<%BANNER%>

Effect of Goal-Setting on Memory Performance in Young and Older Adults: A Functional Magnetic Resonance Imaging (fMRI) Study


PAGE 1

EFFECTS OF GOAL-SETTING ON MEMO RY PERFORMANCE IN YOUNG AND OLDER ADULTS: A FUNCTIONAL MAGNETIC RESONANCE IMAGING (fMRI) STUDY By MICHAEL A. COLE A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2005

PAGE 2

Copyright 2005 by Michael A. Cole

PAGE 3

I dedicate this dissertation to my wife, Heather Graham Cole, and to my daughter, Tiana Cassidy Cole.

PAGE 4

iv ACKNOWLEDGMENTS I would first and foremost like to ack nowledge my wife, Heather Cole, for her unwavering support and patience in my pursuit of doctoral training. I extend my genuine gratitude to Dr. William Perlstein for the valu ed mentoring he has provided. I feel very fortunate for the excellent neuropsychology a nd cognitive neurosci ence training that I received from Dr. Perlstein and other faculty at University of Florida. Much appreciation is also extended to Dr. Robert Spencer, who helped me to build foundational research skills and a fundamental understanding of the br ain during my studies at the University of Colorado. I give many thanks to Drs. Bruce Crosson and Eileen Fennell for their exceptional mentoring in clinical neuropsyc hology, my professional development, and my research endeavors. I would like to ac knowledge Dr. Robin West for her fundamental role in this project and extend my apprecia tion for her important advising in conducting this research. I would also like to thank my dissertation committee for their guidance and support in my doctoral traini ng. I would like to extend my gratitude to Alissa DarkFreudeman, Roger Saldana, Vonetta Jones, and Neha Dixit for their contributions to this research. Special thanks are in order for my parents, Dr. Dennis G. Cole and Mrs. Barbara F. Cole, for their continued support, interest, and encouragem ent in my long road of educational endeavors. This research was supported by the Evelyn F. & William L. McKnight Brain Institute.

PAGE 5

v TABLE OF CONTENTS Page ACKNOWLEDGMENTS.................................................................................................iv LIST OF TABLES............................................................................................................vii LIST OF FIGURES.........................................................................................................viii CHAPTER 1 INTRODUCTION........................................................................................................1 Overview....................................................................................................................... 1 Episodic Memory in Young and Older Adults.............................................................3 Neural Substrates of Episodic Memory Encoding........................................................5 Overview...............................................................................................................5 Temporal Lobes...................................................................................................10 Frontal Lobes.......................................................................................................13 Neural Changes in Aging Affecting Episodic Memory Encoding.............................15 Temporal Lobes...................................................................................................16 Frontal Lobes.......................................................................................................18 Effect of Enhanced Goal-directed Action on Cognition.............................................22 Neural Substrates of Goal-direc ted Action in Cognitive Operations.........................24 Effect of Enhanced Goal-directed Action on Cognition in Older Adults...................30 Summary and Predictions...........................................................................................34 2 METHODS.................................................................................................................36 Overview.....................................................................................................................36 General Methods.........................................................................................................36 Subjects................................................................................................................36 Exclusion Criteria................................................................................................37 Experimental Task and Procedures.....................................................................38 Behavioral Data Analysis....................................................................................44 Magnetic Resonance (MR) Acquisition..............................................................45 FMRI Data Analysis............................................................................................46 Internal activation standard..........................................................................48 Memory Encoding Experiment....................................................................51

PAGE 6

vi 3 RESULTS...................................................................................................................56 Behavioral Performance.............................................................................................56 Functional MRI Findings............................................................................................58 Internal Activation Standard................................................................................58 Age Effect on Encoding Activity........................................................................60 Age Effect on Subsequent Memory Activity......................................................63 Goal-setting Effect on Encoding Activity...........................................................67 Goal-setting Effect on Subs equent Memory Activity.........................................71 Interaction of Goal-setting and Age during Encoding Activity..........................74 Interaction of Goal-setting and Age for Subsequent Memory Activity..............78 4 DISCUSSION.............................................................................................................81 Encoding in Young Adults.........................................................................................82 Encoding in Older Adults and Age-related Changes..................................................83 Goal-setting Influence on Encoding...........................................................................86 Differential Neural Response to Goal -setting in Young and Older Adults................88 Potential Limitations...................................................................................................90 Concluding Remarks..................................................................................................94 APPENDIX A SHIPLEY VOCABULARY TEST.............................................................................96 B WORD LISTS FOR MEMORY TRIALS..................................................................98 C QUESTIONNAIRES..................................................................................................99 D STRATEGY USE QUESTIONNAIRE....................................................................108 E TISSUE-AIR INTERFACE SIGNAL DROP OFF..................................................110 F GLASS BRAIN ILLUSTRATIONS........................................................................111 LIST OF REFERENCES.................................................................................................122 BIOGRAPHICAL SKETCH...........................................................................................137

PAGE 7

vii LIST OF TABLES Table Page 2-1 Mean (Standard Error) Demographic Charact eristics of Experimental Participants.38 3-1 Mean Percent (Standard Error) of Recall Performance for Each Group..................57 3-2 Comparison of Activation during Enc oding in Young and Older Adult Groups Not Receiving Goal-setting.............................................................................................64 3-3 Encoding Related Activity Differences between Young and Older Adult Groups Not Receiving Goal-setting......................................................................................65 3-4 Comparison of Subsequent Memory Effect in Young and Older Adult Groups Not Receiving Goal-setting.............................................................................................66 3-5 Subsequent Memory Differences be tween Young and Older Adult Groups Not Receiving Goal-setting.............................................................................................67 3-6 Comparison of Activation during Enc oding in the Goal and No-goal Groups........72 3-7 Comparison of Subsequent Memory Effect in Goal and No-goal Groups...............75 3-8 Subsequent Memory Effect Differen ces between Goal and No-goal Groups..........76 3-9 Interaction of Age and Goal-setting during Memory Encoding...............................79 3-10 Interaction of Age and Goal-setting during Encoding for Subsequent Memory Effects.......................................................................................................................8 0

PAGE 8

viii LIST OF FIGURES Figure Page 2-1 Graphical representation of the me mory task for Trials 1 through 3.......................42 2-2 GLM predictor model for signal intensity increases corresponding to the motor response of a button press for a single subject.........................................................50 2-3 GLM model for the multiple regression analysis of overall memory encoding for a single subject............................................................................................................52 2-4 GLM model of the multiple regression analysis of memory recall performance correlated with signal intensity................................................................................54 3-1 Summary of memory recall percent in a) young adult/goal, young adult/no-goal, older adult/goal, and older adult/no-goal gr oups, b) memory trials, and c) blocks within each trial........................................................................................................57 3-2 Left primary motor cortex (BA 4) activ ation during the motor task in young adults, older adults, and older adults subtracted from young adults....................................60 3-3 Motor task-related z-transformed signal in tensity change as a function of scan-intrial is displayed.......................................................................................................61 3-4 Effect of subsequent memory in young (n=10) and older (n=10) adults.................68 3-5 Effect of subsequent memory: older adults subtracted from young adults..............69 3-6 Effect of subsequent memory: no-goal group subtracted from the goal group........77 3-7 Coronal slice image illustrates PFC cluster (BA 9) that exhibited significant taskrelated interaction of age by goal-setting.................................................................78 E-1 Illustration of signal drop off in portions of the orbitofrontal cortex in a single subject.....................................................................................................................110 F-1 Glass-brain representati on of regional activations during encoding in older and young adults...........................................................................................................112 F-2 Glass-brain representation of the regions showing signi ficantly different activation between young and older adults during encoding..................................................113

PAGE 9

ix F-3 Glass-brain representati on of encoding activity correlated with subsequent recall performance in young and older adults..................................................................114 F-4 Glass-brain representation showing significant differences between young and older adults for encoding ac tivity correlated w ith subsequent recall performance.115 F-5 Glass-brain representati on of regional activations during encoding in goal and nogoal groups ( P < 0.001, corrected; minimum thres hold of 250 contiguous voxels).116 F-6 Glass-brain representation of the regions showing signifi cantly different levels of activation between goal and nogoal groups during encoding...............................117 F-7 Glass-brain representati on of encoding activity correlated with subsequent recall performance in goal and no-goal groups................................................................118 F-8 Glass-brain representa tion showing significant diffe rences between goal and nogoal groups for encoding activ ity correlated with subse quent recall performance.119 F-9 Glass-brain representation of regions showing a signifi cant interaction between age and goal-setting during encoding...........................................................................120 F-10 Glass-brain representation of regions showing a signifi cant interaction between age and goal-setting for encoding activity that was correlated with subsequent recall performance............................................................................................................121

PAGE 10

x Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy EFFECT OF GOAL-SETTING ON MEMORY PERFORMANCE IN YOUNG AND OLDER ADULTS: A FUNCTIONAL MAGNETIC RESONANCE (FMRI) STUDY By Michael A. Cole August 2005 Chair: William Perlstein Cochair: Bruce Crosson Major Department: Clini cal and Health Psychology Episodic memory decline in late life can be improved by use of explicit goal-setting for performance. The neural correlates that underlie the benefits of goal-setting on memory performance have yet to be ex amined. Event-related functional magnetic resonance imaging (fMRI) was employed to inve stigate the neural correlates of memory encoding as a function of age and memory e nhancement by goal-setting. FMRI data were obtained while 20 young adults (ages 18 – 28) and 20 older adults (ages 60 – 70) performed 3 trials of a list-learning task that was comprised of grocery items. Half of the young adult and half of the older adult groups received goals for performance achievement prior to each of the 3 trials, whereas the other half of the young adult and older adult groups did not receive performan ce goals. FMRI data were analyzed for signal increases related to the encoding period, as well as signal increases that correlated with subsequent recall perf ormance of the word-lists. Young adults remembered a

PAGE 11

xi significantly greater number of words than older adults Significant performance improvement derived from goal-setting was equivalent between young and older adults. Functional MRI findings revealed left lateralized prefrontal cortex (PFC) activation in young adults (not receiving goa ls), which is consistent with predictions of the hemispheric encoding and retrieval asy mmetry (HERA) model. Older adults demonstrated left lateralized PFC activation as well. Consistent with predications of the hemispheric asymmetry reduction in older adu lts (HAROLD) model, the left PFC activity was significantly dampened in older adu lts. The effect of goa l-setting on encoding activity was primarily constrained to the fr ontal lobes. Regions that demonstrated significantly greater ac tivity in the goal group than in the no-goal group included the orbitofrontal cortex (OFC), dorsolateral pref rontal cortex (dlPFC), and Broca’s area. Engagement of these regions likely refl ects increased motivation and increased mnemonic processes, such as subvocal rehear sal. In conjunction with goal-setting, older adults activated several different regions to a greater extent than young adults. As these regions were observed to be activated during encoding in the absence of goal-setting, the differentially greater activation in older adults may reflect in creased resources put toward mnemonic processing in older adults or perhaps decreased overall efficiency.

PAGE 12

1 CHAPTER 1 INTRODUCTION Overview As advances in modern medicine have en abled individuals to reach later and later decades in their lives, much interest has accu mulated into the physical and psychological changes that take place in these late stages of life. Many remain highly inventive and ingenious in their work well into the late st ages of their life, such as Michelangelo, Claude Monet, Frank Lloyd Wright, and Jose ph Campbell, but these are unique examples of notables who unfortunately are not re presentative of typical cognitive aging. A particular area of interest in cognitive aging that has evolved is that of episodic memory, a cognitive domain that expe riences one of the steepest tr ajectories of decline with increasing age (Connor, 2001; Luszcz and Br yan, 1999). The declines in episodic memory performance have been firmly demons trated, but less is known about the neural substrates that underlie this decrease in be havioral performance. Anatomical findings have revealed correlations between volumetric measures of the pref rontal cortex (PFC) and the rate of decline in episodic memo ry (Raz, 2000). Additional evidence suggests decreased activity (e.g., neural engagement) o ccurring in the PFC, as well as the temporal lobes, during memory tasks (Grady, 2000). Some strategic techniques have been s hown to be effective in improving episodic memory performance in both young and older ad ults. One such technique is goal-setting, which provides a challenge to the individual and often re sults in improved episodic memory performance (Linnenbrink et al., 1999; We st et al., in press). Findings show that both young and older adults can improve memo ry performance through the provision of

PAGE 13

2 explicit goals, which clearly ha s a neural origin. Although the neural und erpinnings of the effect of explicit goa l-setting on cognition has yet to be elucidated, ev idence suggests a role for the orbito frontal cortex (OFC) and the dorsolateral prefrontal cortex (dlPFC; Jahanshanhi and Frith, 1998; Tremblay and Schultz, 1999). The effect of goal-setting has generally been found to be as strong in older adults as in young adults (West and Thorn, 2001; West et al., in press). This finding is perhaps surprising in light of th e fact that the two primary areas th at likely participate in the goalsetting effect on memory, the OFC and dlPFC, undergo the steepest rate of degradation toward the latter stages of th e lifespan (Band et al., 2001). It is therefore important to examine if an equal extent of activation exists in the OFC and dlPFC in young and older adults resulting from goal -setting. Alternatively, compen satory mechanisms may be taking place, allowing older adults to gain equal benefit by greater activation and/or recruitment of other regions to compensate for atrophy taking place in the OFC and the dlPFC. The present research first reevaluated previous findings that episodic memory encoding in young adults is primarily lateralized to the left prefrontal cortex and that older adults engage the left prefrontal cortex to a lesser extent. This research also evaluated the role of the fr ontal cortex, specifically the OFC and the dlPFC, in the memory enhancing effects of goal-setting. Prio r studies indicate that both these regions are involved with goal-directed behaviors. Bu t the potential role for these frontal cortex regions in mediating the influence of e xplicit goal-setting on memory performance remains to be examined. It was hypothesized th at the increased chal lenge resulting from explicit goal-setting would be accompanied by significant increases in dlPFC and OFC

PAGE 14

3 activity. Additionally, it was hypothesized that older adults would show greater activation in these regions and perhaps more widespread activation to generate the equal benefit of explicit goal-setting on memory performance, compensating for the age-related declines seen in the OFC and dl PFC (Band et al., 2001). Episodic Memory in Young and Older Adults Different memory domains are not equally vulnerable to the declines occurring in older age. In fact, the inte grity of some memory domain s remains highly intact, while others are particularly suscep tible to aging. A major distin ction between memory types, and domains that demonstrate this differential vulnerability to aging, are the two types of declarative memory: semantic memory a nd episodic memory (Tulving, 1987). Semantic memory pertains to an individual’s general knowledge about the world. This information includes, but is not limited to, vocabulary, f acts, and concepts. Semantic memories are ones that an individual can vol itionally bring to consciousne ss, but typically cannot report where the specific knowledge unit was obtai ned. This knowledge is therefore not associated with specific learning contexts or events, as it is in episodic memories. Overall, studies of semantic memory integrity in older adults show very little difference from young adults in the retrie val and use of semantic info rmation (Madden et al., 1993). On the other hand, older adults do experi ence declines in episodic memory, which refers to the ability to remember specific events situated in time and place. Episodic memory is commonly thought of as the acquisiti on, storage, and retr ieval of information that refers to a specific context and is cons ciously and intentionally recollected (Tulving et al., 1994). Studies consistently show declin es in episodic memory performance in older adults (Craik and Jennings, 1992; Smith, 1996). Findings indicate that older adults often have difficulty with encoding (the initial st orage of information), as well as retrieval

PAGE 15

4 (Craik and Jennings, 1992). This decline has be en observed with virtually every type of stimulus, such as prose passages, single wo rds, spatial locations, pictures, faces, and activities (Burke and Light, 1981; Light, 1991; Smith, 1996). Additionally, age differences have been demonstrated fo r spontaneously used elaborative and organizational strategies to stor e and retrieve information, as ol der adults are less likely to engage in these sorts of mnemonics. The great est age differences are observed in tests of recall. Tests of cued recall and recognition reveal less s ubstantial, but significant differences between young and older adults (Smith, 1996). These age-related changes are not peculiar to the artificial nature of memory tests that are administered in laboratory settings because they also occur in task s that are designed to emulate memory in everyday life (Kirasic et al., 1996). Strong evidence exists that th ese age-related declines in episodic memory are not due to older adults being less motivated to remember unfamiliar or unimportant stimuli of laboratory tasks because the memory decline st ill exists for more naturalistic stimuli. These stimuli include such examples as ha nds in bridge, grocer ies on a shelf (Read, 1987), board positions in chess (Charness, 1981), instructions on bottles of prescription medicine (Morrell et al., 1990), people’s names (Cohen and Faulkner, 1986), and golf shots (Backman and Molander, 1986). A me ta-analysis by Verhaeghen et al. (1993) confirmed results from individual experi ments by reporting a negative relationship between age and episodic memory test pe rformance with recall and recognition. They found that the average older a dult above the age of 60 perfor med at a level between the 16th and 25th percentile of the young adults ’ performance distribution on various measures of recall. This result suggests th at the average older adult’s performance on

PAGE 16

5 episodic memory tasks is approximately one st andard deviation lower than that of the average performance of a young a dult (Verhaeghen et al., 1993). Especially in the last decade, resear chers have begun to examine the neural substrates that might underlie these changes in memory performance across the lifespan. Studies have focused on the neuroanatomical cha nges that exist in older adults in addition to investigations of functiona l activation of brain regions in vivo as a subject performs an episodic memory task. Research identifyi ng these neuroanatomical and functional changes has been successful in implicating likely mechanisms that may account for the decline in episodic memory performance in older age. Neural Substrates of Episodic Memory Encoding Overview For some time it has been known that a pe rson who has suffered a traumatic brain injury can have selective memory loss for events that occurred before (retrograde amnesia) and after (anterograde amnesia) the event that precipitated the traumatic brain injury. This phenomenon has been investig ated thoroughly in animal studies using approaches such as electroconvulsive shock, phy sical trauma to the brain, and drugs that depress neuronal activity or inhibit protein sy nthesis in the brain. C linical studies also indicate that brain trauma can produce amnesia that is partic ularly prevalent for recent events. Findings indicate that more recent me mories are more susceptible to disruption, whereas older memories remain quite inta ct (Kupfermann, 1991). Squire et al. (1975) investigated this phenomenon in patients with depression who received electroconvulsive treatment. They used a memory test that co uld reliably quantify the degree of memory for relatively recent events (1-2 years old), old even ts (3-9 years old), and very old events (916 years old). Patients were asked to name television programs that were broadcast

PAGE 17

6 during a single year between 1957 and 1972. The pa tients were initiall y tested and then tested again (with a different set of television programs ) after the electroconvulsive therapy. After patients received electroconvu lsive therapy, memory for television shows from less than 2 years removed were selec tively impaired, whereas memory performance for more temporally removed programs remain ed consistent with performance levels observed prior to electroconvulsive therapy. The differential susceptibility to disrupti on of memory that is dependent on the time of acquisition brings about possible expl anations for how memories are stored and how neural changes that are associated with memories are maintained for years. One possibility is that the dynam ic change that underlies the initial encoding of a memory persists and represents the l ong-term memory as well, such as a reverberating circuit. Another possibility is that l ong-term memories are related to some plastic rather than dynamic change (e.g., a persistent functional change within the brain). The extent literature provides strong support for the latter possibility. Studies have shown that by silencing the brain through use of deep anes thesia, anoxia, or by cooling the brain, shortterm memories or recent memories are disrupt ed, but older memories are not. Therefore, it can be concluded that at least older memo ries are not mediated by dynamic change, but involve physical changes in the brain. It is thought that the storage of long-term memories is in part mediated by processes such as increased protein synthesis, growth of new synaptic connections, and increased synaptic efficacy (such as long-term potentiation; Kupfermann et al., 1991). A central region in the facil itation of memory storage is the medial temporal lobe. The medial temporal lobe is needed at the time of learning to establish functional

PAGE 18

7 connections with widespread areas of neocor tex, based on neural activ ity that occurs at the time of learning. Medial te mporal lobe lesions spare sh ort-term (immediate) memory, presumably because the neocortex can suppor t short-term memory. Therefore, it is thought that the medial temporal lobe is i nvolved with processing and analyzing. This function begins at the time of learning as it receives highly processed input from neocortical association areas and continues to interact with the neocortex as it processes the information. Deficits from medial tempor al lobe lesions have been described as extreme forgetfulness, and these deficits ar e most salient after some time has elapsed after the point of learning. This dissociation between perc eption and short-term memory and long-term-memory has been well establis hed in humans, monkeys, and rats (Squire and Zola, 1997). In contrast, lesions in the neocortex impair memory abilities at both short and long delays (Goldm an-Rakic, 1987; Fuster, 1995.) The medial temporal lobe is involved in memory for a limited period of time after learning. As time passes, memory is slowly c onsolidated, and information storage in the neocortex becomes independent of the medial temporal lobe system. This is evidenced by the finding that if a medial temporal lobe lesion is sufficiently delayed after learning, memory is not affected. For instance, object discrimination tasks in monkeys demonstrated this temporally graded amnesi a with lesions at different times following learning (Zola-Morgan et al., 1986). In contrast, there is no evidence of temporally graded amnesia in the neocortex (Squire and Zola, 1997). Another important characteristic of the medial temporal lobe is that damage to this region produces memory deficits that are global and multimodal. That is, the memory impairments are present regardless of the type of material to be remembered, such as

PAGE 19

8 objects, words, or designs, or the sensory modality in which information is presented (Baltes, 1993). In contrast, memory deficits associated with neocortical lesions are domain specific. That is, they are specific to the kind of material that is ordinarily processed by the damaged area. The global and multimodal functioning of the medial temporal lobe is an important characteristic for theoretical accounts of me mory consolidation, wh ich assert that the medial temporal lobe directs consolidation in the neocortex by gradually binding together the multiple cortical regions located in di fferent areas, storing a memory for a whole event (Squire and Alvarez, 1995). Much like the medial temporal lobe, th e basal forebrain is also thought to participate in the storage of memories in th e neocortex. Neurons of the basal forebrain are activated by sustained attention in learni ng (Muir et al., 1993), which is a condition during which cortical plasticity often takes place. Evidence suggests that cholinergic and GABAergic neurons projecting from the basal forebrain (and in particular the nucleus basalis) can induce experien ce-induced plasticity changes in, for instance, auditory cortical responses. Response characteristics of the auditory cortex can be altered by repeatedly pairing of sounds with basal fo rebrain stimulation (Kilgard and Merzenich, 1998). Additionally, it has been shown that expe rience-induced plasticity can be blocked by lesioning the basal forebrain (Kilgard a nd Merzenich, 1998) or blocking cholinergic effects (Baskin and Weinberger, 1996). Study of Korsakoff’s syndrome patients has been informative in elucidating the role of the other structures important to memory function. Patients with Korsakoff’s syndrome suffer from similar amnestic features as do patients who have had damage to

PAGE 20

9 temporal lobe structures. Korsakoff’s syndr ome, which is caused by chronic alcoholism and associated nutritional deficiency, is associ ated with signs of frontal lobe dysfunction in addition to severe memory deficits. Patients exhibit pathological changes in diencephalic structures such as the ma mmillary bodies of the hypothalamus and the medial dorsal nucleus of the thalamus. Wa rrington and Weiskrantz (1982) found that when patients with Korsakoff’s syndrome are gi ven a list of words to remember, they do poorly on a simple recall task, but their pe rformance is significantly improved when recall is tested by the use of prompts or pa rtial cues. The authors concluded that this finding represents intact priming in the Korsa koff’s syndrome patients in the presence of poor episodic memory abilities. Evidence for thalamic involvement in memo ry is also derived from patients with infarctions, haemorrhages, mechanical injury, or tumor interfering with the integrity of the thalamus. Studies have shown a role for the thalamus in various cognitive functions related to memory such as the formation of new memories, attention to stimuli and events, and the use of memory strategies (V an Der Werf et al., 2003). Studies in animals indicate that a large lesion of the medial do rsal region of the thalamus is sufficient to produce learning deficits analogous to thos e exhibited by amnesic patients (Kupfermann, 1991; Van Der Werf et al., 2003). Evidence outside of Korsakoff’s syndrome patients exists for a role of mammillary bodies in memory. For instance, two cases of well documented amnesia have been reported in which marked neuronal loss was found within the mammillary bodies (Mayes et al., 1988). In addition, lesioning of th e mammillary bodies in animals results in

PAGE 21

10 significant impairments in the performance of spatial delaye d alternation tasks in rats, cats, and monkeys (Sziklas and Petrides, 1998). The above brief overview of neural subs trates underlying memory function was primarily limited to animal and neuropsycholog ical studies. Differen tiating the role of these regions in encoding versus retrieva l is difficult as the approaches are based primarily on lesion studies. With functional neuroimaging, cerebral metabolic activity is measured “on-line” as a cognitive task is being performed, and, in this way, the brain regions recruited for specific memory processes can be id entified. Functional neuroimaging studies have played an important role in evaluating the neural substrates of memory because, in part, the encoding a nd retrieval processes of memory can be observed independently in real-time. As the present research evaluates episodic memory encoding, the following sections provide a focused review of primarily functional neuroimaging findings relating to the encoding phase of episodic memory. Temporal Lobes As mentioned above, the medial temporal lobe is central to memory function. The report by Scoville and Milner (1957) of impa ired memory in patient H.M., who had a medial temporal lobe resecti on, was the first to highlight this region’s importance in memory. Later neuropsychological, neurosci entific, and psychological research all converged on the medial temporal lobe, es pecially the hippocampal portion, as the site that mediates the storage of memories fo r episodes and factual knowledge of the world (Squire, 1987; Tulving, 1987). With the advent of neuroimaging, a logical region to begin the study of episodic memory would be the me dial temporal lobe. The memory literature would suggest that episodic memory task s would surely induc e activation of the hippocampal complex. Surprisingly, though, ma ny initial studie s did not report

PAGE 22

11 significant activation in this region (Buckner et al., 1995; Petersen et al., 1988; Frith et al., 1991; Demonet et al., 1992; Grasby et al ., 1993, 1994). Several possible explanations exist that could account for these surprising findings. First, signa l difference between experimental conditions is often only 1% to 2%, and even lower in the hippocampus because it resides in a region th at is close to sinus cavities and subsequently subject to more noise. Thus, there may not have been sufficient signal over increased background noise to detect activations that may have been present. Additionally, some researchers have asserted that the nature of hipp ocampal neural functioning itself may have contributed to the inability to detect hi ppocampal activation betw een conditions (Cabeza and Kingstone, 2001). The hippocampus is central to information processing. Therefore it is frequently active, perhaps making it difficult to detect any increases in activation in an episodic memory condition. Many more recent neuroimaging studies, which have investigated hippocampal involvement in episodic memory, have at tempted a more nuanced assessment of hippocampal functioning. It was initially assume d that effort directed toward encoding and recollection should result in increased hippocampal activity. However, neuroimaging studies thus far have not found effort in encoding and recall/recognition to be an important factor as success in recollection. Specifically, hippocampa l activation has been shown to be significantly correlated to succe ss of memory performance (Fernandez et al., 1998; Yancey and Phelps, 2001). A study by Fernandez et al. (1998) found that activation in the posterior hippocampus was significantly corr elated with successful encodi ng of verbal stimuli. In a follow-up study, this same group compared results using the correlational method

PAGE 23

12 between encoding activation a nd performance with the mo re traditional method of cognitive subtraction. The cognitive subtracti on approach entails, in this context, the activation of the episodic memory encoding co mponent process having an identical task, but without the mnemonic functions subtract ed from it (Fernandez et al., 1999). They found that the entorhinal cortex did not re spond transiently as th e study word appeared (as assessed using the cognitive subtraction t echnique), but did correlate positively with subsequent test performance. This study was the first to directly compare these two techniques finding that the corr elation with performance techni que is able to much more reliably detect medial temporal lobe activation related to encoding. With this novel approach, Brewer et al. (1998) demonstrated that the strength of medial temporal lobe activity during enc oding predicts not only what items will be remembered, but also how well they will be remembered. They found that the magnitude of activation in the bilateral parahippocampal cortex predicted which picture stimuli were later remembered well, remembered less we ll, or forgotten. The distinction between remembering well and remembering less well was provided by a subjective report by the participant. Schacter and Wagner (1999) reviewed all fMRI studies of memory encoding and compared them to results of a meta-ana lysis of positron emission tomography (PET) studies of episodic memory encoding (L epage et al., 1998). Schacter and Wagner’s review found that encoding processes resulted in more posterior medial temporal lobe activation across various ma terials and conditions (Sch acter and Wagner, 1999). The meta-analysis by Lepage et al. (1998) of PET studies found a slightly different conclusion: episodic encoding is associated w ith more anterior me dial temporal lobe

PAGE 24

13 activation. Therefore, both methodologies conve rge in identifying the medial temporal lobe as an important region in encoding. The discrepancy between more anterior medial temporal lobe regions being activated in PET studies and more posterior medial temporal lobe regions in fMRI studies can likely be accounted for by the fact that PET is more sensitive to activation in the anterior porti ons of the medial temporal lobe. FMRI is known to be characterized by susceptibility artifacts that can be pronounced in the anterior medial temporal lobe, resulti ng in less sensitivity (Ojemann et al., 1997). Frontal Lobes Lack of activation of the hippocampus in many early neuroimaging studies was perplexing to those performing these experime nts. Equally perplexing was the consistent and robust activation found in the frontal lobe s. Prior to these studies, the frontal lobes were not necessarily thought to be major co ntributors to episodic memory. Patients who suffer damage to the frontal lobes, for instan ce, do not exhibit the pe rvasive and disabling amnesia that is characteristic of patients w ith hippocampal lesions (Wheeler et al., 1995). In functional neuroimaging studi es, a highly consistent, late ralization of frontal lobe function has been found with the left frontal cortex showing pre dominant activation in association with learning or encoding tasks. The primary areas within the frontal cortex where this lateralized contribution to epis odic memory encoding has been found are in ventral regions (Brodmann’s areas 44, 45, a nd 47) and dorsal regions (Brodmann’s areas 9 and 46) of the PFC. These findings have b een integrated in the HERA (hemispheric encoding and retrieval asymmetry) model, which states that the left prefrontal cortex is more involved in episodic en coding (Tulving et al., 1994). Though ventral PFC appears to contribute to episodic memory encoding, it has not been consistently found to be activated acr oss experiments. Wagner et al. (1999) found

PAGE 25

14 significant activation in the left ventral late ral frontal cortex, poste riorly in Brodmann’s area (BA) 44 and anteriorly in BA 45/47, for encoding words that would later be remembered well versus words that would late r be forgotten. Another study by Hensen et al. (1999) found that words that would be late r specifically recalled as having been seen in the study phase exhibited significantly great er correlation with activ ation in the ventral PFC than words being classified as havi ng been seen in the study phase but not specifically remembered. Fernandez et al. (1999) found that only the left BA 45 exhibited transient hemodynamic responses during encoding of verbal information. The dorsal PFC, on the other hand, is reliab ly activated during encoding of episodic memories. Studies have demonstrated activ ation in the left dorsal PFC during the encoding of words (Grady et al., 1998; Kelley et al., 1998; Kopelman et al., 1998; Nyberg et al., 1996; Wagner et al., 1998); word pairs (Dolan and Fl etcher, 1997; Fletcher et al., 1995; Halsband et al., 1998; Kapur et al., 1996); and wo rd lists (Fletcher et al., 1998). The left dorsolateral prefrontal cortex (d lPFC) has also been shown to contribute specifically to the implementation of strate gic processes for the encoding of episodic memories. In a PET study, Fletcher and colleag ues manipulated the level of attention to and the degree of organization of study material (Fletche r et al., 1998). The degree to which the subjects were required to organize wo rd lists semantically was systematically varied across three experimental conditions. Le ft dlPFC activity reache d its highest levels when organizational demands were the greate st. A role for the dlPFC in attention was suggested by dlPFC activity being attenuate d by a concurrent moto r distraction task during the most organizationally demanding task. Subsequent re trieval was also

PAGE 26

15 correspondingly attenuated under this conditi on. Wagner et al. (1999) also found dlPFC activation during the implementation of st rategic memory processes that improved memory performance. Subjects were presented w ith three words that they had to maintain either in the same order for a short period by means of subvocal rehearsal or to reorder along an abstract semantic dimension (e.g., pl easantness). Both activities engaged the dlPFC, but the semantic reordering of the wo rds resulted in greater activation of the dlPFC. The semantic reordering of the words subsequently led to better memory as well. In summary, neuropsychological an d non-primate animal studies have documented the important role of the tempor al lobes, diencephalon, and basal forebrain in memory function. Neuroimaging studies have allowed for a greater ability to evaluate the neural substrates of specific compone nts of memory function, such as encoding. Neuroimaging findings have confirmed the role of temporal lobe structures in encoding. Additionally, neuroimaging studies have reve aled the importance of the left PFC in encoding. Both the temporal lobe and PFC undergo significant changes in aging. The next section will discuss these changes a nd how they relate to episodic memory performance in older adults. Neural Changes in Aging Affecting Episodic Memory Encoding This section describes the multiple neuronal changes that take place in aging in the context of the effects on episodic memory encoding. Two key regions involved in episodic memory (temporal lobe and PFC) a ppear to experience both anatomical and physiological changes with aging that likel y serve as the underpinnings of episodic memory performance declines.

PAGE 27

16 Temporal Lobes Several studies demonstrated modest stru ctural and physiological changes in the temporal lobe regions in agi ng. Post-mortem studies of anim al and human brains reveal age-associated changes in the hippocampal si ze (Geinisman et al., 1995). In one of the most comprehensive human post-mortem studi es, the correlation value between neural counts in the hippocampus and age was r = .21 (West, 1993). Aging also affects the neuronal architecture of the hippocampus. A lthough not as rampant in the normal aging brain as in Alzheimer’s disease, neurofibrilla ry tangles display a similar characteristic regional distribution with the highest concen tration in the hippocampus (Kemper, 1994). The hippocampus shares other architectural deformities, such as Hirano bodies and granulovacular degeneration (Kemper, 1994). The hippocampus is therefore the focal point of several deleterious events associat ed with aging. However, the magnitude of these age-related effects is thought to be within the mild range as compared to effects in other brain regions (Raz, 2000). In keeping with this estimate, in vivo neuroimaging studies reveal only mild age-associated sh rinkage of the broadly defined hippocampal formation (Raz, 2000). Age-associated volume reduction takes pl ace in other temporal lobes regions outside the hippocampus as well, and this reduc tion has been estimated to be about 1% (Haug and Eggers, 1991). The reduction is thought to result primarily from a decrease in neuron size rather than an act ual loss of neurons (Haug a nd Eggers, 1991). Three research groups (Coffey et al., 1992; Cowell et al., 1994; Murphy et al., 1996) found similar correlations between age and temporal l obe volume, having a combined average correlation of r = -.25. Raz et al. (1997) f ound a correlation between inferior temporal lobe volume and age of r = -.32.

PAGE 28

17 These modest structural changes observe d in the temporal lobes are accompanied by physiological changes as well. An fMRI study conducted by Iida ka et al. (2001) provided evidence for functional changes in older adults taking place in the medial temporal lobe during episodic memory enc oding. Young and older ad ults were studied while they encoded pairs of c oncrete and abstract pictures. Age differences were found in correlations between memory performance and amplitude of signal change in the parahippocampal gyrus under both concrete a nd abstract picture c onditions. Specifically, medial temporal lobe regions were demonstr ated to experience reduced activation in older adults during encoding. A decrease in activity in the medial te mporal lobe during episodic encoding was also found in a study by Bennett et al. (2001) PET was used to measure activational differences between young and older particip ants during encoding of simple visual attributes of a sine wave grad ient screen. They found that th ere was a significant decrease in activation in the left medial temporal gyrus in older adults. Application of covariance analysis to human neuroimaging data suggests that age modifies the relationship between the hippo campus/medial temporal lobes and other brain regions. Age-related differences in the relationship between the hippocampus and the cingulate gyrus during episodic memory encoding have been suggested by Grady et al. (1995). An age-associated change in neur al interactions between the hippocampus and the dlPFC during episodic encoding has also been suggested by D'Es posito el al. (1999). However, these findings are controversial becau se they use statistical methods to parse out purported functional interact ions between regions in ne uroimaging studies that are based on independent observations of activations.

PAGE 29

18 Overall, the age-associated decline in episodic memory encoding in older adults appears to be accompanied by a slight decreas e in activation of the hippocampal/temporal lobe regions as well as modest morphological changes. It should be noted that the above functional neuroimaging studies reporting amplitude of ac tivity differences between young and older adults are based on the assump tion that the measured signal intensity differences genuinely reflect differences in neural activit y. Given that these functional neuroimaging studies are based on cerebral blow flow, it is possible that differences between young and older adults ar e indicative of reduced vascul ar responsivity to neural activation that is due to compromised cerebr ovascular function (D’Esposito et al., 1999). This possible interpretation will later be discus sed at length in the context of the present research. Frontal Lobes Episodic memory, as previously stated, expe riences one of the steepest trajectories of decline of all the cognitive domains. Perhaps not coincidentally, the frontal lobes too experience one of the steepest trajectories of decline in older age. Ample evidence suggests that structural and f unctional changes in the frontal lobes experienced in older age do indeed contribute to the decreased memo ry performance observed in older adults. Age-associated volume reduction in the frontal lobes has been estim ated to be from 10% to 17% (Haug and Eggers, 1991). This reduction may result from a reduction in neuron size rather than from an actual loss of neurons, similarly to what is observed in the temporal lobe (Haug and Eggers, 1991). Shrinkage of cells in the frontal lobes appears to begin earlier and is more severe than in any other region. An estimated 22% shrinkage in cells outside of the pyramidal layer of the PFC occurs within the fifth to seventh decades of life. Above the age of 65, the reduction in cell size becomes more

PAGE 30

19 pronounced, reaching a 43% reduction in cells ou tside of the pyramidal cell layer in the prefrontal cortex (Haug and Eggers, 1991). Several studies have also found signifi cant differences between young and older adults in the degree to which they engage these frontal regions during memory encoding. Grady et al. (1995) performed the first st udy demonstrating a functional neuronal basis for why older adults perform worse on memory tasks. This study utilized an episodic memory task that entailed memorizing faces and then later choosing the ones that were previously seen. The left prefrontal cortex was activated during encoding in young adults. In older adults, however, there was no si gnificant activation observed in the left prefrontal cortex during encoding. Studies that followed also found this difference between young and older adults. It should be noted, however, that while finding a significant difference between young and older ad ults in the left PFC, there was still significant activation in the left PFC of ol der adults in studies that followed. This difference may likely be attri buted to the statistical pow er of the experiment in conjunction with the conservative sign ificance threshold used (Cabeza, 2002). Studies that found the dampened, but still significant, left PFC activation in older adults include the previously mentioned fM RI study conducted by Iidaka et al. (2001). Young and older adults were st udied while they encoded pair s of concrete and abstract pictures. Older adults showed significant activation of the left dorsal PFC during encoding of both concrete and abstract pict ures. This activation in older adults was reduced as compared to young adults. Several other studies have looked specif ically at left PFC activation during encoding in older adults as compared to young adults. These studies consistently found

PAGE 31

20 age-related decreases in activation of left prefrontal areas: BA 6 (Cabeza et al., 1997; Cabeza et al., 1997); BA 10 (Madden et al., 1996); BA 45 (Grady et al., 1995); and BA 46 (Cabeza et al., 1997). Cabeza et al. (1997) conducted a unique approach in their PET study of episodic memory in an attempt to explain the diffe rences between memory activations in young adults and those observed in older adults. They aimed to determine if the changes in activation observed in older adults were a resu lt of local neural cha nges or if there were more global changes taking place in the way that regions interact. In order to test this, they performed a path analysis on the area s of activation observed during encoding and retrieval during an episodic memory task. Thei r path analysis indicated that the patterns of activation in older adults reflected a global shift in processing of the memory information. The authors concluded that the ne ural changes in memory encoding that take place in older adults is not limited to a few di screte regions in the brain in isolation, but rather is a global alteration in the way the many neural networks behind memory encoding interact with each other. This conc lusion is perhaps premature, however, given that nothing implicit in the design of thei r study would allow them to exploit possible connectivity changes between brain region activations. Other methods have been used to inves tigate whether global shifts in neural processing can account for the activational changes observed in older adults during mnemonic tasks, or if theses changes are spec ific to only a few discrete brain regions. Evidence for a “common cause” behind the epis odic memory declines was found in some studies (Cabeza et al., 1997; Grady et al., 1995) as evidenced by a decreased activation seen in the older adults’ fusiform and/or th e lingual gyrus. On the other hand, evidence in

PAGE 32

21 other studies suggested specific processe s as a possible etio logy, in which equal activation across age groups in the same regi ons of visual association areas was observed (BA 19 and BA 37; Cabeza et al., 1997; Madde n et al., 1999). Further support for the specific-process hypothesis can be taken from activation changes obser ved in the left and right PFC during retrieval of episodic memo ries. Specifically, significant lateralized activation in the right PFC is observed during retrieval in young adu lts, but older adults show left and right PFC activation during retrieval of episodic information (Cabeza, 2001, 2002; Madden et al., 1999). Based on this finding, the left PFC does not suffer from a generalized reduction in activity during memory tasks in older adults because it is consistently shown to have an increased act ivation in older adults during retrieval of episodic information. Overall, the decrease in activation of the left PFC during encoding observed in older adults is th erefore likely related to process-specific changes observed in aging. The findings of reduced left prefrontal activity during encoding in older adults compared to young adults have been integr ated in the HAROLD (hemispheric asymmetry reduction in old adults) model. The HAROLD model states that young subjects, in line with the HERA model, engage the left fr ontal cortex more heavily during encoding, whereas older subjects experience reduced left prefrontal activity (Cabeza et al., 2001). Numerous pharmacological and behavioral interventions have been developed in order to attempt to address these delete rious changes related to episodic memory performance decline in late life. One such intervention that appears promising is goalsetting.

PAGE 33

22 Effect of Enhanced Goal-dir ected Action on Cognition The use of explicit goals has proven to be an effective approach to improving cognitive performance, especially memory performance. Certain parameters of goals have been identified that result in the goa l being more efficacious. Goal-theory directly addresses these parameters indicating that goals must have a certain degree of specificity, difficulty, and proximity to be effective in optimizing behavioral performance (West and Thorn, 2001). Goals that are difficult and at th e same time attainable tend to motivate improvement in performance (Lee et al., 1989). Research has shown that goals that are not attainable may serve to be more di scouraging than motivating (Bandura, 1989). Support for the importance of specificity of goals has been provided by areas of education (Schunk, 1990) and organizational management (Lee et al., 1989). These authors found that goals directed toward specific performance levels, which contain measurable outcomes, proved to be the mo st efficacious. The greater clarity in the performance level expected for the goal, as well as the goal being clearly measurable, results in greater performance increases than do more general goals in which the outcome is less clearly ascertained. Proximal goals ha ve been found to be more effective than distal goals in increasing an individual’s motivation and e xpectations regarding task performance and self-efficacy on tasks that ar e even easy and intrin sically interesting (Madderlink and Harackiewicz, 1984). Last ly, Bandura (1989) found that short-term goals are more effective than long-term goals because they allow the individual to track more effectively the progress that is made. Goals have the effect of increasing the challenge to the individual for the cognitive task at hand. When a goal is util ized for a specific task, the performance requirements for the individual are raised a bove the otherwise implicit assumption that

PAGE 34

23 she/he is to perform well. Research has show n that this increase in challenge results in performance advantages. It has been shown that individuals will work towards achieving their goals (Bandura and Cervone, 1983; El liott and Dweck, 1988). Stock and Cervone (1990) have also shown that goa ls are strongly related to th e task-associated effort and persistence. Goals provide not only an increase in motivatio n for achieving success in the task at hand, but they also increase cognitive ac tivity in efforts to achieve the goal, that is, greater use of strategies (Elliott and Dw eck, 1988). A study by Hins z and Ployhart (1998) evaluated the effects of goals on the performance of a word pair memory task. They measured “trying” in this task, which they ope rationalized in terms of effort, persistence, attention, and the use of eff ective strategies while performing a task. They found that subjects provided with goals significantly in creased “trying” in the verbal memory task. A study by Juergen et al. (2001) found that goal-setting substantially improved performance in two types of memory tasks. Two conditi ons of goal-setting were employed: do your best versus specific and difficult goals. The first of the two memory tasks was a reading span test that required reading aloud sets of sentences continuously. The subject then had to recall the final words of every sentence each time all sentences of a set were processed. In the second task, a me mory span test with lists of one-syllable words was used. The goal-setting condition re sulted in greater motivation, as well as significantly greater memory performance in both of the memory tasks. Because evidence suggested that increased memory performan ce was not precipitated by different encoding or recall strategies, the authors conclude d that goal-setting re sulted in temporary cognitive arousal. A study by Linnenbrink et al (1999) found that the setting of mastery goals in a working memory paradigm also improved performance. Task-irrelevant

PAGE 35

24 thoughts were decreased and motivation for hi gher achievement was increased in the mastery goals condition, which the authors as serted to be a major contributor to the greater memory success. One’s progress on goal attainment can be monitored either through attending to their progress as they proceed through the task (assuming the performance is readily discernable by the subject) or by means of explicit, external feedback. Cognitive performance can be further improved by providi ng external feedback in addition to the increased challenge in a task provided by goals. In gene ral, explicit feedback in conjunction with goal-setting has been shown to be more effective than goal-setting alone for enhancing performance and efficacy (Bandura, 1989; Bandura and Cervone, 1983). These studies indicate that a specific goal coupled with the ability to monitor performance is the most effect approach to improvement memory performance through goal-setting. The underlying neur al substrates that mediate the effect of goal-setting remain to be elucidated. Seve ral areas of research exist, however, that bring the OFC and the dlPFC into light as being likely ca ndidates in mediating the effect of goal enhancement of memory performance. Neural Substrates of Goal-directed Action in Cognitive Operations The neural substrates supporting the explicit use of goals in the improvement of cognitive performance remain to be clearly elucidated. Ample literature exists, though, evaluating the contribution of goal-directed be havior and willed action to cognition. More specifically, studies have elucid ated the neural substrates i nvolved in motivation, drive, and/or effort arising from implicit goals that influence cogn ition. Interestingly, findings from diverse methodologies and paradigms all converge to suggest a role for the dlPFC and OFC as major participants in goal-direct ed behavior and willed action. These studies

PAGE 36

25 serve to reveal the possible neural substrates supporting the explicit use of goals in the improvement of memory. Stuss and Levine (2002) primarily utilized neuropsychological evidence to inform their conceptualization of vent ral PFC functions in goal-directed behavior. They coined the term “self-regulatory disorder” (SRD) to characterize the clinical manifestation of individuals having suffered an insult to th e ventral PFC. They define SRD as “the inability to regulate behavior according to internal goals and constraints” (Stuss and Levine, 2002, p. 405). Damage to the ventral PFC impairs an individual’s ability to maintain goals internally, which results in highly disorganized be havior. Their findings show that patients with this neuropsychological deficit rema in largely unable to perform complex goal-directed behavior. Barrash et al. (2000) performed a study in cluding 7 participan ts with bilateral ventromedial PFC lesions, 14 participants with PFC lesions but no ventromedial involvement, and 36 participants with nonfront al lesions. Subjects we re administered the Iowa Rating Scales of Personality Change in which informants rated 30 specific characteristics for degree of disturbance a nd change from premorbid personality. They found that only the bilateral ventromedial lesioned group had signifi cant impairments in goal-directed behaviors. Specifically, bila teral ventromedial patients had significant problems in planning, ini tiation, and pe rsistence. Tamm et al. (2002) investigated neural activation in females with Fragile X Syndrome and normal controls while perfor ming a counting Stroop interference task. Fragile X Syndrome is an X-chromosome linke d syndrome that results in mild mental retardation in females. The authors found th at the experimental group had a decreased

PAGE 37

26 ability for goal-directed beha viors and consequently a decr ease in performance, which was associated specifically with a reduced activation in the left orbitofrontal gyrus. Many studies have demonstrated the involveme nt of the OFC in the reward circuit, but a study by Tremblay and Schultz (1999) was able to demons trate its direct involvement in the motivational aspects of appetitive behavior. They investigated goaldirected behaviors in non-human primates, r ecording neuronal activity in the OFC during a spatial delayed responding task. Neurons of the OFC selectively became more excited in response to reward-predicting signals, duri ng the expectation of rewards, and after the receipt of rewards. The authors concluded that neurons in the OFC appear to process the motivational value regarding outcomes of voluntary action. The previous studies demonstrate that the OFC participates in goal-directed actions. As stated earlier, the dlPFC has also been shown to be involved in goal-directed activities. For instance, Jahans hai and Frith (1998) reviewed studies in which they found the dlPFC to play a critical role in willed action, which th ey defined as (1) conscious awareness and attention, (2) c hoice and control, a nd (3) intentionality. Several lines of studies are cited by these authors that meet these criteria for willed action, and in these studies the dlPFC has been found to be th e primary contributor to this function. Frith et al. (1991) used PET to study a moto r task that required the subject to move the first or second finger of the right hand at will in a random order, paced by touches to the fingers made by the experimenter. Th is condition was compared to a control condition that had the subject lif t his/her finger after the expe rimenter touched it. Random finger lifting was associated with significan tly greater dlPFC activation, as compared to the control condition. Another approach to exam ining willed action was to have subjects

PAGE 38

27 make random movements on a joystick in one of four possible directions: up, down, left, and right (Playford et al., 1992). Compared to rest condition, move ments of the joystick resulted in activation of the dlPFC. The task of random number generation ha s also been used to study willed action (Jahanshanhi et al., 1997). Random number ge neration involves operations characteristic of willed action, such as the selection and maintenance of strategies, holding information in attention, suppression of ha bitual counting, internally dr iven response generation, and monitoring of responses. This task, as compar ed to counting, activated the right dlPFC, as well as the right inferior PFC. Transcranial magnetic stimulation (TMS) has been used in research as a transient “lesion” model where it allows for the tem porary disruption of ne ural processing during focal stimulation of a local brain region. Ro et al. (1997) found that the latency increased for volitional saccades made to a central a rrowhead (endogenous go signal) that indicated the location of the required response in the right or left visual fiel d as a result of TMS over the superior PFC. TMS over the superior PFC had no effect on the saccades triggered by a peripheral aste risk (exogenous go signal) that marked the hemifield where a response was required. This indicates that the effect was not a function of disrupting visual tracking abilities due to diffuse effects of the TMS at the frontal eye fields. Additionally, TMS over the parietal cortex ha d no effect on either the volitional or triggered saccades. The transient “lesion” to the dorsal PFC region therefore resulted in an interruption of the willed action of making the saccade. Other groups have looked at the functi oning of the dlPFC in the context of cognitive control. Jonathan Cohen, Earl Miller, William Perlstein, and others have

PAGE 39

28 conceptualized the role of the prefrontal cort ex as being responsible for activities such as internal representation, maintenance, and updating of contextual information in the service of exerting control over thoughts and behavior, or “cognitive control” (Braver et al., 2001; Miller and Cohen, 2001; Perlstein et al., 2002). Contex t is referred to as any task-relevant information that is internally re presented in a way that it can bias processing in the pathways responsible for the performa nce of a task. Goal representations are one form that this information can take, as th ey have influence on planning and cognitive operations. Within this line of thinking, context is viewed as the subs et of representations within working memory that governs how othe r representations are used. These context representations are thought to subserve bot h mnemonic and control functions. Support for this conceptualization and the role of the dl PFC in context maintenance is obtained from several different domains. Neuropsychological ev idence comes in part from findings that led to the development of the term “frontal syndrome,” which refers to a particular impairment in which the normal control over so cial and sexual behavior is dysregulated (Stuss and Benson, 1986). Neuropsychological studi es have demonstrated that patients with PFC lesions show impairments on tasks involving cognitive control, such as the Stroop test and the Wisconsin Card Sorti ng Test. Neurophysiological studies with nonhuman primates have provided direct eviden ce of the dlPFC involvement with cognitive control. In experiments such as the delaye d task paradigm, neurons in the dlPFC have been found to exhibit sustained, stimulus-spe cific activity during the delay periods of simple tasks requiring the active maintenance of task relevant information (Fuster, 1989; Goldman-Rakic, 1987). More recent neuroi maging studies have corroborated these previous findings from the neuropsychologi cal and neurophysiologi cal literature. PFC

PAGE 40

29 activity has been demonstrated during a wide range of tasks involving a cognitive control component (Cabeza and Nyberg, 2000; Cohen et al., 1997; Perlstein et al., 2002). Neuroimaging studies have also confirmed that the dlPFC is specifically involved in active maintenance functions by demonstrating sustained activity in this region during the maintenance period of working memory task s (Braver and Cohen, 2001; Cohen et al., 1997; Perlstein et al., 2002). Significant evidence exists to suggest a role for both the dlPFC and OFC regions in goal-directed behaviors. Notably, Perlstein et al. (2002) provided evidence that both the dlPFC and OFC are sensitive to contextu al motivational characteristics in which higher-level cognitive tasks ar e performed. Overall, these tw o regions appear to subsume slightly different aspects of goal-directed action. The OF C appears to be involved in biasing volitional aspects of behavior by the underlying motivational state (Bechara et al., 2000). The dlPFC, on the other hand, appears to play a role in the more conscious, or salient, aspects of goal-directed behavior, such as keeping the context in mind through active goal representation (Miller and Cohen, 2001). Many studies have evaluated the role of the OFC and dlPFC in goal-directed behavior, willed action and context maintenance, but the role of thes e two regions in the improvement of episodic memory by use of explicit goal-setting has yet to be systematically evaluated. Further, these two likely candidates for contributing to the positive effects of goal-setting undergo the st eepest rate of decline in older adults. Specifically, the dlPFC and the OFC are t hought by many to be the most susceptible regions in the brain to neur onal degeneration with age (Ba nd et al., 2001). Huttenlocher (1979) found a 13% decrease of the synapt ic density in Brodmann’s area (BA) 46.

PAGE 41

30 Ulyings et al. (2000; 2002) show ed that age-related dendritic change in BA 9 and 46 are large in the pyramidal cells of layer V. An approximate loss of 20% of layer V spines in the OFC has been found (Band et al., 2001) Above the age of 65, a pronounced reduction in cell size exists where, for example, ther e is a 25% reduction in cells outside of the pyramidal layer of the OFC (Haug and Eggers 1991). In light of these findings, the potential for the beneficial effects of goa l-setting in young adults extending to older adults may appear dubious. However, as th e next section will detail, enhancing goaldirected action by goal-setting and other techni ques does in fact help older adults in cognitive tasks to an equal extent as young adults. Effect of Enhanced Goal-directed Ac tion on Cognition in Older Adults Evidence suggests that older adults may have greater di fficulties with maintaining the context of a cognitive task (e.g., important features in generating successful responses for the task). However, it has generally b een found that improving the salience of those features important for a task, or increasing the challenge to the i ndividual by techniques such as goal-setting, can compensate for decreased task performance. De Jong (2001) asserted a goal-neglect hypot hesis of age-related decrements in cognitive control. He stated that decrements in cognitive control could be characterized by a reduced capacity for goal selection and goal maintenance in working memory. In particular, De Jong stated th at under conditions of novelty or of weak environmental stimulation (i.e., reduced environmental cues that support cognitive operations), pronounced goal neglect results. Goal neglec t is defined as “disregard of a task requirement even if it has been understood, re sulting in a mismatch between what is known about task requirements and what can be done in principle, and what is actually attempted in behavior” (De Jong, 2001, p. 71). Ev idence indicates that older adults tend

PAGE 42

31 to be more dependent on and more sensitive to means for external support offered by the context of the task so that they might co mpensate for a reduced capacity for cognitive control (Hultsch et al., 1987). Evidence for the goal-neglect hypothesis comes in part from task-switching paradigms. In task-switching paradigms, the task to be performed on each trial is selected by the subject from a set of alternative tasks. The tasks are presente d in an unpredictable order, and each trial starts with the presen tation of a cue that signals the task to be performed. The cue is followed by either a fixed or random delay, which is called the preparation interval. Then the stimulus is pr esented in which the subject must perform the task. This stimulus is typically ambiguous as to which task is to be performed, rendering it necessary to keep track of the task sequence and/or pr ocess the cue effectively (De Jong, 2001). De Jong (2000) found that older adu lts experienced a failure to engage in advance preparation in task-switching in which he proposed an intention-activation account. According to this account, the effec tive utilization of advanced preparation depends on two components. The first component is an explicit goal or intention to be added to the basic goal structure that governs performance in the task-switching paradigm. The second required component is the retrieval and ca rrying out of this intention at the proper time. Success of the in tention retrieval is thought to depend on the activation level of the intention and the charac teristics or triggering power of the retrieval cue. Therefore, De Jong asserted that th e frequent failures to engage in advance preparation in older adults may be due to lo w levels of intention activation, which would reflect goal neglect. Interest ingly, the number of errors comm itted by older adults in the task-switching paradigm can be reduced to the level observed in young adults by straight

PAGE 43

32 forward speed manipulations. Specifically, if the delay period between the cue and the stimulus to perform the task is shortened, older adults perform as well as young adults. By shortening the time delay, there was an increase in challenge for older adults. According to De Jong, this increase in challenge resulted in older adults being better able to utilize available control capabilities in or der to optimize performance. Importantly, the increase in challenge decreased the po ssibility of goal neglect (De Jong, 2001). Several studies have investigated the use of explicit goals to improve performance on a cognitive task in young and older adults. That is, a specific goal for performance achievement on a cognitive task is provided, a nd then the impact of the explicit goal on performance is measured. According to De Jong’s goal-neglec t hypothesis of older adults, the use of an explicit goal woul d provide important exogenous environmental structure that would serve to help older adul ts compensate for a reduced capacity for goal selection and goal maintenance during cogniti ve tasks. One study tested young and older adults in a free-recall task in which one subgroup established a performance goal for blocks of trials and received feedback on a trial-by-trial basis. Th e other subgroup neither established goals nor recei ved feedback (Stadtlander and Coyne, 1990). Memory for random letter strings of 5 letters in 3 blocks of 50 trials was measured. The use of the motivational technique of explicit goal-setting and feedback increased memory performance in both young and older adu lts above that of the no-goal condition. West et al. (2001; 2003) re ported similar findings, showing that an increase in challenge to older adults by means of explic it goals significantly increases performance in a memory task. These studies examined the impact of goal-se tting on memory and memory beliefs across adult age groups. In one study (West et al., 2001), a baseline

PAGE 44

33 memory trial was administered, followed by three additional reca ll trials. All four memory trials entailed memorization of a gr ocery item list and subsequent free recall of the items. Young and older adults were placed in one of three conditions: goal-setting, goal-setting with feedba ck, or no goals. Goal-setting was in itiated after the baseline trial. An assessment of memory beliefs, self -efficacy, and motivation was performed in conjunction with memory performance meas urement. As was expected, young adults remembered a significantly greater number of items from the word list than did older adults across trials. Goal-setting significan tly improved memory performance in both young and older adults. The increase in chal lenge on the memory task provided by the explicit goal resulted in equal memory pe rformance enhancement for young and older adults. Additionally, motivation and self-effi cacy were both positively affected by goalsetting. The results for the goa l-setting plus feedback group were mixed. The variable outcome is thought to be related to the di ffering performance outcomes for each subject and the subsequent individual feedback that they received. For instance, subjects not meeting their goal would receive negative feedback and often would have a poorer performance in the subsequent trial, partic ularly in the older adult group. Overall, the primary finding of this study was that pe rformance was positively affected by goalsetting in both age groups, and self-efficacy and motivation were also higher after goalsetting. In another study by West et al. (in press) two different experiments investigated different levels of goal-sett ing/challenge in older and young adults in a similar shopping list recall task. In both experiments, ol der and young adults completed a baseline shopping list recall task to begin. Three more shopping list recall trials were completed

PAGE 45

34 following the baseline. For goal conditions in both experiments, subj ects were given a specific recall goal based on their own prior pe rformance prior to each of the trials (excluding the baseline trial). Additionally, all subjects in goal conditions received positive feedback for their memory gains over trials. In the first experiment, subjects were assigned a low-challenge or high-cha llenge goal. Both young and older adults clearly benefited from the hi gher challenge goal. Further, older adults experienced an equal degree of memory performance im provement from the higher challenge goal condition as did young adults. In the second experiment of this same study, a moderate challenge goal condition was compared to a no-goal condition. Results indicated that performance gains in the goal-setting condition exceeded the gains in the control group for both young and older adults. The average score gain per trial in the control c ondition (no goals) was 4.6 words for young adults and 2.1 words for older adults In the goal-setting condition, the average score gain for young adults was 5.6 words a nd 4.0 words for older adults. These score gain increases were significant for both young and older adults. Further, a significant difference was not found between young and older adults in these posi tive effects of goalsetting. Summary and Predictions Evidence clearly indicates that older a nd young adults can benefit from explicit goal-setting in episodic memory tasks. The OFC and dlPFC are likel y contributors to the improvement of episodic memory performance associated with explicit goal-setting, but this has yet to be tested. As noted previously, the two regi ons of the brain demonstrating the steepest trajectory of decline in aging ar e the OFC and the dlPFC (Band et al., 2001). This is of substantial interest because if th ese two regions do in fact correlate with the

PAGE 46

35 performance improving effects of goal-setting, th en it would be important to determine if the OFC and dlPFC sustained equal activation increases in young and older adults as a result of goal-setting. Perhaps activity in the OFC and dlPFC in older adults increases to a greater extent than in young adults in order to receive equal benefit from explicit goalsetting, compensating for the age-associated atrophy in these two re gions. Alternatively, compensation could be achieved by greater activation in regi ons involved with motivation and/or mnemonic processes in older adults in order to generate equal benefit from goal-setting conditions as young adults. In order to investigate th ese possibilities, a pilot study was performed in which 20 older adults and 20 young adults were recruite d to undergo fMRI to examine the neural substrates that underlie the memory enha ncing effects of goalsetting. Both young and older adults were assigned to one of two groups: a goal-setting group and a no-goal group. Four memory trials were conducted w ith brain-related encoding activity being measured in the last three trials by fMRI. Ac tivations of the left PFC and the temporal lobe, bilaterally, were anticipated during the verbal memory encoding task. Activation in these two regions was predicted to be damp ened in older adults, corresponding to the decreased memory performance anticipa ted for older adults. Goal-setting was hypothesized to increase performance in both young and older adults. It was hypothesized that gains in me mory performance through goal-setting would be coupled with increased activation in the OFC and dlPFC, among other regions in the PFC. In older adults, compensation for the declin e in these two regions was hypothesized by means of greater activation in the OFC and dlPFC and/or greater activation in other important motivation or mn emonic-related regions.

PAGE 47

36 CHAPTER 2 METHODS Overview A between group design was used with four groups: young adult/goal, young adult/no-goal, older adult/goal, and older a dult/no-goal. Subjects performed a memory task that entailed studyi ng a list of grocery items that th ey were later asked to recall. Brain activity during encoding was measured by fMRI and behavioral performance was measured by recall accuracy. Subjects also pe rformed a motor task that required a button press in response to a visual stimulus in orde r to assess for the possibi lity of a generalized lesser magnitude fMRI hemodynamic response in older adults. Motor response time was measured in the motor task. It should be not ed that a fixed-effects statistical approach was used for the behavioral and fMRI data due to power considerat ions, thus limiting the ability to generalize findings. General Methods Subjects Subjects were 20 young adults (ages 18-28) a nd 20 older adults (ages 60-70) with members of each age group pseudo-randomly a ssigned to one of two goal groups (goal, no-goal). Young adult subjects were primarily un dergraduate and graduate students at the University of Florida. Older adult subj ects were high-functio ning, community-dwelling individuals. Young adults were matched to older adults in ed ucation and self-rated health (see Table 2-1). Self-rated health was assesse d by asking the subject to circle a number on a scale of 1 to 10, 1 being excellent heal th and 10 being very poor health, indicating how healthy they are in general. Older and young adults were administered the Shipley

PAGE 48

37 vocabulary test (Shipley, 1940; Appendix A), on which older adults scored significantly better than young adults, t (1,38) = 7.6, p = .009 (Table 2-1). Exclusion Criteria All subjects were right-handed. Subjects we re excluded if they reported any history of neurological illness (including strokes or tr aumatic brain injury) or psychiatric illness. Subjects were also excluded if taking ps ychoactive or anticholinergic medications. Subjects taking blood pressure medication were not excluded if their blood pressure had been stable for the previous 6 months wh ile taking the medication. Subjects having a history of substance abuse or previous treatment for substance abuse were excluded. Participants with a history of epilepsy or a seizure disorder were excluded. Subjects were administered the Telephone Interview for C ognitive Status (TICS; Brandt et al., 1988) and were excluded if their score reflected the possible presence of a cognitive impairments (as defined by a score of <15th percentile). Subjects were also excluded for any MRI environment contraindications (e .g., cardiac pacemaker, implanted cardiac defibrillator, aneurysm cli p, claustrophobia, non-removabl e ferromagnetic dental work such as bridges, etc.). Lastly, subjects we re not permitted to ingest caffeine or nicotine within a 60 minute time period pr ior to the MRI session. This cr iteria reduced the risk of confounded cerebral blood flow measurements by allowing time for at least partial clearance of any potentially high levels of circulating ca ffeine or nicotine, which are vasoactive agents (caffeine half life = 3.5 hours and nicotine half life = 2 hours). High resolution T1 weighted structural MRI scans were used to determine the presence of any structural abnormalities. On e older adult was exclude d after a significant lesion was identified.

PAGE 49

38 Table 2-1 Mean (Standard Error) Demographic Character istics of Experimental Participants. Young Adults Older Adults N 20 20 Age 22.3 (0.62) 64.8 (0.55) Age Range 18-28 60-70 Sex (Men/Women) 9/11 6/14 Education 15.6 (0.41) 15.5 (0.69) Self-Reported Health 2.55 (0.33) 2.40 (0.35) Vocabularya 33.0 (0.56) 35.2 (0.60) aOlder adults significantly greater than young adults: t (1,38) = 7.6, p = .009. Three older adults and two young adults were excluded due to im age distortion that was caused by an inability to position th e head far enough into the scanner bore longitudinally to be within the optimal image acquisition range. On account of the positioning difficulty, images were subjected to large field inhomogeneity and distortion. One older adult was excluded due to a claust rophobic reaction to th e scanner. Lastly, collection of data from one subject was termin ated in the middle of testing due to an MRI schedule conflict. Experimental Task and Procedures Participants performed a free-recall epis odic memory task modeled after that employed by West (2001; 2002). Word lists, co mprised of categori zable grocery items, were presented. This task was originally developed as follo ws. In order to crea te a large pool of categorizable items, four research assistants listed all items they found in grocery stores and also identified a large num ber of narrowly defined categories for those items. Nine independent raters then determined the mo st suitable category for each item and rated each item as a high, moderate, or low frequenc y exemplar of its category. In order for an

PAGE 50

39 item to be included in the final word list, the word must have met a predetermined criterion rating such that seven of nine ra ters rated the item as a “high frequency” exemplar of its category (West and Thorn, 2001). The episodic memory task was modified from its original form for administration in the imaging context. Four memory trials were performed (Baseline Trial and Trials 1 through 3). All four groups (young adult/goal, young adult/no-goal, older adult/goal, older adult/no-goal) were administered id entical memory trials and word lists. The Baseline Trial was administered prior to entering the scanner and was comprised of a 15-word list presented on a single sheet of paper (Appendix B). Participants were given a one -minute study period, and then were prompted to write out all the words that they could remember on a nu mbered sheet of paper. Instructions given to the participant were as follows: “On this task, you will be asked to study a list of items that can be bought at a store. You are not e xpected to remember every item on the list. Just do your best. I will tell you when to be gin studying the list and I will also let you know when your study time has ended. You will be given approximately 1 to 3 minutes to study the list. You may not write during th e study time. After th e study period is done, I will ask you to write down all of the items that you remember from that list. You will write the remembered items on this page.” Following the Baseline Trial, participan ts were placed in the MR scanner and performed the remaining three memory trials (Trials 1 through 3). Memory trials were computer controlled by the software package PsyScope (Cohen et al., 1993). Participants were presented with the identical 42-item grocery list for Trials 1 through 3 (Appendix B). Included in the 42-item list were all 15 items contained in the Baseline Trial list. For

PAGE 51

40 each of the three trials, word-list content of the 7 blocks was identical. However, the order of presentation for the 6 words within each block wa s randomized for each trial. The word content of each of th e three trials was identical in order to observe continued improvement across trials in conjunction with th e goal-setting conditi on. The content of the word list is detailed in Appendix B. In structions given prior to Trial 1 were as follows: “In these next activities, you will again be asked to study a list of items that can be bought at a store, except this time they wi ll be presented on a screen. After an item is presented, there will sometimes be a short de lay before another item is presented on the screen, and other times there will be a li ttle longer delay. During these delays, you are asked to attend to the target in the center of the screen. When all the items have been presented, I will ask you to tell me back as many as you can remember. You will not be expected to remember all of th em.” Instructions prior to Tr ials 2 and 3 were not stated vocally but appeared on a liquid crystal disp lay (LCD) screen (described below). The instructions read: “Study the following list of words.” The vocal recall of word lists by the subjects and task instruction delivery we re achieved by a bi-directional in-scanner microphone and receiver. The 42-word trials were presented singly in 6-word blocks, for a total of 7 blocks per trial (see Figure 2-1). Words were pres ented singly for a 2 s duration with a 2 s interstimulus interval (ISI). Subjects were presented th e words through a mirror system orientated towards an LCD screen mounted in the radio frequency co il (RF) 4 in. behind the top of the subjects’ head (approximate visual angle = 41 degrees). All 6-word blocks lasted a total duration of 24 s, followed by a 10 s interblock interval. During the interblock interval, subjects were asked to at tend to a fixation point (“+”) in the middle of

PAGE 52

41 the screen. All three trials were initiated by a 16 s baseline fixation period during which the subjects were asked to attend to a fixati on point (“+”) in the middle of the screen. Image acquisition began at the onset of th e baseline fixation period. The first 2 image volumes were automatically discarded (TR=2 s) so the first image that was collected for later analysis was the 3rd image volume. Consequently, 12 s (6 image volumes) of the 16 s baseline fixation period were analyzed. Afte r the seventh block of words, there was a 16 s recovery fixation period when the subject was again asked to attend to the fixation point. Image acquisition terminated at the conc lusion of the recovery period. Free recall of the 42-item list began immediately followi ng the recovery fixation period and subjects were given a maximum of 4 min for recall. During the recall peri od, subjects verbally recalled as many of the 42 items as they coul d remember, in any order. Vocal responses were digitally recorded and immediately sc ored for accuracy. Image volumes were not collected during the reca ll period. In total, 133 images were collected per trial, yielding 399 images for each subject. All groups performed the above describe d memory task. The only procedural difference was between the goal and no-goal groups, which was implemented between the memory trials. Prior to Trials 1, 2 and 3, subjects in the goal group received feedback regarding the number of items they remember ed correctly from the previous trial in addition to a goal statement for that trial: “Your goal is to achieve a 50% improvement in your score.” Prior to Trials 2 and Trial 3, s ubjects in the goal gr oups also received a positive feedback statement (e.g., “Good job th at’s a great score”) and after the goal statement, “Keep trying” was written. Following the final trial, Trial 3, these statements and feedback were given again. All statemen ts and feedback were provided on the LCD

PAGE 53

42 Figure 2-1 Graphical representation of the memory task for Trials 1 through 3. screen. Subjects alerted the experimenter ve rbally when they completed reading each screen. Subjects in the no-goal group did not receive any of the feedback described above. Feedback was not included for the goal group in order to maintain paradigmatic consistency with previous studies finding im proved performance from goal-setting (West et al., 2002, 2003). However, communications w ith the goal group were equated in the no-goal group by inserting the statements : “You have completed trial number x ” and “We are now ready to begi n the next trial.” The Baseline Trial contained only 15 word s, as compared to 42 words in the following 3 trials, so that subjects woul d not be overwhelmed by the task (as was observed in pilot versions of this paradigm ). Also, the initial goal of 50% improvement would be more easily attainable, again reduci ng the risk of overwhelming the subject at the onset of the memory trials. Baseline Fixation Period Word Bloc k 2 s presentation duration p er word 2 s ISI 24 s tota l Interblock Interval Recover y Fixation Period 16 s Recall 10 s WORD 16 s 6-word block & interblock interval are repeated 7 times f or a total o f 42 g rocer y items per trial

PAGE 54

43 All experimentation was pe rformed during one visit to the MRI facility. The first portion of experimentation took place in a test ing room inside the MRI facility (although outside the scanning environment). To begi n, informed consent was obtained from the participant in a manner consistent with the University of Florid a Institutional Review Board regulations. Subjects then completed se veral questionnaires that inquired about their opinions about their memory, demographic information, among other similar information (Appendix C). Then the Base line Trial was conducted with recall immediately following. Following the baseli ne trial, another questionnaire was administered, which was also related to the participants’ opinion of their memory. The next procedure varied depending on wh ether the subject was in the older adult group or the young adult group. Because all 15 words from the Baseline Trial were included in subsequent memory trials, it was important that a consistent time frame be maintained between the administration of the Baseline Trial and the commencement of Trial 1 for all participants. (It should be noted that the subjects were not made aware that the exact same words would be repeated in subsequent trials). The older adults, on average, require a greater amount of time to complete the questionnaires following the Baseline trial. Thus, a temporal window of 30 40 minutes between the completion of the Baseline Trial and the start of Trial 1 was maintained for both young and older adults by giving the young adults a part icipant information form and the vocabulary test to complete before proceeding to the MRI scanner. Older adults, on the other hand, proceeded directly to the MR I scanner after completion of the initial questionnaire. (Older adults instead completed the vocabul ary test and participant information sheet after the MRI testing was complete.) Conse quently, the time between the Baseline Trial

PAGE 55

44 and Trial 1 were matched between young and olde r adults in temporal length and activity, as both were completing questionnaires. Afte r completing their respective questionnaires, subjects were then prepared for the scanning environment and placed in the scanner bore. After localizing sequences were performe d, Trials 1 through 3 were administered. After completion of the three memory trials in the scanner, subjects then completed a motor task consisting of pressi ng the right index finge r button of a Button Response Unit (BRU) every time a large white square appears on the LCD screen. The stimulus duration was 1s followed by a fixati on period of 9 s. Each motor task session contained 9 trials and there we re a total of 3 motor task sessions completed (e.g., 27 total motor responses). This task served as an inte rnal activation standard (described in further detail shortly). Subjects then completed 5 more questi onnaires following the scanning portion of the experiment that again asked for their opin ions about their memory and also how they felt they performed on the memory task. An additional questionnaire asked about strategies utilized to complete the memory task (Appendix D). Behavioral Data Analysis The percent of items recalled from each 6-word block served as the dependent variable. The Baseline Trial was excluded b ecause it was conducted outside the scanner and procedures for the goal and no-goal groups were identical. Due to power considerations, a fixed-effects analysis wa s conducted. The error variance term was estimated on a block by block basis with each block representing an independent observation. The percent of items recalled for each block was evaluated by analysis of variance (ANOVA) with Group (young adults, older adults) and Condition (goal, no-

PAGE 56

45 goal) as the between-subjects f actors and trial (trials 1 – 3) and block (blocks 1 –7) as within-subjects factors. Magnetic Resonance (MR) Acquisition Scanning was conducted in a Siemens Allegra 3.0 Tesla head-only MR superconducting system (128 MHz; 60cm bore) at the University of Florida McKnight Brain Institute MR Facility. Images were ac quired using a Siemens quadrature head radio frequency (RF) coil. A BOLD sensitive echo-planar imaging pulse sequence (EPI; Siemens; TR=2000 ms, TE=30 ms, FOV=240 mm, flip angle=90, 64 x 64 matrix) was used to acquire 31 slices (voxels = 3.75 mm2 in-plane, 3.2 mm thick, 0.32 mm gap) in the axial oblique plane. A 3-plane localizer was first acquired onto which a sagigittal scout was prescribed. The prescription was acquired with 31 contiguous slices perpendicular to the anterior commissure-pos terior commissure line. A doubl e-oblique prescription was collected in order to reduce potential imag e misregistration across subjects due to differences in ventral head orientation. F unctional scanning was synchronized to trial onset (baseline fixation period) and terminated at the end of the recovery fixation period. Following functional scanning, structural images were acquired with a 3-D magnetization-prepared rapi d acquisition gradient echo (MPRAGE) T1-weighted pulse sequence (128 slices, 1.3 mm slice thickne ss, TE=4.13 ms, FOV=240 mm, flip angle=8, 512 x 512 matrix It is important to note that an area of specific interest in this study, the orbital frontal cortex, contains a hi gh potential for signal drop o ff, especially in the high magnetic field strength of 3 tesla. The orbito frontal cortex borders the orbital sinus and the auditory meatus, creating su sceptibility artifacts at the tissue-air interfaces. Indeed,

PAGE 57

46 signal loss was observed in the most ventral and anterior portions of the OFC, as pictured in Appendix E. FMRI Data Analysis Data analysis, registration and visuali zation were performed with the fMRI software package BrainVoyager 2000 (Brain Innovation, Maastricht, The Netherlands). FMRI data reduction The first two volumes (4 s) of each functional scanning run were automatically discarded by the Siemens scanner to allow for T1 equilibrium and thus were not included in any of the analyses Prior to data analysis, functional images were aligned to the last volume for each sl ice in order to minimize the signal changes related to rigid body rotation and translati on during the acquisition. Following movement correction, images were spatially smoothed w ith a Gaussian kernel, FWHM = 8 mm, to accommodate for differences in anatomy. Three-dimensional motion correction and Talairach transformation were performed for th e functional data of each subject. Linear drifts of the signal with respect to time were removed from each pixel’s time course. The 3-D anatomical volumes and 2-D func tional volumes were resliced to a 256 x 256 matrix size. Reslicing of functional volum es took place before statistical analysis. The functional volumes were analyzed in 3-D space. Statistical maps were superimposed onto 3-D anatomical data sets. Since the EPI functional scans and 3-D structural measurements were performed within the same recording session and contained the exact same positioning parameters, co-registration of the respective data sets were performed semi-automatically based on the Siemens sli ce position parameters of the T2*-weighted measurement (number of slices, slice thickne ss, distance factor, pitch angle (axial – coronal angle), FOV, shift mean, off-centre r ead, off-centre phase, in plane resolution) and on parameters of the T1-weighted 3-D measurement (number of sagittal partitions,

PAGE 58

47 shift mean, off-centre read, off-centre phase, resolution) with respect to the initial overview measurement (prescription). Manual co -registering corrections in the x, y, and z planes and in the pitch angle were required to optimize final alignment. For each subject, the structural 3-D data se ts were transformed into Talairach space using a two step process. The first step c onsisted of rotating the 3-D data set of each subject to be aligned with the stereotaxic atlas. For this step the location of the anterior commissure (AC), the posterior commissure (PC), and two rotation parameters for midsagittal alignment had to be specified manually in the high resolution 3-D volumetrics. In the second step the extreme points of the cerebrum were specified. These points together with the AC and PC coordinates were then used to scale the 3-D data sets into the dimensions of the standard brain of the Talairach and Tournaux atlas (Talairach and Tournaux, 1988) using a piecewise affine and continuous transformation for each of the 12 defined cerebral borders. FMRI data analyses. Statistical analyses were performed using BrainVoyager by fitting a general linear model to the individual fMRI time series data (e.g., multiple regression analysis). The predictor variable s were created for each condition creating the idealized time series that represented the response to the condition (e.g. encoding period or motor response) in each group. The hemodyn amic response for predictor variables was estimated by convolving each regressor with a standard gamma variate function that shifted the estimated hemodynamic response ap proximately 6 seconds to account for the expected delay. A weighted sum of the predic tor variables was crea ted that produced the closest match to the actual data time series A parameter estimate (e.g., beta weight) was generated for each model that estimated the strength of covariance between the actual

PAGE 59

48 data and the modeled hemodynamic respons e function. Parameter estimates were calculated for each participant, which we re then subjected to group analysis. For all analyses, statistical maps genera ted for each pattern of interest were thresholded for significance using a cluster-size algorithm (Forman et al., 1995) of 80 voxels, which protects against an inflati on of a false-positive rate with multiple comparisons. The voxel size was 0.9375 mm2 in plane and 1.3 mm thick. Due to power considerations, a fixed-effect analysis appr oach was utilized, which estimates the error variance on a scan to scan basis. Significan t effects are shown onl y if the associated P valued yielded P < 0.01 (Bonferroni corrected for multiple comparisons using the number of voxels exceeding the minimum in brain th reshold signal criteri a intensity of 250). It is important to note that relative di fferences between groups reflect differential activation intensity for a specific prescribed area, and do not indicate a greater or lesser spatial extent of activation. That is, relative differences of activation asserted to be present between groups reflect a significantly gr eater or lesser signal intensity measured for a discrete group of voxels. Internal activation standard In order to establish that differential activation observed in older adults is not simply a generalized change in neural act ivation, an “internal activation standard” (Weinberger et al., 1996) was utilized. Th e blood oxygen level depe ndent (BOLD) signal of fMRI depends on neurovascular coupling, wh ich is a process in which neural activity influences the hemodynamic properties of the surrounding vasculatur e. There could be direct changes in the cerebral vasculature as well as alterations in the complex neurochemical transformation of neural activ ity into changes in blood flow that might affect the measured BOLD response (D ’Esposito, 2003). Age-associated signal

PAGE 60

49 differences have not been demonstrated in components of the BOLD hemodynamic response function in the motor cortex in re sponse to a buttonpress (D'Esposito et al., 1999). Thus, the neural activation in the precen tral gyrus, Brodmann’s area 4 (BA 4), in response to subjects’ buttonpress was utilized as the internal activation standard. Absence of age-related BA 4 signal intensity differences supports the specificity of any differences observed between young and older ad ults in the memory task. The general linear model (GLM ) of the experiment was computed from the 40 (40 subjects; 3 motor trials per subject collaps ed) z-normalized volume time courses. For each time course, the 1 s period of stimulus presentation (white square to which the subject responds with a button press) was defi ned to represent motor activity. The signal values during this phase were considered th e effects of interest. Response latencies in older adults (median = 404 ms, standard er ror = 24 ms) and young adults (median = 397 ms, standard error = 22 ms) occurred well w ithin the temporal frame of the defined predictor response of the GL M model. The baseline fixation period (9 s following the motor stimulus cue) was defined as the ba seline (non-motor) period. For each trial, 90 volumes were collected. There were a total of three trials so 270 volumes were collected for each subject. The resulting predictor was obtained by shifting an ideal box-car response (assuming a value of 1 for the volum es of the respective motor periods and a value of 0 for the baseline periods) by a st andard gamma variate f unction prescribed by Brainvoyager (in order to account for the hemody namic delay). This predictor, which was identical for all subjects across all groups, wa s used to build the design matrix of the experiment (see Figure 2-2). The above steps generated a 4-D functiona l time series (volume time course: 3 x

PAGE 61

50 Figure 2-2. GLM predictor model for signal intensity increases corresponding to the motor response of a button press for a single subject. The model represented here was identical for all subjects. The green shading represents the presence of the predictor, in this case a motor response, and the gray area indicates the rest period. space, 1 x time) for the predictors of the mode l. Statistical analysis of 4-D functional time series included first single-subject multiple regression analysis, followed by multi-subject multiple regression analysis that concatenat ed the single-subject analyses. The global level of the signal time course in each sessi on was considered to be a confounding effect and was entered as such into the GLM mode l. A fixed effects analysis was employed. Statistical maps generated from the multi-subject analyses were projected on the flattened surface of a volumetric rendering of all 40 subjects’ high resolution 3-D volumes averaged together in Talairach space. Three st atistical maps were generated for the motor

PAGE 62

51 predictor model: 1) young adults, 2) older adu lts, and 3) young adults directly compared to older adults (contrast map). Button press response latencies to the visual cue were measured in all subjects. A comparison between young and older adult re sponse latencies was conducted by first determining the median response time for each subject. A between subjects (young and older adults) analysis of va riance was conducted on the medi an reaction time for each subject. The motor task was employed for an additiona l analysis to evaluate the possibility of an altered hemodynamic response in olde r adults taking blood pressure medication. The hemodynamic response during the motor task in the 6 older adults taking blood pressure medication and 6 pseudo-randomly selected older adults not taking blood pressure medication were compared. Memory Encoding Experiment Data were analyzed separately for en coding and for relation of encoding to blockwise memory recall performance. Encoding-related activity. A general linear model (GLM) for the experiment was computed from the 40 (40 subjects; 3 memory trials per subject collapsed) z-normalized volume time courses. For each time course, the 24 s period of word presentation was defined to represent encoding. The signal va lues during these phase s were considered effects of interest. The baseline fixation pe riod (12 s), interblock intervals (10 s), and recovery fixation period (16 s) were defined as the re st (non-encoding) period. The resulting predictor was obtained by shifting an ideal box-car respons e (assuming a value of 1 for the volumes of the respective encodi ng phases and a value of 0 for the baseline time points) by a standard gamma variate f unction prescribed by Br ainvoyager (in order

PAGE 63

52 to account for the hemodynamic delay and form approximate hemodynamic rise and fall times). This predictor, which was identical for all subjects across all groups, was used to build the design matrix of the experiment (see Figure 2-3). Figure 2-3. GLM model for the multiple regre ssion analysis of overall memory encoding for a single subject. The model represente d here was identical for all subjects. The green shading represents the presen ce of the predictor, in this case encoding activity, and the gray area indicates the rest period. Subsequent memory activity The above model reflecte d overall neur al activity corresponding to encoding the 6-word blocks. In order to evaluate how encoding activity related to subjects’ recall performance, an additional GLM model was used that correlated voxel signal intensity with reca ll performance on each of the 6-word blocks. This approach has become known as the ‘subsequent memory’ effects approach to analyzing encoding (Rugg et al., 2002). In the subsequent memory procedure, eventrelated activity elicited by a se ries of study items is contrast ed according to the number of

PAGE 64

53 items remembered or forgotten on a subseque nt memory test. The assumption behind this analysis is that differences in activity that predict successful versus unsuccessful memory reflect the different levels of engagement of processes s upporting effective encoding. As an example of this approach, a subject may have remembered 5 of 6 words on the first block of a trial, so .83 (83% correct) is entered for each of the 12 volumes (covering the 24 s period when the 6 words were presented) for the encoding predic tor. Likewise, recall performance values are entered into the en coding predictor for each of the following 6word blocks. The resulting predictor for the GLM model represents higher signal intensity for blocks in which the subject pe rformed well, and lower signal intensity for blocks in which the subject performed poorl y. In other words, the model represents a correlation of recall performance for each 6-word block and signal intensity increases above baseline during encoding periods. Enc oding time periods are thus set as relative increases in signal intensity corresponding to recall performance for each respective 6word block and are corrected for the he modynamic response delay (see Figure 2-4). For both the encoding activity model and th e subsequent memory activity model, the global level of the signal time course in each session was considered to be a confounding effect and was entered as such into the GLM model. A fixed effects analysis was employed. The creation of other alte rnative GLM models further evaluating encoding was constrained by software limita tions. For instance, creation of a model investigating purely recall performance correlat ion with signal inte nsity was not possible because it was necessary to include rest periods into the model. The above steps generated a 4-D functiona l time series (volume time course: 3 x space, 1 x time) for the predictors of the 2 m odels. Statistical analysis of 4-D functional

PAGE 65

54 Figure 2-4. GLM model of the multiple regression analysis of memory recall performance correlated with signal intens ity. The green shading represents the presence of the predictor, in this case encoding activity correlated with subsequent memory performance, and th e gray area indicate s the rest period. The illustrated model is an exemplar for a single subject. A separate model was calculated for each subject that was based on their individual recall performance during each of the 21 blocks. time series included first single-subject multiple regression analysis, followed by multisubject multiple regression analysis that concatenated the single-subject analyses. Statistical maps generated from the multi-subject analysis were projected on the flattened surface of a volumetric rendering of all 40 subjects’ high resolution 3-D volumes averaged together in Talairach space. Seve n statistical maps were generated for each model (e.g., encoding activity model and subs equent memory activity model): 1) young adult/no-goal, 2) older adult/no-goal, 3) young adult/no-goal compared directly to older adult/no-goal (contrast map), 4) goal group, 5) no-goal group, 6) goal group compared

PAGE 66

55 directly to the no-goal group (contrast map), and 7) interaction of goal-setting group by age.

PAGE 67

56 CHAPTER 3 RESULTS Behavioral Performance Encoding accuracy was scored as the per cent of words correctly recalled for each block of words (21 blocks per subject). Data were analyzed with a fixed-effects ANOVA with between-subjects factors of age (young adults, older adul ts), goal (goal, no-goal) and within subjects factors of trial (trials 1 through 3) and block (blocks 1 through 7). Group means and standard errors are reported in Ta ble 3-1. There was a main effect of age: F [1,756] = 7.31, P = 0.007, Cohen’s d = 0.16, reflecting better recall memory performance in young adults than in the older adults (see Figure 3-1). There was a main effect of goal-setting: F [1,756] = 9.13, P = 0.003, Cohen’s d = 0.18, reflecting better performance in the goal than no-goal group (s ee Figure 3-1). A main effect of trial, F [2,756] = 77.33, P < 0.001, reflected subjects being able to perform progressively better on each successive trial, as lists were identical in content fo r each trial (see Figure 3-1). There was a main effect of block: F [6,756] = 12.16, P < 0.001, reflecting subjects’ tendency to recall words better in the earlier than later bloc ks (e.g., primacy effects; see Figure 3-1). Notably, there was not a significant interaction between goal-setting and age groups: F [1,756] = 0.652, P = 0.420, indicating that ne ither age group benefited disproportionately from the prov ision of goals. Interestingly, the effect of goal-setting in older adults brought their pe rformance up to levels of performance in young adults without goal-setting. Following testing, subjects repo rted all the strategies th ey used for performing the memory task. The number of strategies us ed by subjects was analyzed using an ANOVA

PAGE 68

57 with factors of age (young adults, older adults ) and goal (goal, no-goal) A main effect of goal was not found, as subjects in each group used, on average, the same number of Table 3-1. Mean Percent (Standard Erro r) of Recall Performance for Each Group Goal-setting Age Mean Goal Older Adults 49.5 (1.7) Young Adults 55.0 (1.7) Total 52.2 (1.2) No-Goal Older Adults 46.2 (1.7) Young Adults 49.0 (1.7) Total 47.6 (1.2) Total Older Adults 47.8 (1.2) Young Adults 52.9 (1.2) Total 49.9 (0. 8) 44 46 48 50 52 54 56 58Young Older N o-goa l GoalPercent Recal l a. 25 35 45 55 65 123TrialPercent Recal l b. 30 40 50 60 70 1234567Percent Recal l Blockc. Figure 3-1. Summary of memory recall per cent in a) young adult/goal, young adult/nogoal, older adult/goal, and older adult/no-goal groups, b) memory trials, and c) blocks within each trial. Standard error bars are shown. strategies to approach the task. However, a trend was observed for young adults to use a greater number of strategies in performing the memory task than older adults, F (1,36) =

PAGE 69

58 3.53, P = 0.068. Young adults used a mean of 7.0 strategies (standard error = 0.4), as compared to 6.4 (standard error = 0.3) strategi es used by older adults. Follow-up analysis revealed a significant difference between th e young adult/no-goal (mean = 7.8) and the older adult/no-goal (mean = 6.2) groups, (Student’s t -test, t (1,19) = 13.9, P < 0.001, Cohen’s d = 0.54). Functional MRI Findings Internal Activation Standard The motor task was employed to assess for generalized lesser-magnitude signal increase in older adults. The statistical th reshold that was used for the memory study analyses ( P > 0.01, Bonferroni corrected for multip le comparisons; minimum threshold of 80 contiguous voxels) was used for the moto r task study in order to facilitate cross study comparison of potential lesser-magnit ude signal increases in older adults. Activation in the left precentral gyrus (BA 4) was selectively examined as the task involved the contralateral ri ght index finger response. Left BA 4 activation in young adults (center of gr avity: x = -41, y = -17, z = 48; 3241voxels ; center of gravity refers to the geographical center of a si gnificantly activated cluster, independent of the relative statistical significance magnitude) and older ad ults (center of gravity: x = -41, y = -17, z = 46; 2502 voxels) was significant (see Figure 3-2; P > 0.01, Bonferroni corrected for multiple comparisons; minimum threshold of 80 contiguous voxels). A follow-up analysis, which directly compar ed activation in young adults to older adults, revealed no significant differences between young and ol der adults in BA 4 (see Figure 3-1; P > 0.01, Bonferroni corrected for multiple comparis ons; minimum threshold of 80 contiguous voxels). Figure 3-3 illustrate s the hemodynamic response to the button press in young and older adults. A group by linear trend over scan analysis of signal intensity was not

PAGE 70

59 significant, suggesting that young and older adults did not differ significantly in the hemodynamic response curve characteristics (group x linear trend over scan analysis: F (1,38) = 0.08, P = 0.785). A random effects analysis of the median reaction time of each subject comparing young and older adults revealed that there wa s not a significant difference in the time taken for the button press response after the visual cue ( F [1,38] = 0.07, P = 0.79). The response latency median values were 404 ms (standard error = 24 ms ) for older adults and 397 ms (standard error = 22 ms) for young adults. In order to address concer ns regarding the possibility of altered hemodynamic response in older adults ta king blood pressure medication, their BOLD response during the motor task was compared to 6 pseudo-ra ndomly selected older adults not taking blood pressure medication. There was significant ac tivation in the contra lateral left primary motor cortex (BA 4) in blood pressure medica tion free older adults ( center of gr avity: x = -42, y = -18, z = 48; 1204 voxels). Significant activation of left BA 4 was also observed in older adults taking blood pressure medica tions (center of gravity : x = -37, y = -22, z = 56; 435 voxels). Importantly, there were no si gnificant differences between subjects taking blood pressure medications and subjects that were blood pre ssure medication free older adults in left BA 4, or any other regions ( P > .01, Bonferroni corr ected; threshold of 80 contiguous voxels). As this analysis is susc eptible to a Type II statistical error due to low power (e.g., each group had 6 subjects), a mu ch more liberal threshold was utilized as well. There again was no significant difference observed at a threshold of P < 0.01, uncorrected for multiple comparisons, a nd a minimum of 10 contiguous voxels.

PAGE 71

60 Age Effect on Encoding Activity In order to evaluate the effect of ag ing on encoding, only the young adult and older adult no-goal groups were compared so as not to confound this comparison with the Young Adults Older Adults Young Adults – Older Adults Figure 3-2. Left primary motor cortex (BA 4) activation during the motor task in young adults, older adults, and older adults subtracted from young adults. Statistical threshold was set at P < 0.01 (Bonferroni correcte d for multiple comparisons; minimum of 80 contiguous voxels). Activ ation statistical maps are displayed on the smoothed, averaged T1 image for all subjects in the analysis. Central point of the cross-hairs i ndicates the center of gravity for activation within BA 4.

PAGE 72

61 -0.2 -0.15 -0.1 -0.05 0 0.05 0.1 0.15 0.2 0.25 12345VolumesZ-Transformed Signal Intensity Young Adults Older Adults Motor Response Figure 3-3. Motor task-related z-transformed signal intensity change as a function of scan-in-trial is displayed. Data were obtained from the left BA 4 in young adults (n=20) and older adults (n=20) Scan by group analysis revealed no significant differences in the curvatur e of the hemodynamic response between young and older adults. presence of goals. Brain regions demonstra ting significant activations for encoding in young and older adults are presented in Ta ble 3-2. Direct co mparisons between young and older adults’ encoding-related activation le vels are presented in Table 3-3. Figures F1 and F-2 of Appendix F show the 3-dimensiona l spatial extent of ac tivation in older and young adults, as well as activation di fferences between the two groups. A priori findings The effects of memory encoding were pronounced and statistically significant in bot h older and young adults in the prefrontal cortex. Encoding was associated with several prefrontal foci in both groups. Much of this activity was in the left middle frontal gyrus (BA 6, 8, 9, and 10) Left BA 8 of the medial frontal gyrus was activated only in younger adults. Broca’s area was additionally activated in both

PAGE 73

62 groups. Right middle frontal gyrus activation was observed in both groups, although to a much lesser extent. In the BA 6 portion of the right middle frontal gyrus, activation was greater in older adults. In the BA 8 portion of the right middle front al gyrus, activation was greater in young adults. Left anterior cingulate (BA 32) was activated in young adults, but not older adults. A small, bilatera l region of BA 10 in the middle frontal gyrus was also activated only in younger adults. Ba sed on qualitative observations of the final statistical maps, activation was left laterali zed in young adults, and to a lesser extent in older adults. When activation levels betw een young and older adults were compared directly, regions of the left medial frontal gyrus (BA 6, 8, and 9) were greater in young adults than in older adults. Activations were also greater in young adults in bilateral anterior cingulate cortex (BA 32). Right lateralized medial prefrontal gyrus regions (BA 6 and 8) and middle frontal gyrus (BA 6) demo nstrated greater activation in young adults. Temporal lobe involvement was observed in young and older adults extending from BA 20 to BA 21. This activation was bilate ral in BA 20. Activati on in the bilateral parahippocampal gyrus and left hippocampus was found only in older adults. Activation in the left BA 22 region of the superior te mporal gyrus was only observed in the young adults. Overall, greater spatial extent of activation was observed in the temporal lobe regions for the young adults, but signal intens ity-based comparisons did not reveal any significant differences between the two groups. It should be noted that regions located in th e occipital lobe were activated as well in both young and older adults. The encoding period is time-locked to the presentation of the words to be remembered. Consequently, visual stimulation occurred in conjunction with encoding. As was expected, several occi pital regions were activated (precuneus,

PAGE 74

63 fusiform gyrus, and lingual gyrus). These areas will not be a direct focus of evaluation and therefore are not listed in the following tables. A posteriori findings. Bilateral activation of the angular gyrus (BA 39), as well as left lateralized supramarginal gyrus (BA 40), was found in both groups. Activation was also observed in the transverse temporal gyr us (BA 41), also referred to as the primary auditory receiving cortex. Lastly, activa tion was observed in the precentral and postcentral gyri (BA 3 and 4) in older adul ts. When directly comparing young and older adults, greater activity was observed in bilate ral posterior cingulate cortex (BA 31) and bilateral supramarginal gyrus (BA 40) in young adults. Age Effect on Subsequent Memory Activity (Encoding Activity Correlated with Recall Performance) Brain regions demonstrating significant subsequent memory effects in young and older adults are presented in Table 3-4. Direct comparisons between young and older adults on subsequent memory effects are pres ented in Table 3-5. Fi gures F-3 and F-4 of Appendix F show the 3-dimensional spatia l extent of activat ion in older and young adults, as well as activation diffe rences between the two groups. A priori findings The effects of memory encoding were pronounced and statistically significant for both older and young adults in the prefrontal cortex. Subsequent memory was associated with severa l prefrontal foci in both groups, including the middle frontal gyrus (BA 6, 8, and 9; see Fi gure 3-4). Left anteri or cingulate (BA 32) and right middle frontal gyr us (BA 9) were activated only in young adults. Right supplementary motor area (BA 6) was activated only in older adults. When activation in young and older adults was directly compare d, greater activation in young adults was observed in the medial frontal gyrus (BA 6, b ilaterally; left BA 8 and 9) and anterior

PAGE 75

64 Table 3-2. Comparison of Activation during Encoding in Young and Older Adult Groups Not Receiving Goal-setting Region (BA) Young Older Talairach (x,y,z) Voxels Talairach (x,y,z) Voxels Left Postcentral gyrus (3) -55,-10,46 112 Precentral gyrus (4) -55,-10,35 126 -52,-7,45 257 Middle frontal gyrus (6) -29,5,43 1220 -48,0,38 2278 Medial frontal gyrus (8) -9,30,42 477 Middle frontal gyrus (8) -49,12,40 236 Middle frontal gyrus (9) -48,10,31 734 -49,10,33 973 Middle frontal gyrus (10) -36,47,12 101 Middle temporal gyrus (20) -36,-36,-15207 -35,-36,-15 292 Middle temporal gyrus (21) -60,-27,0 887 -64,-26,0 113 Superior temporal gyrus (22) -59,-33,5 1165 Anterior cingulate cortex (32) -12,19,42 268 Parahippocampal gyrus (36) -34,-29,-16 230 Angular gyrus (39) -41,-61,37 562 -37,-60,37 324 Supramarginal gyrus (40) -47,-59,43 114 -43,-47,42 1054 Inferior frontal gyrus (44) -52,15,9 92 Broca’s area (45) -48,22,5 205 -42,28,16 1623 Hippocampus -25,-21,-4 165 Right Middle frontal gyrus (6) 40,-4,33 257 Middle frontal gyrus (8) 34,24,34 166 Middle temporal gyrus (20) 31,-36,-15 233 32,-37,-15 250 Parahippocampal gyrus (36) 29,-31,-13 431 Angular gyrus (39) 33,-60,38 204 31,-59,36 126 Transverse temporal gyrus (41) 41,-31,9 94 BA = Brodmann’s Area. Talairach = 3-dimensional coo rdinates for the center of gravity in each activation cluster given the stereotactic space of Talairach and Tournoux (1988). Voxel = number of voxels in each cluster exceeding height threshold P < 0.01, Bonferroni corrected in the whole brain volume; exceeding minimum threshold of 80 contiguous voxels.

PAGE 76

65 Table 3-3. Encoding Related Activity Di fferences between Young and Older Adult Groups Not Receiving Goal-setting Region (BA) Young Older Talairach (x,y,z) Voxels Direction of Effect Left Medial frontal gyrus (6) -6,24,46 570 Y > O Medial frontal gyrus (8) -6,32,43 818 Y > O Medial frontal gyrus (9) -6,43,28 830 Y > O Posterior cingulate cortex (31) -7,-63,27 518 Y > O Anterior cingulate (32) -6,29,31 609 Y > O Supramarginal gyrus (40) -58,-46,26 119 Y > O Right Middle frontal gyrus (6) 21,17,47 811 Y > O Medial frontal gyrus (8) 14,28,45 1011 Y > O Medial frontal gyrus (9) 10,41,32 736 Y > O Posterior cingulate cortex (31) 1,-59,27 697 Y > O Anterior cingulate cortex (32) 2,30,29 347 Y > O Supramarginal gyrus (40) 53,-29,33 117 Y > O BA = Brodmann’s Area. Talairach = 3-dimensional coordinate s for the center of gravity in each activation cluster given the stereot actic space of Talairach and Tournoux (1988). Voxel = number of voxels in each cl uster exceeding height threshold P < 0.01, Bonferroni corrected in the whole brain volume; exceeding minimum threshold of 80 contiguous voxels. Y = young adul ts; O = older adults. cingulate (left BA 24 and bilateral BA 32). Fi gure 3-5 illustrates a PFC cluster including portions of the anterior cingulate cortex and medial frontal gyrus that exhibited significant task-related changes in signal intens ity that were correlated with subsequent performance. Extensive temporal lobe involvement, wh ich was mostly left lateralized, was observed in young and older adults. Bilate ral activations were observed in the parahippocampal gyrus (BA 36) in both gr oups. Bilateral activat ion in the middle temporal lobe was observed in BA 20 for bot h groups. Right BA 21 and 22 of the middle temporal gyrus were activated only in young adults. Hippocampal activation was bilateral in young and older adults. The spatial extent of activation was some what greater in young

PAGE 77

66 Table 3-4. Comparison of Subsequent Memory Effect in Young and Older Adult Groups Not Receiving Goal-setting Region (BA) Young Older Talairach (x,y,z) Voxels Talairach (x,y,z) Voxels Left Postcentral gyrus (3) -55,-10,46 84 Precentral gyrus (4) -51,-6,47 167 Middle frontal gyrus (6) -37,4,41 1576 -50,0,41 1438 Middle frontal gyrus (8) -26,21,41 480 -50,13,40 160 Middle frontal gyrus (9) -44,9,35 486 -54,9,34 251 Middle temporal gyrus (20) -35,-36,-15 250 -36,-36,-15 238 Middle temporal gyrus (21) -59,-29,0 573 Superior temporal gyrus (22) -58,-34,4 501 Anterior cingulate cortex (32) -11,19,42 264 Parahippocampal gyrus (36) -37,-31,-13 78 -35,-28,-17 150 Angular gyrus (39) -38,-61,38 391 -36,-60,37 237 Supramarginal gyrus (40) -45,-58,43 102 -42,-48,42 948 Hippocampus -26,-12,-9 1770 -27,-23,-4 422 Right Supplementary motor areas (6) 40,-2,35 156 Middle frontal gyrus (9) 37,24,33 117 Middle temporal gyrus (20) 32,-36,-15 237 32,-37,-15 245 Parahippocampal gyrus (35) 25,-25,-17 104 Parahippocampal gyrus (36) 33,-31,-13 133 28,-31,-13 455 Angular gyrus (39) 35,-61,39 243 32,-60,38 115 Hippocampus 21,-12,-6 1003 BA = Brodmann’s Area. Talairach = 3-dimensional coo rdinates for the center of gravity in each activation cluster given the stereotactic space of Talairach and To urnoux (1988). Voxels = nu mber of voxels in each cluster exceeding height threshold P < 0.01, Bonferroni corrected in the whole brain volume; exceeding minimum threshold of 80 contiguous voxels.

PAGE 78

67 Table 3-5. Subsequent Memory Differences between Young and Older Adult Groups Not Receiving Goal-setting Region (BA) Young Older Talairach (x,y,z) Voxels Direction of Effect Left Medial frontal gyrus (6) -8,29,36 118 Y > O Medial frontal gyrus (8) -7,29,42 321 Y > O Medial frontal gyrus (9) -8,30,32 84 Y > O Anterior cingulate cortex (24) -7,23,25 113 Y > O Posterior cingulate cortex (31)-4,-66,29 89 Y > O Anterior cingulate cortex (32) -7,26,31 838 Y > O Supramarginal gyrus (39) -52,-60,23 83 Y > O Supramarginal gyrus (40) -58,-46,27 85 Y > O Putamen -20,7,5 219 Y > O Right Middle frontal gyrus (6) 27,10,45 180 Y > O Posterior cingulate cortex (31) 2,-56,28 293 Y > O Anterior cingulate cortex (32) 1,26,29 137 Y > O BA = Brodmann’s Area. Talairach = 3-dimensional coo rdinates for the center of gravity in each activation cluster given the stereotactic space of Talairach and To urnoux (1988). Voxels = nu mber of voxels in each cluster exceeding height threshold P < 0.01, Bonferroni corrected in the whole brain volume; exceeding minimum threshold of 80 contiguous voxels. Y = young adults; O = older adults. adults in temporal regions, includ ing the hippocampus (see Figure 3-4). Several occipital regions we re additionally activated (precuneus, fusiform gyrus, and lingual gyrus). A posteriori findings. Activation of the precentral (BA 4) and postcentral (BA 3) gyri was found only in older adults. Bila teral angular gyrus (BA 39) and left supramarginal gyrus activity (BA 40) wa s found in young and older adults. When activation in young and older adu lts was compared directly, bilateral posterior cingulate (BA 31), left supramarginal gyrus (BA 39 a nd 40), and left putamen were found to be greater in young adults. Goal-setting Effect on Encoding Activity In order to evaluate the effect of goa l-setting on encoding, young and older adult

PAGE 79

68 Young Adults Older Adults Middle and Medial Frontal Gyri R L y=9 R L y=9 Hippocampus y=-12 y=-23 Figure 3-4. Effect of subsequent memory in young (n=10) and older (n=10) adults. Coronal slice images illustrate regions in the medial and middle frontal gyri and hippocampus that exhibited signifi cant task-related changes in signal intensity that were correlated with subsequent performance ( P <0.01, Bonferroni corrected; minimum of 80 contiguous voxels). Activation statistical maps are displayed on the smoothed, averaged T1 image for all subjects in the analysis.

PAGE 80

69 Young – Older R L y=26 -0.15 -0.1 -0.05 0 0.05 0.1 0.15 1234567891011121314151617VolumesZ-Transformed Si gnal Intensit y Presentation of Word-list Block Inter-block Interval Younger Adults Older Adults Figure 3-5. Effect of subsequent memory: ol der adults subtracted from young adults. Coronal slice image illustrates PFC cluster (extending from the medial frontal gyrus to the anterior cingulate co rtex) that exhibited significant group differences in signal intensity that were correlated with subsequent performance ( P <0.01, Bonferroni corrected; minimum of 80 contiguous voxels). Activation statistical maps ar e displayed on the smoothed, averaged T1 image for all subjects in the analys is. Task-related z-transformed signal intensity as a function of scan-in-trial is displayed in the graphical illustration. Data were obtained from the contig uous cluster of 7 voxels that had the highest t -values of the overall cluste r (x = -6, y = 26, z = 29).

PAGE 81

70 groups were collapsed across goal (n = 20) and no-goal (n = 20) groups. Brain regions demonstrating significant activations for en coding in the goal and no-goal groups are presented in Table 3-6. Data reflecting signi ficant differences between groups are not presented in tabular format as there wa s only one region demonstrating a significant difference. Figures F-5 and F6 of Appendix F show the 3-di mensional spatial extent of activation in the goal and no-goal groups, as well as ac tivation differences between the two groups. A priori findings In the individual group analys es, activation was observed in bilateral dlPFC regions in the goal and nogoal groups. Specifically, bilateral activation was found in the BA 9 region of the middle fr ontal gyrus, and left BA 46 activation was found in the middle frontal gyrus. Orbitofrontal cortex activation was observed in the left BA 10 region of the middle frontal gyrus and left BA 47 regions of the inferior frontal gyrus in both groups. However, only in th e goal group was right BA 10 activation found. Broca’s area (BA 45), left BA 8 of the mi ddle frontal gyrus, and left supplementary motor area (BA 6) include other frontal regi ons activated in both groups. Right BA 6 was additionally activated in the goal group. The left BA 32 regi on of the anterior cingulate was activated in both groups, whereas activatio n in the right anteri or cingulate (BA 24 and 32) was observed only in the goal gr oup. No significant differences were found between the goal and no-goal groups in any frontal lobe regions. Bilateral temporal lobe involvement was observed in the hippocampus, parahippocampal gyrus (BA 36), and BA 20 of the middle temporal gyrus in both groups. Greater spatial extent of the hippocampal activation was observe d in the goal group; however, differences in signal intensity in th is region were not signi ficant. Activation in

PAGE 82

71 the left insula (BA 13), left superior temporal gyrus (BA 22), and left middle temporal gyrus (BA 21) was observed in both groups as well. When goal and no-goal groups were compared directly, BA 20 activation of th e middle temporal gyrus was found to be greater in the goal group. Several occipital regions we re additionally activated (precuneus, fusiform gyrus, and lingual gyrus). A posteriori findings. Bilateral angular gyrus (BA 39) and supramarginal gyrus (BA 40) activity was observed in both groups. Bilateral posterior ci ngulate cortex (BA 23 and 31) was activated in the goal group, but not in the no-goa l group. Left precentral (BA 4) and postcentral (BA 3) gyri were activ ated in both groups. Activation was also observed in the right transverse tem poral gyrus (BA 41) in the no-goal group. Goal-setting Effect on Subsequent Memory Activity (Encoding Activity Correlated with Recall Performance) Brain regions demonstrating significant subsequent memory activations are presented in Table 3-7. Direct comparisons between goal and no-goal groups’ subsequent memory activation level differences are pres ented in Table 3-8. Fi gures F-7 and F-8 of Appendix F show the 3-dimensional spatial exte nt of activation in the goal and no-goal groups, as well as activation diffe rences between the two groups. A priori findings In the individual group analys es, activation was observed in bilateral dlPFC regions in goal and no-goa l groups. Specifically, there was bilateral activation in the BA 9 regi on of the middle frontal gyrus Only in the goal group was activation in the left BA 46 region of the middle frontal gyrus found. In the goal group, orbitofrontal cortex activation was observed in the left BA 10 region of the middle frontal gyrus and in the left BA 47 regions of the inferior frontal gyrus. Broca’s area (BA 45),

PAGE 83

72 Table 3-6. Comparison of Activation during Encoding in the Goal and No-goal Groups Region (BA) Goal No-goal Talairach (x,y,z) Voxels Talairach (x,y,z) Voxels Left Frontal Lobe Precental gyrus (4) -52,-9,42 736 -54,-9,39 679 Supplementary motor area (6) -36,0,45 3684 -45,0,40 3155 Middle frontal gyrus (8) -39,16,40 333 -41,16,41 317 Middle frontal gyrus (9) -48,13,32 1820 -48,11,32 1671 Middle frontal gyrus (10) -38,45,13 854 -39,45,12 341 Broca’s area (45) -50,22,14 390 -44,24,6 203 Middle frontal gyrus (46) -47,34,16 1059 -45,32,16 432 Anterior cingulate cortex (32) -13,17,41 368 -13,13,45 141 Inferior frontal gyrus (47) -36,27,-1 107 -42,25,0 201 Temporal Lobe Insula (13) -37,12,10 393 -37,22,7 374 Middle temporal gyrus (20) -35,-35,-15 317 -35,-36,-15 288 Middle temporal gyrus (21) -60,-26,0 514 -60,-27,0 764 Superior temporal gyrus (22) -60,-30,5 766 -59,-31,4 801 Parahippocampal gyrus (36) -36,-29,-14 313 -35,-29,-16 193 Hippocampus -35,-21,-9 1991 -25,-22,-5 525 Parietal Lobe Supramarginal gyrus (40) -44,-49,44 1389 -44,-51,43 853 Posterior cingulate cortex (23) -9,-72,12 144 Angular gyrus (39) -41,-61,33 976 -40,-61,37 601 Postcentral gyrus (3) -55,-11,41 266 -56,-11,39 253 Right Frontal Lobe Supplementary motor area (6) 41,-5,34 583 Middle frontal gyrus (9) 37,24,33 680 35,24,33 287 Middle frontal gyrus (10) 30,47,9 150 Anterior cingulate cortex (24) 14,2,38 116 Anterior cingulate cortex (32) 11,17,35 471 Temporal Lobe Middle temporal gyrus (20) 32,-37,-15 250 32,-36,-15 254 Transverse temporal gyrus (41) 42,-29,8 104 Parahippocampal gyrus (36) 31,-30,-14 342 30,-30,-14 427 Hippocampus 28,-23,-4 1100 23,-19,-4 515 Parietal Lobe Angular gyrus (39) 33,-59,36 231 33,-60,37 275 Supramarginal gyrus (40) 39,-47,42 1017 35,-51,43 132 Posterior cingulate cortex (31) 16,-58,24 177 Posterior cingulate cortex (23) 7,-71,12 191 BA = Brodmann’s Area. Talairach = 3-dimensional coo rdinates for the center of gravity in each activation cluster given the stereotactic space of Talairach and To urnoux (1988). Voxels = nu mber of voxels in each cluster exceeding height threshold P < 0.01, Bonferroni corrected in the whole brain volume; exceeding minimum threshold of 80 contiguous voxels.

PAGE 84

73 left BA 8 of the middle frontal cortex, and left supplementary motor area (BA 6) include other frontal regions activated in both gr oups. Right BA 6 and Broca’s homologue (right BA 45) were additionally activated in th e goal group. The left BA 32 region of the anterior cingulate was ac tivated in both groups. When the goal and no-goal groups were comp ared directly, there was significantly greater activity in the bilate ral orbitofrontal cortex (B A 10; x=-32, y=49, z=16), dlPFC (BA 46; x =-47, y=28, z=19), and Broca’s ar ea (BA 45; x=-50, y=19, z=18) in the goalgroup (see Figure 3-6). Bilateral temporal lobe involvement was observed in the hippocampus, parahippocampal gyrus (BA 36), and BA 20 of the middle temporal gyrus in both groups. Activation in the left insula (BA 13), left superior temporal gyrus (BA 22), and left middle temporal gyrus (BA 21) was observed in both groups as well. When goal and nogoal groups were compared directly, BA 20 of the middle temporal gyrus was found to be greater in the no-goal group b ilaterally and activation in the right hippocampus was greater in the goal group. Several occipital regions we re additionally activated (precuneus, fusiform gyrus, and lingual gyrus). A posteriori findings. Bilateral angular gyrus (BA 39) and supramarginal gyrus (BA 40) activity was observed in both groups Left posterior cingulate cortex (BA 30) was activated in the goal group, but not in the no-goal group. Left precentral and postcentral gyri were activated in both groups. Activation wa s also observed in the right transverse temporal gyrus (BA 41). Amygda lar activation was observed in the left

PAGE 85

74 hemisphere of the goal group only. There were no significant differen ces between groups in any of these regions. Interaction of Goal-setting a nd Age during Encoding Activity Brain regions demonstrating significant in teractions during en coding activation are presented in Table 3-9 with parameter es timates listed for each group. Figure F-9 of Appendix F shows the 3-dimensional spatial extent of activation in regions that demonstrated a significant ag e by goal-setting interaction. Regions that were identified as having a significant interaction demonstrated, without exception, the same pattern of relative increases or decreases within age groups. More specifically, regions with a significant interaction showed decreases in activity in young adults as a function of goal-setting, whereas activation in older adults increased as a function of goal-setting. The observed patt ern of activation differences between young and older adults in response to goal-setting indicated that older a dults activated these regions to a greater extent dur ing goal-setting relative to acti vity observed in the no-goal group. Beta weights provided in Ta ble 3-9 are products of the multiple-regression analysis performed on the interaction of age by goal-se tting. Beta weights are interpreted only to the extent that they show relative increases or decreases in encoding-related activity within an age group as a functi on of goal-setting. For regions demonstrating a significant interaction, relative increases or decreases in encoding-related activ ity as a consequence of goal-setting in one age group are then comp ared to the pattern observed in the other age group. A priori findings In frontal regions, activation was observed in bilateral medial frontal gyrus (BA 9; see Figur e 3-7), BA 8 of the left me dial frontal gyrus, and left

PAGE 86

75 Table 3-7. Comparison of Subsequent Memo ry Effect in Goal and No-goal Groups Region (BA) Goal No-Goal Talairach (x,y,z) Voxels Talairach (x,y,z) Voxels Left Frontal Lobe Precentral gyrus (4) -51,-8,44 264 -51,-7,46 272 Supplementary motor area (6) -40,0,39 2059 -45,0,40 2932 Middle frontal gyrus (8) -33,19,40 413 -47,14,41 282 Middle frontal gyrus (9) -47,13,32 1859 -48,9,33 1031 Middle frontal gyrus (10) -36,48,10 1181 Anterior cingulate cortex (32) -13,18,41 368 -13,13,45 130 Broca’s area (45) -37,29,11 9181 -27,32,2 370 Middle frontal gyrus (46) -47,34,16 1045 Inferior frontal gyrus (47) -39,27,-3 129 Temporal Lobe Insula (13) -37,23,7 258 Middle temporal gyrus (20) -36,-34,-15 324 -35,-36,-15 290 Middle temporal gyrus (21) -62,-25,0 212 -60,-28,0 363 Superior temporal gyrus (22) -63,-28,3 145 -58,-30,3 133 Parahippocampal gyrus (36) -36,-28,-15 273 -36,-29,-15 276 Hippocampus -24,-28,0 1141 -29,-14,-9 2315 Parietal Lobe Angular gyrus (39) -40,-61,35 753 -39,-61,37 504 Supramarginal gyrus (40) -43,-52,43 585 -43,-49,43 1119 Postcentral gyrus (3) -55,-10,46 112 Right Frontal Lobe Supplementary motor area (6) 39,-5,34 342 Middle frontal gyrus (9) 39,19,34 609 38,25,34 249 Broca’s homologue (45) 25,31,8 3632 22,33,2 197 Temporal Lobe Middle temporal gyrus (20) 32,-37,-15 252 32,-36,-15 254 Parahippocampal gyrus (36) 31,-30,-14 306 29,-30,-13 537 Parahippocampal gyrus (35) 24,-25,-15 86 Amygdala 29,-4,-17 398 Hippocampus 23,-25,-2 710 25,-16,-8 820 Parietal Lobe Angular gyrus (39) 32,-59,36 220 35,-61,37 378 Supramarginal gyrus (40) 38,-46,40 615 36,-47,44 515 Posterior cingulate cortex (30)22,-67,11 112 BA = Brodmann’s Area. Talairach = 3-dimensional coo rdinates for the center of gravity in each activation cluster given the stereotactic space of Talairach and To urnoux (1988). Voxels = nu mber of voxels in each cluster exceeding height threshold P < 0.01, Bonferroni corrected in the whole brain volume; exceeding minimum threshold of 80 contiguous voxels.

PAGE 87

76 Table 3-8. Subsequent Memory Effect Diffe rences between Goal and No-goal Groups Region (BA) Goal – No-goal Talairach (x,y,z) Voxels Direction of Effect Left Supplementary motor area (6) -41,4,52 210 NG > G Middle frontal gyrus (10) -32,49,16 826 G > NG Middle temporal gyrus (20) -25,-84,-12117 NG > G Broca’s area (45) -39,31,7 1680 G > NG Inferior frontal gyrus (46) -45,40,11 153 G > NG Right Middle frontal gyrus (10) 36,45,20 103 G > NG Middle temporal gyrus (20) 27,-79,-10 150 NG > G Hippocampus 38,25,8 507 G > NG BA = Brodmann’s Area. Talairach = 3-dimensional coo rdinates for the center of gravity in each activation cluster given the stereotactic space of Talairach and To urnoux (1988). Voxels = nu mber of voxels in each cluster exceeding height threshold P < 0.01, Bonferroni corrected in the whole brain volume; exceeding minimum threshold of 80 contiguous voxels. G = goal group, NG = no-goal group

PAGE 88

77 (a) BA 10 Goal – No-goal R L y=49 -0.2 -0.15 -0.1 -0.05 0 0.05 0.1 0.15 0.2 1234567891011121314151617VolumesZ-Transformed Signal Intensity Presentation of Word-list Block Inter-block Interval Goal No-goal (b) BA 45 y=19 -0.15 -0.1 -0.05 0 0.05 0.1 0.15 1234567891011121314151617VolumesZ-Transformed Si gnal Intensit y Presentation of Word-list Block Inter-block Interval Goal No-goal (c) BA 46 y=28 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 1234567891011121314151617VolumesZ-Transformed Si gnal Intensit y Presentation of Word-list Block Inter-block Interval Goal No-goal Figure 3-6. Effect of subsequent memory: nogoal group subtracted from the goal group. Coronal slice images illustrate frontal cortex clusters that exhibited significant activation ( P <0.01, Bonferroni corrected; mini mum of 80 contiguous voxels). Clusters depicted are (a) orbitofronta l cortex (BA 10; x=-32, y=49, z=16); (b) Broca’s area (BA 45; x=-50, y=19, z=18) ; and (c) dorsolate ral prefrontal cortex (BA 46; x =-47, y=28, z=19). Activat ion statistical maps are displayed on the smoothed averaged T1 image for all subjects in the analysis. Taskrelated z-transformed signal intensity as a function of scan-intrial is displayed in the graphical illustrations. Data we re obtained from the contiguous cluster of 7 voxels that had the highest tvalues of the overall cluster.

PAGE 89

78 supplementary motor area (BA 6). Activation in the temporal lobe was limited to the left superior temporal gyrus (BA 22). Significant interactions were not observed in the OFC or dlPFC. Age x Goal-setting Interaction R L y=44 -0.4 -0.3 -0.2 -0.1 0 0.1 0.2Parameter EstimatesYoung Adults/ Goal Young Adults/ No-Goal Older Adults/ Goal Older Adults/ No-Goal Figure 3-7. Coronal slice image illustrates PFC cluster (BA 9) that exhibited significant task-related interaction of age by goal-setting ( P <0.01; Bonferroni corrected, minimum of 80 contiguous voxels). Activ ation statistical maps are displayed on the smoothed, averaged T1 image for all subjects in the analysis. Parameter estimates (e.g., beta weights) for each group are displayed in the graphical illustration. Data were obtained from th e contiguous cluster of 7 voxels that had the highest t -values of the overall cluste r (x=-5, y=44, z=29). Standard error bars are shown. A posteriori findings. An interaction effect was f ound in bilateral supramarginal gyrus (BA 40), left angular gyrus (BA 39), a nd left posterior cingul ate cortex (BA 31). Interaction of Goal-setting and Age for Subsequent Memory Activity (Encoding Activity Correlated with Recall Performance) Brain regions demonstrating significant subsequent memory interactions are presented in Table 3-10 with parameter es timates listed for each group. Figure F-10 of Appendix F shows the 3-dimensional spatial extent of activation in regions that demonstrated a significant age by goal-setting interaction. Beta wei ghts are again used for interpretation of regions that demonstrate a significant interaction.

PAGE 90

79 Table 3-9. Interaction of Age and Goal-setting during Memory Encoding Parameter Estimates (beta weights) Region (BA) Talairach (x,y,z) Voxels YA/GYA/NG OA/G OA/NG Left Supplementary motor area (6) -6,33,39 169 -0.09 0.14 0.00 -0.26 Medial frontal gyrus (8) -8,31,41 277 0.04 0.29 0.10 -0.13 Medial frontal gyrus (9) -5,44,29 348 -0.21 0.08 -0.12 -0.31 Superior temporal gyrus (22) -55,-50,17178 -0.28 0.08 0.04 -0.04 Posterior cingulate cortex (31) -6,-55,25 196 -0.18 0.10 -0.06 -0.18 Angular gyrus (39) -50,-61,18 477 -0.15 0.11 0.12 -0.15 Supramarginal gyrus (40) -59,-40,26 485 -0.25 -0.04 0.020 -0.29 Right Medial frontal gyrus (9) 2,43,31 175 -0.24 0.03 -0.16 -0.34 Supramarginal gyrus (40) 52,-29,33 242 -0.28 -0.15 -0.09 -0.46 BA = Brodmann’s Area. Talairach = 3-dimensional coo rdinates for the center of gravity in each activation cluster given the stereotactic space of Talairach and To urnoux (1988). Voxels = nu mber of voxels in each cluster exceeding height threshold P < 0.01, Bonferroni corrected in the whole brain volume; exceeding minimum threshold of 80 contiguous voxels. YA = young adults, OA = older adult, G = goal group, NG = no-goal group. A priori findings. Frontal regions that demonstr ated a significant interaction included Broca’s homologue (BA 45) and the ante rior cingulate cortex (BA 31). The left superior temporal gyrus (B A 22) was also activated. A posteriori findings. The left angular gyrus (BA 39) demonstrated a significant interaction.

PAGE 91

80 Table 3-10. Interaction of Age and Goal-set ting during Encoding for Subsequent Memory Effects Parameter Estimates (beta weights) Region (BA) Talairach (x,y,z) Voxels YA/GYA/NG OA/G OA/NG Left Superior temporal gyrus (22) -52,-56,16 83 -0.34 0.18 0.09 -0.26 Posterior cingulate cortex (31) -5,-12,45 167 -0.37 -0.12 0.01 -0.44 Angular gyrus (39) -50,-60,19 260 -0.30 .14 .14 -0.33 Right Broca’s homologue (45) 45,34,3 207 -0.21 -0.07 0.09 -0.33 BA = Brodmann’s Area. Talairach = 3-dimensional coo rdinates for the center of gravity in each activation cluster given the stereotactic space of Talairach and To urnoux (1988). Voxels = nu mber of voxels in each cluster exceeding height threshold P < 0.01, Bonferroni corrected in the whole brain volume; exceeding minimum threshold of 80 contiguous voxels. YA = young adults, OA = older adult, G = goal group, NG = no-goal group.

PAGE 92

81 CHAPTER 4 DISCUSSION Despite their high levels of self-reported general hea lth and greater vocabulary skills, older subjects exhibited a pattern of episodic memory perfor mance typical of that reported in behavioral studies of aging (L ight, 1991). In partic ular, older subjects’ performances were below those of young s ubjects on a word list recall task. The provision of goals for performing the task resu lted in a significant increase in the number of words recalled in both young and older adults. Further, the increase in performance as a result of goal-setting was equivalent between young and older adults. These findings are consistent with other studi es that utilized an analogous paradigm (West et al., 2002; 2003). Encoding of the word list was associated w ith highly left-lateralized activity in the prefrontal cortex (PFC) of young adults. PFC activity in older adults was dampened compared to young adults. When correlating r ecall performance with signal intensity (e.g., subsequent memory effects), bilateral temporal lobe and hippocampal activity was also observed in young and older adults. Volume distribution of the activation was larger in young adults; however, a signalintensity based comparison di d not reveal a significant difference. The effect of goal-setting on subsequent memory activation was only found in the frontal lobes. Specifically, regi ons that demonstrated higher levels of activity from goalsetting included the orbitofrontal cortex (OFC), dorsolateral prefrontal cortex (dlPFC), as well as Broca’s area. A signifi cant increase in the right hippoc ampal region as a result of

PAGE 93

82 goal-setting was observed as well. When co mparing young and older adults’ response to goal-setting, older adults were found to recruit several diffe rent regions to a greater extent than young adults. Encoding in Young Adults A leading model describing the cortical neural substrates of encoding is the hemispheric encoding/retrieval asymmetry (HERA) model (Tulving, 1994). The HERA model proposes that the left PFC is more involved than the right PFC in encoding, whereas the right PFC is more involved than the left PFC in episodic memory retrieval. However, more recent studies of spatial (as opposed to verbal) material find greater rightlateralized PFC activity during encoding, presen ting a formidable challenge to the HERA model. Thus, Tulving and colleagues (Habib et al., 2003) recently asserted a revised formulation of the HERA model in order to accommodate these c onflicting results. They assert that encoding and retrieval processes must be systematically varied in their interaction with hemispheric in teractions. Specifical ly, to test for asymmetry in encoding, the following formulation must be met: left hemisphere encoding minus left hemisphere retrieval is greater than right hemisphere encoding minus right hemisphere retrieval. Findings from the present study are generally consistent with the HERA model as strongly left-lateralized PFC activation was observed during encoding. However, the design of the present study doe s not allow for evaluation of retrieval-related neural activation. Consequently, findings only part ially address the HERA model as an evaluation of the present data set with the more stringent reformulation of the HERA model is not possible. The present study found less than robust te mporal lobe activat ion in the encodingrelated activity analysis. Extensive neuropsyc hological and physiologi cal evidence points

PAGE 94

83 to a clear role of temporal lobe involvement in encodi ng; however, many other fMRI studies exist that found little or no encoding-related medial temporal lobe activation (Petersen et al., 1988; Frith et al., 1991; Demonet et al., 1992; Grasby et al., 1993, 1994; Kapur et al., 1994, 1996; Raichle et al., 1994; Sh allice et al., 1994; Tu lving et al., 1994; Fletcher et al., 1995; Nyberg et al., 1996). In addition to the dubious finding of little or no involvement of the medial temporal lobe in encoding, this approach – measuring encoding-related activity – is questionable to the extent th at encoding activities are genuinely being assessed. For instance, enc oding related activity in the hippocampus could reflect a response to novelty unrelated to the encoding of a memory, or it could reflect habituation or reduced attention. To address these importan t issues, the present study also employed an approach that correlated subsequent memory performance with signal intensity during encoding. In this way, one can be more confident that a particular region of activation signifies processes impor tant for encoding an episodic memory. When using the correlative approach between memory performance and signal intensity, robust activation was found in the temporal lobe/hippocampus regions, which is consistent with other studies that employed th is approach (Brewer et al., 1998; Gabrieli et al., 1997; Fernandez et al., 1998, 1999). Activation was observed bilaterally in th e angular gyrus and the supramarginal gyrus, which likely reflects reading processes taking place during the word list presentation (Price, 2000). Ac tivation also occurred in th e transverse temporal gyrus, which may be accounted for by the loud scanner environment. Encoding in Older Adults and Age-related Changes The predicted reduction in left PFC ac tivity observed during encoding for older adults compared with young adults is in line with the HAROLD (h emispheric asymmetry

PAGE 95

84 reduction in older adults) model. The HAROLD model states that young subjects, consistent with the HERA model, engage th e left PFC more extensively during encoding, whereas older subjects show dampened left PFC activity during encoding (Cabeza et al., 2001). Even though the present study, among ma ny others (see Grady, 2000 for review), showed decreased left PFC activation in olde r adults during encoding, it is worthwhile noting two recent studies that showed a simila r level of activation in the left PFC in young and older adults. Daselaar et al. (2003) compared brain activ ity patterns obtained during incidental encoding in wh ich subjects were not asked to learn the words presented, but instead were asked to make pleasant/ unpleasant judgments about the words. They found equivalent left PFC enc oding activation in young subjects and older subjects when correlating encoding activity w ith subsequent memory of the words. Morcum et al. (2003) also found equivalent le ft PFC activation when subjects were asked to make animacy decisions about words. Subjects late r underwent a recognition memory test for these words that was correlated w ith activity during the encoding period. A common thread can be found between thes e two studies that reported equivalent left PFC activation between young and older adults during encoding. Both of these studies involved incidental lear ning, as opposed to the intenti onal learning approach that was utilized by other studies, including the present one. This account for the differential findings is further supported by a study by Buckner and colleagues (Logan et al., 2002), which showed less prefrontal activity in olde r adults compared w ith young adults under intentional learning instructions. Importantl y, this difference was not observed when a semantic-orienting task was used to support episodic encoding (e.g., incidental encoding). This interpretation of the seemingly disparate findings in the left PFC is in agreement

PAGE 96

85 with a production deficiency account of ag e-related impairments in episodic encoding. This account asserts that olde r adults do not employ seman tic elaboration strategies spontaneously but are able to make use of th em when forced to do so (Burke and Light, 1981). The production deficiency account is part icularly interesti ng in light of the differential use of strategi es between young and older adu lts observed in the present study. Older adults utilized less strategies in performing the intentional memory encoding task. The decreased use of strategies is cons istent with the produc tion deficiency account and is likely reflected by the decreased le ft PFC activation observed in older adults. Greater activation was found in the anterior and posterior porti ons of the cingulate cortex in young adults. A significant scan by ag e interaction existed in a cluster of voxels that included regions of the an terior cingulate and me dial frontal gyrus. The greater level of sustained activity in the young adults suggests greater atte ntional resources dedicated to the task and perhaps more concentrated performance monitoring (Tisserand and Jolles, 2003). In regards to the increa sed level of activa tion in the posterio r cingulate in young adults, several studies have asserted a role for this region bein g involved in encoding (Hunkin et al., 2002; Hofer et al., 2003). However, Otte n and Rugg (2001) found that encoding activation in the pos terior cingulate wa s associated with recall failure on subsequent memory tests. At present, the ro le of the posterior cingulate in encoding is unclear and is need of further elucidation. Older adults demonstrated encoding activity in the temporal lobe (primarily in the middle temporal gyrus and hippocampus), which was to a lesser spatia l extent than young adults. There were no significant signal in tensity differences be tween young and older adults, though significant activation present in BA 21 of the middle temporal gyrus and

PAGE 97

86 right hippocampus in the young adults was absent in older adults. It has been suggested that the medial temporal lobe operates by forming associations between sensory, cognitive and emotional processes that make up an episode in memory (Alvarez and Squire, 1994; Eichembaum, 1996). Accordingly, it has been suggested that there is a relation between the amount of medial temporal lobe activity and the number of associations that are formed during encodi ng of study material (Henke et al., 1997, 1999). Given the presence of medial temporal lobe activity during encodi ng, this would imply that the older adults formed fewer memory associations. In addition to the PFC and temporal lobes, older adults engaged other regions that were in common with young adults. Activati on was observed in the angular gyrus and supramarginal gyrus regions in the encodingrelated and subsequent memory analyses. This activation, as previously stated, was likely associated with reading processes required for list learning (Price, 2002). Taken together, data from this pilot study suggest that older subjects engage much of the same neural circuitry as young subj ects when encoding new memories. However, the findings also point to th e possible presence of agerelated differences in both prefrontal and temporal activ ity during episodic encoding. Furt her, these findings suggest that the age-related declines in episodic encoding may be related to strategic encoding and attentional differences, as well as fewer memory associations formed. Goal-setting Influence on Encoding Predicted associations were found betw een OFC and dlPFC activation and the presence of goal-setting during encoding. Additional regions of the prefrontal cortex were also found to be involved with subseque nt memory performance, including the supplementary motor cortex and a large portion of Broca’s area.

PAGE 98

87 In further evaluating areas engaged by goa l-setting, it is impossible to separate which regions underlie the motivation prope rties of goal-setting and which regions underlie increased mnemonic processing, faci litated by increased motivation. However, comparing the results to the extent literatu re provides a likely framework in which the goal-setting effect is mediated. The generally agreed upon role of the OFC in motiv ation and the pattern of activation observed in the OFC suggest that this region supports the motivational component of goal-setting. In the subsequent memory effect analysis of goal-setting, activation in the OFC was not observed in th e no-goal group, indicati ng that this region was likely not associated with successful or unsuccessful encoding. Consequently, it is unlikely that the engagement of the OFC di rectly reflects mnemonic processes. It is likely, however, that the OFC is associated w ith the motivational aspect of goal-directed improvement of the performance, which is s upported by several lines of evidence. Stuss and Levine (2002) put forward the concept of the self-regulatory disorder, which is characterized by an inability to regulate behavior according to internal goals and constraints. They observed this disorder af ter discrete lesioning to the ventral frontal cortex. Humans with ventral frontal lobe da mage can show impairments in a number of tasks in which an alteration of behavioral stra tegy is required in re sponse to a change in environmental context or expectations (D amasio, 1994; Rolls, 2000). Activity in this region in the present study likely reflects th e other end of the spectrum from what these studies report: increased activity in the OFC mediates increases in goaldirected/contextually-guided behavior.

PAGE 99

88 A more difficult region to elucidate the cont ribution to the goal-setting effect is the dlPFC. Extensive work demonstrates the role of the dlPFC in cognitive control (see Miller and Cohen, 2002 for review). Cognitive c ontrol includes the ability to maintain context, or a goal, by which behavior is then biased. Certainly it is plausible that the increased dlPFC activation subserved the func tion of maintaining the idea of the goal online while performing the task. However, being that the dlPFC subserved mnemonic processes as well in the absence of goal-setti ng, it is difficult to ascertain whether this region helped to improve memory performance by enhanced mnemonic processing or maintenance of goal intent, or a combination of these factors. Goal-setting also produced substantially greater activation in Broca’s area and supplementary motor areas. As activation in motor regions and Broca’s area has been implicated in subvocalization (Sweet et al., 2004; Gruber, 200 1), these regions could be involved in an increase in subvocal rehearsal in the goal group as a result of increased motivation. This finding suggests that the goal group may have gained in performance over the no-goal group by greater subvocal rehearsal during the encoding period. Interestingly, none of the frontal regions obs erved to be signifi cantly greater in the goal-setting group in the subsequent memory analysis were significant in the encodingrelated activity analysis. Changes that o ccur during goal-setting are therefore not reflected by general activity during encoding, but more likely reflect neural processes directly related to successf ul or unsuccessful encoding. Differential Neural Response to Goal -setting in Young and Older Adults Older adults did not show the predicted di fferential activation of the OFC or dlPFC in response to goal-setting, as compared to young adults. However, older adults recruited several other regions to a gr eater extent than young adults to achieve the performance

PAGE 100

89 enhancing effects of goal-sett ing. Interestingly, there were no regions that exhibited significantly greater activity in young a dults in conjunction with goal-setting, as compared to older adults. This result is consistent with th e hypothesis that older adults would compensate for the agerelated atrophy in the OFC and dlPFC. Each of the regions demonstrating an age by goal-setting interact ion was previously shown to be involved in mnemonic processes. Therefore, the overall pa ttern of differentially greater activation in older adults may reflect greater increases in mnemonic processing that are precipitated by increases in motivation. However, it is also possible that greater activation in these regions in older adults reflects less effi cient recruitment of neural resources. In the encoding-related activity analysis, increases in activity in the prefrontal, temporal and posterior cingulate regions pe rhaps reflect increased resources committed toward memory encoding processes. Activat ion in the supplementary motor area could reflect increased subvocal reh earsal of the word list by ol der adults. The differential increased activity in older a dults in the angular gyrus a nd the supramarginal gyrus may reflect an increased level of focus on read ing the words as they are presented. Again, these increases in activati on in the above mentioned re gions could also represent inefficient recruitment of neur al resources in older adults. It should be noted that activation in all the previously mentioned regions, with the exception of the temporal lobe and angular gyrus, does not reach significance in the subsequent memory analysis. This may indicate that even though the regions are recruited more heavily as compared to young a dults, only activations in the angular gyrus and superior temporal gyrus have a genuine impact on subsequent recall performance. Additionally, results from the subsequent memo ry analysis revealed that older adults

PAGE 101

90 recruit the anterior cingul ate and Broca’s homologue to a greater extent than young adults, perhaps reflecting increased attention. The suggestion that the older adults had greater increases in use of mnemonic processes than the young adults as a conseque nce of goal-setting is consistent with De Jong’s goal neglect hypothesis. It has been shown that incr easing the challenge to older adults can compensate for difficulties main taining the context of a cognitive task (e.g. compensate for neglect of salient processes necessary for successful task performance; De Jong, 2001). In the present study, it a ppears as though the increased challenge provided by goal-setting enabled older adults to generate di sproportionately greater focus towards important mnemonic processes for the episodic memory task, which helped them improve their performance levels to that of the young adults not re ceiving goal-setting. Potential Limitations Group membership in the pres ent study was limited to 10 su bjects, which gives rise to restricted power. In order to address this power consideration, a fixed-effects analysis approach was utilized for both fMRI and behavioral data. Thus, error variance was estimated on a data point by data point ba sis for each subject with each data point representing an independent observation. In contrast to a random-effects analysis in which generalizations can be inferred about the population(s) studi ed, generalizations from the present research must be considered more cautiously. It is also important to highlight that the goal-setting condition was accompanied by performance feedback, as well as encourag ing statements for achieving the goal. The feedback and encouragement was not cont rolled for in the no-goal group, thus constraining conclusions regard ing goal-setting to this context. The ultimate aim of this research is to provide a better understanding of the positive effect s of the goal-setting

PAGE 102

91 paradigm utilized in the previously menti oned studies by West et al. (2002, 2003) such that it might be utilized in an applied setting. To this end, goal theory asserts that knowledge of progress toward a goal improves th e efficacy of the goal. In regards to the positive encouragement provided, West and colleagues (Dark-Freidman et al., 2004) recently found no memory recall performance di fferences between individuals receiving positive encouragement and progress feedback with goal-setting and individuals who receive “realistic” fee dback in conjunction with goal -setting (e.g., subjects are told whether or not they met their goal and are not given encouragement before goals). Thus, it is unlikely that the encouraging remarks significantly influenced performance outcome. Another consideration of the present findi ngs worth elucidating is that groups, whether it be the goal and no-goal groups or the young and older adults, demonstrated differential behavioral performance on the memo ry task that subsequently could result in different cognitive operations being measured. For instance, the comparison of older adults and young adults could be measur ing differences between poor encoding and successful encoding, as opposed to age-rela ted changes in encoding activity. This potential confound was addressed, in part, by the subsequent memory analyses performed that correlated recall performan ce with signal intensity. Even though this analysis takes into consideration changes in signal intensit y relative to differential performance, it still does not account for the overall poorer perfor mance in older adults and no-goal groups. Follow-up studies would benefit from a design that allows for only correct responses to be subjected to fMRI analysis. In such a cas e, direct comparisons can be made between groups relative to the cognitive process of successful encoding.

PAGE 103

92 Findings of the present research should also be considered in the context of important factors related to the use of fMRI in studying aging. The older adults’ hemodynamic response to several sensorimotor tasks shows that the neural circuitry of healthy older adults may produce greater noise while keeping the signal at the same level as young adults (Raz, 2000). Consequently, a lo wer signal-to-noise ra tio is observed in the older cohort, and fewer (by a factor of four) pixels on their fMRI images pass the threshold for activation (D'Esposito et al., 1999). When performing a group analysis, an activation focus emerges to the extent that each member of the group exhibits greater activation for one conditi on relative to another in a partic ular brain region. In theory, a group of older subjects could have equal spa tial extent and intensity of activation as young subjects, but if the older participants were more variable from one to another in the locations of activation, activa tion in significantly fewer voxels would surpass a statistical threshold. In this case, the group activation would be incorrectly characterized, since greater variability across participants in activation loci would appear as reduced activation (Stebbins et al., 2002). A further methodological consideration, as previously discussed, is that of differential neurovascular coupling between young and older adults. FMRI does not measure brain activity directly, but cha nges in blood oxygen level dependent (BOLD) signal that are strongly associat ed with neural activity (Logot hetis, 2001). As this is the case, it gives rise to the po ssibility that differences be tween young and older adults may not be due to differences in neural activation, but to secondary factor s associated with the BOLD response. The BOLD signal is a reflect ion of interactions between cerebral blood flow, cerebral blood volume, blood oxygen extrac tion, and local metabolism that occur as

PAGE 104

93 a consequence of neural activ ity. The possibility therefore exists that age-associated changes observed in this research could be in fluenced by age-related changes in vascular processes (D'Esposito et al., 1999, 2003). It is noteworthy that the regions of greatest interest in the present research, the PFC a nd OFC, experience the steepest trajectory of atrophy of all brain regions in aging (R az, 2000) and consequently may have a heightened susceptibility to neurovascular changes as well. These concerns regarding greater intra-gr oup variability in the older adults and differential neurovascular coupling between young and older adults were addressed in part by use of the internal activation standard. Activation occurring in response to the motor task was equivalent between young and ol der adults. Demonstration of equivalent activation helps to vitiate these concerns; how ever, it is conceivable that differential neurovascular coupling between young and ol der adults exists throughout the brain. Nonetheless, a clear absence of activati on differences between young and older adults during the motor task (see Figure 3-1) supports the assertion that the findings of the present research do not repr esent neurovascular coupling differences or intra-group variability differences betw een young and older adults. Steps were taken in the data analysis of the present research as well to minimize the concern of intra-group variability and differe ntial neurovascular coupling. An important comparison was that of possible differentia l responses to goal-se tting between young and older adults. Instead of testing for direct group differences in young and older adults who received goals (e.g., young adult/goal subtract ed by older adult/goal), a safer approach was used testing for an age by goal-setti ng interaction. That is differences in the

PAGE 105

94 relative activation between no-goa l and goal groups within e ach age group was evaluated rather than a direct comparison between the two groups. Concluding Remarks The present research findings are consistent with the left lateralized PFC activation proposed by the HERA model of episodic memory encoding. These findings are also consistent with the proposition of the HAROLD model that this lateralized PFC activation during encoding is suppressed in olde r adults. Decreased left PFC activation in older adults was accompanied by a poorer performance on the episodic memory task. Previous findings of the improvement of memory performance by goal-setting in young and older adults were replic ated here. The use of explicit goals to improve memory performance was, as hypothesized, associated with the dlPFC and the OFC. However, especially in the case of the dlPFC, it remains somewhat unclear as to the extent these regions contribute to increased motivati on and/or increased mnemonic processing resulting from increased motivation. Dissoci ation of these roles may be achieved in future studies by parametrically manipul ating the level of motivation, perhaps by differential monetary rewards, and parametric ally manipulating the level of challenge in mnemonic processing. Older adults showed a pattern of differen tial activation in re sponse to goal-setting, as compared to young adults, that entailed gr eater recruitment of several regions during encoding. The pattern of activ ation suggests possible cogn itive processes that are producing the beneficial effect s of goal-setting. For instance, activation in the anterior cingulate cortex likely unde rlies greater attentional resources and performance monitoring committed to the list learning ta sk. Activation of Broca’s homologue could reflect greater use of strategies. These possibil ities could be capitali zed upon to refine and

PAGE 106

95 improve behavioral interventions that improve memory. For instance, goal-setting can be combined with attention focusing training and sp ecific strategies, such as rehearsal, to optimize the efficacy of memory training. In addition, regions disproportionately activated in older adults duri ng goal-setting, as well as region s that demonstrated reduced activation during encoding, are potentially pr omising targets for possible pharmaceutical interventions in normal and clinic ally significant memory declines.

PAGE 107

96 APPENDIX A SHIPLEY VOCABULARY TEST VOCABULARY On this questionnaire, the fi rst word in each line is pr inted in capital letters. Opposite it are four other words. Draw a line under the one word which means the same thing or most nearly the same th ing, as the first word. A sample has been worked out for you, with the right answer underlined. If you don't know, guess Be sure to underline the one word in each line which means the same thing as the first word. sample: LARGE red big silent wet 1. TALK draw eat speak sleep 2. PERMIT allow sew cut drive 3. PARDON forgive pound divide tell 4. COUCH pin eraser sofa glass 5. REMEMBER swim recall number defy 6. TUMBLE drink dress fall think 7. HIDEOUS silvery tilted young dreadful 8. CORDIAL swift muddy leafy hearty 9. EVIDENT green obvious skeptical afraid 10. IMPOSTOR conductor officer book pretender 11. MERIT deserve distrust fight separate 12. FASCINATE welcome fix stir enchant 13. INDICATE defy excite signify bicker 14. IGNORANT red sharp uninformed precise 15. FORTIFY submerge strengthen vent deaden

PAGE 108

97 16. RENOWN length head fame loyalty 17. NARRATE yield buy associate tell 18. MASSIVE bright large speedy low 19. HILARITY laughter speed grace malice 20. SMIRCHED stolen pointed remade soiled 21. SQUANDER tease belittle cut waste 22. CAPTION drum ballast heading ape 23. FACILITATE help turn strip bewilder 24. JOCOSE humorous paltry fervid plain 25. APPRISE reduce strew inform delight 26. RUE eat lament dominate cure 27. DENIZEN senator inhabitant fish atom 28. DIVEST dispossess intrude rally pledge 29. AMULET charm orphan dingo pond 30. INEXORABLE untidy in volatile rigid sparse 31. SERRATED dried notched armed blunt 32. LISSOM moldy loose supple convex 33. MOLLIFY mitigate direct pertain abuse 34. PLAGIARIZE appropriate intend revoke maintain 35. ORIFICE brush hole building lute 36. QUERULOUS maniacal curi ous devout complaining 37. PARIAH outcast priest lentil locker 38. ABET waken ensue incite placate 39. TEMERITY rashness timidity desire kindness 40. PRISTINE vain sound first level

PAGE 109

98 APPENDIX B WORD LISTS FOR MEMORY TRIALS Baseline Trial Tomatoes Cauliflower Broccoli Cucumbers Roast Ribs Bacon Ham Mayonnaise Antacid Detergent Hammer Jelly Beans Thread Magazine Trials 1, 2 and 3 Block 1 Block 2 Block 3 Block 4 Block 5 Block 6 Block 7 Tomatoes Antacid Thread Rolls Broccoli Pears Foil Bacon Detergent Cauliflower Envelope Salt Ham Cologne Mayonnaise Hammer Magazine Popcorn Cola Mop Pineapple Pens Scissors Roast Oranges Bananas Peanuts Beer Cucumbers Jelly Beans Eraser Shampoo Ribs Chili Macaroni Iron Cheese Dog Food Teaspoon Razor Conditioner Scarf

PAGE 110

99 APPENDIX C QUESTIONNAIRES MEMORY QUESTIONNAIRE DIRECTIONS: Different people use their memory in different ways in their everyday lives. For example, some people make shopping lists, whereas others do not. So me people are good at remembering names, whereas others are not. In this questionnaire, we would like you to tell us how you use your memory and how you feel about it. There are no right or wron g answers to these questions because people are different. Please take your time and answer each of these questions to th e best of your ability. Each question is followed by five choices. Draw a circle around the number corresponding to your choice. Mark only one number for each statement. For example: My memory will get worse as 1. agree strongly I get older. 2. agree 3. undecided 4. disagree 5. disagree strongly In this example you could, of course, choose any one of the answers. If you agree strongly with the statement you would circle the “1.” If you disagree st rongly you would circle the “5.” The answers of “2” and “4” indicate less strong agreem ent or disagreement. The “3” gives you a middle choice, but don’t use the “3” unless you really can’t decide on any of the other responses. Keep these points in mind : Answer every question, even if it doesn’t seem to apply to you very well. Answer as honestly as you can what is true for you Please do not mark something because it seems like the “right thing to say.” Agree Agree Undecided Disagree Disagree Strongly Strongly 1. It is important to me to 1 2 3 4 5 have a good memory. 2. I have little control over 1 2 3 4 5 my memory ability. 3. I think a good memory is 1 2 3 4 5

PAGE 111

100 something of which to be proud. 4. I find it harder to 1 2 3 4 5 remember things when I am upset. 5. I get anxious when I am 1 2 3 4 5 asked to remember something. 6. It bothers me when others 1 2 3 4 5 notice my memory failures. 7. My friends often notice my 1 2 3 4 5 memory ability. 8. I am usually uneasy when I 1 2 3 4 5 attempt a problem that requires me to use my memory. 9. I know if I keep using my 1 2 3 4 5 memory I will never lose it. 10. Having a better memory 1 2 3 4 5 would be nice but it is not very important. 11. It doesn’t bother me when 1 2 3 4 5 my memory fails. 12. I can’t expect to be good 1 2 3 4 5 at remembering zip codes at my age. 13. I get upset when I cannot 1 2 3 4 5 remember something. 14. I think it is important to 1 2 3 4 5 work at sustaining my memory abilities. 15. I work hard at trying to 1 2 3 4 5 improve my memory. 16. If I am put on the spot to 1 2 3 4 5 remember names, I know I will have difficulty doing it. 17. I admire people who have 1 2 3 4 5 good memories. 18. I have difficulty remembering 1 2 3 4 5 things when I am anxious. 19. I would feel on edge right 1 2 3 4 5 now if I had to take a memory test or something similar. 20. I often notice my friends’ 1 2 3 4 5 memory ability.

PAGE 112

101 21. As long as I exercise my 1 2 3 4 5 memory it will not decline. 22. I feel jittery if I have to 1 2 3 4 5 introduce someone I just met. 23. It’s important that I am very 1 2 3 4 5 accurate when remembering names of people. 24. When I am tense and uneasy 1 2 3 4 5 at a social gathering, I cannot remember names very well. 25. I’m highly motivated to 1 2 3 4 5 remember new things I learn. 26. It’s up to me to keep my 1 2 3 4 5 remembering abilities from deteriorating. 27. When someone I don’t know 1 2 3 4 5 very well asks me to remember something, I get nervous. 28. Even if I work on it, my memory 1 2 3 4 5 ability will go downhill. 29. I get anxious when I have 1 2 3 4 5 to do something I haven’t done for a long time. 30. It bothers me when I forget 1 2 3 4 5 an appointment. 31. I like to remember things 1 2 3 4 5 on my own, without relying on other people to remind me. 32. I get tense and anxious 1 2 3 4 5 when I feel my memory is not as good as other people’s. 33. It’s important that I am very 1 2 3 4 5 accurate when remembering significant dates. 34. I do not get flustered 1 2 3 4 5 when I am put on the spot to remember new things. 35. I would feel very anxious 1 2 3 4 5 if I visited a new place and had to remember how to find my way back. 36. No matter how hard a 1 2 3 4 5 person works on his memory, it cannot be

PAGE 113

102 improved very much. 37. If I were to work on 1 2 3 4 5 my memory I could improve it. 38. It gives me great satisfaction 1 2 3 4 5 to remember things I thought I had forgotten. 39. I think a good memory 1 2 3 4 5 comes mostly from working at it.

PAGE 114

103 QUESTIONS ABOUT RECENT MEMORY PERFORMANCE On this page, there are some questions asking for your opinions. To answer each question, you should circle the number that best indicates your opinion. Please read each question carefully before you decide how to answer. There are no right or wrong answers on these questions. -----------------------------------------------------------------------------------------------------------------------First, do the EXAMPLE: Please give us your opinion about the weather. What do you think about the weather outside today? 1 2 3 4 5 6 7 Wonderful Horrible -----------------------------------------------------------------------------------------------------------------------------QUESTIONS ABOUT RECENT MEMORY PERFORMANCE 1. How important has it been to you to perform well on memory activities in your everyday life? 1 2 3 4 5 6 7 Not at all Very Important Important 2. How have you performed on most me mory tasks you have done recently? 1 2 3 4 5 6 7 Very Poor Very Good 3. How do you think your memory compares with most other people your age? 1 2 3 4 5 6 7 Much worse Much better 4. How satisfied are you with your recent memory performance? 1 2 3 4 5 6 7 Very Very unsatisfied satisfied

PAGE 115

104 DIRECTIONS: A number of statements which people hav e used to describe themselves are given below. Read each statement and then circle the appropriate number to the right of the statement to indicate how you feel right now that is, at this moment There are no right or wrong answers. Do not spend too much time on any one statement but give the answer which seems to describe your present feelings best. 1. I feel calm 2. I feel secure 3. I am tense 4. I feel strained 5. I feel at ease 6. I feel upset 7. I am presently worrying over possible misfortunes 8. I feel satisfied 9. I feel frightened 10. I feel comfortable 11. I feel self-confident 12. I feel nervous 13. I am jittery 14. I feel indecisive 15. I am relaxed 16. I feel content 17. I am worried 18. I feel confused 19. I feel steady 20. I feel pleasant

PAGE 116

105 PARTICIPANT INFORMATION Please give us the following information about yourself. 1. Your date of birt h: __________ _________________ 2. Sex (circle one): male female 3. Marital status (circle one): single married separated divorced widowed 4. Race: ________ _______________ _____________ 5. Current job status (circle all that a pply): work full-time work part-time retired student Occupation and/or job position that you held for the longest time (please describe): 6. Total years of education, beginning with Grade 1: _____________ _____________ Highest educational de gree or diploma that you hold: _____________ ____________ 7. Health: Circle one number below to indicate how he althy you are in general: 1 2 3 4 5 6 7 8 9 10 Excellent Very poor health health 8. Have you been hospitalized in the last 5 years? (circle one) YES NO If YES, please describe the reason(s) for each hospitalization (if needed, you can use more space on the back of this page): 9. Please list any medications you take regularly. If you do not take any medications, write “none.” If you do take something regularly, please give the names of the medications and describe yo ur reason for taking each one. If you do not know the names, just list the reason. Fo r example -"I take 2 pills each day for high blood pressure." (if needed, you can use more space on the back of this page)

PAGE 117

106 GOAL REVIEW SHEET The experimenter gave you a goal. Please indicate how you feel about this goal. ------------------------------------------------------------------------------------------------------------------1. I am committed to achieving this goal. 1 2 3 4 5 6 7 Strongly agree Strongly disagree 2. It is realistic to expect me to reach this goal. 1 2 3 4 5 6 7 Yes, I can No, I will never reach it easily reach it 3. The goal for me should be changed. 1 2 3 4 5 6 7 Strongly agree Strongly disagree 4. I will make an effort to reach this goal. 1 2 3 4 5 6 7 No, I won’t Yes, I will make any put forth effort maximum effort 5. It is important to me to achieve this goal. 1 2 3 4 5 6 7 Strongly agree Strongly disagree 6. I will never be able to reach this goal. 1 2 3 4 5 6 7 Strongly agree Strongly disagree 7. This goal is personally meaningful to me. 1 2 3 4 5 6 7 Strongly agree Strongly disagree 8. I am motivated to reach the goal. 1 2 3 4 5 6 7 Yes, highly Not at all motivated motivated

PAGE 118

107 QUESTIONS Please circle the answer that best repr esents your opinion for each question. --------------------------------------------------------------------------------------------------------------1. If you had an opportunity to work on another memory lis t today, to try to improve your performance, would you want to do it? 1 2 3 4 5 6 7 Definitely Definitely would try would not try 2. How difficult were the memory tasks that you completed today? 1 2 3 4 5 6 7 Very Very difficult easy 3. How much effort did you make on the memory tasks today? 1 2 3 4 5 6 7 Very high No effort effort at all 4. Did you achieve your goals today? 1 2 3 4 5 6 7 Definitely Definitely not 5. How satisfied are you with your memory performance today? 1 2 3 4 5 6 7 Highly Highly satisfied dissatisfied

PAGE 119

108 APPENDIX D STRATEGY USE QUESTIONNAIRE Techniques or Strategies for Recalling a Shopping List Some people are able to use special techniques to help them to remember. Here is a long list of many memory techniques that could be used. Y ou may have concentrated on the words while you were studying, and did not do anything else. Or you may have tried many different strategies to try to improve your score. Either way is fine. We are interested only in finding out exactly what you did while you were studying. Please place a checkmark by all of the methods that you used while you were studying the shopping lists. -----------------------------------------------------------------------------------------------------------------------1. _____ I concentrated and paid attention to each word. 2. _____ I thought about how I might make a meal out of some items (e.g., “eggs cereal orange juice are my breakfast”). 3. _____ I repeated single words over and over to myself (e.g., “peas ,” “peas ,” “peas ,” “peas ”). 4. _____ I repeated groups or sets of words over and over to myself (e.g., “pears pizza bags ,” “pears pizza bags ,” “pears pizza bags ”). 5. _____ I grouped items together in my mi nd into categories (e.g., meats, beverages, and fruits were grouped). 6. _____ I put together items in my mind that began with the same letter of the alphabet. 7. _____ I put together items that a person might use together in daily life (e.g., use a hammer, pencil, and nail to mount a picture frame in a certain place). 8. _____ I made up sentences or stories to connect the items (e.g., “the candy apple was on a plate pierced by a fork .” 9. _____ In my mind, I pictured each individual item.

PAGE 120

109 10. _____ In my mind, I pictured sets of items together (e.g., sugar in a bowl with a spoon). 11. _____ In my mind, I pictured items inte racting in an active video (e.g., the coffee is being poured into the mug with milk and a chocolate drop). 12. _____ I connected the first letters of th e items (e.g., “HOT reminds me of Honey Onion T-shirt ”). 13. _____ I thought about where the items are located in the grocery store where I usually shop. 14. _____ I looked away from the list and t ested myself to see how many items I could recall. 15. _____ Other method: please describe Now go back through and review th e list of things you checked. Circle the 1 or 2 methods that you used the most often for remembering.

PAGE 121

110 APPENDIX E TISSUE-AIR INTERFACE SIGNAL DROP OFF Regions that border tissueair interfaces, such as the orbitofrontal cortex, experience field inhomogeneities that result in signal loss and mism apping artifacts. The orbitofrontal cortex borders the orbito sinus and the auditory meatus, creating susceptibility artifacts at the tissue-air inte rfaces. Figure F-1 shows the degree of signal loss in the echo planar (functional) scans as compared to the high resolution (MPRAGE) structural image, both in the same 3D orientation in an example subject. High Resolution Structural Image Echo-planar (functional) Image Figure E-1. Illustration of signal drop off in portions of the or bitofrontal cortex in a single subject. The white highlighted area depicts Brodmann’s area 10 and 47 for purposes of anatomical reference and image comparison. Considerable signal loss is observed in ventral and anterior portions of the orbitofrontal cortex.

PAGE 122

111 APPENDIX F GLASS BRAIN ILLUSTRATIONS DESCRIPTION OF GLASS BRAIN ILLUSTRATIONS “Glass-brain” illustrations of the primary statistical analyses visualize activated regions in 3-dimentional space. Glass brain de pictions are constructed using a wire mesh outline of the brain and red-green stereo colori ng of activated regions in efforts to better visualize activated regions in 3-dimentional space. Directionality of effect is not represented in the figures. Presentation of the glass brain illustrations follow the order in which the statistical analyses were presented in the Results chapter: 1) effect of age on encoding, 2) effect of goal-setting on enc oding, and 3) the interaction of age and goalsetting in encoding. Statistical thre shold for glass brain figures are P < 0.01, Bonferroni corrected, and 80 contiguous voxels, unless otherwise noted.

PAGE 123

112 Age Effects on Encoding-related Activity Young Adults Older Adults Figure F-1. Glass-brain repres entation of regional activatio ns during encoding in older and young adults. Coronal Sa g ittal Axial Coronal Sa g ittal Axial

PAGE 124

113 Young Adults – Older Adults Figure F-2. Glass-brain repres entation of the regions show ing significantly different activation between young and ol der adults during encoding. Sa g ittal Coronal Axial

PAGE 125

114 Age Effects on Subsequent Memory Young Adults Older Adults Figure F-3. Glass-brain repres entation of encoding activity correlated with subsequent recall performance in young and older adults. Sagittal Axial Coronal Axial Sagittal Coronal

PAGE 126

115 Young Adults – Older Adults Figure F-4. Glass-brain repr esentation showing significan t differences between young and older adults for encoding activity correlated with subsequent recall performance. Sa g ittal Coronal Axial

PAGE 127

116 Goal-setting Effects on Encoding Goal No-goal Figure F-5. Glass-brain represen tation of regional activations during encoding in goal and no-goal groups ( P < 0.001, corrected; minimum threshold of 250 contiguous voxels). Due to the extensive activati on observed in goal and no-goal groups, it was necessary to set the threshol d, for viewing purposes only, higher for these two groups than the st atistical threshold used for all other statistical maps and tables. The extensive activation is a consequence of increased power and use of the fixed-effects analysis. These analyses are conducted with 20 subjects per group, as compared to the age analyses, which contained 10 subjects per group. It should be noted th at the contrast gla ss brain statistical maps for goal and no-goal groups (e. g., goal – no-goal) employs the other statistical threshold ( P < 0.01, corrected; 80 contiguous voxels). Sagittal Sagittal Coronal Coronal Axial Axial

PAGE 128

117 Goal – No-goal Figure F-6. Glass-brain repres entation of the regions show ing significantly different levels of activation between goal and no-goal groups during encoding. Sagittal Coronal Axial

PAGE 129

118 Goals-setting Effects for Subsequent Memory Goal No-goal Figure F-7. Glass-brain repres entation of encoding activity correlated with subsequent recall performance in goa l and no-goal groups ( P < 0.001, Bonferroni corrected; minimum threshold of 250 contiguous voxels). Sagittal Sagittal Coronal Coronal Axial Axial

PAGE 130

119 Goal – No-goal Figure F-8. Glass-brain repres entation showing significant di fferences between goal and no-goal groups for encoding activity correlated with s ubsequent recall performance. Sagittal Coronal Axial

PAGE 131

120 Interaction of Goal-setting and Age Age x Goal-setting Figure F-9. Glass-brain repres entation of regions showi ng a significant interaction between age and goal-setting during encoding. Sagittal Coronal Axial

PAGE 132

121 Interaction of Goal-setting and Age for Subsequent Memory Age x Goal-setting Figure F-10. Glass-brain repr esentation of regions showi ng a significant interaction between age and goal-setting for encodi ng activity that was correlated with subsequent recall performance. Sa g ittal Coronal Axial

PAGE 133

122 LIST OF REFERENCES Arnsten, A. F. (1993). Catecholamine mechanisms in age-related cognitive decline. Neurobiology of Aging, 14 639-641. Arnsten, A. F., Cai, J. X., Murphy, B. L., & Goldman-Rakic, P. S. (1994). Dopamine D1 receptor mechanisms in the cognitive performance of young adults and aged monkeys. Psychopharmacology, 116 143-151. Arnsten, A. F., Cai, J. X., Steere, J. C., & Goldman-Rakic, P. S. (1995). Dopamine D2 receptor mechanisms contribute to age-re lated cognitive decline: The effects of quinpirole on memory and motor function in monkeys. Journal of Neuroscience, 15 3429-3439. Bacciottini, L., Passani, M. B., Mannaioni, P. F., & Blandina, P. (2001). Interactions between histaminergic and cholinergi c systems in learning and memory. Behavioral Brain Research, 124 (2), 183-194. Backman, L., & Molander, B. (1986). Effects of adult age and level of skill on the ability to cope with high-stress condi tions in a precision sport. Psychology and Aging, 1 334-336. Baddeley, A., Lewis, V., Eldridge, M., & T homson, N. (1984). Attention and retrieval from long-term memory. Journal of Experime ntal Psychology, 113 518-540. Baltes, P. B. (1993) The aging mind: potential and limits. Gerontologist, 333(5) 580594. Band, G. P. H., Ridderinkhof, K. R., & Sega lowitz, S. (2001). Expl aining neurocognitive aging: Is one factor enough? Brain and Cognition, 49 259-267. Bandura, A. (1989). Self-regul ation of motivation and acti on through internal standards and goal systems. In L. A. Pervin (Ed.), Goal concepts in personality and social psychology (pp. 19-85). Hillsdale, NJ: Erlbaum. Bandura, A., & Cervone, D. (1983). Self-e valuative and self-efficacy mechanisms governing the motivational as pects of goal systems. Journal of Personality and Social Psychology, 45 1017-1028. Bao, S., Chang, E., Davis, J., Gobeske, K., & Merzenich, M. (2003). Progressive degradation and subsequent refinement of acoustic representations in the adult auditory cortex. The Journal of Neuroscience, 23 10765-10775.

PAGE 134

123 Barrash, J., Tranel, D., & Anderson, S. W. (2000). Acquired persona lity disturbances associated with bilater damage to the ventromedial prefrontal cortex. Developmental Neuropsychology, 18 (3), 355-381. Baskerville, K., Schweitzer, J., & Herrron, P. (1997). Effects of c holinergic depletion on experience-dependent plasticity in the cortex of the rat. Neuroscience, 80 (4), 1159-1169. Baskin, J., & Weinberger, N. (1996). Induction of a physiological memory in the cerebral cortex by stimulation of the nucleus basalis. Proceedings of the National Academy of Sciences USA, 93 (20), 11219-11224. Bechara, A., Damasio, H., & Damasio, A. R. (2000). Emotion, decision making and the orbitofrontal cortex. Cerebral Cortex, 10 295-307. Brandt, J., Spencer, M., & Folstein, M. ( 1988). The Telephone Inte rview for Cognitive Status. Neuropsychiatry, Neuropsychology and Behavioral Neurology, 1 (2), 111117. Braver, T. S., Barch, D. M., Keys, B. A., Cohe n, J. D., Taylor, S. F., Carter, C. S., Kaye, J. A., Janowsky, J. S., Yesavage, J. A., Mumenthaler, M. S., Jajust, W. J., & Reed, B. R. (2001). Context processing in older adults: evidence for a theory relating cognitive control to neurobiology in healthy aging. Journal of Experimental Psychol ogy: General, 130 (4), 746-763. Braver, T. S., & Cohen, J. D. (2001). Wo rking memory, cognitive control, and the prefrontal cortex: Compuational and empirical studies. Cognitive Processing, 2 25-55. Brewer, J., Zhao, Z., Glover, G., & Gabrieli J. D. E. (1998). Making memories: Brain activity that predicts whether visual expe riences will be remembered or forgotten. Science, 281 1185-1187. Buckner, R. L., Petersen, S., Ojemann, J., Miezin, F., Squire, L., & Raichle, M. E. (1995). Functional anatomical studies of e xplicit and implicit memory retrieval tasks. Journal of Neuroscience, 15 12-29. Burke, D., & Light, L. (1981). Memory and ag ing: the role of retrieval processes. Psychological Bulletin, 90 513-554. Cabeza, R. (2001). Cognitive neuroscience of aging: contributions of functional neuroimaging. Scandanavian Journal of Psychology, 42 (3), 277-286. Cabeza, R. (2002). Hemispheric asymmetry reduction in older adults: the HAROLD model. Psychology and Aging, 17 (1), 85-100.

PAGE 135

124 Cabeza, R., Grady, C. L., Nyberg, L., McInto sh, A. R., Tulving, E., Kapur, S., Jennings, J. M., Houle, S., & Craik, F. I. M. (1 997). Age-related differences in neural activity during memory en coding and retrieval: A positron emission tomography study. Journal of Neuroscience, 17 391-400. Cabeza, R., & Kingstone, A. (2001). Handbook of functional neuroimaging of cognition Cambridge, MA: The MIT Press. Cabeza, R., McIntosh, A. R., Tulving, E., Nyberg, L., & Grady, C. L. (1997). Age-related differences in effective neural c onnectivity during encoding and recall. Neuroreport, 8 3479-3483. Cabeza, R., & Nyberg, L. (2000). Imaging co gnition II: An empirical review of 275 PET and fMRI studies. Journal of Cognitive Neuroscience, 12 (1), 1-47. Carey, J. R., Kimberley, T. J., Lewis, S., Auerbach, E. J., Dorsey, L., Rundquist, P., & Ugurbil, K. (2002). Analysis of fMRI and finger tracking training in subjects with chronic stroke. Brain: A Journal of Neurology, 125 773-788. Charness, N. (1981). Aging and skilled problem solving. Journal of Experimental Psychology, 110 21-38. Chiodo, L., & Berger, T. (1986). Interactions between dopamine and amino-acid induced excitation and inhibition in the striatum. Brain Research, 375 198-203. Coffey, C. E., Wilkinson, W. E., Parashos I. A., Soady, S. A. R., Sullivan, R. J., Patterson, L. J., Figiel, G. S., Webb, M. C., Spritzer, C. E., & Djang, W. T. (1992). Quantitative cerebral anatomy of the aging human brain: A crosssectional study using magnetic resonance imaging. Neurology, 42 527-536. Cohen, G., & Faulkner, D. (1986). Memory for proper names: age differences in retrieval. British Journal of Developmental Psychology, 4 187-197. Cohen, J. D., Perlstein, W. M., Braver, T. S., Nystrom, L. E., Noll, D. C., Jonides, J., & Smith, E. E. (1997). Temporal dynamics of brain activation during a working memory task. Nature, 386 604-608. Connor, L. T. (2001). Memory in old age: patterns of decline and preseveration. Seminars in Speech and Language, 22 (2), 117-125. Cowell, P., Turetsky, B. I., Gur, R. C., Grossman, R. I., Shtasel, D. L., & Gur, R. E. (1994). Sex differences in aging of th e human frontal and temporal lobes. Journal of Neuroscience, 14 4748-4755.

PAGE 136

125 Craik, F. I. M., & Jennings, J. M. (1992). Hu man memory. In F. I. M. Craik & T. A. Salthouse (Eds.), The handbook of aging and cognition. (pp. 51-110). Hillsdale, NJ: Erlbaum. Damasio, A. R. (1994). Descartes' error. New York: Putnam. Dark-Freudeman, A., West, R ., & Bagwell, D. (2004). Factors that predict age-related performance after goal-setting. Paper presented at the Cognitive Aging Conference, Atlanta, GA. Das, A., Dikshit, M., & Nath, C. (2001). Profile of acetylcholinesterase in brain areas of male and female rats of adult and old age. Life Sciences, 68 (13), 1545-1555. Daselaar, S., Veltman, D., Rombouts, S., Raaijmakers, J., & Jonker, C. (2003). Neuroanatomical correlates of episodic encoding and retrieval in young and elderly subjects. Brain, 126 43-56. De Jong, R. (2000). An intention-activation account of residual sw itch costs. In S. Monsell & J. Driver (Eds.), Attention and perform ance XVIII: Control & cognitive processes. Cambridge, MA: MIT Press. De Jong, R. (2001). Adult age differences in goal activation a nd goal maintenance. European Journal of Cognitive Psychology, 13 71-89. de Keyser, J., De Backer, J. P., Vauquelin, G ., & Ebinger, G. (1990). The effect of aging on the D1 dopamine receptors in human frontal cortex. Brain Research, 528 308310. D'Esposito, M., Deouell, L., & Gazzaley, A. (2003). Alterations in the BOLD fMRI signal with ageing and disease: a challenge for neuroimaging. Nature Reviews, 4 863-872. D'Esposito, M., Zarahn, E., Aguirre, G., & Rypm a, B. (1999). The effect of normal aging on the coupling of neural activty to the bold hemodynamic response. Neuroimage, 10 6-14. Dolan, R. J., & Fletcher, P. C. (1997). Disso ciating prefrontal a nd hippocampal function in episodic memory encoding. Nature, 388 (6642), 582-585. Elliott, E. S., & Dweck, C. S. (1988). Go als: An approach to motivation and achievement. Journal of Personality and Social Psychology, 54 5-12. Fernandez, G., Brewer, J., Zhao, Z., Glover, G., & Gabrieli, J. D. E. (1999). Level of sustained entorhinal activity at study correlates with subsequent cued-recall performance: a functional magnetic resonance imagning study with high acquisition rate. Hippocampus, 9 35-44.

PAGE 137

126 Fernandez, G., Weyerts, H., Schrader-Bolsc he, M., Tendolkar, I., Smid, H., Tempelmann, C., Hindchs, H., Scheich, H., Elger, C., Mangun, G., & Heinze, H. (1998). Successful verbal encoding into epis odic memory engages the posterior hippocampus: a parametrically analyzed functional magnetic resonance imaging study. The Journal of Neuroscience, 18 (5), 1841-1847. Ferrucci, L., Del Lungo, I., Guralnik, J. M., Bandinelli, S., Benvenuti, E., Salani, B., Lamponi, M., Ubezio, C., Benvenuti, F., & Baroni, A. (1998). Is the telephone interview for cognitive status a valid alternative in persons who cannot be evaluated by the Mini Mental State Examination? Aging, 10 (4), 332-338. Fletcher, P. C., Frith, C. D., Grasby, P., Shallic e, T., Frackowiak, R. S. J., & Dolan, R. J. (1995). Brain systems for encoding and retr ieval of auditory-verbal memory: An in vivo study in humans. Brain, 118 401-416. Fletcher, P. C., Shallice, T., & Dolan, R. J. (1998). The functional roles of prefrontal cortex in episodic memory: I. Encoding. Brain, 121 (7), 1239-1248. Forman, S. D., Cohen, J. D., Fitzgerald, M., Eddy, W. F., Mintun, M. A., & Noll, D. C. (1995). Improved assessment of signifi cant activation in functional magnetic resonance imaging (fMRI): Use of a cluster-size threshold. Magnetic Resonance in Medicine, 33 636-647. Friston, K. J., Holmes, A. P., Worsley, K. J., Poline, J. P., Frith, C. D., & Frackowiak, R. S. J. (1995). Statistical parametric maps in functional imaging: A general linear approach. Human Brain Mapping, 2 189-210. Frith, C. D., Friston, K. J., Liddle, P. F., & Frackowiak, R. S. I. (1991). Willed action and the prefrontal cortex in man: a study with PET. Proceedings of the Royal Society of London (Biological Sciences), 244 241-246. Fuster, J. M. (1989). The prefrontal cortex: Anatomy, physiology and neuropsychology of the frontal lobe. New York: Raven Press. Geinisman, Y., deToledo-Morell, L., Morell F., & Heller, R. E. (1995). Hippocampal markers of age-related memory dysfunctio n: Behavioral, electrophysiological and morphological perspectives. Progress in Neurobiology, 45 223-252. Goldman-Rakic, P. S. (1987). Circuitry of pr imate prefrontal cort ex and regulation of behavior by representational memory. In F. Plum & V. Mountcastle (Eds.), Handbook of physiologyThe nervous system. (Vol. 5, pp. 373-417). Bethesda, MD: American Physiological Society. Goldman-Rakic, P. S., & Brown, R. M. (1981). Regional changes of monoamines in cerebral cortex and subcortical stru ctures of aging rhesus monkeys. Neuroscience, 6 177-187.

PAGE 138

127 Grady, C. L. (2000). Functional brain imag ing and age-related changes in cognition. Biological Psychology, 54 (54), 259-281. Grady, C. L., McIntosh, A. R., Horwitz, B., Ma isog, J. M., Ungerleider, L. G., Mentis, M. J., Pietrini, P., Schapiro, M. B., & Haxby, J. V. (1995). Age-related reductions in human recognition memory due to impaired encoding. Science, 269 (5221), 218221. Grady, C. L., McIntosh, A. R., Rajah, M. N., & Craik, F. I. (1998). Neural correlates of the episodic encoding of pictures and words. Proceedings of the National Academy of Sciences USA, 95 (5), 2703-2708. Gruber, O. (2001). Effects of domain-specific interference on brain activation associated with verbal working memory task performance. Cerebral Cortex, 11 (11), 10471055. Halsband, U., Krause, B. J., Schmidt, D., Herzog, H., Tellmann, L., & Muller-Gartner, H.-W. (1998). Encoding and retrieval in declarative learning: A positron emission tomography study. Behavioral Brain Research, 97 69-78. Haug, H., & Eggers, R. (1991). Morphometry of the human cortex cerebri and corpus striatum during aging. Neurobiology of Aging, 12 336-338. Henson, R., Rugg, M., Shallice, T., Josephs, O., & Dolan, R. J. (1999). Recollection and familiarity in recognition memory: an event-related functional magnetic resonance imaging study. Journal of Neuroscience, 19 3962-3972. Hinsz, V., & Ployhart, R. (1998). Trying, inte ntions, and the processes by which goals influence performance: An empirical test of the theory of goal pursuit. Journal of Applied Social Psychology, 28 (12), 1051-1066. Hofer, A., Weiss, E., Golaszewski, S., Sieden topf, C., Brinkhoff, C., Kremser, C., Felber, S., & Fleischahacker, W. (2003). Neural correlates of episodic encoding and recognition of words in unmedicated patients during and acute episode of schizophrenia: a functional MRI study. American Journal of Psychiatry, 160 (1802-1808). Hultsch, D. F., Hertzog, C., & Dixon, R. A. (1987). Memory self-knowledge and selfefficacy in the aged. In N. L. Lowe & C. J. Brainerd (Eds.), Cognitive Development in Adulthood: Progress in Cognitive Developmental Research. (pp. 65-92). New York: Springer-Verlag. Hunkin, N., Mayes, A., Gregory, L., Nicholas A., Nunn, J., Brammer, M., Bullmore, E., & Williams, S. (2002). Novelty-related ac tivation within the medial temporal lobes. Neuropsychologia, 40 1456-1464.

PAGE 139

128 Huttenlocher, P. (1979). Synaptic density in human frontal cortex: Developmental changes and effects of aging. Brain Research, 163 195-205. Iidaka, T., Sadato, N., Yamada, H., Murata, T., Omori, M., & Yonekura, Y. (2001). An fMRI study of the functional neuroanato my of picture encoding in younger and older adults. Cognitive Brain Research, 11 1-11. Jahanshanhi, M., Dirnberger, G., Fuller, R., & Frith, C. D. (1997). The functional anatomy of random number generation studied with PET. Journal of Cerebral Blood Flow and Metabolism, 17 643. Jahanshanhi, M., & Frith, C. D. (1998). Willed action and its impairments. Cognitive Neuropsychology, 15 483-533. Jarvenpaa, T., Rinne, J., Raiha, I., Koskenvuo, M., Lopponen, M., Hinkka, S., & Kaprio, J. (2002). Characteristics of two tele phone screens for cognitive impairment. Dementia and Geriatric Cognitive Disorders, 13 (3), 149-155. Kakiuchi, T., Ohba, H., Nishiyama, S., Sat o, K., Harada, N., Nakanishi, S., & Tsukada, H. (2001). Age-related changes in muscarin ic cholinergic receptors in the living brain: A PET study using N-[11C]methyl-4 -piperidyl benzilate combined with cerebral blood flow measurement in conscious monkeys. Brain Research, 916 22-31. Kapur, S., Tulving, E., Cabeza, R., McIntosh, A. R., Houle, S., & Craik, F. I. M. (1996). The neural correlates of in tentional learning of verbal materials: A PET study in humans. Brain Research. Cognitive Brain Research, 4 (4), 243-249. Kelley, W. M., Miezin, F. M., McDermott, K. B., Buckner, R. L., Raichle, M. E., Cohen, N. J., Ollinger, J. M., Akbudak, E., Conturo, T. E., Snyder, A. Z., & Peterson, S. E. (1998). Hemispheric specialization in hum an dorsal frontal cortex and medial temporal lobe for verbal and nonverbal memory encoding. Neuron, 20 927-936. Kemper, T. L. (1994). Neuroanatomical and neuropathological cha nges during aging and in dementia. In M. L. Albert & E. J. E. Knoepfel (Eds.), Clinical neurology of aging, 2nd Edition (pp. 3-67). New York: Oxford University Press. Kilgard, M., & Merzenich, M. (1998). Cortic al map reorganization enabled by nuclues basalis activity. Science, 279 1714-1718. Kirasic, K. C., Allen, G. L., Dobson, S. H ., & Binder, K. S. (1996). Aging, cognitive resources, and declarative learning. Psychology and Aging, 11 658-670. Kopelman, M. D., Stevens, T. G., Foli, S ., & Grasby, P. (1998). PET activation of the medial-temporal lobe in learning. Brain, 121 (5), 875-887.

PAGE 140

129 Kupfermann, I. (1991). Learning and Memory. In E. Kandel & J. Schwartz & T. Jessell (Eds.), Principles of neural science (3 ed., pp. 997-1007). Norwalk, Connecticut: Appleton & Lange. Lee, K. S., Frey, K. A., Koeppe, R. A., Bu ck, A., Mulholland, G. K., & Kuhl, D. E. (1996). In vivo quantification of cerebral muscarinic receptors in normal human aging using positron emission tomography and [11C]tropanyl benzilate. Journal of Cerebral Blood Flow and Metabolism, 16 303-100. Lee, T. W., Locke, E. A., & Latham, G. P. (1989). Goal setting theory and job performance. In L. A. Pervin (Ed.), Goal concepts in personality and social psychology. (pp. 291-326). Hillsdale, NJ: Erlbaum. Lepage, M., Habib, R., & Tulving, E. (1998) Hippocampal PET activations of memory encoding and retrieva l: The HIPER model. Hippocampus, 8 313-322. Light, L. (1991). Memory and aging: four hypotheses in search of data. Annual Reviews in Psychology, 42 333-376. Linnenbrink, E., Ryan, A., & Pintrich, P. (1999) The role of goals and affect in working memory functioning. Learning and Individual Differences, 11 (2), 213-230. Logan, J., & Buckner, R. (2001). Age-related changes in neural correlated of encoding. Eighth Annual Meeting of th e Cognitive Neuroscience Society Logan, J., Sanders, A., Snyder, A., Morris, J., & Buckner, R. (2002). Under-recruitment and nonselective recruitment: dissociable neural mechanisms associated with aging. Neuron, 33 827-840. Logothetis, N. K., Pauls, J., Augath, M., Trinath, T., & Oeltermann, A. (2001). Neurophysiological investigation of the basis of the fMRI signal. Nature, 412 (6843), 150-157. Luciana, M., Collins, P. F., & Depue, R. A. (1998). Opposing roles for dopamine and serotonin in the modulation of huma n spatial working memory functions. Cerebral Cortex, 8 218-226. Luszcz, M. A., & Bryan, J. (1999). Toward understanding age-related memory loss in late adulthood. Gerontology, 45 2-9. Madden, D. J., Gottlob, L. R., Denny, L. L., Tu rkington, T. G., Provenzale, J. M., Hawk, T. C., & Coleman, R. E. (1999). Aging and recognition memory: changes in regional cerebral blood flow associat ed with components of reaction time distributions. Journal of Cognitive Neuroscience, 11 (5), 511-520. Madden, D. J., Pierce, T. W., & Allen, P. A. (1993). Age-related slowing and the time course of semantic priming in visual word identification. Psychology and Aging, 8 490-507.

PAGE 141

130 Madden, D. J., Turkington, T. G., Coleman, R. E., Provenzale, J. M., DeGrado, T. R., & Hoffman, J. M. (1996). Adult age differe nces in regional cerebral blood flow during visual word identification: Evidence from H2 15O PET. NeuroImage, 3 127-142. Madderlink, G., & Harackiewicz, J. M. (1984) Proximal versus distal goal setting and intrinisic motivation. Journal of Personality and Social Psychology, 47 918-928. Mayes, A., Meudell, P., Mann, D., & Picker ing, A. (1988). Loca tion of lesions in Korsakoff's syndrome: neuropsychological and neuropathological data on two patients. Cortex, 24 367-388. McIntosh, A. R., Bookstein, F. L., Haxby, J. V., & Grady, C. L. (1996). Spatial pattern analysis of functional brain imag es using partial least squares. Neuroimage, 3 143-157. Mercado III, E., Bao, S., Orduna, I., Gluck, M ., & Merzenich, M. (2001). Basal forebrain stimulation changes cortical sensitivities to complex sound. Neuroreport, 12 (10), 2283-2287. Miller, E. K., & Cohen, J. D. (2001). An integr ative theory of prefr ontal cortex function. Annual Reviews in Neuroscience, 24 167-202. Morcom, A., Good, C., Frackowiak, R. S. J., & Rugg, M. (2003). Age effects on the neural correlates of su ccessful memory encoding. Brain, 126 213-229. Morrell, R., Park, D., & Poon, L. (1990). Eff ects of labeling techniques on memory and comprehension of prescription info rmation in young and older adults. Journal of Gerontology, 45 P166-P172. Muir, J. L. (1997). Acetylcholine, aging, and Alzheimer's disease. Pharmacology, Biochemistry, and Behavior, 56 (4), 687-696. Muir, J. L., Page, K., Sirinathsinghji, D., R obbins, T., & Everitt, B. (1993). Excitotoxic lesions of basal forebrain cholinergic neurons: effects of learning, memory and attention. Behavioral Brain Research, 57 785-807. Murphy, D. G. M., DecCarli, C., McIntosh, A. R., Daly, E., Mentis, J. J., Pietrini, P., Szczepanik, J., Schapiro, M. B., Grady, C. L., Horwitz, B., & Rapoport, S. I. (1996). Age-related differences in volumes of subcortical nuclei, brain matter, and cerebro-spinal fluid in healthy men as m easured with magnetic resonance imaging (MRI). Archives of General Psychiatry, 53 585-594.

PAGE 142

131 Nelles, G., Widman, G., de Greiff, A., Me istrowitz, A., Dimitrova, A., Weber, J., Forsting, M., Esser, J., & Diener, H. ( 2002). Brain representation of hemifield stimulation in poststroke visual field defects. Stroke: A Journal of Cerebral Circulation, 33 (5), 1286-1293. Nyberg, L., McIntosh, A. R., Cabeza, R., Habib, R., Houle, S., & Tulving, E. (1996). General and specific brain re gions involved in encoding and retrieval of events: What, where, and when. Proceedings of the National Academy of Sciences USA, 93 11280-11285. Ojemann, J., Akbudak, E., Snyder, A. Z., McKi nstry, R., Raichle, M. E., & Conturo, T. E. (1997). Anatomic localization and quantit ative analysis of gradient refocused echo-planar fMRI susceptability artifacts. Neuroimage, 6 156-167. Penit-Soria, J., Audinat, E., & Crepel, F. (1987). Excitation of rat prefrontal cortical neurons by dopamine: An in v itro electrophysiolgical study. Brain Research, 425 263-274. Perlstein, W., Elbert, T., & Ste nger, V. (2002). Dissociation in human prefrontal cortex of affective influences on working memory-related activity. Proceedings of the National Academy of Sciences, USA, 99 (3), 1736-1741. Pineiro, R., Pendlebury, S., J ohansen, B. H., & Matthews, P. M. (2001). Functional MRI detects posterior shifts in primary sens orimotor cortex activation after stroke: evidence of local adaptive reorganization? Stroke: A Journal of Cerebral Circulation, 32 (5), 1134-1139. Playford, D., Jenkins, H., Passignham, R., Nu tt, J., Frackowiak, R. S. I., & Brooks, D. (1992). Impaired mesial frontal and putame n activation in Parkinson's disease: A PET study. Annals of Neurology, 32 151-161. Price, C. (2000). The anatomy of language : contributions from functional imaging. Journal of Anatomy, 197 335-359. Raz, N. (2000). Aging of the brain and its im pact on cognitive perf ormance: Integration of structural and function findings. In F. I. M. Crai k & T. A. Salthouse (Eds.), The handbook of aging and cognition (2 ed., pp. 1-90). Mahwah, NJ: Lawrence Erlbaum Associates. Raz, N., Gunning, F. M., Head, D., Dupuis, J. H., McQuain, J., Briggs, S. D., Loken, W. J., Thornton, A. E., & Acker, J. D. (1997) Selective aging of the human cerebral cortex observed in vivo: diffe rential vulnerability of th e prefrontal gray matter. Cereb Cortex, 7 (3), 268-282. Read, D. (1987). Neuropsychological asse ssment of memory in the elderly. Canadian Journal of Psychology, 41 158-174.

PAGE 143

132 Ro, T., Henik, A., Machado, L., & Rafal, R. D. (1997). Transcranial magnetic stimulation of the prefrontal cortex delays contralateral endogenous saccades. Journal of Cognitive Neuroscience, 9 (4), 433-440. Rolls, E. (2000). Memory systems in the brain. Annual Reviews in Psychology, 51 599630. Ross, M., Yurgelun-Todd, D., Renshaw, P., Maas, L., Mendelson, J., & Mello, N. (1997). Age-related reduction in functional MR I response to photic stimulation. Neurology, 48 173-176. Rugg, M. (2001). When more means less: neural activity related to unsuccessful memory encoding. Current Biology, 11 (19), 1528-1530. Rugg, M., Otten, L., & Henson, R. (2002). The neural basis of episodic memory: evidence from functional imaging. Philosophical Transactions of the Royal Society of London, 357 1097-1110. Savage, C. R., Deckersbach, T., Heckers, S., Wagner, A. D., Schacter, D. L., Alpert, N. M., Fischman, A. J., & Rauch, S. L. (2001). Prefrontal regions supporting spontaneous and directed application of verbal learning strategies: Evidence from PET. Brain, 124 219-231. Sawaguchi, T., & Goldman-Rakic, P. S. ( 1991). D1 dopamine receptors in prefrontal cortex: Involvement in working memory. Science, 251 947-950. Sawaguchi, T., Matsumara, M., & Kubota, K. (1990). Catecholaminergic effects on neural activity related to a delayed res ponse task in monkey prefrontal cortex. Journal of Neurophysiology, 63 1385-1400. Schacter, D. L., & Wagner, A. D. (1999). Medi al temporal lobe activations in fMRI and PET studies of episodic encoding and retrieval. Hippocampus, 9 (1), 7-24. Schunk, D. H. (1990). Goal setting and self -efficacy during self-regulated learning. Educational Psychologist, 25 71-86. Scoville, W., & Milner, B. (1957). Loss of recent memory after bilateral hippocampal lesions. Journal of Neurology, Neur osurgery, and Psychiatry, 20 11-21. Segal, M. A., & Mandler, G. (1967). Directiona lity and organizational processes in paired associate learning. Journal of Experimental Psychology, 74 305-312. Servan-Schreiber, D., Carter, C. S., Bruno, R., & Cohen, J. D. (1998). Dopamine and the mechanism of cognition: II. D-amphetamine effects in human subjects performing a selective attention task. Biological Psychiatry, 43 723-729.

PAGE 144

133 Shipley, W. (1940). A self-administering scal e for measuring intellectual impairment and deterioration. Journal of Psychology, 9 371 377. Skoog, I. (1998). Status of risk factors for vascular dementia. Neuroepidemiology, 17 29. Smith, A. D. (1996). Memory. In J. E. Birren & K. W. Schaie (Eds.), Handbook of the psychology of aging. (pp. 236-250). San Diego: Academic Press. Squire, L. (1987). Memory and brain. New York: Oxford University Press. Squire, L., & Alvarez, P. (1995). Retrograd e amnesia and memory consolidation: a neurobiological perspective. Current Opinion in Neurobiology, 5 169-177. Squire, L., Slater, P., & Chace, P. (1975). Retrog rade amnesia: Temporal gradient in very long term memory following electroconvulsive therapy. Science, 187 77-79. Squire, L., & Zola, S. (1997). Amnesia, memory and brain systems. Philosophical Transactions of the Royal Society of London, 352 1663-1673. Stadtlander, L. M., & Coyne, A. C. (1990). The effect of goal-se tting and feedback on age differences in secondary memory. Experimental Aging Research, 16 (1-2), 9194. Stebbins, G., Carrillo, M., Dorfman, J., Kirkse n, C., Desmond, J., Turner, D., Bennett, D., Wilson, R., Glover, G., & Gabrieli, J. D. E. (2002). Aging effects on memory encoding in the frontal lobes. Psychology and Aging, 17 (1), 44-55. Stock, J., & Cervone, D. (1990). Proximal goal-setting and self-regulatory processes. Cognitive Therapy and Research, 14 438-498. Stuss, D. T., & Benson, D. F. (1986). The frontal lobes. New York: Raven Press. Stuss, D. T., & Levine, B. (2002). Adult clin ical neuropsychology: lessons from studies of the frontal lobes. Annual Reviews in Psychololgy, 53 401-433. Suhara, T., Fukuda, H., Inoue, O., Itoh, T., Suzuki, K., Yamasaki, T., & Tateno, Y. (1991). Age-related changes in human D1 dopamine receptors measured by positron emission tomography. Psychopharmacology, 103 41-45. Sweet, L., Rao, S., Primeau, M., Mayer, A ., & Cohen, R. (2004). Functional magnetic resonance imaging of working memory among multiple sclerosis patients. Journal of Neuroimaging, 14 (2), 150-157.

PAGE 145

134 Syka, J. (2002). Plastic changes in the cen tral auditory system after hearing loss, restoration of function, and during learning. Physiological Reviews, 82 (3), 601636. Sziklas, V., & Petrides, M. (1998). Memory and the region of the mammillary bodies. Progress in Neurobiology, 54 55-70. Tamm, L., Menon, V., Johnston, C. K., Hessl, D. R., & Reiss, A. L. (2002). fMRI study of cognitive interference processing in females with Fragile X Syndrome. Journal of Cognitive Neuroscience, 14 (2), 160-171. Tisserand, D., & Jolles, J. (2003). On the involv ement of prefrontal networks in cognitive ageing. Cortex, 39 1107-1128. Tremblay, L., & Schultz, W. ( 1999). Relative reward preferen ce in primate orbitofrontal cortex. Nature, 398 704-708. Tulving, E. (1987). Introduction: multiple memory systems and consciousness. Human Neurobiology, 6 67-80. Tulving, E., Kapur, S., Craik, F. I. M., Mosc ovitch, M., & Houle, S. (1994). Hemispheric encoding/retrieval asymmetry in episodi c memory: Positron emission tomography findings. Proceedings of the National Academy of Sciences, USA, 91 2016-2020. Ulings, H. B. M., & de Brabander, J. M. (2002). Neural changes in normal human aging and Alzheimer's disease. Brain and Cognition, 49 268-276. Ulings, H. B. M., West, M. J., Coleman, P. D., de Brabander, J. M., & Flood, D. G. (2000). Neuronal and cellular changes in aging brain. In J. Trojanowski & C. Clark (Eds.), Neurodegenerative dementias and pathological mechanisms. New York: McGraw-Hill. Van Der Werf, Y., Jolles, J., Witter, M., & Uy lings, H. (2003). Contributions of thalamic nuclei to declarative memory functioning. Cortex, 39 1047-1062. Verhaeghen, P., Marcoen, A., & Goossens, L. (1993). Facts and fiction about memory aging: a quantitative integrat ion of research findings. Journal of Gerontology: Psychological Sciences, 48 P157-P171. Volkow, N. D., Gur, R. C., Wang, G. J., Fo wler, J. S., Moberg, P. J., Ding, Y. S., Hitzemann, R., Smith, G., & Logan, J. ( 1998). Association between decline in brain dopamine activity with age and cogni tive and motor impairment in healthy individuals. American Journal of Psychiatry, 155 344-349.

PAGE 146

135 Wagner, A. D., Maril, A., & Bjork, R. A. (1999). Executive control and episodic memory encoding: fMRI evidence for separabl e prefrontal control processes. Society for Neuroscience Abstracts, 25 296. Wagner, A. D., Poldrack, R., Eldridge, L., De smond, J., Glover, G., & Gabrieli, J. D. E. (1998). Material-specific la teralization of prefrontal activation during episodic encoding and retrieval. Neuroreport, 9 3711-3717. Wagner, A. D., Schacter, D., Rotte, M., Koustaal, W., Maril, A., Dale, A., Rosen, B., & Buckner, R. L. (1998). Building memories: remembering and forgetting verbal experiences as predicted by brain activity. Science, 281 1188-1191. Warrington, E., & Weiskrantz, L. (1982) Amneisa: A disconnection syndrome? Neuropsychologia, 20 233-248. Wegge, J., Kleinbeck, U., & Schmidt, K. (2001). Goal setting and performance in working memory and short-term memory tasks. In M. Erez & U. Kleinbeck (Eds.), Work motivation in the context of a globalizing economy. (pp. 49-72). Mahway, NJ: Lawrence Erlbaum Associates, Inc. Weinberger, D., Mattay, V., Callicott, J., Kotr la, K., Santha, A., van Gelderen, P., Duyn, J., Moonen, C., & Frank, J. (1996). FMRI a pplications in schizophrenia research. Neuroimage, 4 118-126. West, M. J. (1993). Regionally specific loss of neurons in the aging human hippocampus. Neurobiology of Aging, 14 287-293. West, R. L., & Thorn, R. M. (2001). Goal-setting, self-efficacy, and memory performance in older and younger adults. Exp Aging Res, 27 (1), 41-65. West, R. L., Thorn, R. M., & Bagwell, D. K. (in press). Memory performance and beliefs as a function of goal setting and aging. Psychology and Aging West, R. L., Welch, D. C., & Thorn, R. M. (2001). Effects of goalsetting and feedback on memory performance and beliefs among older and younger adults. Psychol Aging, 16 (2), 240-250. Wheeler, M. A., Stuss, D. T., & Tulving, E. (1995). Frontal lobe damage produces episodic memory impairment. Journal of the International Neuropsychological Society, 1 525-536. Woods, R. P., Cherry, S. R., & Maziotta, J. C. (1992). Rapid automation algorithm for aligning and reslicing PET images. Journal of Computer Assisted Topography, 16 (620-633).

PAGE 147

136 Yancey, S. W., & Phelps, E. A. (2001). Functional neuroimaging and episodic memory: a perspective. Journal of Clinical and Expe rimental Neuropsychology, 23 (1), 3248. Zola-Morgan, S., Squire, L., & Amaral, D. (1986). Human amnesia and the medial temporal region: enduring memory impair ment following a bilateral lesion limited to field CA1 of the hippocampus. Journal of Neuroscience, 6 2950-2967.

PAGE 148

137 BIOGRAPHICAL SKETCH Michael Cole graduated with honors from the University of Colorado, Boulder, in 1996 majoring in psychology. Mi chael continued his studie s at the University of Colorado, Boulder, obtaining a master’s degr ee in behavioral neuroscience in 1999. He began his predoctoral training at the Universi ty of Florida in 2000 in clinical psychology with focuses on clinical neuropsychology a nd cognitive neuroscience. Michael pursued his interest in executive functions and aff ective/motivational influences on cognition in aging and in traumatic brain injury populations using neuroimaging and neuropsychological research techniques. Michael will continue his training at the University of California Los Angeles Clin ical Psychology Internship Program in the clinical neuropsychology track.


Permanent Link: http://ufdc.ufl.edu/UFE0006045/00001

Material Information

Title: Effect of Goal-Setting on Memory Performance in Young and Older Adults: A Functional Magnetic Resonance Imaging (fMRI) Study
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0006045:00001

Permanent Link: http://ufdc.ufl.edu/UFE0006045/00001

Material Information

Title: Effect of Goal-Setting on Memory Performance in Young and Older Adults: A Functional Magnetic Resonance Imaging (fMRI) Study
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0006045:00001


This item has the following downloads:


Full Text












EFFECTS OF GOAL-SETTING ON MEMORY PERFORMANCE IN YOUNG AND
OLDER ADULTS: A FUNCTIONAL MAGNETIC RESONANCE IMAGING (fMRI)
STUDY















By

MICHAEL A. COLE


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


2005

































Copyright 2005

by

Michael A. Cole

































I dedicate this dissertation to my wife, Heather Graham Cole,
and to my daughter, Tiana Cassidy Cole.















ACKNOWLEDGMENTS

I would first and foremost like to acknowledge my wife, Heather Cole, for her

unwavering support and patience in my pursuit of doctoral training. I extend my genuine

gratitude to Dr. William Perlstein for the valued mentoring he has provided. I feel very

fortunate for the excellent neuropsychology and cognitive neuroscience training that I

received from Dr. Perlstein and other faculty at University of Florida. Much appreciation

is also extended to Dr. Robert Spencer, who helped me to build foundational research

skills and a fundamental understanding of the brain during my studies at the University of

Colorado. I give many thanks to Drs. Bruce Crosson and Eileen Fennell for their

exceptional mentoring in clinical neuropsychology, my professional development, and

my research endeavors. I would like to acknowledge Dr. Robin West for her fundamental

role in this project and extend my appreciation for her important advising in conducting

this research. I would also like to thank my dissertation committee for their guidance and

support in my doctoral training. I would like to extend my gratitude to Alissa Dark-

Freudeman, Roger Saldana, Vonetta Jones, and Neha Dixit for their contributions to this

research. Special thanks are in order for my parents, Dr. Dennis G. Cole and Mrs.

Barbara F. Cole, for their continued support, interest, and encouragement in my long road

of educational endeavors. This research was supported by the Evelyn F. & William L.

McKnight Brain Institute.

















TABLE OF CONTENTS

Page

ACKNOW LEDGM ENTS ........................................ iv

LIST OF TABLES ............... .............. .. .......... ............ ........... vii

LIST OF FIGURES .............................. .. .... .............. viii

CHAPTER

1 INTRODUCTION ................... .................. .............. .... ......... .......

O overview ............... ........ ......... ................... .......... ...............
Episodic Memory in Young and Older Adults.............. ........................................3
Neural Substrates of Episodic M emory Encoding................................ ..... .......... 5
Overview ................................................5
Temporal Lobes.............................. ............ .........10
Frontal Lobes................................... ... .... ... ..................13
Neural Changes in Aging Affecting Episodic Memory Encoding .............................15
T em poral L obes ................................................. 16
Frontal Lobes................................... .. ............ ........ 18
Effect of Enhanced Goal-directed Action on Cognition................ ....... ........22
Neural Substrates of Goal-directed Action in Cognitive Operations .......................24
Effect of Enhanced Goal-directed Action on Cognition in Older Adults.................30
Sum m ary and P reduction s ...................................................................................... 34

2 METHODS .........................................................36

Overview...................................... ................... ..................36
General M ethods............................................... 36
Subjects...................................... ................................ ........36
Exclusion Criteria.............. .............. ...............37
Experim mental Task and Procedures .............................................. ......38
Behavioral D ata Analysis ................................... ......................... ........... 44
Magnetic Resonance (MR) Acquisition ........................................45
FM RI Data Analysis.................. ...... ................ 46
Internal activation standard .............................. ............... 48
Memory Encoding Experiment ............ .............................51




v










3 R E SU L T S ....................................................... 56

B behavioral Perform ance .................................................. ............... 56
Functional MRI Findings.......................................58
Internal Activation Standard................. ........................ 58
Age Effect on Encoding Activity ............................................ .......60
Age Effect on Subsequent Memory Activity ...................................... 63
Goal-setting Effect on Encoding Activity ........................................ ........... 67
Goal-setting Effect on Subsequent Memory Activity .......................................71
Interaction of Goal-setting and Age during Encoding Activity .......................74
Interaction of Goal-setting and Age for Subsequent Memory Activity .............78

4 DISCUSSION ......................... ............ ..........81

Encoding in Young Adults ................................................82
Encoding in Older Adults and Age-related Changes....................... ...............83
Goal-setting Influence on Encoding .................... .................86
Differential Neural Response to Goal-setting in Young and Older Adults................88
Potential Limitations........................ .................90
Concluding Remarks .............................................. ..... ...94

APPENDIX

A SHIPLEY VOCABULARY TEST....................................... ...............96

B WORD LISTS FOR MEMORY TRIALS....................................98

C Q U E ST IO N N A IR E S ............................................................................................. 99

D STRATEGY USE QUESTIONNAIRE............................. .......... 108

E TISSUE-AIR INTERFACE SIGNAL DROP OFF........................110

F GLA SS BRAIN ILLU STRATION S ......................................................................111

LIST OF REFEREN CES ..................................... ................... .....122

BIOGRAPHICAL SKETCH .............................................................................137
















LIST OF TABLES


Table Page

2-1 Mean (Standard Error) Demographic Characteristics of Experimental Participants.38

3-1 Mean Percent (Standard Error) of Recall Performance for Each Group................57

3-2 Comparison of Activation during Encoding in Young and Older Adult Groups Not
R eceiv ing G oal-setting ....................................................................................... 64

3-3 Encoding Related Activity Differences between Young and Older Adult Groups
N ot R receiving G oal-setting ......................................................... ............... 65

3-4 Comparison of Subsequent Memory Effect in Young and Older Adult Groups Not
R eceiv ing G oal-setting ....................................................................................... 66

3-5 Subsequent Memory Differences between Young and Older Adult Groups Not
Receiving G oal-setting ..............................................................................67

3-6 Comparison of Activation during Encoding in the Goal and No-goal Groups........72

3-7 Comparison of Subsequent Memory Effect in Goal and No-goal Groups............75

3-8 Subsequent Memory Effect Differences between Goal and No-goal Groups..........76

3-9 Interaction of Age and Goal-setting during Memory Encoding............................79

3-10 Interaction of Age and Goal-setting during Encoding for Subsequent Memory
Effects ........................... ....................... 80
















LIST OF FIGURES


Figure Page

2-1 Graphical representation of the memory task for Trials 1 through 3..................42

2-2 GLM predictor model for signal intensity increases corresponding to the motor
response of a button press for a single subject. ..................................................50

2-3 GLM model for the multiple regression analysis of overall memory encoding for a
single subject. ..... ............. ................ ....... .........52

2-4 GLM model of the multiple regression analysis of memory recall performance
correlated w ith signal intensity. ................................. ............... 54

3-1 Summary of memory recall percent in a) young adult/goal, young adult/no-goal,
older adult/goal, and older adult/no-goal groups, b) memory trials, and c) blocks
within each trial ......................................................57

3-2 Left primary motor cortex (BA 4) activation during the motor task in young adults,
older adults, and older adults subtracted from young adults...............................60

3-3 Motor task-related z-transformed signal intensity change as a function of scan-in-
trial is displayed. ........................................... ........ 61

3-4 Effect of subsequent memory in young (n=10) and older (n=10) adults..............68

3-5 Effect of subsequent memory: older adults subtracted from young adults..........69

3-6 Effect of subsequent memory: no-goal group subtracted from the goal group........77

3-7 Coronal slice image illustrates PFC cluster (BA 9) that exhibited significant task-
related interaction of age by goal-setting. .........................................78

E-1 Illustration of signal drop off in portions of the orbitofrontal cortex in a single
subj ect....................... ......... .... ...................... 110

F-i Glass-brain representation of regional activations during encoding in older and
young adults. ............. ......................... ........ 112

F-2 Glass-brain representation of the regions showing significantly different activation
between young and older adults during encoding................. ............. .......... 113









F-3 Glass-brain representation of encoding activity correlated with subsequent recall
performance in young and older adults. ............. ...........................114

F-4 Glass-brain representation showing significant differences between young and
older adults for encoding activity correlated with subsequent recall performance. 115

F-5 Glass-brain representation of regional activations during encoding in goal and no-
goal groups (P < 0.001, corrected; minimum threshold of 250 contiguous voxels). 116

F-6 Glass-brain representation of the regions showing significantly different levels of
activation between goal and no-goal groups during encoding..........................117

F-7 Glass-brain representation of encoding activity correlated with subsequent recall
performance in goal and no-goal groups............................. ...... ... ....... 118

F-8 Glass-brain representation showing significant differences between goal and no-
goal groups for encoding activity correlated with subsequent recall performance. 119

F-9 Glass-brain representation of regions showing a significant interaction between age
and goal-setting during encoding. ........................................ .. ........ 120

F-10 Glass-brain representation of regions showing a significant interaction between age
and goal-setting for encoding activity that was correlated with subsequent recall
performance..................... ... ..... ................ 121
















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

EFFECT OF GOAL-SETTING ON MEMORY PERFORMANCE
IN YOUNG AND OLDER ADULTS: A FUNCTIONAL MAGNETIC RESONANCE
(FMRI) STUDY

By

Michael A. Cole

August 2005

Chair: William Perlstein
Cochair: Bruce Crosson
Major Department: Clinical and Health Psychology

Episodic memory decline in late life can be improved by use of explicit goal-setting

for performance. The neural correlates that underlie the benefits of goal-setting on

memory performance have yet to be examined. Event-related functional magnetic

resonance imaging (fMRI) was employed to investigate the neural correlates of memory

encoding as a function of age and memory enhancement by goal-setting. FMRI data were

obtained while 20 young adults (ages 18 28) and 20 older adults (ages 60 70)

performed 3 trials of a list-learning task that was comprised of grocery items. Half of the

young adult and half of the older adult groups received goals for performance

achievement prior to each of the 3 trials, whereas the other half of the young adult and

older adult groups did not receive performance goals. FMRI data were analyzed for

signal increases related to the encoding period, as well as signal increases that correlated

with subsequent recall performance of the word-lists. Young adults remembered a









significantly greater number of words than older adults. Significant performance

improvement derived from goal-setting was equivalent between young and older adults.

Functional MRI findings revealed left lateralized prefrontal cortex (PFC) activation in

young adults (not receiving goals), which is consistent with predictions of the

hemispheric encoding and retrieval asymmetry (HERA) model. Older adults

demonstrated left lateralized PFC activation as well. Consistent with predications of the

hemispheric asymmetry reduction in older adults (HAROLD) model, the left PFC activity

was significantly dampened in older adults. The effect of goal-setting on encoding

activity was primarily constrained to the frontal lobes. Regions that demonstrated

significantly greater activity in the goal group than in the no-goal group included the

orbitofrontal cortex (OFC), dorsolateral prefrontal cortex (dlPFC), and Broca's area.

Engagement of these regions likely reflects increased motivation and increased

mnemonic processes, such as subvocal rehearsal. In conjunction with goal-setting, older

adults activated several different regions to a greater extent than young adults. As these

regions were observed to be activated during encoding in the absence of goal-setting, the

differentially greater activation in older adults may reflect increased resources put toward

mnemonic processing in older adults or perhaps decreased overall efficiency.











CHAPTER 1
INTRODUCTION

Overview

As advances in modern medicine have enabled individuals to reach later and later

decades in their lives, much interest has accumulated into the physical and psychological

changes that take place in these late stages of life. Many remain highly inventive and

ingenious in their work well into the late stages of their life, such as Michelangelo,

Claude Monet, Frank Lloyd Wright, and Joseph Campbell, but these are unique examples

of notables who unfortunately are not representative of typical cognitive aging. A

particular area of interest in cognitive aging that has evolved is that of episodic memory,

a cognitive domain that experiences one of the steepest trajectories of decline with

increasing age (Connor, 2001; Luszcz and Bryan, 1999). The declines in episodic

memory performance have been firmly demonstrated, but less is known about the neural

substrates that underlie this decrease in behavioral performance. Anatomical findings

have revealed correlations between volumetric measures of the prefrontal cortex (PFC)

and the rate of decline in episodic memory (Raz, 2000). Additional evidence suggests

decreased activity (e.g., neural engagement) occurring in the PFC, as well as the temporal

lobes, during memory tasks (Grady, 2000).

Some strategic techniques have been shown to be effective in improving episodic

memory performance in both young and older adults. One such technique is goal-setting,

which provides a challenge to the individual and often results in improved episodic

memory performance (Linnenbrink et al., 1999; West et al., in press). Findings show that

both young and older adults can improve memory performance through the provision of









explicit goals, which clearly has a neural origin. Although the neural underpinnings of the

effect of explicit goal-setting on cognition has yet to be elucidated, evidence suggests a

role for the orbito frontal cortex (OFC) and the dorsolateral prefrontal cortex (dlPFC;

Jahanshanhi and Frith, 1998; Tremblay and Schultz, 1999).

The effect of goal-setting has generally been found to be as strong in older adults as

in young adults (West and Thorn, 2001; West et al., in press). This finding is perhaps

surprising in light of the fact that the two primary areas that likely participate in the goal-

setting effect on memory, the OFC and dlPFC, undergo the steepest rate of degradation

toward the latter stages of the lifespan (Band et al., 2001). It is therefore important to

examine if an equal extent of activation exists in the OFC and dlPFC in young and older

adults resulting from goal-setting. Alternatively, compensatory mechanisms may be

taking place, allowing older adults to gain equal benefit by greater activation and/or

recruitment of other regions to compensate for atrophy taking place in the OFC and the

dlPFC.

The present research first reevaluated previous findings that episodic memory

encoding in young adults is primarily lateralized to the left prefrontal cortex and that

older adults engage the left prefrontal cortex to a lesser extent. This research also

evaluated the role of the frontal cortex, specifically the OFC and the dlPFC, in the

memory enhancing effects of goal-setting. Prior studies indicate that both these regions

are involved with goal-directed behaviors. But the potential role for these frontal cortex

regions in mediating the influence of explicit goal-setting on memory performance

remains to be examined. It was hypothesized that the increased challenge resulting from

explicit goal-setting would be accompanied by significant increases in dlPFC and OFC









activity. Additionally, it was hypothesized that older adults would show greater activation

in these regions and perhaps more widespread activation to generate the equal benefit of

explicit goal-setting on memory performance, compensating for the age-related declines

seen in the OFC and dlPFC (Band et al., 2001).

Episodic Memory in Young and Older Adults

Different memory domains are not equally vulnerable to the declines occurring in

older age. In fact, the integrity of some memory domains remains highly intact, while

others are particularly susceptible to aging. A major distinction between memory types,

and domains that demonstrate this differential vulnerability to aging, are the two types of

declarative memory: semantic memory and episodic memory (Tulving, 1987). Semantic

memory pertains to an individual's general knowledge about the world. This information

includes, but is not limited to, vocabulary, facts, and concepts. Semantic memories are

ones that an individual can volitionally bring to consciousness, but typically cannot report

where the specific knowledge unit was obtained. This knowledge is therefore not

associated with specific learning contexts or events, as it is in episodic memories.

Overall, studies of semantic memory integrity in older adults show very little difference

from young adults in the retrieval and use of semantic information (Madden et al., 1993).

On the other hand, older adults do experience declines in episodic memory, which

refers to the ability to remember specific events situated in time and place. Episodic

memory is commonly thought of as the acquisition, storage, and retrieval of information

that refers to a specific context and is consciously and intentionally recollected (Tulving

et al., 1994). Studies consistently show declines in episodic memory performance in older

adults (Craik and Jennings, 1992; Smith, 1996). Findings indicate that older adults often

have difficulty with encoding (the initial storage of information), as well as retrieval









(Craik and Jennings, 1992). This decline has been observed with virtually every type of

stimulus, such as prose passages, single words, spatial locations, pictures, faces, and

activities (Burke and Light, 1981; Light, 1991; Smith, 1996). Additionally, age

differences have been demonstrated for spontaneously used elaborative and

organizational strategies to store and retrieve information, as older adults are less likely to

engage in these sorts of mnemonics. The greatest age differences are observed in tests of

recall. Tests of cued recall and recognition reveal less substantial, but significant

differences between young and older adults (Smith, 1996). These age-related changes are

not peculiar to the artificial nature of memory tests that are administered in laboratory

settings because they also occur in tasks that are designed to emulate memory in

everyday life (Kirasic et al., 1996).

Strong evidence exists that these age-related declines in episodic memory are not

due to older adults being less motivated to remember unfamiliar or unimportant stimuli of

laboratory tasks because the memory decline still exists for more naturalistic stimuli.

These stimuli include such examples as hands in bridge, groceries on a shelf (Read,

1987), board positions in chess (Charness, 1981), instructions on bottles of prescription

medicine (Morrell et al., 1990), people's names (Cohen and Faulkner, 1986), and golf

shots (Backman and Molander, 1986). A meta-analysis by Verhaeghen et al. (1993)

confirmed results from individual experiments by reporting a negative relationship

between age and episodic memory test performance with recall and recognition. They

found that the average older adult above the age of 60 performed at a level between the

16th and 25th percentile of the young adults' performance distribution on various

measures of recall. This result suggests that the average older adult's performance on









episodic memory tasks is approximately one standard deviation lower than that of the

average performance of a young adult (Verhaeghen et al., 1993).

Especially in the last decade, researchers have begun to examine the neural

substrates that might underlie these changes in memory performance across the lifespan.

Studies have focused on the neuroanatomical changes that exist in older adults in addition

to investigations of functional activation of brain regions in vivo as a subject performs an

episodic memory task. Research identifying these neuroanatomical and functional

changes has been successful in implicating likely mechanisms that may account for the

decline in episodic memory performance in older age.

Neural Substrates of Episodic Memory Encoding

Overview

For some time it has been known that a person who has suffered a traumatic brain

injury can have selective memory loss for events that occurred before (retrograde

amnesia) and after (anterograde amnesia) the event that precipitated the traumatic brain

injury. This phenomenon has been investigated thoroughly in animal studies using

approaches such as electroconvulsive shock, physical trauma to the brain, and drugs that

depress neuronal activity or inhibit protein synthesis in the brain. Clinical studies also

indicate that brain trauma can produce amnesia that is particularly prevalent for recent

events. Findings indicate that more recent memories are more susceptible to disruption,

whereas older memories remain quite intact (Kupfermann, 1991). Squire et al. (1975)

investigated this phenomenon in patients with depression who received electroconvulsive

treatment. They used a memory test that could reliably quantify the degree of memory for

relatively recent events (1-2 years old), old events (3-9 years old), and very old events (9-

16 years old). Patients were asked to name television programs that were broadcast









during a single year between 1957 and 1972. The patients were initially tested and then

tested again (with a different set of television programs) after the electroconvulsive

therapy. After patients received electroconvulsive therapy, memory for television shows

from less than 2 years removed were selectively impaired, whereas memory performance

for more temporally removed programs remained consistent with performance levels

observed prior to electroconvulsive therapy.

The differential susceptibility to disruption of memory that is dependent on the

time of acquisition brings about possible explanations for how memories are stored and

how neural changes that are associated with memories are maintained for years. One

possibility is that the dynamic change that underlies the initial encoding of a memory

persists and represents the long-term memory as well, such as a reverberating circuit.

Another possibility is that long-term memories are related to some plastic rather than

dynamic change (e.g., a persistent functional change within the brain). The extent

literature provides strong support for the latter possibility. Studies have shown that by

silencing the brain through use of deep anesthesia, anoxia, or by cooling the brain, short-

term memories or recent memories are disrupted, but older memories are not. Therefore,

it can be concluded that at least older memories are not mediated by dynamic change, but

involve physical changes in the brain. It is thought that the storage of long-term

memories is in part mediated by processes such as increased protein synthesis, growth of

new synaptic connections, and increased synaptic efficacy (such as long-term

potentiation; Kupfermann et al., 1991).

A central region in the facilitation of memory storage is the medial temporal lobe.

The medial temporal lobe is needed at the time of learning to establish functional









connections with widespread areas of neocortex, based on neural activity that occurs at

the time of learning. Medial temporal lobe lesions spare short-term (immediate) memory,

presumably because the neocortex can support short-term memory. Therefore, it is

thought that the medial temporal lobe is involved with processing and analyzing. This

function begins at the time of learning as it receives highly processed input from

neocortical association areas and continues to interact with the neocortex as it processes

the information. Deficits from medial temporal lobe lesions have been described as

extreme forgetfulness, and these deficits are most salient after some time has elapsed

after the point of learning. This dissociation between perception and short-term memory

and long-term-memory has been well established in humans, monkeys, and rats (Squire

and Zola, 1997). In contrast, lesions in the neocortex impair memory abilities at both

short and long delays (Goldman-Rakic, 1987; Fuster, 1995.)

The medial temporal lobe is involved in memory for a limited period of time after

learning. As time passes, memory is slowly consolidated, and information storage in the

neocortex becomes independent of the medial temporal lobe system. This is evidenced by

the finding that if a medial temporal lobe lesion is sufficiently delayed after learning,

memory is not affected. For instance, object discrimination tasks in monkeys

demonstrated this temporally graded amnesia with lesions at different times following

learning (Zola-Morgan et al., 1986). In contrast, there is no evidence of temporally

graded amnesia in the neocortex (Squire and Zola, 1997).

Another important characteristic of the medial temporal lobe is that damage to this

region produces memory deficits that are global and multimodal. That is, the memory

impairments are present regardless of the type of material to be remembered, such as









objects, words, or designs, or the sensory modality in which information is presented

(Baltes, 1993). In contrast, memory deficits associated with neocortical lesions are

domain specific. That is, they are specific to the kind of material that is ordinarily

processed by the damaged area.

The global and multimodal functioning of the medial temporal lobe is an important

characteristic for theoretical accounts of memory consolidation, which assert that the

medial temporal lobe directs consolidation in the neocortex by gradually binding together

the multiple cortical regions located in different areas, storing a memory for a whole

event (Squire and Alvarez, 1995).

Much like the medial temporal lobe, the basal forebrain is also thought to

participate in the storage of memories in the neocortex. Neurons of the basal forebrain are

activated by sustained attention in learning (Muir et al., 1993), which is a condition

during which cortical plasticity often takes place. Evidence suggests that cholinergic and

GABAergic neurons projecting from the basal forebrain (and in particular the nucleus

basalis) can induce experience-induced plasticity changes in, for instance, auditory

cortical responses. Response characteristics of the auditory cortex can be altered by

repeatedly pairing of sounds with basal forebrain stimulation (Kilgard and Merzenich,

1998). Additionally, it has been shown that experience-induced plasticity can be blocked

by lesioning the basal forebrain (Kilgard and Merzenich, 1998) or blocking cholinergic

effects (Baskin and Weinberger, 1996).

Study of Korsakoff s syndrome patients has been informative in elucidating the

role of the other structures important to memory function. Patients with Korsakoff s

syndrome suffer from similar amnestic features as do patients who have had damage to









temporal lobe structures. Korsakoff s syndrome, which is caused by chronic alcoholism

and associated nutritional deficiency, is associated with signs of frontal lobe dysfunction

in addition to severe memory deficits. Patients exhibit pathological changes in

diencephalic structures such as the mammillary bodies of the hypothalamus and the

medial dorsal nucleus of the thalamus. Warrington and Weiskrantz (1982) found that

when patients with Korsakoff s syndrome are given a list of words to remember, they do

poorly on a simple recall task, but their performance is significantly improved when

recall is tested by the use of prompts or partial cues. The authors concluded that this

finding represents intact priming in the Korsakoff s syndrome patients in the presence of

poor episodic memory abilities.

Evidence for thalamic involvement in memory is also derived from patients with

infarctions, haemorrhages, mechanical injury, or tumor interfering with the integrity of

the thalamus. Studies have shown a role for the thalamus in various cognitive functions

related to memory such as the formation of new memories, attention to stimuli and

events, and the use of memory strategies (Van Der Werf et al., 2003). Studies in animals

indicate that a large lesion of the medial dorsal region of the thalamus is sufficient to

produce learning deficits analogous to those exhibited by amnesic patients (Kupfermann,

1991; Van Der Werf et al., 2003).

Evidence outside of Korsakoff s syndrome patients exists for a role of mammillary

bodies in memory. For instance, two cases of well documented amnesia have been

reported in which marked neuronal loss was found within the mammillary bodies (Mayes

et al., 1988). In addition, lesioning of the mammillary bodies in animals results in









significant impairments in the performance of spatial delayed alternation tasks in rats,

cats, and monkeys (Sziklas and Petrides, 1998).

The above brief overview of neural substrates underlying memory function was

primarily limited to animal and neuropsychological studies. Differentiating the role of

these regions in encoding versus retrieval is difficult as the approaches are based

primarily on lesion studies. With functional neuroimaging, cerebral metabolic activity is

measured "on-line" as a cognitive task is being performed, and, in this way, the brain

regions recruited for specific memory processes can be identified. Functional

neuroimaging studies have played an important role in evaluating the neural substrates of

memory because, in part, the encoding and retrieval processes of memory can be

observed independently in real-time. As the present research evaluates episodic memory

encoding, the following sections provide a focused review of primarily functional

neuroimaging findings relating to the encoding phase of episodic memory.

Temporal Lobes

As mentioned above, the medial temporal lobe is central to memory function. The

report by Scoville and Milner (1957) of impaired memory in patient H.M., who had a

medial temporal lobe resection, was the first to highlight this region's importance in

memory. Later neuropsychological, neuroscientific, and psychological research all

converged on the medial temporal lobe, especially the hippocampal portion, as the site

that mediates the storage of memories for episodes and factual knowledge of the world

(Squire, 1987; Tulving, 1987). With the advent of neuroimaging, a logical region to begin

the study of episodic memory would be the medial temporal lobe. The memory literature

would suggest that episodic memory tasks would surely induce activation of the

hippocampal complex. Surprisingly, though, many initial studies did not report









significant activation in this region (Buckner et al., 1995; Petersen et al., 1988; Frith et

al., 1991; Demonet et al., 1992; Grasby et al., 1993, 1994). Several possible explanations

exist that could account for these surprising findings. First, signal difference between

experimental conditions is often only 1% to 2%, and even lower in the hippocampus

because it resides in a region that is close to sinus cavities and subsequently subject to

more noise. Thus, there may not have been sufficient signal over increased background

noise to detect activations that may have been present. Additionally, some researchers

have asserted that the nature of hippocampal neural functioning itself may have

contributed to the inability to detect hippocampal activation between conditions (Cabeza

and Kingstone, 2001). The hippocampus is central to information processing. Therefore it

is frequently active, perhaps making it difficult to detect any increases in activation in an

episodic memory condition.

Many more recent neuroimaging studies, which have investigated hippocampal

involvement in episodic memory, have attempted a more nuanced assessment of

hippocampal functioning. It was initially assumed that effort directed toward encoding

and recollection should result in increased hippocampal activity. However, neuroimaging

studies thus far have not found effort in encoding and recall/recognition to be an

important factor as success in recollection. Specifically, hippocampal activation has been

shown to be significantly correlated to success of memory performance (Fernandez et al.,

1998; Yancey and Phelps, 2001).

A study by Fernandez et al. (1998) found that activation in the posterior

hippocampus was significantly correlated with successful encoding of verbal stimuli. In a

follow-up study, this same group compared results using the correlational method









between encoding activation and performance with the more traditional method of

cognitive subtraction. The cognitive subtraction approach entails, in this context, the

activation of the episodic memory encoding component process having an identical task,

but without the mnemonic functions subtracted from it (Fernandez et al., 1999). They

found that the entorhinal cortex did not respond transiently as the study word appeared

(as assessed using the cognitive subtraction technique), but did correlate positively with

subsequent test performance. This study was the first to directly compare these two

techniques finding that the correlation with performance technique is able to much more

reliably detect medial temporal lobe activation related to encoding.

With this novel approach, Brewer et al. (1998) demonstrated that the strength of

medial temporal lobe activity during encoding predicts not only what items will be

remembered, but also how well they will be remembered. They found that the magnitude

of activation in the bilateral parahippocampal cortex predicted which picture stimuli were

later remembered well, remembered less well, or forgotten. The distinction between

remembering well and remembering less well was provided by a subjective report by the

participant.

Schacter and Wagner (1999) reviewed all fMRI studies of memory encoding and

compared them to results of a meta-analysis of positron emission tomography (PET)

studies of episodic memory encoding (Lepage et al., 1998). Schacter and Wagner's

review found that encoding processes resulted in more posterior medial temporal lobe

activation across various materials and conditions (Schacter and Wagner, 1999). The

meta-analysis by Lepage et al. (1998) of PET studies found a slightly different

conclusion: episodic encoding is associated with more anterior medial temporal lobe









activation. Therefore, both methodologies converge in identifying the medial temporal

lobe as an important region in encoding. The discrepancy between more anterior medial

temporal lobe regions being activated in PET studies and more posterior medial temporal

lobe regions in fMRI studies can likely be accounted for by the fact that PET is more

sensitive to activation in the anterior portions of the medial temporal lobe. FMRI is

known to be characterized by susceptibility artifacts that can be pronounced in the

anterior medial temporal lobe, resulting in less sensitivity (Ojemann et al., 1997).

Frontal Lobes

Lack of activation of the hippocampus in many early neuroimaging studies was

perplexing to those performing these experiments. Equally perplexing was the consistent

and robust activation found in the frontal lobes. Prior to these studies, the frontal lobes

were not necessarily thought to be major contributors to episodic memory. Patients who

suffer damage to the frontal lobes, for instance, do not exhibit the pervasive and disabling

amnesia that is characteristic of patients with hippocampal lesions (Wheeler et al., 1995).

In functional neuroimaging studies, a highly consistent, lateralization of frontal lobe

function has been found with the left frontal cortex showing predominant activation in

association with learning or encoding tasks. The primary areas within the frontal cortex

where this lateralized contribution to episodic memory encoding has been found are in

ventral regions (Brodmann's areas 44, 45, and 47) and dorsal regions (Brodmann's areas

9 and 46) of the PFC. These findings have been integrated in the HERA (hemispheric

encoding and retrieval asymmetry) model, which states that the left prefrontal cortex is

more involved in episodic encoding (Tulving et al., 1994).

Though ventral PFC appears to contribute to episodic memory encoding, it has not

been consistently found to be activated across experiments. Wagner et al. (1999) found









significant activation in the left ventral lateral frontal cortex, posteriorly in Brodmann's

area (BA) 44 and anteriorly in BA 45/47, for encoding words that would later be

remembered well versus words that would later be forgotten. Another study by Hensen et

al. (1999) found that words that would be later specifically recalled as having been seen

in the study phase exhibited significantly greater correlation with activation in the ventral

PFC than words being classified as having been seen in the study phase but not

specifically remembered. Fernandez et al. (1999) found that only the left BA 45 exhibited

transient hemodynamic responses during encoding of verbal information.

The dorsal PFC, on the other hand, is reliably activated during encoding of episodic

memories. Studies have demonstrated activation in the left dorsal PFC during the

encoding of words (Grady et al., 1998; Kelley et al., 1998; Kopelman et al., 1998;

Nyberg et al., 1996; Wagner et al., 1998); word pairs (Dolan and Fletcher, 1997; Fletcher

et al., 1995; Halsband et al., 1998; Kapur et al., 1996); and word lists (Fletcher et al.,

1998).

The left dorsolateral prefrontal cortex (dlPFC) has also been shown to contribute

specifically to the implementation of strategic processes for the encoding of episodic

memories. In a PET study, Fletcher and colleagues manipulated the level of attention to

and the degree of organization of study material (Fletcher et al., 1998). The degree to

which the subjects were required to organize word lists semantically was systematically

varied across three experimental conditions. Left dlPFC activity reached its highest levels

when organizational demands were the greatest. A role for the dlPFC in attention was

suggested by dlPFC activity being attenuated by a concurrent motor distraction task

during the most organizationally demanding task. Subsequent retrieval was also









correspondingly attenuated under this condition. Wagner et al. (1999) also found dlPFC

activation during the implementation of strategic memory processes that improved

memory performance. Subjects were presented with three words that they had to maintain

either in the same order for a short period by means of subvocal rehearsal or to reorder

along an abstract semantic dimension (e.g., pleasantness). Both activities engaged the

dlPFC, but the semantic reordering of the words resulted in greater activation of the

dlPFC. The semantic reordering of the words subsequently led to better memory as well.

In summary, neuropsychological and non-primate animal studies have

documented the important role of the temporal lobes, diencephalon, and basal forebrain

in memory function. Neuroimaging studies have allowed for a greater ability to evaluate

the neural substrates of specific components of memory function, such as encoding.

Neuroimaging findings have confirmed the role of temporal lobe structures in encoding.

Additionally, neuroimaging studies have revealed the importance of the left PFC in

encoding. Both the temporal lobe and PFC undergo significant changes in aging. The

next section will discuss these changes and how they relate to episodic memory

performance in older adults.

Neural Changes in Aging Affecting Episodic Memory Encoding

This section describes the multiple neuronal changes that take place in aging in the

context of the effects on episodic memory encoding. Two key regions involved in

episodic memory (temporal lobe and PFC) appear to experience both anatomical and

physiological changes with aging that likely serve as the underpinnings of episodic

memory performance declines.









Temporal Lobes

Several studies demonstrated modest structural and physiological changes in the

temporal lobe regions in aging. Post-mortem studies of animal and human brains reveal

age-associated changes in the hippocampal size (Geinisman et al., 1995). In one of the

most comprehensive human post-mortem studies, the correlation value between neural

counts in the hippocampus and age was r = -.21 (West, 1993). Aging also affects the

neuronal architecture of the hippocampus. Although not as rampant in the normal aging

brain as in Alzheimer's disease, neurofibrillary tangles display a similar characteristic

regional distribution with the highest concentration in the hippocampus (Kemper, 1994).

The hippocampus shares other architectural deformities, such as Hirano bodies and

granulovacular degeneration (Kemper, 1994). The hippocampus is therefore the focal

point of several deleterious events associated with aging. However, the magnitude of

these age-related effects is thought to be within the mild range, as compared to effects in

other brain regions (Raz, 2000). In keeping with this estimate, in vivo neuroimaging

studies reveal only mild age-associated shrinkage of the broadly defined hippocampal

formation (Raz, 2000).

Age-associated volume reduction takes place in other temporal lobes regions

outside the hippocampus as well, and this reduction has been estimated to be about 1%

(Haug and Eggers, 1991). The reduction is thought to result primarily from a decrease in

neuron size rather than an actual loss of neurons (Haug and Eggers, 1991). Three research

groups (Coffey et al., 1992; Cowell et al., 1994; Murphy et al., 1996) found similar

correlations between age and temporal lobe volume, having a combined average

correlation ofr = -.25. Raz et al. (1997) found a correlation between inferior temporal

lobe volume and age ofr = -.32.









These modest structural changes observed in the temporal lobes are accompanied

by physiological changes as well. An fMRI study conducted by lidaka et al. (2001)

provided evidence for functional changes in older adults taking place in the medial

temporal lobe during episodic memory encoding. Young and older adults were studied

while they encoded pairs of concrete and abstract pictures. Age differences were found in

correlations between memory performance and amplitude of signal change in the

parahippocampal gyrus under both concrete and abstract picture conditions. Specifically,

medial temporal lobe regions were demonstrated to experience reduced activation in

older adults during encoding.

A decrease in activity in the medial temporal lobe during episodic encoding was

also found in a study by Bennett et al. (2001). PET was used to measure activational

differences between young and older participants during encoding of simple visual

attributes of a sine wave gradient screen. They found that there was a significant decrease

in activation in the left medial temporal gyrus in older adults.

Application of covariance analysis to human neuroimaging data suggests that age

modifies the relationship between the hippocampus/medial temporal lobes and other

brain regions. Age-related differences in the relationship between the hippocampus and

the cingulate gyrus during episodic memory encoding have been suggested by Grady et

al. (1995). An age-associated change in neural interactions between the hippocampus and

the dlPFC during episodic encoding has also been suggested by D'Esposito el al. (1999).

However, these findings are controversial because they use statistical methods to parse

out purported functional interactions between regions in neuroimaging studies that are

based on independent observations of activations.









Overall, the age-associated decline in episodic memory encoding in older adults

appears to be accompanied by a slight decrease in activation of the hippocampal/temporal

lobe regions as well as modest morphological changes. It should be noted that the above

functional neuroimaging studies reporting amplitude of activity differences between

young and older adults are based on the assumption that the measured signal intensity

differences genuinely reflect differences in neural activity. Given that these functional

neuroimaging studies are based on cerebral blow flow, it is possible that differences

between young and older adults are indicative of reduced vascular responsivity to neural

activation that is due to compromised cerebrovascular function (D'Esposito et al., 1999).

This possible interpretation will later be discussed at length in the context of the present

research.

Frontal Lobes

Episodic memory, as previously stated, experiences one of the steepest trajectories

of decline of all the cognitive domains. Perhaps not coincidentally, the frontal lobes too

experience one of the steepest trajectories of decline in older age. Ample evidence

suggests that structural and functional changes in the frontal lobes experienced in older

age do indeed contribute to the decreased memory performance observed in older adults.

Age-associated volume reduction in the frontal lobes has been estimated to be from

10% to 17% (Haug and Eggers, 1991). This reduction may result from a reduction in

neuron size rather than from an actual loss of neurons, similarly to what is observed in

the temporal lobe (Haug and Eggers, 1991). Shrinkage of cells in the frontal lobes

appears to begin earlier and is more severe than in any other region. An estimated 22%

shrinkage in cells outside of the pyramidal layer of the PFC occurs within the fifth to

seventh decades of life. Above the age of 65, the reduction in cell size becomes more









pronounced, reaching a 43% reduction in cells outside of the pyramidal cell layer in the

prefrontal cortex (Haug and Eggers, 1991).

Several studies have also found significant differences between young and older

adults in the degree to which they engage these frontal regions during memory encoding.

Grady et al. (1995) performed the first study demonstrating a functional neuronal basis

for why older adults perform worse on memory tasks. This study utilized an episodic

memory task that entailed memorizing faces and then later choosing the ones that were

previously seen. The left prefrontal cortex was activated during encoding in young adults.

In older adults, however, there was no significant activation observed in the left

prefrontal cortex during encoding. Studies that followed also found this difference

between young and older adults. It should be noted, however, that while finding a

significant difference between young and older adults in the left PFC, there was still

significant activation in the left PFC of older adults in studies that followed. This

difference may likely be attributed to the statistical power of the experiment in

conjunction with the conservative significance threshold used (Cabeza, 2002).

Studies that found the dampened, but still significant, left PFC activation in older

adults include the previously mentioned fMRI study conducted by lidaka et al. (2001).

Young and older adults were studied while they encoded pairs of concrete and abstract

pictures. Older adults showed significant activation of the left dorsal PFC during

encoding of both concrete and abstract pictures. This activation in older adults was

reduced as compared to young adults.

Several other studies have looked specifically at left PFC activation during

encoding in older adults as compared to young adults. These studies consistently found









age-related decreases in activation of left prefrontal areas: BA 6 (Cabeza et al., 1997;

Cabeza et al., 1997); BA 10 (Madden et al., 1996); BA 45 (Grady et al., 1995); and BA

46 (Cabeza et al., 1997).

Cabeza et al. (1997) conducted a unique approach in their PET study of episodic

memory in an attempt to explain the differences between memory activations in young

adults and those observed in older adults. They aimed to determine if the changes in

activation observed in older adults were a result of local neural changes or if there were

more global changes taking place in the way that regions interact. In order to test this,

they performed a path analysis on the areas of activation observed during encoding and

retrieval during an episodic memory task. Their path analysis indicated that the patterns

of activation in older adults reflected a global shift in processing of the memory

information. The authors concluded that the neural changes in memory encoding that take

place in older adults is not limited to a few discrete regions in the brain in isolation, but

rather is a global alteration in the way the many neural networks behind memory

encoding interact with each other. This conclusion is perhaps premature, however, given

that nothing implicit in the design of their study would allow them to exploit possible

connectivity changes between brain region activations.

Other methods have been used to investigate whether global shifts in neural

processing can account for the activational changes observed in older adults during

mnemonic tasks, or if theses changes are specific to only a few discrete brain regions.

Evidence for a "common cause" behind the episodic memory declines was found in some

studies (Cabeza et al., 1997; Grady et al., 1995), as evidenced by a decreased activation

seen in the older adults' fusiform and/or the lingual gyrus. On the other hand, evidence in









other studies suggested specific processes as a possible etiology, in which equal

activation across age groups in the same regions of visual association areas was observed

(BA 19 and BA 37; Cabeza et al., 1997; Madden et al., 1999). Further support for the

specific-process hypothesis can be taken from activation changes observed in the left and

right PFC during retrieval of episodic memories. Specifically, significant lateralized

activation in the right PFC is observed during retrieval in young adults, but older adults

show left and right PFC activation during retrieval of episodic information (Cabeza,

2001, 2002; Madden et al., 1999). Based on this finding, the left PFC does not suffer

from a generalized reduction in activity during memory tasks in older adults because it is

consistently shown to have an increased activation in older adults during retrieval of

episodic information. Overall, the decrease in activation of the left PFC during encoding

observed in older adults is therefore likely related to process-specific changes observed in

aging.

The findings of reduced left prefrontal activity during encoding in older adults

compared to young adults have been integrated in the HAROLD (hemispheric asymmetry

reduction in old adults) model. The HAROLD model states that young subjects, in line

with the HERA model, engage the left frontal cortex more heavily during encoding,

whereas older subjects experience reduced left prefrontal activity (Cabeza et al., 2001).

Numerous pharmacological and behavioral interventions have been developed in

order to attempt to address these deleterious changes related to episodic memory

performance decline in late life. One such intervention that appears promising is goal-

setting.









Effect of Enhanced Goal-directed Action on Cognition

The use of explicit goals has proven to be an effective approach to improving

cognitive performance, especially memory performance. Certain parameters of goals

have been identified that result in the goal being more efficacious. Goal-theory directly

addresses these parameters indicating that goals must have a certain degree of specificity,

difficulty, and proximity to be effective in optimizing behavioral performance (West and

Thorn, 2001). Goals that are difficult and at the same time attainable tend to motivate

improvement in performance (Lee et al., 1989). Research has shown that goals that are

not attainable may serve to be more discouraging than motivating (Bandura, 1989).

Support for the importance of specificity of goals has been provided by areas of

education (Schunk, 1990) and organizational management (Lee et al., 1989). These

authors found that goals directed toward specific performance levels, which contain

measurable outcomes, proved to be the most efficacious. The greater clarity in the

performance level expected for the goal, as well as the goal being clearly measurable,

results in greater performance increases than do more general goals in which the outcome

is less clearly ascertained. Proximal goals have been found to be more effective than

distal goals in increasing an individual's motivation and expectations regarding task

performance and self-efficacy on tasks that are even easy and intrinsically interesting

(Madderlink and Harackiewicz, 1984). Lastly, Bandura (1989) found that short-term

goals are more effective than long-term goals because they allow the individual to track

more effectively the progress that is made.

Goals have the effect of increasing the challenge to the individual for the

cognitive task at hand. When a goal is utilized for a specific task, the performance

requirements for the individual are raised above the otherwise implicit assumption that









she/he is to perform well. Research has shown that this increase in challenge results in

performance advantages. It has been shown that individuals will work towards achieving

their goals (Bandura and Cervone, 1983; Elliott and Dweck, 1988). Stock and Cervone

(1990) have also shown that goals are strongly related to the task-associated effort and

persistence. Goals provide not only an increase in motivation for achieving success in the

task at hand, but they also increase cognitive activity in efforts to achieve the goal, that is,

greater use of strategies (Elliott and Dweck, 1988). A study by Hinsz and Ployhart (1998)

evaluated the effects of goals on the performance of a word pair memory task. They

measured "trying" in this task, which they operationalized in terms of effort, persistence,

attention, and the use of effective strategies while performing a task. They found that

subjects provided with goals significantly increased "trying" in the verbal memory task.

A study by Juergen et al. (2001) found that goal-setting substantially improved

performance in two types of memory tasks. Two conditions of goal-setting were

employed: do your best versus specific and difficult goals. The first of the two memory

tasks was a reading span test that required reading aloud sets of sentences continuously.

The subject then had to recall the final words of every sentence each time all sentences of

a set were processed. In the second task, a memory span test with lists of one-syllable

words was used. The goal-setting condition resulted in greater motivation, as well as

significantly greater memory performance in both of the memory tasks. Because evidence

suggested that increased memory performance was not precipitated by different encoding

or recall strategies, the authors concluded that goal-setting resulted in temporary

cognitive arousal. A study by Linnenbrink et al. (1999) found that the setting of mastery

goals in a working memory paradigm also improved performance. Task-irrelevant









thoughts were decreased and motivation for higher achievement was increased in the

mastery goals condition, which the authors asserted to be a major contributor to the

greater memory success.

One's progress on goal attainment can be monitored either through attending to

their progress as they proceed through the task (assuming the performance is readily

discernable by the subject) or by means of explicit, external feedback. Cognitive

performance can be further improved by providing external feedback in addition to the

increased challenge in a task provided by goals. In general, explicit feedback in

conjunction with goal-setting has been shown to be more effective than goal-setting alone

for enhancing performance and efficacy (Bandura, 1989; Bandura and Cervone, 1983).

These studies indicate that a specific goal coupled with the ability to monitor

performance is the most effect approach to improvement memory performance through

goal-setting. The underlying neural substrates that mediate the effect of goal-setting

remain to be elucidated. Several areas of research exist, however, that bring the OFC and

the dlPFC into light as being likely candidates in mediating the effect of goal

enhancement of memory performance.

Neural Substrates of Goal-directed Action in Cognitive Operations

The neural substrates supporting the explicit use of goals in the improvement of

cognitive performance remain to be clearly elucidated. Ample literature exists, though,

evaluating the contribution of goal-directed behavior and willed action to cognition. More

specifically, studies have elucidated the neural substrates involved in motivation, drive,

and/or effort arising from implicit goals that influence cognition. Interestingly, findings

from diverse methodologies and paradigms all converge to suggest a role for the dlPFC

and OFC as major participants in goal-directed behavior and willed action. These studies









serve to reveal the possible neural substrates supporting the explicit use of goals in the

improvement of memory.

Stuss and Levine (2002) primarily utilized neuropsychological evidence to inform

their conceptualization of ventral PFC functions in goal-directed behavior. They coined

the term "self-regulatory disorder" (SRD) to characterize the clinical manifestation of

individuals having suffered an insult to the ventral PFC. They define SRD as "the

inability to regulate behavior according to internal goals and constraints" (Stuss and

Levine, 2002, p. 405). Damage to the ventral PFC impairs an individual's ability to

maintain goals internally, which results in highly disorganized behavior. Their findings

show that patients with this neuropsychological deficit remain largely unable to perform

complex goal-directed behavior.

Barrash et al. (2000) performed a study including 7 participants with bilateral

ventromedial PFC lesions, 14 participants with PFC lesions but no ventromedial

involvement, and 36 participants with nonfrontal lesions. Subjects were administered the

Iowa Rating Scales of Personality Change in which informants rated 30 specific

characteristics for degree of disturbance and change from premorbid personality. They

found that only the bilateral ventromedial lesioned group had significant impairments in

goal-directed behaviors. Specifically, bilateral ventromedial patients had significant

problems in planning, initiation, and persistence.

Tamm et al. (2002) investigated neural activation in females with Fragile X

Syndrome and normal controls while performing a counting Stroop interference task.

Fragile X Syndrome is an X-chromosome linked syndrome that results in mild mental

retardation in females. The authors found that the experimental group had a decreased









ability for goal-directed behaviors and consequently a decrease in performance, which

was associated specifically with a reduced activation in the left orbitofrontal gyrus.

Many studies have demonstrated the involvement of the OFC in the reward circuit,

but a study by Tremblay and Schultz (1999) was able to demonstrate its direct

involvement in the motivational aspects of appetitive behavior. They investigated goal-

directed behaviors in non-human primates, recording neuronal activity in the OFC during

a spatial delayed responding task. Neurons of the OFC selectively became more excited

in response to reward-predicting signals, during the expectation of rewards, and after the

receipt of rewards. The authors concluded that neurons in the OFC appear to process the

motivational value regarding outcomes of voluntary action.

The previous studies demonstrate that the OFC participates in goal-directed actions.

As stated earlier, the dlPFC has also been shown to be involved in goal-directed

activities. For instance, Jahanshai and Frith (1998) reviewed studies in which they found

the dlPFC to play a critical role in willed action, which they defined as (1) conscious

awareness and attention, (2) choice and control, and (3) intentionality. Several lines of

studies are cited by these authors that meet these criteria for willed action, and in these

studies the dlPFC has been found to be the primary contributor to this function.

Frith et al. (1991) used PET to study a motor task that required the subject to move

the first or second finger of the right hand at will in a random order, paced by touches to

the fingers made by the experimenter. This condition was compared to a control

condition that had the subject lift his/her finger after the experimenter touched it. Random

finger lifting was associated with significantly greater dlPFC activation, as compared to

the control condition. Another approach to examining willed action was to have subjects









make random movements on a joystick in one of four possible directions: up, down, left,

and right (Playford et al., 1992). Compared to rest condition, movements of the joystick

resulted in activation of the dlPFC.

The task of random number generation has also been used to study willed action

(Jahanshanhi et al., 1997). Random number generation involves operations characteristic

of willed action, such as the selection and maintenance of strategies, holding information

in attention, suppression of habitual counting, internally driven response generation, and

monitoring of responses. This task, as compared to counting, activated the right dlPFC, as

well as the right inferior PFC.

Transcranial magnetic stimulation (TMS) has been used in research as a transient

"lesion" model where it allows for the temporary disruption of neural processing during

focal stimulation of a local brain region. Ro et al. (1997) found that the latency increased

for volitional saccades made to a central arrowhead endogenouss go signal) that indicated

the location of the required response in the right or left visual field as a result of TMS

over the superior PFC. TMS over the superior PFC had no effect on the saccades

triggered by a peripheral asterisk (exogenous go signal) that marked the hemifield where

a response was required. This indicates that the effect was not a function of disrupting

visual tracking abilities due to diffuse effects of the TMS at the frontal eye fields.

Additionally, TMS over the parietal cortex had no effect on either the volitional or

triggered saccades. The transient "lesion" to the dorsal PFC region therefore resulted in

an interruption of the willed action of making the saccade.

Other groups have looked at the functioning of the dlPFC in the context of

cognitive control. Jonathan Cohen, Earl Miller, William Perlstein, and others have









conceptualized the role of the prefrontal cortex as being responsible for activities such as

internal representation, maintenance, and updating of contextual information in the

service of exerting control over thoughts and behavior, or "cognitive control" (Braver et

al., 2001; Miller and Cohen, 2001; Perlstein et al., 2002). Context is referred to as any

task-relevant information that is internally represented in a way that it can bias processing

in the pathways responsible for the performance of a task. Goal representations are one

form that this information can take, as they have influence on planning and cognitive

operations. Within this line of thinking, context is viewed as the subset of representations

within working memory that governs how other representations are used. These context

representations are thought to subserve both mnemonic and control functions. Support for

this conceptualization and the role of the dlPFC in context maintenance is obtained from

several different domains. Neuropsychological evidence comes in part from findings that

led to the development of the term "frontal syndrome," which refers to a particular

impairment in which the normal control over social and sexual behavior is dysregulated

(Stuss and Benson, 1986). Neuropsychological studies have demonstrated that patients

with PFC lesions show impairments on tasks involving cognitive control, such as the

Stroop test and the Wisconsin Card Sorting Test. Neurophysiological studies with non-

human primates have provided direct evidence of the dlPFC involvement with cognitive

control. In experiments such as the delayed task paradigm, neurons in the dlPFC have

been found to exhibit sustained, stimulus-specific activity during the delay periods of

simple tasks requiring the active maintenance of task relevant information (Fuster, 1989;

Goldman-Rakic, 1987). More recent neuroimaging studies have corroborated these

previous findings from the neuropsychological and neurophysiological literature. PFC









activity has been demonstrated during a wide range of tasks involving a cognitive control

component (Cabeza and Nyberg, 2000; Cohen et al., 1997; Perlstein et al., 2002).

Neuroimaging studies have also confirmed that the dlPFC is specifically involved in

active maintenance functions by demonstrating sustained activity in this region during the

maintenance period of working memory tasks (Braver and Cohen, 2001; Cohen et al.,

1997; Perlstein et al., 2002).

Significant evidence exists to suggest a role for both the dlPFC and OFC regions

in goal-directed behaviors. Notably, Perlstein et al. (2002) provided evidence that both

the dlPFC and OFC are sensitive to contextual motivational characteristics in which

higher-level cognitive tasks are performed. Overall, these two regions appear to subsume

slightly different aspects of goal-directed action. The OFC appears to be involved in

biasing volitional aspects of behavior by the underlying motivational state (Bechara et al.,

2000). The dlPFC, on the other hand, appears to play a role in the more conscious, or

salient, aspects of goal-directed behavior, such as keeping the context in mind through

active goal representation (Miller and Cohen, 2001).

Many studies have evaluated the role of the OFC and dlPFC in goal-directed

behavior, willed action and context maintenance, but the role of these two regions in the

improvement of episodic memory by use of explicit goal-setting has yet to be

systematically evaluated. Further, these two likely candidates for contributing to the

positive effects of goal-setting undergo the steepest rate of decline in older adults.

Specifically, the dlPFC and the OFC are thought by many to be the most susceptible

regions in the brain to neuronal degeneration with age (Band et al., 2001). Huttenlocher

(1979) found a 13% decrease of the synaptic density in Brodmann's area (BA) 46.









Ulyings et al. (2000; 2002) showed that age-related dendritic change in BA 9 and 46 are

large in the pyramidal cells of layer V. An approximate loss of 20% of layer V spines in

the OFC has been found (Band et al., 2001). Above the age of 65, a pronounced reduction

in cell size exists where, for example, there is a 25% reduction in cells outside of the

pyramidal layer of the OFC (Haug and Eggers, 1991). In light of these findings, the

potential for the beneficial effects of goal-setting in young adults extending to older

adults may appear dubious. However, as the next section will detail, enhancing goal-

directed action by goal-setting and other techniques does in fact help older adults in

cognitive tasks to an equal extent as young adults.

Effect of Enhanced Goal-directed Action on Cognition in Older Adults

Evidence suggests that older adults may have greater difficulties with maintaining

the context of a cognitive task (e.g., important features in generating successful responses

for the task). However, it has generally been found that improving the salience of those

features important for a task, or increasing the challenge to the individual by techniques

such as goal-setting, can compensate for decreased task performance.

De Jong (2001) asserted a goal-neglect hypothesis of age-related decrements in

cognitive control. He stated that decrements in cognitive control could be characterized

by a reduced capacity for goal selection and goal maintenance in working memory. In

particular, De Jong stated that under conditions of novelty or of weak environmental

stimulation (i.e., reduced environmental cues that support cognitive operations),

pronounced goal neglect results. Goal neglect is defined as "disregard of a task

requirement even if it has been understood, resulting in a mismatch between what is

known about task requirements and what can be done in principle, and what is actually

attempted in behavior" (De Jong, 2001, p. 71). Evidence indicates that older adults tend









to be more dependent on and more sensitive to means for external support offered by the

context of the task so that they might compensate for a reduced capacity for cognitive

control (Hultsch et al., 1987).

Evidence for the goal-neglect hypothesis comes in part from task-switching

paradigms. In task-switching paradigms, the task to be performed on each trial is selected

by the subject from a set of alternative tasks. The tasks are presented in an unpredictable

order, and each trial starts with the presentation of a cue that signals the task to be

performed. The cue is followed by either a fixed or random delay, which is called the

preparation interval. Then the stimulus is presented in which the subject must perform the

task. This stimulus is typically ambiguous as to which task is to be performed, rendering

it necessary to keep track of the task sequence and/or process the cue effectively (De

Jong, 2001). De Jong (2000) found that older adults experienced a failure to engage in

advance preparation in task-switching in which he proposed an intention-activation

account. According to this account, the effective utilization of advanced preparation

depends on two components. The first component is an explicit goal or intention to be

added to the basic goal structure that governs performance in the task-switching

paradigm. The second required component is the retrieval and carrying out of this

intention at the proper time. Success of the intention retrieval is thought to depend on the

activation level of the intention and the characteristics or triggering power of the retrieval

cue. Therefore, De Jong asserted that the frequent failures to engage in advance

preparation in older adults may be due to low levels of intention activation, which would

reflect goal neglect. Interestingly, the number of errors committed by older adults in the

task-switching paradigm can be reduced to the level observed in young adults by straight









forward speed manipulations. Specifically, if the delay period between the cue and the

stimulus to perform the task is shortened, older adults perform as well as young adults.

By shortening the time delay, there was an increase in challenge for older adults.

According to De Jong, this increase in challenge resulted in older adults being better able

to utilize available control capabilities in order to optimize performance. Importantly, the

increase in challenge decreased the possibility of goal neglect (De Jong, 2001).

Several studies have investigated the use of explicit goals to improve performance

on a cognitive task in young and older adults. That is, a specific goal for performance

achievement on a cognitive task is provided, and then the impact of the explicit goal on

performance is measured. According to De Jong's goal-neglect hypothesis of older

adults, the use of an explicit goal would provide important exogenous environmental

structure that would serve to help older adults compensate for a reduced capacity for goal

selection and goal maintenance during cognitive tasks. One study tested young and older

adults in a free-recall task in which one subgroup established a performance goal for

blocks of trials and received feedback on a trial-by-trial basis. The other subgroup neither

established goals nor received feedback (Stadtlander and Coyne, 1990). Memory for

random letter strings of 5 letters in 3 blocks of 50 trials was measured. The use of the

motivational technique of explicit goal-setting and feedback increased memory

performance in both young and older adults above that of the no-goal condition.

West et al. (2001; 2003) reported similar findings, showing that an increase in

challenge to older adults by means of explicit goals significantly increases performance

in a memory task. These studies examined the impact of goal-setting on memory and

memory beliefs across adult age groups. In one study (West et al., 2001), a baseline









memory trial was administered, followed by three additional recall trials. All four

memory trials entailed memorization of a grocery item list and subsequent free recall of

the items. Young and older adults were placed in one of three conditions: goal-setting,

goal-setting with feedback, or no goals. Goal-setting was initiated after the baseline trial.

An assessment of memory beliefs, self-efficacy, and motivation was performed in

conjunction with memory performance measurement. As was expected, young adults

remembered a significantly greater number of items from the word list than did older

adults across trials. Goal-setting significantly improved memory performance in both

young and older adults. The increase in challenge on the memory task provided by the

explicit goal resulted in equal memory performance enhancement for young and older

adults. Additionally, motivation and self-efficacy were both positively affected by goal-

setting. The results for the goal-setting plus feedback group were mixed. The variable

outcome is thought to be related to the differing performance outcomes for each subject

and the subsequent individual feedback that they received. For instance, subjects not

meeting their goal would receive negative feedback and often would have a poorer

performance in the subsequent trial, particularly in the older adult group. Overall, the

primary finding of this study was that performance was positively affected by goal-

setting in both age groups, and self-efficacy and motivation were also higher after goal-

setting.

In another study by West et al. (in press), two different experiments investigated

different levels of goal-setting/challenge in older and young adults in a similar shopping

list recall task. In both experiments, older and young adults completed a baseline

shopping list recall task to begin. Three more shopping list recall trials were completed









following the baseline. For goal conditions in both experiments, subjects were given a

specific recall goal based on their own prior performance prior to each of the trials

(excluding the baseline trial). Additionally, all subjects in goal conditions received

positive feedback for their memory gains over trials. In the first experiment, subjects

were assigned a low-challenge or high-challenge goal. Both young and older adults

clearly benefited from the higher challenge goal. Further, older adults experienced an

equal degree of memory performance improvement from the higher challenge goal

condition as did young adults.

In the second experiment of this same study, a moderate challenge goal condition

was compared to a no-goal condition. Results indicated that performance gains in the

goal-setting condition exceeded the gains in the control group for both young and older

adults. The average score gain per trial in the control condition (no goals) was 4.6 words

for young adults and 2.1 words for older adults. In the goal-setting condition, the average

score gain for young adults was 5.6 words and 4.0 words for older adults. These score

gain increases were significant for both young and older adults. Further, a significant

difference was not found between young and older adults in these positive effects of goal-

setting.

Summary and Predictions

Evidence clearly indicates that older and young adults can benefit from explicit

goal-setting in episodic memory tasks. The OFC and dlPFC are likely contributors to the

improvement of episodic memory performance associated with explicit goal-setting, but

this has yet to be tested. As noted previously, the two regions of the brain demonstrating

the steepest trajectory of decline in aging are the OFC and the dlPFC (Band et al., 2001).

This is of substantial interest because if these two regions do in fact correlate with the









performance improving effects of goal-setting, then it would be important to determine if

the OFC and dlPFC sustained equal activation increases in young and older adults as a

result of goal-setting. Perhaps activity in the OFC and dlPFC in older adults increases to a

greater extent than in young adults in order to receive equal benefit from explicit goal-

setting, compensating for the age-associated atrophy in these two regions. Alternatively,

compensation could be achieved by greater activation in regions involved with

motivation and/or mnemonic processes in older adults in order to generate equal benefit

from goal-setting conditions as young adults.

In order to investigate these possibilities, a pilot study was performed in which 20

older adults and 20 young adults were recruited to undergo fMRI to examine the neural

substrates that underlie the memory enhancing effects of goal-setting. Both young and

older adults were assigned to one of two groups: a goal-setting group and a no-goal

group. Four memory trials were conducted with brain-related encoding activity being

measured in the last three trials by fMRI. Activations of the left PFC and the temporal

lobe, bilaterally, were anticipated during the verbal memory encoding task. Activation in

these two regions was predicted to be dampened in older adults, corresponding to the

decreased memory performance anticipated for older adults. Goal-setting was

hypothesized to increase performance in both young and older adults. It was

hypothesized that gains in memory performance through goal-setting would be coupled

with increased activation in the OFC and dlPFC, among other regions in the PFC. In

older adults, compensation for the decline in these two regions was hypothesized by

means of greater activation in the OFC and dlPFC and/or greater activation in other

important motivation or mnemonic-related regions.











CHAPTER 2
METHODS

Overview

A between group design was used with four groups: young adult/goal, young

adult/no-goal, older adult/goal, and older adult/no-goal. Subjects performed a memory

task that entailed studying a list of grocery items that they were later asked to recall.

Brain activity during encoding was measured by fMRI and behavioral performance was

measured by recall accuracy. Subjects also performed a motor task that required a button

press in response to a visual stimulus in order to assess for the possibility of a generalized

lesser magnitude fMRI hemodynamic response in older adults. Motor response time was

measured in the motor task. It should be noted that a fixed-effects statistical approach

was used for the behavioral and fMRI data due to power considerations, thus limiting the

ability to generalize findings.

General Methods

Subjects

Subjects were 20 young adults (ages 18-28) and 20 older adults (ages 60-70) with

members of each age group pseudo-randomly assigned to one of two goal groups (goal,

no-goal). Young adult subjects were primarily undergraduate and graduate students at the

University of Florida. Older adult subjects were high-functioning, community-dwelling

individuals. Young adults were matched to older adults in education and self-rated health

(see Table 2-1). Self-rated health was assessed by asking the subject to circle a number

on a scale of 1 to 10, 1 being excellent health and 10 being very poor health, indicating

how healthy they are in general. Older and young adults were administered the Shipley









vocabulary test (Shipley, 1940; Appendix A), on which older adults scored significantly

better than young adults, t (1,38) = 7.6, p = .009 (Table 2-1).

Exclusion Criteria

All subjects were right-handed. Subjects were excluded if they reported any history

of neurological illness (including strokes or traumatic brain injury) or psychiatric illness.

Subjects were also excluded if taking psychoactive or anticholinergic medications.

Subjects taking blood pressure medication were not excluded if their blood pressure had

been stable for the previous 6 months while taking the medication. Subjects having a

history of substance abuse or previous treatment for substance abuse were excluded.

Participants with a history of epilepsy or a seizure disorder were excluded. Subjects were

administered the Telephone Interview for Cognitive Status (TICS; Brandt et al., 1988)

and were excluded if their score reflected the possible presence of a cognitive

impairments (as defined by a score of < 15th percentile). Subjects were also excluded for

any MRI environment contraindications (e.g., cardiac pacemaker, implanted cardiac

defibrillator, aneurysm clip, claustrophobia, non-removable ferromagnetic dental work

such as bridges, etc.). Lastly, subjects were not permitted to ingest caffeine or nicotine

within a 60 minute time period prior to the MRI session. This criteria reduced the risk of

confounded cerebral blood flow measurements by allowing time for at least partial

clearance of any potentially high levels of circulating caffeine or nicotine, which are

vasoactive agents (caffeine half life = 3.5 hours and nicotine half life = 2 hours).

High resolution Tl weighted structural MRI scans were used to determine the

presence of any structural abnormalities. One older adult was excluded after a significant

lesion was identified.











Table 2-1 Mean (Standard Error) Demographic Characteristics of Experimental
Participants.
Young Adults Older Adults
N 20 20
Age 22.3 (0.62) 64.8 (0.55)
Age Range 18-28 60-70
Sex (Men/Women) 9/11 6/14
Education 15.6 (0.41) 15.5 (0.69)
Self-Reported Health 2.55 (0.33) 2.40 (0.35)
Vocabulary 33.0 (0.56) 35.2 (0.60)
aOlder adults significantly greater than young adults: t (1,38) = 7.6, p = .009.


Three older adults and two young adults were excluded due to image distortion that

was caused by an inability to position the head far enough into the scanner bore

longitudinally to be within the optimal image acquisition range. On account of the

positioning difficulty, images were subjected to large field inhomogeneity and distortion.

One older adult was excluded due to a claustrophobic reaction to the scanner. Lastly,

collection of data from one subject was terminated in the middle of testing due to an MRI

schedule conflict.

Experimental Task and Procedures

Participants performed a free-recall episodic memory task modeled after that

employed by West (2001; 2002). Word lists, comprised of categorizable grocery items,

were presented.

This task was originally developed as follows. In order to create a large pool of

categorizable items, four research assistants listed all items they found in grocery stores

and also identified a large number of narrowly defined categories for those items. Nine

independent raters then determined the most suitable category for each item and rated

each item as a high, moderate, or low frequency exemplar of its category. In order for an









item to be included in the final word list, the word must have met a predetermined

criterion rating such that seven of nine raters rated the item as a "high frequency"

exemplar of its category (West and Thorn, 2001).

The episodic memory task was modified from its original form for administration

in the imaging context. Four memory trials were performed (Baseline Trial and Trials 1

through 3). All four groups (young adult/goal, young adult/no-goal, older adult/goal,

older adult/no-goal) were administered identical memory trials and word lists.

The Baseline Trial was administered prior to entering the scanner and was

comprised of a 15-word list presented on a single sheet of paper (Appendix B).

Participants were given a one-minute study period, and then were prompted to write out

all the words that they could remember on a numbered sheet of paper. Instructions given

to the participant were as follows: "On this task, you will be asked to study a list of items

that can be bought at a store. You are not expected to remember every item on the list.

Just do your best. I will tell you when to begin studying the list and I will also let you

know when your study time has ended. You will be given approximately 1 to 3 minutes

to study the list. You may not write during the study time. After the study period is done,

I will ask you to write down all of the items that you remember from that list. You will

write the remembered items on this page."

Following the Baseline Trial, participants were placed in the MR scanner and

performed the remaining three memory trials (Trials 1 through 3). Memory trials were

computer controlled by the software package PsyScope (Cohen et al., 1993). Participants

were presented with the identical 42-item grocery list for Trials 1 through 3 (Appendix

B). Included in the 42-item list were all 15 items contained in the Baseline Trial list. For









each of the three trials, word-list content of the 7 blocks was identical. However, the

order of presentation for the 6 words within each block was randomized for each trial.

The word content of each of the three trials was identical in order to observe continued

improvement across trials in conjunction with the goal-setting condition. The content of

the word list is detailed in Appendix B. Instructions given prior to Trial 1 were as

follows: "In these next activities, you will again be asked to study a list of items that can

be bought at a store, except this time they will be presented on a screen. After an item is

presented, there will sometimes be a short delay before another item is presented on the

screen, and other times there will be a little longer delay. During these delays, you are

asked to attend to the target in the center of the screen. When all the items have been

presented, I will ask you to tell me back as many as you can remember. You will not be

expected to remember all of them." Instructions prior to Trials 2 and 3 were not stated

vocally but appeared on a liquid crystal display (LCD) screen (described below). The

instructions read: "Study the following list of words." The vocal recall of word lists by

the subjects and task instruction delivery were achieved by a bi-directional in-scanner

microphone and receiver.

The 42-word trials were presented singly in 6-word blocks, for a total of 7 blocks

per trial (see Figure 2-1). Words were presented singly for a 2 s duration with a 2 s

interstimulus interval (ISI). Subjects were presented the words through a mirror system

orientated towards an LCD screen mounted in the radio frequency coil (RF) 4 in. behind

the top of the subjects' head (approximate visual angle = 41 degrees). All 6-word blocks

lasted a total duration of 24 s, followed by a 10 s interblock interval. During the

interblock interval, subjects were asked to attend to a fixation point ("+") in the middle of









the screen. All three trials were initiated by a 16 s baseline fixation period during which

the subjects were asked to attend to a fixation point ("+") in the middle of the screen.

Image acquisition began at the onset of the baseline fixation period. The first 2 image

volumes were automatically discarded (TR=2 s), so the first image that was collected for

later analysis was the 3rd image volume. Consequently, 12 s (6 image volumes) of the 16

s baseline fixation period were analyzed. After the seventh block of words, there was a 16

s recovery fixation period when the subject was again asked to attend to the fixation

point. Image acquisition terminated at the conclusion of the recovery period. Free recall

of the 42-item list began immediately following the recovery fixation period and subjects

were given a maximum of 4 min for recall. During the recall period, subjects verbally

recalled as many of the 42 items as they could remember, in any order. Vocal responses

were digitally recorded and immediately scored for accuracy. Image volumes were not

collected during the recall period. In total, 133 images were collected per trial, yielding

399 images for each subject.

All groups performed the above described memory task. The only procedural

difference was between the goal and no-goal groups, which was implemented between

the memory trials. Prior to Trials 1, 2 and 3, subjects in the goal group received feedback

regarding the number of items they remembered correctly from the previous trial in

addition to a goal statement for that trial: "Your goal is to achieve a 50% improvement in

your score." Prior to Trials 2 and Trial 3, subjects in the goal groups also received a

positive feedback statement (e.g., "Good job that's a great score") and after the goal

statement, "Keep trying" was written. Following the final trial, Trial 3, these statements

and feedback were given again. All statements and feedback were provided on the LCD













/ Recall




Recovery Fixation Period
16s

/I nterblock Interval
10 s 6-word block & interblock
Word Block interval are repeated 7 times
2 s presentation duration for a total of 42 grocery
per word, 2 s ISI, 24 s total items per trial
/Baseline Fixation Period
16s


Figure 2-1. Graphical representation of the memory task for Trials 1 through 3.

screen. Subjects alerted the experimenter verbally when they completed reading each

screen. Subjects in the no-goal group did not receive any of the feedback described

above. Feedback was not included for the goal group in order to maintain paradigmatic

consistency with previous studies finding improved performance from goal-setting (West

et al., 2002, 2003). However, communications with the goal group were equated in the

no-goal group by inserting the statements: "You have completed trial number x" and "We

are now ready to begin the next trial."

The Baseline Trial contained only 15 words, as compared to 42 words in the

following 3 trials, so that subjects would not be overwhelmed by the task (as was

observed in pilot versions of this paradigm). Also, the initial goal of 50% improvement

would be more easily attainable, again reducing the risk of overwhelming the subject at

the onset of the memory trials.









All experimentation was performed during one visit to the MRI facility. The first

portion of experimentation took place in a testing room inside the MRI facility (although

outside the scanning environment). To begin, informed consent was obtained from the

participant in a manner consistent with the University of Florida Institutional Review

Board regulations. Subjects then completed several questionnaires that inquired about

their opinions about their memory, demographic information, among other similar

information (Appendix C). Then the Baseline Trial was conducted with recall

immediately following. Following the baseline trial, another questionnaire was

administered, which was also related to the participants' opinion of their memory.

The next procedure varied depending on whether the subject was in the older adult

group or the young adult group. Because all 15 words from the Baseline Trial were

included in subsequent memory trials, it was important that a consistent time frame be

maintained between the administration of the Baseline Trial and the commencement of

Trial 1 for all participants. (It should be noted that the subjects were not made aware that

the exact same words would be repeated in subsequent trials). The older adults, on

average, require a greater amount of time to complete the questionnaires following the

Baseline trial. Thus, a temporal window of 30 40 minutes between the completion of the

Baseline Trial and the start of Trial 1 was maintained for both young and older adults by

giving the young adults a participant information form and the vocabulary test to

complete before proceeding to the MRI scanner. Older adults, on the other hand,

proceeded directly to the MRI scanner after completion of the initial questionnaire.

(Older adults instead completed the vocabulary test and participant information sheet

after the MRI testing was complete.) Consequently, the time between the Baseline Trial









and Trial 1 were matched between young and older adults in temporal length and activity,

as both were completing questionnaires. After completing their respective questionnaires,

subjects were then prepared for the scanning environment and placed in the scanner bore.

After localizing sequences were performed, Trials 1 through 3 were administered.

After completion of the three memory trials in the scanner, subjects then

completed a motor task consisting of pressing the right index finger button of a Button

Response Unit (BRU) every time a large white square appears on the LCD screen. The

stimulus duration was is followed by a fixation period of 9 s. Each motor task session

contained 9 trials and there were a total of 3 motor task sessions completed (e.g., 27 total

motor responses). This task served as an internal activation standard (described in further

detail shortly).

Subjects then completed 5 more questionnaires following the scanning portion of

the experiment that again asked for their opinions about their memory and also how they

felt they performed on the memory task. An additional questionnaire asked about

strategies utilized to complete the memory task (Appendix D).

Behavioral Data Analysis

The percent of items recalled from each 6-word block served as the dependent

variable. The Baseline Trial was excluded because it was conducted outside the scanner

and procedures for the goal and no-goal groups were identical. Due to power

considerations, a fixed-effects analysis was conducted. The error variance term was

estimated on a block by block basis with each block representing an independent

observation. The percent of items recalled for each block was evaluated by analysis of

variance (ANOVA) with Group (young adults, older adults) and Condition (goal, no-









goal) as the between-subjects factors and trial (trials 1 3) and block (blocks 1 -7) as

within-subjects factors.

Magnetic Resonance (MR) Acquisition

Scanning was conducted in a Siemens Allegra 3.0 Tesla head-only MR

superconducting system (128 MHz; 60cm bore) at the University of Florida McKnight

Brain Institute MR Facility. Images were acquired using a Siemens quadrature head radio

frequency (RF) coil. A BOLD sensitive echo-planar imaging pulse sequence (EPI;

Siemens; TR=2000 ms, TE=30 ms, FOV=240 mm, flip angle=900, 64 x 64 matrix) was

used to acquire 31 slices (voxels = 3.75 mm2 in-plane, 3.2 mm thick, 0.32 mm gap) in the

axial oblique plane. A 3-plane localizer was first acquired onto which a sagigittal scout

was prescribed. The prescription was acquired with 31 contiguous slices perpendicular to

the anterior commissure-posterior commissure line. A double-oblique prescription was

collected in order to reduce potential image misregistration across subjects due to

differences in ventral head orientation. Functional scanning was synchronized to trial

onset (baseline fixation period) and terminated at the end of the recovery fixation period.

Following functional scanning, structural images were acquired with a 3-D

magnetization-prepared rapid acquisition gradient echo (MPRAGE) Ti-weighted pulse

sequence (128 slices, 1.3 mm slice thickness, TE=4.13 ms, FOV=240 mm, flip angle=80,

512 x 512 matrix).

It is important to note that an area of specific interest in this study, the orbital

frontal cortex, contains a high potential for signal drop off, especially in the high

magnetic field strength of 3 tesla. The orbito frontal cortex borders the orbital sinus and

the auditory meatus, creating susceptibility artifacts at the tissue-air interfaces. Indeed,









signal loss was observed in the most ventral and anterior portions of the OFC, as pictured

in Appendix E.

FMRT Data Analysis

Data analysis, registration and visualization were performed with the fMRI

software package BrainVoyager 2000 (Brain Innovation, Maastricht, The Netherlands).

FMRT data reduction. The first two volumes (4 s) of each functional scanning run

were automatically discarded by the Siemens scanner to allow for Tl equilibrium and

thus were not included in any of the analyses. Prior to data analysis, functional images

were aligned to the last volume for each slice in order to minimize the signal changes

related to rigid body rotation and translation during the acquisition. Following movement

correction, images were spatially smoothed with a Gaussian kernel, FWHM = 8 mm, to

accommodate for differences in anatomy. Three-dimensional motion correction and

Talairach transformation were performed for the functional data of each subject. Linear

drifts of the signal with respect to time were removed from each pixel's time course.

The 3-D anatomical volumes and 2-D functional volumes were resliced to a 256 x

256 matrix size. Reslicing of functional volumes took place before statistical analysis.

The functional volumes were analyzed in 3-D space. Statistical maps were superimposed

onto 3-D anatomical data sets. Since the EPI functional scans and 3-D structural

measurements were performed within the same recording session and contained the exact

same positioning parameters, co-registration of the respective data sets were performed

semi-automatically based on the Siemens slice position parameters of the T2*-weighted

measurement (number of slices, slice thickness, distance factor, pitch angle (axial -

coronal angle), FOV, shift mean, off-centre read, off-centre phase, in plane resolution)

and on parameters of the Ti-weighted 3-D measurement (number of sagittal partitions,









shift mean, off-centre read, off-centre phase, resolution) with respect to the initial

overview measurement (prescription). Manual co-registering corrections in the x, y, and z

planes and in the pitch angle were required to optimize final alignment.

For each subject, the structural 3-D data sets were transformed into Talairach space

using a two step process. The first step consisted of rotating the 3-D data set of each

subject to be aligned with the stereotaxic atlas. For this step the location of the anterior

commissure (AC), the posterior commissure (PC), and two rotation parameters for

midsagittal alignment had to be specified manually in the high resolution 3-D

volumetrics. In the second step the extreme points of the cerebrum were specified. These

points together with the AC and PC coordinates were then used to scale the 3-D data sets

into the dimensions of the standard brain of the Talairach and Tournaux atlas (Talairach

and Tournaux, 1988) using a piecewise affine and continuous transformation for each of

the 12 defined cerebral borders.

FMRT data analyses. Statistical analyses were performed using BrainVoyager by

fitting a general linear model to the individual fMRI time series data (e.g., multiple

regression analysis). The predictor variables were created for each condition creating the

idealized time series that represented the response to the condition (e.g. encoding period

or motor response) in each group. The hemodynamic response for predictor variables was

estimated by convolving each regressor with a standard gamma variate function that

shifted the estimated hemodynamic response approximately 6 seconds to account for the

expected delay. A weighted sum of the predictor variables was created that produced the

closest match to the actual data time series. A parameter estimate (e.g., beta weight) was

generated for each model that estimated the strength of covariance between the actual









data and the modeled hemodynamic response function. Parameter estimates were

calculated for each participant, which were then subjected to group analysis.

For all analyses, statistical maps generated for each pattern of interest were

thresholded for significance using a cluster-size algorithm (Forman et al., 1995) of 80

voxels, which protects against an inflation of a false-positive rate with multiple

comparisons. The voxel size was 0.9375 mm2 in plane and 1.3 mm thick. Due to power

considerations, a fixed-effect analysis approach was utilized, which estimates the error

variance on a scan to scan basis. Significant effects are shown only if the associated P

valued yielded P < 0.01 (Bonferroni corrected for multiple comparisons using the number

of voxels exceeding the minimum in brain threshold signal criteria intensity of 250).

It is important to note that relative differences between groups reflect differential

activation intensity for a specific prescribed area, and do not indicate a greater or lesser

spatial extent of activation. That is, relative differences of activation asserted to be

present between groups reflect a significantly greater or lesser signal intensity measured

for a discrete group of voxels.

Internal activation standard

In order to establish that differential activation observed in older adults is not

simply a generalized change in neural activation, an "internal activation standard"

(Weinberger et al., 1996) was utilized. The blood oxygen level dependent (BOLD) signal

of fMRI depends on neurovascular coupling, which is a process in which neural activity

influences the hemodynamic properties of the surrounding vasculature. There could be

direct changes in the cerebral vasculature as well as alterations in the complex

neurochemical transformation of neural activity into changes in blood flow that might

affect the measured BOLD response (D'Esposito, 2003). Age-associated signal









differences have not been demonstrated in components of the BOLD hemodynamic

response function in the motor cortex in response to a buttonpress (D'Esposito et al.,

1999). Thus, the neural activation in the precentral gyrus, Brodmann's area 4 (BA 4), in

response to subjects' buttonpress was utilized as the internal activation standard. Absence

of age-related BA 4 signal intensity differences supports the specificity of any differences

observed between young and older adults in the memory task.

The general linear model (GLM) of the experiment was computed from the 40 (40

subjects; 3 motor trials per subject collapsed) z-normalized volume time courses. For

each time course, the 1 s period of stimulus presentation (white square to which the

subject responds with a button press) was defined to represent motor activity. The signal

values during this phase were considered the effects of interest. Response latencies in

older adults (median = 404 ms, standard error = 24 ms) and young adults (median = 397

ms, standard error = 22 ms) occurred well within the temporal frame of the defined

predictor response of the GLM model. The baseline fixation period (9 s following the

motor stimulus cue) was defined as the baseline (non-motor) period. For each trial, 90

volumes were collected. There were a total of three trials so 270 volumes were collected

for each subject. The resulting predictor was obtained by shifting an ideal box-car

response (assuming a value of 1 for the volumes of the respective motor periods and a

value of 0 for the baseline periods) by a standard gamma variate function prescribed by

Brainvoyager (in order to account for the hemodynamic delay). This predictor, which was

identical for all subjects across all groups, was used to build the design matrix of the

experiment (see Figure 2-2).

The above steps generated a 4-D functional time series (volume time course: 3 x









..... ....................................












1' I D.













Figure 2-2. GLM predictor model for signal intensity increases corresponding to the
motor response of a button press for a single subject. The model represented
here was identical for all subjects. The green shading represents the presence
of the predictor, in this case a motor response, and the gray area indicates the
rest period.

space, 1 x time) for the predictors of the model. Statistical analysis of 4-D functional time

series included first single-subject multiple regression analysis, followed by multi-subject

multiple regression analysis that concatenated the single-subject analyses. The global

level of the signal time course in each session was considered to be a confounding effect

and was entered as such into the GLM model. A fixed effects analysis was employed.

Statistical maps generated from the multi-subject analyses were projected on the flattened

surface of a volumetric rendering of all 40 subjects' high resolution 3-D volumes

averaged together in Talairach space. Three statistical maps were generated for the motor









predictor model: 1) young adults, 2) older adults, and 3) young adults directly compared

to older adults (contrast map).

Button press response latencies to the visual cue were measured in all subjects. A

comparison between young and older adult response latencies was conducted by first

determining the median response time for each subject. A between subjects (young and

older adults) analysis of variance was conducted on the median reaction time for each

subj ect.

The motor task was employed for an additional analysis to evaluate the possibility

of an altered hemodynamic response in older adults taking blood pressure medication.

The hemodynamic response during the motor task in the 6 older adults taking blood

pressure medication and 6 pseudo-randomly selected older adults not taking blood

pressure medication were compared.

Memory Encoding Experiment

Data were analyzed separately for encoding and for relation of encoding to

blockwise memory recall performance.

Encoding-related activity. A general linear model (GLM) for the experiment was

computed from the 40 (40 subjects; 3 memory trials per subject collapsed) z-normalized

volume time courses. For each time course, the 24 s period of word presentation was

defined to represent encoding. The signal values during these phases were considered

effects of interest. The baseline fixation period (12 s), interblock intervals (10 s), and

recovery fixation period (16 s) were defined as the rest (non-encoding) period. The

resulting predictor was obtained by shifting an ideal box-car response (assuming a value

of 1 for the volumes of the respective encoding phases and a value of 0 for the baseline

time points) by a standard gamma variate function prescribed by Brainvoyager (in order









to account for the hemodynamic delay and form approximate hemodynamic rise and fall

times). This predictor, which was identical for all subjects across all groups, was used to

build the design matrix of the experiment (see Figure 2-3).

























n Ig- ~ I r Mftr
A!!2 Mn 1_1_ -Fm.

Figure 2-3. GLM model for the multiple regression analysis of overall memory encoding
for a single subject. The model represented here was identical for all subjects.
The green shading represents the presence of the predictor, in this case
encoding activity, and the gray area indicates the rest period.

Subsequent memory activity. The above model reflected overall neural activity

corresponding to encoding the 6-word blocks. In order to evaluate how encoding activity

related to subjects' recall performance, an additional GLM model was used that

correlated voxel signal intensity with recall performance on each of the 6-word blocks.

This approach has become known as the 'subsequent memory' effects approach to

analyzing encoding (Rugg et al., 2002). In the subsequent memory procedure, event-

related activity elicited by a series of study items is contrasted according to the number of









items remembered or forgotten on a subsequent memory test. The assumption behind this

analysis is that differences in activity that predict successful versus unsuccessful memory

reflect the different levels of engagement of processes supporting effective encoding. As

an example of this approach, a subject may have remembered 5 of 6 words on the first

block of a trial, so .83 (83% correct) is entered for each of the 12 volumes (covering the

24 s period when the 6 words were presented) for the encoding predictor. Likewise, recall

performance values are entered into the encoding predictor for each of the following 6-

word blocks. The resulting predictor for the GLM model represents higher signal

intensity for blocks in which the subject performed well, and lower signal intensity for

blocks in which the subject performed poorly. In other words, the model represents a

correlation of recall performance for each 6-word block and signal intensity increases

above baseline during encoding periods. Encoding time periods are thus set as relative

increases in signal intensity corresponding to recall performance for each respective 6-

word block and are corrected for the hemodynamic response delay (see Figure 2-4).

For both the encoding activity model and the subsequent memory activity model,

the global level of the signal time course in each session was considered to be a

confounding effect and was entered as such into the GLM model. A fixed effects analysis

was employed. The creation of other alternative GLM models further evaluating

encoding was constrained by software limitations. For instance, creation of a model

investigating purely recall performance correlation with signal intensity was not possible

because it was necessary to include rest periods into the model.

The above steps generated a 4-D functional time series (volume time course: 3 x

space, 1 x time) for the predictors of the 2 models. Statistical analysis of 4-D functional


































-Isulll d 2 a1 I O a-ll n.a-p lll. .. (Rem be D-:) | gprrcmdei aure .d | |6- 5N2 S BBB SOPM
Figure 2-4. GLM model of the multiple regression analysis of memory recall
performance correlated with signal intensity. The green shading represents the
presence of the predictor, in this case encoding activity correlated with
subsequent memory performance, and the gray area indicates the rest period.
The illustrated model is an exemplar for a single subject. A separate model
was calculated for each subject that was based on their individual recall
performance during each of the 21 blocks.

time series included first single-subject multiple regression analysis, followed by multi-

subject multiple regression analysis that concatenated the single-subject analyses.

Statistical maps generated from the multi-subject analysis were projected on the flattened

surface of a volumetric rendering of all 40 subjects' high resolution 3-D volumes

averaged together in Talairach space. Seven statistical maps were generated for each

model (e.g., encoding activity model and subsequent memory activity model): 1) young

adult/no-goal, 2) older adult/no-goal, 3) young adult/no-goal compared directly to older

adult/no-goal (contrast map), 4) goal group, 5) no-goal group, 6) goal group compared







55


directly to the no-goal group (contrast map), and 7) interaction of goal-setting group by

age.











CHAPTER 3
RESULTS

Behavioral Performance

Encoding accuracy was scored as the percent of words correctly recalled for each

block of words (21 blocks per subject). Data were analyzed with a fixed-effects ANOVA

with between-subjects factors of age (young adults, older adults), goal (goal, no-goal) and

within subjects factors of trial (trials 1 through 3) and block (blocks 1 through 7). Group

means and standard errors are reported in Table 3-1. There was a main effect of age:

F[1,756] = 7.31, P = 0.007, Cohen's d= 0.16, reflecting better recall memory

performance in young adults than in the older adults (see Figure 3-1). There was a main

effect of goal-setting: F[1,756] = 9.13, P = 0.003, Cohen's d= 0.18, reflecting better

performance in the goal than no-goal group (see Figure 3-1). A main effect of trial,

F[2,756] = 77.33, P < 0.001, reflected subjects being able to perform progressively better

on each successive trial, as lists were identical in content for each trial (see Figure 3-1).

There was a main effect of block: F[6,756] = 12.16, P < 0.001, reflecting subjects'

tendency to recall words better in the earlier than later blocks (e.g., primacy effects; see

Figure 3-1). Notably, there was not a significant interaction between goal-setting and age

groups: F[1,756] = 0.652, P = 0.420, indicating that neither age group benefited

disproportionately from the provision of goals. Interestingly, the effect of goal-setting in

older adults brought their performance up to levels of performance in young adults

without goal-setting.

Following testing, subjects reported all the strategies they used for performing the

memory task. The number of strategies used by subjects was analyzed using an ANOVA







57


with factors of age (young adults, older adults) and goal (goal, no-goal). A main effect of


goal was not found, as subjects in each group used, on average, the same number of


Table 3-1. Mean Percent (Standard Error) of Recall Performance for Each Group


Goal-setting
Cio.il


No-Goal


Toi.al


Age
Oldci Adilhli

Told
Older Adults
Young Adults
Total
Oldci Adiulhs

Toiald


a. 58 b.
65
56
YOungY 55

52 45

Older 35
C) 48

46 25
1 2 3
44
No-goal Goal Trial

c.
70

60

50

S40 -

30
1 2 3 4 5 6 7
Block



Figure 3-1. Summary of memory recall percent in a) young adult/goal, young adult/no-
goal, older adult/goal, and older adult/no-goal groups, b) memory trials, and c)
blocks within each trial. Standard error bars are shown.

strategies to approach the task. However, a trend was observed for young adults to use a


greater number of strategies in performing the memory task than older adults, F(1,36) =


Mean
4-'.' 5 I -)
;5 1 I

46.2 (1.7)
49.0 (1.7)
47.6 (1.2)

49" 9 I 21









3.53, P = 0.068. Young adults used a mean of 7.0 strategies (standard error = 0.4), as

compared to 6.4 (standard error = 0.3) strategies used by older adults. Follow-up analysis

revealed a significant difference between the young adult/no-goal (mean = 7.8) and the

older adult/no-goal (mean = 6.2) groups, (Student's t-test, t(1,19) = 13.9, P < 0.001,

Cohen's d= 0.54).

Functional MRI Findings

Internal Activation Standard

The motor task was employed to assess for generalized lesser-magnitude signal

increase in older adults. The statistical threshold that was used for the memory study

analyses (P > 0.01, Bonferroni corrected for multiple comparisons; minimum threshold

of 80 contiguous voxels) was used for the motor task study in order to facilitate cross

study comparison of potential lesser-magnitude signal increases in older adults.

Activation in the left precentral gyrus (BA 4) was selectively examined as the task

involved the contralateral right index finger response. Left BA 4 activation in young

adults (center of gravity: x = -41, y = -17, z = 48; 3241voxels; center of gravity refers to

the geographical center of a significantly activated cluster, independent of the relative

statistical significance magnitude) and older adults (center of gravity: x = -41, y = -17, z

= 46; 2502 voxels) was significant (see Figure 3-2; P > 0.01, Bonferroni corrected for

multiple comparisons; minimum threshold of 80 contiguous voxels). A follow-up

analysis, which directly compared activation in young adults to older adults, revealed no

significant differences between young and older adults in BA 4 (see Figure 3-1; P > 0.01,

Bonferroni corrected for multiple comparisons; minimum threshold of 80 contiguous

voxels). Figure 3-3 illustrates the hemodynamic response to the button press in young and

older adults. A group by linear trend over scan analysis of signal intensity was not









significant, suggesting that young and older adults did not differ significantly in the

hemodynamic response curve characteristics (group x linear trend over scan analysis:

F(1,38)= 0.08, P = 0.785).

A random effects analysis of the median reaction time of each subject comparing

young and older adults revealed that there was not a significant difference in the time

taken for the button press response after the visual cue (F[1,38] = 0.07, P = 0.79). The

response latency median values were 404 ms (standard error = 24 ms) for older adults

and 397 ms (standard error = 22 ms) for young adults.

In order to address concerns regarding the possibility of altered hemodynamic

response in older adults taking blood pressure medication, their BOLD response during

the motor task was compared to 6 pseudo-randomly selected older adults not taking blood

pressure medication. There was significant activation in the contralateral left primary

motor cortex (BA 4) in blood pressure medication free older adults (center of gravity: x =

-42, y = -18, z = 48; 1204 voxels). Significant activation of left BA 4 was also observed

in older adults taking blood pressure medications (center of gravity: x = -37, y = -22, z =

56; 435 voxels). Importantly, there were no significant differences between subjects

taking blood pressure medications and subjects that were blood pressure medication free

older adults in left BA 4, or any other regions (P > .01, Bonferroni corrected; threshold of

80 contiguous voxels). As this analysis is susceptible to a Type II statistical error due to

low power (e.g., each group had 6 subjects), a much more liberal threshold was utilized

as well. There again was no significant difference observed at a threshold of P < 0.01,

uncorrected for multiple comparisons, and a minimum of 10 contiguous voxels.









Age Effect on Encoding Activity

In order to evaluate the effect of aging on encoding, only the young adult and older

adult no-goal groups were compared so as not to confound this comparison with the


Figure 3-2. Left primary motor cortex (BA 4) activation during the motor task in young
adults, older adults, and older adults subtracted from young adults. Statistical
threshold was set at P < 0.01 (Bonferroni corrected for multiple comparisons;
minimum of 80 contiguous voxels). Activation statistical maps are displayed
on the smoothed, averaged T1 image for all subjects in the analysis. Central
point of the cross-hairs indicates the center of gravity for activation within BA







61



0.25 -

0.2 -
.6It
S 015 Young Adults / /
0.15 -
S--- Older Adults //
S 0.1

0.05 -

00 -
1 3 4 5
-0.05 -
U, //

S-0.1 Motor
0 Response Volumes
-0.15 -

-0.2 -




Figure 3-3. Motor task-related z-transformed signal intensity change as a function of
scan-in-trial is displayed. Data were obtained from the left BA 4 in young
adults (n=20) and older adults (n=20). Scan by group analysis revealed no
significant differences in the curvature of the hemodynamic response between
young and older adults.

presence of goals. Brain regions demonstrating significant activations for encoding in

young and older adults are presented in Table 3-2. Direct comparisons between young

and older adults' encoding-related activation levels are presented in Table 3-3. Figures F-

1 and F-2 of Appendix F show the 3-dimensional spatial extent of activation in older and

young adults, as well as activation differences between the two groups.

A priori findings. The effects of memory encoding were pronounced and

statistically significant in both older and young adults in the prefrontal cortex. Encoding

was associated with several prefrontal foci in both groups. Much of this activity was in

the left middle frontal gyms (BA 6, 8, 9, and 10). Left BA 8 of the medial frontal gyms

was activated only in younger adults. Broca's area was additionally activated in both









groups. Right middle frontal gyrus activation was observed in both groups, although to a

much lesser extent. In the BA 6 portion of the right middle frontal gyrus, activation was

greater in older adults. In the BA 8 portion of the right middle frontal gyrus, activation

was greater in young adults. Left anterior cingulate (BA 32) was activated in young

adults, but not older adults. A small, bilateral region of BA 10 in the middle frontal gyrus

was also activated only in younger adults. Based on qualitative observations of the final

statistical maps, activation was left lateralized in young adults, and to a lesser extent in

older adults. When activation levels between young and older adults were compared

directly, regions of the left medial frontal gyrus (BA 6, 8, and 9) were greater in young

adults than in older adults. Activations were also greater in young adults in bilateral

anterior cingulate cortex (BA 32). Right lateralized medial prefrontal gyrus regions (BA

6 and 8) and middle frontal gyrus (BA 6) demonstrated greater activation in young adults.

Temporal lobe involvement was observed in young and older adults extending from

BA 20 to BA 21. This activation was bilateral in BA 20. Activation in the bilateral

parahippocampal gyrus and left hippocampus was found only in older adults. Activation

in the left BA 22 region of the superior temporal gyrus was only observed in the young

adults. Overall, greater spatial extent of activation was observed in the temporal lobe

regions for the young adults, but signal intensity-based comparisons did not reveal any

significant differences between the two groups.

It should be noted that regions located in the occipital lobe were activated as well in

both young and older adults. The encoding period is time-locked to the presentation of

the words to be remembered. Consequently, visual stimulation occurred in conjunction

with encoding. As was expected, several occipital regions were activated (precuneus,









fusiform gyrus, and lingual gyrus). These areas will not be a direct focus of evaluation

and therefore are not listed in the following tables.

A posteriori findings. Bilateral activation of the angular gyrus (BA 39), as well as

left lateralized supramarginal gyrus (BA 40), was found in both groups. Activation was

also observed in the transverse temporal gyrus (BA 41), also referred to as the primary

auditory receiving cortex. Lastly, activation was observed in the precentral and

postcentral gyri (BA 3 and 4) in older adults. When directly comparing young and older

adults, greater activity was observed in bilateral posterior cingulate cortex (BA 31) and

bilateral supramarginal gyrus (BA 40) in young adults.

Age Effect on Subsequent Memory Activity
(Encoding Activity Correlated with Recall Performance)

Brain regions demonstrating significant subsequent memory effects in young and

older adults are presented in Table 3-4. Direct comparisons between young and older

adults on subsequent memory effects are presented in Table 3-5. Figures F-3 and F-4 of

Appendix F show the 3-dimensional spatial extent of activation in older and young

adults, as well as activation differences between the two groups.

A priori findings. The effects of memory encoding were pronounced and

statistically significant for both older and young adults in the prefrontal cortex.

Subsequent memory was associated with several prefrontal foci in both groups, including

the middle frontal gyrus (BA 6, 8, and 9; see Figure 3-4). Left anterior cingulate (BA 32)

and right middle frontal gyrus (BA 9) were activated only in young adults. Right

supplementary motor area (BA 6) was activated only in older adults. When activation in

young and older adults was directly compared, greater activation in young adults was

observed in the medial frontal gyrus (BA 6, bilaterally; left BA 8 and 9) and anterior












Table 3-2. Comparison of Activation during Encoding in Young and Older Adult Groups
Not Receiving Goal-setting


Region (BA)


Left
Postcentral gyms (3)
Precentral gyms (4)
Middle frontal gyms (6)
Medial frontal gyms (8)
Middle frontal gyms (8)
Middle frontal gyms (9)
Middle frontal gyrus (10)
Middle temporal gyrus (20)
Middle temporal gyms (21)
Superior temporal gyms (22)
Anterior cingulate cortex (32)
Parahippocampal gyms (36)
Angular gyrus (39)
Supramarginal gyms (40)
Inferior frontal gyms (44)
Broca's area (45)
Hippocampus

Right
Middle frontal gyms (6)
Middle frontal gyms (8)
Middle temporal gyrus (20)
Parahippocampal gyms (36)
Angular gyrus (39)
Transverse temporal gyms (41)


Young
Talairach Voxels
(x,y,z)


-55,-10,35
-29,5,43
-9,30,42

-48,10,31
-36,47,12
-36,-36,-15
-60,-27,0
-59,-33,5
-12,19,42

-41,-61,37
-47,-59,43
-52,15,9
-48,22,5


126
1220
477

734
101
207
887
1165
268

562
114
92
205


Older
Talairach Voxels
(x,y,z)


-55,-10,46
-52,-7,45
-48,0,38


112
257
2278


-49,12,40 236
-49,10,33 973

-35,-36,-15 292
-64,-26,0 113


-34,-29,-16
-37,-60,37
-43,-47,42

-42,28,16
-25,-21,-4


230
324
1054

1623
165


40,-4,33 257


34,24,34
31,-36,-15


33,-60,38 204
41,-31,9 94


32,-37,-15
29,-31,-13
31,-59,36


BA = Brodmann's Area. Talairach = 3-dimensional coordinates for the center of gravity in each activation
cluster given the stereotactic space of Talairach and Tournoux (1988). Voxel = number of voxels in each
cluster exceeding height threshold P < 0.01, Bonferroni corrected in the whole brain volume; exceeding
minimum threshold of 80 contiguous voxels.












Table 3-3. Encoding Related Activity Differences between Young and Older Adult
Groups Not Receiving Goal-setting
Region (BA) Young Older
Talairach Voxels Direction
(x,y,z) of Effect
Left
Medial frontal gyms (6) -6,24,46 570 Y> O
Medial frontal gyms (8) -6,32,43 818 Y> O
Medial frontal gyrus (9) -6,43,28 830 Y>O
Posterior cingulate cortex (31) -7,-63,27 518 Y>0O
Anterior cingulate (32) -6,29,31 609 Y> O
Supramarginal gyms (40) -58,-46,26 119 Y> O

Right
Middle frontal gyms (6) 21,17,47 811 Y>O
Medial frontal gyms (8) 14,28,45 1011 Y>O
Medial frontal gyms (9) 10,41,32 736 Y > O
Posterior cingulate cortex (31) 1,-59,27 697 Y > O
Anterior cingulate cortex (32) 2,30,29 347 Y> O
Supramarginal gyms (40) 53,-29,33 117 Y> O
BA= Brodmann's Area. Talairach = 3-dimensional coordinates for the center of gravity
in each activation cluster given the stereotactic space of Talairach and Tournoux (1988).
Voxel = number of voxels in each cluster exceeding height threshold P < 0.01,
Bonferroni corrected in the whole brain volume; exceeding minimum threshold of 80
contiguous voxels. Y = young adults; 0 = older adults.

cingulate (left BA 24 and bilateral BA 32). Figure 3-5 illustrates a PFC cluster including

portions of the anterior cingulate cortex and medial frontal gyrus that exhibited

significant task-related changes in signal intensity that were correlated with subsequent

performance.

Extensive temporal lobe involvement, which was mostly left lateralized, was

observed in young and older adults. Bilateral activations were observed in the

parahippocampal gyrus (BA 36) in both groups. Bilateral activation in the middle

temporal lobe was observed in BA 20 for both groups. Right BA 21 and 22 of the middle

temporal gyrus were activated only in young adults. Hippocampal activation was bilateral

in young and older adults. The spatial extent of activation was somewhat greater in young












Table 3-4. Comparison of Subsequent Memory Effect in Young and Older Adult Groups
Not Receiving Goal-setting


Region (BA)


Left
Postcentral gyms (3)
Precentral gyms (4)
Middle frontal gyrus (6)
Middle frontal gyrus (8)
Middle frontal gyms (9)
Middle temporal gyms (20)
Middle temporal gyms (21)
Superior temporal gyrus (22)
Anterior cingulate cortex (32)
Parahippocampal gyms (36)
Angular gyrus (39)
Supramarginal gyrus (40)
Hippocampus

Right
Supplementary motor areas (6)
Middle frontal gyms (9)
Middle temporal gyms (20)
Parahippocampal gyms (35)
Parahippocampal gyms (36)
Angular gyms (39)
Hippocampus


Young
Talairach Voxels
(x,y,z)


-37,4,41
-26,21,41
-44,9,35
-35,-36,-15
-59,-29,0
-58,-34,4
-11,19,42
-37,-31,-13
-38,-61,38
-45,-58,43
-26,-12,-9


1576
480
486
250
573
501
264
78
391
102
1770


Older
Talairach Voxels
(x,y,z)


-55,-10,46
-51,-6,47
-50,0,41
-50,13,40
-54,9,34
-36,-36,-15



-35,-28,-17
-36,-60,37
-42,-48,42
-27,-23,-4


84
167
1438
160
251
238



150
237
948
422


40,-2,35 156


37,24,33 117
32,-36,-15 237


33,-31,-13
35,-61,39
21,-12,-6


133
243
1003


32,-37,-15
25,-25,-17
28,-31,-13
32,-60,38


BA = Brodmann's Area. Talairach = 3-dimensional coordinates for the center of gravity in each activation
cluster given the stereotactic space of Talairach and Tournoux (1988). Voxels = number of voxels in each
cluster exceeding height threshold P < 0.01, Bonferroni corrected in the whole brain volume; exceeding
minimum threshold of 80 contiguous voxels.












Table 3-5. Subsequent Memory Differences between Young and Older Adult Groups Not
Receiving Goal-setting
Region (BA) Young Older
Talairach Voxels Direction
(x,y,z) of Effect
Left
Medial frontal gyms (6) -8,29,36 118 Y> O
Medial frontal gyms (8) -7,29,42 321 Y> O
Medial frontal gyms (9) -8,30,32 84 Y> O
Anterior cingulate cortex (24) -7,23,25 113 Y> O
Posterior cingulate cortex (31) -4,-66,29 89 Y> O
Anterior cingulate cortex (32) -7,26,31 838 Y> O
Supramarginal gyrus (39) -52,-60,23 83 Y> O
Supramarginal gyrus (40) -58,-46,27 85 Y> O
Putamen -20,7,5 219 Y>O

Right
Middle frontal gyrus (6) 27,10,45 180 Y> O
Posterior cingulate cortex (31) 2,-56,28 293 Y> O
Anterior cingulate cortex (32) 1,26,29 137 Y> O
BA = Brodmann's Area. Talairach = 3-dimensional coordinates for the center of gravity in each activation
cluster given the stereotactic space of Talairach and Tournoux (1988). Voxels = number of voxels in each
cluster exceeding height threshold P < 0.01, Bonferroni corrected in the whole brain volume; exceeding
minimum threshold of 80 contiguous voxels. Y = young adults; O = older adults.

adults in temporal regions, including the hippocampus (see Figure 3-4).

Several occipital regions were additionally activated (precuneus, fusiform gyrus,

and lingual gyrus).

A posteriori findings. Activation of the precentral (BA 4) and postcentral (BA 3)

gyri was found only in older adults. Bilateral angular gyrus (BA 39) and left

supramarginal gyrus activity (BA 40) was found in young and older adults. When

activation in young and older adults was compared directly, bilateral posterior cingulate

(BA 31), left supramarginal gyrus (BA 39 and 40), and left putamen were found to be

greater in young adults.

Goal-setting Effect on Encoding Activity

In order to evaluate the effect of goal-setting on encoding, young and older adult
















Middle and
Medial Frontal
Gyri






Hippocampus




Figure 3-4. Effect of subsequent memory in young (n=10) and older (n=10) adults.
Coronal slice images illustrate regions in the medial and middle frontal gyri
and hippocampus that exhibited significant task-related changes in signal
intensity that were correlated with subsequent performance (P<0.01,
Bonferroni corrected; minimum of 80 contiguous voxels). Activation
statistical maps are displayed on the smoothed, averaged T1 image for all
subjects in the analysis.













0.15

S 0.1
Younger Adults
F0.05 3 Older Adults


1 3 \4 6 13 14 15 16 17
E -0.05




Volumes




Figure 3-5. Effect of subsequent memory: older adults subtracted from young adults.
Coronal slice image illustrates PFC cluster (extending from the medial frontal
gyrus to the anterior cingulate cortex) that exhibited significant group
differences in signal intensity that were correlated with subsequent
performance (P<0.01, Bonferroni corrected; minimum of 80 contiguous
voxels). Activation statistical maps are displayed on the smoothed, averaged
T1 image for all subjects in the analysis. Task-related z-transformed signal
intensity as a function of scan-in-trial is displayed in the graphical illustration.
Data were obtained from the contiguous cluster of 7 voxels that had the
highest t-values of the overall cluster (x = -6, y = 26, z = 29).









groups were collapsed across goal (n = 20) and no-goal (n = 20) groups. Brain regions

demonstrating significant activations for encoding in the goal and no-goal groups are

presented in Table 3-6. Data reflecting significant differences between groups are not

presented in tabular format as there was only one region demonstrating a significant

difference. Figures F-5 and F-6 of Appendix F show the 3-dimensional spatial extent of

activation in the goal and no-goal groups, as well as activation differences between the

two groups.

A priori findings. In the individual group analyses, activation was observed in

bilateral dlPFC regions in the goal and no-goal groups. Specifically, bilateral activation

was found in the BA 9 region of the middle frontal gyrus, and left BA 46 activation was

found in the middle frontal gyrus. Orbitofrontal cortex activation was observed in the left

BA 10 region of the middle frontal gyrus and left BA 47 regions of the inferior frontal

gyrus in both groups. However, only in the goal group was right BA 10 activation found.

Broca's area (BA 45), left BA 8 of the middle frontal gyrus, and left supplementary

motor area (BA 6) include other frontal regions activated in both groups. Right BA 6 was

additionally activated in the goal group. The left BA 32 region of the anterior cingulate

was activated in both groups, whereas activation in the right anterior cingulate (BA 24

and 32) was observed only in the goal group. No significant differences were found

between the goal and no-goal groups in any frontal lobe regions.

Bilateral temporal lobe involvement was observed in the hippocampus,

parahippocampal gyrus (BA 36), and BA 20 of the middle temporal gyrus in both groups.

Greater spatial extent of the hippocampal activation was observed in the goal group;

however, differences in signal intensity in this region were not significant. Activation in









the left insula (BA 13), left superior temporal gyms (BA 22), and left middle temporal

gyrus (BA 21) was observed in both groups as well. When goal and no-goal groups were

compared directly, BA 20 activation of the middle temporal gyrus was found to be

greater in the goal group.

Several occipital regions were additionally activated (precuneus, fusiform gyrus,

and lingual gyrus).

A posteriori findings. Bilateral angular gyrus (BA 39) and supramarginal gyrus

(BA 40) activity was observed in both groups. Bilateral posterior cingulate cortex (BA 23

and 31) was activated in the goal group, but not in the no-goal group. Left precentral (BA

4) and postcentral (BA 3) gyri were activated in both groups. Activation was also

observed in the right transverse temporal gyrus (BA 41) in the no-goal group.

Goal-setting Effect on Subsequent Memory Activity
(Encoding Activity Correlated with Recall Performance)

Brain regions demonstrating significant subsequent memory activations are

presented in Table 3-7. Direct comparisons between goal and no-goal groups' subsequent

memory activation level differences are presented in Table 3-8. Figures F-7 and F-8 of

Appendix F show the 3-dimensional spatial extent of activation in the goal and no-goal

groups, as well as activation differences between the two groups.

A priori findings. In the individual group analyses, activation was observed in

bilateral dlPFC regions in goal and no-goal groups. Specifically, there was bilateral

activation in the BA 9 region of the middle frontal gyrus. Only in the goal group was

activation in the left BA 46 region of the middle frontal gyrus found. In the goal group,

orbitofrontal cortex activation was observed in the left BA 10 region of the middle frontal

gyrus and in the left BA 47 regions of the inferior frontal gyrus. Broca's area (BA 45),












Table 3-6. Comparison of Activation during Encoding in the Goal and No-goal Groups


Region (BA)


Goal
Talairach Voxels
(x,y,z)


No-goal
Talairach Voxels
(x,y,z)


Left
Frontal Lobe
Precental gyms (4)
Supplementary motor area (6)
Middle frontal gyrus (8)
Middle frontal gyrus (9)
Middle frontal gyms (10)
Broca's area (45)
Middle frontal gyms (46)
Anterior cingulate cortex (32)
Inferior frontal gyms (47)
Temporal Lobe
Insula (13)
Middle temporal gyms (20)
Middle temporal gyrus (21)
Superior temporal gyms (22)
Parahippocampal gyms (36)
Hippocampus
Parietal Lobe
Supramarginal gyms (40)
Posterior cingulate cortex (23)
Angular gyms (39)
Postcentral gyms (3)

Right
Frontal Lobe
Supplementary motor area (6)
Middle frontal gyms (9)
Middle frontal gyms (10)
Anterior cingulate cortex (24)
Anterior cingulate cortex (32)
Temporal Lobe
Middle temporal gyms (20)
Transverse temporal gyms (41)
Parahippocampal gyms (36)
Hippocampus
Parietal Lobe
Angular gyrus (39)
Supramarginal gyms (40)
Posterior cingulate cortex (31)
Posterior cingulate cortex (23)


-52,-9,42
-36,0,45
-39,16,40
-48,13,32
-38,45,13
-50,22,14
-47,34,16
-13,17,41
-36,27,-1

-37,12,10
-35,-35,-15
-60,-26,0
-60,-30,5
-36,-29,-14
-35,-21,-9

-44,-49,44
-9,-72,12
-41,-61,33
-55,-11,41


736
3684
333
1820
854
390
1059
368
107

393
317
514
766
313
1991

1389
144
976
266


41,-5,34
37,24,33
30,47,9
14,2,38
11,17,35


-54,-9,39
-45,0,40
-41,16,41
-48,11,32
-39,45,12
-44,24,6
-45,32,16
-13,13,45
-42,25,0

-37,22,7
-35,-36,-15
-60,-27,0
-59,-31,4
-35,-29,-16
-25,-22,-5


679
3155
317
1671
341
203
432
141
201

374
288
764
801
193
525


-44,-51,43 853


-40,-61,37
-56,-11,39


35,24,33 287


32,-37,-15 250 32,-36,-15
42,-29,8
31,-30,-14 342 30,-30,-14
28,-23,-4 1100 23,-19,-4


33,-59,36
39,-47,42
16,-58,24
7,-71,12


231
1017
177
191


254
104
427
515


33,-60,37 275
35,-51,43 132


BA = Brodmann's Area. Talairach = 3-dimensional coordinates for the center of gravity in each activation
cluster given the stereotactic space of Talairach and Tournoux (1988). Voxels = number of voxels in each
cluster exceeding height threshold P < 0.01, Bonferroni corrected in the whole brain volume; exceeding
minimum threshold of 80 contiguous voxels.









left BA 8 of the middle frontal cortex, and left supplementary motor area (BA 6) include

other frontal regions activated in both groups. Right BA 6 and Broca's homologue (right

BA 45) were additionally activated in the goal group. The left BA 32 region of the

anterior cingulate was activated in both groups.

When the goal and no-goal groups were compared directly, there was significantly

greater activity in the bilateral orbitofrontal cortex (BA 10; x=-32, y=49, z=16), dlPFC

(BA 46; x =-47, y=28, z=19), and Broca's area (BA 45; x=-50, y=19, z=18) in the goal-

group (see Figure 3-6).

Bilateral temporal lobe involvement was observed in the hippocampus,

parahippocampal gyms (BA 36), and BA 20 of the middle temporal gyrus in both groups.

Activation in the left insula (BA 13), left superior temporal gyrus (BA 22), and left

middle temporal gyrus (BA 21) was observed in both groups as well. When goal and no-

goal groups were compared directly, BA 20 of the middle temporal gyrus was found to be

greater in the no-goal group bilaterally and activation in the right hippocampus was

greater in the goal group.

Several occipital regions were additionally activated (precuneus, fusiform gyrus,

and lingual gyrus).

A posteriori findings. Bilateral angular gyrus (BA 39) and supramarginal gyrus

(BA 40) activity was observed in both groups. Left posterior cingulate cortex (BA 30)

was activated in the goal group, but not in the no-goal group. Left precentral and

postcentral gyri were activated in both groups. Activation was also observed in the right

transverse temporal gyrus (BA 41). Amygdalar activation was observed in the left









hemisphere of the goal group only. There were no significant differences between groups

in any of these regions.

Interaction of Goal-setting and Age during Encoding Activity

Brain regions demonstrating significant interactions during encoding activation are

presented in Table 3-9 with parameter estimates listed for each group. Figure F-9 of

Appendix F shows the 3-dimensional spatial extent of activation in regions that

demonstrated a significant age by goal-setting interaction.

Regions that were identified as having a significant interaction demonstrated,

without exception, the same pattern of relative increases or decreases within age groups.

More specifically, regions with a significant interaction showed decreases in activity in

young adults as a function of goal-setting, whereas activation in older adults increased as

a function of goal-setting. The observed pattern of activation differences between young

and older adults in response to goal-setting indicated that older adults activated these

regions to a greater extent during goal-setting relative to activity observed in the no-goal

group.

Beta weights provided in Table 3-9 are products of the multiple-regression analysis

performed on the interaction of age by goal-setting. Beta weights are interpreted only to

the extent that they show relative increases or decreases in encoding-related activity

within an age group as a function of goal-setting. For regions demonstrating a significant

interaction, relative increases or decreases in encoding-related activity as a consequence

of goal-setting in one age group are then compared to the pattern observed in the other

age group.

A priori findings. In frontal regions, activation was observed in bilateral medial

frontal gyms (BA 9; see Figure 3-7), BA 8 of the left medial frontal gyrus, and left










Table 3-7. Comparison of Subsequent Memory Effect in Goal and No-goal Groups


Region (BA)


Left
Frontal Lobe
Precentral gyms (4)
Supplementary motor area (6)
Middle frontal gyrus (8)
Middle frontal gyrus (9)
Middle frontal gyms (10)
Anterior cingulate cortex (32)
Broca's area (45)
Middle frontal gyrus (46)
Inferior frontal gyms (47)
Temporal Lobe
Insula (13)
Middle temporal gyms (20)
Middle temporal gyrus (21)
Superior temporal gyms (22)
Parahippocampal gyms (36)
Hippocampus
Parietal Lobe
Angular gyrus (39)
Supramarginal gyms (40)
Postcentral gyms (3)

Right
Frontal Lobe
Supplementary motor area (6)
Middle frontal gyms (9)
Broca's homologue (45)
Temporal Lobe
Middle temporal gyms (20)
Parahippocampal gyms (36)
Parahippocampal gyms (35)
Amygdala
Hippocampus
Parietal Lobe
Angular gyrus (39)
Supramarginal gyms (40)
Posterior cingulate cortex (30)


Goal
Talairach Voxels
(x,y,z)


-51,-8,44
-40,0,39
-33,19,40
-47,13,32
-36,48,10
-13,18,41
-37,29,11
-47,34,16
-39,27,-3

-37,23,7
-36,-34,-15
-62,-25,0
-63,-28,3
-36,-28,-15
-24,-28,0


264
2059
413
1859
1181
368
9181
1045
129

258
324
212
145
273
1141


-40,-61,35 753
-43,-52,43 585


39,-5,34
39,19,34
25,31,8


342
609
3632


32,-37,-15 252
31,-30,-14 306


29,-4,-17
23,-25,-2

32,-59,36
38,-46,40
22,-67,11


No-Goal
Talairach Voxels
(x,y,z)


-51,-7,46
-45,0,40
-47,14,41
-48,9,33


272
2932
282
1031


-13,13,45 130
-27,32,2 370


-35,-36,-15
-60,-28,0
-58,-30,3
-36,-29,-15
-29,-14,-9

-39,-61,37
-43,-49,43
-55,-10,46


290
363
133
276
2315

504
1119
112


38,25,34 249
22,33,2 197


32,-36,-15
29,-30,-13
24,-25,-15


25,-16,-8 820


35,-61,37
36,-47,44


BA = Brodmann's Area. Talairach = 3-dimensional coordinates for the center of gravity in each activation
cluster given the stereotactic space of Talairach and Toumoux (1988). Voxels = number of voxels in each
cluster exceeding height threshold P < 0.01, Bonferroni corrected in the whole brain volume; exceeding
minimum threshold of 80 contiguous voxels.












Table 3-8. Subsequent Memory Effect Differences between Goal and No-goal Groups
Region (BA) Goal No-goal
Talairach Voxels Direction
(x,y,z) of Effect
Left
Supplementary motor area (6) -41,4,52 210 NG > G
Middle frontal gyrus (10) -32,49,16 826 G > NG
Middle temporal gyrus (20) -25,-84,-12 117 NG> G
Broca's area (45) -39,31,7 1680 G > NG
Inferior frontal gyms (46) -45,40,11 153 G > NG

Right
Middle frontal gyrus (10) 36,45,20 103 G > NG
Middle temporal gyrus (20) 27,-79,-10 150 NG > G
Hippocampus 38,25,8 507 G>NG
BA = Brodmann's Area. Talairach = 3-dimensional coordinates for the center of gravity in each activation
cluster given the stereotactic space of Talairach and Toumoux (1988). Voxels = number of voxels in each
cluster exceeding height threshold P < 0.01, Bonferroni corrected in the whole brain volume; exceeding
minimum threshold of 80 contiguous voxels. G = goal group, NG = no-goal group
















0.2
S0.15
0.1
i 0.05
7 0
J-0.05 1 2 5 6 7 8 9 10 11 12 13 14 1 67

No- Goal
tl -0.15 No-goal
-0.2- Presen n m B oInter-block Interval

Volumes


0.15

0.1

0.05

0

-0.05

-0.1

-0.15


Goal
I No-goal




1; 2 3 14 6I9 10\ 13 14 5 17


v /


Inter-block Interval


Volumes


1 2 3 4 5 6 7 8 9 10 11 12 13 1^ 16 17
Goal
S- No-goal

X Pe Inter-block Interval

Volumes


Figure 3-6. Effect of subsequent memory: no-goal group subtracted from the goal group.
Coronal slice images illustrate frontal cortex clusters that exhibited significant
activation (P<0.01, Bonferroni corrected; minimum of 80 contiguous voxels).
Clusters depicted are (a) orbitofrontal cortex (BA 10; x=-32, y=49, z=16); (b)
Broca's area (BA 45; x=-50, y=19, z=18); and (c) dorsolateral prefrontal
cortex (BA 46; x =-47, y=28, z=19). Activation statistical maps are displayed
on the smoothed averaged TI image for all subjects in the analysis. Task-
related z-transformed signal intensity as a function of scan-in-trial is displayed
in the graphical illustrations. Data were obtained from the contiguous cluster
of 7 voxels that had the highest t-values of the overall cluster.


I


-

-

-





-










supplementary motor area (BA 6). Activation in the temporal lobe was limited to the left

superior temporal gyrus (BA 22). Significant interactions were not observed in the OFC

or dlPFC.


0.2

0.1

A 0-

-0.1

S-0.2 -

P -0.3

-0.4


Figure 3-7. Coronal slice image illustrates PFC cluster (BA 9) that exhibited significant
task-related interaction of age by goal-setting (P<0.01; Bonferroni corrected,
minimum of 80 contiguous voxels). Activation statistical maps are displayed
on the smoothed, averaged T1 image for all subjects in the analysis. Parameter
estimates (e.g., beta weights) for each group are displayed in the graphical
illustration. Data were obtained from the contiguous cluster of 7 voxels that
had the highest t-values of the overall cluster (x=-5, y=44, z=29). Standard
error bars are shown.

A posteriori findings. An interaction effect was found in bilateral supramarginal

gyrus (BA 40), left angular gyms (BA 39), and left posterior cingulate cortex (BA 31).

Interaction of Goal-setting and Age for Subsequent Memory Activity
(Encoding Activity Correlated with Recall Performance)

Brain regions demonstrating significant subsequent memory interactions are

presented in Table 3-10 with parameter estimates listed for each group. Figure F-10 of

Appendix F shows the 3-dimensional spatial extent of activation in regions that

demonstrated a significant age by goal-setting interaction. Beta weights are again used

for interpretation of regions that demonstrate a significant interaction.


Young Adults/ Older Adults/ Older Adults/
Goal TGoal No-Goal

Young Adults/
No-Goal


Age x Goal-setting
Interaction
R L









y=44












Table 3-9. Interaction of Age and Goal-setting during Memory Encoding
Region (BA) Talairach Voxels Parameter Estimates (beta weights)
(x,y,z) YA/G YA/NG OA/G OA/NG
Left
Supplementary motor area (6) -6,33,39 169 -0.09 0.14 0.00 -0.26
Medial frontal gyrus (8) -8,31,41 277 0.04 0.29 0.10 -0.13
Medial frontal gyrus (9) -5,44,29 348 -0.21 0.08 -0.12 -0.31
Superior temporal gyrus (22) -55,-50,17 178 -0.28 0.08 0.04 -0.04
Posterior cingulate cortex (31) -6,-55,25 196 -0.18 0.10 -0.06 -0.18
Angular gyrus (39) -50,-61,18 477 -0.15 0.11 0.12 -0.15
Supramarginal gyms (40) -59,-40,26 485 -0.25 -0.04 0.020 -0.29

Right
Medial frontal gyms (9) 2,43,31 175 -0.24 0.03 -0.16 -0.34
Supramarginal gyms (40) 52,-29,33 242 -0.28 -0.15 -0.09 -0.46
BA = Brodmann's Area. Talairach = 3-dimensional coordinates for the center of gravity in each activation
cluster given the stereotactic space of Talairach and Toumoux (1988). Voxels = number of voxels in each
cluster exceeding height threshold P < 0.01, Bonferroni corrected in the whole brain volume; exceeding
minimum threshold of 80 contiguous voxels. YA = young adults, OA = older adult, G = goal group, NG =
no-goal group.


A priori findings. Frontal regions that demonstrated a significant interaction

included Broca's homologue (BA 45) and the anterior cingulate cortex (BA 31). The left

superior temporal gyrus (BA 22) was also activated.

A posteriori findings. The left angular gyrus (BA 39) demonstrated a significant


interaction.












Table 3-10. Interaction of Age and Goal-setting during Encoding for Subsequent Memory
Effects
Region (BA) Talairach Voxels Parameter Estimates (beta weights)
(x,y,z) YA/G YA/NG OA/G OA/NG
Left
Superior temporal gyms (22) -52,-56,16 83 -0.34 0.18 0.09 -0.26
Posterior cingulate cortex (31) -5,-12,45 167 -0.37 -0.12 0.01 -0.44
Angular gyrus (39) -50,-60,19 260 -0.30 .14 .14 -0.33

Right
Broca's homologue (45) 45,34,3 207 -0.21 -0.07 0.09 -0.33
BA = Brodmann's Area. Talairach = 3-dimensional coordinates for the center of gravity in each activation
cluster given the stereotactic space of Talairach and Tournoux (1988). Voxels = number of voxels in each
cluster exceeding height threshold P < 0.01, Bonferroni corrected in the whole brain volume; exceeding
minimum threshold of 80 contiguous voxels. YA = young adults, OA = older adult, G = goal group, NG =
no-goal group.













CHAPTER 4
DISCUSSION

Despite their high levels of self-reported general health and greater vocabulary

skills, older subjects exhibited a pattern of episodic memory performance typical of that

reported in behavioral studies of aging (Light, 1991). In particular, older subjects'

performances were below those of young subjects on a word list recall task. The

provision of goals for performing the task resulted in a significant increase in the number

of words recalled in both young and older adults. Further, the increase in performance as

a result of goal-setting was equivalent between young and older adults. These findings

are consistent with other studies that utilized an analogous paradigm (West et al., 2002;

2003).

Encoding of the word list was associated with highly left-lateralized activity in the

prefrontal cortex (PFC) of young adults. PFC activity in older adults was dampened

compared to young adults. When correlating recall performance with signal intensity

(e.g., subsequent memory effects), bilateral temporal lobe and hippocampal activity was

also observed in young and older adults. Volume distribution of the activation was larger

in young adults; however, a signal-intensity based comparison did not reveal a significant

difference.

The effect of goal-setting on subsequent memory activation was only found in the

frontal lobes. Specifically, regions that demonstrated higher levels of activity from goal-

setting included the orbitofrontal cortex (OFC), dorsolateral prefrontal cortex (dlPFC), as

well as Broca's area. A significant increase in the right hippocampal region as a result of









goal-setting was observed as well. When comparing young and older adults' response to

goal-setting, older adults were found to recruit several different regions to a greater extent

than young adults.

Encoding in Young Adults

A leading model describing the cortical neural substrates of encoding is the

hemispheric encoding/retrieval asymmetry (HERA) model (Tulving, 1994). The HERA

model proposes that the left PFC is more involved than the right PFC in encoding,

whereas the right PFC is more involved than the left PFC in episodic memory retrieval.

However, more recent studies of spatial (as opposed to verbal) material find greater right-

lateralized PFC activity during encoding, presenting a formidable challenge to the HERA

model. Thus, Tulving and colleagues (Habib et al., 2003) recently asserted a revised

formulation of the HERA model in order to accommodate these conflicting results. They

assert that encoding and retrieval processes must be systematically varied in their

interaction with hemispheric interactions. Specifically, to test for asymmetry in encoding,

the following formulation must be met: left hemisphere encoding minus left hemisphere

retrieval is greater than right hemisphere encoding minus right hemisphere retrieval.

Findings from the present study are generally consistent with the HERA model as

strongly left-lateralized PFC activation was observed during encoding. However, the

design of the present study does not allow for evaluation of retrieval-related neural

activation. Consequently, findings only partially address the HERA model as an

evaluation of the present data set with the more stringent reformulation of the HERA

model is not possible.

The present study found less than robust temporal lobe activation in the encoding-

related activity analysis. Extensive neuropsychological and physiological evidence points









to a clear role of temporal lobe involvement in encoding; however, many other fMRI

studies exist that found little or no encoding-related medial temporal lobe activation

(Petersen et al., 1988; Frith et al., 1991; Demonet et al., 1992; Grasby et al., 1993, 1994;

Kapur et al., 1994, 1996; Raichle et al., 1994; Shallice et al., 1994; Tulving et al., 1994;

Fletcher et al., 1995; Nyberg et al., 1996). In addition to the dubious finding of little or no

involvement of the medial temporal lobe in encoding, this approach measuring

encoding-related activity is questionable to the extent that encoding activities are

genuinely being assessed. For instance, encoding related activity in the hippocampus

could reflect a response to novelty unrelated to the encoding of a memory, or it could

reflect habituation or reduced attention. To address these important issues, the present

study also employed an approach that correlated subsequent memory performance with

signal intensity during encoding. In this way, one can be more confident that a particular

region of activation signifies processes important for encoding an episodic memory.

When using the correlative approach between memory performance and signal intensity,

robust activation was found in the temporal lobe/hippocampus regions, which is

consistent with other studies that employed this approach (Brewer et al., 1998; Gabrieli et

al., 1997; Fernandez et al., 1998, 1999).

Activation was observed bilaterally in the angular gyms and the supramarginal

gyrus, which likely reflects reading processes taking place during the word list

presentation (Price, 2000). Activation also occurred in the transverse temporal gyrus,

which may be accounted for by the loud scanner environment.

Encoding in Older Adults and Age-related Changes

The predicted reduction in left PFC activity observed during encoding for older

adults compared with young adults is in line with the HAROLD (hemispheric asymmetry









reduction in older adults) model. The HAROLD model states that young subjects,

consistent with the HERA model, engage the left PFC more extensively during encoding,

whereas older subjects show dampened left PFC activity during encoding (Cabeza et al.,

2001). Even though the present study, among many others (see Grady, 2000 for review),

showed decreased left PFC activation in older adults during encoding, it is worthwhile

noting two recent studies that showed a similar level of activation in the left PFC in

young and older adults. Daselaar et al. (2003) compared brain activity patterns obtained

during incidental encoding in which subjects were not asked to learn the words presented,

but instead were asked to make pleasant/unpleasant judgments about the words. They

found equivalent left PFC encoding activation in young subjects and older subjects when

correlating encoding activity with subsequent memory of the words. Morcum et al.

(2003) also found equivalent left PFC activation when subjects were asked to make

animacy decisions about words. Subjects later underwent a recognition memory test for

these words that was correlated with activity during the encoding period.

A common thread can be found between these two studies that reported equivalent

left PFC activation between young and older adults during encoding. Both of these

studies involved incidental learning, as opposed to the intentional learning approach that

was utilized by other studies, including the present one. This account for the differential

findings is further supported by a study by Buckner and colleagues (Logan et al., 2002),

which showed less prefrontal activity in older adults compared with young adults under

intentional learning instructions. Importantly, this difference was not observed when a

semantic-orienting task was used to support episodic encoding (e.g., incidental encoding).

This interpretation of the seemingly disparate findings in the left PFC is in agreement









with a production deficiency account of age-related impairments in episodic encoding.

This account asserts that older adults do not employ semantic elaboration strategies

spontaneously but are able to make use of them when forced to do so (Burke and Light,

1981). The production deficiency account is particularly interesting in light of the

differential use of strategies between young and older adults observed in the present

study. Older adults utilized less strategies in performing the intentional memory encoding

task. The decreased use of strategies is consistent with the production deficiency account

and is likely reflected by the decreased left PFC activation observed in older adults.

Greater activation was found in the anterior and posterior portions of the cingulate

cortex in young adults. A significant scan by age interaction existed in a cluster of voxels

that included regions of the anterior cingulate and medial frontal gyms. The greater level

of sustained activity in the young adults suggests greater attentional resources dedicated

to the task and perhaps more concentrated performance monitoring (Tisserand and Jolles,

2003). In regards to the increased level of activation in the posterior cingulate in young

adults, several studies have asserted a role for this region being involved in encoding

(Hunkin et al., 2002; Hofer et al., 2003). However, Otten and Rugg (2001) found that

encoding activation in the posterior cingulate was associated with recall failure on

subsequent memory tests. At present, the role of the posterior cingulate in encoding is

unclear and is need of further elucidation.

Older adults demonstrated encoding activity in the temporal lobe (primarily in the

middle temporal gyrus and hippocampus), which was to a lesser spatial extent than young

adults. There were no significant signal intensity differences between young and older

adults, though significant activation present in BA 21 of the middle temporal gyrus and









right hippocampus in the young adults was absent in older adults. It has been suggested

that the medial temporal lobe operates by forming associations between sensory,

cognitive and emotional processes that make up an episode in memory (Alvarez and

Squire, 1994; Eichembaum, 1996). Accordingly, it has been suggested that there is a

relation between the amount of medial temporal lobe activity and the number of

associations that are formed during encoding of study material (Henke et al., 1997, 1999).

Given the presence of medial temporal lobe activity during encoding, this would imply

that the older adults formed fewer memory associations.

In addition to the PFC and temporal lobes, older adults engaged other regions that

were in common with young adults. Activation was observed in the angular gyrus and

supramarginal gyrus regions in the encoding-related and subsequent memory analyses.

This activation, as previously stated, was likely associated with reading processes

required for list learning (Price, 2002).

Taken together, data from this pilot study suggest that older subjects engage much

of the same neural circuitry as young subjects when encoding new memories. However,

the findings also point to the possible presence of age-related differences in both

prefrontal and temporal activity during episodic encoding. Further, these findings suggest

that the age-related declines in episodic encoding may be related to strategic encoding

and attentional differences, as well as fewer memory associations formed.

Goal-setting Influence on Encoding

Predicted associations were found between OFC and dlPFC activation and the

presence of goal-setting during encoding. Additional regions of the prefrontal cortex were

also found to be involved with subsequent memory performance, including the

supplementary motor cortex and a large portion of Broca's area.









In further evaluating areas engaged by goal-setting, it is impossible to separate

which regions underlie the motivation properties of goal-setting and which regions

underlie increased mnemonic processing, facilitated by increased motivation. However,

comparing the results to the extent literature provides a likely framework in which the

goal-setting effect is mediated.

The generally agreed upon role of the OFC in motivation and the pattern of

activation observed in the OFC suggest that this region supports the motivational

component of goal-setting. In the subsequent memory effect analysis of goal-setting,

activation in the OFC was not observed in the no-goal group, indicating that this region

was likely not associated with successful or unsuccessful encoding. Consequently, it is

unlikely that the engagement of the OFC directly reflects mnemonic processes. It is

likely, however, that the OFC is associated with the motivational aspect of goal-directed

improvement of the performance, which is supported by several lines of evidence. Stuss

and Levine (2002) put forward the concept of the self-regulatory disorder, which is

characterized by an inability to regulate behavior according to internal goals and

constraints. They observed this disorder after discrete lesioning to the ventral frontal

cortex. Humans with ventral frontal lobe damage can show impairments in a number of

tasks in which an alteration of behavioral strategy is required in response to a change in

environmental context or expectations (Damasio, 1994; Rolls, 2000). Activity in this

region in the present study likely reflects the other end of the spectrum from what these

studies report: increased activity in the OFC mediates increases in goal-

directed/contextually-guided behavior.









A more difficult region to elucidate the contribution to the goal-setting effect is the

dlPFC. Extensive work demonstrates the role of the dlPFC in cognitive control (see

Miller and Cohen, 2002 for review). Cognitive control includes the ability to maintain

context, or a goal, by which behavior is then biased. Certainly it is plausible that the

increased dlPFC activation subserved the function of maintaining the idea of the goal on-

line while performing the task. However, being that the dlPFC subserved mnemonic

processes as well in the absence of goal-setting, it is difficult to ascertain whether this

region helped to improve memory performance by enhanced mnemonic processing or

maintenance of goal intent, or a combination of these factors.

Goal-setting also produced substantially greater activation in Broca's area and

supplementary motor areas. As activation in motor regions and Broca's area has been

implicated in subvocalization (Sweet et al., 2004; Gruber, 2001), these regions could be

involved in an increase in subvocal rehearsal in the goal group as a result of increased

motivation. This finding suggests that the goal group may have gained in performance

over the no-goal group by greater subvocal rehearsal during the encoding period.

Interestingly, none of the frontal regions observed to be significantly greater in the

goal-setting group in the subsequent memory analysis were significant in the encoding-

related activity analysis. Changes that occur during goal-setting are therefore not

reflected by general activity during encoding, but more likely reflect neural processes

directly related to successful or unsuccessful encoding.

Differential Neural Response to Goal-setting in Young and Older Adults

Older adults did not show the predicted differential activation of the OFC or dlPFC

in response to goal-setting, as compared to young adults. However, older adults recruited

several other regions to a greater extent than young adults to achieve the performance









enhancing effects of goal-setting. Interestingly, there were no regions that exhibited

significantly greater activity in young adults in conjunction with goal-setting, as

compared to older adults. This result is consistent with the hypothesis that older adults

would compensate for the age-related atrophy in the OFC and dlPFC. Each of the regions

demonstrating an age by goal-setting interaction was previously shown to be involved in

mnemonic processes. Therefore, the overall pattern of differentially greater activation in

older adults may reflect greater increases in mnemonic processing that are precipitated by

increases in motivation. However, it is also possible that greater activation in these

regions in older adults reflects less efficient recruitment of neural resources.

In the encoding-related activity analysis, increases in activity in the prefrontal,

temporal and posterior cingulate regions perhaps reflect increased resources committed

toward memory encoding processes. Activation in the supplementary motor area could

reflect increased subvocal rehearsal of the word list by older adults. The differential

increased activity in older adults in the angular gyms and the supramarginal gyrus may

reflect an increased level of focus on reading the words as they are presented. Again,

these increases in activation in the above mentioned regions could also represent

inefficient recruitment of neural resources in older adults.

It should be noted that activation in all the previously mentioned regions, with the

exception of the temporal lobe and angular gyrus, does not reach significance in the

subsequent memory analysis. This may indicate that even though the regions are

recruited more heavily as compared to young adults, only activations in the angular gyrus

and superior temporal gyrus have a genuine impact on subsequent recall performance.

Additionally, results from the subsequent memory analysis revealed that older adults