|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
1 CORTICAL SUBCORTICAL FUNCTIONS IN VERB AND NOUN GENERATION B y MICHELLE L. BENJAMIN A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2011
2 2011 Michelle L. Benjamin
3 ACKNOWLEDGMENTS I would l ike to acknowledge my family, friends, dissertation committee and labmates for their steady guidance and support. I extend sincere gratitude to K eith McGregor and Atchar Sudhyadhom for their valuable contributions regarding the technical aspects of the project Lastly, I would especially like to acknowledge Dr. Bruce Crosson for providing remarkable mentorship throughout my doctoral training.
4 TA BLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 3 LIST OF TABLES ................................ ................................ ................................ ............ 6 LIST OF FIGURES ................................ ................................ ................................ .......... 7 ABSTRACT ................................ ................................ ................................ ..................... 8 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 10 2 BASAL GANGLIA ANATOMY, FRONTAL STRIATAL CIRC UITRY, AND THE ROLE OF THE BASAL GANGLIA IN COGNITION ................................ ................. 11 General circuitry and anatomy of the basal ganglia ................................ ................ 11 Organization of f rontal striatal circuitry ................................ ................................ .... 12 Implications for the basal ganglia in language ................................ ........................ 14 3 FUNCTIONAL NEUROIMAGING OF LANGUAGE IN NEUROLOG ICALLY HEALTHY POPULATIONS ................................ ................................ ..................... 19 4 CLINICAL SUPPORT FOR DIStINCTIONS IN VERB NOUN PRODUCTION ........ 24 5 HYPOTHESES ................................ ................................ ................................ ....... 26 will be implicated in aging, as measured using verb verb word generation. ........ 26 will also be implicated in aging, as measured using noun noun word generation. ................................ ................................ ................................ ........... 27 6 METHODS ................................ ................................ ................................ .............. 28 Participants ................................ ................................ ................................ ............. 28 Procedures ................................ ................................ ................................ ............. 29 Neuropsychological testing ................................ ................................ ............... 29 Noun Verb Experimental FMRI Paradigm ................................ ........................ 31 Image Acquisition ................................ ................................ ................................ ... 32 Data analysis ................................ ................................ ................................ .......... 33 Behavioral data ................................ ................................ ................................ 33 Neuroimaging data and FMRI analyses ................................ ........................... 33
5 7 RES ULTS ................................ ................................ ................................ ............... 36 Neuropsychological testing ................................ ................................ ..................... 36 Noun and verb generation outside of the fMRI scanner ................................ .......... 36 FMRI results: Old versus young noun and verb generation comparisons ............... 36 Young versus baseline noun and verb generation comparisons ............................. 38 Old versus baseline noun and verb generation comparisons ................................ 38 A priori ROI AUC time course data ................................ ................................ ......... 39 Noun versus verb generation in young and old adults ................................ ............ 40 8 DISCUSSION ................................ ................................ ................................ ......... 41 APPENDIX: NOUN AND VERB GENERATION STIMULI ................................ ............. 61 LIST OF REFERENCES ................................ ................................ ............................... 63 BIOGRAPHICAL SKETCH ................................ ................................ ............................ 69
6 LIST OF TABLES Table page 8 1 Demographic and neuropsychological testing data for young and old subjects. ................................ ................................ ................................ ............. 49 8 2 FMRI region of interest (ROI) area under the curve time cou rse data. Young versus old t test comparisons for noun and verb generation. ............................. 50 8 3 Correlations between fMRI word generation active ROIs and word generation performance. ................................ ................................ ................................ ...... 51 8 4 FMRI region of interest (ROI) area under the curve time course data. Young t test comparisons for noun and verb generation versus baseline. ..................... 52 8 5 FMRI region of interest (ROI) area under the curve time course data. Old t test comparisons for noun and verb generation versus baseline. ....................... 53 8 6 Anatomy based region of interest (ROI) ar ea under the curve time course data. Young versus old repeated measures analysis of variance (ANOVA) data for lateral and medial frontal ROIs during noun and verb generation. ........ 55 8 7 Anatomy b ased region of interest (ROI) area under the curve time course data. Young versus old repeated measures analysis of variance (ANOVA) data for subcortical ROIs during noun and verb generation. .............................. 56 A 1 Noun and verb generation stimuli ................................ ................................ ....... 61
7 LIST OF FIGURES Figure page 8 1 Schematic diagra m of pre SMA basal ganglia loop ................................ ......... 57 8 2 Functional activity during noun and verb generation for young versus old adults. ................................ ................................ ................................ ................. 58 8 3 Impulse response function time course data for lateral an d medial frontal ROIs. ................................ ................................ ................................ .................. 59 8 4 Impulse response function time course data for subcortical ROIs. ..................... 60
8 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy CORTICAL SUBCORTICAL FUNCTIONS IN VERB AND NOUN GENERATION By Michelle L. Benjamin December 2011 Chair: Bruce Crosson Major: P sychology Neuroimaging studies of word generation to date have focused primarily on phonemic fluency and category member generation for objects. Given that behavioral studies of verba l fluency have found that verbal fluency for actions may strongly impli cate frontal striatal neuroanatomical systems and that imaging studies in healthy aging have implicated bilateral lateral frontal activity in old adults but not young adults, the current study sought to compare noun and verb generation in old and young hea lthy adults using functional magnetic resonance imaging (fMRI). Fifteen old adults (age range: 71 to 89 years) and 15 young adults (age range: 18 to 32 years) were compared for the current fMRI study of covert word generation. Subjects were presented wit h blocks of either three nouns or three verbs and were asked to generate silently an item semantically related to the stimulus within the same grammatical class for each of the three stimuli as they were presented. In whole brain between group analyses, s ignificant group differences were found in the right precentral gyrus for the noun condition and in the right inferior frontal sulcus for the verb condition, with old subjects demonstrating greater activity in these cortical regions. Significant group dif ferences were also found for both generation conditions in left posterior regions typically
9 associated with semantic processing (i.e., left angular gyrus), with old subjects demonstrating greater activity in the angular gyrus in both generation conditions. In addition, a priori anatomy based ROIs showed significant between group differences over the fMRI time course for the subthalamic nucleus (STN) bilaterally and left caudate nucleus during noun generation, with greater STN activity found in old adults b ut greater left caudate activity in young participants. Within group whole brain analyses indicated significant bilateral lateral frontal, medial frontal, and subcortical activity for both age groups during both types of word generation, implicating a bil ateral cortical subcortical network for the current noun and verb generation task regardless of age. Findings from the study suggest that the current paradigm is capable of eliciting an extensive cortical subcortical network for evaluating differences in language production in the healthy aging population and may be applicable for patient populations with frontal striatal involvement.
10 CHAPTER 1 INTRODUCTION The current project studied aging differences in language production via function magnetic resonan ce imaging (fMRI) using a noun and verb generation paradigm. While the current study did not propose to examine performance in clinical populations, the impetus behind the current paradigm was a clinical one. The longer term objective of this study was t o develop a functional neuroimaging language paradigm with adequate sensitivity for detecting subcortical activation during language tasks that can be implemented using fMRI in clinical populations with known neuropathology affecting frontal striatal syste ms. Chapter 2 will first summarize the structures and functions of the basal ganglia. An outline of relevant functional imaging studies of language is provided in Chapter 3. Research studies of neurological patient populations pertinent to the current s tudy are discussed in Chapter 4. Hypotheses for the current study are presented in Chapter 5. Methodology for the current study is outlined in Chapter 6. The results for the fMRI noun and verb generation task are presented in Cha pter 7 with discussion of study findings following in Chapter 8
11 CHAPTER 2 BASAL GANGLIA ANATOM Y, FRONTAL STRIATAL CIRCUITRY, AND THE ROLE OF THE BASAL GANGLIA IN COGNITION The basal ganglia are a set of subcortical anatomical structures that have important neuroregulatory implications for motor and cognition functions. In order to understand the functions of the basal ganglia, one needs to have a clear understanding of the ir anatomy. The following text and Figure 8 1 provide an introduction to basal ganglia anatomy and c General circuitry and anatomy of the basal ganglia Anatomical structures of the basal ganglia. The basal ganglia are a set of subcortical anatomical structures with neuroregulatory impli cations for cognitive and motor functions. The neurons of the basal ganglia themselves do not hold task specific information, but instead function in conjunction with other cortical and subcortical structures via cortical striatal pallidal thalamo cortica l circuits to enhance target behaviors or to suppress competing behaviors (Mink, 1996; Nambu, Tokuno, & Takada, 2002). The basal ganglia consist of the striatum, globus pallidus, subthalamic nucleus (STN), and substantia nigra (SN). The striatum can be f urther delineated anatomically into the neostriatum consisting of the caudate nucleus and putamen and the archistriatum consisting of the nucleus accumbens and olfactory tubercle. The globus pallidus can also be divided into components consisting of the i nternal globus pallidus (GPi) and external globus pallidus (GPe); a ventral pallidum is connected to the archistriatum. Neurotransmission of the basal ganglia. In general, the basal ganglia receive cortical input via excitatory glutaminergic neurot ransmission, and the output of the basal ganglia goes back to the cortex by means of excitatory glutaminergic neurotransmission
12 from the thalamus. STN neurotransmission is excitatory via glutamate to the medial globus pallidus. Neurotransmission from the substantia nigra pars compacta (SNpc), which projects to the neostriatum, can be either inhibitory or excitatory via dopamine. Neurotransmission from the striatum and globus pallidus to target structures is inhibitory via gamma amino butyric acid (GABA). Striatal afferents are the cortex or SNpc, and striatal efferents are the segments of the globus pallidus. The major afferent to the globus pallidus is the striatum, and the target structures of the globus pallidus efferents are the thalamic nuclei. In hibition of the thalamus via the globus pallidus is dependent on input from the striatum. For example, if there is more input from the striatum, this will result in greater inhibition of the globus pallidus. This pallidal inhibition will in turn will res ult in less inhibition of the thalamus, resulting in greater thalamic excitatory output to the cortex. Organization of frontal striatal circuitry Frontal striatal parallel circuits. While the aforementioned information provides a basic schema for the an atomy of the basal ganglia structures, one must understand the concept of parallel functioning in frontal striatal circuitry in order to understand the neuroregulatory role of the basal ganglia. The basic understanding of parallel cortical subcortical cir cuitry is that the cortex forms closed loop circuits with the basal ganglia and thalamus (Alexander, Delong, & Strick, 1986; Middleton & Strick, 2000a, 2000b). There is somatotopic organization within these frontal striatal circuits, in that specific cort ical areas project to specific portions of the STN, striatum, and globus pallidus, which then in turn project to specific thalamic nuclei and then back to the cortex from which the cortical projection originated. Alexander, Delong, and Strick (1986) initi ally described five cortical subcortical closed loop parallel circuits originating from and
13 projecting to areas within the frontal lobes. These five loops consist of the supplementary motor area (SMA), the frontal eye fields (FEF), the dorsolateral prefro ntal cortex (DLPFC), orbitofrontal cortex (OFC), and anterior cingulate cortex (ACC). Middleton and Strick (2000b) further delineated these circuits, and Figure 8 1 provides an example of one such cortical subcortical loop using the pre supplementary moto r area (pre SMA) as an example Temporal aspects of frontal striatal circuits. In addition to the structural organization, there are temporal aspects of the basal ganglia circuitry to consider with regards to the neuroregulatory role of the basal ganglia Traditionally, frontal striatal circuits have been characterized in terms of two pathways -the direct and indirect pathways. The direct pathway is thought to enhance selected behaviors, whereas the indirect pathway is thought to suppress competing beha viors (Mink, 1996) In the direct pathway, the cortical projection to the striatum is excitatory via glutamate. Striatal thalamus and are also inhibitory via GABA. The glutamate back to the cortex. In the indirect pathway, the cortex also projects to the striatum using the excitatory neurotransmitter glutamate. However, the striatum then projects to the GPe using GABA. The outpu t from the GPe then goes to STN and uses the inhibitory neurotransmitter GABA. The output of the STN is excitatory to the GPi using glutamate. GPi then projects to the thalamus and is inhibitory using the inhibitory neurotransmitter GABA. Lastly, the th projections back to the cortex. Mink (1996) outlined in great detail the functions of these
14 structures and circuits and proposed a theoretical center surround model of the excitatory and inhibitory influ ences of the direct and indirect pathways, respectively. In addition to the direct and indirect pathways, Nambu and colleagues (2000, 2002) have more recently described the role of the STN and a third pathway, termed the hyperdirect pathway. This hyperd irect pathway bypasses the striatum and is directed from the cortex directly to the STN, which then projects to the GPi. As previously outlined, the STN also participates in the indirect pathway via excitation to the GPi. However, in the hyperdirect path way, cortical input bypasses the striatum altogether, projecting directly to the STN which then projects to the GPi. Nambu and colleagues (2000) suggest that output resulting from the hyperdirect cortico STN pallidal thalamo cortical loop occurs more quic kly than that exhibited by the direct and indirect pathways. of behavior prior to the enhancement and suppression actions of the direct and indirect pathways, respe ctively. The cortical subcortical loop displayed in Figure 8 1 provides an overview of the projections for the hyperdirect, direct, and indirect pathways. Implications for the basal ganglia in language Knowledge of the aforementioned basal ganglia circ uitry helps to provide a framework for conceptualizing the implications of the basal ganglia in language functions. As previously mentioned, the role of the basal ganglia involves neuroregulation within cortical subcortical networks. One influential func tional neuroimaging study examining the role of the basal ganglia in language functions was a study done by Crosson and colleagues (2003) investigating the contributions of the left and right basal ganglia and frontal activation during word generation in y oung adults. Their study involved examining thr ee types of generation stimuli [ phonological
15 (nonsense syllable), lexical (rhyming), and semantic (category member)] and the findings of the latter two paradigms are the most pertinent to the current researc h proposal. During lexical retrieval (rhyming and category member generation but not phonological generation), neuroanatomical areas consisting of the left pre supplementary motor area (preSMA), dorsal caudate, and ventral anterior (VA) nucleus of the tha lamus were active. In addition, right basal ganglia activity in the caudate and putamen was present during the rhyming and semantic category generation tasks, in the relative absence of right frontal activity. Crosson and colleagues (2003) suggested that the left preSMA caudate VA thalamus serve as a frontal striatal loop involved in word retrieval for language. Furthermore, they proposed that the right basal ganglia activity acts to suppress right frontal activity that may interfere with word production Additional support for the role of the basal ganglia in language paradigms comes from Crosson and colleagues (Crosson, Benjamin, & Levy, 2007) who have postulated that the basal ganglia may be involved in an early suppression, intermediate enhancement, and late suppression cycle of behavior during complex language tasks such as word generation. They suggested that the role of the basal ganglia for cognitive tasks may be similar to the enhancement and early and late suppression cycles suggested by Mink ( 1996) and Nambu and colleagues (2002) for motor behaviors that stem from the hyperdirect, direct, and indirect pathways of the basal ganglia, respectively. Within the language framework, Crosson and colleagues (2007) suggested that for a task such as word generation, there is an initial resetting of the system to allow for a previous item generated from a target category to become inactive. Enhancement for relevant new items for the category then occurs, providing
16 options for the next category exemplar fo r word generation. A second wave of suppression then occurs once a new exemplar is selected; this suppression allows for the new exemplar to be specifically selected and suppresses competing exemplars. Empirical support for the early suppression interm ediate enhancement late suppression cycle of basal ganglia neuroregulation in language functions can be derived from behavioral paradigms. Copland and colleagues (Copland, Chenery, & Murdoch, 2000; Copland, 2000; Copland, 2003) have demonstrated the impli cations of subcortical involvement in complex language tasks during controlled processing using behavioral semantic priming paradigms in patient populations with either ischemic basal se (PD). Copland (2000) administere d a word triplet priming task to PD patients, patients with nonthalamic subcortical lesions of the dominant hemisphere, patients with cortical lesions of the dominant hemisphere, and neurologically healthy controls. At a short stimulus onset asynchrony (S OA), all patient groups performed similarly to controls demonstrating priming effects for concordant, neutral, and discordant triplets as compared to unrelated triplet s, with the concordant word triplets exhibiting the strongest of the priming effects A t the long SOA, healthy adults continued to perform similarly to that of the short SOA with the exception that discordant triplet primes no longer exhibited a priming effect. In contrast at the long SOA, cortical lesion patients continued to prime for all four of the task conditions but lost differences between the priming conditions. More relevant within the context of the current study are the findings from the two patient populations with subcortical involvement, which suggest a loss of enhancement due to aberrant basal ganglia function. Nonthalamic subcortical lesion patients lost all priming effects, consistent with
17 a disruption of the direct loop and subsequent loss of enhancement at longer SOAs. PD patients showed priming only for the concordant c ondition at the long SOA, which suggests an inability to enhance responding for all but the strongest stimuli in Additionally, Copland (200 implications in the late suppression part of the suppress enhance suppress cycle. A dominant subordinate paired priming task using two SOAs was given to PD patients, patients with nonthalamic subcortical lesions of the dominant hemisphere, patients with cortical lesions of the dominant hemis phere, and healthy older controls. At the shorter SOA, all patient groups primed to both subordinate and dominant meanings. At the long SOA, healthy controls and patients with cortical lesions primed only to the dominant meaning for words, a finding simi lar to that found in healthy college age subjects employing this paradigm across multiple SOAs (Simpson & Burgess, 1985). In contrast, both nonthalamic subcortical lesion and PD patients showed priming for both dominant and subordinate meanings at the lon g SOA, suggestive of a failure to suppress the subordinate meaning at the long SOA. More recently, Castner and colleagues (2007) evaluated noun and verb generation brain sti mulation (DBS) surgery as compared to healthy age matched controls. They used a paradigm based on Peran and colleagues (2003) to evaluate noun and verb associate generation for aurally presented nouns and verb targets. Subjects were instructed to generat e either a noun or a verb in response to either a noun or verb stimulus, and the experiment consisted of four types of production pairings: noun noun,
18 verb noun, noun verb, and verb verb. Most relevant to the current study were the findings for the noun n oun and verb verb generation conditions. Castner and colleagues (2007) found that during the DBS off stimulation condition, PD participants exhibited more errors in verb generation but not noun generation as compared to healthy control participants. In a ddition, when STN DBS stimulation was on, PD participants displayed more errors during both noun and verb same pairing conditions (i.e., noun noun, verb verb conditions). T hese findings would suggest a possible enhancement failure that occurs in PD that m ay be specific to verb generation (as displayed by verb verb generation deficits during the off stimulation condition) as well as a suppression failure amplified by STN DBS stimulation (as exhibited in noun and verb exemplar production during the STN DBS o n stimulation state). The study by Castner and colleagues (2007) may provide some insight regarding verbal fluency findings during STN DBS stimulation in PD. While some verbal fluency studies in PD have not found fluency impairments during STN DBS (e.g., Dujardin, 2001), many studies of phonemic and/or semantic fluency in PD with STN DBS have reported impairments on phonemic and/or semantic fluency during STN DBS (e.g., Saint Cyr, Trepanier, Kumar, Lozano, & Lange, 2000; Schroeder, Kuehler, Lange, Hasling er, Tronnier, Krause, Pfister, et al., 2003). Within the context of the findings reported by Castner and colleagues (2007), their findings would appear to suggest that a suppression failure may be induced during fluency tasks via stimulation of the STN.
19 CHAPTER 3 FUNCTIONAL NEUROIMAG ING OF LANGUAGE IN N EUROLOGICALLY HEALTH Y POPULATIONS An appreciation of the aforementioned behavioral research in populations with subcortical pathology is essential for evaluating the implications of the current paradigm. is also important to evaluate applicable neuroimaging findings in the domain of language. Specific neuroanatomical structures and networks, particularly in the left h emisphere, have been implicated in language processing and production. Neuroanatomical areas relevant to the current study include the left lateral frontal cortex (e.g., left inferior frontal gyrus) and left medial frontal cortices. For example, the left inferior frontal cortex, DLPFC (e.g., along the inferior frontal sulcus) and cingulate motor activity has been implicated in semantic tasks since the earliest neuroimaging studies in language (Peterson, Fox, Posner, Mintum, & Raichle, 1988). The role of the left inferior frontal cortex has been specifically implicated in the selection of semantic information among competing alternatives (e.g., Thompson 1997) and switching functions during verbal fluency tasks (e.g., H irschorn & Thompson Schill, 2006). Noun and verb retrieval functional neuroimaging results in neurologically healthy adults are particularly relevant for the current study. While there is limited neuroimaging research involving the study of verb retriev al, in one of the early neuroimaging studies of language, Warburton and colleagues (1996) used positron emission tomography (PET) to evaluate noun and verb retrieval in middle aged adults (mean age of 48 years). Activation was found in left DLPFC, SMA, an d anterior cingulate cortex for word retrieval in general, with more extensive activations patterns in these areas for verb
20 generation. Additionally, while subcortical findings were not discussed within the text of results tables suggest subcortical differences in noun and verb retrieval in middle aged adults and potential sensitivity of these paradigms for activating frontal striatal circuits in word generation. For example, verb generation as compared to rest app eared to show significant activation bilaterally in the caudate head and thalamus, in the absence of right inferior frontal activity. These findings may be suggestive of similarities in verb generation as compared to the frontal striatal implications prov ided from research in other types of word generation, e.g., category exemplar generation (Crosson et al., 2003). Relatedly, Perani and colleagues (1999) investigated verb and noun processing using a lexical decision task and PET. They found that the lan guage tasks evoked an extensive left hemisphere network of frontal and temporal cortex. Perani and colleagues (1999) also noted that while no region showed greater activation during decisions involving nouns as compared to verbs, the converse was not true ; several regions were more active during the verb condition as compared to the noun condition, including left inferior frontal, left middle frontal, and right lenticular regions. Shapiro and colleagues (2005) recently compared noun and verb grammatical o perations, reporting that verbs activated a left frontal cortical network whereas nouns activated a bilateral temporal network. Such results are supportive of recent transcranial magnetic stimulation (TMS) research demonstrating disruption of verb but not noun production when the left prefrontal cortex is suppressed using TMS (e.g., Cappelletti, Fregni, Shapiro, Pascual Leone, & Caramazza, 2008 ).
21 With regards to functional imaging results in aging, Persson, Sylvester, Nelson, Welsh, Jonides, and Reuter L orenz (2004) recently compared verb generation in young and old adults using an fMRI task involving noun stimuli with either few or many competing alternatives for an appropriate verb response. Although the two age groups performed similarly behaviorally in terms of verb generation accuracy and response latency, they showed differences in regional brain activation. For both age groups, high selection demands activated several cortical regions, including bilateral frontal, left anterior frontal, left infer ior temporal and dorsal anterior cingulate cortex. Frontal in the inferior part of prefrontal cortex, with the left frontal region extending dorsally into BA 9. Between gr oup comparisons revealed less activation for old as compared to young adults in the left inferior frontal gyrus, left inferior temporal gyrus, and anterior cingulate; however, greater activation was found in the right inferior frontal gyrus for old adults as compared to young adults. For left inferior frontal gyrus, there was a positive correlation between the fMRI signal and reaction time for young adults, with a negative pattern of correlation observed in old adults. In addition, during task contrasts, young adults showed stronger activity in dorsal frontal regions during the many competing alternatives task. Bilateral basal ganglia activation was found in old participants but not in young participants during contrasts between the many versus few compet ing alternatives conditions, which Persson and colleagues (2004) interpreted to be a compensatory response to the task. non dominant cortex and the basal ganglia as serving com pensatory functions is not an
22 unfamiliar theme in the cognitive neuroscience of aging. Cabeza (2002) posited a model of cognitive performance in aging termed HAROLD (hemispheric asymmetry reduction in older adults). The HAROLD model suggests that old adu lts show less lateralized prefrontal activity during cognitive tasks as compared to young adults. While Cabeza (2002) originally discussed HAROLD as potentially resulting individually from compensation or dedifferentiation or acting in conjunction with on e another, Cabeza and other cognitive neuroscientists studying aging differences in task performance more recently tend to interpret their findings within a compensatory framework (e.g., Cabeza, Daselaar, Dolcos, Prince, Budde, & Nyberg, 2004). However, s everal studies of language function by Crosson and colleagues have suggested that a compensatory role may be less clear of an interpretation for language studies of aging, such as object naming (Wierenga, Benjamin, Gopinath, Perlstein, Leonard, Rothi, et a l., 2008) and category member generation (Benjamin, McGregor, Chang, White, Rackelman, Sherod, et al., 2008; Cohen, Benjamin, McGregor, Chang, White, K. D., Rackelman, et al., 2009; Meinzer, Flaisch, Wilser, Eulitz, Rockstroh, Conway, et al., 2009). For e xample, d uring a picture naming study conducted by Wierenga and colleagues (2008), right inferior frontal, right precentral, and bilateral medial frontal activity was found in old as compared to young adults. When naming accuracy was correlated with the h demonstrated a positive correlation in the left and right inferior frontal gyrus (BA 45, 47) and a negative correlation in the right precentral gyrus and right superior fronta l gyrus. (BA 44, 45, 47) and a negative correlation between accuracy and HDR in the right
23 inferior cortex. These differences within older adults suggested that the ability to recruit the right hemisphere may not be universally compensatory for all older adults and may be region specific appeared to assist in naming performance in g older adults performing older adults while the recruitment of other right frontal regions outside of inferior frontal gyrus
24 CHAPTER 4 CLINICAL SUPPORT FOR DIS TIN CTIONS IN VERB NOUN PRODUCTION In addition to the results from behavioral paradigms in patient populations with frontal striatal involvement already mentioned in Chapter 2, several research groups have demonstrated the importance of prefrontal cortex involvement for ver b retrieval. Damasio and Tranel (1993) reported on the behavioral and neuroanatomical dissociation between noun and verb production in three clinical patients. Two of their patients performed similarly to controls on verb production but were impaired on noun production; these two patients had lesions in the left anterior and middle temporal lobe. In contrast, a patient with a lesion in the left premotor cortex showed the reverse pattern of behavior, demonstrating intact common and proper noun production and impaired verb retrieval. Similarly, Daniele, Giustolisi, Silveri, Colosimo, and Gainotti (1994) described two patients with left frontal lobe atrophy who were impaired on the naming and comprehension of verbs, whereas a third patient with left tempora l lobe atrophy was impaired on noun naming. Thompson (1998) described impaired verb generation performance during high selection demands for individuals with lesions including the posterior region of the l eft inferior frontal gyrus, as compared to persons with left frontal lesions not involving this area and persons with performance, the percentage of damage to BA 44 was a sign ificant predictor of selection related and task errors. regards to basal ganglia implications in verb retrieval. Cotelli and colleagues (2007) recently compared non demen ted PD patients with age matched controls on action and
25 object naming performance. PD subjects showed a deficit in both action and object naming as compared with controls. In addition, PD subjects but not controls performed significantly worse on action t han object naming in that study. Using an action fluency task, Piatt and colleagues (1999a) found that action fluency but not semantic or phonemic fluency discriminated demented PD patients from non demented PD patients and healthy controls and suggested that action fluency may be particularly sensitive to frontal striatal pathophysiology. Piatt and colleagues (1999b) in a study of healthy older adults found significant moderate correlations between action fluency and measures of executive function such a s the Wisconsin Card Sort Test (WCST) and Trails B. As mentioned previously, Castner and colleagues (2007) recently described behavioral evidence of subcortical involvement in noun and verb generation in PD. For patients with subthalamic (STN) deep brain stimulation (DBS), a higher number of errors during verb generation were made both on and off STN stimulation as compared to age matched controls. In addition, PD subjects displayed a higher number of errors for noun production as compared to controls wh ile STN stimulation was on.
26 CHAPTER 5 HYPOTHESES The aforementioned studies suggested that implementing verb and noun generation paradigms may be useful for studying aging and clinical populations of interest with known frontal striatal implicatio ns. The current study used functional magnetic resonance imaging (fMRI) and the noun noun and verb verb conditions for word generation as outlined by Castner and colleagues (2007) to investigate noun and verb generation and the involvement of cortical sub cortical pathways in aging. Specific study hypotheses were as follows: be implicated in aging, as measured using verb verb word generation. During their verb generation paradigm, Persson and colleagues (2004) found less activation for old as compared to young adults in the left inferior frontal gyrus, left inferior temporal gyrus, and anterior cingulate, while greater activation was found in the right inferior frontal gyr us for old adults as compared to young adults during their verb generation task. They interpreted the activation in right frontal cortex as serving a compensatory function for reduced activation in left cortical regions. Additionally, Castner and colleag ues (2007) found that during their STN DBS off stimulation condition, PD participants exhibited a deficit in verb generation but not noun generation, suggesting an enhancement failure in PD that may be specific to verb generation. Therefore, hypotheses fo r the current verb verb word generation were the following: Hypothesis 1a: Significantly less left cortical and subcortical activity was expected for the old as compared to the young subject group. Hypothesis 1b: Significantly more right cortical and sub cortical activity was expected for the old as compared to the young subject group.
27 will also be implicated in aging, as measured using noun noun word generation. Crosson and colleagues (2003) have demonstrated that during category generation for nouns, neuroanatomical areas consisting of the left pre supplementary motor area (preSMA), dorsal caudate, and ventral anterior (VA) nucleus of the thalamus are active. In addition, r ight basal ganglia activity in the caudate and putamen is also present during category generation, in the relative absence of right frontal activity. They suggested that the left preSMA caudate VA thalamus serves as a frontal striatal loop involved in wor d retrieval for language and that right basal ganglia activity may act to suppress right frontal activity that may interfere with language production. In the study conducted by Wierenga and colleagues (2008), right inferior frontal, right precentral, and bilateral medial frontal activity was found in old as compared to young adults during picture naming. Castner and colleagues (2007) found that during STN DBS stimulation, PD participants displayed more errors during both noun and verb same pairing conditi ons (i.e., noun noun, verb verb conditions). These findings would suggest a suppression failure amplified by STN DBS stimulation exhibited in noun and verb exemplar production during the STN DBS on stimulation state. Based on these findings, hypotheses f or noun noun word generation were the following: Hypothesis 2a: Significantly greater left cortical and subcortical activity was expected in the old, as compared to young, subject group. Hypothesis 2b: Significantly greater right cortical and subcortical activity was expected for the old, as compared to the young, subject group.
28 CHAPTER 6 METHODS Participants Fifteen young adults between 18 and 35 years of age (7 female) were recruited from the Gainesville community and 15 older healthy adults at least 70 years of age (8 female) were recruited from the Gainesville community and from a recruitment mailer Americans Independence Center. All of the older participants were Cauc asian, as were 80% of the younger participants. Education level was similar between the two groups, with both groups consisting on average of college educated subjects. Older and younger participants did not differ significantly according to level of edu cation, t (28) = .54, p = .60 (see Table 1) A ll participants spoke English as their first language and were right handed as determined by the Edinburgh Handedness Inventory (EHI; Oldfield, 1971). Medical e xclusionary criteria included history of dementia or Mild Cognitive Impairment, head trauma, neurological disorder (e.g., stroke), learning disability (e.g., dyslexia), psychiatric disorder (e.g., schizophrenia), drug or alcohol abuse, and chronic medical conditions and medical procedures likely to impai r cognition (e.g., renal or hepatic failure, coronary artery bypass grafting). Pharmacological exclusionary criteria included benzodiazapines, antiepileptic, antipsychotic, dopaminergic, and anticholinergic classes of medications, due to their potential e ffects on cognition. Given that the main outcome measures of the current proposal involve fMRI, individuals with pacemakers, metal implants, claustrophobia, or other conditions contraindicated for MRI were excluded from the current study. Young female pa rticipants were excluded if they were pregnant or trying to become pregnant. Three
29 additional young subjects were initially enrolled but excluded from the current analyses due to failing to complet e all testing or fMRI appointment tasks. Seven additional old subjects were initially enrolled but excluded from the study analyses. Two of the excluded old subjects completed testing but did not meet criteria for continuing to participate in the fMRI portion of the study due to testing performance that did not meet healthy control enrollment criteria as outlined below. An additional four old subjects were excluded d ue to subjective memory complaints or the disclosure of neurological problems following their testing session. Lastly, one old subject who partici pated in fMRI was excluded from the current analyses due to an incidental structural MRI finding suggestive of an asymptomatic cortical abnormality. Participants were instructed to abstain from caffeine for at least 2 hours prior to the fMRI scanning sess ion. Informed consent was obtained from participants according to guidelines established by the Health Science Center Institutional Review Board at the University of Fl orida. Participants were paid $25 for participation. Procedures Neuropsychological tes ting Prior enrollment in the fMRI study, participants were administered a standard health history questionnaire and completed a short neuropsychological assessment to screen for possible dementia and Mild Cognitive Impairment anmestic type (aMCI). Amne stic Mild Cognitive Impairment (aMCI) is a precursor to pathological aging and is a high (Peterson et al., 2001). Criteria for aMCI include self and/or corroborated memory comp laint, impaired memory function for age and education level, preserved general cognitive functioning, intact activities of daily living, and impairments not currently
30 meeting criteria for dementia. Notably, none of the participants included in the current study reported subjective language or memory complaints in daily life. Neuropsychological assessment included several tests to assess language, memory, and executive skills, in order to rule out dementia and aMCI and for determining potential group diffe rences in neuropsychological test performance that may either contribute to or be independent of the fMRI findings. The Mini Mental State Exam (MMSE; Folstein, Folstein, & McHugh, 1975) was administered, and all subjects participating in fMRI scored 27 or above on the MMSE, which is a standard research cutoff since MMSE scores below 24 are likely indicative of cognitive impairment. The California Verbal Learning Test, Second Edition (CVLT II; Delis, Kramer, & Kaplan, 2000) was used to assess verbal learni ng and memory. Since patients with mild dementias may perform normally on the MMSE but display memory difficulties, participants scoring greater than 1.5 SD below the mean for age appropriate norms on Long Delay Free Recall and Recognition Memory were ass umed to have a verbal memory deficit and thus were excluded from the fMRI portion of the study. The Boston Naming Test (BNT; Kaplan, Goodglass, & Weintraub, 2001), a 60 item confrontation ies to name objects. The Action Naming Test (ANT; Obler & Albert, unpublished), a 55 item confrontation Several brief word generation tasks were administered to asses s executive language and cognitive functions in the two participant cohorts. The Verbal Fluency subtest of the Delis Kaplan Executive Function System (D KEFS; Delis, Kaplan, & Kramer, 2001) was given in order to assess phonological fluency, semantic fluen cy, and semantic fluency
31 switching abilities. During the phonological fluency portion of Verbal Fluency, participants were given 3 letters (i.e., F, A, S) and asked to generate items beginning with each letter. The semantic fluency portion of Verbal Flue ncy required participants to providing items for 2 categories (e.g., fruits, furnit ure). The Action Fluency Test (Piatt, Fields, Paolo, et al. 1999) was administered to assess action fluency and generation abilities and consists of providing exemplars of actions that people perform. The Homophone Meaning Generation Test (HMGT; Warringt on, 2000) consisting of 8 homophones for which a number of alternate word definitions can be generated (e.g., Noun Verb Experimental FMRI Paradigm Stimuli for the fMRI s tudy were presented visually and projected onto a screen (IFIS) and E Prime version 1 software. The fMRI paradigm consisted of the noun noun and verb verb generation task s outlined by Peran and colleagues (2003) and Castner colleagues (2007). Three fMRI runs consisting of a block design were used. Each run consisted of 8 blocks of noun generation and 8 blocks of verb generation, for a total of 404 seconds per fMRI run. Each of the 16 blocks presented consisted of either 3 nouns or 3 verbs presented in randomized trial blocks, with 24 total noun and 24 total verb generations per run. Thus, a total of 72 nouns and 72 verbs were visually presented during the functional ima ging session. Resting blocks consisting of a visual fixation cross were presented at the beginning and end of each fMRI run and between each 3 item trial block. Resting blocks were pseudo randomized and contained 10, 12, or 14
32 seconds of rest in order to allow the hemodynamic response to return to baseline prior to the next stimuli block presentation. Rest blocks of randomized length were used to mitigate the effects of periodic or quasi periodic physiological noise, and a simple visual fixation rather t han a passive language based control task was used to maximize the possible amount of activity for the experimental condition (Newman, Twieg, & Carpenter, 2001). Each noun or verb within a block was presented for 2 seconds, and an additional 2 seconds wer e allowed for an associated within grammatical class response to be covertly generated by the participant. The timing for the fMRI noun noun and verb verb generation tasks was modified for fMRI presentation but consistent with that used in healthy older a dults during the previous behavioral paradigms (Peran, Rascol, Demonet, et al. 2003; Caster, Chenery, Silburn, et al. 2007). Noun and verb probe stimuli were derived from Lu and colleagues (2002) and from stimuli available through the International Pictur e Naming Project (IPNP, 2009) and equated between the generation conditions for word frequency in the English language (Francis and Kucera, 1982). One additional run of 8 noun and 8 verb blocks was administered outside of the scanner prior to the function al imaging session. The 4 noun verb runs were counterbalanced so that across participants in each age group, each of the 4 runs were used 25% of the time as a practice run and 75% of the time during the fMRI experiment. The additional run outside of the fMRI session allowed both for practice outside of the scanner prior to fMRI data collection and provided an estimate of subject performance on the covert generation task in the scanner. Image Acquisition Images were acquired on a Philip s 3 Tesla Achieva i nstrument with an 8 channel SENSE multiple arrayed head coil. Head motion was minimized using foam padding.
33 Functional images were obtained with a 1 short gradient echo planar functional imaging (EPI) scan (TR=2000ms; TE=30ms; FOV=240mm; matrix size=64 x 64; 3mm x 3mm in plane resolution, FA=80 o ). Thirty eight 3mm thick axial slices covering the whole brain were acquired in the axial plane. A high resolution T1 weighted anatomic scan (TR=8.1ms; TE=3.7ms; FOV=240mm; FA=8 o ; matrix size=240 x 240; 180 1.0m m slices ; sagittal orientation) was obtained prior to functional imaging to provide anatomic reference. A T2 weighted anatomic scan was collected in order to verify that subjects did not have incidental neuroanatomical findings that would preclude them fr om meeting criteria as a healthy control ( TR=2500ms; TE=250ms; FOV=175mm; FA=90 o ; matrix size=224x223; 64 4.5mm slices ; axial orientation ). Head motion was minimized using foam padding. Data analysis Behavioral data Performance accuracy on the noun verb task for the one run completed outside of the scanner was compared between the 15 old and 15 young participants using independent samples t tests. Performance on the memory and language measures was also compared between the two age groups using independe nt samples t tests. Neuroimaging data and FMRI analyses FMRI data were analyzed and overlaid onto structural images using the Analysis of Functional Neuroimaging (AFNI) software from the National Institutes of Health (Cox, 1996). The first 6 images of the 3 functional runs were discarded to ensure attainment of steady state. Time series images were spatially registered in three dimen sional reg program, in order to minimize the effects of head motion. Images were visually inspec ted for gross artifacts and quality control procedures were
34 applied to the data to detect residual motion or susceptibility artifact. For each of the subjects, the 3 imaging runs for the FMRI task were detrended of low frequency signal drifts (Birn, Saad, and Bandettini, 2001) and concatenated into a single time series. Prior to deconvolution and functional analyses, functional images were spatially smoothed with a Gaussian kernel of 4mm full width at half maximum (FWHM) to compensate for variability in s tructural and functional anatomy across participants. For each voxel, the observed fMRI intensity timeseries was modeled as the convolution of the experimental stimulus vector (comprised of 24 blocks of covert noun and 24 covert verb generation) and the e stimated best fit twelve lag impulse response function (IRF), allowing the hemodynamic response (HDR) to return to baseline. Signal timeseries were Z normalized within runs to mitigate signal intensity variance between runs and participants. The depend ent variable for analyses was the area under the curve (AUC) of the adding the deconvolved image intensity at each deconvolved time point of the impulse response. The T1 weig hted anatomic images and the deconvolved functional activation maps were warped to the co ordinates of the co planar stereotaxic atlas of Talairach and Tournoux (1988) as implemented by AFNI and resampled at a 1mm 3 resolution. Within group comparisons of word generation to a baseline of passive viewing was conducted using voxel wise t verb generation separately using the IRF AUC, and independent samples t tests were conducted to compare young vers us old adults on the two generation tasks. A c luster thresholding
35 implemented for whole brain analyses to determine which areas of activation on the t maps were significant at single vo xel p value levels of <.001 for independent measures t tests. Based on the Monte Carlo simulation, the independent samples t tests required contiguous voxel cluster sizes of 78 micro l iters or greater for cortical and subcortical regions of interest across the whole brain to maintain a cluster significance level of p<.05 and thus protect from Type I error. In addition to whole brain fMRI cluster analyses, bilateral subcortical and cortical areas involved in language generation were specifically evaluated a s regions of interest (ROIs) using anatomical masks available in AFNI. ROIs were selected on the basis of previous empirical work from our laboratory (Crosson, Benefield, Sadek, et al. 2003; Crosson, Sadek, Bobholz, et al. 1999) and upon conceptual consid eration of semantically based word generation (Crosson, Benjamin, & Levy, 2007). These anatomy based ROIs included the STN, caudate nucleus, globus pallidus, thalamus, lateral inferior frontal cortex (IFG) corresponding and BA 6), middle frontal gyrus (MFG; in order to include cortex around the inferior frontal sulcus), and medial frontal cortex consisting of pre SMA, SMA, and rostral cingulate zone (BA 6 and BA 32).
36 CHAPTER 7 RESULTS Neuropsychological testing Pe rformance data on the neuropsychological test battery is summarized in Table 8 1. Performances between young and old participants were similar on several measures, including the ANART predicted IQ score, CVLT II age corrected total for 5 learning trials a nd long delay free recall, BNT raw score, ANT raw score, DKEFS age corrected phonemic fluency, category fluency, and category switching, and the HMGT verbal concept generation. Performance differed between the two groups for the age and education correct ed BNT scores, t (28) = 4.16, p <.01, with young adults obtaining lower age and education corrected object naming scores than old adults despite similar BNT raw score performance. Performance differed between the two groups for Action Fluency, t (28) = 2. 80, p <.01, and the DKEFS category fluency raw score, t (28) = 3.71, p <.01, with old adults generating fewer responses than young adults on both tasks. Noun and verb generation outside of the fMRI scanner P erformance accuracy for both noun noun and v erb verb generation was lower for the old adults as compared to the young adults in the noun and verb generation run collected prior to fMRI scanning t Noun (28) = 3.09, p =.004 ; t Verb (28) = 5.04, p < .001, (see Table 8 1). FMRI results: Old versus young noun and verb generation comparisons Whole brain fMRI results comparing old to young for each generation task are summarized in Table 8 2 and displayed in Figure 8 2. During noun generation, the largest significant clusters of activity were found for old adults as compared to young adults in the right precentral gyrus and left angular gyrus. Repeated measures
37 ANOVAs performed on the average IRF for each of these two regions suggests significant Group x Time interactions for the IRF time course during noun generation, F Rprecentral (12, 348 ) = 5.51, p F Langular (12, 348 ) = 3.35, p = .012. Figure 2 indicates that for the right precentral gyrus cluster, old participants showed an increase in activity relative to baseline, whereas young participants showed a decrease in activity relative to the visual fixation baseline. In contrast, both age groups showed some increase in activity in the left angular gyrus, with greater activity for the old adults. Other small er clusters of significant activity for noun generation were found in the low part of the precentral gyrus, the supramarginal gyrus, the superior parietal lobe, and the middle temporal gyrus. During verb generation, the largest significant perisylvian clus ters of activity were found in the junction between right inferior frontal and middle frontal cortex and in left angular gyrus. Repeated measures ANOVA data for these two regions indicates differences in the IRF time course during verb generation for the left angular gyrus but not for the right inferior frontal/middle frontal gyral junction, although individual time Langular (12, 348 F R IFG/MFG (12, 348 ) = 1.73, p = .146. Figure 8 2 indicates that the differences between time points for the right inferior frontal gyrus reflect a main effect for age F RIFG/MFG (12, 28 ) = 4.30, p = .004, rather than an age X time inter action. For the left angular gyrus, old participants show ed strong positive activity relative to baseline, similar to their response for noun generation, but the young adults show ed little evidence of a response, in contrast to a
38 small response in the nou n condition. No areas were significantly greater for young participants as compared to old participants in either the noun or verb generation. Table 8 3 presents performance data for noun and verb generation outside of the scanner for each participant gr oup correlated with AUC data for each of the 4 clusters from Figure 8 2. Pearson product moment correlations were done for each each of the 4 significant clusters. None of these areas was significantly correlated with task performance for either condition for either group. Young versus baseline noun and verb generation comparisons Results for the noun and verb generation tasks as compared to the baseline rest condition for the young participants are outlined in Table 8 4. In sum, cortical activity was fairly similar between noun and verb generation for medial frontal cortex bilaterally (BA 6, 8, and 32) and in both left and right lateral frontal regions (inferior and mi ddle frontal gyri), except that activity does not extend into the middle frontal gyrus on the right. Both tasks elicit activity in the left superior and middle temporal gyri, but there is a small volume of activity in the left supramarginal gyrus for verb but not noun generation. Subcortical activity was also similar between noun and verb generation in the left caudate and right putamen. Old versus baseline noun and verb generation comparisons Results for the noun and verb generation tasks as compared to the baseline rest condition for the old participants are outlined in Table 8 5. In keeping with the comparison of old vs. young across task, these within task comparisons for old subjects show ed both a greater number of clusters and a greater volume of activity within many clusters than young subjects showed. In addition, cortical activity was fairly similar
39 between noun and verb generation and extensive for medial frontal cortex bilaterally (BA 6, 8, and 32) and in both left and right lateral frontal regions (primarily precentral, inferior, and middle frontal gyri). Subcortical activity was found in the left caudate and left pulvinar for both generation conditions and in the right putamen during noun generation. There was ample posterior perisylvian a ctivity (STG, angular gyrus, SMG) in both hemispheres for both tasks. A priori ROI AUC time course data Tables 8 6 and 8 7 and Figures 8 3 and 8 4 summarize significant area under the curve (AUC) repeated measures ANOVA data for young versus old time co urse comparisons in proposed frontal cortical and subcortical regions of interest (ROIs) using AFNI anatomical masks for IFG, MFG, medial frontal, STN, caudate, globus pallidus, and thalamus. For cortical ROIs, significant Group x Time interactions were fo und during noun generation between young and old at p<.05 for right inferior frontal gyrus, F RIFG (12, 348 ) = 2.59, p Lambda = .30, F RMFG (12, 348 ) = 3.29, p = .013, and right medial cortex, W = .32, F RMedial (12, 348 ) = 2.95, p = .02. A significant Group x Time interaction was found during verb generation between young and old at p<.05 for right inferior frontal gyrus, F RIFG (12, 348 ) = 2.51, p = .04. None of these areas significantly differed between the groups using a more conservative p<.0125 cut off to correct for multiple statistical comparisons. For subcortical ROIs, significant Group x Time interactions were found at p<.0125 for the left and right STN F LSTN (12, 348 ) = 3.78, p F RSTN (12, 348 ) = 4.56, p = .002, with old subjects showing greater positive IRFs than young subjects. A Group x Time
40 interaction was also found in the lef Lambda = .23, F LCaudate (12, 348 ) = 4.82, p = .002, with young subjects demonstrating greater positive IRFs than old subjects. While no other subcortical ROIs were significant at the more conservative p< .0125 statistical threshold, Group x Time Lambda = .36, F RCaudate (12, 348 ) = 2.54, p = .04, and in the right STN and right = .38, F RSTN (12, 348 ) = 2.34, p = .05; F LCaudate (1 2, 348 ) = 3.01, p = .02. Right STN verb generation IRFs were similar to those seen in the STN during noun generation, with greater positive IRFs in old versus young adults. The IRF s for the right caudate during noun generation is similar to that seen in the left caudate, with greater positive IRF in young versus old subjects. The IRFs in right thalamus during verb generation suggest differences in time courses for the 2 subject gro ups, with higher positive IRFs in old versus young subjects in the first half of the time course that returns to baseline by image 10, with the young subjects showing a positive biphasic IRF during the 13 image time course. Noun versus verb generation in young and old adults No significant differences were found between the noun and verb generation tasks when all subjects were combined together for statistical analysis, nor were there significant task differences when separate within group analyses were conducted for the old or young participant groups separately to evaluate differences in noun versus verb generation
41 CHAPTER 8 DISCUSSION R esults from the current study suggest aging differences for old as compared to young subjects with regards to noun and verb generation. First, for cortical regions and the fMRI between group comparisons, the current study found greater activation in old as compared to young subjects in the right precentral gyrus during noun generation, the right inferior frontal sulc us during verb generation, and the left angular gyrus during both generation tasks. Findings in the right precentral gyrus during noun generation and the right inferior frontal sulcus during verb generation are consistent with the proposed hypotheses that the old subjects would show greater activity in right frontal cortical regions for both noun and verb generation. However, while it had originally been proposed that less left frontal cortical activity would be present in old adults during verb generatio n and greater left frontal activation during noun generation in old adults based on other studies, there were no between group differences in left frontal cortical activation. While the left angular gyrus and other left posterior language cortices were no t a primary focus of the current study, the findings in the left angular gyrus are not surprising, given its role in semantic processing and imaging results demonstrating positive activation during semantic processing tasks in healthy adults (see Binder, D esai, Graves, and Conant, 2009 for a review). research that suggests greater right hemisphere activity (hemispheric asymmetry reductions) in older adults as compared to youn g for completing cognitive tasks (e.g., Cabeza, 2002). However, specific results from this study suggest that a compensatory interpretation may not be as parsimonious as proposed by some healthy aging
42 researchers. As mentioned in the results, the right p recentral gyrus findings during noun generation stemmed from the combination of greater positive activity in old adults and negative IRF in the young adults for that region. Negative IRF has been reported recently in young adults in the non dominant hemis phere for motor tasks (e.g., McGregor, et al., 2011), suggesting that at least for young adults, suppression of a dominant hemisphere occurs in order for optimal performance on motor tasks whereas the old adults may recruit the non dominant regi on to complete such tasks. This concept of non dominant hemisphere suppression has implications for language research in healthy aging, since while similar to the HAROLD aging model of reduced hemispheric asymmetry with aging, the study by McGregor and co lleagues suggests that that the asymmetry may not necessarily facilitate task performance in older adults but rather, in some tasks, the non suppression may in part interfere with task performance. Additionally, right hemisph ere activation during language studies may not be universally compensatory within older adults or even within frontal cortical regions as a whole. During the picture naming study conducted by Wierenga and colleagues (2008), right inferior frontal, right p recentral, and bilateral medial frontal activation was generally found in old as compared to young adults. However, old adults who performed better on the naming task specifically demonstrated a positive correlation between naming accuracy and the inferio r frontal gyrus bilaterally, while negative correlations were found for that subset of subjects between task accuracy and activation in the right precentral gyrus and right superior frontal gyrus. Furthermore, old adults with lower performance in the nami ng study showed a positive correlation between naming accuracy and left inferior frontal
43 cortex and a negative correlation between naming performance and right inferior frontal activation. These differences within older adults for naming suggests that the ability to recruit the right hemisphere may not be universally compensatory, since the ability to naming performance in old adults, whereas recruitment of other right fro ntal regions were negatively correlated with performance. With regards to the current study, the greater left (but not right) posterior perisylvian activity in old adults suggests that compensatory models such as HAROLD implicating greater use of the task non dominant hemisphere for older adults may apply only for more anterior cortical regions. Within the context of the current study, greater right hemisphere activity for old than young adults involved more anterior regions such as inferior frontal and p recentral cortex, whereas posteriorly greater activity for old than young adults was found primarily in the left hemisphere. The subcortical results for the noun and verb generation tasks are also more complex than as had originally been proposed. F MRI ROI analyses did not demonstrate significant between group differences in subcortical regions at the conservative thresholding criteria used in the current study. However, a priori anatomy based ROI analyses showed greater activity for old adults bilater ally in STN for noun generation, consistent with the idea that greater cortical and subcortical activity would be demonstrated bilaterally in old adults to complete the noun generation task, based neration. In addition, using a less stringent statistical threshold, the right STN activation in verb generation was greater for old subjects as compared to young subjects. Overall, the noun generation
44 STN findings are consistent with study hypotheses th at old adults may show greater activity bilaterally, suggesting greater bilateral activity may be necessary in older adults pathway. Moreover, the greater right subcortical activity in old adults during verb generation appears to be congruent with the premise that if there were less language dominant (left) hemisphere activity, this may require greater right hemisphere activity in order to compensate for an attenuated left h STN findings are in line with recent thought that the STN may participate in cognitive tasks (Crosson et al., 2007) similar to ho w it participates in motor tasks through the Differences in the caudate nucleus bilaterally during noun generation f or the current study sug gest that there is great er caudate ac tivity in young than old adults It may be that the g reater left caudate activity in young adults within the context of noun generation could be interpreted as serving the function to provide adequate s pecific to the left hemisphere during noun generation for that age group. This line of interpretation appears to be support ed by the greater right pre frontal activation found during noun generation for old adults, with the findings in old adults suggestiv e of non dominant hemispheri c recruitment in order to complete the noun generation task. As for the greater right caudate activity in young adults this finding within the non dominant hemisphere for young adults with such an interpretation supported by the
45 negative IRF findings in the right precentral gyrus for young subjects during noun generation, in contrast with the positive IRF data for old subjects in that area. Overall, as the individual gro up comparisons to baseline for both noun and verb generation suggests, both generation tasks evoked increased activity in left lateral frontal regions such as inferior frontal gyrus, medial frontal, and left subcortical regions that have been found to be a ctive during category member (noun) generation studies in healthy young subjects (e.g., Crosson, et al., 2003, Crosson, et al., 2001) as well as healthy aging studies evaluating noun word generation (e.g., Cohen, et al., unpublished) and object picture nam ing (Wierenga, et al., 2008) in healthy old adults. Additionally, for both age groups in the current study, right lateral frontal activity was present in both types of word generation, suggesting that the young subjects as well as the old were using bilat eral lateral frontal networks in order to complete the tasks in the current study. As such, these findings suggest that the current covert word generation paradigms activate a similar network for word generation, regardless of the type of generation, but that the activity may be bilateral for both conditions and both age groups in order to complete this type of word generation task. The exception to this rule of thumb is in those right frontal clusters (precentral/postcentral gyri for noun, inferior front al gyrus for verb) that showed greater activity for old than young adults. In these areas, young adults tend to show decreased activity, while old adults show increased activity. With regards to the left hemisphere activity, the limited spatial segregati on found in the current study for the two types of word production is in line with a recent review of neuroimaging tasks to date evaluating nouns and verbs across various types of language paradigms (see Crepaldi, Berlingeri, Paulesu, & Luzzatti, 2011). O ne
46 previous study of noun and verb semantic fluency in healthy adults (Warburton, et al., 1996) found that verb generation produced greater activation than noun generation in the left inferior frontal gyrus, left SMA, left temporal sulcus, left temporo par ietal junction and left inferior parietal lobe whereas noun generation compared to verbs activated right superior frontal sulcus and left anterior cingulate gyrus However in the current study, no cortical or subcortical area in either hemisphere was mor e active during either generation frontal activation may be bilateral for both conditions and both age groups with the sole exception for each task mentioned above, in order to complete word generation using the current paradigms. The current study sought to establish whether a behavioral noun and verb generation task used in patient populations with frontal striatal dysfunction could be implemented with adequate sensitivit y for evaluating age related differences in noun and verb generation using fMRI. While the current study appeared to be adequate in terms of the subject sample and general experimental design to detect some differences in old versus young adults there we re several limitations o f the current study. First, there are methodological aspects to fMRI in general that pose limitations to studying language paradigms such as the current one within the context of frontal striatal circuitry. The temporal resolution of fMRI does not lend itself well for delineating which pathways may be primarily responsible for the results found during noun and verb generation or how such pathways may work in conjunction with one another during such a paradigm. Given that fMRI task s and the related IRF operate on a scale of seconds whereas the timing as proposed by motor studies of the STN indicates that the cycle of
47 the hyperdirect direct indirect cortical subcortical circuitry operate on a scale of milliseconds, studies such as th e current one have to draw conclusions for results without clear temporal data to delineate the unique contributions of each pathway Secondly, the fMRI experimental paradigm used in the current study was a covert word generation task. Covert word genera tion was chosen so that a block fMRI experimental paradigm presentation could be used, in order to increase statistical power and provide greater sensitivity for detecting group differences in subcortical structures (Crosson, et al., 2003) However, cover t word generation paradigms are limited in that they do not provide direct information on performance accuracy, since the subject is not speaking aloud during the scanning session. Although one experimental run of the generation task was done outside of t he scanner prior to the fMRI session, in order to ensure that participants understood the task and could adequately perform the task, accuracy for the experimental run outside of the scanner can only be interpreted an estimate of fMRI task performance. Th ird while the time allotted for covert generation was similar to that the word generation run outside of the scanner differed from that of the young subjects suggests th at slightly more time may have been needed for similar performance accuracy in generation to occur. The current study suggests that while typical word generation models often use noun generation (e.g., category generation, phonemic fluency), verb generati on is a viable alternative for neuroimaging studies of word generation. While the current study was of a preliminary nature and used whole brain scanning, the between group differences in STN and caudate activity relative to baseline findings for both sub ject
48 groups suggests that subcortical structures are active in this type of word generation. specific functional neuroimaging that encompasses pr imarily subcortical and fro ntal cortices may be viable in future endeavors contrasting noun and verb generation, in order to achieve better spatial delineation of frontal subcortical activation in these tasks. Secondly, the current study did not evaluate differences within grammati cal class for the types of nouns or verbs generated (e.g., living versus nonliving nouns; human versus nonhuman verbs). While the current study did not propose to examine whether there are substrate variations for different types of nouns and verbs genera ted, future directions for these types of paradigms may include evaluating whether the grammatical class of items (i.e., nouns versus verbs), the semantic attributes of classes of items across grammatical classes (i.e., physical attributes versus items sug gesting manipulability such as tools/human actions), or level of selection constraint (i.e., high selection constraint/few alternatives verus low constraint/many alternatives) may produce different patterns of activation (Crosson, et al., 2003) Lastly, t he current study did not further parse subjects within groups by level of performance due to the number of subjects and the fact that accuracy data was not collected during the 3 fMRI runs. However, results from at least one noun picture naming study have suggested that there can be differences within older adults with regards to correlations between behavioral performance and fMRI activation pattern s (e.g., Wierenga et al, 2008), suggesting that this can be an important consideration, should this type of paradigm be done as an overt word generation paradigm in future studies.
49 Table 8 1. Demographic and neuropsychological testing data for young and old subjects. Variable Young Mean (SD) Old Mean (SD) t test or 2 p value Demographics Age (year s) 23.20 (3.76) 76.67 (5.07) t (1,28) = 32.81 p < 0.001*** Education (years) 16.00 (2.07) 15.50 (2.97) t (1,28) = 0.54 p = 0.60 Women/men 7/8 8/7 2 = 0.13 p = 0.72 Global Cognition MMSE 29.48 (1.25) 28.80 (0.86) t (1,28) = 1.74 p = 0.10 ANART 117.19 (4.19) 117.74 (5.53) t (1,28) = 0.30 p = 0.77 Learning and Memory CVLT List 1 5 Total Recall raw score 58.67 (10.73) 45.87 (8.39) t (1,28) = 3.09 p = 0.005** CVLT List 1 5 Total Recall T score 57.60 (12.67 ) 58.13 (9.58) t (1,28) = 0.09 p = 0.93 CVLT Long Delay Free Recall raw score 13.27 (3.20) 9.93 (3.28) t (1,28) = 2.82 p = 0.009** CVLT Long Delay Free Recall z score 0.47 (1.27) 0.50 (0.89) t (1,28) = 0.08 p = 0.93 Language BNT raw score 55.67 (2.09) 56.67 (2.41) t (1,28) = 1.21 p = 0.24 BNT T score 46.20 (6.95) 59.07 (9.77) t (1,28) = 4.16 p < 0.01** ANT raw score 53.87 (2.00) 54.47 (1.92) t (1,28) = 0.84 p = 0.41 DKEFS Letter Fluency raw score 47.07 (11.90) 41.53 (13.84) t (1,28) = 1.18 p = 0.25 DKEFS Letter Fluency scaled score 13.00 (3.32) 12.07 (3.63) t (1,28) = 0.74 p = 0.47 DKEFS Category Fluency raw score 46.67 (7.90) 35.07 (9.06) t (1,28) = 3.74 p = 0.001*** DKEFS Category Fluency scaled score 13.47 (3.20) 11.47 (2.80) t (1,28) = 1.82 p = 0.08 DKEFS Category Switch raw score 15.07 (2.76) 13.13 (2.83) t (1,28) = 1.90 p = 0.07 DKEFS Category Switch scaled score 11.93 (3.20) 11.73 (2.99) t (1,28) = 0.18 p = 0.86 DKEFS Switc h Accuracy raw score 13.60 (3.09) 11.93 (3.34) t (1,28) = 1.42 p = 0.17 DKEFS Switch Accuracy scaled score 11.67 (3.11) 11.33 (3.02) t (1,28) = 0.30 p = 0.77 Action Fluency raw score 22.67 (5.86) 17.40 (4.32) t (1,28) =2.80 p = 0.009** HMGT raw score 27.47 (4.42) 27.60 (5.64) t (1,28) = 0.07 p = 0.94 FMRI task: noun generation raw score 21.93 (1.83) 18.86 (3.28) t (1,28) = 3.09 p = 0.004** FMRI task: verb generation raw score 20.73 (2.22) 15.93 (2.89) t (1,28) = 5.04 p <.001*** BNT = B oston Naming Test; ANT = Action Naming Test; DKEFS = Delis Kaplan Executive Function Systems; HMGT = Homophone Meaning Generation Test. = p<.05. ** = p<. 01. *** = p<.001.
50 Table 8 2. FMRI region of interest (ROI) area under the curve time course data. Young versus old t test comparisons for noun and verb generation. Noun ROIs Old > Young Volume (mm 3 ) BA/ subregion Mean Intensity (x,y,z) Peak Intensity (t value) Verb ROIs Old > Young Volume (mm 3 ) BA/ subregion Mean Intensity (x,y,z) Peak Intensity (t va lue) Frontal Frontal Pre / Postcentral gyrus 759 R BA 4, 6, 43, 3/1/2 53, 11, 31 5.05 IFG, MFG SFG, MFG 154 84 R BA 44, 9 R BA 8 28, 12, 28 19, 18, 40 4.93 5.00 Precentral gyrus 118 R BA 6 62, 4, 12 5.31 Poste rior perisylvian Posterior perisylvian Angular gyrus 656 L BA 39 24, 59, 32 5.94 Angular gyrus 781 L BA 39 24, 58, 33 5.60 MTG 83 L BA 37, 21 55, 51, 4 4.42 SMG 92 R BA 40 57, 30, 47 4.78 85 R BA 40 63, 33, 41 4.38 Other regions Other regions Superior parietal 112 L BA 7 24, 68, 43 4.44 Superior parietal 170 L BA 7 25, 68, 43 4.48 Postcentral gyrus 94 L BA 40 62, 21, 24 4.50 Abbreviations: ROI = Region of interest. SD = Standard deviation L = Left. R = Righ t. IFG = Inferior frontal gyrus MFG = Middle frontal gyrus. Medial = Medial frontal cortex. STG = Superior temporal gyrus. MTG = Middle temporal gyrus. SMG = Supramarginal gyrus. Cluster size > 78 microl iters, p <.001.
51 Table 8 3. Correlations between fMRI word generation active ROIs and word generation performance. Accuracy Noun Generation Verb Generation L angular gyrus R precentral/ postcentral gyrus L angular gyrus R inferior/middle frontal gyrus Young .15 p = .59 .22 p = .42 .34 p = .22 .08 p = .79 Old .41 p = .14 .29 p = .32 .26 p = .37 .06 p = .85 L = Left. R = Right. = p<.05.
52 Table 8 4. FMRI region of interest (ROI) area under the curve time course data. Young t test comparisons for noun and verb generation versus baseline. Noun ROIs Volume (mm 3 ) BA/ subregion Mean Intensity (x,y,z) Peak Intensity (t value) Verb ROIs Volume (mm 3 ) BA/ subregion Mean Intensity (x,y,z) Peak Intensity (t value) Frontal Frontal IFG, insula 2215 L BA 45, 47, 44 35, 18, 2 8.23 IFG, insula 509 370 L BA 45 L BA 47, 45 29, 31, 7 35, 19, 1 6.29 5.82 742 R BA 45, 47 45, 19, 1 6.82 570 R BA 47, 45 44, 20, 1 6.72 IFG, MFG 257 R BA 45, 46 49, 28, 17 6.12 IFG, MFG 3287 L BA 9, 46, 45, 44 35, 25, 24 8.51 MFG 1727 L BA 9, 46 41, 11, 28 6.48 Medial 3802 B BA 6, 8, 32 5, 15, 37 9.45 Medial 3716 B BA 6, 8, 32 5, 16, 37 8.54 Subcortical Subcortical Cau date, striatal bridge s 229 L 15, 1, 15 5.27 Caudate, striatal bridges 446 L 14, 0, 17 7.37 Putamen 168 R 26, 11, 7 6.72 Putamen, s triatal bridges, insula 526 R 22, 10, 16 6.39 Posterior perisylvian Posterior perisylvian STG, MTG 354 L BA 22, 21 57, 43, 5 5.76 STG, MTG 799 L BA 22, 21 57, 45, 6 7.43 Inferior parietal, SMG 90 L BA 40 31, 44, 40 5.02 Abbreviations: ROI = Region of interest. SD = Standard deviation. L = Left. R = Right. B = Bilateral. IFG = Inferior frontal gyrus. MFG = Middle frontal gyrus. Medial = Medial frontal cortex. STG = Superior temporal gyrus. MTG = Middle temporal gyrus. SMG = Suprama rginal gyrus. Cluster size > 78 microliters, p <.001.
53 Table 8 5. FMRI region of interest (ROI) area under the curve time course data. Old t test comparisons for noun and verb generation versus baseline. Noun ROIs Volume (mm 3 ) BA/ subregion Mean Intensity (x,y,z) Peak Intensity (t value) Verb ROIs Volume (mm 3 ) BA/ subregion Mean Intensity (x,y ,z) Peak Intensity (t value) Frontal Frontal IFG, MFG, SFG, precentral 6603 L BA 44, 45, 6, 9 46, 3, 28 7.75 IFG, MFG, precentral 5927 L BA 44, 45, 6, 9 42, 2, 19 8.22 IFG, insula 3264 L BA 47, 45, 44 40, 19, 2 7.31 IFG, insula 1553 105 L BA 47, 45 R BA 45 40, 21, 5 51, 30, 14 7.28 5.38 120 R BA 45 52, 29, 13 5.32 96 R BA 45 29, 9, 25 5.85 IFG, precentral 608 R BA 44, 6 59, 12, 3 6.26 IFG, precentral 467 R BA 44, 6 60, 12, 3 6.02 M FG, precentral 417 R BA 6 49, 3, 43 6.53 MFG, precentral 112 L BA 6 23, 8, 49 5.45 262 R BA 6 49, 2, 44 6.45 171 R BA 6 28, 6, 50 6.25 MFG 706 L BA 9 40, 24, 32 6.99 IFG, MFG 1820 L BA 45, 9, 46 42, 24, 30 6.95 661 L B A 10 40, 43, 4 7.31 MFG 287 L BA 46, 10 41, 44, 5 5.95 337 L BA 10 30, 56, 22 6.57 468 L BA 10 30, 56, 21 7.14 247 R BA 6 27, 9, 45 6.58 118 R BA 9 47, 13, 30 5.21 85 R BA 6 39, 0, 55 6.12 88 R BA 9 45, 37, 30 6.10 MFG, SFG, Precentral 1437 L BA 6, 4 32, 3, 56 6.67 Medial 6466 B BA 6, 8, 32 14, 18, 40 9.42 Medial 8056 B BA 6, 8, 32 0, 23, 38 8.28 117 L BA 6 1, 25, 70 5.23 139 L BA 6 1, 25, 70 5.98 Subcortical Subcortical Caudate, striatal b ridges 269 168 L L 18, 6, 5 13, 7, 17 6.00 5.78 Caudate 269 L 13, 6, 17 6.52 Pulvinar 177 L 23, 29, 10 12.13 Pulvinar 113 L 23, 28, 10 10.16 Putamen, striatal b ridges 88 R 23, 5, 10 5.82 Posterior perisylvian Posterior perisylvian SMG, Angular, SPL, IPL, Postcentral 10247 L BA 40, 39, 7, 2 25, 55, 33 9.06 SMG, Angular, SPL, IPL 7787 L BA 40, 39, 7 21, 51, 30 9.49 STG, MTG, ITG STG 1294 184 L BA 22, 21, 20 L BA 22 51, 46, 9 56, 34, 11 6. 25 5.91 STG, MTG, fusiform, p arahippocampal 7976 L BA 22, 21, 20, 37, 19 13, 44, 17 9.29
54 Table 8 5 Continued Noun ROIs Volume (mm 3 ) BA/ subregion Mean Intensity (x,y,z) Peak Intensity (t value) Verb ROIs Volume (mm 3 ) BA/ subregion Mean I ntensity (x,y,z) Peak Intensity (t value) STG, MTG 798 R BA 22, 21 48, 41, 2 6.60 STG, MTG 400 R BA 22, 21 50, 39, 2 6.89 Angular gyrus, SPL 1705 R BA 39, 7 33, 62, 37 8.97 Angular gyrus, IPL, SPL 1130 R BA 39, 40, 7 36, 64, 41 8.39 SMG, p ostcentral 881 R BA 40 49, 40, 48 8.39 SMG, IPL, postcentral 493 R BA 40 49, 40, 47 6.23 Other areas Other areas Postcentral 467 L BA 40 63, 17, 16 7.93 Postcentral 303 L BA 43, 2 57, 19, 18 6.93 164 L BA 3/1/ 2 50, 20, 34 5.96 191 L BA 43 54, 21, 36 6.30 137 R BA 3/1/2 40, 22, 50 6.64 82 L BA 43 54, 8, 17 4.89 145 R BA 3/1/2 40, 22, 50 6.42 Abbreviations: ROI = Region of interest. SD = Standard deviation. R = R ight. B = Bilateral. IFG = Inferior frontal gyrus. MFG = Middle frontal gyrus. SFG = Superior frontal gyrus. Medial = Medial frontal cortex. SFG = Superior frontal gyrus. STG = Superior temporal gyrus. MTG = Middle temporal gyrus ITG =Inferior te mporal gyrus. SMG = Supramarginal gyrus. SPL = Superior parietal lobe. IPL = Inferior parietal lobe. Cl uster size > 78 microliters, p <.001.
55 Table 8 6. Anatomy based region of interest (ROI) area under the curve time course data. Young versus old r epeated measures analysis of variance (ANOVA) data for lateral and medial frontal ROIs during noun and verb generation. ROI/Condition Group Mean (SD) t test p value ROI/Condition Group Mean (SD) t test p value L IFG Noun Young 3.41 (2.48) t (1,28) = 0.2 5 p = 0.80 R IFG Noun Young 0.99 (2.17) t (1,28) = 0.28 p = 0.78 Old 3.69 (3.36) Old 0.78 (1.88) 0.45 F (1,12) = 1.70 p = 0.15 0.35 F (1,12) = 2.59 p = 0.036* L IFG Verb Young 3.75 (2.85) t (1,28) = 0.18 p = 0.86 R IFG Verb Young 0.88 (2.12) t (1,28) = 0.05 p = 0.96 Old 3.95 (3.23) Old 0.84 (2.02) 0.48 F (1,12) = 1.52 p = 0.21 0.36 F (1,12) = 2.51 p = 0.04* L MFG Noun Young 2.22 (4.32) t (1,28 ) = 0.79 p = 0.44 R MFG Noun Young 0.58 (1.87) t (1,28) = 1.16 p = 0.26 Old 3.19 (1.99) Old 1.34 (1.74) 0.47 F (1,12) = 1.62 p = 0.18 0.30 F (1,12) = 3.29 p = 0.013* L MFG Verb Young 2.28 (4.25) t (1,28 ) = 0.72 p = 0.48 R MFG Verb Young 0.36 (2.0) t (1,28) = 1.31 p = 0.20 Old 3.14 (1.84) Old 1.25 (1.75) 0.56 F (1,12) = 1.11 p = 0.41 0.42 F (1,12) = 1.94 p = 0.10 L Medial Noun Young 0.31 (1. 29) t (1,28) = 0.39 p = 0.70 R Medial Noun Young 0.25 (2.19) t (1,28) = 0.17 p = 0.87 Old 0.08 (2.02) Old 0.13 (1.59) 0.64 F (1,12) = 0.80 p = 0.65 0.32 F (1,12) = 2.95 p = 0.02* L Medial Verb Young 0. 28 (1.50) t (1,28) = 0.50 p = 0.62 R Medial Verb Young 0.31 (2.79) t (1,28) = 0.07 p = 0.94 Old 0.03 (1.78) Old 0.26 (1.54) 0.53 F (1,12) = 1.24 p = 0.33 0.60 F (1,12) = 0.96 p = 0.52 Abbreviations: ROI = Region of interest. SD = Standard deviation. L = Left. R = Right. IFG = Inferior frontal gyrus. MFG = Middle frontal gyrus. Medial = Medial frontal region. = p<.05. ** = p < .0125.
56 Table 8 7. Anatomy based region of interest (ROI) area under the curve time course data. Young versus old repeated measures analysis of variance (ANOVA) data for subcortical ROIs during noun and verb generation. ROI/Condition Group Mean (SD) t test p value ROI/Condition Group Mean (SD) t test p value L STN Noun Young 0.87 (2.42) t (1,28) = 1.36 p = 0.19 R STN Noun Young 0.75 (1.74) t (1,28) = 1.08 p = 0.29 Old 2.06 (2.37) Old 1.59 (2.47) 0.27 F (1,12) = 3.78 p = 0.006** 0.24 F (1,12) = 4.56 p = 0.002** L STN Verb Young 0 .54 (2.29) t (1,28) = 2.02 p = 0.05* R STN Verb Young 0.63 (1.58) t (1,28) = 1.02 p = 0.32 Old 2.42 (2.78) Old 1.46 (2.71) 0.42 F (1,12) = 1.99 p = 0.10 0.38 F (1,12) = 2.34 p = 0.05* L Caudat e Noun Young 1.33 (1.50) t (1,28) = 0.47 p = 0.64 R Caudate Noun Young 0.44 (1.16) t (1,28) = 0.66 p = 0.52 Old 1.08 (1.41) Old 0.17 (1.08) 0.23 F (1,12) = 4.82 p = 0.002** 0.36 F (1,12) = 2.54 p = 0.04* L Caudate Verb Young 1.30 (1.85) t (1,28) = 0.42 p = 0.68 R Caudate Verb Young 0.42 (1.31) t (1,28) = 0.31 p = 0.76 Old 1.05 (1.38) Old 0.29 (0.97) 0.58 F (1,12) = 1.03 p = 0.46 0.43 F (1,12) = 1.87 p = 0.12 L GP Noun Young 0.86 (1.56) t (1,28) = 0.63 p = 0.54 R GP Noun Young 0.80 (1.26) t (1,28) = 0.66 p = 0.51 Old 1.26 (1.89) Old 1.16 (1.71) 0.48 F (1,12) = 1.57 p = 0.19 0.41 F (1,12) = 2.05 p = 0.09 L GP Verb Young 1.07 (1.72) t (1,28) = 0.32 p = 0.75 R GP Verb Young 0.51 (1.42) t (1,28) = 1.14 p = 0.26 Old 1.31 (2.48) Old 1.16 (1.69) 0.42 F (1,12) = 1.96 p = 0.10 0.54 F (1,12) = 1.22 p = 0.34 L Thalamus Noun Young 0.90 (1.38) t (1,28) = 0.78 p = 0.44 R Thalamus Noun Young 0.83 (1.45) t (1,28) = 0.11 p = 0.91 Old 1.26 (1.17) Old 0.92 (2.61) 0.47 F (1,12) = 1.58 p = 0.19 0.49 F (1,12) = 1.48 p = 0.23 L Thalamus Verb Young 0.89 (1.32) t (1,28) = 1.02 p = 0.32 R Thalamus Verb Young 0.70 (1.52) t (1,28) = 0.11 p = 0.92 Old 1.41 (1.48) Old 0.78 (2.69) 0.51 F (1,12) = 1.37 p = 0.27 0.32 F ( 1,12) = 3.01 p = 0.02* Abbreviations: ROI = Region of interest. SD = Standard deviation. L = Left. R = Right. STN = Subthalamic nucleus. GP=Globus pallidus. = p<.05. ** = p < .0125.
57 Figure 8 1. Schematic diagram of pre SMA basal ganglia loop (Crosson, Benjamin, & Levy, 2007). Direction of pathways between structures is indicated by the arrows. Reciprocated indicate excitatory effects of the pathway and negative signs ( ) indicate inhibitory effects of the pathway. The neostriatal component of the loop probably consists of striatal gray bridges, but also may include the caudate nucleus or putamen on either side of the capsule (Inase et al., 1999). Note that dopaminergic projections from the substantia nigra pars compacta to the neostriatum have both excitatory and inhibitor y effects, depending upon the target neurons (Gerfen, 1992).
58 Figure 8 2. Functional activity during noun and verb generation for young versus old adults. Top left: right precentral/postcentral gyrus during noun condition. Top r ight: right inferior frontal gyrus during verb condition. Bottom left: left angular gyrus during noun condition. Bottom right: left angular gyrus during verb condition. Sagittal, axial, and coronal views of the regions of interest (ROIs) with respective regions of interest time course data for fRMI area under the curve. Cluster size > 100 microliters; p<.001. = significant between group time course difference at p < .01.
59 Right Inferior Frontal: Noun Right Inferior Frontal: Verb Right Middle Fron tal: Noun Right Medial Frontal: Noun Figure 8 3. Impulse response function time course data for lateral and medial frontal ROIs. Repeated measures ANOVA (group x 13 timepoints). p<.05.
60 Left STN: Noun Right STN: Noun Right STN: Ver b Right Thalamus: Verb Left Caudate: Noun Right Caudate: Noun Figure 8 4. Impulse response function time course data for subcortical ROIs. Repeated measures ANOVA (group x 13 timepoints). Left caudate, left STN, and right STN noun conditio n significant at p <.0125; other ROIs significant at p<.05
61 APPENDIX NOUN AND VERB GENERATION STIMULI Table A 1. Noun and verb generation stimuli Noun Living Frequency F & K Noun Nonliving Frequency F & K Verb Human Frequency F & K Verb Nonhuman Frequency F & K monkey H 275 gun H 142 reading H 274 rolling H 88 snake H 70 knife H 86 painting H 95 flowing H 40 cat H 42 cup H 58 laughing H 89 sinking H 40 corn H 38 pencil H 38 cooking H 50 melting H 32 fish H 33 fan H 34 whispering H 31 digging H 32 potato M 30 pitcher M 29 waving M 30 boiling M 27 sheep M 24 lock M 28 arresting M 27 rotating M 21 lemon M 16 needle M 21 smoking M 26 nesting M 20 eagle M 12 fork M 20 celebrating M 25 popping M 17 bee M 27 thread M 20 carving M 23 blooming M 17 lion M 26 arrow M 20 wrapping M 23 flooding M 15 peas M 24 ladder M 19 shaving M 23 raining M 14 mouse M 20 basket M 19 whistling M 21 thundering M 14 onion M 19 axe M 19 winking M 18 dripping M 14 elephant M 18 pen M 18 sewing M 18 roaring M 13 rabb it M 16 stove M 17 knitting M 18 buzzing M 13 apple M 15 razor M 15 skipping M 17 snowing M 12 orange M 15 ruler M 13 peeling M 14 erupting M 11 deer M 13 plow M 12 typing M 12 grazing M 9 pineapple M 9 compass M 12 mopping M 9 soaring M 9 turtle M 9 umbrella M 11 clapping M 9 sprouting L 8 pear L 8 microphone L 8 ironing L 8 hatching L 7 tomato L 7 rake L 8 coughing L 8 thawing L 6 duck L 6 saw L 8 vacuuming L 6 freezing L 6 carrot L 5 shovel L 8 twirling L 6 howling L 5 banana L 5 spoon L 6 b uckling L 6 swaying L 5 celery L 4 hammer L 6 cheering L 6 ticking L 5 turkey L 4 hook L 5 raking L 6 overflowing L 4 mushroom L 4 easel L 5 erasing L 5 burrowing L 4 alligator L 4 leash L 4 combing L 5 wagging L 4 butterfly L 3 cannon L 4 scooping L 5 wilting L 3 spinach L 2 mop L 3 hammering L 4 pecking L 3
62 Table A 1 Continued Noun Living Frequency F & K Noun Nonliving Frequency F & K Verb Human Frequency F & K Verb Nonhuman Frequency F & K pumpkin L 2 corkscrew L 3 boxing L 4 clucking L 3 fro g L 2 kettle L 3 saluting L 4 galloping L 3 spider L 2 pitchfork L 2 wrestling L 3 hibernating L 2 watermelon L 1 wrench L 1 wringing L 3 vibrating L 2 broccoli L 1 ladle L 1 skiing L 3 gusting L 2 anteater L 1 scissors L 1 dabbing L 3 migrating L 1 lobster L 1 hinge L 1 skating L 3 deflating L 1 porcupine L 1 blender L 1 juggling L 2 evaporating L 1 squirrel L 1 hoe L 1 zipping L 2 sleeting L 1 zebra L 1 syringe L 1 golfing L 1 croaking L 1 whale L 0 pliers L 1 kneading L 1 barking L 1 giraffe L 0 dishwasher L 1 dribbling L 1 pollinating L 0 lettuce L 0 screwdriver L 0 curtseying L 1 purring L 0 peanut L 0 toaster L 0 surfing L 0 hissing L 0 penguin L 0 stapler L 0 parachuting L 0 snarling L 0 scorpion L 0 spatula L 0 sunbathing L 0 waddlin g L 0
63 LIST OF REFERENCES Alexander, G. E., Delong, M. R., & Strick, P. L. (1986). Parallel organization of functionally segregated circuits linking basal ganglia and cortex. Annual Reviews, 9, 357 381. Benjamin, M. L., McGregor, K. M., Chang, Y., White, K. D., Rackelman, C., Sherod, M., Levy, I., & Crosson, B. A. (2008, February). Hemispheric Asymmetry Reductions in Older Adults during Category Exemplar Generation. Poster presentation for the annual meeting of the International Neuropsychologica l Society, Waikoloa, HI. Binder, J. R., Desai, R. H., Graves, W. W., & Conant, L. L. (2009). Where is the semantic system? A critical review and meta analysis of 120 functional neuroimaging studies. Cerebral Cortex, 19, 2767 2796. Birn, R. M., Saa d, Z. S., & Bandettini, P. A. (2001). Spatial heterogeneity of the nonlinear dynamics in the fMRI BOLD response. NeuroImage, 14, 817 826. Cabeza, R. (2002). Hemispheric asymmetry reduction in older adults: The HAROLD model. Psychology and Aging, 17 85 100. Cabeza, R., Daselaar, S. M., Dolcos, F., Prince, S. E., Budde M., & Nyberg, L. ( 2004). Task independent and task specific age effects on brain activity during working memory, visual attention and episodic retrieval. Cerebral Cortex, 14, 364 3 75. Cappelletti, M., Fregni, F., Shapiro, K., Pascual Leone, A., & Caramazza, A. (2008). Processing nouns and verbs in the left frontal cortex: a transcranial magnetic stimulation study. Journal of Cognitive Neuroscience, 20, 707 720. Castner, J. E., Chenery, H. J., Silburn, P. A., Smith, E. R., Coyne, T. J., Sinclair, F., & Copland, D. A. (2007). The effects of subthalamic deep brain stimulation on disease. Journal of Neurology, Neurosurgery, and Psychiatry. dio:10.1136/jnnp.2007.18729v1. Cohen, M. L., Benjamin, M. L., McGregor, K. M., Yu Ling Chang, Chang, Y. L., White, K. D., Rackelman, C., Sherod, M., Levy, I., Zlatar, Z. Z., & Crosson, B. (February, 2009). Dedif ferentiation of Function in Older Adults during Category Member Generation. Poster presentation for the annual meeting of the International Neuropsychological Society, Atlanta, GA. Copland, D.A. (2000). A Real Time Examination of Lexical Ambiguity Reso lution Following Lesions of the Dominant Nonthalamic Subcortex Doctoral Dissertation: University of Queensland, Brisbane, Australia.
64 Copland, D. A. (2003). The basal ganglia and semantic engagement: Potential insights from semantic priming in individu als with subcortical vascular lesions, Journal of the International Neuropsychological Society, 9, 1141 1052. Copland, D. A., Chenery, H. J., & Murdoch, B. E. (2000). Processing lexical ambiguities in word tri plets: Evidence of lexical semantic deficits following dominant nonthalamic subcortical lesions. Neuropsychology, 14, 379 390. Cotelli, M., Borroni, B., Manenti, R., Zanetti, M., Arevalo, A., Cappa, S. F., & Padovani, A. (2007). Action and object namin European Journal of Neurology, 14, 632 637. Cox, R. W. (1996). AFNI: software for analysis and visualization of functional magnetic resonance neuroimages. Computers and Biomedical Research, 29, 162 173. Cr epaldi, D., Berlingeri, M., Paulesu, E., & Luzzatti, C. (2011). A place for nouns and a place for verbs? A critical review of neurocognitive data on grammatical class effects. Brain and Language, 116, 33 49. Crosson, B., Benefield, H., Cato, M. A., Sadek, J. R., Moore, A. B., Wierenga, C. E., Gopinath, K., Soltysik, D., Bauer, R. M., Auerbach, E. J., Gokcay, D., Leonard, C. M., & Briggs, R. W. (2003). Left and right basal ganglia and frontal activity during language generation: Contributions to lex ical, semantic, and phonological processes. Journal of the International Neuropsychological Society, 9, 1061 1077. Crosson, B., Benjamin, M., Levy, I. (2007). Role of the basal ganglia in language and semantics. In J. Hart, Jr. & M. Kraut (eds.). Neu ral Bases of Semantic Memory New York: Guilford Press. Crosson, B., Sadek, J. R., Bobholz, J. A., Gokcay, D., Mohr, C. M., Leonard, C. M., Maron, L., Auerbach, E. J., Browd, 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. Damasio, A. R. & Tranel, D. (1993). Nouns and verbs are retrieved with differently distributed neural systems. Proceedings of the National A cademy of Science, 90, 4957 4960. Daniele, A., Giustolisi, L., Silveri, M. C., Colosimo, C., & Gainotti, G. (1994). Evidence for a possible neuroanatomical basis for lexical processing of nouns and verbs. Neuropsychologia, 32, 1325 1341. Delis, D. C ., Kramer, J. H., Kaplan, E., & Ober, B. A. (2000). California Verbal Learning Test. Second Edition. San Antonio, TX: Psychological Corporation.
65 Delis, D. C., Kaplan, E., & Kramer, J. H. (2001). Delis Kaplan Executive Function System (D KEFS). San A ntonio, TX: Pearson Assessment. Dujardin, K., Defebrve, L., Krystkowiak, P., Blond, S., & Destee, A. (2001). Influence of chronic bilateral stimulation of the subthalamic nucleus on cognitive function in Journal of Neurology, 248, 603 611. Folstein, M. F., Folstein, S. E., & McHugh, P.R. (1975). "Mini mental state". A practical method for grading the cognitive state of patients for the clinician. Journal of Psychiatric Research, 12, 189 198. Francis, W. & Kucera, H. (1982). Fr equency Analysis of English Usage: Lexicon and Grammar. Boston, MA: Houghton Mifflin Co. Gerfen, C. (1992). The neostriatal mosaic: multiple levels of compartmental organization in the basal ganglia. Annual Review of Neuroscience, 15, 285 320. Hirsc horn, E. A. & Thompson Schill, S. L. (2006). Role of the inferior frontal gyrus in covert word retrieval: Neural correlates of switching during verbal fluency. Neuropsychologia, 44, 2547 2557. Inase, M., Tokuno, H., Nambu, A., Akazawa, T., & Takada, M. (1999). Corticostriatal and corticosubthalamic input zones from the presupplementary motor area in the macaque monkey: comparison with the input zones from the supplementary motor area. Brain Research, 833, 191 201. International Picture Naming Proj ect (IPNP). (2009). http://crl.ucsd.edu/~aszekely/ipnp/1database.html Kaplan, E., Goodglass, H., & Weintraub, S. (2001 ). Boston Naming Test. Second Edition. Austin, TX: PRO ED, Inc. Lu, L. H., Crosson, B., Nadeau, S. E., Heilman, K. M., Gonzalez Rothi, L. J., Raymer, A., Gilmore, R. L., Bauer, R. M., & Roper, S. N. (2002). Category specific naming deficits for objects and actions: Semantic attribute and grammatical role hypothesis Neuropsychologia, 40, 1608 1621. McGregor, K. M, Zlatar, Z., Kleim, E., Sudhyadhom, A., Bauer, A., Phan, S., Seeds, L., Ford, A., Manini, T. M., White, K. D., Kleim, J., & Crosson, B. (2011). Physical activity and neural correlates of aging: A combin ed TMS/fMRI study. Behavioural Brain Research, 222, 158 168. McGregor, K. M, Craggs, J. G., Benjamin, M. L., Crosson, B., & White, K. D. (2009). Age related changes in motor control during unimanual movements. Brain Imaging Behavior, 3, 317 331.
66 M einzer, M., Flaish, T., Wilser, L., Eulitz, C., Rockstroh, B., Conway, T., Gonzalez Rothi, L., & Crosson, B. (2009). Neural signatures of semantic and phonemic fluency in young and old adults. Journal of Cognitive Neuroscience, 21, 2007 2018. Middleton F.A. & Strick, P.L. (2000a). Basal ganglia and cerebellar loops: motor and cognitive circuits. Brain Research Review 31 236 250. Middleton, F. A. & Strick, P. L (2000b). Basal ganglia output and cognition: Evidence from anatomical, behavioral, a nd clinical studies. Brain and Cognition, 42, 183 200. Mink, J. W. (1996). The basal ganglia: Focused selection and inhibition of competing motor programs. Progress in Neurobiology, 50, 381 425. Nambu, A., Tokuno, H., & Takada, M. (2002). Functi onal significance of the cortico subthalamo Neuroscience Research, 43 111 117. Nambu, A., Tokuno, H., Hamada, I., Kita, H., Imanishi, M., Akazawa, T., Ikeuchi, Y., & Hasegawa, N. (2000). Excitatory cortical inputs to pal lidal neurons via the subthalamic nucleus in the monkey. Journal of Neurophysiology, 84, 289 300. Newman, S. D., Twieg, D. B., & Carpenter, P. A. (2001). Baseline conditions and subtractive logic in neuroimaging. Human Brain Mapping, 14, 228 235. Ob ler, L. K. & Albert, M. L. (unpublished). Action Naming Test. http://www.bu.edu/lab/materials.htm Oldfield, R. C. (1971). The assessment and analysis of handedness: the Edinburgh inventory. Neurops ychologia, 9, 97 113. Peran, P., Rascol, O., Demonet, J. F., Celsis, P., Nespoulous, J. L., Dubois, B., & Cardebat, D. (2003). Deficit of verb generation in nondemented patients with Movement Disorders, 18, 150 156. Perani, D. Cappa, S. F., Schnur, T., Tettamanti, M., Collina, S., Rosa, M. M., & Fazio, F. (1999). The neural correlates of verb and noun processing: A PET study. Brain, 122, 2337 2344. Persson, J., Sylvester, C. Y. C., Nelson, J. K., Welsh, K. M., Jonides, J., & Reuter Lorenz, P. A. (2004). Selection requirements during verb generation: differential recruitment in older and young adults. NeuroImage, 23, 1382 1390. Petersen, R. C., Doody, R., Kurz, A., Mohs, R. C., Morris, J. C., Rabins, P. V., Ritchie, K., Rossor, M., Thal, L., & Winblad, B. (2001). Current concepts in mild cognitive impairment. Archives of Neurology, 58, 1985 1992.
67 Peterson, S. E., Fox, P. T., Posner, M. I., Mintum, M., & Raichle, M. E. (1988). Positron emission tomographic studies o f the cortical anatomy of single word processing. Nature, 331, 585 589. Piatt, A. L., Fields, J. A., Paolo, A. M., Koller, W. C., & Troster, A. I. (1999a). Lexical, Journal of Clinical and Experimental Neuropsychology, 21 435 443. Piatt, A. L., Fields, J. A., Paolo, A. M., & Troster, A. I. (1999b). Action (verb naming) fluency as an executive function measure: convergent and divergent evidence of validity. Neuropsycholo gia, 37, 1499 1503. Saint Cyr, J. A., Trepanier, L. L., Kumar, R., Lozano, A. M., & Lang, A. E. (2000). Neuropsychological consequences of chronic bilateral stimulation of the Brain, 123, 2091 2108. Schroe der, U., Keihler, A., Lange, K. W., Haslinger, B., Tronnier, V. M., Krause, M., Pfister, R., Boecker, H., & Ceballos Baumann, A. O. (2003). Subthalamic nucleus stimulation affects a frontotemporal network: A PET study. Annals Neurology, 54, 445 50. Sha piro, K. A., Mottaghy, F. M., Schiller, N. O., Poeppel, T. D., Flub, M. O., Muller, H. W., Caramazza, A., & Krause, B. J. (2005). Dissociating neural correlates for nouns and verbs. NeuroImage, 24, 1058 1067. Simpson, G. B. & Burgess, C. (1985). Act ivation and selection processes in the recognition of ambiguous words. Journal of Experimental Psychology: Human Perception and Performance, 11, 28 39. Talairach, J. & Tournoux, P. (1988). Co Planar Stereotaxic Atlas of the Human Brain. Thieme Medic al Publishers, New York. Thompson T. (1998). Verb generation in patients with focal frontal lesions: A neuropsychological test of neuroimaging findings. Proceedings of th e National Academy of Science, 95, 15855 15860. Thompson left inferior prefrontal cortex in retrieval of semantic knowledge: A reevaluation. Proceedings of the National A cademy of Science, 94, 14792 14797. Warburton, E., Wise, R. J. S., Price, C. J., Weiller, C., Hadar, U., Ramsay, S., & Frackowiak, R. S. J. (1996). Noun and verb retrieval by normal subjects: Studies with PET. Brain, 119, 159 179.
68 Warrington, E. K. ( 2000). Homophone meaning generation: A new test of verbal switching for the detection of frontal lobe dysfunction. Journal of the International Neuropsychological Society, 6, 643 648. Wierenga, C. E., Benjamin, M., Gopinath, K., Perlstein, W. M., Le onard, C. M., Rothi, L. G., Conway, T., Cato, M. A., Briggs, R., & Crosson, B. ( 2008). Age related changes in word retrieval: Role of bilateral frontal and subcortical networks. Neurobiology of Aging, 29 436 51. epub ahead of print 4 December 2006
69 B IOGRAPHICAL SKETCH psychology. Prior to graduate school, Michelle worked as a research assistant at the Wisconsin Early Autism Project in Madison, Wisconsin where she was invol ved in studies evaluating treatment outcome effects for a behavioral modification treatment program for children diagnosed with autism and other pervasive developmental disorders. Michelle then went on to work as a research assistant for the Department of Psychiatry at the University of Iowa. At the University of Iowa, she was involved in research evaluating decisional capacity in schizophrenia, cognitive and psychiatric disease. Michelle began her doctoral training in the Department of Clinical and Health Psychology at the University of Florida in 2002 with a concentration in the area of neuropsychology. During her early graduate studies, Michelle pursued research int erests in pediatric neuropsychology in the area of pediatric traumatic brain injury. She received her Master of Science degree in clinical psychology in 2004 from the University of Florida. During her graduate studies, Michelle pursued interests in subco rtical functions and language through involvement in various research projects investigating healthy and impaired language functions using functional magnetic resonance imaging (fMRI). Michelle completed an internship in c linical p sychology at the Universi ty of Ala bama at Birmingham where she continued to develop her knowledge of clinical neuropsychology. She is completing her postdoctoral studies at the University of Alabama at Birmingham in the Department of Psychiatry and Behavioral Neurobiology.