<%BANNER%>

Neural Change in Patients with Chronic Nonfluent Aphasia after Language Rehabilitation: A Functional Magnetic Resonance ...


PAGE 1

NEURAL CHANGE IN PATIENTS WITH CHRONIC NONFLUENT APHASIA AFTER LANGUAGE REHABILITATI ON: A FUNCTIONAL MAGNETIC RESONANCE IMAGING (FMRI) STUDY By BRADLEY J. DANIELS A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2005

PAGE 2

Copyright 2005 by Bradley J. Daniels

PAGE 3

This thesis is dedicated to the memory of Christopher Ryan Lefever.

PAGE 4

iv ACKNOWLEDGMENTS I would like to thank Dr. Anna Bacon Moore for her mentorship and guidance throughout this project. I woul d also like to acknowledge Dr. Bruce Crosson, Dr. Keith White, Dr. Kaundinya Gopinath, Dr. Kyunk K. Peck, Dr. Leslie J. Gonzales-Rothi, Dr. Richard Briggs, Megan Ga iefsky, Christina Wierenga, Bruce Parkinson, and Keith McGregor for all of their help and support th roughout the past 2 year s. Lastly, I would like to thank my friends, family, and loved ones for their continued support and encouragement. I could not ha ve done this without them.

PAGE 5

v TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iv LIST OF TABLES.............................................................................................................vi LIST OF FIGURES..........................................................................................................vii ABSTRACT.....................................................................................................................vi ii CHAPTER 1 INTRODUCTION......................................................................................................1 Stroke and Aphasia.....................................................................................................1 Intention.....................................................................................................................2 Intention and Language Rehabilitation......................................................................3 Functional Magnetic Resonance Imaging (fMRI).....................................................5 Functional MRI and Aphasia.....................................................................................6 Hypotheses.................................................................................................................9 2 METHODS...............................................................................................................11 Participants...............................................................................................................11 Procedure..................................................................................................................13 Language Rehabilitation.................................................................................13 Image Acquisition...........................................................................................21 Imaging Analyses...........................................................................................23 3 RESULTS.................................................................................................................35 4 DISCUSSION...........................................................................................................44 Conclusions..............................................................................................................44 Limitations...............................................................................................................49 Implications and Future Directions..........................................................................50 LIST OF REFERENCES...................................................................................................52 BIOGRAPHICAL SKETCH.............................................................................................57

PAGE 6

vi LIST OF TABLES Table page 2-1. Participant Demographics........................................................................................27 2-2. Participant Performance on Language-Rel ated Testing Preand Posttreatment.....27 2-3. Participant Baseline Stability and Re sponse to Treatment as Determined by 3 Independent Speech Pathologists.............................................................................27 2-4. Participant C Statistics for Baseline and Treatment.................................................27 3-1. Participant 01’s Representative Activ ity Preand Posttreatment. Regions of interest for our study are indica ted by a broad white outline...................................39 3-2. Participant 02’s Representative Activity Preand Posttreatment............................40 3-3. Participant 03’s Representative Activity Preand Posttreatment............................41 3-4. Participant 04’s Representative Activity Preand Posttreatment............................42 3-5. Participant 05’s Representative Activity Preand Posttreatment............................43

PAGE 7

vii LIST OF FIGURES Figure page 2-1. Participant 01’s Lesion-Sagittal Imag eA broad white outline encompassing the general regions of each participant’s le sion(s) are included in their images............28 2-2. Participant 01’s Lesion-Coronal Image....................................................................28 2-3. Participant 02’s Lesion-Sagittal Image....................................................................28 2-4. Participant 02’s Lesion-Coronal Image....................................................................29 2-5. Participant 03’s Lesion-Sagittal Image....................................................................29 2-6. Participant 03’s Lesion-Coronal Image....................................................................29 2-7. Participant 04’s Lesion-Sagittal Image....................................................................30 2-8. Participant 04’s Lesion-Coronal Image....................................................................30 2-9. Participant 05’s Lesion-Sagittal Image....................................................................30 2-10. Participant 05’s Lesion-Coronal Image....................................................................31 2-11. Formula for the C Statistic.......................................................................................31 2-12. Participant 01’s Performa nce across Treatment Phases...........................................32 2-13. Participant 02’s Performa nce across Treatment Phases...........................................32 2-14. Participant 03’s Performa nce across Treatment Phases...........................................33 2-15. Participant 04’s Performa nce across Treatment Phases...........................................33 2-16. Participant 05’s Performa nce across Treatment Phases...........................................34

PAGE 8

viii Abstract of Thesis Presen ted to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science NEURAL CHANGE IN PATIENTS WITH CHRONIC NONFLUENT APHASIA AFTER LANGUAGE REHABILITATI ON: A FUNCTIONAL MAGNETIC RESONANCE IMAGING (FMRI) STUDY By Bradley J. Daniels May, 2005 Chair: Anna Bacon Moore Major Department: Clini cal and Health Psychology Our study was designed to examine changes in language-related activity in patients with chronic nonfluent aphasia after language rehabilitation, and to further inform the growing debate in aphasia rehabilitation literature on whether improved language function in patients with aphasia involves ac tivation of perilesional regions of the left hemisphere or a reorganization of langua ge function to the language nondominant hemisphere. For our study, 5 patients w ith chronic nonfluent aphasia underwent functional magnetic resonance imaging (fMR I) scanning before and immediately after completing a language rehabilitation program designed to facilitate reorganization of language functions to homologous regions in the nondominant hemisphere. We hypothesized that an increase in activity in homologous langua ge-related medial-frontal and lateral-frontal right-hemisphere regions would be evident. Functional MRI data were acquired on a 3T scanner in an event-related paradigm, while participants completed an overt language task that has been empirically shown to

PAGE 9

ix produce activity in language-re lated regions in normal c ontrols. Deconvolution, a time series analysis of the fMRI signal, was th en conducted to produce a measure of the goodness of fit (R2), of the averaged hemodynamic re sponse for each voxel. Selective detrending was conducted to remove artifact s due to task-correlated motion, and data underwent thresholding to meet criteria for true hemodynamic activity. We also equated for sensitivity to normalize signal-noise differences across multiple fMRI scanning sessions, by deconvolving the fMRI signal from a general-noise test bed. Cluster reports were then run on each participant’s pretreatme nt and posttreatment data to isolate and quantify areas of activity. Changes in activ ity from pretreatment to posttreatment in language-related regions of interest we re analyzed by multiple raters. Results of our study varied across patients, and lent support to both sides of the current debate in aphasia rehabilitation literat ure. We concluded that neural change after language rehabilitation in patients with nonf luent aphasia appears to be primarily influenced by lesion size and location. Ou r findings also suggested that subcortical mechanisms of the left hemisphere may play a role in how the brain reorganizes after language rehabilitation.

PAGE 10

1 CHAPTER 1 INTRODUCTION Stroke and Aphasia A cerebrovascular accident (CVA), or stro ke, occurs when the brain does not receive adequate oxygen because of a disruption in blood supply (e.g., blockage in a vessel or vessel rupture), causing damage to affected brain area s (Snyder & Nussbaum, 1998). Aphasia, a common occurre nce after a stroke, is define d as a partial or total loss of language production (or loss of ability to comprehend spoken or written language) due to disease or brain damage (Broca, 1 861/1997; Finger, 1997). Nonfluent aphasia (impairment in the production of language) is characterized by freque nt pauses, difficulty initiating language, and impaired intonation and grammar; though often with a spared ability to comprehend language. Practical im plications for successful treatments of aphasia are considerable. Approximately 500,00 0 new strokes occur in the United States every year; and as many as 25% of those cas es develop some form of aphasia due to stroke (Nadeau, Rothi, & Crosson, 2000). Currently, an estimated 1,000,000 people living in the United States have aphasia, and most of them developed the impairment as a result of a stroke (National Aphasia Association, 1999). Research and treatments that help to im prove communication abil ity in individuals suffering from aphasia can greatly improve quality of life, d ecrease dependence on others, and decrease the impact on the ove rall economy by helpi ng those individuals successfully return to work (Crosson & Rothi, 1999).

PAGE 11

2 Intention Intention mechanisms are cognitive mechanis ms that help us in our preparation to respond to a given stimulus, and are nece ssary for efficient cognitive processing (Heilman, Watson, & Valenstein, 1993). Language is one cognitive activity that depends upon intentional mechanisms to achieve its goa ls, because intention mechanisms aid in word selection and in the fundamental initi ation of speech (Cross on et al., 2005). One factor that may contribute to difficulty initiating language common in patients with nonfluent aphasia is poor prepar ation to respond (i.e., inade quate intention) (Crosson & Rothi, 1999). It is necessary at this point to describe briefly what is currently known about typical language production. Since the 1860s and Paul Broca’s seminal studies with his patient “Tan,” lesion studies have shown language im pairments after damage to specific lefthemisphere regions (Broca, 1861/1997; Lich theim, 1885). From these studies, Broca (1861/1997), Wernicke (1874), and others develo ped what is commonly referred to as the “classical model” of language, which posits that language function is predominantly housed in the left hemisphere and is divi ded into two major re gions: Broca’s area (housed in the frontal lobe and primarily responsible for the expressive aspects of language), and Wernicke’s area (housed in the posterior regi on of the temporal lobe and responsible for the receptive components of language) (Binder et al., 1997; Lichtheim, 1885; Wernicke, 1874). After the advent of functional neuroimaging, multiple neuroimaging studies (Binder et al., 1997; Foundas, Eure, Luevano, & Weinberger, 1998; Musso et al., 1999) confirmed language func tioning in right-handed individuals to be almost entirely controlled by the left hemi sphere. The primary regions of the left hemisphere thought to be involved in the e xpressive components of language include

PAGE 12

3 Broca’s area, as well as the precentral gyrus (i .e., motor strip), inferior frontal sulcus, and the middle and superior frontal gyri (Binder et al., 1997; Crosson et al., 2001; Foundas et al., 1998; Moore, Crosson, Gockay, Leonar d, & Foundas, 2000; Moore et al., 2001). While these areas are believed to be invol ved in language produc tion, the initiation or intentional component of language is believed to involve the me dial-frontal regions of the left hemisphere, including ci ngulate gyrus and sulcus, paraci ngulate gyrus and sulcus (if present), PreSupplemental Motor Area (Pre -SMA), Supplemental Motor Area (SMA), and the frontal eye fields. (Abdullaev & Po sner, 1998). According to Abdullaev & Posner (1998), normal language begins in le ft-hemisphere medial-f rontal regions and travels outward for approximately 80 ms to th e left lateral-frontal regions for speech production. The middle cerebral artery (MCA) supplies bl ood to the lateralfrontal regions of the brain but not to the medial-frontal regions ; therefore, a left MCA stroke would likely leave left medial-frontal regi ons intact while damaging left lateral-frontal regions. This leads us to believe that nonfluent aphasia is a result of intact medial-f rontal regions of the left hemisphere initiating th e intention of language, and then subsequently attempting to communicate with dysfunctional left lateral-frontal regi ons damaged by the stroke, thereby inhibiting speech output. Intention and Language Rehabilitation Intention mechanisms in the recovery of language function have not been heavily researched to date. Rapid recovery from ak inetic mutism due to unilateral medial-frontal lesion (Damasio & Anderson, 1993) suggests that shifting intention mechanisms to the undamaged hemisphere might be a viable tr eatment strategy in some patients with aphasia. It has also been theorized that at tending to stimuli in he mispace contralateral to

PAGE 13

4 the intact hemisphere or initiating an acti on with the hand contralateral to the intact hemisphere, thereby activating nondominant hemisphere functions engages intention language mechanisms in that hemisphere; which are then able to compensate for dysfunctional language mechanisms in the damaged hemisphere. Previous studies (Hanlon, Brown, & Gerstmann, 1990) showed improvements in naming ability for words learned while gesturing for nonfluent aphasic populations. From these findings, it could be theorized that, since gesturing with the hand contralateral to th e intact hemisphere would cause nondominant right-hemisphere motor regions to become active, simultaneously performing a naming task while gesturing might also then cause nondominant right-hemisphere language regions to become activated. Crosson and Rothi (1999) began developing and applying aphasia treatments specifically designed to recruit intention mechanisms in the intact hemisphe re and to engage thos e regions in language processing. These types of tr eatments are commonly defined as substitutive treatments, meaning that they attempt to reconstruct the impaired language system by substituting an intact mechanism for the damaged one. Rest itutive treatments, by comparison, attempt to reconstitute the impaired system in its orig inal form. Substitutive treatments are more effective than restitutive tr eatments in chronic phases of recovery (Rothi, 1995). From an empirical standpoint, some studies, such as the ones conducted by Code (1982; 1989), actually attempted to apply sim ilar methods to treatments of individual cases. In Code’s patients, stimuli were presen ted to the left visual field, left hand, and/or left ear in an attempt to activate nondominan t-hemisphere language mechanisms. Both of Code’s (1982; 1989) studies reported improvement after a lengthy treatment.

PAGE 14

5 Unfortunately, both studies involved single cases, and allowed for alternative explanations due to the methodology used by the studies (Crosson & Rothi, 1999). Our study was based on a few key assumptions outlined by Crosson & Rothi, (1999): Intention mechanisms support language f unction; and when those regions are damaged or disconnected from language mechanisms, language functions are adversely affected. Alternate intention mechanisms (i.e., in th e intact hemisphere) can be engaged in the service of language functions, impr oving language performance (as shown by Buxbaum, Coslett, Shall, & McNally, 1993). This phenomenon has been used effectively in treatment as suggested by the individual case studies of Buffery & Burton (1982) and Code (1982, 1989). Theref ore, it is believed that some patients with chronic nonfluent aphasia will respond to a treatment designed to activate intention mechanisms in the nondominant hemisphere (Crosson & Rothi, 1999). The goal of our study was to use neuroi maging to reveal changes in neural substrates of language production in patie nts with nonfluent aphasia after their participation in a rehabilitation program. Functional Magnetic Resonance Imaging (fMRI) The foundation of functional neuroimaging th eory is based on the assumption that, as processing demands increase during a task, changes in neural activity also occur in specific brain areas associated with that task (Fiez, 2001). Functional MRI is a noninvasive technique th at vicariously measur es neural activity by measuring changes in the amount of deoxyhemoglobin in the brain ti ssue. When blood flow to a specific area of the brain increases to meet the dema nds of a task, the amount of oxygen-rich hemoglobin supplied to that region also increases; inversely, the amount of deoxyhemoglobin (nonoxygen carrying hemogl obin molecules) decreases (Chen & Ogawa, 2000; Fiez, 2001; Villringer, 2000). Deoxyhemoglobin has certain paramagnetic qualities such that changes in concentration of this com pound are related to changes in

PAGE 15

6 magnetic signal. When red blood cells cont aining deoxyhemoglobin are placed in the magnetic field used for MRI, there is some magnetic field distortion induced by the difference in the magnetic susceptibility relativ e to the surroundings. This susceptibilityinduced field distortion due to the leve l of deoxyhemoglobin is the basis for Blood Oxygen Level Dependent (BOLD) contrast fMRI, and changes in deoxyhemoglobin levels associated with underlying functional ac tivation can be detected in the MRI signal (Chen & Ogawa, 2000). Blood flow changes to a given region of the brain are known as hemodynamic responses. Once a hemodynamic respons e has begun, it will typically peak approximately 6 to 8 seconds afterwards and re turn to baseline leve ls approximately 10 to 12 seconds after initia tion (Fiez, 2001). Functional MRI and Aphasia An ongoing debate in the field of aphasi a rehabilitation involves whether or not improved language function in patients with aphasia involves activat ion of perilesional regions of the left hemisphere (Cao, Vikingstad, George, Johnson & Welch, 1999; Meinzer et al., 2004) or if it involves a reorganization of language function to the language nondominant hemisphere (Abo et al ., 2004; Cappa, 2000; Cr osson et al., 1999; Rosen et al., 2000). Cao et al. (1999) used BOLD fMRI to study cortical language networks in 7 right-handed patients with aphasia during lexicalsemantic processing tasks. All of the patients in the study we re at least 5 months post stroke and had recovered a “substantial” porti on of their language functions since their stroke. The authors found that left-hemisphere perile sional activity was associated with better recovery in their patient population and was inversely related to the amount of activity found in homologous language regions in the right hemisphere (Cao et al., 1999). A

PAGE 16

7 more recent study by Meinzer and colleagues (20 04) studied changes in Delta Dipoles, or focal clusters of slow wave activity in th e delta frequency range (1 to 4 Hz) which are usually located in the vicinity of structur al damage in the brain. Meinzer and his colleagues used magnetoencephalography (MEG ) to study Delta Dipole Density changes in 28 patients with chronic aphasia (>12 mont hs post stroke) before and after intensive speech and language therapy. Results of the st udy showed that decreases in Delta Dipole Density in perilesional regions of the brain were found in a majority of patients. These results emphasize the significance of perilesion al areas in the rehabi litation of aphasia (Meinzer et al., 2004). There are other studies, how ever, that provide contradi ctory findings to those found by Cao et al. (1999) and Mein zer et al. (2004). A study by Musso et al. (1999) used positron emission tomography (PET) to measur e changes in regional cerebral blood flow (rCBF) in 4 patients with Wernicke’s a phasia. Patients underwent a language comprehension task during 12 consecutive s canning sessions. Between sessions, patients participated in brief but intense language comprehension training. All patients improved significantly over the 12 comprehension tasks used during scanning. The authors found that one of the regions that best correlate d with training-induced improvement was the right superior temporal gyrus. From their findings, the authors c oncluded that the right hemisphere plays a significant role in the re covery from aphasia (Musso et al., 1999). Another recent study by Abo et al. (2004) used a word repetition task during fMRI scanning to study language activity in 2 patie nts with aphasia. Functional MRI scans were performed on the patients 60 months post stroke. Results from the study found that both patients showed activation of only the co mpensatory area in the right hemisphere

PAGE 17

8 during the task. It becomes clear from exam ining these studies that current research on this much-debated topic provides support fo r both left-hemisphere perilesional and righthemisphere reorganization-of-function arguments. An alternative, or “middle ground,” explan ation that some studies (including ones previously mentioned) have shown may be that the determination of perilesional activation vs. reorganization to the nondominant hemisphere is mediated by the size and severity of the lesion. In esse nce, the greater the size and seve rity of the lesion in the left hemisphere, the more likely it is for reor ganization of language function over to the nondominant hemisphere to occur (Belin et al., 1996; Cao et al ., 1999; Karbe et al., 1998). More recent studies on this topic are beginning to examin e the role of subcortical structures in predicting how th e brain adapts to injury, and have begun to find predictors of bilateral activation based on whether or not left-hemisphere subcortical structures were affected by the stroke. (Crosson et al., 2005) Using fMRI as a research tool allows scientists the opportunity to attempt to provide answers to these much-debated questions. Functional imaging studies of intenti on mechanisms for language have used primarily word generation tasks (Crosson & Ro thi, 1999). For example, participants are given a letter of the alphabet (e.g., A), and asked to produce as many words as they can think of beginning with that lette r (e.g., apple, actress, aunt), or participants are given a semantic category (e.g., colors) and asked to generate as many exemplars from that category as possible (e.g., red, orange, blue). It has been shown that word generation tasks such as these reliably produce medial -frontal activity in individual controls

PAGE 18

9 (Crosson et al., 1999). Therefore, a word gene ration task designed to track medial-frontal activity preand posttreatment was used in our study. Hypotheses The purpose of our study was to examin e what neuroanato mical patterns of activation are evident in pa tients with chronic nonflu ent aphasia during language generation and over the course of a treatment designed to shift medial-frontal activity from the left to the right hemisphere. Ou r hypotheses were that dominant pretreatment activity would be observed in left medial-fr ontal regions. Right me dial-frontal activity was not expected before initiation of the e xperimental language treatment. We also hypothesized that little, if any, left lateralfrontal activity would be evident in these patients during language production due to left -hemisphere stroke-related damage. Given previous research, we expected that there might be some right lateral-frontal activity during language production in these patients. Another question that our study posed was: among participants enrolled in an experimental language rehabilitation program designed to improve language function by activating intention mechanis ms in the language nondominant hemisphere, what changes in patterns of neural activity were ev ident during language pr oduction posttreatment when compared to pretreatment? One hypothe sis is that participants who responded to treatment would show changes in both righ t medial-frontal and right lateral-frontal homologous language regions (Crosson et al., 1999). The expectation is that righthemisphere medial-frontal regions would show increased activation during language production posttreatment. Predictions about ri ght lateral activity we re not made because it could be that right lateral activity increased posttreatment as a re sult of synergistic coactivation of medial-frontal to lateral-frontal regions. Or, it could be that posttreatment

PAGE 19

10 right lateral-frontal activity would become more efficien t and therefore possibly less distributed as a result of a more efficient “ upstream” activity in the right medial-frontal regions. The practical implications for this type of study are numerous, and include further insight into the neural substr ates of language function in pa tients with chronic nonfluent aphasia and the role that le sion size and location play in reorganization of language functions. Increased knowledge in how these functions work after injury and subsequent rehabilitation can aid in the developmen t of new language reha bilitation treatments designed to maximize treatment outcomes in patients based on the si ze and severity of their individual lesions.

PAGE 20

11 CHAPTER 2 METHODS Participants Ten patients with chronic nonfluent apha sia (6 female, 4 male) with lefthemisphere stroke participated in our study. Due to inherent difficulties associated with fMRI (e.g., scanner problems, excessive motion ar tifact during scanning), data from 5 of these participants (2 male, 3 female; mean age = 55.2 + 9.68 years; mean number of months post stroke = 35.4 + 27.67 months) were available for inclusion in the final analyses presented in our study. Although this may appear to be a relatively high dropout rate, it should be noted that multiple sca nning sessions were re quired for our study, and that the population of intere st included neurologically impaired participants who possessed many deficits (e.g., hemiplegia) that would cause significant discomfort during the extended scanning sessions require d for completion of our study. Inclusion criteria for our study were: Ma les or females over 21 years of age with documented left-hemisphere cerebrovascula r accident (CVA), demonstrated nonfluent aphasia, a minimum of 6 months post str oke, premorbidly right handed, native English speakers, within acceptable limits for height and weight for participation in an fMRI study. Finally, all participants in our study were participants in an experimental language rehabilitation treatment designed to improve language function in pa tients with chronic nonfluent aphasia; this treatment is described in detail later. Partic ipants were excluded from the imaging portion of our study if they were claustroph obic, possessed a history of psychiatric illness, substance abuse, traumatic brain injury, seizures (unrelated to stroke),

PAGE 21

12 dementia, learning disability, neurological diso rders other than stroke, or other criteria that would prevent them from being placed in the fMRI scanner (e.g., being pregnant or having metal inside their body.) For demogra phic information on the participants in our study, see Table 2-1 “Participant 01” ( Figure 2-1 and Figure 2-2 ) was a 48-year old Caucasian male who suffered a left middle cerebral ar tery stroke that primarily affected the left basal ganglia region including the caudate nucleus, globus pallidus, internal capsule, and cerebral peduncle. Participant 01’s stroke also pr oduced an encephalomacic effect that caused damage to the left frontal, te mporal, and parietal lobes. “Participant 02” ( Figure 2-3 and Figure 2-4 ) was a 48-year old Caucasian female who suffered a left middle cerebral artery stroke affecting the frontal, temporal and parietal lobes with some subcortical extensi on upwards into the lenticulostriate endzone. “Participant 03” ( Figure 2-5 and Figure 2-6 ) was a 52-year old Caucasian female who suffered a significant left middle cerebra l artery stroke affecting the frontal, temporal, parietal, and occipital lobes and causing ventriculomegaly. Also, she had undergone a left pari etal craniotomy. “Participant 04” ( Figure 2-7 and Figure 2-8 ) was a 54-year old Caucasian female who suffered a left middle cerebral artery stro ke affecting her frontal and temporal lobes with some midline shift due to the stroke. She had also suffered a previous stroke affecting her left parietal lobe. “Participant 05” ( Figure 2-9 and Figure 2-10 ) was a 74-year old Caucasian male who suffered a left middle cerebral artery stroke affecting the frontal, temporal, and parietal lobes, and extendi ng well into subcortical stru ctures including the caudate

PAGE 22

13 nucleus, internal capsule, insula, cerebral peduncle, and the entire lenticulostriate endzone. Procedure Language Rehabilitation All participants were enro lled in a language rehabilitation treatment designed to improve speech and language function in pa tients with chronic nonfluent aphasia by activating medial-frontal regions with the goal of transferring intention mechanisms from the left-hemisphere medial-fr ontal regions to the right-h emisphere homologue. Although the purpose of our study is to elucidate the neur al substrates of langua ge recovery and not to examine the language treatment study itself, an explanation of the treatment study is necessary because the imaging results occur within the cont ext of a specific language therapy, not across language therap ies in general. The medial-f rontal regions of interest in our study were chosen because previous research involving la nguage rehabilitation populations of patients with a phasia have shown that right lateral-frontal regions show some neural activation during language produc tion in patients with nonfluent aphasia (Belin et al., 1996). It is believed that activation of homo logous medial-frontal regions may facilitate change in right lateral-front al regions as well, thereby producing an increase in speech output due to neural reorganization. The Intention language rehabi litation treatment used in our study was the same as the treatment presented in Richards et al (2002). Briefly, the treatment involved the performance of an overt (out loud) object -naming task while simultaneously producing a meaningless circular gesture with the left hand. The treatment itself consisted of a preliminary baseline phase followed by three ph ases of treatment. To control for the effects of spontaneous recovery, all particip ants enrolled in the language rehabilitation

PAGE 23

14 treatment were at least 6 months post stroke The baseline phase for the study consisted of a minimum of 8 sessions and each of th e three treatment phases consisted of 10 sessions. Each baseline and treatment sess ion for the study lasted approximately fortyfive minutes. After establishing a stable ba seline, progression from phases one through three of the treatment involved transition from movements prompted by external cues (e.g., a tone and a flashing star), to a self-initiated movement sequence. The rationale for differential phases of treatment involved ha ving specific points of demarcation between phases such that the externally guided to internally guided movement progression would allow the participant to eventually learn to pair their movement sequence with the initiation of language. For this prelimin ary study, performance criteria (e.g., minimum percentage of correct res ponses during consecutive treatment sessions) were not established as a rule for transition from one treatment phase to the next; rather, each treatment phase consisted of exactly 10 se ssions. The meaningle ss circular gesture involved in the study was internally guided, gene ralizable to real-world interactions, was not word-related or symbolic in nature, and did not resemble any symbolic action with which the participants would have already been familiar. The same circular gesture was used for all participants. During each of the pretreatment baseline se ssions, participants were seated at a desk directly facing a computer monitor wi th their head and body facing the monitor. The participant then performed a naming ta sk as 40 black-and-white line drawings, approximately 4 inches by 4 inches in size, were displayed on the monitor one at a time. After the presentation of each drawing, the par ticipant named the drawing. For example, if a shoe was presented, the pa rticipant would respond “shoe”. These sessions were used

PAGE 24

15 to establish a baseline rate of percent corre ct of naming accuracy and reaction time. Once a stable baseline was established, the first phase of the Intention treatment would begin. Phase one of the treatment, like the baseli ne sessions, began w ith the participant seated at a desk with their head and body directed towards a computer monitor. The therapist would begin the trial by pressing the mouse bu tton. A 1 inch by 1 inch star would appear on the center of the screen and a 1,000 Hz tone would sound. The color and orientation of the star would vary from tria l to trial. To initiate the presence of the line drawing, each participant w ould lift the lid on a small box located to his or her left with their left hand and press a button inside the box. The button press caused the tone to stop and the star to disappear from the monitor and, after a 2 second delay, a black-andwhite drawing appeared at the center of the monitor and a timer began. If the participant named the picture correctly, the therapist would press the mouse button to end the trial and then stop the timer and remove the line dr awing from the screen. If the participant named the drawing incorrectly, the therapist provided the correct name for the picture while simultaneously making the circular gest ure described above with his or her left hand. The participant was instructed to re peat the corrected pi cture name aloud while also making the same circular gesture. Treatment phases had 50 trials each. Phase two differed from phase one only in that the 1,000 Hz tone was eliminated and a different set of 50 line drawings not pr eviously seen was use d. Incorrect responses were corrected using the same procedure as in phase one of treatment. The same treatment procedures used in phase one and two were used in phase three; however, in this phase of treatment, the flashing star was eliminated and the participant was instructed to perform the same meaningless circular gesture performed in

PAGE 25

16 the previous phases with his or her left hand before the pres entation of the line drawings. Response instructions, and corre ction of incorrect responses, were the same as described above in phases one and two. Once again, a different set of 50 line drawings not previously seen was used. Two overall sets of line drawings were used in the treatment study: one balanced set with both highand low-frequency words, and one set for higher-functioning participants comprised of fifty low-frequency words. The balanced set of line drawings contained 15 high frequency words (27717 occurrences per million), 15 medium frequency words (4-10 occurrences per million), and 20 low frequency words (0-30 occurrences per million), to provide a balan ced set of words and prevent participants from obtaining ceiling effects during treatment. The low frequency set of line drawings contained all low frequency words. Freque ncies of words were based on Francis and Kucera’s “Frequency Analysis of the Englis h Language” (Francis & Kucera, 1982). The determination for whether a participant was considered high functioning and given the low frequency set of line drawings or low f unctioning and given the balanced set of line drawings for the language rehabilitation treatment wa s based upon their overall performance on a set of naming probes from the ba lanced picture set. In other words, if a participant performed on average greater than 70% on the balanced picture set over 4 to 8 naming probes, they were grouped in the highest functioning group and given the low frequency picture set. All participants scor ing below 70% received the balanced set of line drawings. To determine the stability of baseline performance, and overall response to the treatment itself, an independent panel of 3 speech pathologists were recruited to rate each

PAGE 26

17 subject’s data based upon the graphs displaye d in Figures 2-11 through 2-15. Each rater examined the graphs for each participant, wh ich showed the percentage of line drawings named correctly throughout each baseline and treatment session and, using visual inspection, judged each participant for stab ility throughout the baseline sessions and overall response to treatment. Visual insp ection refers to the ma king of judgments of reliability or consistency of intervention eff ects through the use of visual examination of graphed data (Kazdin, 1982). The primar y difference between traditional betweensubjects group research and single-case re search designs is th at, in between-group research, the experimental criterion is met by comparing performance between or within groups though the use of statis tics, whereas, in single-case research, the experimental criterion is met by examining the effects of the treatment at different points over time (Kazdin, 1982). The method of visual insp ection of data to determine performance throughout treatment is considered the most common technique used in single-case research designs (Elder, 1997; Kazdin, 1982), and has been determined by some studies to be highly correlated with single-subject statistical procedures when evaluating treatment effects (Bobrovitz & Ottenbacher, 1998). The C statistic ( Figure 2-11 ) was also used to provide a quantitative measure of baseline stability and treatment perfor mance (Tryon, 1982). The C statistic is a “simplified” time-series analysis that can be used to measure the effect of treatment interventions on studies that ha ve as few as 8 points of data per experimental phase by examining upward trends in performance across time (Tryon, 1982). The numerator of the right-hand term of th e C statistic is the sum of N – 1 squared consecutive differences associated with the time series. The denominator is twice the

PAGE 27

18 sum of N squared deviations of the time-seri es data points from their mean. The final determination for significance using the C sta tistic is calculated by dividing C in the equation above over the standard error of the C statistic, whic h is calculated by taking the square root of N + 2 over (N-1) (N+1) where N equals the number of data points in the time series (Tryon, 1982). If a participant’s baseline sessions did not reach statistical significance (p>.05), it was assumed by the C statistic that they possessed a stable base line. Then, if their overall data from the 3 phases of treatment was consid ered statistically significant (p<.05), they were deemed by the C statistic to be a treatment responder. In single-case research designs, statis tical tests to determine performance throughout treatment are occasionally used; how ever, this practice remains the exception rather than the rule (Kazdin, 1982). Accord ing to Tryon (1982), greater confidence can be placed in scientific data wh en both visual and statistical analysis procedures agree. Studies have shown, however, that the results of single-subject rese arch designs can be directly influenced by the method of statistica l data analysis selected and that multiple single-subject statistical test s-including the C statistic-show relatively low levels of agreement between tests (Nour bakhsh & Ottenbacher, 1994). Al so, it is unclear whether the C statistic is useful with participants w ho display floor effects as the C statistic may become artificial inflated to significance leve l due to low standard error. Therefore, whenever there was a discrepancy between a participant’s ratings of performance by the independent panel of speech pathologists a nd their ratings of perf ormance based on their C statistics, the independent panel was used as the final judgment for our study.

PAGE 28

19 Each participant’s scores on specific language-re lated tests that were administered preand posttreatment, their ratings by an indepe ndent panel of 3 speech pathologists, and their C statistics are displayed in Table 2-2 Table 2-3 and Table 2-4 Participant 01 ( Figure 2-12 ) was judged as having a stable baseline and was considered a treatment responder as judge d by an independent panel of 3 speech pathologists. Participant 01 wa s rated as having a stable baseline and was considered a treatment responder according to his C statisti cs as well. Participant 01 was considered high functioning for our study and so rece ived the low frequency stimuli. Participant 02 ( Figure 2-13 ) was judged to have a st able baseline by 2 out of 3 raters. All raters consider ed her to be a treatment responder. According to her C statistics, she also had a stable baseline and was considered to be a treatment responder. Participant 02 was considered high functioni ng for our study as well and so received the low frequency stimuli. Participant 03 ( Figure 2-14 ), like Participants 01 a nd 02, was judged as having a stable baseline and as responding to treatment by all 3 independent raters. However, it is likely that Participant 03’s baseline stability is best explained by floor effects, as Participant 03 was unable to correctly name any of the 40 presented line drawings during 6 of her 9 baseline sessions. Nevertheless, Participant 03 did begin to show visible improvements in naming ability by Phase two of treatment. Based on her C statistics, Participant 03 was considered to be a treatm ent responder, but was not considered to have a stable baseline. This too, is likely due to her floor effects and the C statistic’s artificial inflation to significance level of her perfor mance throughout the baseline sessions due to the low standard error. This subject shows the limitations of the C-statistic and provides

PAGE 29

20 and example of a situation in which visual inspection by independe nt raters provides a more accurate assessment of st ability versus change. Part icipant 03 was considered low functioning for our study and so received the balanced set of line drawings. Though Participant 04 ( Figure 2-15 ) was judged as possessin g a stable baseline, this judgment is also likely best explained by floor effects. She also did not appear to respond to treatment as determined by the indepe ndent panel of 3 raters. Participant 04’s C statistics were considered stable for her baseline sessions, as was her response to treatment. Recall that a stable response duri ng the treatment phase is consistent with a treatment nonresponder. Participant 04’s poor tr eatment outcome may be in part due to the overall severity of her aphasia. She wa s considered low functioning for our study and therefore received the balan ced set of line drawings. Participant 05 ( Figure 2-16 ) was judged as having an unstable performance throughout the baseline sessions by all 3 raters However, 2 out of 3 raters deemed Participant 05 to be a treatment responder. It is possible that Pa rticipant 05’s unstable baseline may reflect possible continued recovery from stroke However, given that most effects of spontaneous recovery are observed within the first six months after stroke, and that Participant 05’s stroke occurred appr oximately eighty-six months before the initiation of treatment, spontaneous rec overy of language functioning is unlikely. According to his C statistics, Participant 05 a ppeared to have a stable baseline, but was not considered to be a treatme nt responder. Figure 2-10 s uggests that Pa rticipant 05’s performance during treatment appeared to have an upward trend; however, this trend was not statistically significant based on his C statistic. Participant 05 was considered high functioning for our study and so received the low frequency set of line drawings.

PAGE 30

21 Image Acquisition Before beginning the Intention language re habilitation treatmen t, each participant underwent a pretreatment fMRI scan. Partic ipants also underwent a posttreatment scan within 2 weeks of their completion of the treatment. All participant scans were conducted with a 3 Tesla fMRI scanner usi ng a dome-shaped RF quadrature head coil (MRI devices). Functional imaging parameters for all fMRI scans conducted in our study were as follows: single shot spiral scan, gradient echo pulse sequence, TE = 18 ms, TR = 1660 ms, FA = 60 degrees, FOV = 200 mm, matrix = 64 x 64, 32 slices with whole brain coverage, and slice thickness = 4 mm. Struct ural imaging parameters for all fMRI scans conducted in our study were as follows: 3D spoiled GRASS sequence, TE = 6 ms, TR = 23 ms, FOV = 240 mm, matrix size = 256 x 192 and slice thickness = 1.3 mm. An event-related paradigm, which captures single hemodynamic responses by alternating between active stim uli and long interstimulus in tervals (ISI’s), was used during scanning. This was done so that each hemodynamic response could run its course and return to baseline before the next stimul us was presented. This paradigm was chosen over a blocked paradigm as it a llows researchers greater flex ibility and the opportunity to examine overt response patterns, adherence to the required task, and behavioral changes during scanning. In contrast, a block pa radigm attempts to capture a stronger hemodynamic response by alternating between pe riods of multiple “active” stimuli with very short interstimulus interv als, followed by a long period of rest in between blocks. (Fiez, 2001). This produces multiple hemodynamic responses, subsequently producing an additive effect and yielding a stronger activation due to insufficient time for the hemodynamic responses to return to baseline before the next stimulus is presented.

PAGE 31

22 Participants were asked to perform a verb al fluency task requiring the production of a category exemplar during both preand postt reatment scanning sessions. For example, if participants heard the category “types of fish,” they might respond with “shark” or “piranha.” This task was modeled after the word generation task used in Crosson, Sadek, Maron, et al. (2001). Participan ts practiced this task with a clinician prior to their placement inside the scanner to ensure that they fully understood the task. Participants were instructed to give only one response for each category and were reminded to remain completely still during scanning to minimize motion artifact. Participants were also instructed to relax and await the presentation of the next ca tegory after each response. If the participant was unable to hear or unders tand the name of a category during scanning, they were instructed to respond “what,” to prevent the later c oding of a response as “incorrect” or “other” when stimuli we re inaudible or uninterpretable. Category exemplars were presented to part icipants via a magnacoustic digital audio system and nonmagnetic headset with mi crophone. Sound attenuation processes were performed before the functional scans to ensure that participants could hear the stimuli. Five stimulus runs, each containing 9 categor ies with variable rest intervals between stimuli, were used, for a tota l of 45 stimuli per scan. Inte rstimulus intervals (ISI’s) of 21.58 seconds, 23.24 seconds, 24.9 seconds, or 26.56 seconds were pseudorandomly interspersed between categor ies throughout each run. These variable interstimulus intervals between category presentations we re selected based on mean response times outside of the scanner as determined by a se parate but related p ilot study and allowed: Adequate time for the participant’s hem odynamic response, estimated to last as long as twenty seconds, to re turn to baseline level. Variability between intervals required for statistical deconvolution to take place.

PAGE 32

23 Overt responses were recorded directly to a laptop computer via the magnacoustic microphone. All participants received the same stim uli during both preand posttreatment scans, and all runs were presente d in a randomized order across the 2 scans. Participants were debriefed after the comp letion of the scan and were offered an opportunity to view anatomical images of their brain. Imaging Analyses Many different techniques have been used to evaluate the results of fMRI. However, no standardized way of quantif ying, and subsequently qualifying, functional activity has been developed to date. One topic of controvers y in fMRI research involves the collapsing of data across participants. By collapsing data acro ss participants, activity voxels and clusters are more eas ily localized and evaluated; however, when working with stroke patients, each patient’s size and seve rity of lesion will vary significantly. Therefore, collapsing data acro ss these types of patients can corrupt the investigation and understanding of which structures may still be functional in each individual patient. Consequently, data obtained for each participan t in our study were analyzed using withinsubjects analyses. A commercial software package (Cool Edit 2000TM) was used to record participant responses during scanning directly to a lapt op computer. As a consequence of this method, scanner noise was also recorded dur ing participants’ responses. Specialized tools used by Cool Edit 2000TM were then used to reduce the amount of scanner noise in the recorded responses so th at participant responses coul d be heard and analyzed. Participant responses were then coded as co rrect, incorrect/other, or no response (NR). Recorded responses were also analyzed to determine the precise time at which the participant’s response was initiated. By determining the exact time at which each

PAGE 33

24 response was initiated, the image acquisition number for each response was able to be determined. Imaging data for this particular study were analyzed using a paradigm that incorporated all responses given by participan ts, meaning that images were included in analysis for each time the participant responded to a stimulus, regardless of whether or not their actual response was coded as “correc t” for the given category. The reasons an all-response paradigm was used for our study as opposed to a response paradigm analyzing images from only res ponses coded as “correct” were: To allow sufficient responses for analysis. Effort to respond alone would show brain processes of interest for our study. A number of specialized statistical pr ocedures were conducted by qualified lab personnel to prepare each participant’s images for analysis First, the images obtained from each participant’s scans were analyzed using Analysis of Functional Neuroimages (AFNI) software (Cox, 1996) to derive functio nal maps based on response type. Second, deconvolution, a time series an alysis of the fMRI signal, was conducted. For this analysis, specific time intervals for which a hemodynamic response was expected were designated. Deconvolution produces a measure of the goodness of fit (R2), of the averaged hemodynamic response for each voxel. Once deconvolution was completed, selective detrending was conducted for the remo val of artifacts due to task-correlated motion (e.g., speaking). It is expected that task-corre lated motion will occur in approximately the first 5 seconds (corres ponding to approximately 3 images) of the hemodynamic response. Since the hemodyna mic response can last up to 20 seconds, specific measurable signals related to task-c orrelated motion were detected and removed from analysis. A thresholding procedure was then used to exclude activated voxels that did not fit the expected amp litude of change for true hemodynamic activity (e.g., percent

PAGE 34

25 change > 8% for our study). For example, dr aining veins like the s uperior sagittal sinus often produce changes in the Blood Oxygen Level Dependent (BOLD) signal and present as activity on fMRI scans; however, most of these activated voxels are removed through thresholding. After thresholding was comple te, equating for sensitivity was conducted to normalize signal-noise differences across multiple fMRI scanning sessions by deconvolving the fMRI signal from a generalnoise test bed. Smoothing of the images was not performed. Finally, each participant’s f unctional and anatomical images were converted to 1 mm3 voxels and deformed into atlas space (Talairach & Tournoux, 1988) to normalize each brain to a standard size and orientation. This process uses 10 landmark points to fit brains into atlas space. Thes e points include: the midline pos terior, superior margin of the anterior commissure (AC); the inferior margin of the posterior commissure (PC); 2 mid-sagittal points in the interhemispheric fi ssure; the left-most and right-most points in the brain; the most superior and inferior points in the brai n, and the most anterior and posterior points in the brain. Once these pr eliminary procedures we re completed, cluster analyses were performed. Left-hemisphere lateral-frontal regions of interest for our study were as follows: left precentral gyrus (Brodma nn’s area [BA] 4) and left pe rilesional regions (BA’s 6, 8, 9, 44, and 45). Left-hemisphere medial-frontal regions of interest were: left Pre-SMA (medial BA 6), left SMA (medial BA 4), the left paracingulate gyrus (if present) coupled with the top half of the left cingulate sulcus (BA 32), and th e left cingulate gyrus coupled with the bottom half of th e left cingulate sulcus (B A 24). Left cingulate and

PAGE 35

26 paracingulate sulci were also labeled individually if the distinction betw een BA 32 and 24 was unclear. Right-hemisphere lateral-frontal regions of interest were: right precentral gyrus (BA4), right Broca’s homologue, encompassi ng Pars Opercularis and Pars Triangularis (BA’s 44 and 45, respectively), regions of right inferior frontal gyrus outside of Broca’s homologue, right inferior frontal sulcus, right middle frontal gyrus, right superior frontal gyrus, and right superior frontal sulcus (all of which comp rise various portions of BA’s 6, 8, and 9). Right-hemisphere medial-frontal re gions of interest were: right Pre-SMA (medial BA 6), right SMA (medial BA 4), th e right paracingulate gyrus (if present) coupled with the top half of the right cingul ate sulcus (BA 32), and the right cingulate gyrus coupled with the bottom ha lf of the right cingulate sulc us (BA 24). Right cingulate and paracingulate sulci were also labeled individually if the distinction between BA 32 and 24 was unclear. Only activated voxels with an R2 > .16 and a resulting cl uster volume of greater than 100 microliters were captured for analysis. Each participant’s im ages were localized by multiple raters to designate activity clusters in the selected regions of interest for our study. Resulting cluster volumes for prea nd posttreatment activation in the selected regions of interest were then compared.

PAGE 36

27 Table 2-1. Participant Demographics Participant Age Sex Education (in years) Months Post stroke 01 48 Male 13 48 02 48 Female 14 8 03 52 Female 18 15 04 54 Female 14 24 05 74 Male 12 86 Table 2-2. Participant Performance on Language -Related Testing Preand Posttreatment Participant WAB AQ Pre-Tx WAB AQ Post-Tx WAB Comp Pre-Tx WAB Comp Post -Tx BNT Pre-Tx BNT Post-Tx 01 79.6/100 84.6/100 172/200 196/200 50/60 51/60 02 81.4/100 84.1/100 196/200 187/200 44/60 44/60 03 27/100 33.4/100 98/200 111/200 1/60 1/60 04 32.3/100 39.2/100 105/200 140/200 1/60 1/60 05 68.8/100 77/100 154/200 162/200 39/60 47/60 Table 2-3. Participant Baseline Stability a nd Response to Treatment as Determined by 3 Independent Speech Pathologists Participant-Phase Rater 1 Rater 2 Rater 3 Interrater Agreement 01-Baseline Stable Stable Stable 100% 01-Treatment Responder Responder Responder 100% 02-Baseline Stable Stable Unstable 66.66% 02-Treatment Responder Responder Responder 100% 03-Baseline Stable Stable Stable 100% 03-Treatment Responder Responder Responder 100% 04-Baseline Stable Stable Stable 100% 04-Treatment NonResponder NonResponder NonResponder 100% 05-Baseline Unstable Unstable Unstable 100% 05-Treatment Responder Responder NonResponder 66.66% Table 2-4. Participant C Statistics for Baseline and Treatment Participant C-Statistic Baseline Stable Baseline C-Statistic Treatment Treatment Responder 01 .1111 (p>.05) Yes .5523 (p<.01) Yes 02 .0871 (p>.05) Yes .8422 (p<.01) Yes 03 -1.2209 (p<.01)No .8039 (p<.01) Yes 04 -.125 (p>.05) Yes .0205 (p>.05) No 05 .9625 (p>.05) Yes .0233 (p>.05)) No

PAGE 37

28 Figure 2-1. Participant 01’s Lesion-Sagitta l ImageA broad white outline encompassing the general regions of each participant’s lesion(s) are included in their images. Figure 2-2. Participant 01’s Lesion-Coronal Image Figure 2-3. Participant 02 ’s Lesion-Sagittal Image

PAGE 38

29 Figure 2-4. Participant 02’s Lesion-Coronal Image Figure 2-5. Participant 03’s Lesion-Sagittal Image Figure 2-6. Participant 03’s Lesion-Coronal Image

PAGE 39

30 Figure 2-7. Participant 04 ’s Lesion-Sagittal Image Figure 2-8. Participant 04’s Lesion-Coronal Image Figure 2-9. Participant 05 ’s Lesion-Sagittal Image

PAGE 40

31 Figure 2-10. Participant 05’s Lesion-Coronal Image Figure 2-11. Formula for the C Statistic

PAGE 41

32 0 10 20 30 40 50 60 70 80 90 100 123456789123456789101234567891012345678910 Phase and DayPercentage Correct BASELINEPHASE 1PHASE 2PHASE 3 Figure 2-12. Participant 01’s Perf ormance across Treatment Phases 0 10 20 30 40 50 60 70 80 90 100 123456789123456789101234567891012345678910 Phase and DayPercentage Correct BASELINEPHASE 1PHASE 2PHASE 3 Figure 2-13. Participant 02’s Perf ormance across Treatment Phases

PAGE 42

33 0 10 20 30 40 50 60 70 80 90 100 123456789123456789101234567891012345678910 PHASE AND DAYPERCENTAGE CORRECT BASELINEPHASE 1PHASE 2PHASE 3 Figure 2-14. Participant 03’s Perf ormance across Treatment Phases 0 10 20 30 40 50 60 70 80 90 100 123456789123456789101234567891012345678910 Phase and DayPercentage Correct BASELINEPHASE 1 PHASE 2PHASE 3 Figure 2-15. Participant 04’s Perf ormance across Treatment Phases

PAGE 43

34 0 10 20 30 40 50 60 70 80 90 100 1234567891234567891012345678910123456789 Phase and DayPercentage Correct BASELINEPHASE 1PHASE 2PHASE 3 Figure 2-16. Participant 05’s Perf ormance across Treatment Phases

PAGE 44

35 CHAPTER 3 RESULTS Activated cluster voxels in th e regions of interests for our study were calculated for each participant’s pretreatment and posttreatme nt scans. Though changes in individual regions were calculated, the overall goal of our study was to examine global changes in medial-frontal and lateral-frontal regions of the left and right hemisphere after treatment. In addition, there are data to show that larger-scale regions of interest produce greater reliability (Sabsevitz et al., 2003). Therefore, activity totals for each overall region were also calculated and are presented along with images of each participant’s activity in Tables 3-1 through 3-5. Participant 01 ( Table 3-1 ) made 39 out of a possible 45 responses during the pretreatment scan; therefore, 39 responses we re entered into deconvol ution analysis. The maximum R2 value for Participant 01’s pretreatme nt data was .57. Overall, Participant 01’s pretreatment data showed bilateral activity in the lateral-frontal regions, with a slight favoring of left perilesional regions (9,907 activated voxels) over ri ght lateral-frontal regions (6,105 activated voxels). A small am ount of left medial-frontal activity was observed (334 activated voxels), whereas no ri ght medial regions showed significant activity pretreatment. Posttreatment, Participant 01 made 45 out of a possible 45 respons es; therefore, all 45 responses were entered into deco nvolution analysis. The maximum R2 value for Participant 01’s posttreatment data was .58. Participant 01 continued to show bilateral activation in lateral-frontal regions posttr eatment; however, greater activity was observed

PAGE 45

36 in right lateral-frontal regions (8,806 activated vo xels) than left perilesional regions (4,617 activated voxels) posttreatment There was little change in medial-frontal activity in either the left (220 activa ted voxels) hemisphere or the ri ght hemisphere (0 activated voxels) from pretreatment to posttreatment. Pretreatment, Participant 02 ( Table 3-2 ) made 44 out of a possible 45 responses; therefore, 44 responses were entered in to deconvolution analysis. The maximum R2 value for Participant 02’s pretreatment data was .44. Participant 02 showed significantly greater activity in ri ght lateral-frontal ( 2,041 activated voxels) and right medial-frontal regions (1,455 activated voxels) th an in left perilesional (236 activated voxels) and left medial-frontal (523 activated voxels) regions pretreatment. Participant 02 made 45 out of 45 resp onses during her posttreatment scan; therefore, all 45 responses were included in the deconvolution analysis Participant 02’s maximum R2 value posttreatment was .40. Posttreat ment, Participant 02 showed bilateral activations, though she primarily showed activ ity in left-hemisphere regions, with significantly increased activity in left per ilesional regions (3,031 activated voxels). Participant 02 also showed an increase in left medial-frontal ac tivity (1,882 activated voxels). Participants 02’s right lateral activity remain ed relatively unchanged (2,084 activated voxels) from pretreatment to pos ttreatment, whereas her right medial-frontal activity decreased si gnificantly (600 activ ated voxels). Participant 03 ( Table 3-3 ) made 34 out of 45 responses during her posttreatment scan; however, 7 responses had to be rem oved from analysis when scanner problems occurred during the particular run in whic h those responses were made; therefore, 27 responses were included in the pretreatment deconvolution analysis. Participant 03’s

PAGE 46

37 maximum R2 value pretreatment was .55. Partic ipant 03 showed no medial-frontal activity in either hemisphere during her pr etreatment scan. She showed some left perilesional activity (1,615 activated voxels); however, she showed an unusually large amount of right lateral-frontal activity (125,305 activated voxel s), the vast majority of which (123,734) was found in the right motor st rip. Investigation of this unusually large cluster of activity revealed what appeared to be true hemodynamic activity; therefore, the cluster was included in analysis. Participant 03 made 45 out of 45 responses during her posttreatment scan; however, for consistency, the same run that had to be deleted from analysis during Participant 03’s pretreatment scan was deleted from her posttre atment scan. Therefore, only 36 responses were included in Participant 03’s deconvoluti on analysis posttreatment Participant 03’s maximum R2 value posttreatment was .33. Partic ipant 03 showed a significant decrease in overall activity compared to her pretreatment scan. Activ ity was evidenced only in the right hemisphere posttreatment. Some ri ght medial-frontal activity was evident posttreatment (150 activated voxe ls), whereas none was evid ent pretreatment. Right lateral-frontal activity decrea sed significantly posttreatme nt (1,455 activated voxels). Participant 04 ( Table 3-4 ) made 44 out of 45 responses during her pretreatment scan; therefore, 44 responses were included in her pretreatment deconvolution analysis. Participant 04’s maximum R2 value pretreatment was .40. Participant 04 showed very little overall activity during he r pretreatment scan. In fact, only Participant 04’s left medial-frontal regions (295 activated voxe ls) showed significant activity. Participant 04 made 45 out of 45 resp onses during her posttreatment scan; therefore, all responses were included in her posttreatment deconvolution analysis.

PAGE 47

38 Participant 04’s maximum R2 value posttreatment was also .40. Overall, Participant 04 showed evidence of some bilateral activation in her lateral-frontal regions posttreatment with more activation evident in left perilesional regions ( 3,186 activated v oxels) than in right lateral-frontal regions (1,755 activated voxels). Participant 04 also showed an increase in left medial-front al activity (1,607 activated voxels) compared to pretreatment. Consistent with her pretreatment scan, no right medial-frontal activity was observed posttreatment. Pretreatment, Participant 05 ( Table 3-5 ) made 43 out of 45 re sponses; therefore, 43 responses were entered into dec onvolution analysis. The maximum R2 value for Participant 05’s pretreatment data was .32. Pa rticipant 05 showed bilateral medial-frontal activations pretreatment (122 activated voxels left-hemisphere; 290 activated voxels right hemisphere). However, Participant 05 show ed no significant left perilesional or right lateral-frontal activation pretreatment. Participant 05 responded 36 out of 45 times posttreatment; therefore, 36 responses were entered into Participan t 05’s posttreatment deconvolu tion analysis. The maximum R2 value for Participant 05’s posttreatment da ta was .30. Participant 05 showed only left medial-frontal activity (1 30 activated voxels) during hi s posttreatment scan.

PAGE 48

39 Table 3-1. Participant 01’s Re presentative Activity Preand Posttreatment. Regions of interest for our study are indicate d by a broad white outline. Left Lateral-frontal Pretreatment 9,907 Activated Voxels Left Lateral-frontal Posttreatment 4,617 Activated Voxels Left Medial-frontal Pretreatment 334 Activated Voxels Left Medial-frontal Posttreatment 220 Activated Voxels Right Lateral-frontal Pretreatment 6,105 Activated Voxels Right Lateral-frontal Posttreatment 8,106 Activated Voxels Right Medial-frontal Pretreatment 0 Activated Voxels Right Medial-frontal Posttreatment 0 Activated Voxels

PAGE 49

40 Table 3-2. Participant 02’s Representa tive Activity Preand Posttreatment Left Lateral-frontal Pretreatment 236 Activated Voxels Left Lateral-frontal Posttreatment 3,031 Activated Voxels Left Medial-frontal Pretreatment 523 Activated Voxels Left Medial-frontal Posttreatment 1,882 Activated Voxels Right Lateral-frontal Pretreatment 2,041 Activated Voxels Right Lateral-frontal Posttreatment 2,084 Activated Voxels Right Medial-frontal Pretreatment 1,455 Activated Voxels Right Medial-frontal Posttreatment 600 Activated Voxels

PAGE 50

41 Table 3-3. Participant 03’s Representa tive Activity Preand Posttreatment Left Lateral-frontal Pretreatment 1,615 Activated Voxels Left Lateral-frontal Posttreatment 0 Activated Voxels Left Medial-frontal Pretreatment 0 Activated Voxels Left Medial-frontal Posttreatment 0 Activated Voxels Right Lateral-frontal Pretreatment 125,305 Activated Voxels Right Lateral-frontal Posttreatment 1,455 Activated Voxels Right Medial-frontal Pretreatment 0 Activated Voxels Right Medial-frontal Posttreatment 150 Activated Voxels

PAGE 51

42 Table 3-4. Participant 04’s Representa tive Activity Preand Posttreatment Left Lateral-frontal Pretreatment 0 Activated Voxels Left Lateral-frontal Posttreatment 3,186 Activated Voxels Left Medial-frontal Pretreatment 295 Activated Voxels Left Medial-frontal Posttreatment 1,607 Activated Voxels Right Lateral-frontal Pretreatment 0 Activated Voxels Right Lateral-frontal Posttreatment 1,755 Activated Voxels Right Medial-frontal Pretreatment 0 Activated Voxels Right Medial-frontal Posttreatment 0 Activated Voxels

PAGE 52

43 Table 3-5. Participant 05’s Representa tive Activity Preand Posttreatment Left Lateral-frontal Pretreatment 0 Activated Voxels Left Lateral-frontal Posttreatment 0 Activated Voxels Left Medial-frontal Pretreatment 122 Activated Voxels Left Medial-frontal Posttreatment 130 Activated Voxels Right Lateral-frontal Pretreatment 0 Activated Voxels Right Lateral-frontal Posttreatment 0 Activated Voxels Right Medial-frontal Pretreatment 290 Activated Voxels Right Medial-frontal Posttreatment 0 Activated Voxels

PAGE 53

44 CHAPTER 4 DISCUSSION A word generation task requiring the overt production of a category exemplar was used during fMRI scanning to reveal change s in neural substrates associated with language production in five patients with chronic nonfluent aphasia before and after undergoing an experimental langua ge rehabilitation program de signed to recruit intention mechanisms in the medial-frontal regions of the language nondominant hemisphere, thereby improving language function. It wa s hypothesized that changes in language abilities posttreatment would be shown by ch anges in functional activation in the right medial-frontal and lateral-fr ontal language homologous regions of the brain. It was hoped that the results of our study would also he lp to inform the current debate in aphasia rehabilitation literature as to whether or not improvements in language function in patients with aphasia are associated with act ivation or perilesional regions of the lefthemisphere or if they involve reorgani zation of language function to the language nondominant right hemisphere. Conclusions The hypotheses for our study were not fully supported, and resu lts varied from patient-to-patient such that some patients (P articipants 02 and 05) generally lent support to the argument that improved language func tion in patients with nonfluent aphasia is associated with activation in perilesional re gions of the left-hem isphere, while others (Participants 01 and 03) supporte d the argument that reorgani zation of language function over to the language nondominant right hemi sphere takes place. A summary of each

PAGE 54

45 participant’s overall changes in activity from pretreatment to posttreatment, and a brief discussion of possible explanations for these findings, is discussed below. Participant 01 was deemed a treatment responder by all 3 independent raters, and his C statistic. This participant showed b ilateral activation of lateral-frontal language regions pretreatment. Posttreatment, Part icipant 01 continued to show bilateral activations; however, a general shift in overall activation from left-hemisphere perilesional regions to right hemisphere language homologous regions did take place. These findings generally lend support to the re organization of function theory (Abo et al., 2004; Musso et al, 1999); however, due to the heavily subcortical na ture of Participant 01’s lesion, these results are most consistent with the findings of Crosson et al. (2005), which suggest that bilatera l activations may occur if left-hemisphere subcortical structures are affected by the stroke. Participant 02 was considered to be a tr eatment responder by all 3 raters and her C statistic as well. She showed relativel y dominant right-hemisphere language activity during her pretreatment scan. Posttreatment Participant 02, like Participant 01, showed more bilateral activation; however, greater activity was observed in the left-hemisphere perilesional regions than in the right hemisphere posttreatme nt. Participant 02 also had some subcortical damage from her lesion, and her observed bilateral activations are also consistent with those found in Crosson et al. (2005). However, her general tendency towards left-hemisphere activation posttreatment is somewhat contradictory to the righthemisphere favoring of Participant 01 and mi ght be explained by the less severe damage to left-hemisphere subcortical structures in Participant 02.

PAGE 55

46 Participant 03, like Partic ipants 01 and 02, was deemed a treatment responder by both the independent panel of raters and he r C statistic. Pretreatment, Participant 03 showed primarily right-hemisphere language ac tivity with some left lateral-frontal activity. Posttreatment, she showed only righ t-hemisphere language activity. Pa rticipant 03’s findings generally lend support to the argum ents of Musso et al. (1999) and Abo et al. (2004) that improvements in language function are associat ed with reorganization of function over to the right hemisphere. Ba sed on the size of Participant 03’s lesion, however, it could be argued that reorganizati on to the right hemisphere occurred simply because there were insufficie nt perilesional area s left intact by her stroke. These arguments would be in line with th e findings of Cao et al. (1999). Participant 04 was not considered to be a treatment responder by all 3 raters nor by her C statistic. Participant 04 showed no ri ght-hemisphere activation pretreatment. She did show some left-hemisphere medial-front al activation, however. Posttreatment, Participant 04 showed primarily left-hemis phere perilesional ac tivations, though she did show some right lateral-frontal activation as well. In genera l, her results are consistent with the findings of Meinzer et al. (2004) and Cao et al. ( 1999) that suggest perilesional activation is seen after rehabi litation in patients with aphasi a; however, it should be noted that, since Participant 04 was not considered to be a treatment responder, her results may not be generalizable to those who did respond to treatment. Participant 05 was considered to be a treatment responder by 2 out of 3 independent raters; however, Participant 05 was not considered to be a treatment responder by one rater nor by his C statistic. Participant 05 showed some left and righthemisphere medial-frontal activity pretreatme nt. Posttreatment, however, only a slight

PAGE 56

47 increase in left-hemisphere medial-front al activity was observed. Participant 05’s findings, like those of Partic ipant 04, generally lend s upport to the arguments for perilesional activation but may also not be ge neralizable as it is possible that he did not possess a stable baseline before th e initiation of treatment. Based on the results of our study, it is conc luded that the most important factor in determining whether changes in activation afte r language rehabilitation occur in perilesional regions of the left-hemisphere or right-he misphere language homologous regions appears to be the size, location, and overall severity of the lesion itself. These findings seem most consistent with the work of Cao et al. (1999). It is hypothesized from these findings that the brain may in fact follow a general hierarchical pattern of reorganization after brain injury. This theo ry posits that, if sufficient left-hemisphere perilesional regions remain in tact after stroke, a general increase in activity will be observed in these perilesional regions af ter improvements in language functioning. However, if the size of the individual’s lesion is too significant, or there are insufficient perilesional regions availa ble to resume language func tioning, a shift of language functions to right-hemisphere homologous la nguage regions will be observed. Finally, consistent with the findings of Crosson et al. (2005), if left-hemisphere subcortical regions are affected by the stroke, bilateral activ ations of language re gions may in fact be observed. Another particular topic of interest rega rding the results of our study involves the issue of sexand age-related differences in lateralization. Recent functional neuroimaging studies on lateralization have le d to the development of the Hemispheric Asymmetry Reduction in Older Adults (HAR OLD) hypothesis due to general findings

PAGE 57

48 that brain activity in indivi duals tends to become less lateralized as age increases (Cabeza, 2002; Reuter-Lorenz et al., 2000). Using the HAROLD hypothesis, it could be theorized that older participants in our study would be more likely to show reorganization of language functioning over to the nondominant right hemisphe re or bilateral activations due to naturally decreased lateralization eff ects with age. The findings of our study are somewhat ambivalent in regards to this hypot hesis with some patients generally lending support to this hypothesis (Participants 01, 02, and 03) and others whose data do not support this hypothesis (Partici pants 04 and 05); however, it s hould be noted that, with the exception of one outlier (Participant 05) all of the subjects in our study were relatively similar in age (betw een 48 and 54). It is clear from these studies that the HAROLD hypothesis should also be taken into considera tion when evaluating the results of studies relating to lateralization and reorganization of neural functioning. Inherent sex differences of lateralization should also be considered when reviewing the results of our study. Research findings have shown that women consistently perform better than men on verbal tasks (McGlone, 1980) and that men often perform better than women on tasks requiring more right-hemi sphere dominant visuospatial abilities (Halpern, 2000; Voyer, Voyer, & Bryden, 1995) Based on these findings, it could also be theorized that men, given their greater pe rformance on right-hemisphere visuospatial tasks, would be more likely than women to show reorganization of language functioning over to the right hemisphere after treatmen t in our study, whereas women, with greater performance on verbal tasks, which are strongly lateralized to the left-hemisphere in most individuals, would be more li kely to show perilesional ac tivity in the left-hemisphere

PAGE 58

49 after treatment. Further research with signi ficantly larger samples is necessary before any theoretical conclusions on this issue can be made. Limitations There are several limitations to our study th at limit its scope a nd generalizability. First, the unique nature of the experiment al language rehabilitation program that was used in conjunction with our study makes it unclear whether similar results would be found using other language reha bilitation ther apies. Also, given that the language rehabilitation program used was specifical ly designed to improve language output, it is doubtful that patients with other forms of aphasia (e.g., fluent aphasia) would show similar results to the particip ants in our study. However, it is interesting to note that recent work by Cato, Parkinson, Wierenga, & Cr osson (2004) suggest that one of the best predictors of outcome in the experimental treatment these participants underwent is language comprehension. Another possible limitation for our study involves the difference between the category exemplar task used during both the preand posttreatment fMRI scanning sessions and the naming task used during the la nguage rehabilitation program itself. As stated previously, the categor y exemplar task was used fo r each participant’s fMRI scanning sessions because previous research in this field had shown that the task reliably activated medial-frontal regions in normal controls (Crosson et al., 1999). However, it is unknown at this time whether having participan ts perform the naming task that was used during the language rehabilitation treatment during each of their scanning sessions would yield similar results. Our study is also limited by its small and relatively homogenous sample size. Due to several confounding factors including the hi gh costs of fMRI scanning, the dearth of

PAGE 59

50 patients with chronic nonfluent aphasia in th e regional area who meet inclusion criteria for this type of study, and the timeand personnel-intensive natu re of the language rehabilitation therapy itself, it is a challenge to produce studies of this type with large and diverse sample sizes. Also, due to the ri sks of confounding variables, many different groups of individuals (e.g., l earning disabled, individuals wi th neurological disorders other than stroke, individuals who were not fMRI compatible, etc.) were excluded from our study. It is unclear if the findings pr esented here would also apply to those populations. Implications and Future Directions The practical implications for research in th is general field of study are significant. One example of the benefits of th is type of research is that these studies provide us with a better understanding of how language functions occur in individuals both normally and after brain injury. Knowledge of this type is vital to the future developments of many fields of study including Education, Ne urology, Neuropsychology, Neurosurgery, and Rehabilitation. The most direct benefit from co ntinued research in this area is the future development of new rehabilitation programs and techniques tailored to the needs of each individual patient based on their lesion size, location, and severity. This personalized approach to treatment will maximize treatmen t effectiveness and help ensure sustained improvements in language functioning, whic h, as discussed previously, can have a tremendous impact on each individual’s quali ty of life and overall independence, subsequently decreasing their impact on the economy by increasi ng their likelihood of returning to work and decreasing their needs for government assistance. Future directions for research in this field of study should include conducting studies with larger sample sizes, so that mo re adequate conclusions can be drawn. Also,

PAGE 60

51 studies involving comparisons of different groups of patients, with each group containing patients with highly similar le sions would prove be neficial as they would allow us to examine how each group performs and examin e if specific lesion sizes and locations successfully predict changes in activity in specific brain regi ons after language rehabilitation. Further resear ch on the role subcortical mechanisms play as determinants of cortical reorganization after language re habilitation is also strongly recommended. Also, further research on the actual methods of the treatment paradigm used in our study is warranted. Possible directions for such research could include altering the treatment such that individual participants do not progr ess from Phase one to Phase two or from Phase two to Phase three of the treatment until upward trends in their performance are observed in each Phase. Longitudinal resear ch on treatment outcomes from patients who participated in the language rehabilitation tr eatment used in our study and whether their improvements made in treatment are still pr esent at set periods of time (e.g., 1 month, 6 months, 2 years) after treatment would al so prove highly informative. Additional outcomes research on possible behavioral a nd environmental predic tors of treatment responders vs. individuals who did not appear to respond to treatment might also prove beneficial.

PAGE 61

52 LIST OF REFERENCES Abdullaev, Y. G., & Posner, M. I. (1998). Event-related brain potential imaging of semantic encoding during processing single words. NeuroImage, 7 1-13. Abo, M., Senoo, A., Watanabe, S., Miyano, S., Doseki, K., Sasaki, N., Kobayashi, K., Kikuchi, Y., & Yonemoto, K. (2004). Langua ge-related brain function during word repetition in post stroke aphasics. Neuroreport, 15 (12), 1891-1894. Anderson, S. W., & Damasio, S. W. (1993). The frontal lobes. Clinical Neurospychology (3rd Edition). New York: Oxford University Press, 409-460. Belin, P., Van Eeckhout, P., Zilbovicius, M ., Remy, P., Francois, C., Guillaume, S., Chain, F., Rancurel, G., & Samson, Y. ( 1996). Recovery from nonfluent aphasia after melodic intonation therapy: A PET study. Neurology, 47 1504-1511. Binder, J. R., Frost, J. A., Hammeke, T. A., Cox, R. W., Rao, S. M., & Prieto, T. (1997). Human brain language areas identified by functional magnetic resonance imaging. The Journal of Neuroscience, 17 353-362. Bobrovitz, C. D., & Ottenbacher, K. J. ( 1998). Comparison of vi sual inspection and statistical analysis of single-subject data in rehabili tation research. American Journal of Physical Medici ne and Rehabilitation, 77 (2) 94-102. Broca, P. (1997). On the speech center. In L. Benjamin, Jr. (Ed.), A history of psychology: Original sources and contem porary research (pp.77-81). New York: The McGraw-Hill Companies, Inc. (Original work published 1861). Buffery, W. H., & Burton, A. (1982). Info rmation processing and redevelopment: towards a science of neuropsychological rehabilitation. In A. Burton (ed.) The Pathology and Psychology of Cognition London: Methuen. Buxbaum, L. J., Coslett, H. B., Shall, R. R., & McNally, B. (1993). Hemispatial factors in mirror writing. Neuropsychologia, 31 1417-1421. Cabeza, R. (2002). Hemispheric Asymmetr y Reduction in Older Adults: The HAROLD Model. Psychology and Aging, 17 (1), 85-100. Cao, Y., Vikingstad, E. M., George, K. P., Johnson, A. F., & Welch, K. M. A. (1999). Cortical language activation in stroke patients recovering from aphasia with functional MRI. Stroke, 30 2331-2340.

PAGE 62

53 Cappa, S. F. (2000). Neuroimaging of recovery from aphasia. Neuropsychological Rehabilitation, 10 365-376. Cato, M., Parkinson, R., Wierenga, C., & Cro sson, B. (2004). Predicting rehabilitative treatment success in chronic nonfluent aphasia: Lesion and performance characteristics. [Abstract to be published online] International Neuropsychological Society. Chen, W., & Ogawa, S. (2001). Principle of BOLD-functional MRI. In C. T. W. Moonen & P. A. Bandettini (Eds.), Functional MRI (pp. 103-113). Berlin: Springer. Code, C. (1982). Neurolinguistic analys is of recurrent utterance in aphasia Cortex, 18 141-152. Code, C. (1989). Language, aphasia, and the right hemisphere Publisher unknown. Cox, R.W. (1996). AFNI: Softwa re for analysis and visualiz ation of functional magnetic resonance Neuroimages. Computational Biomedical Research, 29 162-173. Crosson, B., Moore, A. B., Gopinath, K., White K. D., Wierenga, C. E., Gaiefsky, M. E., Peck, K. K., Soltysik, D., Briggs, R. W., & Rothi, L. J. G. (2005). Role of the right and left hemispheres in recovery of f unction during treatment of intention in aphasia. Journal of Cognitive Neuroscience, 17 (3) 392-406. Crosson, B., & Rothi, L.J.G. (1999). Subprojec t: Treatment of attent ion and intention in aphasia. VA RR & D grant number C2602V, NIDCD grant number 03888., 253276. Bethesda, MD: National Institute on Deafness and Other Communication Disorders: National Institutes of Health. Crosson, B., Sadek, J.R., Bobholz, J.A., Gockay, D., Mohr, C.M., Leondard, C.M., Maron, L., Auerbach, E.J., Brows, S.R., Freeman, A.J., & Briggs, R.W. (1999). Activity in the paracingulate and cingulate sulci during word gene ration: An fMRI study of functional anatomy. Cerebral Cortex, 9, 307-316. Crosson, B., Sadek, J. R., Maron, L., Gockay, D., Mohr, C. M., Auerbach, E.J. Freeman, A. J., Leonard, C. M. & Briggs, R. W. (2001). Relative shif t in activity from medial to lateral-frontal co rtex during internally vers us externally guided word generation. Journal of Cognitive Neuroscience, 13 272-283. Elder, J. H. (1997). Single Subject E xperimentation in Psychiatric Nursing. Archives of Psychiatric Nursing, 11 (3) 133-138. Fiez, J. A. (2001). Neuroimaging studies of speech: An overview of techniques and methodological approaches. Journal of Communication Disorders, 34 445-454.

PAGE 63

54 Finger, S. (1997). Cortical localization a nd cerebral dominance: The work of Paul Broca. In L. Benjamin, Jr. (Ed.), A hist ory of psychology: Original sources and contemporary research (pp. 77-81). New Yo rk: The McGraw-Hill Companies, Inc. (Original work published 1994). Foundas, A. L. Eure, K. F., Luevano, L. F., & Weinberger, D. R. (1998). MRI asymmetries of broca’s area: The pars triangularis and pars opercularis. Brain and Language, 64 282-295. Francis, W. N., & Kucera, H. (1982). Frequency analysis of English usage: Lexicon and grammar Boston: Houghton Mifflin Co. Gaiefsky, M. (2003). FMRI of overt language production in aphasia rehabilitation: The contribution of the language nondominant hemisphere. Unpublished Masters Thesis, University of Florida 1-39. Halpern, D. F. (2000). Sex differences in cognitive abilities Mahwah, NJ: Lawrence Erlbaum Associates, Inc. Hanlon, R. E., Brown, J. W., & Gerstmann, L. J. (1990). Enhancement of naming in nonfluent aphasia through gesture. Brain and Language, 38 (2) 298-314. Hausmann, M., Gunturkun, O., & Corballis, M. C. (2003). Age-related changes in hemispheric asymmetry depend on sex. Laterality, 8 (3), 277-290. Heilman, K.M.,Watson, R.T., & Valenstein, E. (1993). Neglect and related disorders. Clinical Neuropsychology (3rd Edition). New York: Oxford University Press, 279336. Karbe, H., Thiel, A., Weber-Luxenberger, G., Herholz, K., Kessler, J., & Heiss, W. D. (1998). Brain plasticity in post stroke aphasi a: What is the contribution of the right hemisphere? Brain and Language, 64 215-230. Kazdin, A, (1982). Single-case research design: methods for clinical and applied settings. New York: Oxford University Press. Lichtheim, L. (1885). On aphasia. Brain, 7 433-484. McGlone, J. (1980). Sex differences in human brain asymmetry: A critical survey. Behavioral and Brain Sciences, 3 215-263. Meinzer, M., Elbert, T., Wienbruch, C., Djundja, C., Barthel, G., & Rockstroh, B. (2004). Intensive language training enhances brai n plasticity in chronic aphasia. BMC Biology, 2 20.

PAGE 64

55 Moore, A. B., Crosson, B., Gockay, D., Leonard, C. M., & Foundas, A. L. (2000). Defining pars triangularis. Society for Neuroscience Abstracts, 30 Part 1, 1247. Moore, A. B., Loftis, C., Crosson, B., Gockay, D., Leonard, C. M., & Foundas, A. L. (2001). Localization of functional activity in pars triangularis. Journal of the Intenational Neuropsycho logical Society, 7, 197. Musso, M., Weiller, C., Kiebel, S., Muller, S. P., Bulau, P., & Rijntjes, M. (1999). Training-induced brain pl asticity in aphasia. Brain, 122 1781-1790. Nadeau, S., Rothi, L. J. G., & Crosson, B. (2000) Preface. In S. Nadeau, L. J. G. Rothi, & B. Crosson (Eds.), Aphasia and language: Theory to practice New York: Guilford Press. National Aphasia Association. (1999). Aphasia Fact Sheet Retrieved March 5, 2005, from http://www.aphasia.org/NAAfactsheet.html Nourbakhsk, M. R., & Ottenbacher, K. J. ( 1994). The statistical analysis of singlesubject data: a comparative examination. Physical Therapy, 74 (8) 768-776. Reuter-Lorenz, P. A., Jonides, J., Smith, E. S., Hartley, A., Miller, A., & Marshuetz, C. (2000). Age differences in the frontal late ralization of verbal and spatial working memory revealed by PET. Journal of Cognitive Neuroscience, 12 174-187. Richards, K., Singletary, F., Rothi, L.J.G ., Koehler, S., & Crosson, B. (2002). The activation of intentiona l mechanisms through utilization of nonsymbolic movements in aphasia rehabilitation. Journal of Rehabilitation Research and Development, 39 445-454. Rosen, H. J., Pertersen, S. E., Linenweber, M. R., Snyder, A. Z., White, D. A., Chapman, L., Dromerick, A. W., Fiez, J. A., & Corb etta, M. (2000). Ne ural correlates of recovery from aphasia after damage to left inferior frontal cortex. Neurology, 55 1883-94. Rothi, L.J.G. (1995). Behavioral compensa tion in the case of treatment of acquired language disorders resulting from brain damage. Compensating for psychological deficits and declines: Managing losses and promoting gains. Mahwah, NJ: Lawrence Erlbaum Associates, Inc. 219-230. Sabsevitz, D. S., Swanson, S. J., Hammeke, T. A., Spanaki, M. V., Possing, E. T., Morris, III, G. L., Mueller, W. M., & Binder, J. R. (2003). Use of preoperative functional neuroimaging to predict language de ficits from epilepsy surgery. Neurology, 60 1788-1792.

PAGE 65

56 Snyder, P. J., & Nussbaum, P.D. (Eds.) ( 1998). Clinical Neurops ychology: A pocket handbook for assessment. Washington, DC: American Psychological Association Press Inc. Talaraich, J., & Tournoux, P. (1988). Co -Planar Stereotaxic Atlas of the Human Brain. New York: Thieme Medical Publishers. Tryon, W. W. (1982). A simplified time-seri es analysis for evaluating treatment interventions. Journal of Applied Behavior Analysis, 15 423-429. Villringer, A. (2001). Physiological Change s During Brain Activation. In C. T. W. Moonen & P. A. Bandettini (Eds.), Functional MRI (pp. 3-13). Berlin: Springer Voyer, D., Voyer, S., & Bryden, M. P. (1995). Magnitude of sex differences in spatial abilities: A meta-analysis and cons ideration of critical variables. Psychological Bulletin, 117 250-270. Wernicke, C. (1874). Der Aphasische Symptomencomplex. Breslau: Cohn and Weigert.

PAGE 66

57 BIOGRAPHICAL SKETCH Bradley J. Daniels was born in Orlando, Florida, and gradua ted Summa Cum Laude from the University of Cent ral Florida in May 2003 with a B.A. in psychology. He immediately entered graduate school in th e Department of Clinical and Health Psychology at the University of Florida, wh ere he is pursuing a Ph.D. in clinical psychology, specializing in cl inical neuropsychology.


Permanent Link: http://ufdc.ufl.edu/UFE0010549/00001

Material Information

Title: Neural Change in Patients with Chronic Nonfluent Aphasia after Language Rehabilitation: A Functional Magnetic Resonance Imaging (fMRI) Study
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0010549:00001

Permanent Link: http://ufdc.ufl.edu/UFE0010549/00001

Material Information

Title: Neural Change in Patients with Chronic Nonfluent Aphasia after Language Rehabilitation: A Functional Magnetic Resonance Imaging (fMRI) Study
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0010549:00001


This item has the following downloads:


Full Text












NEURAL CHANGE IN PATIENTS WITH CHRONIC NONFLUENT APHASIA
AFTER LANGUAGE REHABILITATION: A FUNCTIONAL MAGNETIC
RESONANCE IMAGING (FMRI) STUDY














By

BRADLEY J. DANIELS


A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE

UNIVERSITY OF FLORIDA


2005

































Copyright 2005

by

Bradley J. Daniels



























This thesis is dedicated to the memory of Christopher Ryan Lefever.














ACKNOWLEDGMENTS

I would like to thank Dr. Anna Bacon Moore for her mentorship and guidance

throughout this project. I would also like to acknowledge Dr. Bruce Crosson, Dr. Keith

White, Dr. Kaundinya Gopinath, Dr. Kyunk K. Peck, Dr. Leslie J. Gonzales-Rothi,

Dr. Richard Briggs, Megan Gaiefsky, Christina Wierenga, Bruce Parkinson, and Keith

McGregor for all of their help and support throughout the past 2 years. Lastly, I would

like to thank my friends, family, and loved ones for their continued support and

encouragement. I could not have done this without them.
















TABLE OF CONTENTS

page

A C K N O W L E D G M E N T S ................................................................................................. iv

LIST OF TABLES ............. ................... .. .......... .................................... vi

LIST OF FIGURES ............. ................. .. .......... .................................. vii

A B S T R A C T ..................................................................................................................... v iii

CHAPTER

1 IN TR O D U C TIO N ...................................... ............... ..................... ... .... .. 1

Stroke and Aphasia ..... .. ............. .. ........ ........................................... 1
Intention ..... . ........ ..................................................................... . ............ .2
Intention and Language R rehabilitation ................................................. ...............3...
Functional M agnetic Resonance Imaging (fM RI) ...............................................5...
Functional M R I and A phasia ....................................... ...................... ...............6...
H ypotheses .............. ....................................................................9...........

2 M ETH O D S ..........................................................................................................11

P a rtic ip a n ts ............................................................................................................... 1 1
P ro c e d u re ..................................................................................................................1 3
Language R ehabilitation...........................................................................13
Im age A acquisition ........................................................................................21
Im aging A nalyses ........................................................................................23

3 R E SU L T S .............. .........................................................................................35

4 D ISCU SSIO N ..................................................................................................... 44

C conclusions ............................................................................................. . 44
L im ita tio n s ................ ..... ........................................................................................4 9
Implications and Future Directions ...................... ...................................50

LIST O F R EFEREN CE S ................................................................................................52

BIO GRAPH ICAL SK ETCH ..........................................................................................57



v















LIST OF TABLES


Table page

2-1. P articipant D em graphics ........................................ ........................ ................ 27

2-2. Participant Performance on Language-Related Testing Pre- and Posttreatment .....27

2-3. Participant Baseline Stability and Response to Treatment as Determined by 3
Independent Speech Pathologists ........................................................ ................ 27

2-4. Participant C Statistics for Baseline and Treatment...........................................27

3-1. Participant 01 's Representative Activity Pre- and Posttreatment. Regions of
interest for our study are indicated by a broad white outline..............................39

3-2. Participant 02's Representative Activity Pre- and Posttreatment .........................40

3-3. Participant 03's Representative Activity Pre- and Posttreatment .........................41

3-4. Participant 04's Representative Activity Pre- and Posttreatment .........................42

3-5. Participant 05's Representative Activity Pre- and Posttreatment .........................43















LIST OF FIGURES

Figure page
2-1. Participant 01's Lesion-Sagittal Image- A broad white outline encompassing the
general regions of each participant's lesion(s) are included in their images............28

2-2. Participant 01's Lesion-Coronal Im age..................................................... 28

2-3. Participant 02's Lesion-Sagittal Im age ............................................... ................ 28

2-4. Participant 02's Lesion-Coronal Im age..................................................... 29

2-5. Participant 03's Lesion-Sagittal Im age ............................................... ................ 29

2-6. Participant 03's Lesion-Coronal Im age..................................................... 29

2-7. Participant 04's Lesion-Sagittal Im age ............................................... ................ 30

2-8. Participant 04's Lesion-Coronal Im age..................................................... 30

2-9. Participant 05's Lesion-Sagittal Im age ............................................... ................ 30

2-10. Participant 05's Lesion-Coronal Im age..................................................... 31

2-11. F orm ula for the C Statistic ........................................ ....................... ................ 3 1

2-12. Participant 01's Performance across Treatment Phases......................................32

2-13. Participant 02's Performance across Treatment Phases......................................32

2-14. Participant 03's Performance across Treatment Phases......................................33

2-15. Participant 04's Performance across Treatment Phases......................................33

2-16. Participant 05's Performance across Treatment Phases......................................34















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 Science

NEURAL CHANGE IN PATIENTS WITH CHRONIC NONFLUENT APHASIA
AFTER LANGUAGE REHABILITATION: A FUNCTIONAL MAGNETIC
RESONANCE IMAGING (FMRI) STUDY

By

Bradley J. Daniels

May, 2005

Chair: Anna Bacon Moore
Major Department: Clinical and Health Psychology

Our study was designed to examine changes in language-related activity in patients

with chronic nonfluent aphasia after language rehabilitation, and to further inform the

growing debate in aphasia rehabilitation literature on whether improved language

function in patients with aphasia involves activation of perilesional regions of the left

hemisphere or a reorganization of language function to the language nondominant

hemisphere. For our study, 5 patients with chronic nonfluent aphasia underwent

functional magnetic resonance imaging (fMRI) scanning before and immediately after

completing a language rehabilitation program designed to facilitate reorganization of

language functions to homologous regions in the nondominant hemisphere. We

hypothesized that an increase in activity in homologous language-related medial-frontal

and lateral-frontal right-hemisphere regions would be evident.

Functional MRI data were acquired on a 3T scanner in an event-related paradigm,

while participants completed an overt language task that has been empirically shown to









produce activity in language-related regions in normal controls. Deconvolution, a time

series analysis of the fMRI signal, was then conducted to produce a measure of the

goodness of fit (R2), of the averaged hemodynamic response for each voxel. Selective

detrending was conducted to remove artifacts due to task-correlated motion, and data

underwent thresholding to meet criteria for true hemodynamic activity. We also equated

for sensitivity to normalize signal-noise differences across multiple fMRI scanning

sessions, by deconvolving the fMRI signal from a general-noise test bed. Cluster reports

were then run on each participant's pretreatment and posttreatment data to isolate and

quantify areas of activity. Changes in activity from pretreatment to posttreatment in

language-related regions of interest were analyzed by multiple raters.

Results of our study varied across patients, and lent support to both sides of the

current debate in aphasia rehabilitation literature. We concluded that neural change after

language rehabilitation in patients with nonfluent aphasia appears to be primarily

influenced by lesion size and location. Our findings also suggested that subcortical

mechanisms of the left hemisphere may play a role in how the brain reorganizes after

language rehabilitation.














CHAPTER 1
INTRODUCTION

Stroke and Aphasia

A cerebrovascular accident (CVA), or stroke, occurs when the brain does not

receive adequate oxygen because of a disruption in blood supply (e.g., blockage in a

vessel or vessel rupture), causing damage to affected brain areas (Snyder & Nussbaum,

1998). Aphasia, a common occurrence after a stroke, is defined as a partial or total loss

of language production (or loss of ability to comprehend spoken or written language) due

to disease or brain damage (Broca, 1861/1997; Finger, 1997). Nonfluent aphasia

(impairment in the production of language) is characterized by frequent pauses, difficulty

initiating language, and impaired intonation and grammar; though often with a spared

ability to comprehend language. Practical implications for successful treatments of

aphasia are considerable. Approximately 500,000 new strokes occur in the United States

every year; and as many as 25% of those cases develop some form of aphasia due to

stroke (Nadeau, Rothi, & Crosson, 2000). Currently, an estimated 1,000,000 people

living in the United States have aphasia, and most of them developed the impairment as a

result of a stroke (National Aphasia Association, 1999).

Research and treatments that help to improve communication ability in individuals

suffering from aphasia can greatly improve quality of life, decrease dependence on

others, and decrease the impact on the overall economy by helping those individuals

successfully return to work (Crosson & Rothi, 1999).









Intention

Intention mechanisms are cognitive mechanisms that help us in our preparation to

respond to a given stimulus, and are necessary for efficient cognitive processing

(Heilman, Watson, & Valenstein, 1993). Language is one cognitive activity that depends

upon intentional mechanisms to achieve its goals, because intention mechanisms aid in

word selection and in the fundamental initiation of speech (Crosson et al., 2005). One

factor that may contribute to difficulty initiating language common in patients with

nonfluent aphasia is poor preparation to respond (i.e., inadequate intention) (Crosson &

Rothi, 1999).

It is necessary at this point to describe briefly what is currently known about typical

language production. Since the 1860s and Paul Broca's seminal studies with his patient

"Tan," lesion studies have shown language impairments after damage to specific left-

hemisphere regions (Broca, 1861/1997; Lichtheim, 1885). From these studies, Broca

(1861/1997), Wernicke (1874), and others developed what is commonly referred to as the

"classical model" of language, which posits that language function is predominantly

housed in the left hemisphere and is divided into two major regions: Broca's area

(housed in the frontal lobe and primarily responsible for the expressive aspects of

language), and Wernicke's area (housed in the posterior region of the temporal lobe and

responsible for the receptive components of language) (Binder et al., 1997; Lichtheim,

1885; Wernicke, 1874). After the advent of functional neuroimaging, multiple

neuroimaging studies (Binder et al., 1997; Foundas, Eure, Luevano, & Weinberger, 1998;

Musso et al., 1999) confirmed language functioning in right-handed individuals to be

almost entirely controlled by the left hemisphere. The primary regions of the left

hemisphere thought to be involved in the expressive components of language include









Broca's area, as well as the precentral gyrus (i.e., motor strip), inferior frontal sulcus, and

the middle and superior frontal gyri (Binder et al., 1997; Crosson et al., 2001; Foundas et

al., 1998; Moore, Crosson, Gockay, Leonard, & Foundas, 2000; Moore et al., 2001).

While these areas are believed to be involved in language production, the initiation or

intentional component of language is believed to involve the medial-frontal regions of the

left hemisphere, including cingulate gyrus and sulcus, paracingulate gyrus and sulcus (if

present), PreSupplemental Motor Area (Pre-SMA), Supplemental Motor Area (SMA),

and the frontal eye fields. (Abdullaev & Posner, 1998). According to Abdullaev &

Posner (1998), normal language begins in left-hemisphere medial-frontal regions and

travels outward for approximately 80 ms to the left lateral-frontal regions for speech

production.

The middle cerebral artery (MCA) supplies blood to the lateral-frontal regions of

the brain but not to the medial-frontal regions; therefore, a left MCA stroke would likely

leave left medial-frontal regions intact while damaging left lateral-frontal regions. This

leads us to believe that nonfluent aphasia is a result of intact medial-frontal regions of the

left hemisphere initiating the intention of language, and then subsequently attempting to

communicate with dysfunctional left lateral-frontal regions damaged by the stroke,

thereby inhibiting speech output.

Intention and Language Rehabilitation

Intention mechanisms in the recovery of language function have not been heavily

researched to date. Rapid recovery from akinetic mutism due to unilateral medial-frontal

lesion (Damasio & Anderson, 1993) suggests that shifting intention mechanisms to the

undamaged hemisphere might be a viable treatment strategy in some patients with

aphasia. It has also been theorized that attending to stimuli in hemispace contralateral to









the intact hemisphere or initiating an action with the hand contralateral to the intact

hemisphere, thereby activating nondominant hemisphere functions, engages intention

language mechanisms in that hemisphere; which are then able to compensate for

dysfunctional language mechanisms in the damaged hemisphere. Previous studies

(Hanlon, Brown, & Gerstmann, 1990) showed improvements in naming ability for words

learned while gesturing for nonfluent aphasic populations. From these findings, it could

be theorized that, since gesturing with the hand contralateral to the intact hemisphere

would cause nondominant right-hemisphere motor regions to become active,

simultaneously performing a naming task while gesturing might also then cause

nondominant right-hemisphere language regions to become activated. Crosson and Rothi

(1999) began developing and applying aphasia treatments specifically designed to recruit

intention mechanisms in the intact hemisphere and to engage those regions in language

processing. These types of treatments are commonly defined as substitutive treatments,

meaning that they attempt to reconstruct the impaired language system by substituting an

intact mechanism for the damaged one. Restitutive treatments, by comparison, attempt to

reconstitute the impaired system in its original form. Substitutive treatments are more

effective than restitutive treatments in chronic phases of recovery (Rothi, 1995).

From an empirical standpoint, some studies, such as the ones conducted by Code

(1982; 1989), actually attempted to apply similar methods to treatments of individual

cases. In Code's patients, stimuli were presented to the left visual field, left hand, and/or

left ear in an attempt to activate nondominant-hemisphere language mechanisms. Both of

Code's (1982; 1989) studies reported improvement after a lengthy treatment.









Unfortunately, both studies involved single cases, and allowed for alternative

explanations due to the methodology used by the studies (Crosson & Rothi, 1999).

Our study was based on a few key assumptions outlined by Crosson & Rothi,

(1999):

* Intention mechanisms support language function; and when those regions are
damaged or disconnected from language mechanisms, language functions are
adversely affected.

* Alternate intention mechanisms (i.e., in the intact hemisphere) can be engaged in
the service of language functions, improving language performance (as shown by
Buxbaum, Coslett, Shall, & McNally, 1993). This phenomenon has been used
effectively in treatment as suggested by the individual case studies of Buffery &
Burton (1982) and Code (1982, 1989). Therefore, it is believed that some patients
with chronic nonfluent aphasia will respond to a treatment designed to activate
intention mechanisms in the nondominant hemisphere (Crosson & Rothi, 1999).

The goal of our study was to use neuroimaging to reveal changes in neural

substrates of language production in patients with nonfluent aphasia after their

participation in a rehabilitation program.

Functional Magnetic Resonance Imaging (fMRI)

The foundation of functional neuroimaging theory is based on the assumption that,

as processing demands increase during a task, changes in neural activity also occur in

specific brain areas associated with that task (Fiez, 2001). Functional MRI is a

noninvasive technique that vicariously measures neural activity by measuring changes in

the amount of deoxyhemoglobin in the brain tissue. When blood flow to a specific area

of the brain increases to meet the demands of a task, the amount of oxygen-rich

hemoglobin supplied to that region also increases; inversely, the amount of

deoxyhemoglobin (nonoxygen carrying hemoglobin molecules) decreases (Chen &

Ogawa, 2000; Fiez, 2001; Villringer, 2000). Deoxyhemoglobin has certain paramagnetic

qualities such that changes in concentration of this compound are related to changes in









magnetic signal. When red blood cells containing deoxyhemoglobin are placed in the

magnetic field used for MRI, there is some magnetic field distortion induced by the

difference in the magnetic susceptibility relative to the surroundings. This susceptibility-

induced field distortion due to the level of deoxyhemoglobin is the basis for Blood

Oxygen Level Dependent (BOLD) contrast fMRI, and changes in deoxyhemoglobin

levels associated with underlying functional activation can be detected in the MRI signal

(Chen & Ogawa, 2000).

Blood flow changes to a given region of the brain are known as hemodynamic

responses. Once a hemodynamic response has begun, it will typically peak

approximately 6 to 8 seconds afterwards and return to baseline levels approximately 10 to

12 seconds after initiation (Fiez, 2001).

Functional MRI and Aphasia

An ongoing debate in the field of aphasia rehabilitation involves whether or not

improved language function in patients with aphasia involves activation of perilesional

regions of the left hemisphere (Cao, Vikingstad, George, Johnson & Welch, 1999;

Meinzer et al., 2004) or if it involves a reorganization of language function to the

language nondominant hemisphere (Abo et al., 2004; Cappa, 2000; Crosson et al., 1999;

Rosen et al., 2000). Cao et al. (1999) used BOLD fMRI to study cortical language

networks in 7 right-handed patients with aphasia during lexical-semantic processing

tasks. All of the patients in the study were at least 5 months post stroke and had

recovered a "substantial" portion of their language functions since their stroke. The

authors found that left-hemisphere perilesional activity was associated with better

recovery in their patient population and was inversely related to the amount of activity

found in homologous language regions in the right hemisphere (Cao et al., 1999). A









more recent study by Meinzer and colleagues (2004) studied changes in Delta Dipoles, or

focal clusters of slow wave activity in the delta frequency range (1 to 4 Hz) which are

usually located in the vicinity of structural damage in the brain. Meinzer and his

colleagues used magnetoencephalography (MEG) to study Delta Dipole Density changes

in 28 patients with chronic aphasia (>12 months post stroke) before and after intensive

speech and language therapy. Results of the study showed that decreases in Delta Dipole

Density in perilesional regions of the brain were found in a majority of patients. These

results emphasize the significance of perilesional areas in the rehabilitation of aphasia

(Meinzer et al., 2004).

There are other studies, however, that provide contradictory findings to those found

by Cao et al. (1999) and Meinzer et al. (2004). A study by Musso et al. (1999) used

positron emission tomography (PET) to measure changes in regional cerebral blood flow

(rCBF) in 4 patients with Wernicke's aphasia. Patients underwent a language

comprehension task during 12 consecutive scanning sessions. Between sessions, patients

participated in brief but intense language comprehension training. All patients improved

significantly over the 12 comprehension tasks used during scanning. The authors found

that one of the regions that best correlated with training-induced improvement was the

right superior temporal gyms. From their findings, the authors concluded that the right

hemisphere plays a significant role in the recovery from aphasia (Musso et al., 1999).

Another recent study by Abo et al. (2004) used a word repetition task during fMRI

scanning to study language activity in 2 patients with aphasia. Functional MRI scans

were performed on the patients 60 months post stroke. Results from the study found that

both patients showed activation of only the compensatory area in the right hemisphere









during the task. It becomes clear from examining these studies that current research on

this much-debated topic provides support for both left-hemisphere perilesional and right-

hemisphere reorganization-of-function arguments.

An alternative, or "middle ground," explanation that some studies (including ones

previously mentioned) have shown may be that the determination of perilesional

activation vs. reorganization to the nondominant hemisphere is mediated by the size and

severity of the lesion. In essence, the greater the size and severity of the lesion in the left

hemisphere, the more likely it is for reorganization of language function over to the

nondominant hemisphere to occur (Belin et al., 1996; Cao et al., 1999; Karbe et al.,

1998).

More recent studies on this topic are beginning to examine the role of subcortical

structures in predicting how the brain adapts to injury, and have begun to find predictors

of bilateral activation based on whether or not left-hemisphere subcortical structures were

affected by the stroke. (Crosson et al., 2005) Using fMRI as a research tool allows

scientists the opportunity to attempt to provide answers to these much-debated questions.

Functional imaging studies of intention mechanisms for language have used

primarily word generation tasks (Crosson & Rothi, 1999). For example, participants are

given a letter of the alphabet (e.g., A), and asked to produce as many words as they can

think of beginning with that letter (e.g., apple, actress, aunt), or participants are given a

semantic category (e.g., colors) and asked to generate as many exemplars from that

category as possible (e.g., red, orange, blue). It has been shown that word generation

tasks such as these reliably produce medial-frontal activity in individual controls









(Crosson et al., 1999). Therefore, a word generation task designed to track medial-frontal

activity pre- and posttreatment was used in our study.

Hypotheses

The purpose of our study was to examine what neuroanatomical patterns of

activation are evident in patients with chronic nonfluent aphasia during language

generation and over the course of a treatment designed to shift medial-frontal activity

from the left to the right hemisphere. Our hypotheses were that dominant pretreatment

activity would be observed in left medial-frontal regions. Right medial-frontal activity

was not expected before initiation of the experimental language treatment. We also

hypothesized that little, if any, left lateral-frontal activity would be evident in these

patients during language production due to left-hemisphere stroke-related damage. Given

previous research, we expected that there might be some right lateral-frontal activity

during language production in these patients.

Another question that our study posed was: among participants enrolled in an

experimental language rehabilitation program designed to improve language function by

activating intention mechanisms in the language nondominant hemisphere, what changes

in patterns of neural activity were evident during language production posttreatment

when compared to pretreatment? One hypothesis is that participants who responded to

treatment would show changes in both right medial-frontal and right lateral-frontal

homologous language regions (Crosson et al., 1999). The expectation is that right-

hemisphere medial-frontal regions would show increased activation during language

production posttreatment. Predictions about right lateral activity were not made because

it could be that right lateral activity increased posttreatment as a result of synergistic co-

activation of medial-frontal to lateral-frontal regions. Or, it could be that posttreatment









right lateral-frontal activity would become more efficient and therefore possibly less

distributed as a result of a more efficient "upstream" activity in the right medial-frontal

regions.

The practical implications for this type of study are numerous, and include further

insight into the neural substrates of language function in patients with chronic nonfluent

aphasia and the role that lesion size and location play in reorganization of language

functions. Increased knowledge in how these functions work after injury and subsequent

rehabilitation can aid in the development of new language rehabilitation treatments

designed to maximize treatment outcomes in patients based on the size and severity of

their individual lesions.














CHAPTER 2
METHODS

Participants

Ten patients with chronic nonfluent aphasia (6 female, 4 male) with left-

hemisphere stroke participated in our study. Due to inherent difficulties associated with

fMRI (e.g., scanner problems, excessive motion artifact during scanning), data from 5 of

these participants (2 male, 3 female; mean age = 55.2 + 9.68 years; mean number of

months post stroke = 35.4 + 27.67 months) were available for inclusion in the final

analyses presented in our study. Although this may appear to be a relatively high dropout

rate, it should be noted that multiple scanning sessions were required for our study, and

that the population of interest included neurologically impaired participants who

possessed many deficits (e.g., hemiplegia) that would cause significant discomfort during

the extended scanning sessions required for completion of our study.

Inclusion criteria for our study were: Males or females over 21 years of age with

documented left-hemisphere cerebrovascular accident (CVA), demonstrated nonfluent

aphasia, a minimum of 6 months post stroke, premorbidly right handed, native English

speakers, within acceptable limits for height and weight for participation in an fMRI

study. Finally, all participants in our study were participants in an experimental language

rehabilitation treatment designed to improve language function in patients with chronic

nonfluent aphasia; this treatment is described in detail later. Participants were excluded

from the imaging portion of our study if they were claustrophobic, possessed a history of

psychiatric illness, substance abuse, traumatic brain injury, seizures (unrelated to stroke),









dementia, learning disability, neurological disorders other than stroke, or other criteria

that would prevent them from being placed in the fMRI scanner (e.g., being pregnant or

having metal inside their body.) For demographic information on the participants in our

study, see Table 2-1.

"Participant 01" (Figure 2-1 and Figure 2-2) was a 48-year old Caucasian male who

suffered a left middle cerebral artery stroke that primarily affected the left basal ganglia

region including the caudate nucleus, globus pallidus, internal capsule, and cerebral

peduncle. Participant 01's stroke also produced an encephalomacic effect that caused

damage to the left frontal, temporal, and parietal lobes.

"Participant 02" (Figure 2-3 and Figure 2-4) was a 48-year old Caucasian female

who suffered a left middle cerebral artery stroke affecting the frontal, temporal and

parietal lobes with some subcortical extension upwards into the lenticulostriate endzone.

"Participant 03" (Figure 2-5 and Figure 2-6) was a 52-year old Caucasian female

who suffered a significant left middle cerebral artery stroke affecting the frontal,

temporal, parietal, and occipital lobes and causing ventriculomegaly. Also, she had

undergone a left parietal craniotomy.

"Participant 04" (Figure 2-7 and Figure 2-8) was a 54-year old Caucasian female

who suffered a left middle cerebral artery stroke affecting her frontal and temporal lobes

with some midline shift due to the stroke. She had also suffered a previous stroke

affecting her left parietal lobe.

"Participant 05" (Figure 2-9 and Figure 2-10) was a 74-year old Caucasian male

who suffered a left middle cerebral artery stroke affecting the frontal, temporal, and

parietal lobes, and extending well into subcortical structures including the caudate









nucleus, internal capsule, insula, cerebral peduncle, and the entire lenticulostriate

endzone.

Procedure

Language Rehabilitation

All participants were enrolled in a language rehabilitation treatment designed to

improve speech and language function in patients with chronic nonfluent aphasia by

activating medial-frontal regions with the goal of transferring intention mechanisms from

the left-hemisphere medial-frontal regions to the right-hemisphere homologue. Although

the purpose of our study is to elucidate the neural substrates of language recovery and not

to examine the language treatment study itself, an explanation of the treatment study is

necessary because the imaging results occur within the context of a specific language

therapy, not across language therapies in general. The medial-frontal regions of interest

in our study were chosen because previous research involving language rehabilitation

populations of patients with aphasia have shown that right lateral-frontal regions show

some neural activation during language production in patients with nonfluent aphasia

(Belin et al., 1996). It is believed that activation of homologous medial-frontal regions

may facilitate change in right lateral-frontal regions as well, thereby producing an

increase in speech output due to neural reorganization.

The Intention language rehabilitation treatment used in our study was the same as

the treatment presented in Richards et al. (2002). Briefly, the treatment involved the

performance of an overt (out loud) object-naming task while simultaneously producing a

meaningless circular gesture with the left hand. The treatment itself consisted of a

preliminary baseline phase followed by three phases of treatment. To control for the

effects of spontaneous recovery, all participants enrolled in the language rehabilitation









treatment were at least 6 months post stroke. The baseline phase for the study consisted

of a minimum of 8 sessions and each of the three treatment phases consisted of 10

sessions. Each baseline and treatment session for the study lasted approximately forty-

five minutes. After establishing a stable baseline, progression from phases one through

three of the treatment involved transition from movements prompted by external cues

(e.g., a tone and a flashing star), to a self-initiated movement sequence. The rationale for

differential phases of treatment involved having specific points of demarcation between

phases such that the externally guided to internally guided movement progression would

allow the participant to eventually learn to pair their movement sequence with the

initiation of language. For this preliminary study, performance criteria (e.g., minimum

percentage of correct responses during consecutive treatment sessions) were not

established as a rule for transition from one treatment phase to the next; rather, each

treatment phase consisted of exactly 10 sessions. The meaningless circular gesture

involved in the study was internally guided, generalizable to real-world interactions, was

not word-related or symbolic in nature, and did not resemble any symbolic action with

which the participants would have already been familiar. The same circular gesture was

used for all participants.

During each of the pretreatment baseline sessions, participants were seated at a

desk directly facing a computer monitor with their head and body facing the monitor.

The participant then performed a naming task as 40 black-and-white line drawings,

approximately 4 inches by 4 inches in size, were displayed on the monitor one at a time.

After the presentation of each drawing, the participant named the drawing. For example,

if a shoe was presented, the participant would respond "shoe". These sessions were used









to establish a baseline rate of percent correct of naming accuracy and reaction time. Once

a stable baseline was established, the first phase of the Intention treatment would begin.

Phase one of the treatment, like the baseline sessions, began with the participant

seated at a desk with their head and body directed towards a computer monitor. The

therapist would begin the trial by pressing the mouse button. A 1 inch by 1 inch star

would appear on the center of the screen and a 1,000 Hz tone would sound. The color

and orientation of the star would vary from trial to trial. To initiate the presence of the

line drawing, each participant would lift the lid on a small box located to his or her left

with their left hand and press a button inside the box. The button press caused the tone to

stop and the star to disappear from the monitor and, after a 2 second delay, a black-and-

white drawing appeared at the center of the monitor and a timer began. If the participant

named the picture correctly, the therapist would press the mouse button to end the trial

and then stop the timer and remove the line drawing from the screen. If the participant

named the drawing incorrectly, the therapist provided the correct name for the picture

while simultaneously making the circular gesture described above with his or her left

hand. The participant was instructed to repeat the corrected picture name aloud while

also making the same circular gesture. Treatment phases had 50 trials each.

Phase two differed from phase one only in that the 1,000 Hz tone was eliminated

and a different set of 50 line drawings not previously seen was used. Incorrect responses

were corrected using the same procedure as in phase one of treatment.

The same treatment procedures used in phase one and two were used in phase

three; however, in this phase of treatment, the flashing star was eliminated and the

participant was instructed to perform the same meaningless circular gesture performed in









the previous phases with his or her left hand before the presentation of the line drawings.

Response instructions, and correction of incorrect responses, were the same as described

above in phases one and two. Once again, a different set of 50 line drawings not

previously seen was used.

Two overall sets of line drawings were used in the treatment study: one balanced

set with both high- and low-frequency words, and one set for higher-functioning

participants comprised of fifty low-frequency words. The balanced set of line drawings

contained 15 high frequency words (27-717 occurrences per million), 15 medium

frequency words (4-10 occurrences per million), and 20 low frequency words (0-30

occurrences per million), to provide a balanced set of words and prevent participants

from obtaining ceiling effects during treatment. The low frequency set of line drawings

contained all low frequency words. Frequencies of words were based on Francis and

Kucera's "Frequency Analysis of the English Language" (Francis & Kucera, 1982). The

determination for whether a participant was considered high functioning and given the

low frequency set of line drawings or low functioning and given the balanced set of line

drawings for the language rehabilitation treatment was based upon their overall

performance on a set of naming probes from the balanced picture set. In other words, if a

participant performed on average greater than 70% on the balanced picture set over 4 to 8

naming probes, they were grouped in the highest functioning group and given the low

frequency picture set. All participants scoring below 70% received the balanced set of

line drawings.

To determine the stability of baseline performance, and overall response to the

treatment itself, an independent panel of 3 speech pathologists were recruited to rate each









subject's data based upon the graphs displayed in Figures 2-11 through 2-15. Each rater

examined the graphs for each participant, which showed the percentage of line drawings

named correctly throughout each baseline and treatment session and, using visual

inspection, judged each participant for stability throughout the baseline sessions and

overall response to treatment. Visual inspection refers to the making of judgments of

reliability or consistency of intervention effects through the use of visual examination of

graphed data (Kazdin, 1982). The primary difference between traditional between-

subjects group research and single-case research designs is that, in between-group

research, the experimental criterion is met by comparing performance between or within

groups though the use of statistics, whereas, in single-case research, the experimental

criterion is met by examining the effects of the treatment at different points over time

(Kazdin, 1982). The method of visual inspection of data to determine performance

throughout treatment is considered the most common technique used in single-case

research designs (Elder, 1997; Kazdin, 1982), and has been determined by some studies

to be highly correlated with single-subject statistical procedures when evaluating

treatment effects (Bobrovitz & Ottenbacher, 1998).

The C statistic (Figure 2-11) was also used to provide a quantitative measure of

baseline stability and treatment performance (Tryon, 1982). The C statistic is a

"simplified" time-series analysis that can be used to measure the effect of treatment

interventions on studies that have as few as 8 points of data per experimental phase by

examining upward trends in performance across time (Tryon, 1982).

The numerator of the right-hand term of the C statistic is the sum of N 1 squared

consecutive differences associated with the time series. The denominator is twice the









sum of N squared deviations of the time-series data points from their mean. The final

determination for significance using the C statistic is calculated by dividing C in the

equation above over the standard error of the C statistic, which is calculated by taking the

square root of N + 2 over (N-1) (N+I) where N equals the number of data points in the

time series (Tryon, 1982).

If a participant's baseline sessions did not reach statistical significance (p>.05), it

was assumed by the C statistic that they possessed a stable baseline. Then, if their overall

data from the 3 phases of treatment was considered statistically significant (p<.05), they

were deemed by the C statistic to be a treatment responder.

In single-case research designs, statistical tests to determine performance

throughout treatment are occasionally used; however, this practice remains the exception

rather than the rule (Kazdin, 1982). According to Tryon (1982), greater confidence can

be placed in scientific data when both visual and statistical analysis procedures agree.

Studies have shown, however, that the results of single-subject research designs can be

directly influenced by the method of statistical data analysis selected and that multiple

single-subject statistical tests-including the C statistic-show relatively low levels of

agreement between tests (Nourbakhsh & Ottenbacher, 1994). Also, it is unclear whether

the C statistic is useful with participants who display floor effects as the C statistic may

become artificial inflated to significance level due to low standard error. Therefore,

whenever there was a discrepancy between a participant's ratings of performance by the

independent panel of speech pathologists and their ratings of performance based on their

C statistics, the independent panel was used as the final judgment for our study.









Each participant's scores on specific language-related tests that were administered pre-

and posttreatment, their ratings by an independent panel of 3 speech pathologists, and

their C statistics are displayed in Table 2-2, Table 2-3, and Table 2-4.

Participant 01 (Figure 2-12) was judged as having a stable baseline and was

considered a treatment responder as judged by an independent panel of 3 speech

pathologists. Participant 01 was rated as having a stable baseline and was considered a

treatment responder according to his C statistics as well. Participant 01 was considered

high functioning for our study and so received the low frequency stimuli.

Participant 02 (Figure 2-13) was judged to have a stable baseline by 2 out of 3

raters. All raters considered her to be a treatment responder. According to her C

statistics, she also had a stable baseline and was considered to be a treatment responder.

Participant 02 was considered high functioning for our study as well and so received the

low frequency stimuli.

Participant 03 (Figure 2-14), like Participants 01 and 02, was judged as having a

stable baseline and as responding to treatment by all 3 independent raters. However, it is

likely that Participant 03's baseline stability is best explained by floor effects, as

Participant 03 was unable to correctly name any of the 40 presented line drawings during

6 of her 9 baseline sessions. Nevertheless, Participant 03 did begin to show visible

improvements in naming ability by Phase two of treatment. Based on her C statistics,

Participant 03 was considered to be a treatment responder, but was not considered to have

a stable baseline. This too, is likely due to her floor effects and the C statistic's artificial

inflation to significance level of her performance throughout the baseline sessions due to

the low standard error. This subject shows the limitations of the C-statistic and provides









and example of a situation in which visual inspection by independent raters provides a

more accurate assessment of stability versus change. Participant 03 was considered low

functioning for our study and so received the balanced set of line drawings.

Though Participant 04 (Figure 2-15) was judged as possessing a stable baseline,

this judgment is also likely best explained by floor effects. She also did not appear to

respond to treatment as determined by the independent panel of 3 raters. Participant 04's

C statistics were considered stable for her baseline sessions, as was her response to

treatment. Recall that a stable response during the treatment phase is consistent with a

treatment nonresponder. Participant 04's poor treatment outcome may be in part due to

the overall severity of her aphasia. She was considered low functioning for our study and

therefore received the balanced set of line drawings.

Participant 05 (Figure 2-16) was judged as having an unstable performance

throughout the baseline sessions by all 3 raters. However, 2 out of 3 raters deemed

Participant 05 to be a treatment responder. It is possible that Participant 05's unstable

baseline may reflect possible continued recovery from stroke. However, given that most

effects of spontaneous recovery are observed within the first six months after stroke, and

that Participant 05's stroke occurred approximately eighty-six months before the

initiation of treatment, spontaneous recovery of language functioning is unlikely.

According to his C statistics, Participant 05 appeared to have a stable baseline, but was

not considered to be a treatment responder. Figure 2-10 suggests that Participant 05's

performance during treatment appeared to have an upward trend; however, this trend was

not statistically significant based on his C statistic. Participant 05 was considered high

functioning for our study and so received the low frequency set of line drawings.









Image Acquisition

Before beginning the Intention language rehabilitation treatment, each participant

underwent a pretreatment fMRI scan. Participants also underwent a posttreatment scan

within 2 weeks of their completion of the treatment. All participant scans were

conducted with a 3 Tesla fMRI scanner using a dome-shaped RF quadrature head coil

(MRI devices). Functional imaging parameters for all fMRI scans conducted in our study

were as follows: single shot spiral scan, gradient echo pulse sequence, TE = 18 ms, TR =

1660 ms, FA = 60 degrees, FOV = 200 mm, matrix = 64 x 64, 32 slices with whole brain

coverage, and slice thickness = 4 mm. Structural imaging parameters for all fMRI scans

conducted in our study were as follows: 3D spoiled GRASS sequence, TE = 6 ms, TR=

23 ms, FOV = 240 mm, matrix size = 256 x 192 and slice thickness = 1.3 mm.

An event-related paradigm, which captures single hemodynamic responses by

alternating between active stimuli and long interstimulus intervals (ISI's), was used

during scanning. This was done so that each hemodynamic response could run its course

and return to baseline before the next stimulus was presented. This paradigm was chosen

over a blocked paradigm as it allows researchers greater flexibility and the opportunity to

examine overt response patterns, adherence to the required task, and behavioral changes

during scanning. In contrast, a block paradigm attempts to capture a stronger

hemodynamic response by alternating between periods of multiple "active" stimuli with

very short interstimulus intervals, followed by a long period of rest in between blocks.

(Fiez, 2001). This produces multiple hemodynamic responses, subsequently producing

an additive effect and yielding a stronger activation due to insufficient time for the

hemodynamic responses to return to baseline before the next stimulus is presented.









Participants were asked to perform a verbal fluency task requiring the production of

a category exemplar during both pre- and posttreatment scanning sessions. For example,

if participants heard the category "types of fish," they might respond with "shark" or

"piranha." This task was modeled after the word generation task used in Crosson, Sadek,

Maron, et al. (2001). Participants practiced this task with a clinician prior to their

placement inside the scanner to ensure that they fully understood the task. Participants

were instructed to give only one response for each category and were reminded to remain

completely still during scanning to minimize motion artifact. Participants were also

instructed to relax and await the presentation of the next category after each response. If

the participant was unable to hear or understand the name of a category during scanning,

they were instructed to respond "what," to prevent the later coding of a response as

"incorrect" or "other" when stimuli were inaudible or uninterpretable.

Category exemplars were presented to participants via a magnacoustic digital audio

system and nonmagnetic headset with microphone. Sound attenuation processes were

performed before the functional scans to ensure that participants could hear the stimuli.

Five stimulus runs, each containing 9 categories with variable rest intervals between

stimuli, were used, for a total of 45 stimuli per scan. Interstimulus intervals (ISI's) of

21.58 seconds, 23.24 seconds, 24.9 seconds, or 26.56 seconds were pseudorandomly

interspersed between categories throughout each run. These variable interstimulus

intervals between category presentations were selected based on mean response times

outside of the scanner as determined by a separate but related pilot study and allowed:

* Adequate time for the participant's hemodynamic response, estimated to last as
long as twenty seconds, to return to baseline level.

* Variability between intervals required for statistical deconvolution to take place.









Overt responses were recorded directly to a laptop computer via the magnacoustic

microphone. All participants received the same stimuli during both pre- and

posttreatment scans, and all runs were presented in a randomized order across the 2 scans.

Participants were debriefed after the completion of the scan and were offered an

opportunity to view anatomical images of their brain.

Imaging Analyses

Many different techniques have been used to evaluate the results of fMRI.

However, no standardized way of quantifying, and subsequently qualifying, functional

activity has been developed to date. One topic of controversy in fMRI research involves

the collapsing of data across participants. By collapsing data across participants, activity

voxels and clusters are more easily localized and evaluated; however, when working with

stroke patients, each patient's size and severity of lesion will vary significantly.

Therefore, collapsing data across these types of patients can corrupt the investigation and

understanding of which structures may still be functional in each individual patient.

Consequently, data obtained for each participant in our study were analyzed using within-

subjects analyses.

A commercial software package (Cool Edit 2000TM) was used to record participant

responses during scanning directly to a laptop computer. As a consequence of this

method, scanner noise was also recorded during participants' responses. Specialized

tools used by Cool Edit 2000TM were then used to reduce the amount of scanner noise in

the recorded responses so that participant responses could be heard and analyzed.

Participant responses were then coded as correct, incorrect/other, or no response (NR).

Recorded responses were also analyzed to determine the precise time at which the

participant's response was initiated. By determining the exact time at which each









response was initiated, the image acquisition number for each response was able to be

determined. Imaging data for this particular study were analyzed using a paradigm that

incorporated all responses given by participants, meaning that images were included in

analysis for each time the participant responded to a stimulus, regardless of whether or

not their actual response was coded as "correct" for the given category. The reasons an

all-response paradigm was used for our study as opposed to a response paradigm

analyzing images from only responses coded as "correct" were:

* To allow sufficient responses for analysis.

* Effort to respond alone would show brain processes of interest for our study.

A number of specialized statistical procedures were conducted by qualified lab

personnel to prepare each participant's images for analysis. First, the images obtained

from each participant's scans were analyzed using Analysis of Functional Neuroimages

(AFNI) software (Cox, 1996) to derive functional maps based on response type. Second,

deconvolution, a time series analysis of the fMRI signal, was conducted. For this

analysis, specific time intervals for which a hemodynamic response was expected were

designated. Deconvolution produces a measure of the goodness of fit (R2), of the

averaged hemodynamic response for each voxel. Once deconvolution was completed,

selective detrending was conducted for the removal of artifacts due to task-correlated

motion (e.g., speaking). It is expected that task-correlated motion will occur in

approximately the first 5 seconds (corresponding to approximately 3 images) of the

hemodynamic response. Since the hemodynamic response can last up to 20 seconds,

specific measurable signals related to task-correlated motion were detected and removed

from analysis. A thresholding procedure was then used to exclude activated voxels that

did not fit the expected amplitude of change for true hemodynamic activity (e.g., percent









change > 8% for our study). For example, draining veins like the superior sagittal sinus

often produce changes in the Blood Oxygen Level Dependent (BOLD) signal and present

as activity on fMRI scans; however, most of these activated voxels are removed through

thresholding. After thresholding was complete, equating for sensitivity was conducted to

normalize signal-noise differences across multiple fMRI scanning sessions by

deconvolving the fMRI signal from a general-noise test bed. Smoothing of the images

was not performed.

Finally, each participant's functional and anatomical images were converted to 1

mm3 voxels and deformed into atlas space (Talairach & Tournoux, 1988) to normalize

each brain to a standard size and orientation. This process uses 10 landmark points to fit

brains into atlas space. These points include: the midline posterior, superior margin of

the anterior commissure (AC); the inferior margin of the posterior commissure (PC); 2

mid-sagittal points in the interhemispheric fissure; the left-most and right-most points in

the brain; the most superior and inferior points in the brain, and the most anterior and

posterior points in the brain. Once these preliminary procedures were completed, cluster

analyses were performed.

Left-hemisphere lateral-frontal regions of interest for our study were as follows:

left precentral gyms (Brodmann's area [BA] 4) and left perilesional regions (BA's 6, 8, 9,

44, and 45). Left-hemisphere medial-frontal regions of interest were: left Pre-SMA

(medial BA 6), left SMA (medial BA 4), the left paracingulate gyrus (if present) coupled

with the top half of the left cingulate sulcus (BA 32), and the left cingulate gyrus coupled

with the bottom half of the left cingulate sulcus (BA 24). Left cingulate and









paracingulate sulci were also labeled individually if the distinction between BA 32 and

24 was unclear.

Right-hemisphere lateral-frontal regions of interest were: right precentral gyrus

(BA4), right Broca's homologue, encompassing Pars Opercularis and Pars Triangularis

(BA's 44 and 45, respectively), regions of right inferior frontal gyrus outside of Broca's

homologue, right inferior frontal sulcus, right middle frontal gyrus, right superior frontal

gyrus, and right superior frontal sulcus (all of which comprise various portions of BA's 6,

8, and 9). Right-hemisphere medial-frontal regions of interest were: right Pre-SMA

(medial BA 6), right SMA (medial BA 4), the right paracingulate gyrus (if present)

coupled with the top half of the right cingulate sulcus (BA 32), and the right cingulate

gyrus coupled with the bottom half of the right cingulate sulcus (BA 24). Right cingulate

and paracingulate sulci were also labeled individually if the distinction between BA 32

and 24 was unclear.

Only activated voxels with an R2 > .16 and a resulting cluster volume of greater

than 100 microliters were captured for analysis. Each participant's images were localized

by multiple raters to designate activity clusters in the selected regions of interest for our

study. Resulting cluster volumes for pre- and posttreatment activation in the selected

regions of interest were then compared.









Table 2-1. Participant Demographics
Participant Age Sex Education Months Post
(in years) stroke
01 48 Male 13 48
02 48 Female 14 8
03 52 Female 18 15
04 54 Female 14 24
05 74 Male 12 86

Table 2-2. Participant Performance on Language-Related Testing Pre- and Posttreatment
Participant WAB WAB WAB WAB BNT BNT
AQ AQ Comp Comp Pre-Tx Post-Tx
Pre-Tx Post-Tx Pre-Tx Post -Tx
01 79.6/100 84.6/100 172/200 196/200 50/60 51/60
02 81.4/100 84.1/100 196/200 187/200 44/60 44/60
03 27/100 33.4/100 98/200 111/200 1/60 1/60
04 32.3/100 39.2/100 105/200 140/200 1/60 1/60
05 68.8/100 77/100 154/200 162/200 39/60 47/60

Table 2-3. Participant Baseline Stability and Response to Treatment as Determined by 3
Independent Speech Pathologists
Participant-Phase Rater 1 Rater 2 Rater 3 Interrater
Agreement
01-Baseline Stable Stable Stable 100%
01-Treatment Responder Responder Responder 100%
02-Baseline Stable Stable Unstable 66.66%
02-Treatment Responder Responder Responder 100%
03-Baseline Stable Stable Stable 100%
03-Treatment Responder Responder Responder 100%
04-Baseline Stable Stable Stable 100%
04-Treatment NonResponder NonResponder NonResponder 100%
05-Baseline Unstable Unstable Unstable 100%
05-Treatment Responder Responder NonResponder 66.66%

Table 2-4. Partici ant C Statistics for Baseline and Treatment
Participant C-Statistic Stable C-Statistic Treatment
Baseline Baseline Treatment Responder
01 .1111 (p>.05) Yes .5523 (p<.01) Yes
02 .0871 (p>.05) Yes .8422 (p<.01) Yes
03 -1.2209 (p<.01) No .8039 (p<.01) Yes
04 -.125 (p>.05) Yes .0205 (p>.05) No
05 .9625 (p>.05) Yes .0233 (p>.05)) No























Figure 2-1. Participant 01's Lesion-Sagittal Image- A broad white outline encompassing
the general regions of each participant's lesion(s) are included in their images.


Figure 2-2. Participant 01's Lesion-Coronal Image


Figure 2-3. Participant 02's Lesion-Sagittal Image























igure 2-4. Participant 02's Lesion-Coronal Image


Figure 2-5. Participant 03's Lesion-Sagittal Image


Figure 2-6. Participant 03's Lesion-Coronal Image
























Figure 2-7. Participant 04's Lesion-Sagittal Image


Figure 2-8. Participant 04's Lesion-Coronal Image


figure z-r. artlcipant uD s Lesion-bagittal image



















irticipant 05's Lesion-Coronal Image


C 1. -


Figure 2-11. Formula for the C


(Xi Xi+)


N

2 (iX X)i

Statistic













BASELINE PHASE 1 PHASE PHASE






^--^Az


1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10
Phase and Day

Figure 2-12. Participant 01's Performance across Treatment Phases


100

90

80

70

S 60

0

40
0.

30

20

10

0


1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10
Phase and Day

Figure 2-13. Participant 02's Performance across Treatment Phases


BASELINE PHASE 1 PHASE PHASE















BASELINE PHASE 1 PHASE 2 PHASE 3


1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 910
PHASE AND DAY


Figure 2-14. Participant 03's Performance across Treatment Phases



100 lO

BASELINE PHASE 1 PHASE 2 PHASE 3
LnnII_ I


1u



1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10
Phase and Day


Figure 2-15. Participant 04's Performance across Treatment Phases







































1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9
Phase and Day

Figure 2-16. Participant 05's Performance across Treatment Phases














CHAPTER 3
RESULTS

Activated cluster voxels in the regions of interests for our study were calculated for

each participant's pretreatment and posttreatment scans. Though changes in individual

regions were calculated, the overall goal of our study was to examine global changes in

medial-frontal and lateral-frontal regions of the left and right hemisphere after treatment.

In addition, there are data to show that larger-scale regions of interest produce greater

reliability (Sabsevitz et al., 2003). Therefore, activity totals for each overall region were

also calculated and are presented along with images of each participant's activity in

Tables 3-1 through 3-5.

Participant 01 (Table 3-1) made 39 out of a possible 45 responses during the

pretreatment scan; therefore, 39 responses were entered into deconvolution analysis. The

maximum R2 value for Participant 01's pretreatment data was .57. Overall, Participant

01 's pretreatment data showed bilateral activity in the lateral-frontal regions, with a slight

favoring of left perilesional regions (9,907 activated voxels) over right lateral-frontal

regions (6,105 activated voxels). A small amount of left medial-frontal activity was

observed (334 activated voxels), whereas no right medial regions showed significant

activity pretreatment.

Posttreatment, Participant 01 made 45 out of a possible 45 responses; therefore, all

45 responses were entered into deconvolution analysis. The maximum R2 value for

Participant 01's posttreatment data was .58. Participant 01 continued to show bilateral

activation in lateral-frontal regions posttreatment; however, greater activity was observed









in right lateral-frontal regions (8,806 activated voxels) than left perilesional regions

(4,617 activated voxels) posttreatment. There was little change in medial-frontal activity

in either the left (220 activated voxels) hemisphere or the right hemisphere (0 activated

voxels) from pretreatment to posttreatment.

Pretreatment, Participant 02 (Table 3-2) made 44 out of a possible 45 responses;

therefore, 44 responses were entered into deconvolution analysis. The maximum R2

value for Participant 02's pretreatment data was .44. Participant 02 showed significantly

greater activity in right lateral-frontal (2,041 activated voxels) and right medial-frontal

regions (1,455 activated voxels) than in left perilesional (236 activated voxels) and left

medial-frontal (523 activated voxels) regions pretreatment.

Participant 02 made 45 out of 45 responses during her posttreatment scan;

therefore, all 45 responses were included in the deconvolution analysis. Participant 02's

maximum R2 value posttreatment was .40. Posttreatment, Participant 02 showed bilateral

activations, though she primarily showed activity in left-hemisphere regions, with

significantly increased activity in left perilesional regions (3,031 activated voxels).

Participant 02 also showed an increase in left medial-frontal activity (1,882 activated

voxels). Participants 02's right lateral activity remained relatively unchanged (2,084

activated voxels) from pretreatment to posttreatment, whereas her right medial-frontal

activity decreased significantly (600 activated voxels).

Participant 03 (Table 3-3) made 34 out of 45 responses during her posttreatment

scan; however, 7 responses had to be removed from analysis when scanner problems

occurred during the particular run in which those responses were made; therefore, 27

responses were included in the pretreatment deconvolution analysis. Participant 03's









maximum R2 value pretreatment was .55. Participant 03 showed no medial-frontal

activity in either hemisphere during her pretreatment scan. She showed some left

perilesional activity (1,615 activated voxels); however, she showed an unusually large

amount of right lateral-frontal activity (125,305 activated voxels), the vast majority of

which (123,734) was found in the right motor strip. Investigation of this unusually large

cluster of activity revealed what appeared to be true hemodynamic activity; therefore, the

cluster was included in analysis.

Participant 03 made 45 out of 45 responses during her posttreatment scan; however,

for consistency, the same run that had to be deleted from analysis during Participant 03's

pretreatment scan was deleted from her posttreatment scan. Therefore, only 36 responses

were included in Participant 03's deconvolution analysis posttreatment. Participant 03's

maximum R2 value posttreatment was .33. Participant 03 showed a significant decrease

in overall activity compared to her pretreatment scan. Activity was evidenced only in the

right hemisphere posttreatment. Some right medial-frontal activity was evident

posttreatment (150 activated voxels), whereas none was evident pretreatment. Right

lateral-frontal activity decreased significantly posttreatment (1,455 activated voxels).

Participant 04 (Table 3-4) made 44 out of 45 responses during her pretreatment

scan; therefore, 44 responses were included in her pretreatment deconvolution analysis.

Participant 04's maximum R2 value pretreatment was .40. Participant 04 showed very

little overall activity during her pretreatment scan. In fact, only Participant 04's left

medial-frontal regions (295 activated voxels) showed significant activity.

Participant 04 made 45 out of 45 responses during her posttreatment scan;

therefore, all responses were included in her posttreatment deconvolution analysis.









Participant 04's maximum R2 value posttreatment was also .40. Overall, Participant 04

showed evidence of some bilateral activation in her lateral-frontal regions posttreatment,

with more activation evident in left perilesional regions (3,186 activated voxels) than in

right lateral-frontal regions (1,755 activated voxels). Participant 04 also showed an

increase in left medial-frontal activity (1,607 activated voxels) compared to pretreatment.

Consistent with her pretreatment scan, no right medial-frontal activity was observed

posttreatment.

Pretreatment, Participant 05 (Table 3-5) made 43 out of 45 responses; therefore, 43

responses were entered into deconvolution analysis. The maximum R2 value for

Participant 05's pretreatment data was .32. Participant 05 showed bilateral medial-frontal

activations pretreatment (122 activated voxels left-hemisphere; 290 activated voxels right

hemisphere). However, Participant 05 showed no significant left perilesional or right

lateral-frontal activation pretreatment.

Participant 05 responded 36 out of 45 times posttreatment; therefore, 36 responses

were entered into Participant 05's posttreatment deconvolution analysis. The maximum

R2 value for Participant 05's posttreatment data was .30. Participant 05 showed only left

medial-frontal activity (130 activated voxels) during his posttreatment scan.










Table 3-1. Participant 01's Representative Activity Pre- and Posttreatment. Regions of
interest for our study are indicated by a broad white outline.


Left Lateral-frontal Pretreatment
0 007 ArtivtoA Vnxv-olc


Left Medial-frontal Pretreatment
2/1A Aolr+;o+IA Unvolo


Right Lateral-frontal Pretreatment
6 105 Activated Voxels


Right Medial-frontal Pretreatment
0 ACitjavtPd VnYlW


Left Lateral-frontal Posttreatment
A A1 7 ArtijntfPA /Vncwle


Left Medial-frontal Posttreatment
ODn Aotr-;oxIA U/nv-olo


Right Lateral-frontal Posttreatment
o 1 rn- A A T__ -,


Right Medial-frontal Posttreatment
0 AiCtCiv7d Vncxlv










Table 3-2. Participant 02's Representative Activity Pre- and Posttreatment


Left Lateral-frontal Pretreatment
236 Activated Voxels


Left Medial-frontal Pretreatment
S7? ActixajtPdl /VnvYle


Right Lateral-frontal Pretreatment
? 041 ActijvltdCI VnWole


Right Medial-frontal Pretreatment
1 455 Activated Voxels


Left Lateral-frontal Posttreatment
3.031 Activated Voxels


Left Medial-frontal Posttreatment
1 RR9 ACtiv7atpCd V/nvYle


Right Lateral-frontal Posttreatment
9 OR4 ActivatdCl VnvPol


Right Medial-frontal Posttreatment
,ann A -*_-_-J 1,--1-










Table 3-3. Participant 03's Representative Activity Pre- and Posttreatment


Left Lateral-frontal Pretreatment
1.615 Activated Voxels


Left Medial-frontal Pretreatment
0 AC tivt\ d cIVYol V


Right Lateral-frontal Pretreatment
1?S '1S Activatedt Vnvole


Right Medial-frontal Pretreatment
0 ACtivnf7tP VncWl


Left Lateral-frontal Posttreatment
0 Activated Voxels


Left Medial-frontal Posttreatment
0 AC tivXtPCd IVnVol


Right Lateral-frontal Posttreatment
1 4SS Activ+tdCI VWnvol


Right Medial-frontal Posttreatment
1 SO ACifvX7toPC VnWPle










Table 3-4. Participant 04's Representative Activity Pre- and Posttreatment


Left Lateral-frontal Pretreatment
0 Activated Voxels


Left Medial-frontal Pretreatment
790 ActiXaJtPdl /VnvYle


Right Lateral-frontal Pretreatment
0 Activated Voxels


Right Medial-frontal Pretreatment
0 ACtiX7qtPd VnpWl


Left Lateral-frontal Posttreatment
3.186 Activated Voxels


Left Medial-frontal Posttreatment
1 607 ACtiv7atpdc VnvYole


Right Lateral-frontal Posttreatment
1 755 Activated Voxels


Right Medial-frontal Posttreatment
0 ACtiPvtpd Vncxlv









Table 3-5. Participant 05's Representative Activity Pre- and Posttreatment


Left Lateral-frontal Pretreatment
0 Activated Voxels


Left Medial-frontal Pretreatment
19? A tiX7tad Pc VnVple


Right Lateral-frontal Pretreatment
0 Activated Voxels


Right Medial-frontal Pretreatment
290 Activated Voxels


Left Lateral-frontal Posttreatment
0 Activated Voxels


Left Medial-frontal Posttreatment
1 O0 A rCti vtd 1cI vpn ol


Right Lateral-frontal Posttreatment
0 Activated Voxels


Right Medial-frontal Posttreatment
0 Activated Voxels














CHAPTER 4
DISCUSSION

A word generation task requiring the overt production of a category exemplar was

used during fMRI scanning to reveal changes in neural substrates associated with

language production in five patients with chronic nonfluent aphasia before and after

undergoing an experimental language rehabilitation program designed to recruit intention

mechanisms in the medial-frontal regions of the language nondominant hemisphere,

thereby improving language function. It was hypothesized that changes in language

abilities posttreatment would be shown by changes in functional activation in the right

medial-frontal and lateral-frontal language homologous regions of the brain. It was

hoped that the results of our study would also help to inform the current debate in aphasia

rehabilitation literature as to whether or not improvements in language function in

patients with aphasia are associated with activation or perilesional regions of the left-

hemisphere or if they involve reorganization of language function to the language

nondominant right hemisphere.

Conclusions

The hypotheses for our study were not fully supported, and results varied from

patient-to-patient such that some patients (Participants 02 and 05) generally lent support

to the argument that improved language function in patients with nonfluent aphasia is

associated with activation in perilesional regions of the left-hemisphere, while others

(Participants 01 and 03) supported the argument that reorganization of language function

over to the language nondominant right hemisphere takes place. A summary of each









participant's overall changes in activity from pretreatment to posttreatment, and a brief

discussion of possible explanations for these findings, is discussed below.

Participant 01 was deemed a treatment responder by all 3 independent raters, and

his C statistic. This participant showed bilateral activation of lateral-frontal language

regions pretreatment. Posttreatment, Participant 01 continued to show bilateral

activations; however, a general shift in overall activation from left-hemisphere

perilesional regions to right hemisphere language homologous regions did take place.

These findings generally lend support to the reorganization of function theory (Abo et al.,

2004; Musso et al, 1999); however, due to the heavily subcortical nature of Participant

01's lesion, these results are most consistent with the findings of Crosson et al. (2005),

which suggest that bilateral activations may occur if left-hemisphere subcortical

structures are affected by the stroke.

Participant 02 was considered to be a treatment responder by all 3 raters and her C

statistic as well. She showed relatively dominant right-hemisphere language activity

during her pretreatment scan. Posttreatment, Participant 02, like Participant 01, showed

more bilateral activation; however, greater activity was observed in the left-hemisphere

perilesional regions than in the right hemisphere posttreatment. Participant 02 also had

some subcortical damage from her lesion, and her observed bilateral activations are also

consistent with those found in Crosson et al. (2005). However, her general tendency

towards left-hemisphere activation posttreatment is somewhat contradictory to the right-

hemisphere favoring of Participant 01 and might be explained by the less severe damage

to left-hemisphere subcortical structures in Participant 02.









Participant 03, like Participants 01 and 02, was deemed a treatment responder by

both the independent panel of raters and her C statistic. Pretreatment, Participant 03

showed primarily right-hemisphere language activity with some left lateral-frontal

activity. Posttreatment, she showed only right-hemisphere language activity. Participant

03's findings generally lend support to the arguments of Musso et al. (1999) and Abo et

al. (2004) that improvements in language function are associated with reorganization of

function over to the right hemisphere. Based on the size of Participant 03's lesion,

however, it could be argued that reorganization to the right hemisphere occurred simply

because there were insufficient perilesional areas left intact by her stroke. These

arguments would be in line with the findings of Cao et al. (1999).

Participant 04 was not considered to be a treatment responder by all 3 raters nor by

her C statistic. Participant 04 showed no right-hemisphere activation pretreatment. She

did show some left-hemisphere medial-frontal activation, however. Posttreatment,

Participant 04 showed primarily left-hemisphere perilesional activations, though she did

show some right lateral-frontal activation as well. In general, her results are consistent

with the findings ofMeinzer et al. (2004) and Cao et al. (1999) that suggest perilesional

activation is seen after rehabilitation in patients with aphasia; however, it should be noted

that, since Participant 04 was not considered to be a treatment responder, her results may

not be generalizable to those who did respond to treatment.

Participant 05 was considered to be a treatment responder by 2 out of 3

independent raters; however, Participant 05 was not considered to be a treatment

responder by one rater nor by his C statistic. Participant 05 showed some left and right-

hemisphere medial-frontal activity pretreatment. Posttreatment, however, only a slight









increase in left-hemisphere medial-frontal activity was observed. Participant 05's

findings, like those of Participant 04, generally lend support to the arguments for

perilesional activation but may also not be generalizable as it is possible that he did not

possess a stable baseline before the initiation of treatment.

Based on the results of our study, it is concluded that the most important factor in

determining whether changes in activation after language rehabilitation occur in

perilesional regions of the left-hemisphere or right-hemisphere language homologous

regions appears to be the size, location, and overall severity of the lesion itself. These

findings seem most consistent with the work of Cao et al. (1999). It is hypothesized from

these findings that the brain may in fact follow a general hierarchical pattern of

reorganization after brain injury. This theory posits that, if sufficient left-hemisphere

perilesional regions remain intact after stroke, a general increase in activity will be

observed in these perilesional regions after improvements in language functioning.

However, if the size of the individual's lesion is too significant, or there are insufficient

perilesional regions available to resume language functioning, a shift of language

functions to right-hemisphere homologous language regions will be observed. Finally,

consistent with the findings of Crosson et al. (2005), if left-hemisphere subcortical

regions are affected by the stroke, bilateral activations of language regions may in fact be

observed.

Another particular topic of interest regarding the results of our study involves the

issue of sex- and age-related differences in lateralization. Recent functional

neuroimaging studies on lateralization have led to the development of the Hemispheric

Asymmetry Reduction in Older Adults (HAROLD) hypothesis due to general findings









that brain activity in individuals tends to become less lateralized as age increases

(Cabeza, 2002; Reuter-Lorenz et al., 2000). Using the HAROLD hypothesis, it could be

theorized that older participants in our study would be more likely to show reorganization

of language functioning over to the nondominant right hemisphere or bilateral activations

due to naturally decreased lateralization effects with age. The findings of our study are

somewhat ambivalent in regards to this hypothesis with some patients generally lending

support to this hypothesis (Participants 01, 02, and 03) and others whose data do not

support this hypothesis (Participants 04 and 05); however, it should be noted that, with

the exception of one outlier (Participant 05), all of the subjects in our study were

relatively similar in age (between 48 and 54). It is clear from these studies that the

HAROLD hypothesis should also be taken into consideration when evaluating the results

of studies relating to lateralization and reorganization of neural functioning.

Inherent sex differences of lateralization should also be considered when reviewing

the results of our study. Research findings have shown that women consistently perform

better than men on verbal tasks (McGlone, 1980) and that men often perform better than

women on tasks requiring more right-hemisphere dominant visuospatial abilities

(Halpern, 2000; Voyer, Voyer, & Bryden, 1995). Based on these findings, it could also

be theorized that men, given their greater performance on right-hemisphere visuospatial

tasks, would be more likely than women to show reorganization of language functioning

over to the right hemisphere after treatment in our study, whereas women, with greater

performance on verbal tasks, which are strongly lateralized to the left-hemisphere in most

individuals, would be more likely to show perilesional activity in the left-hemisphere









after treatment. Further research with significantly larger samples is necessary before

any theoretical conclusions on this issue can be made.

Limitations

There are several limitations to our study that limit its scope and generalizability.

First, the unique nature of the experimental language rehabilitation program that was

used in conjunction with our study makes it unclear whether similar results would be

found using other language rehabilitation therapies. Also, given that the language

rehabilitation program used was specifically designed to improve language output, it is

doubtful that patients with other forms of aphasia (e.g., fluent aphasia) would show

similar results to the participants in our study. However, it is interesting to note that

recent work by Cato, Parkinson, Wierenga, & Crosson (2004) suggest that one of the best

predictors of outcome in the experimental treatment these participants underwent is

language comprehension.

Another possible limitation for our study involves the difference between the

category exemplar task used during both the pre- and posttreatment fMRI scanning

sessions and the naming task used during the language rehabilitation program itself. As

stated previously, the category exemplar task was used for each participant's fMRI

scanning sessions because previous research in this field had shown that the task reliably

activated medial-frontal regions in normal controls (Crosson et al., 1999). However, it is

unknown at this time whether having participants perform the naming task that was used

during the language rehabilitation treatment during each of their scanning sessions would

yield similar results.

Our study is also limited by its small and relatively homogenous sample size. Due

to several confounding factors including the high costs of fMRI scanning, the dearth of









patients with chronic nonfluent aphasia in the regional area who meet inclusion criteria

for this type of study, and the time- and personnel-intensive nature of the language

rehabilitation therapy itself, it is a challenge to produce studies of this type with large and

diverse sample sizes. Also, due to the risks of confounding variables, many different

groups of individuals (e.g., learning disabled, individuals with neurological disorders

other than stroke, individuals who were not fMRI compatible, etc.) were excluded from

our study. It is unclear if the findings presented here would also apply to those

populations.

Implications and Future Directions

The practical implications for research in this general field of study are significant.

One example of the benefits of this type of research is that these studies provide us with a

better understanding of how language functions occur in individuals both normally and

after brain injury. Knowledge of this type is vital to the future developments of many

fields of study including Education, Neurology, Neuropsychology, Neurosurgery, and

Rehabilitation. The most direct benefit from continued research in this area is the future

development of new rehabilitation programs and techniques tailored to the needs of each

individual patient based on their lesion size, location, and severity. This personalized

approach to treatment will maximize treatment effectiveness and help ensure sustained

improvements in language functioning, which, as discussed previously, can have a

tremendous impact on each individual's quality of life and overall independence,

subsequently decreasing their impact on the economy by increasing their likelihood of

returning to work and decreasing their needs for government assistance.

Future directions for research in this field of study should include conducting

studies with larger sample sizes, so that more adequate conclusions can be drawn. Also,









studies involving comparisons of different groups of patients, with each group containing

patients with highly similar lesions would prove beneficial as they would allow us to

examine how each group performs and examine if specific lesion sizes and locations

successfully predict changes in activity in specific brain regions after language

rehabilitation. Further research on the role subcortical mechanisms play as determinants

of cortical reorganization after language rehabilitation is also strongly recommended.

Also, further research on the actual methods of the treatment paradigm used in our study

is warranted. Possible directions for such research could include altering the treatment

such that individual participants do not progress from Phase one to Phase two or from

Phase two to Phase three of the treatment until upward trends in their performance are

observed in each Phase. Longitudinal research on treatment outcomes from patients who

participated in the language rehabilitation treatment used in our study and whether their

improvements made in treatment are still present at set periods of time (e.g., 1 month, 6

months, 2 years) after treatment would also prove highly informative. Additional

outcomes research on possible behavioral and environmental predictors of treatment

responders vs. individuals who did not appear to respond to treatment might also prove

beneficial.















LIST OF REFERENCES


Abdullaev, Y. G., & Posner, M. I. (1998). Event-related brain potential imaging of
semantic encoding during processing single words. Neurolmage, 7, 1-13.

Abo, M., Senoo, A., Watanabe, S., Miyano, S., Doseki, K., Sasaki, N., Kobayashi, K.,
Kikuchi, Y., & Yonemoto, K. (2004). Language-related brain function during word
repetition in post stroke aphasics. Neuroreport, 15 (12), 1891-1894.

Anderson, S. W., & Damasio, S. W. (1993). The frontal lobes. Clinical
Neurospychology (3rd Edition). New York: Oxford University Press, 409-460.

Belin, P., Van Eeckhout, P., Zilbovicius, M., Remy, P., Francois, C., Guillaume, S.,
Chain, F., Rancurel, G., & Samson, Y. (1996). Recovery from nonfluent aphasia
after melodic intonation therapy: A PET study. Neurology, 47, 1504-1511.

Binder, J. R., Frost, J. A., Hammeke, T. A., Cox, R. W., Rao, S. M., & Prieto, T. (1997).
Human brain language areas identified by functional magnetic resonance imaging.
The Journal of Neuroscience, 17, 353-362.

Bobrovitz, C. D., & Ottenbacher, K. J. (1998). Comparison of visual inspection and
statistical analysis of single-subject data in rehabilitation research. American
Journal of Physical Medicine and Rehabilitation, 77 (2), 94-102.

Broca, P. (1997). On the speech center. In L. Benjamin, Jr. (Ed.), A history of
psychology: Original sources and contemporary research (pp.77-81). New York:
The McGraw-Hill Companies, Inc. (Original work published 1861).

Buffery, W. H., & Burton, A. (1982). Information processing and redevelopment:
towards a science of neuropsychological rehabilitation. In A. Burton (ed.) The
Pathology and Psychology of Cognition. London: Methuen.

Buxbaum, L. J., Coslett, H. B., Shall, R. R., & McNally, B. (1993). Hemispatialfactors
in mirror writing. Neuropsychologia, 31, 1417-1421.

Cabeza, R. (2002). Hemispheric Asymmetry Reduction in Older Adults: The HAROLD
Model. Psychology andAging. 17 (1), 85-100.

Cao, Y., Vikingstad, E. M., George, K. P., Johnson, A. F., & Welch, K. M. A. (1999).
Cortical language activation in stroke patients recovering from aphasia with
functional MRI. Stroke, 30, 2331-2340.









Cappa, S. F. (2000). Neuroimaging of recovery from aphasia. Neuropsychological
Rehabilitation, 10, 365-376.

Cato, M., Parkinson, R., Wierenga, C., & Crosson, B. (2004). Predicting rehabilitative
treatment success in chronic nonfluent aphasia: Lesion and performance
characteristics. [Abstract to be published online]. International
Neuropsychological Society.

Chen, W., & Ogawa, S. (2001). Principle of BOLD-functional MRI. In C. T. W.
Moonen & P. A. Bandettini (Eds.), Functional MRI (pp. 103-113). Berlin:
Springer.

Code, C. (1982). Neurolinguistic analysis of recurrent utterance in aphasia. Cortex, 18,
141-152.

Code, C. (1989). Language, aphasia, and the right hemisphere. Publisher unknown.

Cox, R.W. (1996). AFNI: Software for analysis and visualization of functional magnetic
resonance Neuroimages. Computational Biomedical Research, 29, 162-173.

Crosson, B., Moore, A. B., Gopinath, K., White, K. D., Wierenga, C. E., Gaiefsky, M. E.,
Peck, K. K., Soltysik, D., Briggs, R. W., & Rothi, L. J. G. (2005). Role of the right
and left hemispheres in recovery of function during treatment of intention in
aphasia. Journal of Cognitive Neuroscience, 17 (3), 392-406.

Crosson, B., & Rothi, L.J.G. (1999). Subproject: Treatment of attention and intention in
aphasia. VA RR & D grant number C2602V, NIDCD grant number 03888., 253-
276. Bethesda, MD: National Institute on Deafness and Other Communication
Disorders: National Institutes of Health.

Crosson, B., Sadek, J.R., Bobholz, J.A., Gockay, D., Mohr, C.M., Leondard, C.M.,
Maron, L., Auerbach, E.J., Brows, S.R., Freeman, A.J., & Briggs, R.W. (1999).
Activity in the paracingulate and cingulate sulci during word generation: An fMRI
study of functional anatomy. Cerebral Cortex, 9, 307-316.

Crosson, B., Sadek, J. R., Maron, L., Gockay, D., Mohr, C. M., Auerbach, E.J. Freeman,
A. J., Leonard, C. M. & Briggs, R. W. (2001). Relative shift in activity from
medial to lateral-frontal cortex during internally versus externally guided word
generation. Journal of Cognitive Neuroscience, 13, 272-283.

Elder, J. H. (1997). Single Subject Experimentation in Psychiatric Nursing. Archives of
Psychiatric Nursing, 11 (3), 133-138.

Fiez, J. A. (2001). Neuroimaging studies of speech: An overview of techniques and
methodological approaches. Journal of Communication Disorders, 34, 445-454.









Finger, S. (1997). Cortical localization and cerebral dominance: The work of Paul
Broca. In L. Benjamin, Jr. (Ed.), A history of psychology: Original sources and
contemporary research (pp. 77-81). New York: The McGraw-Hill Companies, Inc.
(Original work published 1994).

Foundas, A. L. Eure, K. F., Luevano, L. F., & Weinberger, D. R. (1998). MRI
asymmetries of broca's area: The pars triangularis and pars opercularis. Brain and
Language, 64, 282-295.

Francis, W. N., & Kucera, H. (1982). Frequency analysis of English usage: Lexicon and
grammar. Boston: Houghton Mifflin Co.

Gaiefsky, M. (2003). FMRI of overt language production in aphasia rehabilitation: The
contribution of the language nondominant hemisphere. UnpublishedMasters
Thesis, University of Florida, 1-39.

Halpern, D. F. (2000). Sex differences in cognitive abilities. Mahwah, NJ: Lawrence
Erlbaum Associates, Inc.

Hanlon, R. E., Brown, J. W., & Gerstmann, L. J. (1990). Enhancement of naming in
nonfluent aphasia through gesture. Brain and Language, 38 (2), 298-314.

Hausmann, M., Gunturkun, 0., & Corballis, M. C. (2003). Age-related changes in
hemispheric asymmetry depend on sex. Laterality, 8 (3), 277-290.

Heilman, K.M.,Watson, R.T., & Valenstein, E. (1993). Neglect and related disorders.
ClinicalNeuropsychology (3rd Edition). New York: Oxford University Press, 279-
336.

Karbe, H., Thiel, A., Weber-Luxenberger, G., Herholz, K., Kessler, J., & Heiss, W. D.
(1998). Brain plasticity in post stroke aphasia: What is the contribution of the right
hemisphere? Brain and Language, 64, 215-230.

Kazdin, A, (1982). Single-case research design: methods for clinical and applied
settings. New York: Oxford University Press.

Lichtheim, L. (1885). On aphasia. Brain, 7, 433-484.

McGlone, J. (1980). Sex differences in human brain asymmetry: A critical survey.
Behavioral and Brain Sciences, 3, 215-263.

Meinzer, M., Elbert, T., Wienbruch, C., Djundja, C., Barthel, G., & Rockstroh, B. (2004).
Intensive language training enhances brain plasticity in chronic aphasia. BMC
Biology, 2, 20.









Moore, A. B., Crosson, B., Gockay, D., Leonard, C. M., & Foundas, A. L. (2000).
Defining pars triangularis. Society for Neuroscience Abstracts, 30, Part 1, 1247.

Moore, A. B., Loftis, C., Crosson, B., Gockay, D., Leonard, C. M., & Foundas, A. L.
(2001). Localization of functional activity in pars triangularis. Journal of the
International Neuropsychological Society, 7, 197.

Musso, M., Weiller, C., Kiebel, S., Muller, S. P., Bulau, P., & Rijntjes, M. (1999).
Training-induced brain plasticity in aphasia. Brain, 122, 1781-1790.

Nadeau, S., Rothi, L. J. G., & Crosson, B. (2000). Preface. In S. Nadeau, L. J. G. Rothi,
& B. Crosson (Eds.), Aphasia and language: Theory to practice. New York:
Guilford Press.

National Aphasia Association. (1999). Aphasia Fact Sheet. Retrieved March 5, 2005,
from http://www.aphasia.org/NAAfactsheet.html.

Nourbakhsk, M. R., & Ottenbacher, K. J. (1994). The statistical analysis of single-
subject data: a comparative examination. Physical Therapy, 74 (8), 768-776.

Reuter-Lorenz, P. A., Jonides, J., Smith, E. S., Hartley, A., Miller, A., & Marshuetz, C.
(2000). Age differences in the frontal lateralization of verbal and spatial working
memory revealed by PET. Journal of Cognitive Neuroscience, 12, 174-187.

Richards, K., Singletary, F., Rothi, L.J.G., Koehler, S., & Crosson, B. (2002). The
activation of intentional mechanisms through utilization of nonsymbolic
movements in aphasia rehabilitation. Journal of Rehabilitation Research and
Development, 39, 445-454.

Rosen, H. J., Pertersen, S. E., Linenweber, M. R., Snyder, A. Z., White, D. A., Chapman,
L., Dromerick, A. W., Fiez, J. A., & Corbetta, M. (2000). Neural correlates of
recovery from aphasia after damage to left inferior frontal cortex. Neurology, 55,
1883-94.

Rothi, L.J.G. (1995). Behavioral compensation in the case of treatment of acquired
language disorders resulting from brain damage. Compensating for psychological
deficits and declines: Managing losses and promoting gains. Mahwah, NJ:
Lawrence Erlbaum Associates, Inc. 219-230.

Sabsevitz, D. S., Swanson, S. J., Hammeke, T. A., Spanaki, M. V., Possing, E. T., Morris,
III, G. L., Mueller, W. M., & Binder, J. R. (2003). Use of preoperative functional
neuroimaging to predict language deficits from epilepsy surgery. Neurology, 60,
1788-1792.









Snyder, P. J., & Nussbaum, P.D. (Eds.) (1998). Clinical Neuropsychology: A pocket
handbook for assessment. Washington, DC: American Psychological Association
Press Inc.

Talaraich, J., & Tournoux, P. (1988). Co-Planar Stereotaxic Atlas of the Human Brain.
New York: Thieme Medical Publishers.

Tryon, W. W. (1982). A simplified time-series analysis for evaluating treatment
interventions. Journal of Applied Behavior Analysis, 15, 423-429.

Villringer, A. (2001). Physiological Changes During Brain Activation. In C. T. W.
Moonen & P. A. Bandettini (Eds.), FunctionalMRI (pp. 3-13). Berlin: Springer

Voyer, D., Voyer, S., & Bryden, M. P. (1995). Magnitude of sex differences in spatial
abilities: A meta-analysis and consideration of critical variables. Psychological
Bulletin, 117, 250-270.

Wernicke, C. (1874). Der Aphasische Symptomencomplex. Breslau: Cohn and Weigert.















BIOGRAPHICAL SKETCH

Bradley J. Daniels was born in Orlando, Florida, and graduated Summa Cum Laude

from the University of Central Florida in May 2003 with a B.A. in psychology. He

immediately entered graduate school in the Department of Clinical and Health

Psychology at the University of Florida, where he is pursuing a Ph.D. in clinical

psychology, specializing in clinical neuropsychology.