1 was determined to have anomic aphasia with an aphasia quotient of 88.4 as determined by the Western Aphasia Battery Revised (WAB R:

PAGE 29

29 Ker t e sz, 2006). Based on the judgments o f the author and an experienced speech language pathologist, he displayed no indications of apraxia of speech (AOS) (e.g., slow rate with inability to increase rate w hile maintaining phonemic integrity, successive self initiated trials to repair error, phoneme distortions, etc. (McNeil, Robin, & Schmidt 2009), so the Apraxia Battery for Adults 2 (ABA 2: Dabul, 2000) was not administered to reduce test burden on the participant. While the ABA 2 would provide a formal confirmation of a lack of apraxia of speech, we did not administer it due to the participants growing levels of fatigue and frustration at the lengthy testing process Participant 2 (P2) is a 55 year old, right handed Caucasian male. Many years prior to his stroke, he learned some German and Japanese which he has since rarely spoke n and after his stroke, he learned some rudimentary American Sign Language to help him communicate He suffered an ischemic st roke approximately 6 years ago, but had no other history of neurological disorder as per personal report. We are still awaiting his medical records for a more detailed history of the event. Previously, P2 worked as an instructor for the military and showed a high proficiency for computers. Since his stroke, he continues to work on computers and participates in a local aphasia support group. P2 currently lives with his wife who assists with activities of daily living as necessary. He was last enrolled in speech language therapy over 3 years before the start of treatment. During our pretreatment testing, P2 was determined to have Brocas aphasia with an aphasia quotient of 51.2 as determined by the WAB R. Additionally he was tested using the ABA 2 and was found to have mild moderate apraxia of speech. Participant 3 (P3) is a 54 year old, monolingual, right handed Caucasian male with 14 years of education. He suffered an ischemic stroke approximately 4 years ago

PAGE 30

30 but had no other history of neurological disorder as per personal report. We are still awaiting his medical records for a more detailed history of the event. Prior to his stroke, he was employed as manager of an IT department. Since his stroke, he has att empted to re teach himself how to program computers. He also occasionally plays guitar. He was last enrolled in speechlanguage therapy 10 months before the start of treatment. During our pretreatment testing, P 3 was determined to have conduction aphasia w ith an aphasia quotient of 84.8 as determined by the Western Aphasia Battery Revised. Additionally the ABA 2 revealed moderateto severe AOS. In previous VNeST studies, more then mild AOS was an exclusion criterion. However, since speech and typing were included in the current study, he was enrolled. In consideration of his AOS, the treatment was slightly modified for him; see the treatment section for specific details regarding protocol modification. One additional participant was consented for the study but was not enrolled due to self report of a learning disability, which was an exclusionary criterion. Study Design A multiple baseline, singlesubject design was used. Three baseline sessions were conducted with both spoken and typed responses to sente nce probe stimuli in addition to a digit span control task. Additional testing was conducted during two other sessions for a total of five pretreatment testing sessions. All pretreatment testing was performed inperson for P1; one session of pretreatment testing was performed via the Internet for P2 and P3 due to transportation limitations. Treatment consisted of twenty four two hour sessions all administered over the Internet. We intended to do therapy three times/week for all participants, and P2 and P3 were able to adhere well to that schedule. However, f or P1 sessions were scheduled twice per week due to scheduling

PAGE 31

31 conflicts which included community activities, local politics, and frequent travel. The effect of treatment was measured during two sess ions concurrent with the treatment phase, after eight and sixteen treatment sessions, and again during post treatment testing. Stimuli development Sentence P robe S timuli The pictures used to elicit sentences during baseline, treatment, and post treatment testing were developed for previous VNeST studies (Edmonds, et al., 2009; Edmonds & Babb, 2011). The sentence probe pictures include two practice pictures and 28 probe pictures, each designed to elicit a single, simple sentence (e.g. The chef boils the corn.). In previous VNeST studies, these pictures were used to elicit only spoken responses; however, they were used for both spoken and typed responses in the current study Examples of these pictures can be seen in Appendix A Treatment S timuli The verbs and semantic judgment sentences used for this study were developed for previous VNeST studies (Edmonds, et al 200 9 ; Edmonds & Babb, 2011 ). Fourteen pairs of semantically related verbs (e.g., hold/carry) were used in this study. Twenty sentences containing each verb were also used for the semantic judgment portion of the treatment (e.g., The suitcase carries the tourist.). These semantic judgment sentences were also developed for the previous VNeST studies. The application of these items during treatment is detailed below. VNeST C C omputer P rogram The VNeST C computer program was written in the University of Florida Aphasia & Bilingualism Lab by the author and a research assistant in the lab. The computer

PAGE 32

32 program was created using the Java comp uter language and the NetBeans Integrated Development Environment (Oracle, 2012). The program was designed to function according to the previously published VNeST protocol (Edmonds, et al 200 9 ; Edmonds & Babb, 2011) with some modifications for use with t he computer. The program was written to first display a start up screen where the participant ID must be entered. It is checked against an ID list stored in an external text file. This ID must match information in the file before the program will advance. The VNeST C program then retrieves stimuli from a Microsoft Excel file where each verbs information is kept on a separate sheet. Each sheet contains the verb as well as all related sentences used for semantic judgments during treatment. Over the course of the session, the responses entered into the program are stored in a second Excel file. This file reports the final responses to each section of the program as well as the participant ID and the time at the beginning of each section. Each stage of the program is discussed in depth under the Treatment section of the procedures section. Pictures of each stage of the computer program as well as sample input and output data files are provided Appendix B Computer Q uestionnaire A computer questionnaire was developed and administered to the younger and older adult participants to serve two functions. The initial section served the purpose of gathering information about frequency of computer use, comfort with computers and depth of computer knowledge as per self report using a series of checkboxes. The second section was used to obtain feedback regarding the VNeST C computer program. These questions pertained to the graphical interface used, speed of the programs performance and overall function of the therapy program, again using a series of checkboxes. Underneath this section, participants were given a space to write

PAGE 33

33 in any additional suggestions and commentary regarding the VNeST C program. A copy of the questionnaire is provided in Appendix C Procedure Young and Older Adults Each younger and older adult attended one, twohour session. Each participant was tested for inclusion using the Cognitive Linguistic Quick Test ( Helm Estabrooks, 2001) to screen language and cognitive function as well as Pyramids and Pal m Trees ( Howard & Patterson, 1992) and Kissing and Dancing ( Bak & Hodges, 2003) to test semantic processing. After the completion of testing, each participant underwent one treatment session of VNeST C with two verb topics (e.g. kick, read) which were ran domly selected. During the course of the session, the participants were guided through the first verb of a simulated treatment session by the researcher to familiarize them with the program interface. Participants were then asked to attempt the second verb without guidance from the researcher; however, the researcher remained in the room to provide assistance as necessary. After the completion of the second verb, the participant was given the computer questionnaire to rate their experience with the program. Treatment Participant s Pre treatment t esting In order to determine the type and severity of aphasia, the participants were given Part 1 of the Western Aphasia Battery Revised ( Kertesz, 2006) as well as the Reading and Writing sections from Part 2. An Object and Action Naming Battery (Druks & Masterson, 2000) List A was administered as a measure of oral naming, and List B (matched for frequency and age of acquisition) was administered to evaluate typed

PAGE 34

34 naming. Spoken and typed discourse was elicited using the pictures from Nicholas and Brookshire (1993). The pictures and prompts were divided in half according to Nicholas and Brookshires protocol with one half used for verbal responses and the other half for typed responses to allow for cross modality comp arisons of variety of discourse types (e.g., procedural, single picture description, sequential picture description, etc.). The participants were also tested using a selection of 30 words from the Johns Hopkins University Dysgraphia and Dyslexia Battery ( Goodman & Carmazza, 1986) to determine if there were modality specific production impairments. These thirty words, seen in Appendix D were given over the course of three pretreatment sessions to test repetition, writing to dictation, written copying, typing to dictation and typed copying. By administering these words across verbal, written and typed modalities, comparisons regarding relative ability in each modality could be made. In addition, the participants completed Pyramids and Palm Trees ( Howard & Patterson, 1992) and Kissing and Dancing ( Bak & Hodges, 2003) to evaluate semantic knowledge and function, Cognitive Linguistic Quick Test ( Helm Estabrooks, 2001), Northwestern Assessment of Verbs and Sentences (NAVS) Argument Structure Production Test and NAVS Comprehension Test ( Thompson, 2011 ), Test of Everyday Attention ( Robertson, Ward, Ridgeway, & Nimmo Smith, 1994 ), Communication Effectiveness Index ( Lomas, Pickard, Bester, Elbard, Finlayson, & Zochaib, 1989), and PALPA Subtest 8: Nonword Repetition task ( Kay, Lesser, & Coltheart, 1992). See the results section for scores from pretreatment testing. Baseline testing Each treatment participant completed a total of five pretreatment testing sessions. During three of the sessions, the participant was tested using the sentence

PAGE 35

35 probe stimuli and digit span control task. In each of these sessions, the two practice stimuli were presented first, followed by the 28 test stimuli in random order. Each picture was presented as a slide in a Microsoft PowerPoint file. The participant was asked to verbally produce a single sentence to describe the picture. For each picture, the researcher indicated the agent, action and patient for each sentence by pointing along with prompts such as Please make a sentence about him (point to agent) and what he is doing here (point to action) to this (point to patient). If the participant provided an incomplete responses (e.g., a sentence without a verb), the researcher probed once more for that aspect of the picture. If the response given was too general (e.g., man for cowboy), a more specific response was requested (e.g., Is there a more specific word you could use for him?) Likewise, if the response were too specific (e.g., John Wayne for cowboy), a more general respo nse was requested (e.g., Instead of giving me a name, what would you call him in general?) However, only one prompt was given, and the response provided after the prompt was recorded. By probing for more or less specific responses, the researcher was able to test whether the semantic concept of the target word (e.g., cowboy) was able to be accessed instead of a generic response (e.g., man) or exemplar (e.g., John Wayne). After the participant provided a verbal response to the sentence, he was then prompt ed to provide a typed response for the same picture. The participant typed this response into the notes section of the PowerPoint slide located underneath the probe picture. The initial prompt and pointing were repeated for the typing portion of each pict ure and similar clarification prompts were given as necessary. The verbal responses were transcribed from the recorded videos and the typed responses were copied from

PAGE 36

36 the PowerPoint file for later scoring. After the completion of the sentences, the partic ipant was asked to perform PALPA subtest 13 a digit span recall task, as a control task. Participant 2 performed PALPA subtest 8, a nonword repetition task because he was unable to perform a digit span task. These two subtests were chosen as control tasks for this study because digit recall and nonword repetition were not targeted by the treatment though they both require spoken production, which was trained in this study The digit span task is primarily a function of working memory while nonword repetit ion does not access the lexical semantic systems targeted in this treatment, making them sufficient control tasks for this study. Treatment After scoring the baseline probes, 10 pairs of verbs were selected for each participant from the 14 sentence pairs g iven during baseline testing. Pairs where a participant scored at ceiling were excluded from selection for treatment. One half of each pair (e.g., kick and throw) was directly treated (e.g., kick) while the semantically related untrained verb (e.g., throw) was not treated in order evaluate potential generalization on these stimuli. These pairs were split and organized such that the trained and untrained groups started at comparable levels of sentence accuracy and there were no ascending baselines. Prior to the start of each session, the researcher selected the verbs for treatment during the session such that each of the 10 verbs was treated in turn before being repeated. The selection of verbs was done by moving the target treated verbs to the first slots of the spreadsheet containing all treatment stimuli; this allowed the VNeST C computer program to read the appropriate information for the current therapy session.

PAGE 37

37 Treatment was performed via the Internet using Adobe Connect software (Adobe, 2012), allowing the participant to access VNeST C from home. The participant was given an Adobe Connect log in, which was used to connect to sessions that were set up by the researcher. These sessions were limited in access so that only the participant, the researcher and a supervisor had access to view each session. Across one side of the screen, the Adobe Connect system allowed for set up of audio and video communication during each session as well as a text based chat window in ca ses where typed visual information was required. The researcher then shared his/her desktop screen that was running the VNeST C program, allowing the participant to see the screen and type directly into the program from a remote location. Each step of trea tment was guided by the researcher or a trained research assistant. The researcher or research assistant prompted the participant about what actions needed to be taken during each stage of the treatment and controlled advancing from one step of the program to the next. The researcher or research assistant also provided verbal feedback to indicate both appropriate responses as well as responses that required additional clarification. Treatment occurred twice per week for twelve weeks or three times per week for eight weeks (depending on the participants availability) a total of 24 sessions, with each session lasting for approximately two hours. Sessions were recorded using the recording software within Adobe Connect and Flash V ideo files were then created and stored locally for future review. Step 1: agent verb patient triad c reation The VNeST C treatment protocol is broken into four steps, similar to those found in previous VNeST studies (Edmonds, et al., 2009; Edmonds & Babb, 2011) Pictures of the in terface at each step of treatment can be found in Appendix B During the first step

PAGE 38

38 of the treatment, the participant was given a prompt at the top of the window which said Who (Verb) What where (Verb) is replaced by the appropriate treatment target (e.g ., Boils, Reads). The participant was then asked to verbally generate an appropriate agent and patient pair (i.e., Who read? What do they read?) (e.g., Student Reads Books). Once the participant provided an agent or patient, they were then asked to type the word into the appropriate box in either the left (who/agent) or right (what/patient) column. In the case where a participant could not provide their own agent or patient, the researcher would begin by providing semantic or contextual cues such as Think of someone who might (boil, read, etc.) something as part of their job. If the participant still could not produce his own response, the researcher then provided two options to select from, one of which was appropriate and one of which was related but n ot appropriate for the given verb (e.g., Chef and Waiter for boils). After the participant verbalizes their chosen word, if the participant could not correctly type the response in two attempts, the researcher then proceeded to assist with the spelling of the word letter by letter as the participant typed it. After the word was typed, the participant then read aloud the spelling of the typed word again before progressing with treatment. Repeating the letter names rather than their associate phonemes was chosen due to the fact that the typing modality is more dependent on searching for the right letter rather than phoneme. O nce the participant provided at least three sets of agents and patients for the verb, one of which was always the target used in the sent ence probes for testing, the participant then read the responses aloud (e.g., Chef boils corn). The

PAGE 39

39 participant then chose one of the scenarios for the next step (Which one would you like to talk more about?) For P1 and P2, this was the end of Step 1. However, an intervening step was introduced for P3. For him, prior to beginning step 2, a Microsoft Word document was displayed onscreen with the chosen sentence, typed by the clinician. At this point, P3 read the sentence aloud 3 times. The program was then minimized and P3 repeated the sentence three more times before moving on to step 2. This was done to increase the amount of verbal production performed during treatment due to P 3s comorbid apraxia of speech. The intent of this is to improve the over all accuracy of verbal production by having P3 repeat sentences of differing complexities and lengths. This is based on the theory that more complex production practice may lead to higher improvement in people with apraxia of speech (Maas, Barlow, Robin, & Shapiro, 2002). Step 2: wh q uestions During step two, the participant was asked whquestions in order to allow him to elaborate on the selected response from step one. Specifically, the information related to where, when and why was elicited in order to create a detailed sentence (e.g., Chef boils corn in the kitchen at lunchtime for the customers.). (Participants were not required to produce complete prepositional phrases, as are shown in the example. Rather, the ideas were most important. However, i f they were able to the prepositions and other grammatical components, they were not discouraged). As with all steps in this protocol, a verbal response was first required, after which the response was typed with cues and spelling corrections occurring as in step one. After the participant typed the responses, he read aloud the entire response one last time before progressing to step three.

PAGE 40

40 Step 3: semantic j udgments During the third step of treatment, the researcher read aloud sentences selected from a li st at random by the VNeST C computer program, each containing the verb currently being treated. (P3 read them aloud himself). As shown in Appendix B the sentence was also displayed on the screen along with buttons that said wrong order and all correct. First the participant was asked to make a judgment as to whether the sentence made sense or not (e.g., The chef boils corn, or The mouse boils corn.). If the participant answered incorrectly, the researcher asked for an explanation and discussed why their answer was incorrect. Once the correct answer was established, if the sentence was not plausible, the participant was then asked to identify what part of the sentence was incorrect. There were four sentence conditions: the agent was wrong, the patient was wrong, the agent and patient were reversed (e.g., The corn boils the chef.), the sentence was correct. If the participant could not independently identify which part was incorrect, the researcher asked a series of questions to help establish the answer For example, for The corn boils the chef, the researcher would ask Does corn boil something by itself? Ten semantic judgments were performed for each verb before proceeding to the next step. Step 4: r ecall and review To begin the final step, the researcher minimized the VNeST C program and asked the participant to recall which verb they had been practicing. Once answered correctly, the researcher then returned the display to normal. For P3, in lieu of the standard verb recall step, he was asked to type any sentence containing the target verb plus an agent, patient and prepositional phrase. The participant was then asked again to produce up to three agent verb patient triads (but the pairs did not have to be the

PAGE 41

41 same that had been generated in step 1). However, unlike step one, the participant did not receive cues and was not required to type answers. Once the participant produced at least one set of responses independently, if they could not produce others, this step was discontinued. After completing this step, the VNeST C program would provide the next verb and return to step one of the protocol. For each session, the four step protocol was repeated until four verbs were completed or two hours of treatment had occurred, whichever happened first During the course of treatment, P1 completed 50 verbs in 20 sessions, P2 completed 79 verbs in 24 sessions, and P3 completed 67 verbs in 24 sessions. Treatment P robes After eight and 16 sessions of treatment, the sentence probe pictures task and digit span control task were administered to measure potential generalization. During these sessions, the two practice pictures were given first, then the 20 pictures containing the trained and semantically related verbs were presented in a random order with the same protocol that was used during baseline testing. After the sentence probes, the digit span task was administered. No treatment was performed during these sessions. Po st t reatment T esting After the completion of 24 sessions, 3 sessions of post treatment testing will be performed. The protocol was the same as that conducted for baseline and sentence probes. The results of these tests were graphed along with the pretrea tment and midtreatment sentence probe tests. Additionally, the same tests administered during pretreatment testing were administered again during post treatment. Results from post treatment testing can be seen in results section

PAGE 42

42 Treatment R eliability T reatment reliability was performed by the Principal Investigator of the study. Ninety five percent of the treatment sessions were observed to ensure protocol was adhered to, and 99% treatment reliability was observed (number of correct treatment steps/all treatment steps). It was observed that the computer program helps with treatment reliability, as it moves the clinician through the treatment steps, thereby reducing human error of forgetting steps or doing steps out of order. Scoring Reliability Sentence probe scoring was performed by the researcher or a trained research assistant. Audio and video recordings taken during tests were reviewed during scoring as appropriate. Discourse transcription was performed by a trained research assistant with audio and v ideo recordings taken during the testing sessions being reviewed as necessary.

PAGE 43

43 CHAPTER 3 RESULTS Younger and Older Adults Research Question 1 With regard to Research Question 1 (usability as rated by cognitively normal young adults), five participants (mean age = 26.8 years, range: 20 to 32 years) completed the computer survey provided at the end of their session. All participants responded to the question of usability by selecting that they could use the program with no difficulty, indicating that the young adults found the program usable and thus satisfying the criterion for question one. Additionally, in the comment space provided on the sheet, four of the users made suggestions to improve the interface. One participant suggested that we adapt the program to change font sizes automatically based on the amount of text the user types, one suggested changing the color palette to be less monochromatic, one suggested that the computer should respond more quickly, and one suggested that we cause the program to change verb conjugation based on the agent (e.g., I knit or He knits). Research Question 2 To answer Research Question 2 (usability as rated by older adults), three older adults (mean age = 64.3 years, range: 56 to 71) also completed the same computer survey at the end of their session. Two of the participants responded to the question of usability indicating that they had little difficulty with the program while one indicated no difficulty meeting the criteria for usability In the commentary spac e provided, one participant suggested that the computer should respond quicker and that using the tab

PAGE 44

44 key to move between boxes should be more responsive. The other older adult participants did not make any suggestions. Participant 1 Research Question 3 To answer Research Question 3 (Was there a reduction of cues during treatment?), each treatment session was analyzed for the level of cueing required during the first two non target agent and patient productions for each verb. Responses were grouped into cyc les where each of the 10 trained verbs w as practiced once. Figure 3 1 shows changes in cueing requirements for each of P1s completed cycle s of treatment. Participant 1's required cueing of agents remained unchanged between cycle 1 and cycle 5. There was a slight improvement during cycle 3, but gains were lost in cycle 4, which correlated with medical and personal issues occurring for Participant 1. As seen in Figure 31, P1 did improve his ability to independently produce patients without cues over the co urse of treatment. Research Question 4 In order to investigate Research Q uestion 4 (Did typing accuracy improve during treatment?), data was gathered during treatment sessions regarding typing accuracy of target agents and patients (e.g., carpenter and wood for the verb measures). Independently typed participant responses were grouped into either 0 errors, 1 error or 2+ errors. Figure 32 shows the number of typing errors for each cycle of treatment. As Figure 32 indicates there was a decrease in the number of words with two or more errors in production of agents, which converted to single letter errors. The number of errorfree words remained virtually unchanged. For patient responses, Participant 1

PAGE 45

45 st arted with a high number of error free words during cycle 1, but that number of errorless productions was not achieved again during the course of treatment. Research Question 5 During the pretreatment testing sessions, 3 baselines were taken for verbal and typed responses to the sentence probe pictures. A nonascending baseline was established for the complete sentences as seen in Figure 33 An additional post hoc analysis was done to determine if isolated components of the target sentences (i.e., agent verb and patient) showed additional changes. These can be seen in Tables 31 and 32. According to Cohens d calculations P1 showed a large improvement on untrained sentences in the verbal modality between baseline and post treatment, ( d = 4.62) and moderate improvement on trained sentences ( d = 1.15). For typed responses to the sentence probes, Cohens d calculations were not performed as all three baseline values and the post treatment value were all zero for trained and untrained sentences. P1s score on the digit span control task was quite variable during baseline, and as a result a fourth baseline was collected. As seen in Figure 34, his post treatment score was equal to his final baseline, indicating that the control was relatively stable over the course of treatment, an indication of experimental control. In addition to the sentence probes a series of other tests were administered during pretreatment testing. The results of these tests can be seen in Table 33 These tests will be readministered during the post treatment testing sessions and will be analyzed at that time.

PAGE 46

46 Participant 2 Research Question 3 As with Participant 1, treatment sessions with Participant 2 were analyzed for change in the number and types of cues given during verbal produc tion of nontarget agents and patients. As seen in Figure 35 Participant 2 showed notable improvement in the number of independent production of both agents and patients over the course of therapy. Most notably, during cycle 1, half of all attempts requi red maximum cueing while subsequent cycles only required a maximum cue during 10% of attempts. Research Question 4 As with Participant 1, sessions were analyzed for the number of errors during typing production of target agent and patients As depicted in Figure 36, Participant 2 showed some improvement in the accuracy of typed agents. The number of words with two or more errors decreased and converted to either one or no errors per production. Patients, on the other hand, remained relatively stable from c ycle 1 to cycle 7 Research Question 5 During the pretreatment testing sessions, 3 baselines were taken for verbal and typed responses to the sentence probe pictures. A nonascending baseline was established for the complete sentences as seen Figure 37 An additional post hoc analysis was done to determine if isolated components of the target sentences (i.e., agent, verb and patient) showed additional changes. These can be seen in Tables 34 and 35. P2 showed notable increases in almost all facets of the sentence probe task. P2 showed improvement on all aspects of the spoken sentence probe task with effect sizes ranging from d = 2.31 for the untrained agents (a 6.66% increase) to d = 12.70 for

PAGE 47

47 trained verbs (a 36.67% increase). Additionally, accuracy on trained agents increased by 50% and trained patients increased by 40%. Cohens d values cannot be calculated for most of typed responses because there was no variation in the baselines, resulting in no standard deviation. Of those that could be calculate d, trained agents and patients showed very large improvement s ( d = 25.40 and 23.09, respectively ). Additional gains were made thus far in trained verbs and sentences as well as untrained agents and patients. The trained verbs have increased from 0% correct at baseline to 50% correct in post treatment testing Likewise, untrained agents increased from 0% to 50%. As shown in Figure 38, P2s control task, the PALPA Nonword repetition task, did not show any inc rease over baseline at the time of the first treatment probe, indicating experimental control. In addition to the sentence probes a series of other tests were administered during pretreatment and post treatment testing. The results of these tests can be seen in Table 36 P2 showed clinically significant improvement on the WAB AQ (from 51.2 to 58.2) with increases in spontaneous speech, repetition and naming. Additionally, P2 also improved on the writing section of the WAB. P2 also improved to within norm al limits on the attention subscore and from severe to moderate executive function impairment on the CLQT, resulting in an improvement overall to mild impairment. He also made notable improvements on the spoken portion of the Object and Action Naming Battery. The results in Table 36 show in detail how widespread his improvement was during post treatment testing.

PAGE 48

48 Participant 3 Research Question 3 As with Participant 1, treatment sessions with Participant 3 were analyzed for change in the number and types of cues given during verbal production of nontarget agents and patients. Results can be seen in Figure 39 Over the course of six cycles of treatment, P3 show ed stability in the number of cues for patients over the course of therapy though his responses w ere mostly independent from the beginning However the number of cues for agents has improved, and during the final cycle of treatment, P3 was able to independently produce all agents. Research Question 4 The research assistant performing treatment did not collect typing data for the first two cycles of treatment. As a result, R esearch Q uestion 4 begins with cycle 3 of treatment As seen in Figure 310, no major changes can be seen between cycles three and six. Typing accuracy of both agents and patients im proved by one answer between cycles three and six; however, it is unclear if this is within natural variation as we do not have baseline information. Research Question 5 During the pretreatment testing sessions, 3 baselines were taken for verbal and typed responses to the sentence probe pictures. A nonascending baseline was established for the complete sentences as seen Figure 311. An additional post hoc analysis was done to determine if isolated components of the target sentences (i.e., agent, verb and patient) showed additional changes. These can be seen in Tables 37 and 38.

PAGE 49

49 P3 has shown improvement in both the spoken and typed modalities as of the first treatment p robe session. The only aspect s of the spoken probes which did not improve were untrained patients which has remained at 90% accuracy and untrained verbs which reduced from 50% to 40% accuracy Trained agents ( d = 3.49), verbs (d = 9.24) and patients ( d = 4 00) have all shown notable increases. Likewise, untrained agents ( d = 4.00) increased, as well. As a result both trained ( d = 5.77) and untrained ( d = 1.15) sentences have improved for the verbal modality. For typed responses trained verbs improved ( d = 6 .00) as did trained patients ( d = 2.00) with a corresponding improvement in trained typed sentences ( d = 6.35). Untrained agents verbs and patients have improved ( d = 2.31, 4.36, and 1.15 respectively) This resulted in a notable improvement in untrained sentence production ( d = 8.00). As shown in Figure 312, P3 has not shown any increase in ability on the digit span control task as of the first treatment probe, indicating experimental control. In addition to the sentence probes a series of other tests were administered during pretreatment and post treatment testing. The results of these tests can be seen in Table 39 P3 reported that he was tired during post treatment testing as he had to drive over 3 hours for the testing session. This fatigue was re flected in some of his test results, most notably a decrease from normal to mild impairment on the CLQTs memory component. He did not show any increases on the WAB AQ, but did improve on the writing section. P3 also showed improved typing accuracy for single word naming on the Object and Action Naming Battery as well as Typing to Dictation using the Johns Hopkins words seen in Appendix D The remainder of P3s standardized testing showed little to no change.

PAGE 50

50 Figures and Tables Figure 31 P1s cue distribution for both A) agents and B) patients during treatment.

PAGE 51

51 Figure 32. P1s typing errors of target A) agents and B) patients during treatment.

PAGE 52

52 Figure 33. Participant 1s spoken and typed sentence probe results. Figure 34. Participa nt 1s results on a control task PALPA Subtest 13.

PAGE 53

53 Figure 35. P2s cue distribution for A) agents and B) patients during treatment.

PAGE 54

54 Figure 36. P2s errors of target A) agents and B) patients during treatment.

PAGE 55

55 Figure 37. Participant 2s spoken and types sentence probe results. Figure 38. Participant 2s results on a control task PALPA Subtest 8.

PAGE 56

56 Figure 39. P3s cue distribution for A) agents and B) patients during treatment.

PAGE 57

57 Figure 310. P 3s errors of target A) agents and B) patients during treatment

PAGE 58

58 Figure 311. Participant 3s spoken and types sentence probe results. Figure 312. Participant 3s results on a control task PALPA Subtest 13.

PAGE 59

59 Table 31. P1s spoken agent verb and patient production during sentence probes Spoken c omponents Baseline average Post Tx Probe % Change Cohens d Trained a gents 9.33 9.00 3.33% 0.58 Untrained a gents 8.00 8.00 0.00% 0.00 Trained v erbs 5.67 7.00 13.33% 2.31 Untrained v erbs 6.67 8.00 13.33% 4.62 Trained p atients 6.00 7.00 10.00% N/A* Untrained p atients 7.33 9.00 16.67% 2.18 N/A Cohens d could not be calculated because baseline standard deviation was 0. Table 32. P1s typed agent, verb and patient production during sentence probes. Typed c omponents Baseline average Post Tx p robe % Change Cohens d Trained a gents 2.33 4.00 16.67% 2.18 Untrained a gents 1.67 2.00 3.33% 0.58 Trained v erbs 0.33 2.00 16.67% 5.77 Untrained v erbs 2.00 1.00 10.00% 2.00 Trained p atients 3.67 3.00 6.67% 0.87 Untrained p atients 3.67 6.00 23.33% 4.04

PAGE 60

60 Table 33. List of results from Participant 1s preand post treatment testing scores. Test Max Score Pre tx Score Post tx Score WAB Spontaneous 20 18 DNT WAB Comprehension 200 182 DNT WAB Repetition 100 94 DNT WAB Naming 100 77 DNT WAB AQ 100 88.4 DNT WAB Reading 100 70 69 WAB Writing 100 67.5 60 O&A Nouns Spoken 81 75 DNT O&A Verbs Spoken 50 40 DNT O&A Nouns Typed 81 15 DNT Pyramids & Palmtrees 52 47 46 Kissing & Dancing 52 40 39 Hopkins repetition 30 30 DNT Hopkins WTD 30 9 11 Hopkins Writing Copy 30 25 DNT Hopkins Type Dictation 30 11 7 Hopkins Type Copy 30 22 22 CLQT Attention 215 102 138 CLQT Memory 185 131 146 CLQT Executive 40 13 10 CLQT Language 37 24 28 CLQT Visuospatial 105 53 54 CLQT Clock drawing 13 4 12 CLQT Overall 4 2.6 3.2 Table 34. P2s spoken agent, verb and patient production during sentence probes. Spoken components Baseline average Post Tx Probe % Change Cohens d Trained agents 4.00 9.00 50.00% 10.00 Untrained agents 3.33 4.00 6.66% 2.31 Trained verbs 4.33 8.00 36.67% 12.70 Untrained verbs 2.00 4.00 20.00% 4.00 Trained patients 5.00 9.00 40.00% 8.00 Untrained patients 3.67 6.00 23.33% 8.08

PAGE 61

61 Table 35. P2s typed agent verb and patient production during sentence probes. Typed components Baseline average Post Tx Probe % Change Cohens d Trained agents 0.67 8.00 7 3.33% 25.40 Untrained agents 0.00 5.00 5 0.00% N/A* Trained verbs 0.00 5.00 5 0.00% N/A* Untrained verbs 0.00 0.00 0.00% N/A* Trained patients 1.33 8 .00 6 6.67% 23.09 Untrained patients 5.00 6 .00 10.00% N/A* N/A Cohens d could not be calculated because baseline standard deviation was 0. Table 36. List of results from Participant 2s preand post treatment testing scores. Test Max Score Pre tx Score Post tx Score WAB Spontaneous 20 10 12 WAB Comprehension 200 170 132 WAB Repetition 100 34 49 WAB Naming 100 37 56 WAB AQ 100 51.2 58.2 WAB Reading 100 80 80 WAB Writing 100 48 55.5 O&A Nouns Spoken 81 39 52 O&A Verbs Spoken 50 18 DNT O&A Nouns Typed 81 11 19 O&A Verbs Typed 50 2 DNT Pyramids & Palmtrees 52 48 48 Kissing & Dancing 52 DNT 45 Hopkins repetition 30 14 18 Hopkins WTD 30 3 3 Hopkins Writing Copy 30 30 29 Hopkins Type Dictation 30 3 5 Hopkins Type Copy 30 29 30 CLQT Attention 215 176 182 CLQT Memory 185 87 108 CLQT Executive 40 18 21 CLQT Language 37 14 18 CLQT Visuospatial 105 85 91 CLQT Clock drawing 13 11 11 CLQT Overall 4 2.2 2.6

PAGE 62

62 Table 37. P3s spoken agent, verb and patient production during sentence probes. Spoken components Baseline average Post Tx Probe % Change Cohens d Trained agents 7.33 10.00 26.67% 3.49 Untrained agents 7.00 9.00 20.00% 4.00 Trained verbs 7.33 10.00 26.67% 9.24 Untrained verbs 5.00 4.00 10.00% 2.00 Trained patients 8.00 10.00 20.00% 4 .00 Untrained patients 9.00 9.00 0.00% N/A* N/A Cohens d could not be calculated because baseline standard deviation was 0 Table 38. P3s typed agent, verb and patient production during sentence probes. Typed components Baseline average Post Tx Probe % Change Cohens d Trained agents 9.00 9.00 0.00% 0.00 Untrained agents 9.33 10.00 6.67% 2.31 Trained verbs 7.00 10.00 30.00% 6.00 Untrained verbs 5.67 9.00 40.00% 4.36 Trained patients 9.00 10.00 10.00% 2.00 Untrained patients 9.67 10.00 3.33% 1.15

PAGE 63

63 Table 39. List of results from Participant 3s preand post treatment testing scores. Test Max Score Pre tx Score Post tx Score WAB Spontaneous 20 17 18 WAB Comprehension 200 200 183 WAB Repetition 100 68 66 WAB Naming 100 86 84 WAB AQ 100 84.8 84.3 WAB Reading 100 96 100 WAB Writing 100 60 79.5 O&A Nouns Spoken 81 59 DNT O&A Verbs Spoken 50 36 DNT O&A Nouns Typed 81 73 76 O&A Verbs Typed 50 36 45 Pyramids & Palmtrees 52 50 48 Kissing & Dancing 52 49 48 Hopkins repetition 30 15 17 Hopkins WTD 30 20 23 Hopkins Writing Copy 30 28 29 Hopkins Type Dictation 30 25 30 Hopkins Type Copy 30 30 30 CLQT Attention 215 196 185 CLQT Memory 185 143 127 CLQT Executive 40 27 24 CLQT Language 37 23.5 23 CLQT Visuospatial 105 98 87 CLQT Clock drawing 13 10 13 CLQT Overall 4 3.4 3.2

PAGE 64

64 CHAPTER 4 DISCUSSION The primary aim of this study was to investigate the effects of a computerized semantic language therapy on lexical access in people with aphasia. This was tested using in a progression of five research questions. The results of this study suggest that VNe ST C may have a positive effect for some people with aphasia. In this discussion section, each research question will be discussed in turn, followed by an overall general discussion of the entire study. Discussion of Research Questions Research Question 1 Research Q uestion 1 asked whether the VNeST C computer program would be usable as reported by neurotypical younger adults. All five young adults who were trained on the program reported that they would have no difficulty using the program, which met our cr iterion for high usability within the group. The additional suggestions given as part of the post session survey to improve the interface such as automatically resizing text were taken under consideration and will be implemented in future iterations of the VNeST C program. These suggestions were not implemented prior to the start of treatment for this pilot study as they were not crucial to the function of the program as each text box was already set up to accommodate 16 characters, which was enough for vir tually all anticipated answers. During the course of the testing session with the VNeST C program, none of the younger adults requested additional assistance with the program after their initial training was completed.

PAGE 65

65 Research Question 2 Once the criterio n for R esearch Q uestion 1 was met, R esearch Q uestion 2 was investigated. Research Q uestion 2 pertained to usability of the VNeST C program as reported by older, cogni tively normal adults. The three older adults tested to this point have met the criterion f or high usability two reported little difficulty and one reported no difficulty. As with the younger adults tested in Research Q uestion 1, the suggestions provided on the survey will be taken under advisement when the program is modified in the future. A s with the suggestions from the younger adults, these were not immediately implemented as they were not crucial to the function of the program. During the course of the testing session with the VNeST C program, none of the older adults requested additional assistance with the program after their initial training was completed. Research Q uestion 3 Research Q uestion 3 pertains to the level of cueing required during treatment, with the expectation that cueing would decrease over the course of treatment. T he overall trend shown during treatment is a mixture of a stable level of cueing and more independent productions. P1 improved on cueing required for patients while P3 improved on cueing required for agents. P2 improved on cueing for both cases. For P1s a gent production and P3s patient production during treatment, which have remained stable, this stability may be due to different causes. P3 was already near ceiling with independent production of patients, making significant changes nearly impossible. For P1, who showed some improvement before regressing again, the return to baseline may be due to a multitude of factors which will be further discussed in the general discussion section below. Overall, this indicates that the treatment is working to improve i ndependent lexical access within the confines of the treatment program.

PAGE 66

66 Participants have the benefit of a visual cue of the verb onscreen to help them retrieve associated words, which would not be available in real life situations or the generalization probes so while this improvement is a promising sign, generalization to other areas is both a more traditional and a more functional measure of this treatments effectiveness. Research Q uestion 4 Research Q uestion 4 deals with improvement of typing accuracy of the target items during treatment. As shown in the results section, P1s accuracy has actually decreased over the course of treatment, which may be due to uncharacteristically accurate production during the first cycle buy may also be due to confounding cognitive impairments and personal distractions that occurred during the latter portion of the treatment. P2 showed an improvement in typing accuracy of agents and stable production of patients over the course of four cycles, and P3 showed no major cha nges in typing accuracy between cycles three and six. The trend of improvement for P2 does suggest that improvement in typing accuracy may occur for some people with aphasia as a result of treatment with VNeST C but this may correlate with the original level of impairment Further testing and additional participants would be required to discern why P2 showed typing improvement while P1 and P3 did not. Research Q uestion 5 Discussion of Research Question 5 will be looked at in terms of sentence probes first and then additional testing afterwards. E ach of the three participants has shown some level of improvement over baseline in the sentence probe task. As discussed before, P1, who showed the least overall improvement, may have had his gains limited by a var iety of personal and medical situations, the extent of which will be addressed

PAGE 67

67 below in the general discussion. P2 and P3 have both shown notable improvements in the treated portions of the sentence probe task, similar to those reported in previous tradit ional VNeST studies. Additionally, b oth P2 and P3 showed some generalization to portions of the untrained sentence probe task. P2 was able to verbally produce some of the untrained targets; however, he did not improve in the typed modality. While he did im prove on untrained, typed agents and patients, he did not improve on untrained, typed verbs. For P3, the increases were greater for typed sentences than for spoken, which may be due to his apraxia of speech. As discussed further below, P3s improvements on single word naming do indicate some limited generalization in the spoken modality. The improvements in spoken accuracy are comparable to those seen in the second VNeST study (Edmonds & Babb, 2011). The additional improvement in typing accuracy suggests t hat this treatment is beneficial to multiple modalities of language production. In addition, both P2 and P3s spouses independently contacted the researcher to relay the fact that their husbands showed noticeable improvements in everyday verbal conversation most notably that P2 and P3 were attempting to speak in complete ( subject verb object ) sentences This suggests that the treatment may have at some level generalized beyond the level of the sentence probe task to more functional communication. With regard to the additional testing, P1 only showed improvement on the CLQTs attention, memory and language sections but no others. P2 also showed improvement on the attention and executive function sections of the CLQT. This suggests that treatment may have result ed in additional cognitive improvements over and above the intended language effects. The treatment requires the participants to

PAGE 68

68 attend to the computer program and specific directions and prompts for two hours at a time while performing a number of language tasks across a number of language modalites. Maintaining and dividing attention across these various tasks may have contributed to the improved attention scores. The improvements seen in memory for P1 and executive function for P3 could be a result of these attentional improvements as the tasks on the CLQT often test a combination of two or three cognitive skills at the same time (e.g., attention, executive function and visuospatial skills for Mazes, or attention, memory and language for Story Retelling). The additional improvements to singleword naming suggest that this treatment has generalized and thus had the theorized effect of improving lexical access for untrained words. The additional improvement by P2 on the WAB AQ further suggests a significant level of generalization as his score improved by seven points. While P1 and P3 did not show as great or diverse an improvement on the standardized tests, they also did not start off with as severe a language impairment as P2 based on the ir Wes tern Aphasia Battery scores. General D iscussion Both the younger and older normal adults indicated that the VNeST C computer program was quite usable. While they were able to provide some suggestions for improvement, not uncommon for the first version of any computer program, none of the changes that were suggested were a result of major usability issues. Furthermore, this study shows that three people with aphasia, each having a different level of linguistic and cognitive impairments, were able to use the program with no serious difficulties. The three people with aphasia had to log into Adobe Connect and then use the program for two hours per session. While there were occasional connectivity issues and

PAGE 69

69 problems with forgotten passwords, these were not m ajor deterrents for treatment. As this is a pilot study for the VNeST C program, overall viability of this treatment system is as much a concern as effectiveness of the treatment itself, and this study indicates that the program is a viable teletherapy opt ion for people with aphasia. Future refinements to the system to improve usability will only make the system easier to operate, but even in their absence there were no barriers to treatment encountered over the course of this study. When comparing the fir st and last treatment cycles for improvements on intreatment cueing levels and typing accuracy, there were no major improvements for P1. He began to show some improvement on the intreatment measures about halfway into the study as indicated by the intre atment results; however, right as he began to show improvements, a series of personal issues and illnesses disrupted the continuity of his treatment. As a result, he missed five out of a series of nine sessions, some of which were made up before he was dis charged from the study. The combination of these personal distractions and the lack of treatment continuity seem to have caused a loss of any gains that were made up until that point. While P1 did not make notable improvements in regards to R esearch Q uest ions 3 and 4, he did show slight improvement on the spoken portion of the sentence probe task. P1s improvement on untrained sentences came from an increase in the correct number of untrained verbs and patients produced. The fact that he did not improve as much on the trained items is directly related to his lack of improvement on the trained verbs. While agents and patients are independently produced by the participant during the treatment, the verb is only read from the prompt or recalled after a short delay. The

PAGE 70

70 lack of independent production of the verb may have limited P1s improvement of verb production on the sentence probe task. The difficulties with session attendance towards the end of treatment were further complicated by P1's nonlinguistic cogn itive impairments. Lambon Ralph and colleagues (2010) found that cognitive skill and level of language impairment are both predictors of treatment outcomes; as people with aphasia who also exhibited other cognitive symptoms did not improve as much after therapy as those who had a purely linguistic impairment. During pre treatment testing, P1 was unable to complete 5 sections of the Test of Everyday Attention (TEA: Robertson, et al ., 1996), and received scores of mild or moderate impairment on all sections o f the CLQT. His most impaired subscore of the CLQT was found to be Executive Functioning, which was corroborated during therapy by frequent displays of perseveration and impulsivity. During treatment, P1 would frequently attempt to say or type the previous word, resulting in triads such as carpenter measures carpenter. Subsequent attempts would still result in this perseveration, at which point the researcher would have to minimize the program and change the subject in order to stop the perseverative errors. P1 also displayed impulsivity during the sessions as he would frequently try to get ahead of the researcher when repeating the spelling of words that P1 had mistyped. Many times, when P1 would try to spell the word without the researcher's assistance, there would be errors and the process would have to be repeated, causing frustration for P1. As Lambon Ralph, et al (2010) indicated with their participants P1 also received littleto no gains as a result of treatment. There were a constellation of issue s that could have limited his improvement, not the least of which were cognitive issues.

PAGE 71

71 With regard to P2 and P3, b oth participants exhibited a reduction in the level and number of cues required during treatment, and P2 additionally showed a marked improvement in typing accuracy during treatment. P2 and P3 also showed notable improvements with regards to accuracy on the sentence probes task. Th is improvement also showed generalization into other testing in administered during post treatment sessions. The generalization to the handwriting modality, which was not treated, is supported by the aforementioned theory from Cameron, Cubelli & Della Sala (2002). The treatment improved the access between the semantic and lexical systems through the orthographic buffer, which resulted in improvements to both the typing (letter detection) and writing (graphememotor) systems. The disparate outputs for writing and typing in Cameron, Cubelli & Della Salas model provide reason for why both typing and handwriting improved on the Johns Hopkins testing battery, but did not improve to the same degree. The lack of treatment directed towards the graphemewriting buff er limited the gains to handwriting as opposed to the direct treatment of letter detection (typing). Based on these results, this st udy is showing promising preliminary results that suggest that VNeST C is a viable teletherapy option for people with aphasia. Three people with aphasia, all presenting with different impairment profiles, were able to complete remote treatment sessions and garner some improvement to language and/or cognition. Further refinement of the treatment protocol as well as improvement s to the VNeST C interface, such as those suggested in Research Questions 1 and 2, could produce additional benefits for other users of the program.

PAGE 72

72 Future D irections of S tudy First and foremost, as this is a pilot study, further investigation of the effect of treatment are needed. This study encompassed three people with very different impairment profiles; future studies would need to expand the number of participants to ensure that the treatment system would provide benefits for a wide variety of people with aphasia. With regards to the future of the program, the addition of the suggestions provided by participants in this study would be a good place to start. Minor adjustments to the size of text and input areas of the interface could be completed without additional usability testing being required. Another possibility for this program would be a spinoff based on the same software which includes some automatic cueing systems built into it. A more automated version of the software could be used for home treatment, either in conjunction with clinical therapy or as a stepdown option after the conclusion of standard clinical therapy. In addition to providing cues, a home version of the program would be required to recognize and correct errors and provide appropriate feedback. This level of change to the software would require significant reprogramming and testing before being used as a clinical tool.

PAGE 73

73 APPENDIX A EXAMPLE PICTURES FROM THE SENTENCE PROBE TASK Figure A 1. The picture for The mailman is delivering the letter.

PAGE 74

74 Figure A 2. The picture for The fireman is holding the cat. Figure A 3. The picture for The snake is biting the hand.

PAGE 75

75 APPENDIX B PICTURES OF THE VNeST C COMPUTER PROGRAM INTERFACE AND OUTPUT Figure B 1. The triad creation step of VNeST C. Figure B 2. The Wh question step of VNeST C.

PAGE 76

76 Figure B 3. The semantic judgement step of VNeST C. Figure B 4. The review step of VNeST C.

PAGE 77

77 Figure B 5. Sample output data from the Triad, Whquestion and Recall steps Figure B 6. Sample output data from the semantic judgment step

PAGE 78

78 APPENDIX C COMPUTER QUESTIONNAIRE

PAGE 79

79

PAGE 80

80 APPENDIX D LIST OF SELECTED WORDS FROM THE JOHNS HOPKINS DYSLEXIA AND DYSGRAPHIA BATTERY Copy Evil Iron Jury Open Ruin Party Ready Seven Cable Fluid Igloo Center Pretty Reason Excess Fumble Pigeon Brother Problem Special Curtain Leopard Sincere Complete Mountain Question Frequent Nuisance Scramble

PAGE 81

81 LIST OF REFERENCES Adrian, J., Gonzalez, M., Buiza, J., & Sage, K. (2011). Extending the use of Spanish computer assisted anomia rehabilitation program (CARP 2) in people with aphasia. Journal of Communication Disorders 44, 666677. American Heart Association (2011). Heart disease and stroke statistics 2011 update. Circulation 123, e18e209. Archibald, L., Orange, J., & Jamieson, J. (2009). Implementation of computer based language therapy in aphasia. Therapeutic Advances in Neurologic Disorders, 2(5), 299311. B ak, T., & Hodges, J.R. (2003). Kissing and dancing a test to distinguish the lexical and conceptual contributions to noun/verb and action/object dissociation. Preliminary results in patients with frontotemporal dementia. Journal of Neurolinguistics 16, 169 181. Brennan, D., Georgeadis, A., & Baron, C. (2002). Telerehabilitation tools for the provision of remote speechlanguage treatment. Topics in Stroke Rehabilitation, 8(4), 7178. Cameron, A., Cubelli, R, & Della Sala, S. (2002). Letter assembling and handwriting share a common allographic code. Journal of Neurolinguistics, 15. 91 97. Dabul, B. (2000). Apraxia Battery for Adults 2nd Edition. Austin, TX: Pro ed. Dechene, L., Tousignant, M., Boissy, P., Macoir, J., Heroux, S., Hamel, M., Briere, S ., & Page, C. (2011). Simulated inhome teletreatment for anomia. International Journal of Telerehabilitation, 3(2). Druks, J., & Masterson, J. (2000). An Object & Action Naming Battery. Hove: Psychology Press. Duffy, J.R. (1997). Telemedicine and diagnosis of speech and language disorders. Mayo Clinic Proceedings, 72, 1116 1122. Duffy, J.R. (2005). Motor speech disorders: Substrates, differential diagnosis, and management St. Louis, MO: Mosby. Edmonds, L., & Babb, M. (2011). Effect of verb network strengthening treatment in moderateto severe aphasia. American Journal of SpeechLanguage Pathology, 20, 131145.

PAGE 82

82 Edmonds, L., Nadeau, S., & Kiran, S. (2009). Effect of Verb Network Strengthening Treatment (VNeST) on lexical retrieval of content words in sentences in persons with aphasia. Aphasiology, 23(3), 402424. Ferretti, T. R., McRae, K., & Hatherell, A. (2001). Integrating verbs, situation schemas, and thematic role concepts. Journal of Memory and Language, 44, 516 547. Goodglass, H., Kaplan, E., & Barresi, B. (2000). Boston Diagnostic Aphasia Examination (3rd Ed.). San Antonio, TX: The Psychological Corporation. Goodman, R., & Carmazza, A. (1986). The Johns Hopkins Dyslexia Battery. Baltimore, MD: Johns Hopkins University. Helm Estabrooks, N (2001). Cognitive Linguistic Quick Test. United States of America: The Psychological Corporation, a Harcourt Assessment Company. Howard, D., & Patterson, K. (1992). Pyramids and Palm Trees. London: Harcourt Assessment. Kaplan, E., Goodglass, H., & Wei ntraub, S. (1983). Boston Naming Test Philadelphia: Lea & Febiger. Katz, R.C. (2010). Computers in the treatment of chronic aphasia. Seminars in Speech and Language, 31, 34 41. Katz, R.C., & Nagy, V.T. (1982). A computerized treatment system for chronic aphasic adults. In: Brookshire RH, ed. Clinical Aphasiology: Conference Proceedings. Minneapolis, MN: BRK Publishers, 153160. Katz, R.C., & Wertz, R.T. (1997). The efficacy of computer provided reading treatment for chronic aphasic adults. Journal of Sp eech, Language, and Hearing Research, 40, 493507. Kay, J., Lesser, R., & Coltheart, M. (1992). PALPA Psycholinguistic Assessments of Language Processing in Aphasia. East Sussex, England: Lawrence Erlbaum Associates. Kertesz, A. (1982). Western Aphasia B attery Austin, TX: Pro ed. Kertesz, A. (2006). Western Aphasia Battery Revised Austin, TX: Pro ed. Kully, D. (2000). Telehealth in speech pathology: applications to the treatment of stuttering. Journal of Telemedicine and Telecare, 6(2), 3941.

PAGE 83

83 Lambon Ralph, M., Snell, C., Fillingham, J., Conroy, P., & Sage, K. (2010). Predicting the outcome of anomia therapy for people with aphasia post CVA: both language and cognitive status are key predictors. Neuropsychological Rehabilitation, 20(2), 289305. Lomas, J., Pickard, L., Bester, S., Elbard, H., Finlayson, A., & Zochaib, C. (1989). The communication effectiveness index: Development and psychometric evaluation of a functional communication measure for adult aphasia. Journal of Speech and Hearing Dis orders 54, 113124. Loverso, F L., Prescott, T. E., & Selinger, M. (1988). Cueing verbs: A treatment strategy for aphasic adults (CVT). Journal of Rehabilitation Research, 25(2), 47 60. Loverso, F., Prescott, T., Selinger, M., Wheeler, K., & Smith, R. (1982) The application of microcomputers for the treatment of aphasic adults. In: Brookshire RH, ed. Clinical Aphasiology: Conference Proceedings. Minneapolis, MN: BRK Publishers: 189195. Maas, E., Barlow, J., Robin, D., & Shapiro, L. (2002). Treatment of sound errors in aphasia and apraxia of speech: Effects of phonological complexity. Aphasiology 16(4 6), 609622. Maher, L., Kendall, D., Swearengin, J., Rodriguez, A., Leon, S., Pingel, K., Holland, A., & Rothi, L. (2006). A pilot study of use dependent learning in the context of Constraint Induced Language Therapy. Journal of the International Neuropsychological Society, 12 (6), 843 852. McNeil, M.R., Robin, D.A., & Schmidt, R.A. (2009). Apraxia of speech. In McNeil, M.R. (Ed.), Clinical Man agement of Sensorimotor Speech Disorders (2nd ed., pp. 249268). New York: Thieme Medical Publishers. Mills, R.H. (1982). Microcomputerized auditory comprehension training. In: Brookshire RH, ed. Clinical Aphasiology: Conference Proceedings. Minneapolis, MN: BRK Publishers, 147152. Nicholas, L.E., & Brookshire, R.H. (1993). A system for quantifying the informativeness and efficiency of the connected speech of adults with aphasia. Journal of Speech and Hearing Research, 36, 338 350. Porch, B.E. (1981). The Porch Index of Communicative Ability Palo Alto, CA: Consulting Psychologists Press. Prescott, T. E., Selinger, M., & Loverso, F. L. (1982). An analysis of learning, generalization, and maintenance of verbs by an aphasic patient. In Clinical Aphasiol ogy Conference, BRK Publishers, 178182.

PAGE 84

84 Rapcsak, S., & Beeson, P. (2002). Neuroanatomical correlates of spelling and writing. In A. Hillis (ed.) The Handbook of Adult Language Disorders 71 99. New York, NY: Psychology Press. Robertson, I., Ward, T., R idgeway, V., & Nimmo Smith, I. (1994). The Test of Everyday Attention Bury St. Edmonds, England: Thames Valley Test Company. Robey, R.R. (2004). A fivephase model for clinical outcome research. Journal of Communication Disorders 37, 401411. Sarno, M. (2007). Quality of life in aphasia in the first post stroke year. Aphasiology, 11(7), 665679. Seron, X., Deloche, G., Moulard, G., Rousselle, M. (1980). A computer based therapy for the treatment of aphasic subjects with writing disorders. Journal of Speech and Hearing Disorders, 45(1), 4558. Theodoros, D., Constantinescu, G., Russell, T., Ward, E., Wilson, S., & Wootton, R. (2006). Treating the speech disorder in parkinsons disease online. Journal of Telemedicine and Telecare, 12(3), 88 91. Theodoros, D., Hill, A., Russell, T., Ward, E., & Wootton, R. (2008). Assessing acquired language disorders in adults via the internet. Telemedicine and eHealth, 14(6), 552561. Thompson, C.K. (2011). Northwestern Assessment of Verbs and Sentences. Evanston, IL: Aphasia and Neurolinguistics Laboratory, Northwestern University School of Communication. Wertz, R.T., & Katz, R.C. (2004). Outcomes for computer provided treatment for aphasia. Aphasiology, 35(2), 303314.

PAGE 85

85 BIOGRAPHICAL SKETCH Daniel Furnas was born in Winter Park, Florida. After graduating from Lake Brantley High School in 2001, he attended University of Florida where he received a degree in East Asian l anguage and l iterature and minored in computer and i nformation science and t echnology in 2005. Afterwards, he worked as a bench scientist and laboratory department manager until he returned to school in fall 2010. He graduated with a Master of Arts in communication sciences and d isorders in 2012.