Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Explicit/implicit memory performance and memory strategies in Alzheimer patients Gallie, Karen A. 1993

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_1993_fall_phd_gallie_karen.pdf [ 10.97MB ]
Metadata
JSON: 831-1.0098809.json
JSON-LD: 831-1.0098809-ld.json
RDF/XML (Pretty): 831-1.0098809-rdf.xml
RDF/JSON: 831-1.0098809-rdf.json
Turtle: 831-1.0098809-turtle.txt
N-Triples: 831-1.0098809-rdf-ntriples.txt
Original Record: 831-1.0098809-source.json
Full Text
831-1.0098809-fulltext.txt
Citation
831-1.0098809.ris

Full Text

EXPLICIT/IMPLICIT MEMORY PERFORMANCE AND MEMORY STRATEGIES IN ALZHEIMER PATIENTS  by KAREN ANN GALLIE B.Sc., University of Victoria, 1980 M.A., University of British Columbia, 1985  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES NEUROSCIENCE PROGRAM  We accept this thesis as conforming to the required standard  THE UNIVERSITY OF BRITISH COLUMBIA September 1993 © Karen A. Gal ie, 19 3  In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission.  (Signature)  Department of  ^Net/ROSCie/ICe Myrath  The University of British Columbia Vancouver, Canada  Date  DE-6 (2/88)  ii ABSTRACT Research has shown that implicit memory can be spared in some amnesic patients, even when explicit memory is severely impaired. In contrast, results from studies with Alzheimer patients are mixed. This investigation re-explored these latter findings in a series of experiments, many that differed from those used in previous work, to find out how the performance of AD patients with different levels of functional impairment compares to that of non-institutionalized controls. The first part of this dissertation explored Alzheimer patients (AD) and age-matched controls' performance on four parallel forms of explicit and implicit memory tests employing written and spoken words, pictures, and common objects (e.g., toothbrush). For each test, subjects encoded critical targets by first identifying, and then generating a personal meaning for each item. The second part of this dissertation consists of three experiments that explored whether encoding and retrieval (i.e., test) strategies could be used to elevate the explicit memory performance of AD patients. The average age of the 20 AD patients (9 possible and 11 probable using NINCDS-ADRDA criteria; 13 female, 7 male) was 70.8 years (range = 50-84 years, SR = 8.2) with a mean of 12.5 years of education (range = 6-18 years, SD = 3.3). The Functional Rating Scale (Tuokko & Crockett, 1989) indicated that 6 of these  iii  functionally impaired. The average age of the 40 control subjects (27 female, 13 male) was not statistically different from that of the AD patients (67.8 years, range = 51-89 years, 2p = 9.9) although controls reported a significantly greater number of years'of education (i.e., H = 14.83 years, range = 8-27 years, = 4.2). The overall hypothesis of part one was that differences in the memory performance of AD patients versus controls would be smaller on implicit than explicit tests. Results showed that this hypothesis was supported for tests of written and spoken words (i.e., using category completion and category cued recall tests) and for common objects in a perceptual learning condition (i.e., the same objects re-presented over two trials using a tactile identification and a tactile recognition test). In contrast, this hypothesis was not supported for tests using picture materials (i.e., using a picture fragment and a picture recognition test) or for common objects in a skill-based learning condition (i.e., different objects presented over two trials in a tactile identification and a tactile recognition test). These findings support previous reports that AD patients exhibit impaired performance on picture fragment completion tests. In contrast, these results do not support previous work which has indicated that performance on semantic priming tests is also impaired in these patients. This study extends current work by finding that implicit memory for spoken words and common objects need not be impaired in these patients. In addition,  iv  that AD patients with mild, moderate, and severe levels of functional impairment can show a similar magnitude of priming on some tests. Results from this study are discussed in terms of the influence that the encoding task, stimuli, and test form may have had on memory performance. The overall hypothesis for part two was that the explicit memory performance of AD patients would be elevated using specific encoding and retrieval conditions. This hypothesis was supported in the levels-of-processing and partially supported for the Subject Performed-Experimenter Performed (i.e., SPT-EPT) and the multisensory (i.e., See, Say, and Do) experiments. Results from this study extends previous work by showing that AD patients' explicit memory abilities can be significantly elevated when performance and meaning-generated encoding strategies are used in combination with a cued recall or recognition test retrieval condition. In addition, the ability of these encoding and retrieval strategies to raise the explicit memory performance of AD patients was found to decline as these patients' level of functional impairment increased. These findings are discussed in terms of the possible mechanisms by which these encoding and retrieval strategies influenced the memory performance of AD patients.  TABLE OF CONTENTS ABSTRACT^  page ii  TABLE OF CONTENTS^  page v  LIST OF TABLES^  page vii  LIST OF FIGURES^  page ix  LIST OF APPENDICES^  page x  ACKNOWLEDGEMENTS^  page xi  CHAPTER ONE: INTRODUCTION AND OVERVIEW Study One Study Two Summary of Results Remaining Chapters  page page page page page  1 3 6 7 9  CHAPTER TWO: INTRODUCTION TO ALZHEIMER'S DISEASE Chapter Overview^ page 11 Behavioral and Anatomical Characteristics page 11 Neuropathological Development and Changes in Memory Associated with Damage to the: Hippocampus ^page 14 Hippocampus and Surrounding Areas^page 22 Outside the Hippocampus^page 28 CHAPTER THREE: EXPLICIT AND IMPLICIT MEMORY ^page Chapter Overview^ page Introduction^ page Multiple Systems Model ^ page Process Model^ page Neuropsychological Investigations ^page Cognitive Investigations^page Study One^ page  34 34 34 36 37 44 59 67  CHAPTER FOUR: MEMORY STRATEGIES^ page 70 Chapter Overview^ page 70 Introduction^ page 70 Retrieval Strategies ^ page 70 Encoding Strategies^ page 71 Subject Performed Tasks (SPTs) ^page 73 Levers af Procesoing ^page^77 Theoretical Explanations^page 82 Study Two^ page 84 -  —  —  —  vi CHAPTER FIVE: METHOD Chapter Overview Subjects Recruitment and Selection Criteria Comparison of Controls and AD Patients Medications Demographics Explicit and Implicit Memory Tests Materials Procedure Memory Strategy Experiments Materials Procedure  page page page page page page page page page page page page page  86 86 86 86 90 90 91 91 91 95 105 105 107  CHAPTER SIX: RESULTS Chapter Overview Overall Design and Analyses Memory Tests Tests of Written Word Materials Tests of Spoken Word Materials Tests of Picture Materials Tests of Object Materials  page page page page page page page  120 120 122 123 129 133 138  page page page page  145 146 151 157  CHAPTER SEVEN: DISCUSSION Chapter Overview Part One Part Two Limitations of Work Future Work Contributions  page page page page page page page  164 164 165 177 187 190 191  REFERENCES^ APPENDICES^  page 193 page 218  Memory Strategy Experiments Levels of Processing SPT/EPT Multisensory  vii  List of Tables  PAGE  Table 1^Demographic Characteristics of Controls and^113 Alzheimer Patients. Table 2^Medications taken by Controls and Alzheimer^114 Patients. Table 3^Statistical Comparison of Controls versus AD Patients.^  115  Table 4^Explicit and Implicit Memory Tests Used in Study One.^  116  Table 5^Counterbalancing Methods Used for Explicit and Implicit Memory Tests. ^  117  Table 6 Summary of Memory Strategy Experiments Used in Study Two.^  118  Table 7 Counterbalancing Methods Used for Memory Strategy 119 Experiments. TableG-8a Controls' and AD Patients' Performance on ^306 Category Cued Recall and Category Completion Tests for Written Word Materials. TableG-8b Two Factor Repeated Measures MANCOVA of Controls' 307 and AD Patients' Performance on Category Cued Recall and Category Completion Tests for Written Word Materials. TableG-9a Controls' and AD Patients' Performance on ^308 Category Cued Recall and Category Completion Tests for Spoken Word Materials. TableG-9b Two Factor Repeated Measures MANCOVA of Controls' 309 and AD Patients' Performance on Category Cued Recall and Category Completion Tests for Spoken Word Materials. TableG-10a Controls' and AD Patients' Performance on ^310 Picture Recognition and Picture Fragment Completion Tests. TableG-10b Two Factor Repeated Measures MANCOVA of ^312 Controls' and AD Patients' Performance on Picture Recognition and Picture Fragment Completion Tests.  viii  List of Tables ^ TableG-ila Controls' and AD Patients' Performance on 313 Tactile Recognition and Tactile Identification Tests. TableG-lib Two Factor Repeated Measures MANCOVA of Control 315 and AD Patients' Performance on the Tactile Recognition and Tactile Identification Tests (Old Materials Condition). TableG-lic Two Factor Repeated Measures MANCOVA of Control 316 and AD Patients' Performance on the Tactile Recognition and Tactile Identification Tests (New Materials Condition). TableG -12a Controls' and AD Patients' Performance in the ^317 Levels of Processing Memory Strategy Experiment. ^ TableG-12b Three Factor Repeated Measures MANCOVA of 318 Controls' and AD Patients' Performance in the Levels of Processing Experiment. TableG-13a Controls' and AD Patients' Performance in the^319 SPT-EPT Memory Strategy Experiment. TableG-13b Three Factor Repeated Measures MANCOVA of ^320 Controls' and AD Patients' Performance in the SPT-EPT Memory Strategy Experiment. TableG -14a Controls' and AD Patients' Performance in the ^322 Multisensory Memory Strategy Experiment. TableG-14b Three Factor Repeated Measures MANCOVA of ^323 Controls' and AD Patients' Performance in the Multisensory Memory Strategy Experiment.  ix List of Figures PAGE Figure 1 Diagram of the Brain and Hippocampal Formation. ^21 Figure 2 Controls' and Alzheimer Patients' Performance on 125 Category Cued Recall and Completion Tests for Written Word Materials. Figure 3 Controls' and Alzheimer Patients' Performance on 130 Category Cued Recall and Completion Tests for Spoken Word Materials. Figure 4 Controls' and Alzheimer Patients' Performance on 135 Picture Recognition and Fragment Completion Tests. Figure 5 Controls' and Alzheimer Patients' Performance on 141 Tactile Recognition and Identification Tests. Figure 6 Controls' and Alzheimer Patients' Performance on 148 Free Recall and Recognition of targets in the Levels of Processing Memory Strategy Experiment. Figure 7 Controls' and Alzheimer Patients' Free and Cued 153 Recall of targets in the Subject Performed Memory Strategy Experiment (SPT-EPT). Figure 8 Alzheimer Patients and Controls' Free and Cued ^159 Recall of targets in the Multisensory Memory Strategy Experiment.  LIST OF APPENDICES  Appendix A  ^  Letters of contact and consent for ^page 218 Alzheimer patients, controls and Alzheimer caregivers. NINCDS-ADRDA guidelines for making ^page 224 a diagnosis of Alzheimer's Disease and Functional Rating Scale criteria for measuring stage of Functional Impairment.  Appendix B  Appendix C  ^  Procedure and questions used to ^page 228 select volunteers to act as non-demented control subjects.  Appendix D^Materials used in Explicit and Implicit Memory Tests.  page 231  Appendix E^Test booklet to record subject responses.  page 290  Appendix F^Materials used in Memory Strategy Experiments.  page 301  Appendix G^Tables reporting descriptive and inferential statistics.^ page 305  xi  ACKNOWLEDGEMENTS This thesis would not have been possible without the personal and financial commitment of several individuals and organizations. First, without the support of Dr. B. Lynn Beattie, Director of the Clinic for Alzheimer Disease and Related Disorders-University Hospital-UBC site, this study would never have happened. Dr. Beattie provided generous access to the clinic and integral help with patient recruitment. I would also like to acknowledge the work of Dr. Peter Graf, my program supervisor; Despite our very different academic backgrounds Peter persisted in his attempts to navigate me through what turned out to be a difficult process for both of us. Finally, I am indebted to Dr. Holly Tuokko for her guidance during the arduous task of data collection and to Dr. Jonathan Berkowitz for statistical advice. Financial support for this study came from several sources. Two research grants from Sigma Xi and funding from the I.O.D.E. supported project and equipment expenses. The Alzheimer Society of B.C. provided generous assistance that supported me for two years during the collection of data for this study. I am also indebted to UCLA and the University of Southern California for travel support that allowed me to interact with researchers dealing with issues directly related to this study. Last, but of equal significance is the personal support I received from friends that kept me writing during the difficult final stages of this work. Special thanks to Drs. Paula Brook and Sandra Clark, Ken Gallie and Eunice Williams.  1 CHAPTER 1: INTRODUCTION AND OVERVIEW  Patients with Alzheimer's disease (AD) provide a unique model for investigating the relationship between human memory and the brain. This is because the neuropathology caused by this disease follows a consistent pattern that begins in the hippocampus before spreading to specific cortical association sites (Ball, 1977; Damasio, Van Hoesen, & Hyman, 1990; Lewis, Campbell, Terry, & Morrison, 1987). This allows the researcher to investigate an intentional or explicit form of memory that is associated with hippocampal functioning (Squire, 1992a; ZolaMorgan & Squire, 1992, p. 333) and to chart how this type of memory changes as damage accumulates in the hippocampus and surrounding cortical areas (cf. Squire, 1992a). It also allows the investigator to examine a non-intentional or implicit form of memory that research suggests may depend on modality-specific cortical association areas (Squire, 1992a; Tulving & Schacter, 1990). To date, our interpretation of the way memory is changed by AD has been influenced by our existing knowledge of the functions that these anatomical regions are presumed to have on memory performance. For example, AD patients' ability to perform explicit memory tests is always impaired in comparison to that of • neurally-intact controls and ^  2 to damage to hippocampal regions (Damasio, Van Hoesen, & Hyman, 1990; Van Hoesen & Damasio, 1987). In addition, AD patients' ability to perform implicit memory tests (with the exception of motor-based tasks) also appears to be impaired (Butters, Heindel, & Salmon, 1990; Salmon & Heindel, 1992). The retained ability to perform implicit motor-based tests has been related to the fact that the cortical areas associated with somatosensory and motor processing, along with related subcortical areas, are spared by AD (Damasio et al., 1990; Eslinger & Damasio, 1986). These findings and their interpretations have been integral to the development of a deficit memory model of AD -- that is, memory  .  abilities are irrevocably lost as a direct cause of neuronal death. A new phase in our thinking about memory processes may be emerging since researchers are now advocating that greater consideration be given to the role that influences like attention and stimulus characteristics have on memory performance (Light & La Voie, in press; Nebes, 1992; Zola-Morgan, 1993). These ideas seem particularly well-suited to study with AD patients who, it is now being recognized, also experience attentional, as well as other types of general cognitive impairments (Filoteo, Delis, Massman, Demadura, Butters, & Salmon, 1992; Graf, Tuokko, & Gallie, 1990; Salmon & Heindel, 1992).  3 Study One  The first study of this investigation explored the memory performance of AD patients using an encoding method that ensured that these attentionally-impaired patients had attended to, and  processed critical target stimulus (refer to chapters 3 & 5). Research suggests that AD patients will not spontaneously encode new information and most previous investigations have not used study methods that compensate for these patients' attentional deficits (Rohling, Ellis, & Scogin, 1991; Strauss, Weingartner, & Thompson, 1985). Memory performance was examined on eight different explicit and implicit tests. Based on previous research, the premise was that explicit tests engaged primarily hippocampal-related  activity and implicit tests required predominantly nonhippocampal or cortical association activity (Squire, 1992b). Because priming or implicit test performance may be dependent on modality-specific perceptual processes (Squire, 1992a) the materials used in these tests were chosen for their ability to tap different sensory modalities. For example, written word and picture materials were chosen since they primarily tap a visual type of processing, whereas spoken word and tactilely presented objects required that subjects engage primarily auditory and somatomotor processing, respectively.  4  Research from cognitive psychology shows that the same test forms should be used when comparing explicit and implicit test performance since variations in the way that targets are studied and retrieved can have very different effects on memory performance (Howard, Fry, & Brune, 1991; Roediger, 1990a, 1990b). For this reason the explicit and implicit tests for each type of stimulus material (i.e., written and spoken word, picture, and objects) were designed so that the same test forms were used. Specifically, the same encoding conditions and materials were used for each set of explicit and implicit tests but they differed in the retrieval conditions that occurred. For example, the explicit forms of each test employed retrieval instructions that encouraged the subject to intentionally recollect previously presented stimuli and the implicit forms of each test did not (refer to Chapter 3). The overall hypothesis of this study was that differences in the memory performance of AD patients and controls would be smaller on implicit than explicit tests. This hypothesis was based on the idea that the hippocampus, which is integral to performing explicit tests, is destroyed early in AD, whereas areas outside the hippocampus, important to performing implicit tests are damaged later (Ball, 1987; Price, Davis, Morris, & White, 1991; refer to Chapter 2). This hypothesis was examined  explicit and implicit tests for-muditarally,^ -  5 visually, and somatomotor processed materials. Follow-up analyses were used to determine whether the explicit and implicit test performance of AD patients was significantly different from controls. The results from this study and the second study in this investigation were based on the performance of 20 patients diagnosed with possible or probable AD using NINCDS-ADRDA criteria (McKhann et al., 1984) and 40 non-institutionalized, age-matched controls. The 20 AD patients represented approximately equal numbers of mildly (n = 6), moderately (n = 7), and severely (n = 7) functionally impaired (F.I.) patients as indexed using the Functional Rating Scale developed by Tuokko and Crockett (FRS: 1989; see Appendix B for a copy of the FRS). Employing approximately twice the number of AD patients that have been included in most extant studies on implicit and explicit memory, as well as obtaining equal representation from different F.I. levels, decreased the chances of this being a biased patient sample. Secondly, it allowed the investigation of whether the implicit test performance of AD patients remains stable or is variant across F.I. groups (F.I. was presumed to provide a rough index of the extent of neuropathology as per Price, Davis, Morris, & White, 1991 and Morris et al., 1991). In previous studies we have found that the explicit test performance of patients declines from mild to severe F.I. groups, as one  6 would expect in a condition where there is an increasing amount of neuropathology to the hippocampus and surrounding cortical areas (Gallie, Tuokko, & Graf, 1991; Tuokko, Gallie, & Crockett, 1990; cf. Squire, 1992a). Using the same patients and controls in study one and in study two ensured that performance variation due to individual differences or different levels of neuropathology was held constant across the memory tests and experiments reported here.  Study Two The second study in this investigation included three different experiments that explored the ability of various types of encoding and retrieval strategies to elevate the explicit memory performance of AD patients. These experiments also enabled a closer examination of some of the assumptions and results that were obtained from the first study. The first experiment examined whether requiring AD patients to study critical targets for their meaning significantly elevated their memory test performance. This was examined in a levels of processing design modelled after the work of Craik and Lockhart (1972; Lockhart & Craik, 1990). The second and third experiments examined whether AD patients recollected more critical targets they had encoded using a performance versus a non-performance based encoding task. The second experiment was based on the subject-performed, experimenter-performed work of Cohen (1983;  7 Cohen & Bryant, 1991) and Backman and Nilsson (Backman & Nilsson, 1989; Nilsson & Backman, 1991). The third experiment was a multisensory experiment and was not based on previous work. All three experiments allowed me to investigate an additional question and this was whether the memory problems of AD patients were irrevocably lost or impaired due to a combined retrieval and encoding deficit (Gallie, Graf, & Tuokko, 1990). The idea for this work was based on existing knowledge of the brain areas that are damaged in AD and their presumed functional roles, as well as theories found in the cognitive literature. Specifically, the hippocampus plays a role in the retrieval of information and other brain areas such as the locus coereleus, basal forebrain, hippocampus, and parietal lobe play a role in the attentional focusing or ability to encode target stimuli (Squire, 1992a; Mountcastle, 1978; Van Hoesen & Damasio, 1987). Thus, the premise in study two was that when AD patients were engaged in tasks that assisted them to encode and retrieve information this would elevate their memory performance. Summary of Results The results obtained from study one provided evidence that: (1) contrary to previous findings, implicit memory for written and spoken word materials is not always impaired in AD patients. (2) support was found for Larry Squire's (1992a) idea that priming or implicit memory performance may be AAtflummmera by slight changes in the form in which the stimulus is presented at study and test. This was shown by the fact that AD patients' performance -  8 on the implicit test of object materials was not significantly different from that of controls when priming was based on a re-presentation of the same materials over two trials. However, significant differences were found when priming was based on the study-test performance of two different sets of object stimuli. (3) as found in previous studies, AD patients' implicit test performance for picture materials was significantly impaired when compared to that of controls. Results suggest that the heavy visuospatial and abstract reasoning demands of this task may have contributed to this finding. (4) unlike explicit test performance, the implicit test performance of mildly, moderately, and severely F.I. patients was similar on several tests used in this study (note, this performance was not at floor). The results obtained from study two provided evidence that: (1) the encoding method employed in study one that required subjects to identify and then generate the meaning of critical targets significantly elevated the memory performance of AD patients. .  (2) somatomotor encoding strategies are more effective than non-performance strategies in elevating the memory performance of AD patients in certain conditions. (3) contrary to pre-existing ideas, the explicit memory performance of AD patients can be significantly elevated when a combination of encoding and retrieval strategies are used. However, the extent to which these strategies elevate performance appears to depend on the patients' F.I. level.  9 Overview of Remaining Chapters This thesis includes six additional chapters. Chapter 2 examines some of the advancements in our knowledge of how AD changes the brain. Emphasis is placed on the brainbehavior relationship of AD and the specific changes in intentional or explicit memory performance that occur in this condition. Chapter 3 reviews the neuropsychological investigations that have provided evidence of the brain regions that might be responsible for performing specific types of implicit memory tests. Work that has focused on the cognitive processes that are involved in performing these memory tests is considered and the overall hypothesis of study one is introduced. Chapter 4 introduces the concept of encoding and retrieval (i.e., test types) strategies. The limited studies that have used memory strategies with AD patients are reviewed, together with the theories addressing how these strategies might work. Study two and its overall hypothesis is then introduced. Chapter 5 describes the method used in this dissertation. The first section discusses the criteria used to select AD patients and controls and the demographic characteristics of these subject groups. The next sections describe how materials were chosen and the procedures that were employed in the memory test and strategy experiments.  10  Chapter 6 reports the results of this investigation. The first section describes the statistical design and general approach taken in analyzing results. The second and third sections report the results of the memory test and strategy experiments, respectively. Chapter 7 is a general discussion of the results that were obtained. In this chapter the main findings from study one and from study two are discussed. The limitations as well as the way this investigation guides my future work is also considered. This chapter ends with a summary of the global and specific contributions of this work.  11  CHAPTER 2: INTRODUCTION TO ALZHEIMER'S DISEASE This chapter examines some of the behavioral and neuroanatomical characteristics of Alzheimer's Disease. Emphasis is placed on brain-behavior relationships and the intentional or explicit memory changes that occur in this condition. Behavioral and Neuroanatomical Characteristics Alois Alzheimer first described the condition we now call Alzheimer's Disease at a German Psychiatric conference in 1906 (Alzheimer, 1907; Hippius, 1990). In his presentation Alzheimer reported the clinical symptoms of a 51 year old patient, Frau A.D., who over a four year period showed changes in both the type and severity of her behavioral symptoms. Frau A.D.'s symptoms began with problems in remembering recent events, like what she had bought at the store that day. With time these memory problems changed and she developed an inability to recollect information learned many years previous, such as her house address. During the time Alzheimer observed this patient he noticed that she also developed other symptoms. One symptom was visuospatial confusion when attempting to find the rooms in the house she had lived in for many years. The other symptom was the development of a progressive language impairment shown by Frau A.D.'s increasing inability to find the correct words to express herself.  12 Alzheimer's presentation of Frau A.D. became significant when he suggested that the behavioral symptoms observed in this patient might be related to the results of her autopsy (Hippius, 1990). Using a recently developed silver stain, Alzheimer had detected neuritic plaques and neurofibrillary tangles in the cerebral cortex and hippocampus of this patient (Hafner, 1990). Contemporary research has provided additional details about these plaques and tangles, and their presumed role in causing the behavioral symptoms of AD patients. Neuritic Plaques and Neurofibrillary Tangles Neuritic plaques (NPs) are extracellular brain lesions that are currently regarded as the primary sites of damage caused by AD (Mann, 1989). In contrast, neurofibrillary tangles (NFT) are intracellular lesions that occur within neurons thought to synapse with cells involved in NP formation (Jellinger, 1990; Mann, 1989). Together the presence of NPs and NFTs are required before a definite diagnosis of AD can be made (Iqbal, 1991; Jellinger, 1990). Although there is no consensus on what causes NPs and NFTs to form, this has not stopped researchers from  acquiring a basic understanding of their structure (Iqbal, 1991; Jellinger, 1990). Structure. Neuritic or senile plaques (NPs) are approximately 50 - 200 micrometer in size and most have a central _corexd'_anyloid  protein fLandomr& Kidd, 19239; Mann, 1989). All  13 plaques are surrounded by neurites that contain paired helical filaments similar to those found in NFTs (Mann, 1989). Presently there is no consensus regarding whether NPs and NFTs are related structures (Arnold, Hyman, Flory, Damasio, & Van Hoesen, 1991; Mann, 1989). The method by which NPs develop is not known, but one of the more prominent theories suggests that it is the presence of amyloid protein that triggers NP formation (see Blass, Li-wen Ko, & Wisniewski, 1991; Landon & Kidd, 1989). The mechanism by which amyloid enters the brain is not known, but its presence is thought to sever axonal and dendritic connections and eventually cause neuronal death (Mann, 1989). As these neurons die they move towards the amyloid to form the plaque (Wisniewski, 1988). Neurofibrillary tangles (NFTs) are the other histological sign of AD (Flament-Durand & Brion, 1990; Lewis, Campbell, Terry, & Morrison, 1987). These structures are composed of paired 10 nm filaments that wind about each other with a periodicity of 80 nm (Mann, 1989). As NFTs develop within the neuron they disrupt axonal transport and metabolic functioning (Flament-Durand & Brion, 1990). At some point these tangles impede the neuron's ability to maintain its metabolic functions, and this results in cell death (Flament-Durand & Brion, 1991; Mann, 1989). Thus, in combination, the development of NPs and NFTs cause brain neurons to  ^TTbIeddho-GA 6 ca, 1988; Wisniewski, 1988). -  -  14 Since neurons are postmitotic, dying cells are not replaced and this causes serious disruptions in interneuronal communication (Damasio, Van Hoesen, & Hyman, 1990; Kandel & Schwartz, 1985). Although surviving neurons may compensate for cell death via axonal sprouting, the ability to reinstate interneuronal communication is limited (Gertz & Cervos-Navarro, 1990). Evidence that axonal sprouting provides limited functional compensation is provided in the following discussions of the areas that NPs and NFTs develop and the parallel disruption/distortion of behavior (Arnold et al., 1991; Van Hoesen & Damasio, 1987).  NPs and NFTs and the Behavioral Symptoms of AD Patients NP and NFT Development Begins in Hippocampal Regions. Autopsy results show that NPs and NFTs always occur in the hippocampus of demented, but not in that of non-demented elderly (Hyman, Van Hoesen, Kramer, & Damasio, 1986; Mann & Esiri, 1988; Morris et al., 1991; Price et al., 1991). These consistent findings have led to the accepted notion that NPs and NFTs signal the presence of AD (Arnold, Hyman, Flory, Damasio, & Van Hoesen, 1991; Lewis, Campbell, Terry, & Morrison, 1987). Evidence that AD originates in hippocampal and/or closely located brain structures (i.e., the olfactory bulbs and amygdala) has been slower to develop since this information has accrued rom  •b  rice e^1991). Initially, autopsies  15 performed on mildly cognitively impaired patients (which clinicians assume are in the early stages of neuropathology) had shown that NP and NFT development was restricted to hippocampal regions (Ball, 1977; Ball et al., 1985; Price et al., 1991). Autopsy results based on moderately and severely impaired patients revealed a pattern of pathology that suggested that NPs and NFTs had originated in the hippocampus before spreading to other brain regions (Price et al., 1991). Recent advances in magnetic resonance imaging methods have allowed investigators to obtain more accurate images of the hippocampal region (deLeon, George, Stylopoulous, Miller, & Smith, 1990a). These advances have provided researchers with direct evidence that the hippocampus is the first area in which brain changes (i.e., atrophy) occur in AD patients (deLeon et al., 1990b). The Hippocampus and Memory. Milner's extensive testing of patient H.M. (Milner, 1958), Squire's work with patient R.B. (Squire, 1992a; Zola-Morgan, Squire, & Amarall, 1986) together with research employing animal models (i.e., Mahut, Zola-Morgan, & Moss, 1982; Mishkin, 1978) have consistently shown that the hippocampus is integral to intentional or explicit memory performance (see Squire, 1992a for an extensive review on this topic). This would suggest that if the pathology associated with AD originates in hippocampal regions, then memory problems should  16 be one of the first symptoms to be exhibited by these patients. This correspondence between brain pathology and behavioral symptoms is supported by clinical research (Van Hoesen & Damasio, 1987; Morris & Rubin, 1991). Specific analyses of the hippocampal regions in which NPs and NFTs form, and the pattern in which they develop appears to be closely correlated to the intentional memory changes that occur in AD patients: for example, the initial fluctuations in the ability to recollect recently acquired information that eventually develop into a consistent problem when recollecting information acquired in the distant past (McKhann et al., 1984; Reisberg, 1983).  Impairment in Intentional and Short Term Memory Abilities. Neuropathological investigations show that one of the first areas that NPs and NFTs develop is in the CA1 subfield of the hippocampus (Morris et al., 1991; Price et al., 1991). Since the initial brain changes caused by AD are restricted to the CA1 region (Morris et al., 1991; Price et al., 1991) this permits a comparison to be made with patient R.B. The most extensively documented case of bilateral damage restricted to the entire rostral-caudal extent of CA1 was patient R.B. (Zola-Morgan, Squire, & Amaral, 1986). Comprehensive neuropsychological testing of this patient showed that his memory problems  were gemerally-confIned to an inability^to intentionally -  -  -  17 recollect recently acquired information (Zola-Morgan, Squire, & Amaral, 1986). For example, R.B. suffered from a short-term memory impairment for all types of stimuli whether it was auditory, visual, pictorial, or somatosensory in nature. This was shown by this patient's impaired performance on tests of story and word recall, immediate reproduction of the ReyOsterreith figure, as well as an inability to remember which of his fingers had recently been touched by the examiner (i.e., behind a partition; Zola-Morgan, Squire, & Amaral, 1986). In contrast, R.B. did not appear to have a profound longterm memory impairment. For example, his performance was similar or better than that of controls on the famous faces and the television test that required that he correctly identify celebrities and television programs that had been prominent before his ischemic attack (Cohen & Squire, 1981; Zola-Morgan, Squire, & Amaral, 1986). R.B. also exhibited a retained ability to engage in the free recall and the recognition of both public and personal events that had occurred before his accident (Squire, 1987). However, there was some evidence that this patient might have had a slight retrograde amnesia for events that had occurred one to two years before his accident (ZolaMorgan, Squire, & Amaral, 1986, p. 2954). This period of retrograde amnesia corresponds to that found in patients undergoimselem:ta=lonstuas4mv-therapy and to ide m5 regarding -  18 consolidation and the ongoing role of the hippocampus for a time after the initial learning event (Squire, 1987; Squire & Cohen 1979; Squire, Slater, & Chace, 1975). Additional testing of R.B. did not detect any significant cognitive impairment other than for memory. For example, this patients' performance on most scales of the Weschler Adult Intelligence Scale was above normal for both the verbal and the performance components of this test (Zola-Morgan, Squire, & Amaral, p. 2954). Use of a non-traditional memory test (i.e., word stem completion) further indicated that R.B. showed an ability to non-intentionally recollect recently studied items (Zola-Morgan, Squire, & Amaral, 1986, p. 2964). Thus, the neuropsychological assessment of R.B. indicated that his cognitive impairment was generally restricted to an inability to intentionally recollect recently experienced information. Memory deficits similar to R.B.'s have now been documented in patients with damage to CA1 and surrounding hippocampal areas (see Squire, 1992a; Victor & Agamanolis, 1990). The consistency of these findings support the idea that the memory impairments documented in R.B. were related to damage to the hippocampus and that this area performs "some computation upon newly processed information" (Squire, 1987, p. 194). Furthermore, since the first memory problems that AD patients encounter is an inability to intentionally recdII dbt newly encountered information, it -  19 seems reasonable to assume, based on our present knowledge, that these changes are related to NP and NFT development in the CA1hippocampal area (Damasio et al., 1990). Insidious Onset. Differences between R.B.'s relatively sudden, and AD patients' gradual inability to intentionally recollect newly acquired information can be explained in terms of the processes that damage CA1 regions. The ischemic episodes experienced by R.B. appear to have occurred over a relatively short time period (i.e., months) in comparison to AD which is estimated to take anywhere between 3 to 15 years to completely develop (Huff, Growdon, Corkin, & Rosen, 1987; Van Dijk, Dippel, & Habbema, 1991). Additionally, the ischemic episodes experienced by R.B. appear to have totally destroyed his CA1 regions (Zola-Morgan, Squire, & Amaral, 1986). In contrast, an incremental development of NFTs and NPs occur in the CA1 regions of AD patients (Price et al., 1991). Since the functional ability of CA1 regions presumably depends on the remaining number of unaffected neurons (see Damasio, Van Hoesen, & Hyman, 1990; Van Hoesen & Damasio, 1987) an incremental accumulation of neuropathology should result in an insidious, rather than an abrupt onset of memory problems. This is the pattern of memory impairment that is observed in AD patients (McKhann et al., 1984).  20 Increasing Severity. The severity of R.B.'s intentional memory deficits did not appear to drastically change over time, unlike that of AD patients (McKhann et al., 1984; Reisberg, 1983; Van Hoesen & Damasio, 1987). Recent work with monkeys in combination with neuropathological investigations of AD patients reveals some of the neural substrates that might be responsible for the increasing memory difficulties encountered by these patients (Squire, 1992a; Price et al., 1991). A brief review of the neuroanatomy of the hippocampus and surrounding structures is required to understand this work. Anatomical Relationships. The hippocampus is a member of the functional unit referred to as the hippocampal formation (Barr & Kiernan, 1988). This formation, depicted in Figure 1, consists of several structures including the parahippocampal gyrus that curves into the subiculum, which in turn merges to form the hippocampal and CA1 regions (Barr & Kiernan, 1988). The hippocampal formation is connected to the amygdala and its overlying perirhinal cortex via the entorhinal cortex (Barr & Kiernan, 1988; Carpenter, 1985). As the entorhinal cortex curves to form the parahippocampal gyrus, it brings with it major fibre bundles that form the perforant pathway that runs within the  1 R.B. may have experienced a more severe memory impairment during the first year of his ischemic attack than in subsequent • • • •:•P• owever, this pattern is opposite to that observed in AD patients where the memory impairment increases with time (McKhann et al., 1984).  21  Optic Tract  Ilk  -  Caudate Nucleus  Inferior Born of the Lateral Ventricle  Fimbria/Fornix  / Rippocampus Subiculum  &  if \ /:,^Perforant Pathway  / '' rhi.. _ ..41  •  cortex  Amygdala  Perirhinal Cortex  Parahippocampal Cyrus  Figure 1. Schematic drawing in coronal perspective of the Rippmcampal Formation with medial surface at left. • Location of Amgdala not exact.  approximate area of basal •anglia  PRIMARY MOTOR AREA  SOMATOSENSORY PARIETAL LOBE  OCCIPITAL CORTEX approximate area Locus ceruleus of cerebellum  BASAL FOREBRAIN  BERSCBEL'S GYROS  (nucleus basali■ of Meynert)  (auditory) TEMPORAL LOBE approximate area of spinal cord  Lateral View of the Brain (stiple indicates distribution and location of Neurofibrillary Plaques and Tangles as per Brun. 1983)  22 hippocampus (Carpenter, 1985). Based on investigations with monkeys, it appears that the perforant pathway is comprised of fibres that originate from isocortical association areas and carry most of the auditory, visual, somatosensory and motorrelated information that enters the hippocampus (Arnold et al., 1991; Carpenter, 1985; Hyman, Van Hoesen, Kromer, & Damasio, 1986; Insausti, Amaral, & Cowan, 1987; Squire, 1992a). The perforant pathway, in a series of intrinsic intrahippocampal connections, ends in the subiculum and CA1 regions (Arnold et al., 1991). From the subiculum and CAI regions, a reciprocal pathway exits the hippocampus carrying information back to the cortex, thalamus, and hypothalamus (Arnold et al., 1991; Braak & Braak, 1990). Information from other brain areas (primarily subcortical) enter and exit the opposite end of the hippocampal formation through the fimbria/fornix (Barr & Kiernan, 1988). Animal model investigations proffer more direct information on some of the neuroanatomical structures responsible for variations in the severity of memory impairments (Squire, 1992a; Squire & Zola-Morgan, 1991). In these investigations monkeys and rats have been subjected to bilateral, circumscribed lesions limited to particular structures, or to a combination of neuroanatomical structures located in the medial temporal lobe (Squire & Zola-Morgan, 1991; Zola-Morgan & Squire, 1992). The effects^ of these anatomical removals on memory performance are -  -  23 then analyzed by having these lesioned animals perform memory tasks in which amnesic patients exhibit similar memory performance (Squire & Zola-Morgan, 1991; Zola-Morgan, 1993). In many of these animal studies the delayed nonmatching-to sample (DNMS) task has been employed (Squire & Zola-Morgan, 1991). Both AD patients and lesioned monkeys show similar performance on DNMS tasks (Albert & Moss, 1984; Albert, Moss, & Milberg, 1989). In combination, the results obtained from these animal investigations have shown that it is the extent of damage to cortical areas surrounding the hippocampus and amygdala (i.e., entorhinal and perirhinal cortex, respectively) that is correlated to the level of performance on the DNMS test (Squire & Zola-Morgan, 1991; Zola-Morgan, Squire, Amaral, & Suzuki, 1989). For example, monkeys with bilateral damage to the hippocampus, amygdala, and the cortical areas surrounding these structures (designated as a H+A+ lesion) exhibit a more severe memory impairment than animals with bilateral lesions restricted to the hippocampus and its surrounding cortex (i.e., an H+ lesion) (Mishkin, 1978; Zola-Morgan & Squire, 1985, 1986). The main difference in these two models is damage to the amygdala and surrounding perirhinal cortex. Since animals with bilateral lesions restricted to the amygdala (i.e., an A lesion) show normal performance on the DNMS task (Zola-Morgan, Squire, & Amaral, 1989) damage to the perirhinal cortex appears responsible  24 for the increased memory deficit (Squire, 1992a; Squire & ZolaMorgan, 1991; Zola-Morgan, Squire, Amaral, & Suzuki, 1989). Additional support for the theory that damage to cortical areas surrounding the amygdala and hippocampus influence memory severity is provided by other studies. For example, the DNMS performance of animals with H+A+ lesions has been found to be similar to H++ animals when damage is restricted to the hippocampus and to cortical areas surrounding the hippocampus and amygdala (Glower, Zola-Morgan, & Squire, 1990). Similarly, the DNMS performance of animals with lesions restricted to the perirhinal-parahippocampal gyrus (i.e., the PRPH lesion) is similar to that of animals with H+A+ lesions (Zola-Morgan, Squire, Amaral, & Suzuki, 1989). In combination, these findings have led animal investigators to conclude that damage to the perirhinal and surrounding cortical areas -- not the amygdala -- are responsible for the greater memory impairment observed in H+A+ animals (Squire, 1992a; Squire & Zola-Morgan, 1991, Zola-Morgan & Squire, 1992). Subsequent re-analysis of the lesioned areas produced in these investigations support this conclusion. Specifically, that "the perirhinal cortex [is] the only area where the H++ and PRPH monkeys sustained more damage than the H+ and the H+A monkeys" (Squire & Zola-Morgan, 1991, p. 1383). As Squire (1992a) has astutely-Gernmented- -"herein-lies-  25 more amnesic than other amnesic study patients, including R.B. H.M. sustained an H+A+ lesion, but R.B. sustained a lesion involving only a portion of the hippocampus" (p. 201) that did not include adjacent perirhinal and entorhinal cortices. Studies based on monkeys indicate that approximately two thirds of all cortical input to the hippocampus travels through the entorhinal cortex (Insausti, Amaral, & Cowan, 1987; Squire 1992a; Zola-Morgan, 1993). Thus, depending on the extent of damage to the entorhinal cortex and areas that project to it (i.e., the perirhinal cortex) this would result in a related decline in the amount of information entering the hippocampus. This would be exhibited as a memory impairment that was presumably proportional to the extent of damage that had occurred (Hyman, Van Hoesen, Kromer, & Damasio, 1986; Squire, 1992a). Price et al. (1991) have recently uncovered a pattern of neuropathology in AD patients that corresponds to the anatomical damage induced to the perirhinal and entorhinal areas in animal investigations. In a rare investigation that compared the distribution of NFTs and NPs in AD patients at different levels of memory severity, Price et al. (1991) found evidence that NFTs and NPs spread from the CA1 region to the perirhinal and entorhinal areas surrounding the hippocampus. The pattern of neuropathology found by Price et al., suggests that the number of NPs and NFTs in perirhinal and  26 entorhinal areas corresponds to the degree of memory impairment exhibited by AD patients. To illustrate, Price et al. found an average of 10.5 NPs and NFTs per mm2 in the perirhinal and entorhinal cortex of mildly memory impaired patients (n = 6; f age = 86 years). In comparison, an average of 16.55 NPs and NFTs per mm2 were found in the perirhinal and entorhinal cortex of moderately and severely memory impaired patients (n = 6; } age = 76.17 years). In contrast, very few NFTs and no NPs were found in the same anatomical areas of young (n = 5; f age = 59.2 years) and older (n = 8; f age = 79.13 years) non-demented elders (Price et al., 1991). Together Price's findings support the idea that there is a correlation between the number of NPs and NFTs that develop in the perirhinal and entorhinal cortices and the severity of the memory impairment experienced by AD patients. Price et al. also found around .15 NFTs per mm2 in the young non-demented subjects compared to 4.2 NFTs per mm2 in the older non-demented subjects. These results are similar to those found by other investigators who have shown that while neuropathological changes are not found in the hippocampus and parahippocampal regions of thirty year olds, they can be found in between 5 - 15% of the brains sampled from the fifth, and in around 50% of the brains sampled from the seventh decade of life (Squire, 1987; Tomlinson, Blessed, & Roth, 1968; Tomlinson, 1972), Together tbese find ings suggest that the subtle, but -  -  -  27 increasing memory impairment experienced by some "healthy" people with advancing age might be attributed, at least in part, to an increase in the number of NFTs that accumulate in these brain areas (Davis & Bernstein, 1992; Price et al., 1991; Squire, 1987). Long-term Memory Impairment. Other AD researchers have chosen to focus their attention on the neuropathological development that occurs within the neural pathways of the hippocampal formation (see Damasio et al., 1990). What this work has shown is that NFTs selectively accumulate in the cells of origin of the perforant pathway (i.e., the entorhinal cortex) while NPs accumulate in this pathway's area of termination (i.e., the hippocampal formation; Hyman, Van Hoesen, Kromer, & Damasio, 1986). These findings have led Damasio and his associates (1990; Van Hoesen et al., 1986) to suggest that while damage is occurring to the cortical areas surrounding the hippocampus (i.e., the entorhinal and perirhinal cortices) there is a concurrent process of neuropathological destruction that is occurring within the hippocampus. The destruction of the major neural pathways within the hippocampus serves to completely disconnect it from other brain regions. That is, information is essentially prevented from entering and leaving the hippocampus through the entorhinal and fornix routes.  28 As Damasio et al. (1990) have noted, the memory deficits that are observed in AD patients are more extensive than that seen in most amnesic conditions (p. 93). Specifically, the development of a long-term memory impairment that is (at some point in the development of AD) transposed onto a pre-existing problem when intentionally recollecting recently acquired information. The exact reasons for the gradual development of this long-term memory impairment is not known, but it seems likely that this is due to the far reaching areas that are directly, and indirectly influenced by the disconnection of the hippocampus from cortical areas (Damasio et al., 1990). Neuropathological Developments Outside the Hippocampus. Proponents of an anatomical pathogenesis of AD suggest that it is during the process of hippocampal disconnection that NPs and NFTs spread from the hippocampus to other brain regions via corticocortical pathways 2 (Damasio, Van Hoesen, & Hyman, 1990; Esiri, 1989; Jellinger, 1990; Lewis, Campbell, Terry, & Morrison, 1987; Rogers & Morrison, 1985). Cortico-cortical pathways originate from the cell bodies of pyramidal neurons and serve to indirectly connect the hippocampus to many cortical and subcortical areas (Van Hoesen & Damasio, 1987). The cortico-cortical progression  This is one of several theories regarding the development of AD (see Hardy & Davies, 1988; Toledano-Gasca, 1988). To date the anatomical pathogenesis model provides the most consistent explanation for the pattern of neuropathology observed in AD patients (Mann, 1991). 2  29 theory is supported by the fact that the brain areas that incur the largest amount of NFT development are those that have the heaviest connections to these pathways (Lewis et al., 1987; Van Hoesen & Damasio, 1987). For example, in later stages of AD the areas with the largest numbers of NPs and NFTs are the temporal and parietal lobes, basal forebrain and hippocampal regions (Brun, 1983; Lewis, et al., 1987; see Figure 1). There is also a discrete pattern of neuropathology found within cortical areas (Lewis et al., 1987; Van Hoesen & Damasio, 1987). That is, within a cortical area, it is the region that receives the most input from cortico-cortical pathways that have the largest numbers of NFTs. To illustrate, Lewis et al. (1987) compared the amount of neuropathology that occurred in the primary visual cortex (i.e., Brodmann's area 17), to the adjacent secondary visual association area (i.e., Brodmann's area 18) and to the tertiary or higher-order visual association areas (i.e., Brodmann's area 20; see Figure 1). These areas receive an incremental amount of cortico-cortical input as one goes from Brodmann's area 17 to 20, which is presumably related to the increasing complexity in the information these areas process (Kandel, 1985; Lewis et al., 1987). Lewis et al. found that when comparing the brains of 8 AD patients (stage of disease not specified) there was a corresponding incremental relationship in the number of NFTs that were found. For example, per 250  30 micrometer area there were, on average, around .9 NFTs found in the primary, 19.7 NFTs in the secondary, and 35.5 NFTs in the tertiary visual areas. Similar findings have been reported for other areas including the auditory cortex (Esiri, Pearson, & Powell, 1986; Pearson, Esiri, Hiorns, Wilcock, & Powell, 1985). Reaffirming this relationship between cortico-cortical connections and AD is the fact that the brain areas that are not as heavily connected to these pathways incur less neuropathology (Van Hoesen & Damasio, 1987). For example, the relative sparing of Brodmann's area 3, 1, and 2 for the somatic modality and Brodmann's area 4 and 6 of the motor cortices illustrates this (Brun & Englund, 1981; Brun, 1983; Lewis et al., 1987; Van Hoesen & Damasio, 1987). To synopsize, the neuropathology associated with AD appears to follow a consistent pattern that begins in the hippocampus before spreading along cortico-cortical pathways to other brain regions. This pattern of pathology serves to eventually disconnect both the hippocampus and the tertiary cortical association areas responsible for visual and auditory processing, from other brain areas (Damasio et al., 1990; Esiri et al., 1986; Lewis et al., 1987). In addition, other brain centers are also involved. For example, cortical areas including the temporal, parietal, and the basal aspects of the frontal lobe (i.e., nucleus basalis_of_14eynert4-as-well as^areas  31 including the locus coeruleus, basal ganglia, and the thalamus are also invaded by NFTs and NPs (Brun & Englund, 1981; Davies & Maloney, 1976; Damasio et al., 1990; Forno, 1978; Rudelli, Ambler, & Wisniewski, 1984). The brain areas outside the hippocampus that are affected by AD are thought to be responsible for the fact that the memory deficits of these patients are more extensive than those observed in most amnesic patients (Damasio et al., 1990). Some of these regions are believed to work together in the coordinated activity of general purpose cognitive functions that have not been considered to directly influence the memory impairment of AD patients. For example, the locus coeruleus, nucleus basalis of Meynert, hippocampus, frontal, and parietal lobes have been ascribed various roles in attention (Filoteo et al., 1992; Mountcastle, 1978). As subsequent sections will examine, attention and memory are inseparately linked in that, amoung other things, if the organism has not attended to (or focused attention on) critical targets, later recollection of the target will not occur (Graf, Tuokko, & Gallie, 1990). Thus, there is a possibility that some of the memory impairments observed in AD patients may be influenced by these patients' attentional deficits. In contrast, other areas of the brain appear to be relatively- -spared-by- PM and this is presumably because they -  -  -  32 receive less input from cortico-cortical pathways (Van Hoesen & Damasio, 1987). These areas include those responsible for somatomotor type processing and this has been correlated to AD patients' ability to learn some motor-based tasks (Damasio et al., 1990; Eslinger & Damasio, 1986). What implications does this pattern of neuropathology have for the memory researcher? While many researchers have used the functional or cognitive impairment level of AD patients as a rough index of the stage of neuropathological progression of this disease, in reality this provides only a gross measure. There is currently no practical method for determining (short of autopsy) the precise locations and amount of neuropathology present in AD patients while they are alive. Thus, unlike memory researchers using animal models, those investigating AD patients are left with the dilemma of what brain areas are correlated to the memory impairment. Comparisons of the memory performance of AD patients with patients with localized lesions has provided a wealth of information with respect to the intentional memory impairments that occur as a result of damage to hippocampal regions. However, AD patients are unique in that they provide the researcher with a model for investigating the implications that other brain regions also have on memory performance. For the pattern of neuropathology that occurs in AD however, one would need a " tout" that could separate hippocampal from non-  -  -  33 hippocampal functions as well as the influences that attention and modality specific processing might have on memory performance. The next chapter introduces such a tool which is contained under the framework known as explicit and implicit memory.  34 CHAPTER 3: EXPLICIT AND IMPLICIT MEMORY The focus of this chapter is on the brain regions and the cognitive processes that might be required to perform implicit (non-intentional) memory tests. Neuropsychological investigations that use a "multiple systems" model and cognitive studies using a "process" interpretation of the dissociations that occur in explicit and implicit test performance are reviewed. In the final section I show how ideas generated by research guided by these two models were used in the first study of this investigation.  Introduction Many ways have been used to categorize and investigate memory (Lovelace, 1990; Richardson-Klavelin & Bjork, 1988). For example, memory has been described in terms of the duration of time that has elapsed between the "learned" event and its subsequent recollection (i.e., short term vs. long term memory). In recent years an additional way of classifying memory has become prominent because of its usefulness in investigating both the neural and the cognitive processes of memory; this is the distinction made between explicit and implicit memory (Masson & Graf, in press). The terms explicit and implicit were first introduced by William McDougall (1924) and describe a memory dissociation that had been known to clinicians for many years (i.e., Freud, 1890;  35 Korsakoff, 1889). For instance, the neurologist Claparede (1911, 1951) observed that an amnesic patient would not shake his hand after he had pricked her with a pin, even though she did not know why she should refuse. These patients appeared to retain an implicit recollection of an event they could not explicitly recount. The terms explicit and implicit were later re-introduced by Graf and Schacter (1985) as "descriptive concepts that focus on the person's psychological experience at the time of retrieval" (Schacter, 1987, p. 501). Graf & Schacter (1985) defined implicit memory as being revealed when previous experiences facilitated performance on tasks that did not require a conscious or intentional recollection of those experiences. In contrast, explicit memory was revealed when performance on a task required the intentional recollection of a previous event. By definition implicit memory was indexed by implicit memory tests and explicit memory by explicit tests (Graf & Schacter, 1985). Priming, or the facilitation in performance as a result of recently encountered information (Shimamura, 1986, in press) was used as an index of implicit memory . The terms explicit and implicit (as introduced by Graf & Schacter) were intended to refer only to the cognitive processes engaged by the subject during the retrieval of information. However,iTivestigators interested in the biological substrates of  36 memory have found the explicit and implicit distinction to be useful in their work 3 (Squire, 1992a). As a result, two main ways for interpreting dissociations between explicit and implicit test performance have evolved (Schacter, 1987; Squire, 1987). One of these methods has been referred to as the "multiple systems" model (Mishkin, Malamut, & Bachevalier, 1984; Squire, 1987; Shimamura, in press) and the other as the "process" view (Roediger, 1990a, 1990b). Multiple Systems Researchers interested in the biological substrates of memory have mainly adopted the "multiple systems" model (e.g., Cohen & Squire, 1980; Schacter, 1989; Squire, 1992a; Tulving, 1986). The main assumption of this model is that different brain structures are required to perform explicit and implicit tests (Shimamura, in press). This idea receives support from amnesic patients like R.B. who can perform implicit but not explicit memory tests (Zola-Morgan, Squire, & Amaral, 1986). In this case explicit tests are viewed as tapping the integrity of areas such as the hippocampus and diencephalic regions that are typically destroyed in amnesia (Squire, 1992a). In contrast, implicit tests are viewed as tapping non-hippocampal regions such as the  3 As Squire, Knowlton and Musen (1993) have discussed, Graf and Schacter's (1985) definition of explicit and implicit depend on the - arse- of language to -influence the subject's -fiFtent at recollection. This is not useful in animal investigations where the focus must be on the tasks that are used.  37 cortical association areas (Squire, 1987; Squire, Knowlton, & Musen, 1993). Performance dissociations also occur on implicit tests (cf. Butters, Heindel, & Salmon 1990) and Squire (1992a) has suggested that this is because priming occurs in many different brain areas. Assumptions are made that the ability to find dissociations in implicit test performance is related to the geographical closeness (i.e., functional separateness) of the neural systems required to perform these tests (Shimamura, in press; Squire, 1987). Process Model In contrast, the process model is used primarily by cognitive psychologists who employ neurally-intact subjects for their investigations (Craik, 1983; Graf & Mandler, 1984; Roediger, Weldon, & Challis, 1989). The main assumption of the process model is that different cognitive processes, not neural regions, are required to perform explicit and implicit tests. Dissociations in test performance are therefore viewed as tapping the different cognitive processes required to perform these tests (Roediger, 1990b; Roediger, Weldon, & Challis, 1989). Unlike systems researchers whose focus is on isolating anatomical regions, the goal of the cognitive psychologist is to identify the cognitive activities that influence test performance tHoediger_, _I-990br-Shimamura- in prebb). They do tnis by  38 manipulating the encoding, retrieval, and test materials (i.e., written words, pictures) that are used and measure the effects this has on performance (Roediger, 1990b). Transfer Appropriate Processing Framework. Many cognitive psychologists use the theoretical framework called Transfer Appropriate Processing to guide their experimental interpretations (TAP; Graf & Gallie, 1992; Roediger, 1990b). The concept of TAP was first introduced by Morris, Bransford, and Franks (1977) who postulated that performance on various tasks was maximized when the cognitive operations engaged at study recapitulated those that were engaged at test. To illustrate, the most effective way to learn how to paraglide is to paraglide, not to read a book on the subject. Placed in other terms, performance on an implicit memory task will be maximized when the cognitive processes engaged at study are those that are engaged at test. As Graf and Ryan (1990) have argued, in this general form TAP is unable to explain differences between explicit and implicit test performance because it does not state which cognitive processes are involved. In recent years several proposals regarding the nature of these processes have been made including the widely held view of Roediger and his colleagues (Roediger, 1990b; Roediger & Blaxton, 1987). Roediger views -  cognitive processing requirements as varying along a continuum -  39 where at one end, to perform a memory test requires that the subject focus on the physical features of the test stimulus (i.e., engage in data driven processes) and at the other end, the focus is on the conceptual features of the stimulus (i.e., engage in conceptually driven processes). However, as Graf and his colleagues have shown, when the only difference between implicit and explicit tests are the test instructions provided to subjects (i.e., the test forms do not differ) performance dissociations are found that are not predicted by Roediger's classification (Graf & Mandler, 1984; Graf & Schacter, 1987). Thus, it appears that Roediger's data driven - conceptually driven distinction may not capture all the critical features that distinguish the cognitive processes required for performing implicit and explicit tests. Alternatively, Graf and his colleagues (Graf & Gallie, 1992; Graf & Mandler, 1984; Mandler, 1980) have focused their interpretations in terms of the requirement for integrative and elaborative-type processing. The advantages of this proposal is that it is able to address cognitive, neural, and test-related considerations. Integrative-type Processina. By definition, integration results from processing that bonds the features of a target into a pre-existing whole or unitized representation (Graf & Gallie, 19921. For example_, main v • -.^  are  40 required to complete the following; PSYCH OGY. Elaborative-type Processing. In contrast, elaborative-type processing results in the formation of a new relation between a target and its pre-existing mental contents (Graf & Gallie, 1992). This can occur when a target is encoded in relation to relevant prior knowledge. For example, when you learn that a new dopaminergic (DA) receptor has just been discovered (i.e., DA 100) and you place this information into your pre-existing knowledge of DA receptors. Explicit and Implicit Tests. The assumption that is made is that memory tests vary in the degree to which they engage elaborative or integrative processes (Graf & Gallie, 1992). Explicit tests such as free and cued recall are viewed as requiring mainly elaborative-type processing since they require the retrieval of targets that have been associated, or tagged, to specific prior knowledge (Roediger, Weldon, & Challis, 1989). In contrast, implicit tests such as word stem completion are viewed as requiring more integrative-type processing (Graf & Gallie, 1992). However, within each test type (i.e., explicit, implicit) there will be variations in the amount of integrative or elaborative processing that is required (Graf & Gallie, 1992). For example, free recall tests are viewed as requiring more elaborative type processing than cued recall tests (Schacter, 1987).  41 Which Model? There has been an ongoing debate regarding the "best" model to use in the interpretation of explicit and implicit memory test performance (see Roediger, 1990a, 1990b; Schacter, 1990b). The major issue with the process model is seen as its reliance on the TAP framework (Schacter, 1990a). TAP does not enable the investigator to test experimental hypotheses since all findings can be explained by this theoretical framework (Graf & Gallie, 1992; Schacter, 1992). In contrast, cognitive psychologists do not see the point of postulating different neural systems to account for findings that are accountable by one processing system (Roediger, 1990b; Roediger, Weldon, & Challis, 1989). In reality, each model is equally able to explain the main findings in the literature (Tulving & Schacter, 1990; Schacter, 1990b). For example, the fact that amnesic patients' performance on explicit tests is much lower than that of controls can be explained with a systems or process interpretation. A systems based explanation would be that the brain regions required to perform explicit tests (i.e., hippocampus and/or diencephalon) are intact in the control but not the amnesic patient. In contrast, a process account would postulate that the control, but not the amnesic patient, was able to engage in the elaborativetype processing required to perform this test. In other terms, the explicit test was too difficult for the amnesic  patient to ^  42 perform. Similarly, both a systems and a process interpretation can be used to guide interpretations regarding implicit test dissociations. A systems interpretation would be that different neural systems are required to perform different implicit tests and dissociations simply identify the integrity of these systems. For example, Huntington Disease (HD) patients can perform implicit tests of visually-presented words but not motor-based tasks (Butters, Heindel, & Salmon, 1990). A systems explanation for this finding would be that the neural areas required to process visually-presented words are intact in HD patients. In contrast, those areas required to perform the motor-based task are somehow impaired. This latter interpretation would appear to correlate closely with the loss of neurons that occur in the basal ganglia of HD patients (although some cortical atrophy may also occur; Kandel & Schwartz, 1985). In contrast, a process interpretation would explain these performance differences in terms of the encoding, retrieval, and/or stimulus characteristics of these tests. Impaired test performance with visually presented word stimuli would be viewed in terms of how this test differed from the motor-based task. For example, is there something inherently different in the way a visually versus motor-based task is encoded? retrieved? etc.  43 As Tulving and Schacter (Schacter, 1992; Tulving & Schacter, 1990) have maintained, there is no incompatibility between systems or process approaches, they just serve to ask different questions about memory. Squire has made similar comments by advocating that the focus should not be a philosophical debate of how to best classify memory, but rather, on finding the best way to approach empirical questions regarding brain function (Squire, 1987, p. 160). As the next sections show, research using systems and process interpretations can provide information on how implicit memory is affected in AD patients. Neuropsychological investigations reveal the implicit tests (and therefore the methods) in which AD patients perform poorly in comparison to controls, and some of the neuroanatomical areas that might be responsible. Cognitive investigations (with older adults) suggest that impaired performance on some implicit tests may be due to the cognitive processes that are involved. The focus of these sections is on test results since the theory that accompanies these findings is not well established. This state of affairs is reflected in a recent quote by Roediger on implicit memory (1990b, p. 1054) The new mass of knowledge is still formless, incomplete, lacking the essential threads of connection, displaying misleading signals at every turn, riddled with blind alleys. There are fascinating ideas all over the place, irresistible experiments beyond numbering, all sorts of new ways into theme  44 The following sections attempt to bring some cohesion to two different bodies of research that (1) are both in the early stages of development, and (2) use memory tests to ask different questions about memory.  Neuropsychological Investigations Dissociations in Explicit and Implicit Test Performance. Warrington and Weiskrantz (1968, 1970) were one of the first groups to discover that different brain regions might be required to perform explicit and implicit memory tests. In one of a series of experiments, four amnesic patients (three patients with Korsakoff's syndrome causing damage to diencephalic regions and one patient with a temporal lobectomy) were provided with lists of words to read. At test these patients were provided with words in which pieces of the printed letters were missing. When asked to use these fragments as clues to remembering previously studied targets (i.e., engage in intentional recollection) amnesic patients were found to remember significantly fewer items in comparison to control subjects (Warrington & Weiskrantz, 1968, 1970). In contrast, when the same patients were allowed to treat the task as a guessing game (i.e., engage in non-intentional recollection) they showed performance that was similar to controls (Warrington & Weiskrantz, 1968, 1970). Subsequent research has shown that patients with amnesia  caused by etiologies other  than Korsakoff's and temporal  45 lobectomies (i.e., chronic electroconvulsive therapy, encephalitis, ischemic attacks) show a consistent ability to perform implicit, but not explicit tests (Graf, Squire, & Mandler, 1984; Squire, Shimamura, & Graf, 1985; Warrington & Weiskrantz, 1978). Research has also shown that it doesn't matter what modality information is presented in (i.e., auditory, visual, somatomotor) amnesic patients can show normal performance on the implicit, but not the explicit version of the same test (Graf, Shimamura, & Squire, 1985; Moscovitch, 1982; Johnson, Kim, & Risse, 1985). It is from these consistent findings that researchers adopting the systems model have concluded that the medial temporal lobe and/or diencephalic areas -- areas typically damaged in amnesic patients -- are required to perform explicit, but not implicit memory tests (Keane, Gabrieli, Fennema, Growdon, & Corkin, 1991; Salmon & Heindel, 1992; Shimamura, in press). In contrast, a systems interpretation would suggest that areas outside the temporal and thalamic regions are involved in implicit test performance' (Butters, Heindel, & Salmon, 1990; Schacter, 1992). This idea has been supported by the results of investigations with patients that have brain damage to areas outside, or in addition to hippocampal and diencephalic regions (Butters, Heindel, & Salmon, 1990). Much of this information has  These_ideas have recently been supported by Positron Panissinn Tomography (e.g., Squire, Ojeman, Miezin, Petersen, Videen, & Raichle, 1992b).  46 been developed from the work of a collaborative team of researchers who have compared the test performance of patients in the early stages of Huntington's disease (HD) and AD (see Butters, Heindel, & Salmon, 1990). In these studies HD patients are used as brain models where damage occurs primarily to the basal ganglia 5 (i.e., subcortical damage) and AD patients where damage is to the cortical association areas (i.e., cortical damage). Dissociations in the implicit test performance of these two patient groups are used to deduce the neural substrates that might be required to perform these tests.  Dissoc ations in Implicit Test Performance. One of the i  first groups to compare the implicit test performance of AD and HD patients was Shimamura, Salmon, Squire, & Butters (1987). In this study Shimamura et al. gave eight patients in mild to moderate AD, eight HD, and seven Korsakoff (KS) patients and their age-matched controls a word stem completion test. The same stimuli and procedures used in a previous study where amnesic patients' performance was found to be similar to that of controls was used (Graf, Squire, & Handler, 1984). At study Shimamura et al. asked subjects to read 10 words (e.g., motel, abstain) and rate how much they liked each word  5 In reality, the neuronal changes that occur in the brains of Huntington and Alzheimer patients are not restricted to these areas. Cortical atrophy can occur in Huntington patients and subcortical damage occurs in Alzheimer patients (see Brun, 1983; Kandel & Schwartz, 1985; Price et al., 1991).  .  47  using a five point scale (i.e., 1 = dislike extremely to 5 = like extremely). Subjects were then shown 20, three-letter word stems (e.g., MOT^, ABS^) and requested that they complete each stem with the first word that came to mind. For each stem there were at least 10 endings that would complete the word (i.e., mother, motive, motel, motor, etc.). Ten of the stems could be completed with words presented at study. The remaining ten stems were new and were used to provide an index of what subject performance might be like without previous exposure to words (i.e., baseline performance). Subjects were also given the Rey Auditory Verbal Learning Test to obtain an index of explicit memory performance (Rey, 1964). Shimamura et al. found that the AD, HD, and KS patients all showed an impaired level of performance on the Rey Auditory Verbal Learning Test. On the word stem completion test the HD and KS patients completed between 30 to 40% more of the stems from the studied (i.e., target) versus the non-studied condition (i.e., baseline) and this performance was similar to that of agematched controls. In contrast, AD patients completed approximately 10% more of the studied than the non-studied items and this performance was significantly lower than this group's age-matched controls. However, baseline guessing rates were found to be similar for the AD (i.e., baseline = 6%) and the HD and KS patient groups (i.e.,  baseline = 5-11%1.  48 Shimamura et al. noted that the AD patients had been able to complete 98 percent of the stems with words, just not with the words that had been previously studied. This finding suggested that the AD patients' poor test performance was not due to an inability to perform the basic task of completing word stems. Other investigators have reported similar findings using different AD patients and the word stem completion test (Bondi & Kaszniak, 1991; Heindel, Salmon, Shults, Walicke, & Butters, 1989; Randolf, 1991; Salmon, Shimamura, Butters, & Smith, 1988). When considering the brain regions that might be required for performing the word stem completion task, Shimamura et al. reasoned that since the AD, but not the KS patients had showed impaired levels of priming, this likely reflected damage to regions other than the medial temporal lobe or diencephalic structures (p. 350). In turn, since HD patients had showed priming levels that were similar to their controls it appeared that the integrity of the basal ganglia was not critical for normal performance on the word stem completion test. This meant that the impaired performance of AD patients on the word stem test was likely a result of damage to cortical areas (p. 350). When considering what AD patients' impaired word stem completion performance meant in terms of memory stores, Shimamura et al. employed ideas found in the spreading activation theory of semantic processing (cf. Collins_kLoftus, 1975; Nebes, 1992). ^  49 As conceptualized by Tulving (1986) semantic memory is the body of knowledge that people possess about words, concepts, their meanings, associations, and the rules that govern the manipulation of these words and concepts. Information in semantic memory is viewed as being arranged as an organized network of concept nodes that are interconnected to each other on the basis of their semantic relationships. The ability to locate items in this network has been viewed as being impeded by either (1) a breakdown in the structure of semantic knowledge (i.e., a loss rather than an inability to access information; cf. Salmon & Heindel, 1992), (2) an inability to access items as a result of memory-related processes (cf. Nebes, 1992), or more recently, (3) an inability to access items due to non-memory processes (i.e., attentional and/or visuospatial impairments; cf. Nebes, 1992). A main assumption of investigators is that the presentation of either a word or of a picture target automatically activates its mental representation in the semantic network (Martin, 1992). Another assumption is that this activation automatically spreads to related concept nodes, and by doing so serves to increase target retrievability (Martin, 1992). Presumably this activation establishes a new and highly specific mental representation of the target. This new representation could account for the specificity and temporal persistence of priming events (Schacter, 1990b). The spreading activation to related concept nodes would  50 presumably correspond to priming that can occur for items that are semantically related to the original target (i.e., target = orange and the subject responds with lemon). Shimamura reasoned that the AD patients poor performance on the word stem completion test was more likely due to a memory rather than to a non-memory impairment (i.e., a global intellect or cognitive deficit). This was because AD patients had been able to perform the basic task of completing word stems, had showed baseline performances that had been similar to the other subject groups, and that their level of dementia had not appeared to be related to their priming performance. Shimamura et al. felt that the AD patients' memory impairment was due to an inability to activate nodal representations and that this might be due to a breakdown in the structure of lexical knowledge caused by damage to cortical brain areas. Salmon, Shimamura, Butters, and Smith (1988) continued their study of implicit memory in AD patients by giving many of the same patients from the Shimamura study a word pair task. This test was thought to provide an index of memory for semantic relationships since it required that subjects focus on the relationship between word pairs. Nine AD and 10 HD patients, along with similar numbers of age-matched controls, were asked to rate a set of 12 word pairs (i.e., bird-robin, needle-thread) twice on the basis of how  51 related each word was to each other (i.e., 1 = not related to 5 = very closely related). Testing was started immediately upon completion of the study phase. At test subjects were told that single words would be presented visually, and that they were to respond verbally with the first word that came to mind in response to each stimulus word (i.e., bird  , needle^ ).  As in the Shimamura et al. study, subjects were shown words belonging to word pairs they had studied (the target condition) along with words they had not studied (the baseline condition). Results showed that only the AD patients showed impaired priming on this free association task when compared to their control group. Specifically, Salmon et al. found that AD patients completed only about 2% more of the word associations from the target than from the baseline condition (baseline = 18%). In contrast, AD patients' age-matched controls completed 30% more of the target than the baseline word associations (baseline = 18%). These findings were much different than for those found for the HD patients. HD patients correctly completed around 25% more of the target than the baseline associates (baseline = 17%). HD patients' performance was similar to their age-matched controls who showed a 20% level of priming above baseline (i.e., baseline = 21%). Since the baseline performance of HD, AD, and their age-matched controls was not significantly different, the magnitude of these priming scores was based on  52 similar performance levels. This provided stronger evidence that priming was impaired in AD patients. Salmon et al. concluded that the AD patients' impairment on the word association test must reflect damage to brain regions that were not disrupted in HD patients who were able to perform this test (p. 492). Furthermore, in view of the language disruptions and the neuropathological changes that occur in the temporo-parietal cortices of AD patients this was the area that was felt to be responsible for their impaired test performance (p. 492). Salmon et al. also concluded that the cue "bird" had been unable to evoke the paired associate "robin" because the association between the two words has been greatly "weakened" (p. 490). Salmon et al. based this idea on the fact that AD patients had been able to provide appropriate responses for non-studied associates and this suggested that the semantic memory network had not been totally destroyed. Salmon et al. went no further in their interpretations of why this system was weakened but the underlying assumption of neuropsychologists would be that neuronal loss caused by AD would somehow be responsible (i.e., a disruption in the memory network rather than an inability to access it; see also Salmon & Heindel, 1992). Heindel, Salmon, and Butters (1990) continued this series of investigations by administering a picture fragment completion  53 test to 12 AD and 12 HD patients and their age-matched controls. Their hypothesis was that if lexical, semantic, and pictorial priming tasks share a common neurological substrate, then AD patients, but not HD patients, would show impaired performance on the picture test (p. 284). At study subjects were asked to name items depicted in a set of 15 simple line drawings of common objects. At test the experimenter placed a binder in front of each subject with the request that they state the first thing they thought of when they saw the fragmented stimuli. Heindel et al. results showed that HD patients had similar levels of priming when compared to their controls. In contrast, AD patients had significantly lower levels of priming which is a finding that was later corroborated by Bondi and Kaszniak (1991). Since AD patients had correctly identified 99.3% of the study pictures Heindel et al. reached a similar conclusion to that offered by Shimamura et al., and Salmon et al. Specifically, that AD patients' performance on the picture fragment completion task was because they were impaired in their ability to activate pre-existing representations of pictures. Additional evidence for dissociations in implicit test performance was provided by Heindel, Salmon, Shults, Walicke, and Butters (1989). In this study they gave a rotary pursuit and a word stem completion test to 16 AD, 13 HD, 8 demented and 9 non-demented Parkinson patients  54 (PD) and a group of 10 middle-aged and 12 elderly controls (agematched to patient groups). The rotary pursuit task consisted of having subjects maintain a stylus on a rotating target for 6 blocks of 4 trials lasting 20 seconds. To standardize the different performance abilities of each subject group the rotation speed of the target was varied. This speed was determined from practice trials and set at a level where that group, on average, had maintained the stylus on target for at least 25% of the trial duration. The word stem completion test was the same as that used in the Shimamura et al. investigation. Heindel et al. found that whereas AD patients showed impaired performance on the word stem completion task, this was not the case for the pursuit rotor task. The average time that AD patients maintained the stylus on the rotating target steadily increased from the first to the sixth block of trials. These results implied that AD patients had "learned" or retained some benefits from previous exposure to the task. In contrast, Heindel et al. found opposite results for the HD group. Whereas HD patients showed similar performance to their age-matched controls on the word stem test, they showed little improvement in their ability to maintain the stylus on the rotating target over the six blocks of trials. In combination with the findings from the AD patients, these results were used to suggest that motor skill learning might be mediated by the  55 corticostriatal system (which is damaged in HD patients) whereas lexical priming may depend on the integrity of neocortical association areas (which are destroyed in AD patients). The additional finding that the demented PD patients showed impaired performance on both the word stem and rotary pursuit tests seemed to confirm this conclusion since the brain pathology of these patients included areas that are affected in both AD and HD patients (Agid, Ruberg, Dubois, & Pillon, 1987; Hakin & Matheson, 1979; Gaspar & Gray, 1984). Similar findings to that of Heindel et al. (1989) have now been reported by investigators who have either employed AD patients (Eslinger & Damasio, 1986), AD and non-demented PD patients (Bondi & Kaszniak, 1991) or AD and HD patients (Heindel, Butters, & Salmon, 1988). To summarize, studies adopting a systems model approach have provided information on some of the memory tests that AD patients perform poorly in comparison to controls. Because these studies focus on mapping test performance to functionally impaired brain areas (i.e., lost neuronal function) they have not been designed to ask questions other than those that would address a structurally damaged network model (i.e., a memory deficit model). That is, neuropsychological studies have not generally focussed on whether access to the memory network is impaired by non-memory factors or whether access can be achieved through other routes. As such, these studies have been unable to provide  56 direct explanations for the reason AD patients can perform the task but not the memory demands of tests. As well, these studies do not address why the memory performance of some AD patients fluctuate during the day. For example, some patients may show greater memory impairments in a morning than during an afternoon test session (personal observations). Additionally, with some investigators now deviating from traditional neuropsychological methods and using different tests and changing the methods that are used different findings are beginning to appear. One study that has shown that the word stem completion performance of AD patients may not be impaired, relative to controls, was conducted by Partridge, Knight, and Feehan (1990). In this study 15 AD patients and 15 age matched controls performed a word stem task similar to that used in the Shimamura et al. (1987), Salmon et al. (1988), and Heindel et al. (1989) investigations. The basic difference in the Partridge et al., study was that, instead of encoding targets by providing a pleasantness rating, subjects responded with the targets' meaning. Partridge et al. found no statistical difference between their AD patients' and their controls' performance on the word stem completion task. Since both groups' baseline performance was not statistically different this confirmed that the priming levels of each group were based on similar performance levels.  57 Unlike previous studies, Partridge found that their AD patients had slightly higher priming levels than that of their controls (i.e., 31.4% vs. 26%). Since this patient group may have been more demented than those in the Shimamura et al., the Salmon et al., and the Heindel et al., investigations (p. 117) the reason for this finding remains unknown. One possibility is that since the controls used in the Partridge study were obtained from nursing homes, they may have been a low functioning control group in comparison to those employed in the other studies. The absence of information on the functional abilities of Partridge's controls as well as precise information on the impairment levels of the AD patients that are used in these type of studies precludes a decision about whether Partridge's findings might be due to the subjects or to the encoding method that she used. A second study to show that priming for word stimuli may not be impaired in AD patients employed a word identification task (see Keane, Gabrieli, Fennema, Growdon, & Corkin, 1991). In this investigation Keane et al. had 12 AD patients and 12 age-matched controls view a series of 32 words, presented one at a time. At study each word appeared on the computer and subjects were asked to read it to the experimenter. At test subjects were told that they would now perform a different task in which words would be flashed on the computer screen and subjects were to simply identify the word.  58 Keane et al. found that AD patients and controls identified studied words faster than non-studied words and that the magnitude of the priming effects did not differ between subject groups. Thus, the Keane et al. results showed that AD patients can show a similar magnitude of priming to controls. Adopting a systems-based explanation Keane et al. decided that their results provided evidence for the existence of an implicit memory system mediated by the occipital lobe (p. 340). In contrast, it is less clear how the postulation of an additional neural system could account for the Partridge et al. results given that they simply changed the way in which AD patients studied target information. The next section shows how recent information from the field of cognitive psychology provide alternative, but complementary explanations for these findings.  Cognitive Investigations DissociAtions in Explicit and Implicit Test Performance.  A  second body of research that focuses on explicit and implicit test dissociations is found in the gerontological literature. The typical study compares the test performance of neurologically intact adults around the age of 70, to younger adults in their early 20's. The age-correlated incidence of AD (Treves, 1991) and the similar pattern of brain changes that occur in the elderly as compared to AD patients (albeit not to the same degree; Corkin, 1982; Squire, 1987) makes this work pertinent to  59 this investigation. Irrespective of the type of explicit test that is used (i.e., free and cued recall, recognition) the memory performance of older adults is found to consistently decline with advancing age (see Craik, 1978; Light, 1991; Poon, 1985, for reviews). This explicit memory pattern is similar to that discussed for AD patients and has also been related to the brain changes that occur around hippocampal and surrounding cortical regions (i.e., Davis & Bernstein, 1992; Price et al., 1991). However, unlike that found for AD patients the age-related performance impairments on explicit tests disappear when implicit versions of the same test are used. To date this finding has been established in a large number of studies using different types of tests and test materials. For example, older and younger adults' performance has been found to be similar in studies that have used the word identification (Light & Singh, 1987), word stem completion (Java & Gardiner, 1991; Light & Singh, 1987), category completion (Light & Albertson, 1989) and the picture fragment completion test (Mitchell, in press). Similarly, the performance of old and young adults has been found to be similar when word and picture stimuli have been used (Java & Gardiner, 1991; Mitchell, in press). There does however appear to be one implicit test where the priming levels of older adults are not always similar to younger- -  60 groups, and this is the paired associates test (Howard, 1988; Light & Burke, 1988; Light & LaVoie, in press). The paired associates test taps memory for newly experienced, unrelated word associations (i.e., queen-stairs, author-project; Howard, 1988). A cognitive interpretation for the reason performance on this test differs from the category completion or word identification test is that, while all are implicit tests, they differ in the amount of elaborative-type processing they require. That is, the paired associates test requires more elaborative (i.e., attentionally demanding) processing than the category completion test and older adults can meet the processing demands of the latter but not former test. This does however leave the question of why older adults are able to perform some, but not all forms of the paired associates test. According to Light and LaVoie (in press) the few studies that have used the paired associates test have used different encoding methods, materials (i.e., words vs. non-words) and nonidentical test forms (i.e., implicit and explicit). Since the cognitive processes required to perform this test has differed on several conditions that are important to the cognitive psychologist (i.e., encoding, retrieval, material type) this makes it impossible to pinpoint the source of these performance differences. A recent investigation by Howard, Fry, and Brune (1991) using similar tests and materials, but varying the  61 encoding conditions, indicates that one reason for these performance variations may be the extent to which targets are encoded for their meaning.  Dissociations in Implicit Test Performance. In the Howard et al. study 20 young (approximately 20 years old) and 20 older participants (approximately 69 years of age) studied pairs of unrelated nouns (i.e., dog-apple). The same test form was used to index explicit and implicit memory so that the main difference in cognitive processing requirements was whether the subject intentionally or non-intentionally recollected targets. For example, subjects were provided with one word and a word stem (i.e., queen-sta^) and then asked to either complete the stem with the first word that came to mind (i.e., the implicit or word stem completion test) or to fill in the blank with a word they had previously studied (i.e., the explicit or wordstem cued recall test). In experiment one subjects were asked to listen to simple sentences containing unrelated word pairs and then make a quick addition to this sentence (i.e., add a short phrase). For example, for the sentence containing the unrelated nouns house and  zky,  an addition might be ... are common Qbjects. Experiment  two was different in that word pairs were presented to the subject who used them to compose their own sentences. Subjects subsequently rated how difficult that sentence had been for them  62 to develop (i.e., 1 = easy to 5 = difficult). For example, house : sky; My friend painted her house in a colour that reminds me of the gky after a thunderstorm. Howard et al. postulated that this latter task forced their subjects to attend more to the meaning of the critical targets than the encoding task used in experiment one. Results showed that in experiment one the older subjects recollected significantly fewer word stem completion and cued recall stems than did the younger subjects. In contrast, in experiment two the older subjects' performance on the word stem completion test was not statistically different from that of the younger subjects. Nevertheless, performance on the word stem cued recall test remained significantly lower than that of the young subjects. Together these findings suggested that the encoding task used in experiment two may have been responsible for the elevation in the implicit test performance of the older subjects. The general gerontological literature supports the assumption that the encoding tasks may have differently influenced the implicit test performance of older adults. There is a large body of literature that shows that aging is accompanied by attentional deficits which may be related to the fact that older adults do not spontaneously engage in elaborative-type encoding processes (Craik & Rabinowitz, 1984;  63 Kausler & Litchty, 1988; Plude & Hoyer, 1985; Rankin & Collins, 1986). Since memory and attention are inseparately linked the attentional problems that accompany aging could be influencing memory test performance at either encoding and/or retrieval (Graf, Tuokko, & Gallie, 1990). To explain further, to remember a particular event requires that subjects be able to focus and sustain attention during encoding processes. Specifically, the subject must attend to those features of the target that distinguish it from non-essential characteristics while concurrently monitoring (i.e., attending to) other processing activities. Similar attentional demands characterize memory test performance (i.e., retrieval). That is, the subject must focus and maintain attention on the goal of the task, select an appropriate retrieval strategy and monitor its effectiveness, while ignoring irrelevant information in the test environment (Graf, Tuokko, & Gallie, 1990, p. 528).  What connection does the relatio ship between attention and wemory have on ttle _results obtained by Howard et al.?  A  processing model interpretation is that the encoding method that Howard et al. used in their second experiment was able to guide the older subject in attending to targets at encoding (i.e., it assisted them in completing the encoding processing requirements). However, it wasn't until this encoding method was  64 combined with a retrieval condition (i.e., non-intentional recollection) that was also less attentionally demanding (i.e., required more integrative processes) that the overall processing demands of the memory task reached a level that the older subjects could perform at levels that were similar to younger adults.  How are these findings relevant to AD patients? First consider the question of attention. To date there are three main lines of evidence to support the claim that AD patients experience attentional deficits. The first comes from clinical observations that some AD patients will not spontaneously encode target information. However, when the experimenter guides them through this processing by asking them questions about the target they often can complete the task. In addition, some AD patients exhibit behavioral fluctuations that have been said to reflect transient fluctuations in attentional focusing (cf. Graf, Tuokko, & Gallie, 1990). Second, as discussed in Chapter 2, the neuroanatomical damage associated with AD infiltrates areas that have been ascribed various roles in attention (cf. Filoteo et al., 1992; Mountcastle, 1978). Third, there have been a few investigators that have directly looked at the attentional abilities of AD patients using different types of tasks. For example, in one study I used a computer to assess AD patients' ability to perform a vigilance, selective, and divided attention  65 task (Gallie, Graf, & Tuokko, 1991). The findings from this study indicated that as the attentional demands of the task were increased, AD patients experienced greater difficulty in completing these tasks in comparison to controls. In a second investigation I followed the memory and attention performance of more than 100 AD patients during an 18 month period (Gallie, Tuokko, & Graf, 1991). Results from this study indicated that AD patients experienced attentional deficits that were more severe than that of controls and that there was a relationship between explicit memory performance, attention, and the patient's level of F.I. That is, as AD patients changed from mild to severe levels of F.I., their performance on memory and attention tasks progressively declined. Other investigators have chosen to look at specific aspects of attention and a recent investigation by Filoteo et al. (1992) has found that AD patients encounter problems when shifting their attention across target stimuli.  How might these attentional deficits relate to previous findings of impaired implicit memory performance in AD patients? Most investigators have simply required that AD patients encode targets by rating how pleasant that target is to them. For example, patients are asked to respond by saying five if they agree that the target is pleasant or one if they think the target is extremely unpleasant. While studies have found that the pleasantness encoding task results in normal priming levels for  66 KS and HD patients, work by Strauss, Weingartner, and Thompson (1985) and later, similar findings by Rohling, Ellis, and Scogin (1991) show that only KS and HD patients retain the ability to spontaneously and effortfully encode targets. Thus, the possibility exists that one reason why studies have found that AD patients show impaired priming on some tests is that they have not used an appropriate encoding method with these patients. To synopsize, the information contained in this chapter provides early evidence that the explicit/implicit distinction may be the tool, or Rosetta stone, that enables neuroscientists' to unravel further relationships between memory and the brain. To date this framework has shown that the priming abilities of AD patients (unlike some other neurologically impaired groups) are impaired on implicit tests of word and picture, but not motorbased materials. Damage to temporal and parietal cortical areas with relative sparing of motor and somatosensory areas may be the reason for these findings. In turn, we reviewed evidence that the cognitive processes engaged by older subjects at encoding, retrieval, and in performing different implicit tests can change their memory performance. These findings show that memory performance changes when different cognitive processes are activated on the same biological substrate. This idea is the premise on which study one was based. Study One  67 The goal of study one was to re-examine the implicit test performance of AD patients. This study extends existing research in that it combines both a process and a systems approach. This was carried out by administering eight explicit and implicit memory tests that were developed on the following principles.  First, given that AD patients suffer from attentional deficits an encoding method was used with all memory tests that ensured that patients had attended to, and processed, critical targets. This method simply required that subjects identify and then tell the experimenter what the target meant to them. Existing research shows that AD patients can effectively generate the meaning of targets until the later stages of their disease (albeit with diminishing comprehension; Nelson & McKenna, 1978).  Second, in order to tap hippocampal and non-hippocampal functioning explicit and implicit memory tests were used. To ensure that the primary differences in the processes these tests tapped were those that were engaged at retrieval, parallel forms of each test were developed.  Third, because existing research suggests that priming may be dependent on modality-specific processing the test materials used in this study were chosen for their ability to tap different sensory modalities. For example, written word and pictures were chosen for their ability to tap a primarily visual type of processing. Spoken words and tactilely presented objects were  68 used to engage primarily auditory and somatomotor processing. To ensure that maximum priming levels were achieved by subjects, all materials were presented in the same modality at study and test. Thus, unlike previous studies that have mixed the modalities of their study and test stimuli (i.e., study visually presented stimuli and test auditorally) this study did not.  The overall hypothesis of study one was that differences in the memory performance of AD patients and controls would be smaller on implicit than on explicit tests using written and spoken word, picture, and object materials. This hypothesis was based on two main findings regarding the neuroanatomy of AD and one regarding the effects of encoding processes on implicit test performance. First, previous research suggests that the neuropathology of AD begins in the hippocampus before spreading to other brain regions that may be integral for performing implicit memory tests. Thus, the explicit test performance of patients should always be lower than that of controls. second, the hippocampus processes multimodal information and so explicit test performance will be lower in AD patients compared to controls for all the test materials used (i.e., even for object materials although the neuropathology of motor and somatomotor areas is generally preserved in AD patients). Third, the implicit test performance of AD patients should be similar to that of controls on all tests for the following reasons -- first,  69 an encoding method was used that ensured that targets had been encoded, and second, damage to those cortical areas presumed responsible for priming appears to occur mainly in advanced stages of this disease (cf. Price et al., 1991).  70  CHAPTER 4: MEMORY STRATEGIES This chapter introduces the concept of encoding and retrieval strategies (i.e., test types). The limited studies that have used memory strategies with AD patients are then reviewed, together with the theories addressing how these strategies might work. Study two and its overall hypothesis is then introduced.  Introduction Memory strategies can be defined as any activity that is used to increase the amount of information that can be explicitly remembered (Bellezza, 1988). To date, systematic studies on how these strategies should be classified have not been conducted (West & Tomer, 1989) and for the purposes of this chapter they will be divided into those that are used during the acquisition (i.e., encoding) and those used during the retrieval of information.  Retrieval Strategies Retrieval strategies basically act to provide information about the to-be-remembered target and they have both a practical and a formal use (Bellezza, 1988). The practical use of retrieval strategies includes leaving notes (to ourselves) on the fridge or on calenders. In this application the effectiveness of these retrieval strategies depends on the extent to which we have  71 provided enough clues to reaccess the original memory event. The formal use of retrieval strategies occurs with the selection of different test types and this application is more familiar to the educator or clinician. For example, depending on the patient (or the class) the wise clinician (and educator) might choose a recognition (i.e., multiple choice test) rather than a free recall test (i.e., or essay) to ensure that the patient's (or class's) performance is not at floor. As in the previous example, the principle at work is that more retrieval clues are provided by a recognition (or cued recall) than by a free recall test and this may assist some patients (and students) to recollect information that may not have been accessible without this assistance. The underlying assumption is that a memory for an event has occurred, but there is a difficulty in reaccessing it. To compare this situation with the semantic network model of the previous chapter, there is either a problem in the memory or non-memory (i.e., attention) related processes that are required to relocate the mental representation of the "learned" event. Encoding Strategies In contrast, encoding strategies influence the cognitive processes that subjects engage to study targets (Bellezza, 1988). Many types of encoding strategies exist, including those where the subject simply repeats information (i.e., repetition) to where more effort is employed (Bellezza, 1988). In this latter  72 case generating the meaning of a target or composing a mental image of the event (Bellezza, 1988). Research on memory strategies has been primarily focused on the use of different encoding strategies with young children (Matlin, 1983). However, there is now an increasing interest in the use of memory strategies with healthy older adults (Hill, Storant, & Simeone, 1990; Norris & West, 1991). Together this work has shown that strategies are most effective when they are matched to the cognitive abilities of the subject group they are to be used with (Duke, Weathers, Caldwell, & Novack, 1990). To illustrate, rehearsal strategies are more effective in elevating the memory performance of young children, but not that of older adults (Matlin, 1983). Since older adults have acquired a more expansive knowledge of the world, a strategy requiring them to think more about the meaning of targets in relationship to what they already know is the more effective strategy for them (Matlin, 1983; Poon, 1985). In contrast, there have been very few investigations that have used memory strategies with AD patients (Camp & McKitrick, 1992). Several reasons can be cited for this, including the evidence that AD patients' memory abilities are progressively and irrevocably lost (cf. McKhann et al., 1984). While it has been recognized that AD patients' performance on recognition and cued recall tests is often higher than that found using free recall  73 tests, there have been few systematic studies on this effect (cf. Bellchambers, 1990). The next section reviews the equally limited work on the use of encoding strategies with AD patients. This section is selective in that it reviews those strategies that engage the abilities that are best retained in AD patients; that is, their ability to perform tasks and generate the meaning of targets (cf. Eslinger & Damasio, 1986; Nelson & McKenna, 1978). Due to the limited work with AD patients relevant studies using healthy older adults are selectively reviewed. Subject Performed Tasks (SPTs) One encoding strategy that emphasizes AD patients' retained motor abilities is referred to as subject performed tasks (Karlsson et al., 1989; Nyberg, Nilsson, & Backman, 1991). Subject performed tasks (or SPTs) are comprised of simple actionfocused sentences like "Put the pen in the case" that the subject performs, and then is later asked to recollect (see Cohen, 1981). Investigators can judge the effectiveness of SPTs by comparing subjects' performance on similar tasks that do not include motor enactment. For example, one condition is to simply ask subjects to listen to, or to read the command sentences without performing the task (i.e., VTs or verbal tasks). Another requires that subjects simply watch the experimenter perform the minitasks (referred to as EPTs or experimenter-performed tasks).  74 Investigators have found that both young and old subjects remember more SPTs than VTs (Engelkamp & Cohen, 1991). Thus, it appears that performing a task provides superior recollection of the event compared to simply reading or listening to it, irrespective of the subject's age (Backman & Nilsson, 1984). SPTs also appear to be a particularly effective strategy for older adults since most studies have found that old and young adults remember equivalent numbers of these tasks (Backman & Nilsson, 1985; Cohen & Stewart, 1982). SPTs also benefit cognitively impaired young adults who often recollect similar numbers when compared to non-impaired controls (Cohen & Bean, 1983). In combination, this research suggests that SPTs may be effective in elevating the memory performance of cognitivelyimpaired older adults, such as AD patients. To date only two groups of investigators have used SPTs with AD patients and they report very different results. The first group was Karlsson et al. (1989) who found that AD patients remembered significantly more SPTs than VTs. However, AD patients did not remember as many SPTs as did their age-matched controls. When Karlsson et al. (1989) gave these AD patients additional clues (i.e., retrieval clues) regarding the SPTs they had performed, this more than doubled the number of items they were able to remember. However, the combined use of an encoding  75 and retrieval strategy still did not raise AD patients' performance to levels that were comparable to those of their agematched controls. In this study Karlsson et al. (1989) also investigated if AD patients in different levels of impairment might be experiencing different memory benefits from the SPTs. Their results showed that this was the case. Mildly impaired AD patients (n = 5) remembered approximately 36 percent of the SPTs they had performed with cueing. In contrast, the moderately impaired AD patients (n = 8) remembered approximately 28 percent of the SPTs with cueing, and severely impaired patients (n = 7) remembered approximately 20 percent. The only other study to investigate SPT versus VT performance in AD patients was Dick, Kean, and Sands (1989). In their investigation AD patients (n = 18) performed a total of 36 minitasks (in comparison to Karisson et al. who used 25). Dick et al. results showed no significant difference in the number of SPTs and VTs that were freely recalled by their AD patients. The discrepancy in the results reported by Dick et al. (1989) and Karisson et al. (1989) may be attributed to the different methods that they used. Recent investigations using healthy older adults have found that memory for SPTs declines when the experimental conditions are made more demanding (Engelkamp & Cohen, 1991; Kausler,  76 Lichty, Hakami, & Freund, 1986; Norris & West, 1991). For example, when the number of SPTs are increased older adults' memory performance for SPTs decline (Engelkamp & Cohen, 1991). Since the method used by Dick et al. (1989) employed 36 SPTs in comparison to Karlsson et al. (1989) who used 25 this may be one reason for the different findings they report. Another reason may be the retrieval conditions that were used. Dick et al. required subjects to freely recall SPTs and VTs, whereas Karlsson et al. (1989) cued patients for target items. Research shows that cued recall conditions always elevate the amount of information that AD patients remember in comparison to when they are required to freely recall an event (Bird & Luszcz, 1991; Grober, Gitlin, Bang, & Buschke, 1992). Thus, when one considers that the Karlsson et al. study used a lower number of SPTs in addition to a cued recall test condition, this may have produced a memory task that was easier for AD patients than did the Dick et al. (1989) study. Dick et al. (1989) also compared the performance of AD patients in an SPT versus EPT task condition. In a free recall test condition Dick et al. again found no difference in the number of SPTs and EPTs that were recollected by AD patients (p. 80). Once again, the large number of tasks and the retrieval condition Dick et al. used is a likely factor in producing this result.  77 To conclude, the two studies that have looked at the way that SPTs change AD patients' memory performance have produced conflicting results. There is some evidence that requiring patients to perform a reasonable number of tasks and providing them with retrieval clues can elevate their explicit memory performance. In contrast, when a larger number of items are performed in a free recall test condition, the memory performance of AD patients is not elevated in comparison to VT or to EPT encoded tasks.  Levels of Processing AD patients retain the ability to generate the meaning of words until the later stages of their disease (Bayles & Kaszniak, 1987; Nelson & McKenna, 1978). This finding, in combination with research guided by the levels of processing framework, suggests that an encoding strategy that requires these patients to generate the meaning of targets might be useful in elevating their memory performance. Briefly, the levels of processing framework (Craik & Lockhart, 1972; Lockhart & Craik, 1990) views memory as the product of cognitive processes. One can engage in a type of sensory-level processing that does not result in a long-lasting memory trace; for example, when studying words for the number of vowels or letters it contains. In contrast, a person can engage  78 in a semantic-level of processing, and this occurs when targets are studied for their meaning. Semantic processing is viewed as resulting in a longer-lasting memory trace than sensory-level processing. Support for Craik and Lockhart's (1972; Lockhart & Craik, 1990) idea that semantically encoded information increases the durability of a memory trace (i.e., increases the amount of information that is remembered) comes from several experiments. The first was a series of experiments conducted by Craik and Tulving (1975). In their first experiment Craik and Tulving (1975) found that more words were remembered when they were studied for their meaning than for their physical characteristics. College students were asked to view targets that were exposed by a tachistoscope, and as each target was presented, they were asked to answer one of three questions. To encourage a non-semantic type of processing subjects were asked "Is the word in capital letters?". To encourage subjects to engage in processing that Craik and Tulving (1975) considered as intermediate (i.e., between non-semantic and semantic processing) they used the question, "Does the word rhyme with ^?". The question "Is the word a type of ^ ?" was used to encourage subjects to process the target word at a more semantic level. At test, subjects were asked to remember all the target words they had viewed. Irrespective of whether free recall, cued  79 recall, or recognition test conditions were used, young adults showed that they remembered significantly more targets from the semantic than from the non-semantic encoding conditions. Older adults also remember more words that they encode for their semantic versus their non-semantic characteristics (Craik, 1978; Erber, Herman, & Botwinick, 1980). However, research shows that several conditions must be present before a semantic encoding strategy will raise the memory performance of older adults to levels that are equivalent to younger adults (Craik, 1978). One condition is that older adults must be tested in a recognition or cued recall format and the second is that they are guided at encoding (Craik, 1978; Craik & Byrd, 1982). This guidance can be provided by simply having the subject state what the target means to them (Rankin & Collins, 1986). To date, there have been a limited number of studies that have used semantic encoding strategies with AD patients (Bird & Luszcz, 1991). Similar to the case of SPTs, some studies have found that AD patients recollect more meaning than non-meaning (or phonetically encoded targets) and others have not. This situation is illustrated by the following two investigations. Martin, Brouwers, Cox, and Fedio (1985) had AD patients (n = 14) and age-matched control subjects (n = 11) respond to 9 target words by either telling the experimenter where the object could be found (semantic condition), or what the word rhymed with  80 (phonetic condition). As expected, control subjects remembered significantly more words than AD patients in both the semantic and the phonemic encoding conditions. Of importance to this discussion is that AD patients recollected more items they had encoded in the semantic than in the non-semantic condition. In contrast, Corkin (1982) reported the absence of a levels effect in AD patients (i.e., more meaning than non-meaning targets being recollected). This study differed from the Martin et al. study in several ways including (1) the number of targets that were employed (n = 30), (2) the way that targets were encoded, and (3) the retrieval condition that was used. Corkin had mildly (n = 11), moderately (n = 8), and severely impaired AD patients (n = 4) make yes/no judgements as to whether (1) a man or woman would say the target word (i.e., the nonmeaning condition), (2) whether the target word rhymed with "X" (i.e., the phonetic condition), or whether (3) the target was a type of "X" (the semantic condition). At test, AD patients were only required to recognize target from distractor words. Similar to the Martin et al. investigation, Corkin (1982) found that AD patients did not remember as many targets as did control subjects. Of importance to this discussion is that, unlike the Martin et al. investigation, Corkin found that AD patients did not show a levels effect (p. 157).  81  Since Corkin (1982) used mildly, moderately, and severely impaired AD patients, this made it possible to examine whether her severely impaired patients were influencing her findings. Results from these comparisons showed that the meaning (versus non-meaning) performance of mildly impaired AD patients was boosted by approximately ten percent, moderately impaired patients showed a five percent increase, and severely impaired patients showed no increase. In combination these findings indicate that all patients were performing at levels that were at or near floor. This makes it unlikely that the inclusion of the severely impaired patients accounted for this study's results. There are additional possibilities for why Corkin may not have found a levels effect and one is that her patients may have, on average, been more impaired than those used in the Martin et al. study. Since Martin et al. did not ascertain the impairment levels of his patients it is not possible to comment further on this possibility. What does appear obvious is that similar encoding conditions were not used in these two investigations. By requesting subjects to make a yes/no response Corkin was unable to ensure that her subjects had encoded targets at the encoding level directed by her questions. This was unlike Martin et al. who required that subjects make completed responses to the experimenter. An additional factor is that Martin et al. employed a smaller number of targets than did Corkin which,  82 similar to that observed in the SPT task, may have influenced test results. This idea is supported by the work of Cermak and Reale (1978) who found that amnesic patients exhibited a levels effect only when target items were kept to a minimum (i.e., n = 12).  Theoretical Explanations SPT encode strategies. Several theories have been used to explain why memory for SPTs are higher than VTs (Engelkamp & Cohen, 1991). Although there is little consistency in these theories they do share one common theme -- that SPTs engage fewer attentionally-demanding processes than VTs (Backman & Nilsson, 1985; Cohen, 1985; Engelkamp & Zimmer, 1985). To date Backman and Nilsson (1984, 1985) have offered the most comprehensive explanation for why SPTs might require fewer attentionally demanding processes than VTs. Backman and Nilsson propose that performing a task (i.e., SPTs) engages several sensory modalities. For example, visual processing as you reach for the target, and somatomotor processes as you grasp and perform the task. In some cases auditory, olfactory, and taste modalities may also be activated. This multimodal processing presumably acts to (1) guide the encoding of targets and (2) produce a stronger representation of the target event (Backman & Nilsson, 1989). A stronger representation of the event (i.e., SPT) is presumed to occur because there are many variables on  83 which the target is processed; for example, on the basis of its colour, weight, shape, texture, smell, etc. This type of multiinformational representation presumably makes retrieval of an SPT event easier since a memory of the target can be reactivated from several sources (Backman & Nilsson, 1989). In contrast, VTs require more attentionally-demanding processes than SPTs. This is because VTs (and presumably EPTs) are typically processed unimodally via visual or auditory modalities (Backman & Nilsson, 1989). Unimodal processing is presumed to provide less guidance at encoding as well as to produce a more limited representation of the target event (Backman & Nilsson, 1989). A restricted representation of the VT event occurs because the informational features on which they are encoded are limited to semantic, phonemic and/or non-semantic event. Assumptions that VTs engage more attentionally-demanding processing than SPTs is supported by the results that have been obtained from vigilance and divided attention tasks (Backman & Nilsson, 1989).  Meaning encode strategies. Cohen, Sandler, and Schroeder (1987) have offered their explanation why the semantic encoding of targets elevates memory performance. When subjects generate the meaning of a target this guides their processing while at the same time producing a "richer" representation of the event. The richer representation makes the target easier to retrieve and it  84 lowers the attentional demands required to recollect the event. In contrast, encoding targets for their phonemic and physical attributes does not provide a similar level of guidance at encoding. This results in a mental representation that is not as easily retrieved, and thus, increases the attentional demands for recollecting the event.  Test or retrieval strategies. According to Perlmutter (1978) cued recall and recognition tests elevate memory performance because they provide information that helps to locate the target, and this reduces the attentional demands of the task. Decreasing the number of items to-be-remembered also lowers the attentional demands of the task since fewer targets presumably need to be located (Backman & Nilsson, 1989).  Study Two The goal of study two was to examine the effects that different encoding and retrieval strategies had in elevating the explicit memory performance of AD patients. This study extends previous research in several ways. First, existing research has provided conflicting results that make it impossible to determine with any certainty whether the SPT or the meaning encoded strategies significantly elevate the explicit memory performance of AD patients. Two different memory strategy experiments were designed to address this situation. One experiment examined the controversy between meaning and non-meaning encoded tasks, and  85 the second addressed SPT versus EPT tasks. These two experiments extended previous research by (1) requiring that patients in known levels of functional impairment (2) encode an ideal number of targets (i.e., n = 12) in (3) conditions that enabled the separate study of the effects that the encoding strategy and the retrieval (i.e., test) condition, as well as the combined effects that the encoding and the retrieval strategy had on memory performance.  Second, this study extends existing research by investigating the effects of a new encoding strategy in elevating  AD patients' memory performance. This strategy required subjects to engage increasing amounts of multimodal processing in conditions that required subjects to either silently read (the See condition), read and say (the Say condition), or see, say, and perform (the Do condition) tasks. Similar methods (i.e., number of stimuli, retrieval conditions) to that employed in the previous two experiments were used in this third experiment.  The overall hypothesis of study two was that AD patients would recollect more targets from the meaning, SPT, and Do encode conditions and that more targets would be recollected in either the recognition and the cued recall, than in the free recall test condition.  86  CHAPTER 5: METHOD This chapter has three main sections. The first describes how volunteers were recruited and the subject sample that was used. The next section discusses the materials used for the explicit and implicit memory tests and the method used to administer these tests. The final section describes the materials used for the memory strategy experiments and the method used to administer these experiments. All tables referred to in these sections are located at the end of the chapter.  Subjects Two groups of subjects participated in this investigation. One group consisted of Alzheimer patients from the Clinic for Alzheimer and Related Disorders, University Hospital-UBC site and the second group served as non-demented, non-institutionalized, age-matched controls. Both groups were volunteers who met the following general criteria: (1) they resided in the lower mainland or Vancouver Island region, (2) they had hearing and vision levels appropriate for psychological testing, and (3) they were manifestly right-handed (i.e., wrote with their right hand).  Alzheimer Patients Patient Recruitment. Information from the Clinic for Alzheimer Disease and Related Disorders was used to identify potential AD patients to participate in this study. After  87 qualified patients were identified (see criteria below), letters inviting them to be a member of this study were sent to their designated caregiver (see Appendix A for copies of these letters). Of the 250 patients contacted by letter, 12 agreed to participate. Eight additional patients were recruited by Dr. B. Lynn Beattie, Director of the Clinic for Alzheimer Disease and Related Disorders, who asked qualifying patients and their families to participate.  Patient Selection Criteria. To be asked to participate in this study patients had to meet the following criteria: (1) they had visited the Clinic for Alzheimer Disease and Related Disorders in the last 18 months, (2) they had a diagnosis of possible or probable Dementia of the Alzheimer's Type using NINCDS-ADRDA criteria (McKhann, Drachman, Foistein, Katzman, Price, & Stadlan, 1984; see Appendix B for NINCDS-ADRDA criteria), (3) there was no evidence of multiple strokes, cardiac insufficiency, or previous head injury recorded in their medical records, (4) they could perform basic psychological tests as indicated by the results from the neuropsychological part of their clinical assessment, and (5) both the patient and their family agreed that the patient could participate. As part of their clinic assessment all patients had their level of functional impairment evaluated using the Functional Rating Scale (FRS: Tuokko & Crockett, 1989). The FRS has sound  88 psychometric properties (i.e., an inter-rater reliability of 0.94 and validity of 0.84; Tuokko & Crockett, 1991) and scores from this scale were used to categorize patients into mild, moderate, and severe levels of functional impairment (F.I.). Categorizing patients on the basis of their FRS score enabled the further investigation of possible differences in memory performance and strategy effects due to F.I. Of the 20 AD patients who participated in this study 6 were mildly, 7 were moderately, and 7 were severely F.I. using criteria that we have established in other investigations (see Gallie, Tuokko, & Graf, 1991; Tuokko, Gallie, & Crockett, 1990). See Appendix B for a copy of the FRS and description of how scores were assigned to patients and used to categorize them into F.I. groups. Table 1 summarizes the characteristics of the Alzheimer patients who participated in this investigation. As can be seen from this table, Alzheimer patients had an average age of 70.8 years (range = 50-84) and 12.5 years of education (range = 6-18). There were almost twice as many women as men and for four patients English was not their first language. Additionally, the number of patients with possible and probable diagnoses of AD were not statistically different (see Table 3). Table 1 also shows that patients with mild, moderate, and severe levels of F.I. generally had similar characteristics. The main exception was the mildly impaired group which was slightly younger, had  89 more women than men, and took less medication. Control Subjects Control Subject Recruitment. Volunteers were recruited from talks given to local seniors' groups, radio interviews, and ads placed in community newsletters and "UBC Reports". After indicating their interest in participating all volunteers were later telephoned and asked several questions to determine their suitability for the study. (See Appendix C for the procedure and questions asked during the telephone interview). Control Subject Selection Criteria. Only individuals who were: (1) functioning independently in their home and not receiving institutional care or services, (2) provided a selfreport of good health including eyesight and hearing, (3) had no history of psychiatric or neurological disturbances including depression, memory problems, or major head injury, (4) were 50 years of age or older, and (5) were manifestly right handed (i.e., wrote with their right hand) were selected to act as controls. From a total of 80 initial volunteers, the first 40 who met selection criteria were chosen to act as non-demented controls for this investigation'. In total, approximately 75% of It was not possible to corroborate control subjects' responses since most lived alone. Thus, to ensure that controls were not demented they were asked a series of questions that are good indicators of normal functioning (i.e., absence of selfreported memory problems, current employment, active social life, non-institutional living arrangements, etc.; Reisberg, 1983). In addition, controls' test performances were examined for low scores that might indicate dementia. If volunteers had failed to pass  90 these people had responded to ads placed in UBC Reports. Twelve control subjects accepted a ten dollar honorarium at the end of the study (28 subjects declined the honorarium).  Control Subject Characteristics. Table 1 summarizes the demographic characteristics of control subjects who participated in this investigation. As can be seen from this table, controls reported an average age of 67.8 years (range = 51-89) and 14.83 years of education (range = 8-27). As in the AD patient group there were twice as many women as men in this control group.  Comparison of Control and AD Patient Characteristics Medications. All participants were asked about the medications they were taking to ensure the absence of a druginduced memory impairment. The CPA's Compendium of Pharmaceuticals and Specialties (CPA, 1992) and Goodman and Gilman's Pharmacological Basis of Therapeutics (Gilman, Goodman, Rall, & Murad, 1985) indicated that none of the medications reported by participants were likely to exert anti-mnemonic effects. Medications fell into the six general categories listed in Table 2 (i.e., analgesic/anti-inflammatory, anti-arrythmic, antidepressant, anti-epileptic, anti-hypertensive, and hormonal). As this table shows, although the control group reported taking a larger number of medications than did AD patients (i.e., 25 vs.  these measures they would have been excluded from the study.  91 7, respectively) there was a similar percentage of individuals in each group who were not taking drugs. Table 2 also shows that controls and AD patients took similar percentages of all drugs (within 5 percent) with the exception of analgesics/antiinflammatories and anti-hypertensives (controls took 15% and 10% more of these drugs, respectively). In the case of analgesic/anti-inflammatories this difference was due to a large number of controls using aspirin for arthritic purposes. Statistical analysis indicated that there were no significant differences in the total number of medications taken by the control and the AD patient group (see Table 3).  Demographic Characteristics. Statistical comparisons of control and AD patients' demographic characteristics are reported in Table 3. As the results of these analyses indicate, no statistical differences were found in the age, proportion of female and male participants, or occupation of control and AD patients. The only characteristic on which these groups were statistically different was that controls reported an average of 2.33 more years of education than AD patients (i.e., t(58) = 2.16, g < .04).  Materials and Procedures for Explicit and Implicit Memory Test Materials The same materials were used to develop explicit and implicit versions of each memory test (i.e., for written and  92 spoken words, pictures, and objects). This was possible since the main difference between explicit and implicit tests was the instructions provided to subjects at recollection.  Written and Spoken Words. To investigate explicit and implicit memory for written and spoken words, category cued recall and category completion tests were developed, respectively (see Appendix D). These tests were chosen since they enabled the same materials and study procedures to be used to index explicit and implicit memory performance. The 32 categories required to develop these tests were selected from Battig and Montague's (1969) well-established norms. To be selected, categories had to meet the following criteria: (1) they were common in everyday language (e.g., type of weather), (2) they were unlikely to produce ambiguous responses (e.g., type of fruit was chosen but pot type of ,  science), and (3) categories were unlikely to produce cohortbiased responses (e.g., type of tree was chosen but pot type of dance).' Once 32 categories were selected they were randomly organized to form eight sets of four categories. Next, three  ' Results from Howard's (1980) study where adults from different age groups made responses to 21 of Battig and Montague's categories were used to select 13 non age-biased categories. The remaining categories that were required were chosen based on the decisions of five independent raters that they were unlikely to produce cohort-biased responses.  93 words that had been provided as responses to each category by at least 30 percent of Battig and Montague's (1969) subjects were chosen (see Appendix D for the percentage of subjects responding to each item). In total, 96 words were selected from Battig and Montagues's (1969) word category norms (i.e., 3 words/category x 4 categories/set x 8 sets = 96 words in total). These words were then printed in 8 mm block lettering on blank index cards (7.5 cm by 13 cm; see Appendix D). These stimulus cards were used for written and spoken word target and baseline conditions.  Pictures. To investigate explicit and implicit memory for pictures, picture recognition and picture fragment completion tests were developed, respectively (see Appendix D). These tests were chosen since they enabled the same materials and study procedures to be used to index explicit and implicit memory performance. To develop these tests 48 pictures from Snodgrass and Corwin's (1988) well-known fragmented picture materials were used (see Appendix D).'  8 To develop these fragmented pictures Snodgrass and Vanderwaart (1980) had 219 subjects rate 260 black and white pictures on four variables known to be of central relevance to memory and cognitive functions (i.e., name and image agreement, familiarity, and visual complexity). Snodgrass and Corwin (1988) then changed these pictures using a computer that randomly deleted successively greater areas of critically important visual information. Snodgrass and Corwin (1988) then used this process to develop pictures that had eight levels of increasingly more complete visual information ranging from 8% at level one to 100% at level eight (see Appendix D for stimuli and information regarding fragmentation levels).  94 To be selected pictures had to meet the following criteria: (1) they were correctly named by more than 70% of Snodgrass and Corwin's subjects, and (2) they had received ratings of 2 or greater for both item familiarity and image consistency (scores ranged from 1 = very unfamiliar to 5 = very familiar). Overall, selected pictures were consistently named by a large percentage of subjects (i.e., 1 = 93.9%), were very familiar (11 rating = 3.67), and closely matched subjects' mental image of how the item should look (M rating = 3.72; refer to Appendix D for ratings of individual pictures). These pictures were then semi-randomly divided to make four sets of 12 pictures that had similar naming consistency, item familiarity, and image consistency characteristics (see Appendix D). Assessing explicit and implicit memory required that pictures be in two different forms, a 100% intact form (for the picture recognition test) and eight levels of picture fragments (for the picture fragment completion test). Thus, each of the 48 pictures was developed into a 100% intact and a fragmented version set (with eight levels of fragments per picture). Each of the 48 intact pictures (7 cm by 7.5 cm in dimension) was mounted on 12.5 cm by 13 cm black cardboard. In contrast, the eight fragmented versions of each picture (7 cm by 7.5 cm in dimension) were arranged on 21 by 28 cm paper with the least completed image in the lower right hand quadrant and successive  95 fragment levels continuing to the upper left corner (for examples see Appendix D).  Objects. To investigate explicit and implicit memory for objects, a tactile recognition and a tactile identification test was developed, respectively (see Appendix D). These tests were chosen since they allowed the same materials and study procedures to be used to index explicit and implicit memory performance. The 48 stimuli required to develop these tests were chosen using the following criteria: (1) they were common everyday objects (e.g., spoon), (2) items weren't sharp to the touch, and (3) they were small enough to fit behind a 90 cm by 46 cm by 20.5 cm curtained partition. Chosen objects were then randomly arranged to make four sets of 12 items (see Appendix D). Table 4, located at the end of this chapter, provides a summary of the explicit and implicit memory tests described in this section.  Procedure General. Since many of the volunteers selected to participate said they would not travel to U.B.C. all subjects were given the choice of being examined at UBC or in their home. Approximately 90% of the AD patients and 40% of the controls were tested in their homes. All subjects were tested on an individual basis.  96 Because fatigue factors affect the memory performance of AD patients and older subjects, all participants were given the choice of being tested in one or two sessions. Approximately 90% of the AD patients and 20% of the control subjects were tested in two sessions. The average time it took control subjects to complete tests and experiments was approximately 4 hours compared to 8 hours (with breaks) for AD patients. A pilot study where a counterbalanced order of tests and experiments had been used indicated that most patients could not complete testing unless tests were staggered in an easy-difficult order. Thus, a standard order of test presentation was adopted that is common to many AD investigations (cf. Heindel et al., 1989; Salmon et al., 1988). The following order of test presentation was followed for both AD patients and controls: SPT-EPT experiment, memory tests for pictures, levels of processing experiment, memory tests for written and spoken words, memory tests for objects, and the multisensory experiment (see Appendix E for listing of test order). At the beginning of the first test session all participants were verbally told of their rights as UBC subjects and were provided with a brief explanation of the study and the purpose of their involvement (i.e., to study memory in older adults). They were then asked to read a letter that contained the same information and sign it to indicate their consent to participate.  97 AD patients and controls were provided with the same consent letter. Legal guardians of AD patients were provided with a different letter that requested that they ensure that the patient in their care understand the purpose of the study (see Appendix A). All participants were told they would receive a summary of the results upon completion of the investigation. Explicit and Ii,plicit Memory Tests. All memory tests began with two practice items that were used with all subjects (see Appendix E). Practice items were used to ensure that participants understood task requirements and had an opportunity to ask questions before studying critical target stimuli. Subjects who were unable to complete practice items were given assistance and the experimenter returned to that item when the other practice item had been completed. If any subject had been unable to complete the practice items with assistance, they would have been excluded from the study. The same procedure was used to administer target stimuli and all subject responses were recorded by the experimenter in that subject's test booklet (see Appendix E). Before beginning each test the experimenter shuffled cards containing the names or pictures of the stimuli that would be used. The order in which items appeared on these cards determined the sequence of stimulus presentation. The exception was object materials, which were randomly selected by the  98 experimenter from behind a partition that prevented the items from being viewed. Similar study procedures were followed for all memory tests. Subjects were first asked to identify targets before telling the experimenter what that item meant to them. Requiring that subjects identify study items enabled the experimenter to be sure that subjects could provide the correct name for stimuli, which is a problem for some AD patients and older subjects (see Bayles & Kaszniak, 1987). Thus, if subjects could name stimuli at study, but not a test, this suggested the presence of a memory, rather than a language impairment. Requiring that subjects identify stimuli also enabled the experimenter to check that subjects were attending to critical rather than non-critical targets. In addition, having subjects generate a personal meaning for each critical target allowed the experimenter to check that subjects had focused on the target long enough to ensure that items had been processed. A main premise of this study was that the reason most previous investigations had not found implicit test performance to be similar between AD patients and controls was that they had not employed study methods which ensured that patients had focused on, and successfully encoded target stimuli. Implicit memory tests were performed before the explicit test for each material type (i.e., implicit test for pictures  99 before the explicit test for pictures). This order of testing is a standard method used to minimize the chance that subjects will discover that they can perform implicit tests by intentionally recollecting previously presented items (see Graf, Squire, & Mandler, 1984; Greene, 1988).  Category Cued Recall and Category Completion Tests for Written and Spoken Word Materials. For these tests the experimenter told subjects they would either hear or be shown words written on cards. For each word that was presented they were to first repeat the word, and then tell the experimenter what that word meant to them. For example, if the experimenter said the word car, they might say, "Car, I bought my first car when I was 16, it was... etc." Cards designated as written practice items were visually presented to the subject; those selected as spoken practice items were read to the subject. Upon completion of the four practice items, the study phase for the category completion tests for written and spoken words began. All subjects were assigned eight sets of materials that had been counterbalanced across the target and the baseline conditions of the category cued recall and the category completion tests for written and spoken word materials (see Table 5). During the study phase the 24 critical stimuli assigned to act as targets for the category completion tests for written and spoken word materials were studied using the same  100 procedure as practice items. Immediately upon completion of the study condition, the test phase for the category completion tests for written and spoken word materials began. At test, subjects were told they would now start a different task. For this task they would either be visually or verbally presented with different category labels to which they were to respond with the first three exemplars that came to mind. The experimenter emphasized that subjects should provide only those words that first occurred to them since this task was like a word association game. If subjects stated that they were attempting to intentionally recollect previously presented targets they were discouraged to do so by being told that this was a completely new task. During the test phase subjects were presented with 16 different category labels. Four of these labels corresponded to the 12 critical targets assigned to the category completion test for written word materials, and four categories matched the 12 critical targets assigned to the category completion test for spoken word materials. The remaining eight categories represented 24 stimulus words that had not been presented at study and were assigned to act as baseline measures for that subject. In each case, the modality in which the category label was presented (i.e., visual or verbal) corresponded to the original format in which the target had been presented.  101  Target scores were the number of critical target stimuli that were recollected from the written and the spoken word conditions. Baseline scores corresponded to the number of responses that matched critical baseline stimuli selected from Battig and Montague's (1969) norms (see Appendix D). The same general procedure used for category completion tests was repeated for the category cued recall tests with the exception of the test instructions that were provided to subjects. For each category label subjects were asked to tell the experimenter whether they had studied any words that were exemplars of that category, and if so, what those three words had been. Subjects were told that if they had previously been shown words that belonged to that category label there would be three words to remember. This information was provided to produce a test environment that was similar to that employed for the category completion test. Picture Recognition and Picture Fragment Completion Tests. The same general procedure used to administer the category cued recall and completion tests was followed for the picture recognition and the picture fragment completion tests. A total of four sets of 12 picture stimuli were counterbalanced across the picture recognition and the picture fragment completion target and baseline conditions (see Table 5). After completing the two standard practice items (see Appendix E) subjects were  102 then shown the 12 stimuli assigned to act as targets for the picture recognition, and the 12 stimuli assigned to act as targets for the picture fragment completion tests. Immediately after subjects named these critical target stimuli and told the experimenter what these items meant to them, the test phase for the picture fragment completion test began. At test, subjects were told that they would now start a different task. For this task they would be shown pictures that varied in their amount of completeness and the experimenter would ask them what they thought each stimulus would make when it was completed. Similar to previous implicit tests, the experimenter emphasized that this was a game that was not related to anything else the subject had previously done. At test the experimenter exposed each fragmented stimuli by lifting one of the eight flaps of a cardboard sheet that covered the eight fragmented versions of each picture (see Appendix D). The experimenter began by exposing the most incomplete level of each picture (i.e., fragment level one) and continued to uncover successive fragment levels (i.e., level two, etc.) until subjects had correctly identified the stimulus. After the stimulus had been correctly identified the experimenter then proceeded to the next fragment level. If the subject again produced a correct response the first level in which that stimulus was correctly identified was recorded. The process of proceeding to the next  103 fragment level was employed to assure that subjects were certain of their responses. This procedure was repeated for the 12 critical target and the 12 baseline stimuli assigned to the picture fragment test condition for that subject. Target scores were the average fragment level at which all targets had been correctly identified (i.e., the sum of the levels in which fragmented pictures had been correctly identified divided by 12). Baseline scores were calculated in a similar manner based on performance of the stimuli that had been assigned to that subject's baseline condition. Similar procedures to that used for the picture fragment completion test were repeated for the picture recognition test. Two exceptions were that, (1) at test the set of intact (100% complete) pictures corresponding to the 12 critical target and the 12 baseline stimuli assigned to that subject were used, and (2) subjects were requested to look at each picture and make a yes/no response regarding whether they remembered having been shown that stimulus before. The number of target and distractor pictures correctly identified as well as the false positive responses made to distractor stimuli were recorded.  Tactile Recognition and Tactile Identification Tests. The same general procedure described in previous sections was used to administer the tactile recognition and the tactile identification tests. A total of four sets of 12 object stimuli were  104 counterbalanced across the tactile recognition and the tactile completion target and baseline conditions (see Table 5). After completing the two standard practice items (see Appendix E) participants were then administered the 24 stimuli assigned to act as targets for the tactile recognition and the tactile identification tests. Subjects were asked to put both of their hands behind a curtain that prevented them from viewing the objects. They were then told that the experimenter would place an object in their hands, one item at a time, and that they should identify the item as quickly as possible and then tell the experimenter what that item meant to them. For example, "Baseball Hat", "You wear one of these to shade your eyes from the sun". Subjects were told to identify each object as quickly as possible since the experimenter was measuring the time it took for them to identify each item. Using a stopwatch the experimenter measured the time (in milliseconds) that elapsed between placing objects in the subject's hands until the item was correctly named. After completing the study phase the test phase of the tactile identification test was started. Subjects were informed that they would now be asked to identify more objects. The experimenter placed 24 stimuli, one item at a time, in the subject's hands with the request that they identify the object as quickly as possible. The 24 items consisted of the 12 target and  105 the 12 baseline items assigned to the tactile identification condition for that subject. The time until a correct identification was made by the subject was again recorded. For the tactile recognition test a similar procedure to that employed for the tactile identification test was used. The exception was that subjects were simply asked to tell the experimenter if the item that was placed in their hands was an item they had previously felt. Subjects were asked to make yes/no responses to this request as the 12 target and the 12 distractor items were randomly placed in their hands. The total number of target and distractor objects correctly identified as well as false positive responses made to distractor items were recorded. Materials and Procedures Used in the Memory Strategy Experiments Materials This section contains information on the three memory strategy experiments that comprised the second study of this investigation. The materials that were used in these experiments are described first. This is followed by a description of the procedures used to conduct these experiments. yevels of Processing. For this experiment two sets of six words having similar letter length (mean = 5.5) and number of vowels (mean = 2.5) were used. These two sets of materials were counterbalanced across meaning and letter encode conditions.  106 Nine additional words with similar numbers of letters, vowels, and general meaning to the target stimuli served as distractors that were presented at test. All words were printed in 8 mm block lettering on blank index cards (7.5 cm by 13.5 cm) (see Appendix F). SPT-EPT Experiment. For this experiment 12 items from Cohen's (1981) list of minitasks were selected. The criteria used to select these minitasks were that they: (1) were easy to perform and (2) were simple in meaning, for example, "Stretch the elastic band" and "Nod your head". Twelve minitasks were randomly divided to make two sets of six tasks. These two sets of materials were counterbalanced across SPT and EPT conditions (see Appendix F). Multisensory Strategy. For this experiment four sets of materials consisting of three sentences with similar verb stems were created. For example, one set of sentences was: (a) touch your nose, (b) touch your mouth, and (c) touch your ear. The second set of sentences was (a) make a sign for goodbye, (b) make A sign for yes, and (c) make a sign for no. These sentences were  printed in 8 mm block lettering on white paper that was mounted on black cardboard (10 cm by 34 cm). One sentence was taken from each of the four sets of materials to make three final sets of materials containing four different sentences (i.e., set one = touch your nose, make a sign for goodbye, put the pencil in the  107 case, make the clock say 3 o'clock). Each set of four sentences was assigned to either a (1) see, say and DO, (2) see and SAY or (3) a SEE only encode condition (see Appendix F). Procedure Similar to memory tests, all memory strategy experiments began with two practice items to ensure that subjects understood procedures. Assistance was provided as required, stimulus presentation was randomized, and the experimenter recorded all subject responses in that subject's test booklet (see Appendix E). Jevels of Processing Experiment. This experiment investigated whether differences would occur in AD patients' abilities to remember words studied for their meaning versus the number of letters they contained. This is referred to as the meaning versus the letter encode strategy conditions. These encoding conditions were modelled after Craik and Lockhart's (1972; Lockhart & Craik, 1990) levels of processing framework. This experiment also investigated if AD patients would recollect more targets under recognition versus free recall test conditions. Controls' performance in these encoding and test conditions was also examined. During the practice and study phases of this experiment subjects were told they would be shown cards that had different words written on them. As in previous cases, the order of  108 stimulus presentation was random. Subjects were first asked to identify the stimulus and then the experimenter would either ask them to count the number of letters in the word, or to tell the experimenter what the word meant to them. Subjects were first asked to identify the target word to ensure that the subject had focused on the stimulus. Each subject was assigned one set of six stimuli to act as targets for the meaning encode and one set for the letter encode condition (see Appendix F). These two sets of materials were counterbalanced across each encoding condition (see Table 7). Immediately after the study phase finished the test phase began. At test the experimenter asked subjects to freely recall all the targets they had just studied. If the subject was unable to recollect all the targets, the experimenter then shuffled the study and distractor stimulus cards together and then presented each card to the subject, one card at a time. Subjects were asked to respond by saying "yes" when the experimenter displayed a word they had seen at study, or "no" when it was a word they had not studied. This represented the recognition test of this experiment. The number of meaning encoded words that were correctly identified in the free recall and then the recognition condition were recorded. Similarly, the number of letter encoded words that were remembered in the free recall and then the recognition  109 condition was recorded. Any target recollected in both free recall and recognition conditions was included only in the free recall score.  SPT-EPT Experiment. This experiment investigated whether differences would occur in AD patients' abilities to remember minitasks they performed (Subject-Performed Minitask, SPT) versus those they watched the experimenter perform (i.e., EPT). This is referred to as the SPT versus the EPT encode strategy conditions. These encoding conditions were modelled after Cohen (1983) and Backman and Nilsson's (1984) work. This experiment also investigated if AD patients would recollect more targets under cued than free recall test conditions. Controls' performance in these encoding and test conditions was also examined. At practice, and during the study phase of this experiment, subjects were told that they would be performing some simple tasks. Subjects were told that the experimenter had these tasks written on cards, some of which the experimenter would do and some that the subject would be asked to perform (see Appendix F). Before they could perform these tasks, however, subjects would need to identify the physical objects they would use to perform these tasks. The experimenter then pointed to each object, one at a time, and asked the subject to name each item. This ensured that failure to recollect minitasks was more likely due to problems in recollecting the event rather than an inability to  110  produce the name for the objects involved in task performance. At study, all commands were read by the experimenter who had randomized their order of presentation by shuffling the stimulus cards. Each subject was assigned one set of materials that served as the targets for the SPT encode condition, and one set of materials for the EPT encode condition. These two sets of materials were counterbalanced across each encoding condition (see Table 7). Upon completion of the study phase, the objects used to perform the minitasks were removed from the subject's view. Testing began immediately after the study phase and subjects were asked to recollect all the minitasks that either they, or the experimenter, had performed. Subjects were told they did not need to remember who did what task, just what had been performed. This represented the free recall condition of this experiment. Tasks that were not remembered by the subject were cued by the experimenter. This was done by placing the objects used to perform these tasks in front of the subject and the experimenter provided verbal cues for items that had not been recollected. For example, the minitask "Cross your fingers" was verbally cued by saying, "One of the commands said to cross something, do you remember what that was?" (see Appendix F for cues). The number of SPT-encoded minitasks freely recalled and then the remaining items recollected with cues were recorded.  111  Similarly, the number of EPT-encoded minitasks that were recollected in the free recall and then the cued recall conditions were also recorded. Multisensory Experiment. This experiment investigated whether AD patients would remember more sentences if they (a) read the sentence to the experimenter and then performed what the sentence said (the DO condition), than if they (b) just read the sentence to the experimenter (the SAY condition), or if they (c) read the sentence silently to themselves (the SEE only condition). This is referred to as the Do, Say, and See encode conditions, respectively. This experiment also investigated whether AD patients would recollect more sentences in cued than free recall conditipns. At practice and during the study phase of this experiment subjects were told they would be shown cards with sentences written on them, some of which they would be asked to read silently to themselves, some of which they would be asked to read aloud to the experimenter, and some that they would be asked to read aloud and then do what the card said. Stimulus presentation was randomized. Objects required to perform these sentences were placed in front of the subject. The experimenter pointed to each of the objects, one item at a time, and asked subjects to name the presented stimulus. As in previous experiments, this procedure was conducted to ensure that if stimuli were not  112 recollected it was more likely to be due to a memory, rather than word finding difficulty. After the study phase of this experiment was completed objects were removed from the subject's view. Each subject was assigned one set of materials to act as targets for the See condition, one set for the Say condition, and one for the Do condition (see Appendix F). These three sets of materials were counterbalanced across each encode condition (see Table 7). At test, subjects were asked to freely recall all the sentences they had been shown. Sentences that were not remembered by the subject were then cued by the experimenter (i.e., the cued recall condition). This was done by placing the objects used to perform the sentence commands in front of the subject and the experimenter then provided verbal cues for any targets not recollected. For example, the sentence that said to "Touch your ear" was cued by the experimenter saying "One sentence said to touch something, do you remember what that was?" The number of See, Say, and Do encoded sentences that were freely recalled were recorded. Similarly, the number of See, Say, and Do encoded sentences that remained and were recollected with cueing were also recorded. See Appendix F for a listing of the materials used in memory strategy experiments and Tables 6 and & 7 for a summary and listing of counterbalancing procedures.  113 Table 1 Subject Characteristics. AD Patients Controls (n=40I  All (n=20I  Mild* (n=6I  Moderate* (n=7I  Severe* (n=7I  Will M^67.8  70.8  68.2  70.0  73.9  9.9  8.2  10.9  6.5  7.2  =^  50-84  50-79  58-79  62-84  EDUCATION M^14.83  12.5  12.3  13.4  11.7  ZD^4.2  3.3  2.6  4.1  3.2  Range^8-27  6-18  10-17  6-18  8-16  5 1  4 3  4 3  Range^51-89  GENDER  Women^27 Men^13  13 7  AVERAGE NUMBER OF MEDICATIONS IS^0.63  0.35  0.16  0.42  0.42  5.12^0.95  0.59  0.41  0.55  0.55  3 2 1  4 1 2  4 1 2  3 4  3 4  OCCUPATION  White Collar+ 24 Blue Collar++ 10 Housewife^6  11 4 5  DIAGNOSIS Possible^n/a Probable^n/a  9 11  Note. * Level of Functional Impairment + White Collar (e.g., teacher, nurse) ++Blue Collar (e.g., painter, bus driver)  114  Table 2 Medications Taken by Controls and AD Patients. Medication  Analgesic/ Anti-inflammatory  Controls (n=40)  AD Patients (n=20)  N^I  N  I  10  25  2  10  Anti-arrythmic  3  7.5  1  5  Anti-depressant  3  7.5  2  10  Anti-epileptic  1  2.5  1  5  Anti-hypertensive  4  10  0  0  Hormonal*  4  10  1  5  Total number of Medications  25  Number of Individuals Taking Medications  15  7  37.5  6  30  Note. This table summarizes medications taken by AD patients as recorded in their medical files and updated by a spouse or sibling collaborator. For control subjects this information was obtained on a self-report basis. * Medications in this category included estrogen and thyroid replacement.  115  Table 3 Comparison of Control (n=40) versus AD Patients' (n=20) Characteristics.  Variable^Statistic^Probability AGE^t= -1.18 ++  0.24  58  EDUCATION^t= 2.16 ++  0.04*  58  2 GENDER^X  1.00  1  0.24  58  0.55  0.76  2  0.95  0.62  2  0.00 +  I MEDICATIONS j,= 1.18 ++ OCCUPATION^  x  2  ;X:=  POSSIBLE/^ PROBABLE Dx.^2  Note. * Significant at p <. 05 + Yates Correction For Continuity (Yates, 1934) ++ Based on Pooled Variance Estimates to adjust for unequal sample sizes.  116  Table 4 Explicit and Implicit Memory Tests Used in this Investigation.  Material Type^Test Type^Test Name Written words^Implicit^Category completion Explicit^Category cued recall Spoken words^Implicit^Category completion Explicit^Category cued recall Pictures^Implicit^Picture fragment completion Explicit^Picture recognition Objects^Implicit^Tactile identification Explicit^Tactile recognition  117  Table 5 Counterbalancing Method Used for Explicit and Implicit Tests.  Four sets of written word, spoken word, picture, and object materials were counterbalanced across Implicit and Explicit Memory Test, target and baseline conditions. Note, for written and spoken word tests the same materials were used so that a total of eight sets of items were counterbalanced across written and spoken word conditions. Counterbalancing was conducted across test conditions for the AD patient and the control subject groups. The following diagram shows the counterbalancing procedure that was used. IMP designates the implicit and EXP the explicit test conditions. denotes target and B denotes baseline. Each line denotes the counterbalancing schedule for one AD patient or one control subject. Thus, line one is for subject one, line two for subject two, etc. Each number (i.e., 1, 2, 3, 4, etc.) denotes the material set number (see Appendix D for items contained in each set of materials).  Imp/Exp for Written Words  Imp/Exp for spoken Words  Imp/Exp for Pictures  INE 1  EKE T$  IKE^EKE T T  TB  EXP TB  ME^EKE TB TB  1 2 3 4 5 6 7 8  34 43 56 65 78 87 12 21  5 6 7 8 1 2 3 4  1 2 3 4  3 4 1 2  1 2 3 4  2 1 4 3 6 5 8 7  etc.  6 5 8 7 2 1 4 3  7 8 1 2 3 4 5 6  8 7 2 1 4 3 6 5  2 3 4 1  etc.  4 1 2 3  Imp/Exp for Objects  2 3 4 1  etc.  3 4 1 2  4 1 2 3  118  Table 6 Summary of the Memory Strategy Experiments Used in this Investigation.  Each of the three memory strategy experiments listed below had (1) two different encoding conditions, and (2) two different test retrieval conditions. The exception was the multisensory experiment which had three different encoding conditions.  Name^Experimental Procedure Levels of^In this experiment subjects studied Processing^twelve words using two different encoding strategies. (1) Six words were studied by telling the experimenter what the word meant (= meaning encode condition), and six words by counting the number of letters the word contained (= letter encode condition). (2) Retrieval conditions began with a free recall then recognition condition. SPT-EPT Strategy^In this experiment subjects studied twelve minitasks using two different encoding strategies. (1) Six minitasks were studied by having the subject perform the minitask (= SPT or active encode), and six minitasks by watching this experimenter perform tasks (= EPT or passive encode strategy). (2) Retrieval conditions began with a free recall and then cued recall condition. Multisensory Strategy  In this experiment subjects studied twelve simple sentences. (1) Four sentences were studied having subjects read sentences silently to themselves (= See encode), four sentences were studied having subjects read sentences aloud to the experimenter (= Say encode) and four sentences were studied having subjects read the sentence to the experimenter and perform what the sentence said (i.e., touch your nose = Do encode condition). (2) Retrieval conditions began with a free recall then cued recall condition.  119  Table 7 Counterbalancing Method Used for the Memory Strategy Experiments.  In the (1) Levels of Processing and (2) SPT-EPT Memory Strategy Experiments two sets of materials were counterbalanced across encoding conditions. Three sets of materials were counterbalanced across the Multisensory Strategy Experiment. Similar to Implicit and Explicit Tests counterbalancing was between subjects. That is, counterbalancing of AD patients was maintained separate from control subjects. Each line denotes the counterbalancing schedule for one AD patient or one control subject. Thus, line one is for subject one, line two for subject two, etc. Each number (i.e., 1, 2, 3) denotes the material set number. See Appendix F for items contained in each set of materials. Levels of Processing SPT-EPT ^Multisensory Experiment ^Bxperiment^Experiment Mean / Letter^Subject/Experimenter^See/ Say / Do 1^2^ 1^2^1^2^3 2^1^ 2^1 ^2^3^1 etc.^ etc.^ 3^1^2 etc.  120 CHAPTER 6: RESULTS This chapter has three main sections. The first section describes the design and the analyses of results. The next section reports the results obtained from explicit and implicit memory tests. In the final section I examine the data from the memory strategy experiments. Tables reporting descriptive and inferential statistics are located in Appendix G.  Overall Design and Analyses Repeated measures Multivariate Analyses of Covariance (MANCOVA: Subjects' level of education as the covariate) with one between-subjects factor (Subjects: AD patients, Control Subjects) and one within-subjects factor (Test Type: Explicit, Implicit) were used, where described, to analyze the memory test results'. Similar statistical designs were used to analyze the results from the memory strategy experiments with the exceptions that the within-subjects factor for the levels of processing and the SPTEPT experiments were Encoding Strategy: Retrieval Test, and each of these factors had two levels. For the multisensory experiment there were three levels to the Encoding Strategy factor (i.e., the Do, Say, and See conditions).  All multivariate analyses of covariance (MANCOVA) tests were repeated using multivariate analyses of variance (MANOVA) to determine covariate effects on results. In no case did MANCOVA and MANOVA analyses result in different conclusions being reached. Thus, the 2.33 years more education reported by control subjects did not appear to significantly influence the memory test results or the memory strategy experiment results.  121 Repeated Measures MANCOVA analyses were used since they provided the most powerful and relevant statistical tool for addressing the hypotheses of memory tests and strategy experiments (see Davidson, 1972; Kirk, 1968; O'Brien & Kaiser, 1985). To guard against making Type 1 errors Bonferonni corrections were applied. For study one the alpha level per Mancova analyses was set at .01 (i.e., 5 analyses/.05) and for study two alpha was .0167 (i.e., 3 analyses/.05) following the procedure recommended by Rosenthal and Rosnow (1991).  Analyses Significant main effects were followed by post hoc tests and significant interactions were analyzed with main effect tests (as per Winer, Brown, & Michels, 1991). To correct for unequal subject numbers the unweighted means approach using dummy variables (as discussed in SPSS-X, 1990) was used in each MANCOVA analysis (as per Howell, 1987; Rosenthal & Rosnow, 1991; Winer, Brown, & Michels, 1991). To test for homogeneity of variance Box's test for multivariate F's and then the Bartlett-Box test" for univariate F's were performed (as per Glass & Hopkins, 1984; Howell, 1987). When heterogeneity of variance occurred, Box's (1954) procedure for adjusting the degrees of freedom for F  " Several persons have made contributions to this test which has also been referred to as the Box test, Bartlett-Kendall, and Scheffe test (Glass & Hopkins, 1984). This test was chosen since it has special applicability when group n's are not equal (Glass & Hopkins, 1984).  122 was employed as described by Howell" (1987, p. 297). When studying patients of mixed dementia severity it is important to determine if outlying scores have biased average group results. Scrutiny of the results obtained in the memory test and the strategy experiments showed that the standard deviations of most variables were within acceptable ranges (i.e., the majority are within 3 $Ds). Examination of raw scores confirmed the general absence of extreme scores except when noted in the discussion of the results. Additionally, checks on normality were performed by establishing that AD patients' score distributions were similar in shape to those of control subjects and devoid of extreme platykurtic and leptokurtic characteristics following the procedure described by Glass and Hopkins (1984). Study One Explicit and Implicit Memory Tests The focus of this part of the investigation was to explore AD patients and non-demented, age-matched control subjects' performance on four sets of explicit and implicit memory tests (see Table 4 of Method section for listing). The overall hypothesis was that differences in the memory performance of AD  n Box (1953, 1954) has shown that when both variance and group size are unequal, a valid and conservative test of the significance of group differences can be carried out if the degrees of freedom for F mites are altered from df=(k-1, k(n-1)) to df=(1, n-1). If this adjustment leads to a significant result then group means are significantly different regardless of the variances.  123 patients and controls would be smaller on implicit than explicit tests, using written and spoken word, picture, and object materials. A strict scoring criterion was applied to all memory tests. Responses were considered correct only if they matched the essential form of the original stimulus. For example, changes in tense and plural forms were acceptable, but not substitute words with similar meaning. Category Cued Recall and Category Completion Tests for Written Word Materials Critical Measures. For the category cued recall and the category completion test the critical dependent measure was the number of written words identified in the target and the baseline conditions. Items were scored as correct if they matched one of the 12 words presented at study (the target condition) or one of the 12 words designated to act as baseline for that subject. Corrected scores were computed by subtracting target from baseline performance. This is the standard method employed for obtaining an index of priming or implicit memory performance (see Graf, Shimamura, & Squire, 1985; Salmon, Shimamura, Butters, & Smith, 1988). The corrected score provides several advantages to using the target score in the analyses of results. One advantage is that corrected scores provide an index of the magnitude of priming that is not influenced by differences in the subjects'  124 baseline scores (Salmon et al., 1988). A second advantage is that corrected scores are not as biased by guessing and/or chance performance on tests (see Graf et al., 1985; Salmon et al., 1988). Average percentage corrected scores were computed by dividing the corrected score by 12 and multiplying by 100. AD Patients and Control Subjects. Figure 2 shows the average percentage corrected scores for the AD patients and the control subjects on the category cued recall test and on the category completion test for written word materials. As expected, the average percentage corrected category cued recall performance of AD patients was much lower than that of the agematched control subjects (i.e., N = 15.4%, 5D = 2.28 vs. M = 72.3%, 5D = 1.83). In contrast, the average percentage corrected scores on the category completion test was more similar between AD patients and control subjects (M = 10.8%, 5D = 2.20 vs. M = 13.5%, 5D = 2.39, respectively). Some additional points must be made regarding the results ' obtained from the category cued recall and the category completion tests. The first is that patients' and controls' performance on both tests were not at floor. This was confirmed by inspecting the group where the smallest difference between target and baseline performance had occurred to confirm that a significant difference existed between scores [i.e., AD patients on the category completion test, t(19) = 2.36, p < .032].  0 0  0  o  -  'S V  0 CO  0  co  Control Subjects (n-40) All AD Patients (n-20) Mild (n-6) Moderate (n.7) Severe (n-7)  0 o to  0 0 • 0  v.  co  'et  0 Ns'  O  O  0  O  0  csi  Control (n•40)  All AD Patients (n-20)  Mild^Mod.^Severe (n-6)^(n.7)^(n-7)  irA Control^All AD^Mild^Mod.^Severe Patients (n-40)^(n-20)^(n-6)^(n-7)^(n-7)  Explicit Test (Category Cued Recall) ^  Implicit Test (Category Completion)  Figure 2. Percentage corrected score of Alzheimer patients and of Control subjects on explicit and implicit tests for written word materials. Mildly, moderately and severely impaired patient groups' performance are shown separately. The error bars represent standard deviation scores.  126 Second, statistical inspection showed that there was a significant difference between the baseline performance of controls and AD patients on the category completion [E(1,58) = 10.56, R < .005] but not on the category cued recall test [E(1,58) = .13, R > .5]. Since the corrected scores obtained for the controls and the AD group on the category completion test was not based on similar baseline levels, some caution is required in the interpretation of this test result. That is, although the magnitude of the corrected scores may be similar they are derived from different performance levels that may reflect different memory processes. For instance, controls may have been engaging in the explicit recollection of items while performing implicit tests although the method used in this study (i.e., presenting implicit before explicit tests) would have minimized this effect. It should be noted that the magnitude of the corrected scores obtained on the category completion test was similar to those obtained in a pilot investigation where a similar task was used with different subjects (see Graf, Tuokko, & Gallie, 1990). To examine if AD patients showed similar average percentage corrected scores to controls on the category completion test, despite lower performance on the cued recall task, a repeated measures Multivariate Analysis of Covariance (covariate factor =  127 education) was performed. The results of this MANCOVA l2 showed significant main effects for subject group" f(1,57) = 51.14, p < .001, MSe = 306.47 and test type f(1,58) = 110.33, p < .001, MSe = 385.07. A significant interaction between group and test type was also detected, f(1,58) = 80.70, g < .001, MSe = 281.67. When the group by test interaction was analyzed with main effect tests, significant differences were evident between control and AD patients' corrected performance on the category cued recall test f(1,58) = 157.08, p < .001, MSe = 621.07, but not on the category completion test, were found (see Table G-8b). Further examination confirmed these findings by showing that there was a larger effect of test type on subjects' performance on the category cued recall (eta-squared = .73) than on the category completion test (eta-squared = .004). The next step was to decide whether the significantly lower corrected scores of the AD patients on the category cued recall test was influenced by the instructions provided at test, or to an inability to perform the basic task of providing category exemplars. Since AD patients and controls had provided a statistically similar number of responses to baseline items (as previously reported) this suggested that these patients could  12 Results of the Box M and Bartlett Box tests confirmed that homogeneity of variance assumptions were met in this analysis. " One degree of freedom is lost from the within source of variation due to the covariate of education.  128 perform the non-memory, or task-related requirements of the test (see Salmon et al., 1988, p. 487; Shimamura et al., 1987, p. 349). To summarize, these results supported the hypothesis that differences in the memory performance of AD patients and controls would be smaller on implicit than explicit tests of written word materialS. These results indicate that when AD patients were provided with test instructions that did not require them to intentionally recollect written word materials, they showed memory performance that was similar to non-demented, noninstitutionalized, age-matched controls. In contrast, when the same patients were instructed to intentionally recollect the same materials, they showed a corrected score performance that was significantly lower than controls. Mildly, Moderately. and Severely Impaired AD Patients. A secondary question was how AD patients with different levels of functional impairment (F.I.) would perform on the category completion test, despite expected variations in their performance on the category cued recall test (see Gallie, Tuokko, & Graf, 1991; Tuokko, Gallie, & Crockett, 1990). Figure 2 shows that patients that were in mild, moderate, and severe levels of F.I. displayed similar average percentage corrected scores for the category completion test. In contrast, there was a greater variation in the average percentage corrected performances of  129 these patients on the category cued recall test. The small number of patients in each F.I. group precluded the use of inferential analyses. However, similar findings were found on subsequent implicit tests suggesting that the consistent levels of performance in mildly, moderately, and severely F.I. patients was not artifactual in nature (note that this performance was not at floor). Refer to Table G-8a for a complete listing of these results. Category Cued Recall and Cateaory Completion Tests for Spoken  Word Materials Critical Measures. These were the same as described for written word materials with the exception that stimuli were verbally presented.. AD Patients and Control Subjects. Figure 3 shows the average percentage corrected scores for the AD patients and the  control subjects on the category cued recall and on the category completion test for spoken word materials. Similar to the results obtained using written word materials, the average percentage corrected performance on the category cued recall test for AD patients was lower than that of age-matched controls (i.e., M = 20.4%, SD = 2.56 vs. M = 70%, SD = 2.67, respectively). In contrast, the performance of AD patients and controls was closer on the category completion test (M = 9.2%, SD = 1.55 versus M = 14.4%, SD = 2.48, respectively).  0 -  0 0 -  O  O  0  7  , ,  Control Subjects (n-40) All AD Patients (n-20) Mild (n-6) Moderate (n-7) Severe (n-7)  CO  CO  ar  8 O  to  O  4.)  rn al Co rU  O  0  O  0  Control (n-40)  All AD Patients (n-20)  Mild^Mod.^Severe (n.6)^(n-7)^(n-7)  Explicit Test (Category Cued Recall)  Control^All AD Mild^Mod.^Severe Patients (n-40)^(n-20)^(n-6)^(n-7)^(n-7)  implicit Test (Category Completion)  Figure 3. Percentage corrected score of Alzheimer patients and of Control subjects on explicit and implicit tests for spoken word materials. Mildly, moderately and severely impaired patient groups' performance are shown separately. The error bars represent standard deviation scores.  131 Two additional points must be made regarding the results obtained from the category cued recall and the category completion tests. The first is that patients' and controls' performance on both tests were not at floor. This was confirmed using the same method that was previously described (i.e., AD patients' category completion performance, t(19) = 2.78, p < .013). Second, statistical inspection showed that there was a significant difference between the baseline performance of controls and AD patients on the category completion [F(1,58) = 11.95, p < .002] but not on the category cued recall test [E(1,58) = .053, p > .5]. Thus, the magnitude of the average percentage corrected scores obtained for AD patients and controls on the category completion test were not based on similar performance levels. To examine whether AD patients showed similar average percentage corrected scores to controls on the category completion test, despite lower performance on the category cued recall test, a Multivariate Analysis of Covariance (covariate = education) was performed. The results of this MANCOVA showed a significant main effect for subject group E(1,57) = 36.44, p < .001, MSe = 232.71. Since the Bartlett-Box test indicated a failure to meet assumptions of homogeneity of variance on the factor of implicit test type, Box's procedure for adjusting F critical  was applied (Box, 1954). With Box's adjustment the main  132 effect of test type continued to be highly significant, E(1,57) = 80.68, 12 < .001, MSe = 429.34 together with a significant interaction between group and test type, E(1,57) = 35.52, R < .001, MSe = 189.04. When the group by test interaction was analyzed with main effect tests and Box's adjustment, significant differences between the controls' and the AD patients' average percentage corrected performance on the category cued recall test Z(1,58) = 68.01, p < .001, MSe = 472.03, but not on the category completion test were found (see Table G-9b). Eta-squared analysis confirmed that the effect of test type on subjects' performance was much larger for the category cued recall (.54) than on the category completion test (.0fl. Statistical analysis of baseline performance showed that AD patients and controls performed at similar levels on the category cued recall test (values previously reported) suggesting that patients were able to perform the basic task of providing category exemplars but had difficulty when intentionally recollecting target stimuli. To summarize, the results reported in this section supported the hypothesis that differences in the memory performance of AD patients and controls would be smaller on implicit than explicit tests of spoken word materials. Similar to conclusions reached from tests using written word materials, it appeared that when AD patients were directed to non-intentionally recollect targets  133 they showed memory performance that was similar in magnitude to non-demented controls. In contrast, the same patients encountered difficulty when directed to intentionally recollect the same stimulus materials.  Mildly, Moderately, and Severely Impaired AD Patients. A secondary question was whether mildly, moderately, and severely F.I. patients would exhibit similar average percentage corrected scores on the category completion test despite expected differences on the category cued recall test. As Figure 3 shows, a similar profile of results to that found with written word materials was obtained. The average percentage corrected score on the category completion test was similar for mildly, moderately, and severely F.I. patients (note that this performance was not at floor). However, this was not the case for the category cued recall test where variations in performance across F.I. groups can be observed. (Refer to Table G-9a for a complete listing of these results).  Picture Recognition and Picture Fragment Completion Tests Critical Measures. For the recognition test the critical dependent measure was the number of pictures correctly identified in the target (n = 12) and the baseline conditions (n = 12). In the picture fragment completion test the critical dependent measure was the fragment level (of 8 possible levels) at which the target (n = 12) and baseline items (n = 12) were correctly  134 identified. For each of these tests, responses were scored as correct if they matched stimuli from the target and baseline conditions assigned to each subject. Average percentage corrected scores were computed for both the picture recognition and the picture fragment completion tests. In the case of the picture recognition test, average percentage corrected scores were computed by subtracting target (hits) from false positive (or false alarm) responses, dividing by 12 and multiplying by 100. Subtracting target from false positive scores is the method recommended for correcting AD patients' performance on recognition tests (see Braconnier, Cole, Spera, & De Vitt, 1982; Snodgrass & Corwin, 1988). Average percentage corrected scores for the picture fragment completion test were computed by subtracting target from baseline scores, dividing by 12 and multiplying by 100.  AD Patients and Control Subjects. Figure 4 shows the average percentage corrected scores for the AD patients and the control subjects on the picture recognition and the picture fragment completion tests. As expected, the average percentage corrected picture recognition score of AD patients was much lower than that of controls (i.e., }1 = 49.2%, alp = 3.78 vs. 11 = 97.5%,  5.p = .61, respectively). AD patients' average percentage corrected picture fragment completion score was also lower than controls (i.e., X = 6.1%, SD = .64 vs. M = 13.1%, Sp = .65, respectively). These performance scores were similar in  0 0  0  0  Control Subjects (n-40) All AD Patients (n-20) Mild (n•6) Moderate (n.7) Severe (n.7)  0 co  0 03  :12 0  40  le, o 0, o,  e ti) > < 0  0 NT  O  O CM  01  0  0  Control^All AD^Mild Mod. Severe Patients (n•40)^(n.20)^(n-6)^(n.7)^(n.7)  Explicit Test (Picture Recognition)  1011 Control (n-40)  All AD Patients (n•20)  Mild^Mod.^Severe (n.6)^(n.7)^(n-7)  Implicit Test (Picture Fragment Completion)  Figure 4. Alzheimer patients' and Control subjects' performance on explicit and implicit tests for picture materials. Percentage corrected scores are shown. Mildly, moderately and severely impaired patient groups' performance are shown separately. The error bars represent standard deviation scores.  136  magnitude to those obtained by Bondi and Kaszniak (1991) using the picture fragment completion test with a pleasantness encode condition (i.e., AD patients: n = 12, N = 4.2%; Controls: n = 16, M = 16%; M. Bondi, personal communication, June, 1992). Patients' and controls' performance was not at floor on the picture recognition test (i.e., AD patients' target vs. baseline performance, E(1,38) = 31.57, R < .001). However, AD patients' performance was at floor on the picture fragment completion test, t(19) = 1.87, R > .98. This finding corresponds to a similar problem reported by Heindel et al. (1990) using a picture fragment completion test with 5 levels of picture fragments. Thus, using a test with 8 different fragment levels still did not provide enough discriminating ability to detect differences in AD patients' performance. The baseline performance of AD patients and controls was significantly different on both the picture recognition [E(1,58) = 16.81, p < .001], and the picture fragment completion tests [E(1,58) = 12.17, R <.002]. These findings are similar to those reported by Bondi and Kaszniak (1991) and Heindel et al. (1990) and they suggest that AD patients were encountering greater difficulty than controls in performing the task components of the picture recognition and picture fragment completion tests. To examine the hypothesis that differences in the memory performance between AD patients and controls would be smaller on  137 the picture fragment completion than the picture recognition test a repeated measures Multivariate Analyses of Covariance (covariate factor = education) was performed on subjects' average percentage corrected scores. The results of this MANCOVA showed significant main effects for subject group E(1,57) = 90.36, g < .001, MSe = 18233.71. Since the Bartlett-Box test indicated a failure to meet assumptions of homogeneity of variance on the factor of explicit test type, Box's procedure for adjusting F critical  was applied (Box, 1954). With Box's adjustment the main  effect of test type continued to be highly significant [i.e., E(1,57) = 534.56, 12 < .001, MSe = 105139.41] together with a significant interaction between group and test type, E(1,57) = 60.57, 2 < .001, MSe = 11913.92. When the group by test interaction was analyzed with main effect tests, significant differences between control and AD patients' performance on both the picture recognition [E(1,57) = 90.90, 12 < .001, MSe = 31148.15] and the picture fragment completion test were found [E(1,58) = 9.36, p < .003, MSe = 489.55; Refer to Table G-10b]. These findings did not support the hypothesis that there would be smaller differences between the memory performance of AD patients and controls on the implicit but not the explicit test of picture materials. Eta-squared analyses indicated that there had been a larger correlation between subject performance on the explicit (eta-  138 squared = .61) than the implicit test (eta-squared = .14). These findings in conjunction with those reported on differences in baseline and floor effects suggest that the failure to find support for the hypothesis was due to the tests that were used. Specifically, that the picture fragment completion test was too difficult for AD patients and the picture recognition test was too easy for controls. Together this produced a test environment where there were large average corrected score performance differences between the AD patient and control group. Mildly, Moderately, and Severely Impaired AD Patients. As Figure 4 shows, the average percentage corrected score performance on the picture fragment completion test were similar for mildly, moderately, and severely F.I. patients but this performance was at floor. In contrast, performance on the picture recognition test was not at floor and varied with F.I. group. Patients in the mildly F.I. group showed, on average, much higher average percentage corrected scores on this test than did the moderately and severely impaired patients. (Refer to Table G-10a for a complete listing of these results). Tactile Recognition and Tactile Identification Tests Critical Measures. For the tactile recognition test the critical dependent measure was the number of objects correctly identified in the target (n = 12) and the baseline conditions (n  139 = 12). For the tactile identification test the critical measures were similar with the exception that the time required (in milliseconds or ms) to identify stimuli were recorded. Average percentage corrected scores for the tactile recognition test were computed using the same method employed for the picture recognition test. For the tactile identification test timed responses were obtained for target items assigned to both the tactile recognition and tactile identification tests enabling two different corrected scores to be computed. A corrected score based on the re-identification of the same targets (i.e., targets assigned to the study and test conditions for the tactile identification test; the old materials condition) and a corrected score based on targets assigned to the tactile recognition test compared to targets assigned to the tactile identification test (i.e., the new materials condition). Average percentage corrected scores were then computed for the old and new materials condition based on a proportion of baseline performance (i.e., the second presentation of targets assigned to the tactile identification test acted as baseline for both the new and old material conditions, refer to Table G-lla) which was then multiplied by 100. There were large variations in the timed responses of subjects on the tactile identification test. To decrease this variation a log transformation of the data was performed but this  140 resulted in the loss of subject data and mean square values of zero in the MANCOVA analyses. Average median timed responses were then computed as used in previous studies that have obtained timed response data from AD patients (see Knopman, 1991). Using the average median timed responses decreased some of the variation in subject responses without the subsequent loss of subject data.  AD Patients and Control Subjects. Figure 5 shows the average percentage corrected performance scores of AD patients and control subjects on the tactile recognition and the tactile identification test". As this figure shows, there were large differences in the performance of AD patients and control subjects on both memory tests. For the tactile recognition test AD patients recognized 49.17% (02 = 3.96) of the targets compared to control subjects whose performance was at ceiling with 97.25% (ap = .94). This ceiling effect likely masked the true magnitude of the differences in group performance. For the tactile identification test results depended on whether priming was based on "old" or "new" stimulus materials (see Table G-lla). AD patients showed a similar average percentage corrected score to controls when performance was based on the old material condition (i.e., if = 59.29%, 0 = 454.41 vs.  " Four AD patients were unable to complete the tactile identification test and so their average performance scores were removed as per Howell, (1987).  0 0  Control Subjects (n 40) All AD Patients (n-16) -  Mild (n-6) Moderate (n.5) Severe (n.5)  0  03  cS  0  0  cy.  f■••••••••■■•  0 0  0  ^ ^ Mild Mod.^Severe Control^All AD Patients (n.40)^(n.16)^(n.6)^(n.5)^(n.5)  Explicit Test^ (Tactile Recognition)^  Priming 1 (Old Materials)  Priming 2 (New Materials)  Implicit Test (Tactile Identification)  Figure 5. Alzheimer patients' and Control subjects' performance on explicit and implicit tests tor objects. Percentage corrected scores are Mildly. moderately and severely impaired patient groups' pertormance are shown separately. The error bars represent standard deviation scores.  142 N = 64.87%, SD = 103.04, respectively). In contrast, when performance was from the new materials condition control subjects showed much higher average percentage corrected scores than did AD patients (i.e., M = 73.96%,  an = 84.80 vs. M = 13.19%, an =  577.84). Additional statements must be made about these findings. The first is that despite computing average median timed responses a large variation in subject responses remained. Similar findings have been reported by other investigators (e.g., Nissen & Knoopman, 1987) and it suggests that these findings are highly variable, especially in patients in severe levels of F.I. The large variances in these findings makes it impossible to determine whether AD patients' performance was at floor or whether significant differences existed between group baseline performances. The first question that was addressed was whether the average percentage corrected performance of AD patients were similar to that of controls on the tactile identification test for "old materials", despite lower performance on the tactile recognition test. A repeated measures MANCOVA with Box's adjustment detected only a significant group by test type interaction [f(1,53) = 9.56, R < .003, 4Se = 16976.41]. When the group by test interaction was analyzed with main effect tests significant differences between the controls' and the AD  143 patients' average percentage corrected performance on the tactile recognition [i.e., E(1,53) = 77.55, R < .001, MSe = 30880.21] but not on the tactile identification test was found (see Table G11b). Eta-squared analyses confirmed that the effect of test type on subject performance was much larger for the tactile recognition (.59) than on the tactile identification test (.007). To summarize, these results supported the hypothesis that differences in the memory performance of AD patients and that of controls would be smaller on the implicit test of "old materials" than on the explicit test. In contrast, much different results were found when tactile identification performance was based on the "new materials" condition. A repeated measures MANCOVA of the average percentage corrected performance scores of AD patients and controls on the tactile recognition and the tactile identification test of the "new materials" condition detected a significant main effect for subject group [E(1,53) = 14.44, R < .001, MSe = 9878.21 and a Box-adjusted effect of test type [E(1,53) = 80.68, R < .001, MSe = 429.34. A significant interaction between group and test type [E(1,53) = 35.52, p < .001, MSe = 18851.02 was also found. When the group by test interaction was analyzed with main effect tests significant differences between the controls' and the AD patients' average percentage corrected performance on the tactile recognition [E(1,53) = 77.55, p < .001, MSe = 30880.21]  144 and the tactile identification test were found [i.e., E(1,54) = 9.36, p < .003, MSe = 8470.10]. Eta-squared analyses indicated that the effect of test type on group performance had been much larger for the tactile recognition (.59) than on the tactile identification test, but that the effect of the "new materials" condition had been much greater than for the "old materials" condition (i.e., .15 vs. .007, respectively). Mildly, Moderately, and Severely Impaired AD Patients. As Figure 5 shows, the performance of mildly, moderately, and severely F.I. patients on the tactile identification test was very different to that found for the previous implicit tests. For example, the new material condition of the tactile  identification test was the only case where mildly F.I. patients outperformed moderately F.I. patients, who in turn outperformed the severely F.I. group. Two main trends were found when comparing the performance of these patients on the new and old material conditions. The first was that the performance of mildly F.I. patients did not differ across the new and old material conditions. Thus, changing whether the same or different objects were to be identified did not affect mildly F.I. patients' performance. This was not the case for moderately and severely F.I. patients whose average percentage corrected performance was much higher in the old than in the new materials condition. In fact, severely F.I. patients  145 showed a negative performance in the new materials condition suggesting that this task was more difficult for them to perform than the old materials condition. Closer inspection reveals that the mildly F.I. patients showed a higher average percentage corrected performance than did controls in both the old and new material conditions. This counter-intuitive result occurred because AD patients took longer to identify items than did controls and so there was a greater margin for improvement in their performance. Similar results have been reported in other studies that have used timed tasks with AD patients and controls (cf. Eslinger & Damasio, 1986).  Study Two Memory Strategy Experiments The focus of this part of the investigation was to examine the effects that different encoding strategies (or study tasks) and different retrieval cues (or test types) had on the explicit memory performance of AD patients. This was addressed in three different experiments where different study and test types were used. Strict scoring criteria were applied to all memory strategy experiments. For example, in the case of single word targets, changes in tense and plural forms were acceptable, but not substitute words with similar meaning. In the case of simple sentence commands, subjects' responses had to contain the correct  146 action verb and target nouns to be considered correct.  Levels of Processing Experiment This experiment was designed to see if AD patients would remember more target words that they had studied on the basis of meaning than for the number of letters the word contained. Similarly, this experiment was designed to see if subjects would recollect more targets in a recognition, rather than a free recall test condition.  Critical Measures. The critical scores for the encode condition were the number of target words recollected from the six words studied for their meaning and the six words studied for the number of letters they contained. Previous work had shown that using more than 12 stimuli would not result in a levels effect (i.e., more meaning than letter encoded targets being recollected) in memory-impaired patients (Cermak & Reale, 1978). The critical scores for the retrieval condition were the number of words freely recalled and recognized from the meaning and letter encode conditions. Subjects were engaged in the free recall of targets before a recognition phase for the remaining targets was begun. This meant that the total number of targets available to be recognized was dependent on the number that had not been freely recalled, and so a conditional or average percentage corrected score was computed. This score was obtained by subtracting the number of  147 targets that were recognized by the number of false alarms, dividing by the number of remaining targets, and multiplying by 100. The average percentage corrected free recall score was obtained by dividing the number of target words that were freely recalled by the number available for recall, and multiplying by 100.  AD Patients and Control Subjects. Figure 6 shows the average percentage corrected scores of the AD patients and the control subjects on the free recall and recognition of items in the levels of processing experiment. As this figure shows, AD patients freely recalled, on average, 15.8% more targets from the meaning than from the letter study task condition (M = 27.5% versus X = 11.67%, respectively). Similar results can be observed in the control subjects' performance. Controls freely recalled approximately 23.8% more targets from the meaning than letter study task condition (i.e., M = 64.67% vs. M = 40.83%, respectively)". Similar trends were found for the recognition test. AD patients recognized around 4.4% more targets from the meaning than letter study task (i.e., X = 66.7% vs. fl = 62.26%). Similarly, control subjects recognized approximately 11.35% more targets from the meaning than letter study task (i.e., X = 98.11%  IS Standard deviations are not reported in this section since these are within group comparisons. Refer to relevant tables for these values.  0 0  0 0  Encode Condition  rdprii 0  Meaning Letter  0  co  co  :0))  0  10  0 co  O  0  0  0  r71  0  Control^All^Mild^Mod.^Severe AD (n.40)^(n-20)^(n-6)^(n-7)^(n-7)  Free Recall  0  Control (n-40)  All^Mild^Mod.^Severe AD (n-20)^(n-6)^(n-7)^(n-7)  Recognition  Figure 6. Percentage scores of Alzheimer patients (AD: all stages) and Control subjects on tree recall and recognition of items in the levels of processing expenment. The error bars represent standard deviation scores.  149 vs. M = 86.76%, respectively). It must be noted that controls' performance was at ceiling in the meaning encode condition and so these results may not provide an accurate index of the magnitude of the performance differences between the meaning and letterencoded tasks. Figure 6 also shows that both AD patients and controls recognized more targets than they freely recalled. For example, AD patients had an average percentage corrected recognition score of 64.46% in comparison to 19.58% in the free recall test condition (see Table G-12a). Control subjects showed a similar effect with an average percentage corrected recognition score of 92.43% compared to 52.75% in the free recall condition. Finally, the results contained in Figure 6 indicated that, as expected, control subjects showed higher average percentage corrected scores in all study task and test conditions than did AD patients. To examine the hypothesis that AD patients and controls would recollect more target stimuli from the meaning than from the letter encode conditions, and from recognition than from the free recall test conditions, a three factor repeated measures MANCOVA (covariate factor = education) was performed. In this analysis the between-subjects factor was group (AD patients and controls) and the two within-subjects factors were study task (meaning, letter encode), and test type (recognition, free  150 recall). Results from this analysis indicated that there were failures to meet multivariate assumptions of homogeneity of variance (i.e., Box M test = E(10,6902) = 7.98, R < .001). When this result was further inspected with Bartlett-Box tests it appeared that these violations occurred on the recognition test factor (i.e., meaning recognition f(1,7249) = 70.74, 112 < .001; letter recognition E(1,7249) = 8.05, 2 < .005). To correct for these violations Box's procedure for adjusting the degrees of freedom for F  critical  was employed on factors involving test type  (Box, 1954). Significant main effects of group E(1,57) = 39.68, < .001, MSe = 42093.20, study task, f(1,58) = 39.72, R < .001,  MSe = 13546.87, and a Box adjusted test type factor, E(1,57) = 165.54, R < .001, MSe = 103840.83 were found. No interaction effects were significant (see Table G-12b). These results supported the hypothesis that AD patients would have a higher average percentage corrected score performance on the meaning than the letter study tasks and from recognition than free recall tests. A secondary and expected finding was that, although the meaning study task and recognition test condition resulted in a significant elevation of AD patients' explicit memory performance, their performance was never raised to levels that were statistically similar to controls.  151  Mildly, Moderately. and Severely Impaired AD Patients.  A  secondary question was whether AD patients in different levels of F.I. would also have a higher average percentage corrected performance in the meaning than from the letter encode condition, and in the recognition than from the free recall test condition. Figure 6 shows this expected profile of results. These results can be summarized in three ways. Irrespective of the level of F.I., all AD patients had higher average percentage corrected scores: (1) in the meaning than in the letter encode condition and (2) in the recognition than in the free recall test condition, and (3) there was a general trend for mildly F.I. patients to outperform moderately and severely F.I. patients. With the exception of the recognition test moderately F.I. patients showed a higher average percentage corrected performance than patients in severe levels of F.I.  SPT-EPT Experiment This experiment was designed to see if AD patients would remember more minitasks that they had performed than the experimenter had performed. These conditions are referred to as the subject-performed task (i.e., SPT) and experimenter-performed task conditions (i.e., EPT), respectively. Similarly, this experiment was designed to see if AD patients would recollect more targets under cued than free recall conditions.  152 Critical Measures. The critical scores for the encode condition were the number of minitasks recollected from the six SPT and the six EPT items. This stimulus number was chosen based on the results of a pilot study indicating the presence of floor and ceiling effects at smaller and larger numbers. The critical scores for the retrieval condition were the number of SPT and EPT minitasks that were recollected with cueing in the free recall test condition. Similar to the levels of processing experiment, subjects were first engaged in the free recall of targets before the cued recall phase began. Thus, similar procedures to those used in the previous experiment were employed to calculate average percentage corrected scores for the free and the cued recall test conditions. For example, the average percentage corrected free recall score was the number of SPT and of EPT minitasks freely recalled, divided by 6, and multiplied by 100. The average percentage corrected cued recall score was the number of the SPT and of the EPT minitasks recollected with cues, divided by the number of remaining targets and multiplied by 100. X51 Patients and Control Subjects. Figure 7 shows the average percentage corrected scores of AD patients and control subjects on the free and cued recall of targets in the SPT-EPT memory strategy experiment. As this figure indicates, AD patients freely recalled the same number of average percentage  ^  0  8  o-  Encode Condition  V o  Subject Performed (SPT) Experimenter Performed (EPT)  0 03  0  0  0  O  O  ^ All • Control AU^Mild^Mod.^Severe AD AD ^(n.40)^(n-20) (n•40)^(n-20)^(n.6)^(n.7)^(n.7) Control  ^  Free Recall  ^  Mild^Mod.^Severe (n.6)^(n•7)^(n.7)  Cued Recall  Flours 7. Percentage scores of Alzheimer patients (AD: alt stages) and Control subjects on tree recall and cued recall of items M the subject performed memory strategy experiment (SPT:EPT). The error bars represent standard deviation scores.  154 corrected SPT and EPT encoded minitasks (i.e., 12.5% and 12.5%). In contrast, these patients recollected around 9.52% more SPT than EPT corrected minitasks in the cued recall test condition (i.e., }1 = 56.19% vs. M = 46.67%, respectively). Two comments can be made about these results. The first is that AD patients' performance was elevated by the provision of additional retrieval clues from the free to the cued recall test condition for both the SPT and EPT encoded conditions. This conclusion was derived by comparing the average corrected performance scores on the SPT and the EPT tasks in free and cued recall conditions (i.e., 12.5% and 12.5% versus 56.19% and 46.67%, respectively). The second was that the cued recall test revealed differences in the AD patients' abilities to recollect SPT versus EPT encoded tasks (i.e., M = 56.19% vs. If = 46.67%, respectively). Control subjects consistently recollected more corrected targets from the SPT than the EPT study task condition, irrespective of the test condition. As Figure 7 shows, controls recollected around 9% more SPT than EPT encoded tasks in both free and cued recall test conditions (i.e., in free recall: 70.0% versus 61.67%; in cued recall: 79.44% versus 68.69%, respectively). Similar to AD patients, controls also benefited from the provision of additional clues, albeit, not to the same extent. Comparison of free and cued recall performance shows an  155 average elevation of around 8.0% in the average percentage corrected scores of the cued versus the free recall test conditions (i.e., 74.06% versus 65.83%, respectively; see Table G-13a). In contrast, AD patients' explicit memory performance was raised by approximately 39% in the cued versus the free recall test condition (i.e., 51.43% versus 12.5%). To examine the hypothesis that AD patients and controls would recollect more SPT than EPT encoded minitasks, and more critical targets in the cued than in the free recall test condition, a three factor repeated measures MANCOVA was performed. In this analysis the between-subjects factor was subjects (AD patients, controls) and the two within-subjects factors were study task (SPT, EPT encode), and test type (i.e., cued and free recall). Results from this analysis shows that multivariate and univariate assumptions of homogeneity of variance were met. A significant main effect for group, E(1,57) = 58.79, R < .001, MSe = 79227.64, study task, E(1,58) = 8.34, R < .005, MSe = 3325.02, and test type, E(1,58) = 70.80, p < .001, MSe = 38940.02 were found. Unlike the levels of processing  experiment, a significant group by test type interaction occurred, E(1,58) = 15.92, R < .001, MSe = 8755.21. When this group by test type interaction was graphed it showed that AD patients' free recall of the same percentage of SPT and EPT encoded targets accounted this effect. This  156 conclusion was confirmed by the results of main effect tests that showed that AD patients and controls recollected significantly more SPT than EPT encoded targets in the cued recall test condition; f(1,19) = 63.15, p < .001, MSe = 31733.89 and E(1,39) = 14.09, p < .001, Hag = 8075.07, respectively. In contrast, main effect tests showed that only controls recollected more SPT than EPT average percentage corrected targets in the free recall test condition fE(1,39) = 9.01, p < .005, Nag = 4375.07). Main effect tests also revealed that AD patients and controls each recollected more average percentage corrected targets in the cued than in the free recall condition. For example, AD patients and controls recollected more SPTs in cued, rather than free recall conditions; E(1,19) = 49.19, p < .001, 14Se = 20400.28 and E(I,39) = 12.47, p < .001, Mae = 5335.56,  respectively. AD patients also recollected significantly more EPTs in cued versus free recall conditions, E(I,19) = 22.09, p < .001, MSe = 11902.50. This was not the case for controls and is due to the fact that there was a smaller proportion of items that were left to be recollected in the cued recall condition (see Table G-13b). To summarize, these results show that AD patients recollected significantly more targets in cued than free recall test conditions. These results also show that although AD patients recollected significantly more SPTs than EPTs in the  157 cued recall test condition, this was not the case for the free recall test. Thus, the hypothesis that AD patients would remember more SPTs than EPTs and more items in the cued than in the free recall test conditions was only partially supported. This hypothesis was also only partially supported for controls.  Mildly. Moderately. and Severely Impaired AD Patients. A secondary question was whether AD patients in different levels of F.I. would show similar trends to recollect more targets from SPT than EPT, and in the cued versus the free recall test condition. Figure 7 shows that patients in all stages of F.I. recollected more corrected targets in the cued than in the free recall test condition. However, the SPT encode condition produced selective effects. Only mildly and moderately F.I. patients recollected more SPT than EPT encoded targets in the cued recall test condition. Since moderately and severely F.I. patients' performance in the free recall test was at floor little significance can be placed in the fact that SPT performance was slightly higher than EPT encoded targets in this test. (Refer to Table G-13a for a complete listing of these results).  Multisensory Experiment This experiment was designed to see if AD patients and control subjects would remember more sentence targets that they read aloud to the experimenter and then performed (the See, Say, &  Dca condition)  than they read aloud to the experimenter (the See  158 & Say condition) or read silently to themselves (the See only condition). Similarly, this experiment was designed to see if subjects would recollect more targets in the cued than in the free recall test condition. Critical Measures. The critical score was the number of sentences recollected from (1) DQ, (2) Say, and (3) See conditions. There were four target sentences per encode task condition. As in the previous experiment, subjects were asked to freely recall critical targets before the cued recall test began. Average percentage corrected scores were computed using the same method described in earlier sections with the exception that there were four targets per study task condition. AD Patients and Control Subjects. Figure 8 shows the average percentage corrected scores of AD patients and control subjects on the free and the cued recall tests of Do, Say, and See encoded targets. As Figure 8 shows, AD patients recollected more Do encoded, than Say or See encoded targets, in both free and cued recall conditions. For example, AD patients freely recalled 16.25% Do, 6.25% Say, and 5% See encoded targets compared to 44.77% Do, 12% Say, and 11.84% See encoded targets in the cued recall test condition. As these results show, there was little difference in AD patients' recollection of Say and See encoded targets, irrespective of the test condition. Comparison  8  Encode Condition See. Say and Do See and Say See  2  0 3  3  0  ti  0  K  0  0  Control^All^Mild^Mod. AD (n.40)^(n-20) ^(n-6)^(n.7)  Free Recall  Severe (ri." 7)  Control  All^Mild^Mod.^Severe AD (n-40)^(n-20)^(n.6)^(n.7)^(n-7)  Cued Recall  Figure 8. Percentage corrected score of Alzheimer patients (AD: all stages) and Control subtects on free recall and cued recall of items n the multIsensory strategy experiment. The error bars represent standard deviation scores  160 of AD patients' cued versus free recall test performance shows that cueing elevated the recollection of Do, but not of Say or See encoded targets. AD patients recollected 44.77% of Do encoded targets in the cued recall and 16.25% in the free recall test condition -- a difference of around 28.5%. In contrast, AD patients' performance on Say and See encoded targets was at floor in the free recall test condition. That is, performance was 6.25% and 5%, respectively and only slightly raised in the cued recall test condition (i.e., around 12% for both conditions). Similar trends to those described for AD patients were found in the controls' performance in this experiment, with two exceptions. Controls' performance was never at floor and there was a similar performance in the Do, Say, and See encoded conditions in both the free and cued recall test conditions. Controls did not show the same elevation in performance as did AD patients in the cued recall test condition because their high performance in the free recall condition meant that there were fewer remaining items to be recollected in the cued recall test condition. For example, controls recollected 72.5% versus 68% of the Do encoded targets in the free versus cued recall test conditions, 42% versus 36% in say encoded targets, and 34.5% versus 41% in see encoded conditions (see Table G-14a). To examine the hypothesis that subjects would recollect more (a) Do, than Say or See encoded targets, and (b) more targets in  161 cued rather than free recall test conditions, a three factor MANCOVA was performed. In this analysis the between-subjects factor was subjects (AD patients and controls) and the two within-subjects factors were study task with three levels (i.e., Do, Say, and See), and test type with two levels (i.e., free and cued recall). This analysis indicated that there were failures to meet assumptions of homogeneity of variance and Bartlett-Box tests showed that these violations occurred on the See free recall task factor (i.e., E(1,7249) = 11.67, R < .001). To correct for these violations Box's procedure for adjusting for degrees of freedom for F criticia was employed on factors involving the study task and test factor (Box, 1954; Howell, 1987). A significant main effect of group, E(1,57) = 47.26, p < .001, Hag = 91084.47, study task, E(1,115) = 28.46, p < .001, MSe = 24390.05, and test type E(1,57) = 9.41, p < .003, MSe = 8336.81. No interaction effects were significant. Since the performance of controls and AD patients were similar for the See and Say encode conditions post hoc tests did not need to be performed to determine where the significant differences in study task performance had occurred. Visual inspection of the average percentage corrected scores showed that more Do than Say or See encoded targets were recollected in both free and cued recall test conditions. Thus, it appeared that requesting subjects to see, say and perform simple sentences  162 produced a significant elevation in explicit test performance than if subjects read the sentence to themselves or silently to the experimenter. The hypothesis in this experiment was only partially supported in that subjects did not recollect significantly more items in the See than Say conditions.  Mildly, Moderately, and Severely Impaired Patients.  A  secondary question was whether AD patients in different levels of F.I. would recollect more corrected targets from Do, than Say or See conditions and from cued than free recall tests. Figure 8 shows that there was a consistent trend for all patients to recollect more Do than Say or See encoded targets. Whether more Say than See encoded corrected targets were recollected was dependent on the patients' level of F.I. For example, whereas mildly F.I. patients recollected more say than see encoded average percentage corrected targets in both cued and free recall test conditions, this was not the case for moderately and severely F.I. patients. Both moderately and severely F.I. patients showed Do, Say, and See performance that was at floor in the free recall test. In the cued recall test condition the moderate F.I. group's performance in the Say condition was raised from floor while the performance of the severe F.I. group remained at floor. Thus, as in previous experiments there was evidence of a slide-rule type influence in the effects that  163 encoding and test conditions had on the memory performance of patients in different levels of F.I. (Refer to Table G-14a).  164  CHAPTER 7: DISCUSSION This chapter has six sections. The first reviews the general research objectives and motivation for this investigation. The second section reviews the hypothesis and main findings from part one, how they relate to previous work, and the new things they may be telling us about implicit memory in AD patients. The third section contains the same information as it pertains to part two of the investigation. In the next section I consider some of the factors that limit the generalizability of the results from parts one and two. The fifth section considers future research, and the last section summarizes the general and specific contributions of this work. Research Objectives and Motivation The main goal of this investigation was to re-explore the implicit memory abilities and the effects of different memory strategies on the explicit memory performance of AD patients. To address the first part of this goal I examined AD patients' and controls' memory for different types of materials with implicit and explicit tests that differed in the instructions provided at retrieval. The second part of this goal was addressed with three memory strategy experiments that had different encoding and retrieval conditions.  165 The first part of this investigation was motivated by mixed reports regarding whether implicit memory is spared or impaired in AD patients. Variations in these findings may have been caused by using tasks that differed in their ability to guide these attentionally-impaired patients during the encoding of critical targets. The motivation behind the second part of this investigation was similar to that of the first. A limited number of studies have shown that the explicit memory performance of AD patients can be elevated in some, but not in other conditions. The main premise of the second study was that strategies that reduced the attentional demands of tasks at study and at retrieval would elevate the explicit memory performance of AD patients.  Part One The overall hypothesis of this study was that differences in the memory performance of AD patients and controls would be smaller on implicit, than explicit tests, using written and spoken word, picture, and object materials. This hypothesis was guided by two main ideas: First, that the neuropathology of AD begins in hippocampal regions before advancing to brain areas which have been ascribed roles in implicit test performance. Thus, a greater magnitude of impairment on explicit than on implicit tests should be observed in most AD patients. Second, that when an encoding method that compensated for AD patients'  166 attentional impairments was used, their implicit test performance would be raised to levels that were similar to that of controls. Results from part one indicated that the explicit test performance of AD patients was significantly lower than that of controls on all tests, irrespective of the materials that were used. This finding was expected since the hippocampus, which plays a role in the explicit recollection of multimodal information appears to be damaged early in AD (Ball, 1977; Barr & Kiernan, 1988). In contrast, whether AD patients' implicit test performance was found to be impaired depended on the materials that were used.  Implicit Test Results AD patients' performance on category completion tests appeared not to be significantly different from that of controls. This test required that subjects provide exemplars to categories, and thus tapped a semantic type of knowledge similar to the word association test used by Salmon et al. (1988). However, in contrast to the present study, Salmon et al. did not find that priming was spared in their AD patients. Several reasons can be offered for the different findings of the present experiment. One is that semantic priming was found to be spared in AD patients because an encoding method that required that patients identify and generate the meaning of critical targets (instead of a pleasantness rating as in the  167 Salmon study) was used. The premise here is that when a target is not completely encoded its subsequent recollection is impeded (Graf, Tuokko, & Gallie, 1990). In the case of AD patients the premise was that their inability to spontaneously encode targets (cf. Rohling, Ellis, & Scogin, 1991) would impede their performance, making them appear to be more impaired on implicit tests than they really were. However, when an encoding method that guided AD patients during encoding was used, their memory performance would be elevated. There are at least two reasons why the meaning encode condition could not be the sole factor responsible for finding that priming for written and spoken words was spared in AD patients. One reason is that this explanation would imply that cortical association areas are not required since patients with damage to this brain area (Brun, 1983) could perform these implicit tests. Since Positron Emission Tomography investigations have shown that cortical association areas are activated when controls perform wordstem completion tests (cf. Squire et al., 1992b) this makes it unlikely that requiring AD patients to encode targets for their meaning is the sole reason for this study's findings. Second, AD patients' performance on the implicit tests used in this study was not always found to be similar to that of controls.  168 An alternative reason that accounts for encoding method, neuroanatomical substrates, and the fact that priming varied with the test that was used is the following. Although implicit tests may index similar types of memory (i.e., semantic as in the case of category completion and word association), different cognitive processes may be required to perform these tests. Whether priming is found to be spared or impaired in AD patients may be influenced by the particular combination of cognitive processes required to perform the specific test, as well as on the type of materials that are used (i.e., pictures, words). Evidence that the word association test employed by Salmon et al. and the category completion tests used in this investigation may have tapped different cognitive processes comes from a consideration of baseline performance. Salmon et al. (1988) found that the baseline performance of their AD patients and controls was not statistically different. That is, both AD patients and controls were able to provide a similar number of words to match previously unstudied word pairs. In contrast, this was not the case for the category completion tests. AD patients provided significantly fewer exemplars to unstudied categories than did controls, for both written and spoken word conditions. In combination, the above findings suggest that the task requirements of the word association test may have been easier  169 for the AD patients to achieve than those of the category completion test. Thus, while both of these tests tap a semantic type of word knowledge, they may have done this via different cognitive processes. Further support for the idea that AD patients' performance on implicit tests may be influenced by factors other than the stimulus materials that are used (i.e., pictures or words) comes from a consideration of subjects' performance on the other tests in this investigation. Comparison of the performance of AD patients and that of controls on the picture fragment completion test suggests that picture completeness, rather than picture materials per se, influenced this study's findings. Consideration of the baseline performance of AD patients and controls suggest that whereas both groups were able to correctly identify a similar number of completed picture stimuli, this was not the case when incompleted or fragmented pictures were used. AD patients required significantly more information (i.e., stimuli that were in a more completed form) than controls did to identify picture stimuli. Other investigators have also found that AD patients experience a difficulty in processing fragmented pictures. For instance, Kirshner, Webb, and Kelly (1984) found that AD patients  were able to identify completed, but not fragmented versions of the same items. Furthermore, AD patients may be unique in their inability to identify fragmented pictures. As evidence, Bondi  170 and Kaszniak (1991) examined the average fragment level in which AD, PD, and control subjects were able to identify a series of 540 pictures. They found that AD patients required a significantly more completed picture than PD patients and controls did to identify the same targets. Closer inspection of the stimuli contained in Appendix D reveals a likely reason for these findings -- fragmented pictures are more abstract and visuospatial in form than the completed version of the same picture. Since problems in visuospatial and abstract reasoning characterize AD patients (Reisberg, 1983) the fragmented condition of these pictures rather than "an inability to activate pre-existing picture representations" (Heindel et al., 1990) may be the reason that implicit memory for pictures was found to be impaired in the AD patients in this, and in other investigations (cf. Bondi & Kaszniak, 1991; Heindel et al., 1990). A recent study supports this conclusion. When completed pictures were employed the implicit test performance of AD patients was not found to be different from that of controls (Gabrieli, unpublished). Additional evidence that stimulus characteristics can influence implicit test performance comes from the tactile identification test. When performance on this test was based on the re-identification of the same objects (i.e., old materials  171 condition) AD patients showed a magnitude of priming that did not appear significantly different from controls. In contrast, when the identification of different objects was required (i.e., the new materials condition) the same AD patients' performance was now found to be impaired. Since the same encoding method, test, and type of stimuli were present in both test conditions, the reason(s) for these performance differences is not clear. One possibility is that the re-identification of the same items made the old materials condition an easier task than the new materials condition; this idea is supported by the moderately and severely F.I. patients' performance. Both the moderately and the severely F.I. patients showed better performance in the old versus the new materials condition. One interpretation of these findings is that the cognitive processes required to perform the old materials condition was more attainable by these patients than those for the new materials condition. In contrast, there was no difference in the mildly F.I. patients' performance in either the new or the old materials condition, presumably because the cognitive processes required to perform both of these tasks did not exceed these patients' abilities".  " The assumption made here is that the neuropathology of AD is less advanced in mildly, than in moderately or severely F.I. patients. As a result of this more limited pathology, the mildly F.I. patients were able to meet the cognitive processing demands of both versions of the tactile identification test.  172 Neuroanatomical Considerations What implications do the results of this study have for current ideas regarding the neuroanatomical areas responsible for priming? Any attempt to address this question is speculative given the absence of pathological data regarding the patients in this investigation. However, if one is willing to accept that a consistent pattern of neuropathology is caused by AD (cf. Lewis et al., 1987; Price et al., 1991) some reflection on this question is possible. Several sources have provided evidence that cortical association areas may be integral to performing implicit tests of lexical, semantic, and picture materials (cf. Butters et al., 1990; Heindel et al., 1990; Squire et al., 1992b; Tulving & Schacter, 1990). In the case of AD, cortical association areas are eventually destroyed (Brun, 1983). Thus, to find that AD patients' performance on category completion tests was of a magnitude that was not statistically different from that of controls seems incongruous, until one considers the pattern of damage that occurs to cortical association areas. That is, the primary and secondary cortical association areas of AD patients appear to incur less pathology than the tertiary cortical association areas (cf. Esiri et al., 1986; Lewis et al., 1987; Van Hoesen & Damasio, 1987). Thus, the possibility exists that when the appropriate level of cognitive demands is made (i.e., by  173 the combination of test type, encoding method, and stimuli that are used) the primary and secondary cortical association areas of some AD patients may be able to support a magnitude of priming that is similar to that achievable by controls (L.R. Squire, personal communication, April, 1993). It is also possible to be more specific regarding the influence that encoding method and stimuli characteristics may have had on neural substrates. With respect to encoding method, studies have shown that maximal activation of the nervous system occurs when subjects generate the meaning, rather than provide a pleasantness or rhymed response to stimuli (Cohen & Waters, 1985). Thus, when AD patients generated the meaning of critical targets, this presumably elevated the level of neural activity in brain areas including those responsible for attentional focusing (i.e., locus coeruleus, nucleus basalis of Meynert, etc.). Under "normal"" circumstances these brain areas may not have been activated to a level capable of supporting the mental processes required to perform various implicit tests. Support for this proposal comes from the Partridge et al. (1990) study which found intact priming in AD patients when a meaning encode method was used, in contrast to other investigators using the same test and a pleasantness encode task (cf. Shimamura et al., 1987; Salmon et  " Where "normal" refers to the absence of specific encoding instructions.  174 al., 1988; Heindel et al., 1990). Whereas a meaning encoding method may have served to change the pattern of neural activity in AD patients, the same may be true of the stimuli that was used. A completed picture may have made less processing demands on the AD brain than the same picture in its fragmented form. While the mechanisms by which this difference in cognitive processes may have been mediated has yet to be determined, one possibility is the number of brain areas that were involved. That is, intact pictures may primarily tap the neural processing of areas such as the parietal, occipital, and temporal lobes whereas fragmented pictures might, in addition, make heavy demands on areas such as the frontal lobes (Lezak, 1983). In the case of the AD brain where neuropathology occurs in all these areas, the mental processes required to identify the fragmented (in contrast to the completed) picture stimuli may have surpassed the AD patients' neuronal processing abilities. Additionally, the progressive loss of neurons that occurs in AD patients' brains would presumably correspond with a decline in neural processing capabilities. This event presumably accounts for the fact that patients in mild levels of F.I. (i.e., in early stages of neuronal loss) outperformed those patients in later stages of F.I.  175 Alternatively, principles of Transfer Appropriate Processing must also be considered. Specifically, the stimulus form that was presented at study was not identical to that presented at retrieval in the picture fragment completion and in the new materials condition of the tactile identification test. This suggests that test performance might have been lowered on these tests since, under the theoretical framework of TAP, an exact recapitulation of the cognitive operations engaged at study and at test would not have occurred. This event may have been translated into more neural work for the AD brain in contrast to if the same stimulus form had been employed at study and at test. Positron Emission Tomography that have found that less neural activity is required to re-process identical (vs. non-identical) stimuli supports this idea (cf. Squire, 1992a; Squire, et al., 1992b). To review, the findings from this study have been interpreted from the view that implicit tests can vary in the cognitive processing (and thus the neural processing) demands they place on the AD patient. While this perspective is able to account for the discrepant findings regarding if implicit memory is retained in AD patients (cf. Keane et al., 1991; Partridge et al., 1990) additional work on this issue is required. Specifically, a systematic investigation of the influence that encoding and retrieval methods, as well as stimulus format has on  176 the test performance and brain activity of AD patients is required to determine the validity of my interpretations since other explanations are possible. Additional Possibilities There are of course additional factors that could explain the findings obtained in this study. One is that the AD patients in this investigation were somehow different from those that have been used in other studies. For example, there may have been a different pattern and/or a less advanced degree of neuropathology present in this study's patients that could account for the present study's finding of spared priming in AD patients. This question of patient similarity pervades all AD investigations (L. R. Squire, personal communication, April, 1993) and is not an easy issue to address. However, while information on the neuropathology of patients is not available to most researchers an indirect index via level of functional impairment is obtainable. The inclusion of mildly, moderately, and severely F.I. patients suggests that the average level of impairment in the patients used in this study may have been greater than in other investigations where mildly and moderately impaired patients have been used. As a result, had the findings from this study been solely attributable to a unique patient sample, then priming should have been found to be impaired rather than spared.  177 Additional explanations for this study's findings also exist. For example, the implicit tests used in this study may have not been sensitive to differences in control and patient performance and this could account for finding preserved priming in AD patients. In the case of the picture fragment completion test, AD patients' priming performance was at floor and this makes it impossible to make conclusive statements regarding this test result. Further, as is the case with any timed measure of priming, the large variability in subject performance on the tactile identification test may have masked the true magnitude of the differences between the controls and the AD patients. Thus, the findings from this study will need to be replicated before it will be possible to determine whether the semantic network. of AD patients is lost, or whether impaired priming reflects an inability to enter and access this network due to memory and nonmemory related factors (i.e., attention). Part Two The overall hypothesis of this study was guided by the idea that the attentional deficits encountered by AD patients can influence their explicit memory performance. Thus, it was proposed that strategies that changed the way that targets were encoded and retrieved (i.e., by presumably lowering the attentional processing demands that were required) should result in an elevation of the explicit memory performance of AD  178 patients. The above hypothesis was tested in three different experiments. The first experiment used a levels of processing framework where target words were either studied for their meaning or for the number of letters they contained (i.e., the non-semantic condition). Existing theories suggest that encoding targets for their meaning (vs. non-meaning) might serve to lower the attentional processing demands required to encode targets (Cohen, Sandler, & Schroeder, 1987). The second experiment employed a Subject Performed-Experimenter Performed (i.e., SPTEPT) strategy where mini-tasks were either performed by the subject or by the experimenter. Existing theories suggest that performing a task (in contrast to watching it being performed) engages multimodal processing that serves to lower the  attentional demands of the task (Backman & Nilsson, 1985; Cohen, 1985; Engelkamp & Zimmer, 1985). The third or muitisensory experiment was new. More multimodal processing was assumed to occur when subjects read aloud and performed simple sentences (i.e., the DO condition) than when they either read aloud (i.e., the Say condition) or silently read a target (i.e., the See condition). Thus, it was presumed that AD patients would recollect more DO than Say or See encoded targets since the DO condition should lower the attentional demands at encoding more than the latter conditions did.  179 For the levels of processing experiment a recognition and a free recall test served as the retrieval conditions. For the SPT-EPT and multisensory experiments a cued recall and a free recall test served as the two retrieval or test strategy conditions. Previous work suggested that recognition and cued recall conditions provided more specific clues to re-locating targets than did a free recall condition (Perlmutter, 1978). In essence, the recognition and cued recall conditions could be viewed as lowering the attentional demands of the memory task more than did a free recall test. Thus, it was assumed that AD patients would retrieve more targets in the recognition and cued recall conditions than when free recall test strategies were used.  Levels of processing experiment. Results from this part of study two extended previous research by showing that AD patients could recollect a larger number of targets that they had encoded on the basis of their meaning (vs. non-meaning) and in a recognition (vs. free recall) test condition. The overall memory performance of AD patients was also found to be highest when a combination of meaning encode and recognition conditions had been employed. Previous investigators have not been able to find evidence that AD patients would recollect more targets that had been studied for their meaning (i.e., a levels effect; cf. Corkin,  180 1982) or to find a significant change in performance when different tests were used (i.e., in a free versus cued recall test condition; cf. Martin et al., 1985). Several factors could account for the reason this study found that the explicit memory performance of AD patients was significantly elevated when a meaning encode and a recognition strategy was employed. Variations in the encoding method used in this study and that used by Corkin (1982) might account for some of the differences found in the present experiment. That is, it is likely that Corkin was unable to find a levels effect (i.e., the recollection of more meaning than non-meaning encoded targets) in her AD patients because, in effect, there was no levels condition in her study. To illustrate, the encoding method used by Corkin required that subjects provide a yes/no response to questions such as "Is the word a type of fruit ?" A similar situation to that previously discussed for implicit tests is likely to have been present. That is, since AD patients are unable to engage in the spontaneous elaboration of targets (cf. Rohling et al., 1991) requiring them to simply provide a yes/no response to a sentences such as "Is the word a type of fruit ?" would not have ensured that they had thought about the target's meaning. That is, patients might have responded yes or no without contemplating the targets' meaning. In contrast, in the present study subjects were required to generate the meaning of critical targets which  181 confirmed that AD patients had thought about the meaning of targets. In comparison, Martin et al. (1985) who used a similar encoding method to that used in the present experiment and also found a levels effect, did not find a significant test effect. That is, the explicit memory performance of AD patients did not change when a free and a cued recall test condition was used. This differs from what was found in this study where AD patients recollected significantly more targets in a recognition than in a free recall test condition. A comparison of the tests that were used in the Martin et al. versus the present study provides one possibility for the discrepant findings. Martin et al. required that AD patients engage in the free recall of items before they were cued for the correct response. In this experiment AD patients were engaged in the free recall of items before they were asked to recognize the target presented at study. The recognition condition would have provided different "clues" about the stimulus than would a cued recall test. To illustrate, presenting the stimulus that was shown at study would make the process of re-locating a mental representation of the event easier simply because a matching of the study to test stimulus would be taking place. In contrast, a cue such as "What type of fruit were you shown ?" would not be a clue that presumably matched the mental representation of the study target.  182 This latter condition would therefore theoretically require more cognitive effort or attentional processing, and thus be a more difficult task for a cognitively impaired subject to perform. In the case of a patient experiencing serious retrieval difficulties, the clues provided by a recognition versus a cued recall test might make the difference in whether the target was successfully recollected. SPT EPT experiment. ResUlts from this experiment indicated that AD patients recollected more minitasks that they performed (i.e., SPTs) than the experimenter had performed (i.e., EPTs) in a cued recall test condition. However, AD patients freely recalled the same number of SPTs and EPTs. Closer inspection of the results presented two reasons for this latter finding. First, although both the moderately and the severely F.I. patients freely recalled more SPT than EPT encoded targets, their performance was at floor. These floor effects may have masked the magnitude of the SPT encoding effect. Second, inspection of the mildly F.I. patients' free recall performance shows that these patients recalled more EPTs than SPTs. While the reason for this latter finding could be attributable to many factors, including low subject numbers, it is possible that watching the experimenter perform the tasks was a somehow unique or unusual event. Since the mildly F.I. patients were presumably less cognitively impaired than the moderately and severely F.I.  183 patients the "uniqueness" of the EPT tasks may have, in effect, produced a more memorable event that had significance only for the mildly F.I. group. The present experiment confirmed the findings reported in the only other investigation of this type. In this study Dick et al. (1989) also found that there was no difference in the number of SPTs and EPTs that AD patients recollected in a free recall condition. Thus, it appears that any memory benefits that AD patients might obtain by performing a task may only be realized if they are also provided with recognition or cued recall tests. That is, when compared to the presumed attentional processing lowering capabilities of a meaning encode condition, the SPT encode strategy may not be as effective. However, when the SPT and the cued recall conditions were combined, the explicit memory performance of some AD patients (i.e., those in mild and moderate stages of F.I.) was maximally raised. This finding appears to correspond with previous research in which we have found that impairments in retrieval are present in all AD patients, but problems at encoding may only typify those in later stages of F.I. (Tuokko, Gallie, & Crockett, 1990). Thus, in order to consistently raise the explicit memory performance of any AD patient a retrieval strategy may be required. However, the combined use of an effective encoding and retrieval strategy should produce the greatest elevation in the explicit memory  184 performance of this patient group.  Multisensory experiment. The results obtained from this experiment indicated that the AD patients in this study recollected more targets that they had encoded in the DO than in the Say or See condition. Also evident was the fact that the AD patients' overall performance in the Say and See encode conditions was not that different. Thus, seeing, saying, and performing the target seems to have enhanced explicit memory performance more than if AD patients simply read the target. As in the previous experiment, when retrieval in a cued rather than a free recall format was required, AD patients' explicit memory performance was maximally elevated.  Implications The results from the three experiments contained in study two have provided some consistent findings. First, they have shown that certain encoding and retrieval strategies can significantly elevate AD patients' explicit memory performance. That is, cued recall and recognition tests appear to be more effective in elevating explicit memory performance than free recall conditions. Furthermore, the meaning and motor performance encoding strategies may be more effective than nonmeaning and non-motor performance events. While the exact processes by which the above encoding and retrieval strategies work is not known, one proposal has been  185 that they work by lowering the attentional or cognitive processing requirements of the task (Backman & Nilsson, 1985; Cohen et al., 1987; Perlmutter, 1978). For the SPT and DO encode tasks,.the combination of multimodal processing and the fact that the brain areas associated with motor performance are relatively spared in AD patients may be some of the factors responsible for these memory-enhancing effects. In contrast, since the nonmeaning, See, Say, and free recall conditions were not as effective in elevating memory performance they presumably did not influence cognitive processing in the same way as the former tasks. A second finding that was consistent across the three experiments was that the combined use of specific encoding and retrieval strategies resulted in a maximal elevation of the explicit memory performance of AD patients. However, by themselves it was the cued and recognition retrieval strategies that were more effective than the encoding strategies in raising the explicit memory performance of AD patients. This finding appears to correspond to an issue that has been previously discussed. That is, retrieval deficits may occur early, whereas encoding difficulties may arise in later stages of AD (Tuokko, Gallie, & Crockett, 1990). Thus it makes sense that, used by themselves, the recognition and cued recall retrieval strategies will always be more effective in elevating the explicit memory of  186 AD patients than the sole use of a meaning or performance-based encoding strategy. While the exact reasons for this relationship is unknown it might correspond to the neuropathological development of AD. That is, since AD begins in the hippocampus (which is integral for the intentional retrieval of information) the sole use of either a recognition or cued recall retrieval strategy will always be more effective than either a meaning or performance-based encoding strategy. In turn, since the neuropathology of AD eventually progresses to include brain areas associated with various attention-related processes (Fedio et al., 1992; Mountcastle, 1978), the use of either a meaning or performance-based encoding strategy should also be effective in raising the explicit memory performance of some AD patients (especially in combination with certain retrieval tests). In addition, the fact that neuronal loss increases as AD progresses may be one reason that the benefits derived from these strategies declined as the level of patient F.I. changed from mild to severe stages. To summarize, together the results from studies one and two suggest that the memory abilities of AD patients may be better preserved than previous research has indicated. Continued investigation of the effects that different cognitive processes have on the memory performance of AD patients will be required to continue the quest for a more accurate brain - memory model of  187 this disease.  Limitations of this Work The environment one works in and the subjects one chooses to study provides every researcher with a unique spectrum of challenges. Working with AD patients is certainly no exception and many of the limitations present in this study are a direct reflection of the characteristics of this patient group. I will selectively discuss the limitations of this work in terms of three main areas: (1) the method that was used, (2) subject characteristics, and (3) statistical considerations.  Limitations Attributable to Method Two aspects of the method limit my findings: i) the number of stimuli that were used, and ii) the order of test and experiment presentation. With respect to stimuli, 12 targets were employed in each test and experiment and this limits the generalizability of the results. The main reason for employing 12 items was that pilot work had indicated that floor effects occurred when more targets were included. Multiple block stimulus presentations could have been employed to resolve this problem but wasn't for the following reasons. The focus of this investigation was exploratory and so I wanted to use a large number of tests and experiments with the AD patients. The large number of tests in combination with the high fatigability of the AD patients meant that they would not have completed the  188 investigation if several blocks of stimulus presentations had been used. Secondly, a standard order of test and experiment presentation was followed and thus, there is the possibility that order effects are present in the data. A randomized test order was not employed for reasons similar to that already discussed. Pilot work had indicated that AD patients could not complete the tests and experiments unless they were staggered in an alternating easy/difficult order. However, since most neuropsychological investigations include a standard order of test presentation the method used in this study is consistent with current research practice. Subject Characteristics All researchers encounter a problem in recruiting large numbers of AD patients for two main reasons. First, although many patients visit Alzheimer clinics only a small proportion receive a diagnosis of AD, and so there are few patients to select from. Second, once diagnosed, AD patients are reluctant to participate in research that requires that they engage in tasks that they will have difficulty performing. For these main reasons research in this area have been based on small numbers of AD patients. While there have been few studies that have employed more than 20 patients (the sample size employed in this study) this still remains a small number of subjects in terms of  189 the statistical power it might provide (cf. Cohen, 1988). Thus, the possibility is that there was insufficient power to accurately detect performance differences that might have occurred. However, since the size of the effect that was being indexed (i.e., memory performance) was large, it is unlikely that low subject numbers unduly limited the ability to detect performance differences when they existed. A second source of error is the possibility that some of the patients did not have Alzheimer's Disease. Since a definite diagnosis of AD is made at autopsy (cf. McKhann et al., 1984) and the patients in this study are still alive, it has not been possible to check the accuracy of their diagnosis. Thus, the possibility exists that the memory deficits observed in patients could be attributable to brain areas not associated with AD. Since all the patients included in this study were recruited from a clinic that has between an 85 to 90 percent diagnostic accuracy rate (B.L. Beattie, personal communication, April, 1993) it is unlikely that many of the patients in this study were not of the Alzheimer's type. Statistical Considerations There were floor and ceiling effects in some of the test and experimental conditions in this investigation. While this situation is common in AD research, it would have masked the true magnitude of the performance differences between patients and  190 controls. Since these performance differences were generally large, it is unlikely that any masking of the true magnitude of this difference would have led to different conclusions from those that have been reported. In addition, Bonferroni family-wise error rates were used to reduce the likelihood of committing a Type I error. In turn, the largest sample size that was reasonably possible, together with the inferential test that was most relevant for the data (i.e., repeated measures Multivariate Analysis of Covariance), was employed to reduce the possibility of making Type II errors.  Future Work There are three main directions that I would like to pursue in future work. The first, and perhaps the most obvious, is to directly investigate the assumption that I, and others have made that the encoding method used by AD patients has a large impact on their implicit test performance (cf. Graf, Tuokko, & Gallie, 1990; Partridge et al., 1990; Salmon & Heindel, 1992). For example, will AD patients show differences in performance when they are provided with no directions, requested to provide a pleasantness rating, or to identify and generate their own meaning for the critical targets used on various implicit memory tests? The second and third studies represent life long research projects. One project would be to administer the tests developed  191 for study one to HD, KS, and PD and other memory-impaired patient groups. Since many of the implicit tests that were employed in this study have not been used elsewhere, a study of this type would contribute to current efforts towards establishing unique test signatures for these patient types. The other project would be to continue to investigate, in a more direct manner, the relationship between cognitive or mental processes and the neural substrates on which they are orchestrated. This would be done by systematically varying specific encoding and retrieval conditions as well as stimulus formats and correlating subject performance on these tasks with brain activity as indexed by techniques such as Positron Emission Tomography. This general line of inquiry is already extending our current understanding of brain : behavior relationships. For example, Larry Squire and his colleagues (1992b) have shown that subjects activate different brain regions when they intentionally versus non-intentionally recollect events. In addition Ken Paller (1987, 1990) has found that different patterns of neural activity are correlated with the processing of abstract and concrete words.  Main Contributions of this Work This investigation makes both a specific and a general contribution to the field of neuropsychology. With respect to its specific contribution, it has shown that the implicit and the  192 explicit memory abilities of AD patients may be better retained than previous research has indicated. That is, by simply changing the way that targets were encoded, the clues that were provided at retrieval, as well as the stimulus materials that were used, the memory performance of AD patients were significantly changed. With respect to this study's general contribution, the findings from this investigation suggest that future research will need to deviate from the standard tests and methods that have been previously employed if we are to extend our current understanding of the brain - implicit memory relationships of patients with Alzheimer's disease.  193  REFERENCES Adams, R.A., & Victor, M. (1985). Principles of Neurology (3rd ed.). New York: McGraw Hill. .  Agid, Y., Ruberg, M., Dubois, B., & Pillon, B. (1987). Anatomoclinical and biochemical concepts of subcortical dementia. In S.M. Stahl, S.D. Iversen, and E.C. Goodman (Eds.), Cognitive Neurochemistry (pp. 248-271). Oxford: Oxford University Press. Albert, M., & Moss, M. (1984). The assessment of memory disorders in patients with Alzheimer's disease. In L. Squire & N. Butters (Eds.), The Neuropsycholoav of Memory (pp. 236-246). New York: Guilford Press. Albert, M., Moss, M., & Milberg, W. (1989). Memory testing to improve the differential diagnosis of Alzheimer's Disease. In R. Iqbal, H.M. Wisniewski, & B. Winblad (Eds.), Alzheimer's disease and related disorders (pp. 55-69). New York: Alan R. Liss. Alzheimer, A. (1907). Veber eine eigenartige Erkrankung der hirnrinde. Allg Z Psychiatr, 645, 146-148. Arnold, S.E., Hyman, B.T., Flory, J., Damasio, A.R., & Van Hoesen, G. (1991). The Topographical and Neuroanatomical Distribution of Neurofibrillary Tangles and Neuritic Plaques in the Cerebral Cortex of Patients with Alzheimer's Disease. Cerebral Cortex, 1, 103-116. Backman, L., & Nilsson, L.G. (1984). Aging effects in free recall: An exception to the rule. Human Learning, 2, 53-69. Backman, L., & Nilsson, L.G. (1985). Prerequisites for lack of age differences in memory performance. Experimental Aging Research, 11, 67-73. Backman, L., & Nilsson, L.G. (1989). Encoding of verbal and nonverbal features of action events under conditions of focused and divided attention. In A.F. Bennett & K.M. McConkey (Eds.), Cognition in Individual and Social Contexts (pp.295-304). North Holland: Elsevier Publishers. Ball, M.J. (1977). Neuronal loss, neurofibrillary tangles and granulovacuolar degeneration in the hippocampus with aging and dementia. Acta Neuropathology, 22, 111-118.  194 Ball, M.J., Hachinski, V., Fox, A., Kirshen, A.J., Fishman, M., Blume, W., Kral, V.A., & Fox, H. (1985). A new definition of Alzheimer's disease: A hippocampal dementia. Lancet, 1, 14-16. Barr, M.L., & Kiernan, J.A. (1988). The Human Nemvous System: An Anatomical Viewpoint (5th ed.). New York: J.B. Lippincott Company. Battig, W.F., & Montague, W.E. (1969). Category norms for verbal items in 56 categories: A replication and extension of the Connecticut category norms. Journal of Experimental Psychology, la, 1-46. Bayles, K.A., & Kaszniak, A.W. (1987). Communication and Cognition in Normal Aging and Dementia. Toronto: Little, Brown & Company. Bayles, K.A., Tomoeda, C.K., Kaszniak, A.W., & Trosset, M.W. (1991). Alzheimer's disease's effects on semantic memory: Loss of structure or impaired processing? Journal of Cognitive Neuroscience, 2, 166-182. Belichambers, M.G. (1990). An Investigation of Remote Memory and the Effectiveness of Cues on the Retrieval of Old Memories in Dementia of the Alzheimer Type. Unpublished Doctoral Dissertation, Ohio State University. Bellezza, F.S. (1988). Mnemonic devices and memory schemas. In M.A. McDaniel & M. Pressley (Eds.), Duagery and Related Mnemonic Processes: Theories, Individual Differences and Applications (pp. 34-55). New York: Springer-Verlag. Bird, M., & Luszcz, M. (1991). Encoding specificity, depth of processing and cued recall in Alzheimer's disease. Journal of Clinical and Experimental NeuroTsychology, 12, 508-520. Blass, J.P., Li-wen Ko, B.V.M., & Wisniewski, H.M. (1991). Pathology of Alzheimer's Disease. In M. Davidson (Ed.), The Psychiatric Clinics of North America (pp. 397-420). London: W.B. Saunders Company. Box, G.E.P. (1953). Non-normality and tests on variances. Diometrika, A2, 318-335.  195 Box, G.E.P. (1954). Some theorems on quadratic forms applied in the study of analysis of variance problems. Effect of inequality of variance in the one-way classification. Annals of Mathematical Statistics, 25, 290-302. Braak, H., & Braak, E. (1990). Morphology of the cerebral cortex in relation to Alzheimer's disease. In K. Maurer, P. Riederer, & H. Beckmann (Eds.), Alzheimer's Disease, Epidemiology. Neuropathology, Neurochemistry and Clinics (pp. 85-92). New York: Springer-Verlag. Braconnier, R.J., Cole, J.0., Spera, K.F., & De Vitt, D.R. (1982). Recall and recognition as diagnostic indices of malignant memory loss in senile dementia: A Bayesian analysis. Experimental Aging Research, 1, 189-193. Brun, A. (1983). An Overview of Light and Electron Microscopic Changes. In B. Reisberg (Ed.), Alzheimer's Disease: The Standard Refe ence (pp. 37-47). New York: The Free Press. Brun, A., & Englund, E. (1981). Regional pattern of degeneration in Alzheimer's Disease: neuronal loss and histopathological grading. Histopatholoay,^549-564. Butters, N., Heindel, W.C., & Salmon, D.P. (1990). Dissociation of implicit memory in dementia: Neurological implications. Bulletin of the Psychonomic Society, 21, 359-366. Camp, C.J., & McKitrick, L.A. (1992). Memory interventions in Alzheimer's type dementia populations: Methodological and theoretical issues. In R.L. West & J.D. Sinnot (Eds.), Everyday Memory and Aging: Current Research and Methodology (pp. 155-172). New York: Springer-Verlag. Canadian Pharmaceutical Association. (1992). Compendium of Pharmaceuticals and Specialties (27th ed.). Ottawa: Canadian Pharmaceutical Association. Carpenter, M.B. (1985). Core Textbook of Neuroanatomy. London: Williams and Wilkins. Cermak, L.S. & Reale, L. (1978). Depth of Processing and Retention of Words by Alcoholic Korsakoff Patients. Journal "^  165-174.  • •L •^r^  -  I^-^I^I •^- i •^• 1111  196 Claparede, E. (1951). Recognition and 'me-ness'. In D. Rapaport (Ed.), Organization and pathology of thought (pp. 58-75). New York: Columbia University Press. (Reprinted from Archives de Psychologie, 1911, 11, 79-90). Glower, R.P., Zola-Morgan, S., & Squire, L.R. (1990). Lesions of the perirhinal cortex, but not lesions of the amygdala, exacerbate memory impairment in monkeys following lesions of the hippocampal formation. Society for Neuroscience Abstracts, ^616. Cohen, J. (1988). Statistical power analysis 'or the behavioral sciences. San Diego, CA: Academic Press. Cohen, N.J., & Squire, L.R. (1980). Preserved Learning and retention of pattern analyzing skill in amnesia: Dissociation of knowing how and knowing that. Science, 210, 207-209. Cohen, N.J., & Squire, L.R. (1981). Retrograde amnesia and remote memory impairment. peuropsychologia, 119, 337-356. Cohen, A.A., & Waters, W.F. (1985). Psychophysiological correlates of levels and stages of cognitive processing. Neuropsychologia, 22, 243-256. Cohen, R.L. (1981). On the generality of some memory laws. Scandinavian Journal of Psychology, 22, 267-281. Cohen, R.L. (1983). The effect of encoding variables on the free recall of words and action events. Memory and Cognition, 11, 575-582. Cohen, R.L., & Bean, G. (1983). Memory in educable mentally retarded adults: Deficit in subject or experimenter? Intelligence, 2, 287-298. Cohen, R.L., & Bryant, S. (1991). The role of duration in memory and metamemory of enacted instructions. Psychological Research, al, 175-182. Cohen, R.L., Sandler, S.P., & Schroeder, K. (1987). Aging and Memory for Words and Action Events: Effects of Item Repetition and List Length. psychology and Aging, 2, 280285. Cohen, R.L., & Stewart, M. (1982). How to avoid developmental effects in free recall. Scandinavian Journal of Psychology, 22, 9-16.  197 Collins, A.M., & Loftus, E.F. (1975). A spreading activation theory of semantic processing. psychological Review, 12, 407-428. Corkin, S. (1982). Some relationships between global amnesias and the memory impairments in Alzheimer's Disease. In S. Corkin, K.L. Davis, J.H. Growdon, E. Usdin, & R.L. Wurtmen (Eds.), Aging. vol. 19. Alzheimer's Disease: A Report of Progress, (pp. 149-164). New York: Raven Press. Craik, F.I.M. (1978). Age differences in memory: The roles of attention and depth of processing. In L.W. Poon, J.L. Fozard, L.S. Cermak, D. Arenberg, & L.W. Thompson (Eds.), New Directions in Memory and Aging: Proceedings of the George A. Talland Memorial Conference (pp. 95-112). Hillsdale, New Jersey: Erlbaum. Craik, F.I.M. (1983). On the transfer of information from temporary to permanent memory. philosophical Transactions of the Royal Society of London, p302, 341-359. Craik, F.I.M., & Byrd, M. (1982). Aging and cognitive deficits: The role of attentional resources. In F.I.M. Craik & S. Trehub (Eds.), Aging and Cognitive Processes (pp. 191-208). New York: Plenum Press. Craik, F.I.M., & Lockhart, R.S. (1972). Levels of processing: A framework for memory research. Journal of Verbal Learning and Verbal Behavior, 11, 671-684. Craik, F.I.M., & Rabinowitz, J.C. (1984). Age differences in the acquisition and use of verbal information. In H. Houma & D.G. Bouwhuis (Eds.), Attention and performance (pp. 417499). New York: Erlbaum. Craik, F.I.M., & Tulving, E. (1975). Depth of processing and the retention of words in episodic memory. Journal of experimental Psycholoay: General, 104, 268-294. Damasio, A.R., Van Hoesen, G.W., & Hyman, B.T. (1990). Reflections on the selectivity of neuropathological changes in Alzheimer's disease. In M.P. Swartz (Ed.), Modular Deficits in Alzheimer Type Dementia (pp. 83-100). Cambridge, Massachusetts: MIT Press. Davies, P., & Maloney, A.J.F. (1976). Selective loss of central cholinergic neurons in Alzheimer's Disease. Lancet, 2, 1403.  198 Davis, H.P., & Bernstein, P.A. (1992). Age-Related Changes in Explicit and Implicit Memory. In L.R. Squire & N. Butters (Eds.), Neuropsychology of Memory, (pp. 249-261). New York: Guildford Press. deLeon, M.J., George, A.E., Stylopoulous, L.A., Miller, D.C., & Smith, G. (1990a). In vivo studies of hippocampal atrophy in Alzheimer's disease. In K. Maurer, P. Riederer, & H. Beckmann (Eds.), Alzheimer Disease, Epidemiology. Neuropathology. Neurochemistry. and Clinics (pp. 523-529). New York: Springer-Verlag. deLeon, M.J., George, A.E., Stylopoulous, L.A., Miller, D.C., & Smith, G. (1990b). Computed tomographic hippocampal atrophy: An early marker for Alzheimer's disease. In J.M. Orgogozo & C.G. Gottfries (Eds.), progress in Dementia Research (pp. 55-62). New York: The Parthenon Publishing Group. Dick, M.B., Kean, M.L., & Sands, D. (1989). Memory for action events in Alzheimer type dementia: Further evidence of an encoding failure. Brain and Cognition, 2, 71-87. Duke, L.W., Weathers, S.L., Caldwell, S.G., & Novack, T.A. (1990). Cognitive rehabilitation after head injury trauma: Toward an integrated cognitive/behavioral perspective on intervention. In C.J. Long, L.K. Ross, & M.G. Mutchnick (Eds.), Head Trauma: Acute Care to Recovery (pp. 1-30). New York: Plenum Press. Engelkamp, J., & Cohen, R.L. (1991). Current issues in memory of action events. Psychological Research, La, 175-182. Engelkamp, J., & Zimmer, H.D. (1985). Motor programs and their relation to semantic memory. ggrman Journal of Psychology, 2, 239-254. Esiri, M.M. (1989). Patterns of cortical and subcortical pathology in Alzheimer's disease. In D.C. Davies (Ed.), Alzheimer's Disease: Towards an Understanding of the Aetiology and Pathogenesis (pp. 33-41). London: John Libbey. Esiri, M.M., Pearson, R.C.A., & Powell, T.P.S. (1986). The cortex of the primary auditory area in Alzheimer's Disease. Brain Research, 366, 385-387.  199 Eslinger, P.J., & Damasio, A.R. (1986). Preserved motor learning in Alzheimer's disease: Implications for anatomy and behavior. Journal of Neuroscience, A, 3006-3009. Filoteo, J.V., Delis, D.C., Massman, P.J., Demadura, T., Butters, N., & Salmon, D.P. (1992). Directed and divided attention in Alzheimer's Disease: Impairment in shifting of attention to global and local stimuli. Journal of Clinical and Experimental Neuropsycholoay, A, 871-883. Flament-Durand, J., & Brion, J.P. (1990). Morphology of neurofibrillary tangles and senile plaques. In K. Maurer, P. Riederer, and H. Beckmann (Eds.), Alzheimer's Disease. Epidemioloay, Neuropatholoay, Neurochemistry. and Clinics (pp. 105-122). New York: Springer-Verlag. Forno, L.S. (1978). The locus coeruleus in Alzheimer's Disease. Journal of Neuropathology, 22, 614. Gallie, K.A., Graf, P., & Tuokko, H. (1990, November). Study Strategies and their ability to offset memory losses caused by Alzheimer's Disease. Paper presented at the Canadian Association of Gerontology/Acg, Victoria. Gallie, K.A., Graf, P., & Tuokko, H. (1991, June). Memory and attentional processes in healthy elderly versus those with Alzheimer's disease. Paper presented at the Canadian Psychological Association, Calgary. Gallie, K.A., Tuokko, H., & Graf, P. (1991, February). Longitudinal changes in memory and attention in Alzheimer's disease. Paper presented at the International Neuropsychological Society, San Antonio, Texas. Gaspar, P., & Gray, F. (1984). Dementia in Idiopathic Parkinson's Disease. Acta Neuropathologia, AA, 43-52. Gerzt, H.J., & Cervos-Navarro, J. (1990). Neuronal Plasticity of the septo-hippocampal pathway in patients suffering from dementia of Alzheimer type. In K. Maurer, P. Riederer. and H.Beckman (Eds.), Alzheimer's Disease. Epidemiology, Neuropathology, Neurochemistry, and Clinics (pp. 99-104). New York: Springer-Verlag. Gilman, A.G., Goodman, L.S., Rall, T.W., & Murad, F. (1985). The Pharmacological Basis of Therapeutics. Toronto: MacMillan Publishing Company.  200 Glass, G.V., & Hopkins, K.D. (1984). Statistical Methods in Education and Psychology. New Jersey: Prentice Hall. Grady, C.L., Grimes, A.M., Patronas, N., Sunderland, T., Foster, N.L., & Rapoport, S.I. (1989). Divided attention as measured by dichotic speech performance in Dementia of the Alzheimer Type. Archives of Neuroloay, Ai, 317-320. Graf, P., & Gallie, K.A. (1992). A transfer appropriate processing account for memory and amnesia. In N. Butters & L.R. Squire, (Eds.), Neuropsycholoay of Memory (pp. 241248). New York: Guilford Press. Graf, P., & Mandler, G. (1984). Activation makes words more accessible, but not necessarily more retrievable. Journal of Verbal Learning and Verbal Behavior, al, 553-568. Graf, P., Mandler, G., & Haden, P.E. (1982). Simulating amnesic symptoms in normal subjects. Science, 218, 1243-1244. Graf, P., & Masson, M.E.J. (in press). Introduction: Looking back and into the future. In P. Graf & M. Masson (Eds.), Implicit Memory: New Directions in Cognition. Development and Neuropsychology. Hillsdale, New Jersey: Lawrence Erlbaum. Graf, P., & Ryan, L. (1990). Transfer-appropriate processing for implicit and explicit memory. Journal of Experimental Psychology: Learning, Memory, and Cognition,^978-992. Graf, P., & Schacter, D.L. (1987). Selective effects of interference on implicit and explicit memory for new associations. Journal of Experimental Psychology: Learning, Memory. and Cognition, 12, 45-53. Graf, P., & Schacter, D.L. (1985). Implicit and Explicit memory for New Associations in Normal and Amnesic Subjects. Journal of Experimental Psychology: Learning. Memory, and Cognition, 11, 501-518. Graf, P., Shimamura, A.P., & Squire, L.R. (1985). Priming across modalities and priming across category levels: Extending the domain of preserved function in amnesia. Journal of Experimental Psychology: Learning. Memory, and Cognition, 11, 386-396.  201 Graf, P., Squire, L.R., & Mandler, G. (1984). The information that amnesic patients do not forget. Journal of Experimental Psychology: Learning, Memory and Cognition, 1Q, 164-178. Graf, P., Tuokko, H., & Gallie, K.A. (1990). Attentional deficits in Alzheimer's disease and related dementias. In J.T. Enns (Ed.), The Development of Attention: Research and Theory (pp. 527-544). Amsterdam, Elsevier Science Publisher. Green, R.L. (1988). Word stems as cues in recall completion tasks. Ouarterly Journal of Experimental Psychology, 38A, 663-673. Grober, E., Gitlin, H.L., Bang, S., & Buschke, H. (1992). Implicit and explicit memory in young, old, and demented adults. Journal of Clinical and Experimental Neuropsycholocm, j4., 298-316. Grosse, D.A., Wilson, R.S., & Fox, J.H. (1990). Preserved wordstem completion priming of semantically encoded information in Alzheimer's disease. psychology and Aging, 5, 304-306. Hafner, H. (1990). Epidemiology of Alzheimer's disease. In K. Maurer, P. Riederer, & H. Beckmann (Eds.), Alzheimer's Disease. Epidemiology, Neuropathology, Neurochemistry, and Clinics (pp. 24-39). New York: Springer-Verlag. Makin, A.M., & Matheson, G. (1979). Dementia in Parkinson Disease: A Neuropathologic Study. Neurology, 22, 12091214. Hardy, J.A., & Davies, D.C. (1988). Alzheimer's disease. British Journal of Hospital Medicine, 22, 372-377. Hardy, J.A., Mann, D.M.A., Wester, P., & Winblad, B. (1986). An integrative hypothesis concerning the pathogenesis and progression of Alzheimer's disease. Neurobiology of Aging, 2, 489-502. Heindel, W.C., Butters, N., & Salmon, D.P. (1988). Impaired learning of a motor skill in patients with Huntingtons' disease. behavioral Neuroscience, 122, 141-147. Heindel, W.C., Salmon, D.P., & Butters, N. (1990). Pictorial priming and cued recall in Alzheimer's and Huntington's disease. Brain and Cognition, la, 282-295.  202 Heindel, W.C., Salmon, D.P., Shults, C.W., Walicke, P.A., & Butters, N. (1989). Neuropsychological evidence for multiple implicit memory systems: A comparison of Alzheimer's, Huntington's, and Parkinson's disease patients. The Journal of Neuroscience, „a, 582-587. Hill, R.D., Storandt, M., & Simeone, C. (1990). The effects of memory skills training and incentives on free recall in older learners. Journal of Gerontology, 45, 227-232. Hippius, H. (1990). Aloys Alzheimer 14.06 1984-19.12 1915. In K. Maurer, P. Riederer, & H. Beckmann, (Eds.), Alzheimer's Disease: Epidemiology, Neuropathology. Neurochemistry and Clinics (pp XIII-XIX). New York: Springer-Verlag. Howard, D.V. (1980). Category norms: A comparison of the Battig and Montague (1969) norms with the responses of adults between the ages of 20 and 80. Journal of Gerontology, 15, 225-231. Howard, D.V. (1988). Implicit and explicit assessment of cognitive aging. In M.L. Howe & E.J. Brainerd (Eds.), Cognitive Development in Adulthood: Progress in Cognitive Development Research (pp. 3-37). New York: Springer-Verlag. Howard, D.V., Fry, A.F., & Brune, C.M. (1991). Aging and Memory for New Associations: Direct Versus Indirect Measures. Journal of Experimental Psychology: Learning. Memory, and Cognition, 12, 779-792. Howell, D.C. (1987). Statistical methods for Psychology. Boston: Duxbury Press. Huff, F.J., Growdon, J.H., Corkin, S., & Rosen, T.J. (1987). Age at Onset and Rate of Progression of Alzheimer's Disease. Journal of the American Geriatrics Association, 25, 27-30. Hultsch, D.F., Masson, M.E.J., & Small, B.J. (1991). Adult age differences in direct and indirect tests of memory. journals of Gerontology, 45, 22-30. Hyman, B.T., Van Hoesen, G.W., Kromer, L.J., & Damasio, A.R. (1986). Perforant pathway changes and the memory impairment of Alzheimers' Disease. Annals of Neurology, 2Q, 472-481. Insausti, R., Amaral, D.G., & Cowen, W.M. (1987). The Entorhinal Cortex of the Monkey: II. Cortical Afferents. Journal of Comparative Neurology, 264, 356-395.  203 Iqbal, K. (1991). Prevalence and neurobiology of Alzheimer's disease: Some highlights. In K. Iqbal, D.R.C. McLachlan, B. Winblad, & H.M. Wisniewski (Eds.), Alzheimer's Disease: Basic Mechanisms, Diagnosis and Therapeutic Strategies (pp. 1-5). Toronto: John Wiley and Sons. Java, R.I., & Gardiner, J.M. (1991). Priming and aging: Further evidence of preserved memory function. American Journal of Psychology, 104, 89-100. Jellinger, K. (1990). Morphology of Alzheimer's disease and related disorders. In K. Maurer, P. Riederer, & H. Beckmann (Eds.), Alzheimer's Disease, Epidemiology. Neuropathology, Neurochemistry and Clinics (pp. 61-77). New York: SpringerVerlag. Johnson, M.K., Kim, J.K., & Risse, G. (1985). Do alcoholic Korsakoff's syndrome patients acquire affective reactions? Journal of Experimental Psychology: Learning. Memory, and Cognition, 11, 27-36. Kandel, E.R., & Schwartz, J.H. (1985). Principles of Neural Science (2nd ed.). New York: Elsevier. Karisson, T., Backman, L., Herlitz, A., Nilsson, L.G., Winblad, B., & Osterlind, P.O. (1989). Memory improvement at different stages of Alzheimer's disease. Neuropsychologia, 2.2, 737-742. Katzman, R. (1986). Alzheimer's disease. The New England Journal of Medicine, 314, 964-973. Kausler, D.H., & Litchty, W. (1988). Memory for activities: Rehearsal independence and aging. In M.L. Howe & C.L. Brainerd (Eds.), Cognitive Development in Adulthood: Progress in Cognitive Development Research (pp. 93-131). New York: Springer-Verlag. Kausler, D.H., Lichty, W., Hakami, M.K., & Freund, S.J. (1986). Activity duration and adult age differences for activity performance. Psychology and Aging, 1, 80-81. Keane, M.M., Gabrieli, J.D.E., Fennema, A.C., Growdon, J.H., & Corkin, S. (1991). Evidence for a Dissociation between Perceptual and Conceptual Priming in Alzheimer's Disease. Behavioral Neuroscience, 2, 326-342. Kirk, R.E. (1968). Experimental Design for the Behavioral Sciences. Belmont, California: Brooks Cole Publishers.  204 Kirshner, H.S., Webb, W.G., & Kelly, M.P. (1984). The Naming Disorder of Dementia. Neuropsychologia, 22, 23-30. Kirsner, K., Milech, D., & Standen, P. (1983). Common and modality-specific processes in the mental lexicon. Memory and Cognition, 11, 621-630. Knopman, D. (1991). Long-term retention of implicitly acquired learning in patients with Alzheimer's disease. Journal of Clinical and Experimental Neuropsycholoay, 12, 880-894. Knopman, D.S., & Nissen, M.J. (1987). Implicit learning in patients with probable Alzheimer's Disease. Neuroloay, 784-788.  2z,  Korsakoff, S.S. (1889). Medical-psychological study of a form of diseases of memory. Revue Philosophique, 2a, 501-530. Landon, M., & Kidd, M. (1989). The structure of protein chemistry of senile plaques and cerebrovascular amyloid. Alzheimers' disease: Towards an understanding of the aetioloay and pathogenesis (pp. 65-75). London: John Libbey. Lassman, H., Fisher, P., Bancher, C., & Jellinger, K. (1990). Immunocytochemical and ultrastructural pathology of nerve cells in Alzheimer's disease. Epidemiology, Neuropatholoay, Neurochemistry, and Clinics (pp. 171-179). New York: Springer-Wein. Lewis, D.A., Campbell, M.J., Terry, R.D., & Morrison, J.H. (1987). Laminar and Regional Distributions of Neurofibrillary Tangles and Neuritic Plaques in Alzheimer's Disease: A Quantitative Study of Visual and Auditory Cortices. The Journal of Neuroscience, Z, 1799-1808. Lezak, M.D. (1983). Neuropsychological Assessment. New York: Oxford University Press. Light, L.L. (1991). Memory and aging: Four hypothesis in search of data. Annual Review of Psvcholoay, AZ, 333-376. Light, L.L., & Albertson, S.A. (1989). Direct and indirect tests of memory for category exemplars in young and older adults. psychology and Aging, A, 487-492.  205 Light, L.L., & Burke, D.M. (1988). Patterns of Language and Memory in Old Age. In L.L. Light & D.M. Burke (Eds.), Language. Memory. and Aging (pp. 244-271). Cambridge University Press. Light, L.L., & LaVoie, D. (in press). Direct and indirect measures of memory in old age. In P. Graf & M. Masson (Eds.), Implicit Memory: New Directions in Cognition. Development and Neuropsycholoay. Hillsdale, New Jersey: Erlbaum. Light, L.L., & Singh, A. (1987). Implicit and explicit memory in young and older adults. Journal of Experimental Psychology: Learning, Memory and Cognition, 12, 531-541. Light, L.L., Singh, A., & Capps, J.L. (1986). The dissociation of memory and awareness in young and older adults. Journal of Clinical and Experimental Neuropsvcholoay, 1, 62-74. Lockhart, R.S., & Craik, F.I.M. (1990). Levels of processing : A retrospective commentary on a framework for memory research. Canadian Journal of Psychology, AA, 87-112. Lovelace, E.A. (1990). Basic concepts in cognition and aging. In E.A. Lovelace (Ed.), Aging and Cognition: Mental Processes, Self Awareness and Interventions (pp. 1-28). North Holland: Elsevier Science Publishers. Mahut, H., Zola-Morgan, S., & Moss, M. (1982). Hippocampal resections impair associative learning and recognition memory in the monkey. Journal of Neuroscience, 2, 12141229. Mandler, G. (1980). Recognizing: The judgement of previous occurrence. Psychological Review, gz, 252-271. Mandler, G. (1988). Memory: Conscious and unconscious. In P.R. Soloman, G.R. Goethals, C.M. Kelly, & R.B. Stephens (Eds.), Memory: Interdisciplinary Approaches (pp. 84-106). New York: Springer-Verlag. Mann, D.M.A. (1991). Is the pattern of nerve loss in ageing and Alzheimer's Disease a real or only an apparent selectivity? Neurobioloay of Aging, 12, 340-342. Mann, D.M.A. (1985). The Neuropathology of Alzheimer's Disease: A Review with Pathogenetic, Aetiological and Therapeutic Considerations. Mechanisms of Ageing and Development, fl, 213-255.  206 Mann, D.M.A. (1989). The pathogenesis and progression of the pathological changes of Alzheimer's disease. In D.C. Davies (Ed.), Alzheimer's Disease: Towards an Understanding of the Aetiology and Pathogenesis (pp. 43-53). London: John Libbey. Mann, D.M.A., & Esiri, M.M. (1988). The site of the earliest lesions in Alzheimer's disease. New England Journal of Medicine, 318, 12. Martin, A. (1992). Degraded Knowledge Representations in Patients with Alzheimer's Disease: Implications for Models of Semantic and Repetition Priming. In L. R. Squire & N. Butters (Eds.), Neuropsycholoay of Memory (pp. 220-231). New York: Guilford Press. Martin, A., Brouwers, P., Cox, C., & Fedio, P. (1985). On the nature of the verbal memory deficit in Alzheimer's disease. Brain and Language, 2A, 323-341. Masson, M., & Graf, P. (in press). Introduction: Looking back and into the future. In P. Graf and M. Masson (Eds.), Implicit Memory: New Directions in Cognition, Development and Neuropsycholoay. Hillsdale, New Jersey: Lawrence Erlbaum. Matlin, M. (1983). Cognition. New York: CBS College Publishers. McDougall, W. (1924). Outline of Psychology. New York: Charles Scribner & Sons. McKhann, G., Drachman, D., Foistein, M., Katzman, R., Price, D., & Stadlan, E.M. (1984). Clinical diagnosis of Alzheimer's disease: Report of the NINCDS-ADRDA Work Group and Human Services Task Force on Alzheimer's Disease. Neurology, 2A, 939-944. Milner, B. (1958). Psychological deficits produced by temporal lobe excision. Research publications. Association for Research in Nervous and Mental Disease, 2A, 244-257. Mishkin, M. (1978). Memory in Monkeys severely impaired by combined but not by separate removal of amygdala and hippocampus. Nature, 273, 297-298.  207 Mishkin, M. (1989). Cerebral Organization Underlying Cognition. proceedings of the Seventeenth Annual International Neuropsychological Association (pp. 15). Vancouver, British Columbia, Canada: Swets & Zeitlinger B.V. Lisse. Mishkin, M., Malamut, B., & Bachevalier, J. (1984). Memories and Habits: Two Neural Systems. In J.L. McGaugh, N.M. Weinberger, & G. Lynch (Eds.), Neurobiology of Learning and Memory (pp. 65-77). New York: Guilford Press. Mitchell, D.B. (in press). Implicit and Explicit Memory for Pictures: Multiple Views Across the Lifespan. In P. Graf & M.E.J. Masson (Eds.), Implicit Memory: New Directions in Cognition. Development. and Neuropsycholocry.•Hillsdale, NJ: Erlbaum. Morris, C.D., Bransford, J.D., & Franks, J.J. (1977). Levels of processing versus transfer appropriate processing. Journal of Verbal Learning and Verbal Behavior, 1A, 519-533. Morris, R.G., & Kopelman, M.D. (1986). The memory deficits in Alzheimer type dementia: A review. The Ouarterly Journal of Experimental Psychology, 384, 575-602. Morris, J.C., McKeel, D.W., Storandt, M., Rubin, E.H., Price, J.L., Grant, E.A., Ball, M.J., & Berg, L. (1991). Very Mild Alzheimer's disease: Informant-based clinical, psychometric, and pathologic distinction from normal aging. Neurology, Ai, 469-478. Morris, J.C., & Rubin, E.H. (1991). Clinical diagnosis and course of Alzheimer's disease. In M. Davidson (Ed.), The Psychiatric Clinics of North America (pp. 223-236). London: W.B. Saunders Company. Moscovitch, M. (1982). Multiple Dissociations of function in amnesia. In L.S. Cermak (Ed.), Human Memory and Amnesia. New York: Erlbaum. Moscovitch, M., Winocur, G., & McLachlan, D. (1986). Memory as assessed by recognition and reading time in normal and memory impaired people with Alzheimer's Disease and other neurological disorders. Journal of Experimental Psychology: General, 115, 331-347. Mountcastle, V.B. (1978). Brain Mechanisms for Directed Attention. Journal of the Royal Society of Medicine, 21, 14-28.  208 Nebes, R.D. (1992). Semantic Memory Dysfunction in Alzheimer's Disease: Disruption of Semantic Knowledge or Information Processing Limitation? In L.R. Squire & N. Butters (Eds.), Neuropsycholoay of Memory (pp. 233-240). New York: Guilford Press. Nelson, H.E., & McKenna, P. (1978). The use of current reading ability in the assessment of dementia. Dritish Journal of Social and Clinical Psychology, 14, 259-267. Nelson, D.L., Reed, V.S., & McEvoy, C.L. (1977). Learning to order pictures and words: A model of sensory and semantic encoding. Journal of Expe imental Psychology: Human Learning and Memory, 2, 485-497. Nilsson, L.G., & Backman, L. (1991). Encoding dimensions of subject-performed tasks. Psychological Research, 52, 212218. Norris, M.P., & West, R.L. (1990). Adult age differences in activity memory: Cue and strategy utilization. In T.M. Hess (Ed.), Aging and Cognition: Knowledge Organization and Utilization (pp. 1-31). North Holland: Elsevier Science Publishers. Norris, M.P., & West, R.L. (1991).* Age differences in the recall of actions and cognitive abilities: The effects of presentation rate and object cues. Psychological Research, A2, 188-194. Nyberg, L., Nilsson, L.G., & Backman, L. (1991). A component analysis of action events. psychological Research, 12, 219225. Paller, K.A. (1990). Recall and stem-completion priming have different electrophysiological correlates and are modified differently by directed forgetting. Journal of Experimental Psychology: Learning. Memory. and Cognition, IA, 1021-1032. .  Paller, K.A., Kutas, M., Shimamura, A.P., & Squire, L.R. (1987). Brain responses to concrete and abstract words reflect processes that correlate with later performance on a test of stem-completion priming. In R. Johnson, Jr., J. W. Rohrbaugh, & R. Parasuraman (Eds.), Current trends in Eventrelated research (pp. 360-365). North Holland: Elsevier Science Publishers.  209  Parkin, A.J. (in press). Implicit memory across the lifespan. In P. Graf & M. Masson (Eds.), Implicit Memory: New Directions in Cognition. Development and Neuropsycholoay. Hillsdale, New Jersey: Lawrence Erlbaum. Partridge, F.M., Knight, R.G., & Feehan, M. (1990). Direct and indirect memory performance in patients with senile dementia. Psychological Medicine, 22., 111-118. Pearson, R.C.A., Esiri, M.M., Hiorns, R.W., Wilcock, G.K.,. & Powell, T.P.S. (1985). Anatomical correlates of the distribution of the pathological changes within the neocortex in Alzheimer's disease. proceedings of the National cademy of Sciences, 12, 4531-4534. Perlmutter, M. (1978). What is memory aging the aging of? Developmental Psychology, 1A, 330-345. Plude, D.J., & Hoyer, W.J. (1985). Attention and performance: Identifying and localizing age deficits. In N. Charness (Ed.), Aging and human performance (pp. 47-100). New York: Wiley. Poon, L.W. (1985). •Differences in human memory with aging: Nature, causes, and clinical implications. In J.E. Birren & K.W. Schaie (Eds.), Handbook of the psychology of aging (2nd ed., pp. 427-462). New York: Van Nostrand Reinhold. Price, J.L., Davis, P.B., Morris, J.C., & White, D.L. (1991). The Distribution of Tangles, Plaques and Related Immunohistochemical Markers in Healthy Aging and Alzheimer's Disease. Neurobioloay of Aging, 12, 295-312. Randolph, C. (1991). Implicit, explicit and semantic memory functions in Alzheimer's disease and Huntington's disease. Journal of Clinical and Experimental Neuropsycholoay, 11, 479-494. Rankin, J.L., & Collins, M. (1986). Adult age differences in memory elaboration. Journal of Gerontology, AQ, 451-458. Reisberg, B. (1983). Alzheimer's Disease: The Standard Reference. New York: The Free Press. Rey, A. (1964). L'examen clinique en psychologie. [The clinical  psychology examination]. Paris: Presses Universitaires de France.  210 Richardson-Klavehn, A., & Bjork, R.A. (1988). Measures of memory. Annual Review of Psvcholoay, 22, 475-543. Roediger, H.L. (1990a). Implicit memory: A commentary. bulletin of the Psychonomic Society, 21, 373-380. Roediger, H.L. (1990b). Implicit memory: Retention without remembering. American Psychologist, A5, 1043-1056. Roediger, H.L., Srinivas, K., & Weldon, M.S. (1989). Dissociations between implicit measures of retention. In S. Lewandowsky, J.C. Dunn, & K. Kirsner (Eds.), implicit Memory: Theoretical Issues (pp. 67-84). Hillsdale, New Jersey: Lawrence Erlbaum Associates. Roediger, H.L., Weldon, M.S., & Challis, B.H. (1989). Explaining dissociations between implicit and explicit measures of retention: A processing account. In H.L. Roediger & F.I.M. Craik (Eds.), Varieties of memory and, consciousness: Essays in honor of Endel Tulvinq (pp. 3-41). Hillsdale, New Jersey: Lawrence Erlbaum Associates. Rogers, J., & Morrison, J.H. (1985). Quantitative morphology and regional and laminar distributions of senile plaques in Alzheimer's disease. Journal of Neuroscience, 5, 2801-2808. Rohling, M.L., Ellis, N.R., & Scogin, F. (1991). Automatic and Effortful Memory Processes in Elderly Persons with Organic Brain Pathology. Journal of Gerontology: Psychological ,Sciences, 46, 137-143. Rosen, W.G., & Mohs, R.C. (1982). Evolution of cognitive decline in dementia. In S. Corkin (Ed.), Alzheimer's Disease: A Report of Progress (pp. 183-188). New York: Raven Press. Rosenthal, R., & Rosnow, R.L. (1991). Essentials of Behavioral Research and Data Analysis. Montreal: McGraw-Hill. Rudelli, R.D., Ambler, M.W., & Wisniewski, H.M. (1984). Morphology and distribution of Alzheimer neurotic and amyloid plaques in striatum and diencephalon. Acta Meuropathologia, AA, 273-286. Salmon, D.P., & Heindel, W.C. (1992). Impaired Priming in Alzheimer's Disease: Neuropsychological Implications. In L.R. Squire & N. Butters (Eds.), Neuropsycholoay of Memory (pp. 179-187). London: Guilford Press.  211 Salmon, D.P., Shimamura, A.P., Butters, N., & Smith, S. (1988). Lexical and semantic priming deficits in patients with Alzheimer's Disease. Journal of Clinical and Experimental Neuropsychology, 12, 477-494. Schacter, D.L. (1987). Implicit Memory: History and Current Status. Journal of Experimental Psychology: Learning, Memory, and Cognition, la, 501-518. Schacter, D.L. (1989). On the relation between memory and consciousness: Dissociable interactions and conscious experience. In H.L. Roediger & F.I.M. Craik (Eds.), Varieties of memory and consciousness: Essays in honor of Endel Tulvina (pp. 355-389). Hillsdale New Jersey: Lawrence Erlbaum Associates. Schacter, D.L. (1990a). Introduction to Implicit Memory: Multiple perspectives. bulletin of the Psychonomic Society, 2A, 338-340. Schacter, D.L. (1990b). Perceptual representation systems and implicit memory: Toward a resolution of the multiple memory systems debate. ,Annals of the New York Academy of Sciences, 608, 543-571. Schacter, D.L. (1992). Understanding Implicit Memory: A Cognitive Neuroscience Approach. American Psychologist, AZ, 559-569. Schacter, D.L., & Graf, P. (1989). Modality specificity of implicit memory for new associations. Journal of Experimental Psychology: Learning, Memory. and Cognition, 2, 432-444. Shallice, T. (1990). From Neuropsycholoay to Mental Structure. Cambridge: Cambridge University Press. Shimamura, A.P. (1986). Priming effects in amnesia: Evidence for a dissociable memory function. The Ouarterly Journal of Experimental Psychology, 381, 619-644. Shimamura, A.P. (in press). Neuropsychological analysis of implicit memory: History, Methodology and Theoretical Interpretations. In P. Graf & M. Masson (Eds.), Implicit Memory: New Directions in Cognition, Development and Neuropsychology. Hillsdale, New Jersey: Lawrence Erlbaum.  212 Shimamura, A.P., Salmon, D.P., Squire, L.R., & Butters, N. (1987). Memory dysfunction and word priming in dementia and amnesia. Behavioral Neuroscience, 101, 347-351. Shimamura, A.P., & Squire, L.R. (1984). Paired-associate learning and priming effects in amnesia: A neuropsychological study. Journal of Experimental Psychology: General, 213, 556-570. Snodgrass, J.G., & Corwin, J. (1988). Perceptual identification thresholds for 150 fragmented pictures from the Snodgrass and Vanderwart picture set. perceptual and Motor Skills, 0., 3-36. Snodgrass, J.G., & Corwin, J. (1988). Pragmatics of measuring recognition memory: Applications to dementia and amnesia. Journal of Experimental Psychology: General, 117, 34-50. Snodgrass, J.G., & Vanderwart, M. (1980). A standardized set of 260 pictures: Norms for name agreement, familiarity, and visual complexity. 'Journal of Experimental Psychology: Human Learning and Memory, 1, 174-215. SPSS-X Inc. (1990). Statistical Package for the Social Sciences: User's Guide. New York: McGraw-Hill. Squire, L.R. (1982). Comparisons between forms of amnesia: Some deficits are unique to Korsakoff's syndrome. Journal  of Experim-ntal Psychology: Learning. Memory. and Cognition, 8, 560-571.  Squire, L.R. (1987). Memory and Brain. New York: Oxford University Press. Squire, L.R. (1992a). Memory and the hippocampus: A synthesis from findings with rats, monkeys, and humans. Psychological Review, 22, 195-231. Squire, L.R. (1992b). Declarative and Nondeclarative Memory: Multiple Brain Systems Supporting Learning and Memory. Journal of Cognitive Neuroscience, A, 232-243. Squire, L.R., & Butters, N. (1992). Neuropsycholoay of Memory (pp. 1-3). New York: Guilford Press. Squire, L.R., Knowlton, B., & Musen, G. (1993). The Structure and Organization of Memory. Annual Review of Psychology, AA, 453-495.  213 Squire, L.R., Ojeman, J.G., Miezin, F.M., Petersen, S.E., Videen, T.O., & Raichle, M.E. (1992b). Activation of the hippocampus in normal humans: A functional neuroanatomical study of memory. Proceedings of the National Academy of ,Science, 12, 1837-1841. Squire, L.R., Shimamura, A.P., & Graf, P. (1985). Independence of recognition memory and priming effects: A neuropsychological analysis. Journal of Experimental psychology: Learning. Memory. and Cognition, U, 37-44. Squire, L.R., Slater, P.C., & Chace, P.M. (1975). Retrograde amnesia: Temporal gradient in very long-term memory following electroconvulsive therapy. Science, 187, 77-79. Squire, L.R., & Zola-Morgan, S. (1991). The Medial Temporal Lobe Memory System. Science, 253, 1380-1385. Squire, L.R., Wetzel, D., & Slater, P.C. (1978). Anterograde amnesia following ECT: An analysis of the beneficial effects of partial information. Neuropsychologia, lk, 339-348. Strauss, M.E., Weingartner, H., & Thompson, K. (1985). Remembering words and how often they occurred in memory impaired patients. Memory and Cognition, 11, 507-510. Stuss, D.T., Ely, P., Hugenholtz, H., Richard, M.T., LaRochelle, S., Poirier, C.A., & Bell, I. (1985). Subtle neuropsychological deficits in patients with good recovery after closed-head injury. Neurosurgery, 12, 41-47. Terry, R.D., & Katzman, R. (1983). Senile Dementia of the Alzheimer Type. Annals of Neurology, 1A, 497-506. Toledano-Gasca, A. (1988). Hypotheses concerning the aetiology of Alzheimer's Disease. pharmacopsychiatry, 21, 17-25. Tomlinson, B.E. (1972). Morphological brain changes in nondemented old people. In H.M. VanPrag and A.F. Kalverbov (Eds.), Aging of the Central Nervous System. Biological ancl Psychological Aspects (pp. 38-57). Bonn: Harlem Press. Tomlinson, B.E., Blessed, G., & Roth, M. (1968). Observations on the Brains of Non-demented old people. Journal of Neurological Science, 2, 331-356. Treves, T.A. (1991). Epidemiology of Alzheimer's disease. In M. Davidson (Ed.), The Psychiatric Clinics of North America (pp. 251-265). London: W.B. Saunders Company.  214 Tulving, E. (1986). What kind of hypothesis is the distinction between episodic and semantic memory? Journal of experimental Psycholoay: Learning, Memory, and Cognition. 12, 307-311. Tulving, E., & Schacter, D.L. (1990). Priming and Human Memory Systems. Science, p47, 301-306. Tuokko, H., & Crockett, D. (1989). Cued recall and memory in dementia. Journal of Clinical and Experimental Neuropsychology, 11, 278-294. Tuokko, H., & Crockett, D. (1991). Assessment of everyday functioning in normal and malignant memory disordered elderly. In D.E. Tupper & K.D. Cicerone (Eds.), The Neuropsychology of Everyday Life: Issues in Development and Rehabilitation (pp. 135-182). Boston: Kluver Academic Publishers. Tuokko, H., Gallie, K.A., & Crockett, D.J. (1990). Patterns of memory deterioration in normal and memory impaired elderly. Developmental Neuropsychologv, A, 291-300. van Dijk, P.T.M., Dippel, D.W.J., & Habbema, J.D. (1991). Survival of patients with dementia. Journal of the American Geriatric Society, 22, 603-610. Van Hoesen, G.W., & Damasio, A.R. (1987). Neural Correlates of cognitive impairment in Alzheimer's Disease. In V.B. Mountcastle, F. Plum, & S.R. Geiger (Eds.), Handbook of Physioloay: The Nervous System (pp. 871-898). Maryland: American Physiological Society. Victor, M., & Agamanolis, D. (1990). Amnesia due to Lesions confined to the Hippocampus: A Clinical-Pathologic Study. Journal of Cognitive Neuroscience, 2, 246-257. Warrington, E.K., & Weiskrantz, L. (1968). New method of testing long-term retention with special reference to amnesic patients. Nature, 217, 972-974. Warrington, E.K., & Weiskranz, L. (1970). Amnesia: Consolidation or retrieval? Nature, 224, 628-630. Welch, B.L. (1947). The generalization of students' problems when several different population variances are involved. eiometrika, 21, 29-35.  215 West, R.L., & Tomer, A. (1989). Everyday memory problems of healthy older adults: Characteristics of a successful intervention. In G.C. Gilmore, P.J. Whitehouse, & M.L. Wykle (Eds.), Memory, Aging and Dementia: Theory, Assessmeqt and Treatment (pp. 74-98). New York: Springer Publishing. Winer, B.J., Brown, D.R, & Michels, K.M. (1991). Statistical Principles in Experimental Design (3rd ed.). New York: McGraw-Hill Inc. Wisniewski, H.M., Currie, J.R., Wegiel, J., Silverman, W.P., Merz, G.S., Jenkins, E.C., & Iqbal, K. (1990). Pathological, biochemical and genetic aspects of Alzheimer's Disease. In J.M. Orgogozo & C.G. Gottfries (Eds.), Progress in Dementia Research (pp. 37-62). New Jersey: The Parthenon Publishing Group. Wisniewski, H.M., & Terry, R.D. (1973). Reexamination of the pathogenesis of the senile plaque. In H.M. Zimmerman (Ed.), Progress in Neuropathology, (pp. 1-26). New York: Grune and Stratton. Woods, R.T., & Piercy, M. (1974). A similarity between amnesic memory and normal forgetting. Neuropsychologia, la, 437445. Yates, F. (1934). Contingency tables involving small numbers and the Chi-square test. Journal of the Royal Statistical Society Supplement, 1, 217-235. Zola-Morgan, S. (1993, March). UBC Neuroscience Lecture Series. Zola-Morgan, S., & Squire, L.R. (1985). Medial Temporal lesions in monkeys impair memory in a variety of tasks sensitive to human amnesia. Behavioral Neuroscience, 22, 22-34. Zola-Morgan, S., & Squire, L.R. (1986). Memory impairment in monkeys following lesions of the hippocampus. Behavioral Neuroscience, 100, 155-160. Zola-Morgan, S., Squire, L.R., & Amaral, D.G. (1986). Human Temporal Region: Enduring Memory Impairments Following a Bilateral Lesion limited to Field CAI of the Hippocampus. The Journal of Neuroscience, A, 2950-2967. Zola-Morgan, S., & Squire, L.R. (1992). The Components of the Medial Temporal Lobe Memory System. In N. Butters and L.R. Squire, (Eds.), Neuropsycholoay of Memory (pp. 325-335). New York: Guilford Press.  216 Zola-Morgan, S., Squire, L.R., & Amaral, D.G. (1986). Lesions of the amygdala that spare adjacent cortical regions do not impair memory or exacerbate the impairment following lesions of the hippocampal formation. Journal of Neuroscience, 2, 1922-1936. Zola-Morgan, S., Squire, L.R., Amarall, D.G., & Suzuki, W.A. (1989). Lesions of Perirhinal and Parahippocampal Cortex that Spare the Amygdala and Hippocampal Formation Produce Severe Memory Impairment. The Journal of Neuroscience, 2, 4355-4370.  217  APPENDICES  218  Appendix A Letters of Contact and Consent  Table of Contents  (1) Letter inviting AD patients and Control Subjects to participate in study (includes information about the study and consent form). (2) Letter containing information about study for the patient's caregiver (includes consent form).  219  Date Dear Patient/Control Subject: This letter invites you to participate in a project that concerns MEMORY FOR FACTS AND EPISODES IN OLDER ADULTS. The project is part of my doctoral dissertation, and is conducted under the supervision of Dr. Holly Tuokko and Dr. Peter Graf. Both Dr. Tuokko and Graf are on the faculty of the University of British Columbia. You can reach Dr. Tuokko at 822-7535 and Dr. Graf at 822-6635. I am Karen Gallie -- a senior graduate student; you can reach me at 822-2140. The project focuses on memory for common words and pictures; it is part of a larger study that examines changes in memory that occur in adulthood and in various clinical conditions. There are two parts to the project, each of which lasts about 45 minutes. The two parts will be conducted on different days. If you would like to participate, I --Karen Gallie -- will contact you to arrange an appointment that is convenient to you. The project involves a series of short tasks that serve different purposes. The main purpose is to find out about two types of memory: the first has to do with such things as remembering items from a grocery list or the time of an appointment, whereas the second has to do with remembering things that were learnt a long time ago, such as the rules of a game. To find out about these two types of memory, you will do a series of paper and pencil task that require saying and writing common words in response to questions such as "what things belong to the category sports?" For another task we show you cards with words or pictures of common objects such as a desk or a pencil and then ask you to remember these items. Your performance on these tests gives an indication of how different types of memory function in adulthood. ' Your results will be treated in strict confidence. For this purpose the test forms will contain no information that would allow anyone else to find out about who you are. The results will be used to gain a better understanding of memory and to prepare scientific reports about memory functions in adulthood. Although you will participate as an individual, your individual results will not be published in any form; they will be combined with the results of a large group of adults. If you are interested, we will send you a copy of our findings at the conclusion of the study.  220 Participation in this project is completely voluntary and does not affect in any way your access to treatment or other programs. Participation requires about 2 times 45 minutes of your time with each 45 minute session arranged at your convenience. I (Karen Gallie) will visit you in your home to conduct each session. If at any time during the project you would like to stop, you are free to do so. Your withdrawal will be kept in strict confidence, will not be reported to anyone, and you will not be penalized in any way. If you have any further questions about this project and about its implications I will be glad to talk with you. You can contact me at 822-2140; or you can call Dr. Graf at 822-6635 or Dr. Tuokko at 822-7535. IF YOU WOULD LIKE TO PARTICIPATE PLEASE SIGN THIS FORM AND CONTACT ME BY CALLING 822-2140, OR DR. GRAF AT 822-6635. Sincerely,  Peter Graf Karen Gallie^ Doctoral Student^Associate Professor  If you would like to participate in this project, please print and sign your name below and return the bottom part of this form to me in the enclosed envelope. Your signature indicates that you have received and read a copy of this form.  Name:  ^ (please print)  Date: ^  Signature: ^  221  Date Dear Caregiver: Our files at the Clinic for Alzheimer's Disease and Related Disorders list you as the primary caregiver of ^ . With the enclosed letter we are inviting ^ to participate in a new research project that concerns MEMORY FOR FACTS AND EPISODES IN OLDER ADULTS. We are writing to you at this time to solicit your help to ensure that ^ understands fully the content of the enclosed letter which describes the nature and purpose of the research project, where and when it will be conducted, and that participation is entirely voluntary; in other words, we request your help in obtaining informed consent from the person in your care. The enclosed letter presents an outline of the project, describes the nature of the tasks that are involved, how much time is required of each participant, that participation is entirely voluntary, and it indicates what we plan to do with the results. If ^ has problems reading and/or understanding this letter, we would very much appreciate if you assist them in any way possible. The project is conducted under the supervision of Dr. Holly Tuokko and myself, Dr. Peter Graf. Both of us are on the faculty of the University of British Columbia. If you have any questions you can reach us by telephone: for Dr. Tuokko call 822-7535, and for Dr. Graf call 822-6635. If ^ decides to participate in our project after reading the enclosed letter will you please ask them to sign at the bottom of the page in the space provided. We would also appreciate your signature on the bottom of this letter to indicate to us that you have read the enclosed letter and that you have assisted the person in your care to make an informed decision about participating in our project. If the person in your care would like to participate in our project, PLEASE CONTACT ME BY CALLING 822-2140 OR 822-6635.  222 We appreciate your assistance in this matter. Sincerely,  Karen Gallie^Peter Graf Ph.D. candidate^Associate Professor  223  I have read the enclosed project description and assisted ^ to understand it and make an informed decision about participating in it.  Name:  (please print)  Date: ^  ^Signature: ^  224 Appendix B  Guidelines Used to Establish Presence of AD and Level of Functional Impairment  Table of Contents (1) NINCDS-ADRDA guidelines for making a Possible or Probable Diagnosis of AD. (2) Copy of Functional Rating Scale and description of procedures used to assign patients FRS scores and categorize them into mild, moderate, and severe levels of functional impairment.  Criteria for clinical diagnosis of Alzheimer's disease from the NINCDS-ADRDA  1. The criteria for the clinical diagnosis of PROBABLE Alzheimer's disease include: — dementia established by clinical examination and documented by the MiniMental Test (Folstein et al., 1975), Blessed Dementia Scale (Blessed et al., 1968), or some similar examination, and confirmed by neuropsychological tests — deficits in two or more areas of cognition — progressive worsening of memory and other cognitive functions — no disturbance of consciousness — owes between ages 40 and 90, most often after age 65 — absence of systemic disorders or other brain diseases that in and of themselves could act punt for the progressive deficits in memory and cognition 11. The diagnosis of PROBABLE Alzheimer's disease is supported by: — progressive deterioration of specific cognitive functions such as language (aphasia), motor skills (apraxia), and perception (agnosia) — impaired activities ofdaily living and altered patterns of behavior — family history of similar disorders, particularly if confirmed neuropathologically — laboratory results of: — normal lumbar puncture as evaluated by standard techniques — normal pattern or nonspecific changes in EEG, such as increased slow-wave activity — evidence of cerebral atrophy on CT with progression documented by serial observation Ill. Other clinical features consistent with the diagnosis of*PROBABLE Alzheimer's disease, after exclusion of causes of dementia other than Alzheimer's disease, include: — plateaus in the course ofprogression of the illness — associated symptoms of depression; insomnia; incontinence; delusions; illusions; hallucinations; catastrophic verbal emotional, or physical outbursts; sexual disorders; and weight loss — other neurological abnormalities in some patients, especially with more advanced disease, and including motor signs such as increased muscle tone, myoclonus, or gait disorder — seizures in advanced disease — CT normal for age^' IV. Features that make the diagnosis of PROBABLE Alzheimer's disease uncertain or unlikely include: — sudden, apoplectic onset — focal neurologic findings such as hemiparesis, sensory loss, visual field deficits, and incoordination early in the course of the illness — seizures or gait disturbances at the onset or very early in the course of the illness V. Clinical diagnosis of POSSIBLE Alzhiemer's disease: — may be made on the basis of the dementia syndrome, in the absence of the neurologic, psychiatric, or systemic disorders sufficient to cause dementia, and in the presence of variations in the onset, in the presentation, or in the clinical course — may be made in the presence ofa second systemic or brain disorder sufficient to produce dementia, which is not considered to be the cause ofthe dementia — should be used in research studies when a single, gradually progressive severe cognitive deficit is identified in the absence of other identifiabk cause VI. Criteria for diagnosis of DEFINITE Alzheimer's disease arc: — the clinical criteria for probable Alzheimer's disease — histopathologic evidence obtained from a biopsy or autopsy VII. Classification of Alzheimer's disease for research purposes should specify features that may differentiate subtypes of the disorder, such as: — familial occurrence — onset before age of 65 — presence of trisomy-21 — coexistence of other relevant conditions such as Parkinson's disease :  225  ^ NEWRY ^ CUESTIONAII.E (1) (7)  NEWAY  SOCIAL/COMMUNITY AND  Memory losses utitch  Neither patient nor relatives suer. of any deficit  Variable levels of functioning repotted by patient er relatives no objective evidence of deficits in employmeet or social . Megatons  Patient or relative aware of decreased performaoce In demendleg omplopmeet or social settings. spews menial to casual infection  Ne thanes, noted by patieet or relative  Slightly decreased involvement la household teaks and hobbles  Engages le social activities to this home but defialte impaimmeat in some household tests. same complicated hobbies and laterests aband000d  Only dimple chores/hobbies preserved. most complicated hobbles/interests abondowed  No isidepeedeet involvemeat to home sr hobbies  fully capable of self-care  Occasional problems with self-care reported by pottiest/ relatives or observed  Needs prompting to complete tasks adequately (1.e. dress)ng. feeding. hygiene)  Requires supervisiee in dressing. feeding. homilies. and keeping track of personal effects  Needs coaster* superyule. and assistaace. with fee tag. dressialk er hygiene. etc.  No disturbance of temples. reported by patient or relettlai  Subjective complaint of. er relative reports. levovseo deficits. usually limited to ward timeless or Ramie,  Fattest er relative resorts verloble disturbeaces or sock skills as articulation Of nosing. occasteoal lihguage Westmont evident during esamlostlem.  Patient er relative reports consistent imposts disturtaace. language dIsturberce evident em examinatiem  Severe lopeirmeat of receptive aer/er sapressive language. preductien of welatellitIbl* speech  Solves everyday problems adequately  Variable lopelrmoat of problem-2000h  Difficulty in mealtime temples problems  Narked Impstrnoot em temples problmasolving tasks  11061114, es 1101.0 PfebION St 1W47 level, Via sad error behevier often ebs4OVed  Imfrogeent Changes in  Frequeet thong's In affect reported by patient or relative. noticeeble to objective observer  Sestolood elterotiems of effect. imemirmd contact with reality observed er reported  Usually disorleated to time *ad often to Plato  Oriented arty to person or net et 411  CARE  LANGUAGE SKILLS  PIONItEN SOLVING AND  ORIENTATION  interfere with daily living. sere apparent for recent events.  Moderato mosery loss. only highly leaned arterial retained. new material rapidly lost  Uwe?* memory loss. needle to recall rel evant aspects of curreat life. witty Mnch, recall of past life.  Patient or relative are  Narked Impairment of  aware of cooing deter-  ioration. eses not pear mensal to objective observer. unable to Perform job. little 1o/dependent functioning outside home  differences  REASONING  AFFECT  NYE' (S)  Variable ssipteis reported by patient or relative. seemingly unrelated to level of functioning  AND HOMES  PERSONAL  ^ MODERATE ^ (4)  No deficit Or inceasisteet forgetfulness evident only ee clinical interview  OCCUPATIONAL  NOW  ^ NILO^ (3)  Me change in effect reported by patient or relative  Apiropriatit concern with reseset to symptematesey  Fully oriented  Occasional difficulties with time relatierships  affect (e.g. Irrita-  bility) reported by flatfeet or relative. mould appear morsel to objective observer Herbed difficulty with time relationships  social functioning. ao ladopoadoet functioning outside hems  (Tuokko & Crockett, 1989)  227 Assignment of FRS Scores to AD Patients  Each patient was assigned an FRS score ranging from one (normal), to five (severe), for each of the eight dimensions of this scale. FRS scores were obtained from a consensual team score obtained from the clinic's geneticist, geriatrician, neurologist, neuropsychologist, psychiatrist, and social worker. These eight FRS scores were then averaged to produce one estimate of the patient's overall level of functional impairment (average FRS  score = AFRAS), or dementia severity.  The following cut-off scores were used to assign patients to functional impairment groups: Mildly impaired: 2.13 < AFRAS < 2.63;  Moderately impaired: 2.63 < AFRAS < 3.13; and Severely  impaired: 3.13 or greater.  228  Appendix C  Telephone Interview  Table of Contents  (1) Procedure and questions used to screen volunteers who acted as non-demented control subjects.  229  Procedure and questions used to screen volunteers who would act as controls in this investigation.  Procedure  People who were interested in acting as volunteers were contacted by telephone. During the telephone conversation this experimenter began by identifying herself and her affiliation with UBC and confirming that the person she was speaking to was interested in participating in a memory study. If the respondent said yes the nature of the study was explained to them including the fact that the investigation was designed to learn more about memory in people with Alzheimer's disease including strategies that we were developing to help them remember more information. They were also informed of their rights as UBC subjects and the fact that if they participated they would receive a ten dollar honorarium and a copy of the study's results upon completion. They were then told that to conduct this investigation we required healthy older adults to act as controls. Controls would need to be healthy people who were 50 years of age or older who did not have any health or memory problems so that we could establish how normal people performed on our tasks in comparison to people with Alzheimer's disease. To be considered as a control it would be necessary for volunteers to meet very strict  230 selection criteria and this meant that I would need to ask them some very personal questions. They were told that they did not have to answer any of these questions if they did not want to and they could ask me any questions they wanted in return. The following are questions that were covered during the course of the informal telephone interview. In most cases volunteers spontaneously provided information that made direct questioning unnecessary. In no case did a volunteer refuse to answer any of the following questions.  Questions 1). How old are you? (Volunteers had to be 50 years of age or older). 2). How would you describe your memory abilities? (If they reported concern but could not name a serious memory loss event they were considered normal. If they reported a repetitive problem remembering events they were not considered). 3). Do you, or have you in the past, suffered from a neurological or psychiatric problem such as depression, or stroke, or have you been in an accident where you hit your head? (If they reported yes they were not included as a control). 4). How would you describe your current health? (If they described their health as being poor they were not accepted to act as a control). 5). Are you currently taking any medications? (If they reported continuously taking more than three medications they were thanked for their time but told that, due to the nature of the study, they could not act as a control).  231  Appendix D Materials Used for Explicit and Implicit Memory Tests  Table of Contents (1) Written and Spoken Word Materials (2) Picture Materials (3) Object Materials  232 Implicit and Explicit Memory Test Materials Spoken and Written Word Materials Categories and three items named in response to it are contained below. Numbers found in brackets indicate the frequency in which that word was provided by 442 university students. Please refer to the method section for the procedure used to select these categories from Battig and Montague's norms (1969). SET1  SET 2  STATE PENNSYLVANIA (30%) TEXAS (33%) ILLINOIS (42%)  VEHICLE BOAT (32%) MOTORCYCLE (39%) BICYCLE (44%)  SHIP CRUISER (32%) DESTROYER (36%) SAILBOAT (40%)  MONEY PENNY (55%) QUARTER (59%) DIME (59%)  FLOWER ORCHID (30%) DAISY (40%) CARNATION (41%)  WEAPON RIFLE (37%) GUN (89%) KNIFE (92%)  FOOTGEAR SLIPPERS (36%) SANDALS (50%) SOCKS (58%) Set average = 39%  COUNTRY MEXICO (31%) SPAIN (36%) ITALY (36%) Set average = 50.7%  SET 3  aBT__1  STONE PEARL (40%) SAPPHIRE (55%) EMERALD (74%)  KITCHEN UTENSIL SPATULA (35%) POT (46%) PAN (54%)  RELATIVE NIECE (31%) NEPHEW (32%) GRANDFATHER (68%)  METAL ZINC (30%) TIN (39%) SILVER (57%)  233 READING MATERIAL PAMPHLET (45%) NEWSPAPER (67%) BOOK (83%)  PART OF BUILDING ROOM (36%) CEILING (38%) FLOOR (54%)  CLOTH LINEN (32%) NYLON (48%) RAYON (50%) Set average = 52%  WEATHER SLEET (39%) HAIL (47%) SNOW (60%) Set average = 44.6%  SET 5  SET 6  CITY BALTIMORE (37%) SAN FRANCISCO (39%) LOS ANGELES (43%)  BIRD CANARY (30%) BLUEBIRD (31%) CROW (34%)  TREE BIRCH (30%) ELM (48%) PINE (48%)  MUSICAL INSTRUMENT OBOE (33%) TROMBONE (39%) SAXOPHONE (40%)  GIRL'S NAME JANE (30%) ANN (36%) SUE (37%)  DWELLING CAVE (35%) TENT (43%) APARTMENT (71%)  VEGETABLE ASPARAGUS (31%) SPINACH (37%) LETTUCE (43%) Set average = 38%  FISH TUNA (31%) PERCH (32%) SALMON (32%) Set average = 37.6%  ^ SET 7 ^ TIME ^ WEEK (63%) ^ DECADE (72%) ^ MONTH (73%) ^ DISTANCE ^ MILLIMETER (41%) ^ KILOMETER (48%) ^ CENTIMETER (58%)  SET 8 MILITARY TITLE ADMIRAL (33%) CORPORAL (38%) MAJOR (55%) FRUIT LEMON (30%) GRAPEFRUIT (35%) PLUM (38%)  234 4 FOOTED ANIMAL PIG (32%) ELEPHANT (41%) TIGER (46%)  CLOTHING TIE (31%) SWEATER (37%) HAT (45%)  COLOUR BROWN (49%) PINK (50%) WHITE (62) Set average = 52.9%  SPORT GOLF (35%) SOCCER (36%) SWIMMING (63%) Set average = 39.7%  235  Picture Materials Items selected from Snodgrass & Corwin's fragmented picture norms (1988) that were consistently named by 70% or more of their 219 subjects and were given a rating of 2 or more for item familiarity and image agreement. A score of 1 indicates that the picture was unfamiliar and there was no image match, and a score of 5 indicates that the picture was very familiar and was a good match to the subject's mental image of that item. Refer to the method section for the procedure used to select and administer these items. Percentage Naming Consistency  Item Familiarity (Mean score)  Image Consistency (Mean score)  100 98 100 95 100 98 98 100 98 79 98 86 95.8%  2.58 4.12 4.75 4.32 4.60 4.88 2.88 4.42 3.58 3.65 3.48 2.68 3.83  4.33 4.05 4.33 3.18 3.05 4.15 3.78 4.40 3.98 2.98 3.70 4.18 3.84  98 100 95 83 93 100 76 98 95 90 90 98 93j  3.98 4.72 4.18 4.40 4.52 3.28 2.05 3.58 4.38 4.45 2.70 3.64 3.82  4.05 3.65 3.76 4.02 3.78 3.58 3.58 4.20 2.65 3.25 3.52 3.28 3.61  SET 1 BALLOON BELT BOOK DESK DOG EYE MUSHROOM PENCIL RULER SUITCASE VEST VIOLIN AVERAGE SET 2 APPLE BED BOWL BREAD COMB FISH GORILLA GUITAR HOUSE KNIFE MOUNTAIN SHIRT AVERAGE  236  Item Familiarity (Mean score)  Image Consistency (Mean score)  90 100 95 93 90 100 93 100 98 83 90 95 93.9%  2.28 4.50 2.75 4.68 4.59 3.25 4.50 2.22 2.90 4.48 4.02 2.78 3.58  4.50 4.08 3.85 3.85 2.71 4.35 4.10 4.10 4.26 2.78 2.80 2.72 3.68  88 100 81 89 100 100 98 88 100 71 100 100 92.9%  3.78 3.42 4.70 3.88 3.55 2.48 3.18 3.18 2.95 2.22 4.08 3.95 3.45  3.40 4.35 3.10 2.59 4.18 3.60 3.65 3.51 4.20 4.44 3.92 3.92 3.73  Percentage Naming Consistency SET 3 AXE BUS DUCK FRIDGE HAIR LEMON LIPS OWL PIPE SWEATER TRUCK VASE AVERAGE  SET 4 BIKE BROOM CAR COAT FOOTBALL FROG HAT LOCK RABBIT ROLLING PIN TOASTER UMBRELLA AVERAGE  237  Picture Stimulus Set 1 The following pictures are from Snodgrass and Vanderwarts' (1980) norms. These pictures have been scanned into an Apple MacIntosh computer and subjected to a fragmentation algorithm program. This program produced a fragment series for each picture by cumulatively deleting randomly selected 16 X.16 pixel blocks (see Snodgrass & Corwin, 1988). The percentage of deleted blocks followed an exponential function. Refer to the next page for the amount of picture displayed at each fragment level.  LEVEL 8 (100% intact) ^  LEVEL 7 (70% intact)  I  41  243  mo  ti  ti  f •  LC ■•  244  L  /^  N (  O  • a /  •  •  •  •  247  •r  Ij )  a-  = r  r 1  J^  =  r  = ^ -  1  r  248  •  ;  4  t  -; 1...^  •-; it■ ..: I •t '  ......:  ,*^/1^ ■^  .^ i^ I  4  t  1 ■  1  ^  I^ t  I  i  -  )1^ -la di/  ' AP/  250  Picture Stimulus Set 2  251  I  \  A  1101•1 IC. .11.:A ..... Os. h ......■^... 7 ...  41  . .. i' l 252 . ..-  I  h  13  • ..._ -7  -  aro  arms  253  •""`  ,e■  •  41•1111M  I  254  ...•■••• •■••••••■•  ro.  \\V4^  a'"  255  ea. ...  :  4  ;4: •  ..-  it  46 1  ■ ,/ ' • ' . '^.* .  •  -4 •  4  .A • •  t'; • • •  4  -  A  11%24  .  'I  L^1 1  '1  •  44  • V  IE Ella^j 259  J  •••■  ...•  -• • ,^  f^::.  `i oriT E l iffl = 4.1,7 j  •M•M  a  th: I I,ArIt = Awls  .-_- J  •  .*  MUM^ amea  is =311  •  • Nol. MEM  •  I  a  •••=1  -  is  a  •  ■•■■■•  1111111M•  As  • • \  260  r" •  •••  .00  •  .0"  /  .  ,  •••  /  '..  4 J  •  •  •• 1 I  k■  I /^ k■  •  k■  ,  ,^ ■^  a  ■  It.  kJ\  '  /^^ /  J  .  or,  "16  ^  .4.  ^  A-  A.  I  263  Picture Stimulus Set 3  265  a  IV  l  •  )  276  Picture Stimulus Set 4  288  V  1  289 Object Materials  Items were selected because they were common articles that weren't sharp on touch and could be placed behind a 46 cm by 20.5 cm partition. Refer to the method section for the procedure used to select and administer these items. SET 1^  SET 2  BOOT^ CLOTHESPIN^ CUP^ ELASTIC BAND^ FLAG^ HAMMER^ HANGER^ SCREW^ SOAP^ SOCK^ SPOON^ TOOTHBRUSH^  BANDAID BELT BOOK BOWL CANDLE DOLL FORK HAT PENCIL SCISSORS SCREWDRIVER THIMBLE  SET 3^  SET 4  BOW^ BRUSH^ ENVELOPE^ GLASS^ GLOVE^ KEY^ KNIFE^ NAILFILE^ PIPE^ ROLLERSKATE^ RULER^ TOILET PAPER^  BALL BASKET BOTTLE BOX CAR GLASSES LOCK PAINTBRUSH PAPERCLIP PEN PLIERS SHOE  290  Appendix E  Test Booklet Used to Record Responses  Type ^ Name ^ Years of education ^ Age^ Occupation^ Address ^ Phone Number ^ Medications ^  291  Results:  1. 2. 3.  4.  5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15.  Experimenter Subject Picture Completion Picture Recognition Letters Meaning Auditory Category Production Visual Category Production Tactile Production Tactile Recognition Auditory Category Recall Visual Category Recall See See/Say See/Say/Do  1/. Exp/Sub.^2/. Picture Comp/Recog. set.  Exp.^Implicit^T Sub. Explicit^T  4/. Auditory/Visual Cat. ^Product. Implicit  B  5/. Tactile Ident/Recog. Implicit Explicit  D  D  6/. Auditory/Visual Cat.^Recall. Explicit  B  7/. Multi-Modal see ^ see/say see/say/do  D  3/. Letters/Meaning  I letters meaning  292  EXPERIMENTER/SUBJECT Procedure:  Instructions: STUDY: In this first task we're going to do some things together. I have some commands written on these cards, some of which I will do, and some of which you will do. For some of the tasks we'll need to use the items that are in front of us. When we're finished I'm going to ask you to tell me all the things we did together. You don't have to remember who did what, just what was done. Any questions? Good, now for all of the tasks that you'll be doing today we'll always start with some practice items so you understand what it is that I want you to do. Here are the practice items for this task, I'll do the first one and I'll have you watch me do it. The first command is "raise your hands", now watch me as I do it. Good! Now you do the next one "put the glasses in the case" - Good - you've got the idea. Now we'll do the following in the same way, but before we start I'd like you to tell me the names of these items (point to each of the props and have them tell you its name (i.e., elastic band, matches and matchbox, paper, pen, thimble, toothpick). TEST: Now I'd like you to tell me all of the things that we just did together. 1/ Free Recall 1) 2) 3) 4) 5) 6) 7) 8) 9)  2/ Cue (physical items)^3/ Cue (words)  PICTURE COMPLETION/RECOGNITION Procedure:  A. Implicit/Explicit Targets ^,^] B. Implicit Test (target b distracter) C. Explicit Test (target & distracter)  293  ]  ]  Instructions: STUDY: Now I'm going to show you some pictures that I'd like you to look at. For each picture I want you to tell me what it is, and something about it. The more personal meaning you can tell me about the picture the better. For example, this is a picture of a ball and its the type of ball that I'd take to the beach. The next picture is an elephant, and the elephants are my favorite animal in the zoo. You do the next two (airplane, ashtray). That's good. Now I want you to do the same thing for these next pictures. (show the 24 pictures constituting implicit and explicit targets).  IMPLICIT TEST: Now we're going to do a different task. I'm going to show you some pictures that vary in their amount of completeness and I want you to tell me when you think you know what the picture is. (Now show them the elephant practice item and illustrate how the picture starts out with not very much information and gradually increases in the amount of completeness. Tell them that some of the items will be more difficult than others and that it is important that they clear their minds and tell you whatever they think the picture is. Tell them that this is a new task and that it is important that they tell you whatever first pops to mind without thinking back to anything done previously). Record all responses and level. When they've successfully identified item proceed to the next level to ensure confidence of response. If they appear to be trying to remember the pictures at study discourage them by repeating instructions (i.e. it is important that you don't try to think of anything we've previously done. Just tell me what first comes to mind).  EXPLICIT TEST: • Now I'm going to show you some more pictures, but this time all I want you to do is say yes, if this was one of the pictures I've just shown you, or no if you don't remember seeing this picture before.  Explicit Test  Implicit Test Item  Level^  294  Item^ItPL012  1)  1)  2)  2)  3)  3)  4)  4)  5)  5)  6)  6)  7)  7)  8)  8)  9)  9)  10)  10)  11)  11)  12)  12)  13)  13)  14)  14)  15)  15)  16)  16)  17)  17)  18)  18)  19)  19)  20)  20)  21)  21)  22)  22)  23)  23)  24)  24)  295  LETTERS/MEANING Procedure:^fletters [ meaning [  ]  ]  Instructions: STUDY: For this task I'm going to show you some cards that have different words written on them. For all of the words I'd like you to first tell me what the word is. Then, I'll either ask you to count the number of letters in the word, or I'll ask you to tell me what that word means to you. When we're finished I'll ask you to tell me the words that were written on the cards. Now let's try some practice cards- (Bus- please tell me the # of letters this word has. That's good. Now the next word is-show car- I'd like you to tell me what meaning this word has for you). That's good, now I want you to do the same thing for these next cards. TEST: 1/ Words recalled ^2/ 1) 2) 3) 4) 5) 6) 7) 8) 9) 10) 11) 12)  Recognition.  296  CATEGORY PRODUCTION/RECOGNITION ^ IMPLICIT CONDITION Procedure: A. Auditory/Visual Targets^, B. Auditory/Visual Test^( . Instructions: STUDY: For this task I'm either going to say some words, or show you some that are written on cards. For each word that is presented, I'd like you to first repeat the word, and then I'd like you to tell what that word means to you. For example, if I said the word car, what would you say. That's good, or you could have also said something like, Car, I bought my first car when I was sixteen. If I showed you a card with the word cow written on it, what would you say ? Good! Now I'd like you to do the same thing for these words. (present setf auditorally, and ^visually, total=24 words). TEST:  Now I'm going to show you cards that have different categories written on them, and I'd like you to tell me the first three words that are members of that category that come to mind when you see that category. Here's the first category, what three words come to mind? Category^Items (3) 1) 2) 3) 4) 5) 6) 7) 8)  13) 14) 15) 16)  297  TACTILE IDENTIFICATION/RECALL Procedure:^A. Implicit/Explicit Targets [ , ] B. Implicit Test (target & distracter) C. Explicit Test (target E. distracter) Instructions: STUDY:  Now I'd like you to put both of your hands behind this curtain - That's Good-. I'm going to place an object in your hands and I'd like you to tell me what it is, and something about it. For example, let's practice with this item, tell me what it is, and something about it, [practice=loonie, packet of kleenex]. (Present set # for implicit and # for explicit target, total=24items). IMPLICIT TEST: Now I'm going to place some more objects in your hands and I'd like you to tell me what you think it is. I'll be timing you with this stopwatch so I'd like you to tell me what the item is as soon as you know. After you've identified the object then have you tell me something about it or what you'd use it for. (Present implicit target and distracter #'es b , total=74).  EXPLICIT TEST: Now I'm going to place some more objects in your hands and I'd like you to answer yes, if this was an item I gave you to feel before, or no, if you don't remember feeling this item before. (Present explicit target and distracter res , total=24).  ITEMS:^  298  SET1: boot, clothespin, cup, elastic band, flag, hammer, hanger, screw, soap, sock, spoon, toothbrush. SET2: bandaid, belt, book, bowl, candle, doll, fork, hat, pencil, scissors, screwdriver,. thimble. SET3: bow, brush, envelope, glass, glove, key, knife, nailfile, pipe, rollerskate, ruler, toilet paper. SET4: ball, basket, bottle, box, car, glasses, lock, paintbrush, paperclip,^pen,^pliers, shoe. Implicit Test  Explicit Test  Item^Response 1^Response 2 1) 2)  Item^Yes/No 1) 2)  3) 4)  3) 4)  5) 6)  5) 6)  7) 8)  7) 8)  9) 10)  9) 10)  11) 12)  11) 12)  13) 14)  13) 14)  15) 16)  15) 16)  17) 18)  17) 18)  19) 20)  19) 20)  21) 22)  21) 22)  23) 24)  23) 24)  Procedure:  EXPLICIT CONDITION A. Auditory/Visual Targets B. Auditory/Visual Test [ ,  299  Instructions: STUDY: This task is similar to one that we did before. I'm either going to say some words, or show you some that are written on cards . Like before I'd like you to first repeat the word and then tell me something about that word. For example, if I said the word car you'd say, car, and then you might say something like I bought my first car when I was sixteen. Any questions? TEST: Now I'm going to show you cards that have different categories written on them. First I'd like you to tell me if you've just studied any words that are members of that category, and then I'd like you to tell me what three words you just studied that belong to this category. Category^Items (3) 1) 2) 3) 4)  5) 6) 7) 8)  9) 10) 11) 12)  13) 14) 15) 16)  MULTISENSORY EXPERIMENT  ^  Procedure:^see [ ] see/say [ ] see/say/do [ ]  300  .  Instructions: STUDY: In this task I'm going to show you some cards - some of which I will ask you to read silently to yourself, some of which I will ask you to read out loud to me, and some that I'll have you read out loud to me and then do what the card says. For this task you may need to use the items you see in front of you (have them identify items as in first task). Do you have any questions? Good- now let's try a practice set. For this card I'd like you to read it silently to yourself (practice #1), I'd like you to read this card out loud to me (practice #2), and for this card I'd like you to read it out loud to me and then do what it says (practice #3). Do you have any questions? Good, now let's begin. When we're finished I'm going to ask you to remember as many of the items as you can. 1/ Free Recall  1) 2) 3) 4) 5) 6) 7) 8) 9) 10) 11) 12)  2/ Cue (physical items)^3/ Cue (words)  301  Appendix F Materials and Cues Used in Memory Strategy Experiments  Table of Contents (1) Levels of Processing Experiment (2) SPT-EPT Experiment (i.e., Subject-Experimenter Performed Tasks) (3) Multisensory Experiment  302 Memory Strategy Materials Levels of Processing Experiment Words contained in material sets 1 and 2 are target items that were selected because they had similar numbers of letters and vowels. Depending on the set, words were either studied for the number of letters or personal meaning they had for the subject. Sets 1 and 2 were counterbalanced across letter and meaning conditions for control and patient groups. The distractor set is the nine words that were used in the recognition part of this experiment. Distractor items were chosen because they either had similar numbers of letters, vowels or meaning to target items. Refer to the method section for the procedures used to select and administer these items. SET 1^  SET 2  TIN^ JANE^ RIFLE^ VIOLET^ SWEATER^ KEROSINE^  PIG PINE LEMON BEETLE CEILING VIRGINIA  DISTRACTOR SET JILL BILL DOG RAIN BIKE SWING PURPLE SPIDER CLOTHING  SPT-EPT Experiment Commands contained in material sets 1 and 2 were either performed by the subject or the experimenter. Command sets were counterbalanced across subject and experimenter conditions so that one subject would perform items in set 1 and watch the experimenter perform set 2 and the next subject would perform set 2 and watch set 1 performed etc. Refer to the method section for the procedures used to select and administer these items.  303  SET a) b) c) d) e) f)  ^  SET2  BLINK THREE TIMES POINT TO YOUR MOUTH CLAP YOUR HANDS BREAK THE TOOTHPICK PUT THE THIMBLE ON YOUR FINGER STRETCH THE ELASTIC BAND  g) h) i) j) k) 1)  NOD YOUR HEAD CROSS YOUR FINGERS SCRATCH YOUR NOSE PUT THE MATCH IN THE BOX DRAW A CIRCLE WITH THE PEN KNOCK ON THE TABLE  OBJECTS USED TO PERFORM THESE MINITASKS ELASTIC BAND MATCHES AND BOX PAPER AND PEN FINGER THIMBLE TOOTHPICKS CUES USED IN CUED RECALL CONDITION (e.g., cue a is for item a above) a) ONE OF THE COMMANDS SAID WAS? b) ONE OF THE COMMANDS SAID REMEMBER WHAT THAT WAS? c) ONE OF THE COMMANDS SAID WHAT THAT WAS? d) ONE OF THE COMMANDS SAID WHAT THAT WAS? e) ONE OF THE COMMANDS SAID REMEMBER WHAT THAT WAS? f) ONE OF THE COMMANDS SAID WHAT THAT WAS? g) ONE OF THE COMMANDS SAID WHAT THAT WAS? h) ONE OF THE COMMANDS SAID WHAT THAT WAS? i) ONE OF THE COMMANDS SAID WHAT THAT WAS? j) ONE OF THE COMMANDS SAID REMEMBER WHERE THAT WAS? k) ONE OF THE COMMANDS SAID WHAT THAT WAS? 1) ONE OF THE COMMANDS SAID REMEMBER WHERE THAT WAS?  TO BLINK, DO YOU REMEMBER WHAT THAT TO POINT TO SOMETHING, DO YOU TO CLAP SOMETHING, DO YOU REMEMBER TO BREAK SOMETHING, DO YOU REMEMBER TO PUT THE THIMBLE SOMEWHERE, DO YOU TO STRETCH SOMETHING, DO YOU REMEMBER TO NOD SOMETHING, DO YOU REMEMBER TO CROSS SOMETHING, DO YOU REMEMBER TO SCRATCH SOMETHING, DO YOU REMEMBER TO PUT THE MATCH SOMEWHERE, DO YOU TO DRAW SOMETHING, DO YOU REMEMBER TO KNOCK ON SOMETHING, DO YOU  304 Multisensory Experiment Commands contained in material sets 1, 2, and 3 were either performed by silently reading the item, reading the command to the experimenter, or reading the item aloud and performing the command. Similar to previous tests these material sets were counterbalanced across performance conditions. Refer to the method section for the procedures used to select and administer these items. SET1 a) TOUCH YOUR NOSE b) MAKE A SIGN FOR GOODBYE c) PUT THE PENCIL IN THE CASE d) MAKE THE CLOCK SAY 3 O'CLOCK SET2 a) TOUCH YOUR EAR b) MAKE A SIGN FOR YES c) PUT THE PEN IN THE DRAWER d) MAKE THE CLOCK SAY 6 O'CLOCK SET3 a) TOUCH YOUR MOUTH b) MAKE A SIGN FOR NO c) PUT THE SCISSORS IN THE BOX d) MAKE THE CLOCK SAY 9 O'CLOCK OBJECTS USED TO PERFORM THESE SENTENCES BOX CLOCK CUP PEN PENCIL PENCIL CASE SCISSORS CUES USED IN CUED RECALL CONDITION (e.g., cue a is for sentences a above) a) ONE SENTENCE SAID TO TOUCH SOMETHING, DO YOU REMEMBER WHAT THAT WAS?  b) ONE SENTENCE SAID TO MAKE A SIGN FOR SOMETHING, DO YOU REMEMBER WHAT THAT WAS? C) ONE SENTENCE SAID TO PUT SOMETHING SOMEWHERE, DO YOU REMEMBER WHAT THAT WAS? d) ONE SENTENCE SAID TO MAKE THE CLOCK SAY SOMETHING, DO YOU REMEMBER WHAT THAT WAS?  305  Appendix G  Tables  306  Table G -8a AD Patients* and Control Subjects' Performance on Category Cued Recall and Category Completion Tests for Written Words. *Mild, Moderate, and Severe indicates level of Functional Impairment. AD Patients Controls^All^Mild* Moderate* Severe* (n=40) (n=7) 01=20) (n=6) (n=7) EXPLICIT MEMORY/ Category Cued Recall Corrected Score  8.68  SD^(1.83) % Score^72.33 Target M^8.95  5D^(1.75) % Score^74.58  Baseline  M^0.27 D^(0.85) % Score^2.25  1.85 (2.28)  15.42  4.50 (1.98)  37.50  0.43 (0.79)  1.00 (1.52)  2.05 (2.50) 17.08  4.83 (2.22) 40.25  0.43 (0.79) 3.58  1.29 (1.89) 10.75  0.20 (0.62) 1.67  0.33 (0.82) 2.75  0.00 (0.00) 0.00  0.29 (0.76) 2.42  1.30 (2.20)  1.33 (3.39)  1.28 (1.38)  1.28 (1.98)  3.58  8.33  IMPLICIT MEMORY/  Category Completion  Priming or Corrected Score  M^1.62 52^(2.39)  Target  % Score^13.50  10.83  11.08  10.67  10.67  M^3.80  2.45 (2.16) 20.42  2.83 (2.63) 23.58  2.14 (1.46) 17.83  2.42 (2.57) 20.17  1.15 (1.18) 9.58  1.50 (1.38) 12.50  0.86 (0.69) 7.17  1.14 (1.46) 9.50  ED^(2.36) % Score^31.67  Baseline  M^2.18 SD^(1.11) % Score^18.17  307  Table G-8b Two Factor Repeated Measures MANCOVA* with Subjects as the Between Factor (AD Patients, Control Subjects) and Test Type (Explicit, Implicit) as the Within Factor for Written Word Materials. Source of Variation  df  Mean Square  Homogeneity Tests Box M 3,34728 Bartlett-Box(Explicit) 1,^7249 (Implicit) 1,^7249  1.13 1.24 .16  .334 .265 .681  Between Factor Effects Within cells Regression Constant Group  57 1 1 1  5.99 .87 82.58 306.47  .15 13.78 51.14  .704 .001 .001  Within Factor Effects Within cells Test Type Group by Test Type  58 1 1  3.49 385.07 281.67  110.33 80.70  .001 .001  621.07 1.41  157.08 0.26  .001 .613  Main Effects Tests Explicit Test by Group 1,58 Implicit Test by Group 1,58  *Note. Subjects' level of education is the covariate, Category Cued Recall is the explicit, and Category Completion is the implicit memory test.  308  Table G-9a AD Patients* and Control Subjects' Performance on Category Cued Recall and Category Completion Tests for Spoken Words. *Mild, Moderate, and Severe indicates level of Functional Impairment. AD Patients Controls^All^Mild* Moderate* Severe* (n=40) (n=20) (n=6) (n=7) (n=7) EXPLICIT MEMORY/  Category Cued Recall  Corrected Score SD % Score  8.40 (2.67)  2.45 (2.56)  4.17 (2.71)  1.14 (1.86)  9.50  19.00  2.28 (2.49)  70.Q0  20.42  34.75  8.53 (2.39) 71.08  2.55 (2.63) 21.25  4.17 (2.72) 34.75  1.14 (1.86) 9.50  2.57 (2.69) 21.42  0.13 (0.52) 1.08  0.10 (0.45) 0.83  0.00 (0.00) 0.00  0.00 (0.00) 0.00  0.29 (0.76) 2.42  Priming or Corrected Score M 1.73 1.10 SD (2.48) (1.55) 9.17 % Score 14.41  1.17 (1.72)  1.20 (1.64)  1.29 (1.79)  Target M SD % Score Baseline M  5D  % Score  IMPLICIT MEMORY/  Category Completion  Target M  9.75  10.00  10.75  % Score  3.63 (2.32) 30.25  2.10 (1.51) 17.50  2.67 (1.21) 22.25  1.86 (1.35) 15.50  1.86 (1.95) 15.50  Baseline M SD % Score  1.90 (1.01) 15.83  1.00 (0.92) 8.33  1.50 (0.84) 12.50  0.66 (0.82) 5.50  0.57 (0.97) 4.75  an  309 Table G-9b Two Factor Repeated Measures MANCOVA* with Subjects as the Between Factor (AD Patients, Control Subjects) and Test Type (Explicit, Implicit) as the Within Factor for Spoken Word Materials. Source of Variation  df  Mean Square  F  p  Homogeneity Tests 3,34728 Box M Bartlett-Box(Explicit) 1,^7249 (Implicit) 1,^7249  1.86 .03 4.82  .134 .844 .028  Between Factor Effects Within cells Regression Constant Group  57 1 1 1  6.39 15.62 40.51 232.71  2.45 6.34 36.44  .123 .015 .001  Within Factor Effects Within cells Test Type Group by Test Type  58 1** 1**  5.32 429.34 189.04  80.68 35.52  .001 .001  472.03 5.21  68.01 1.06  .001 .308  Main Effects Tests Explicit Test by Group 1,58 Implicit Test by Group 1,58**  Note. *Subjects' level of education is the covariate, Category  Cued Recall is the explicit, and Category Completion is the implicit memory test. **Box-adjusted values reported in the results section.  310 Table G -10a AD patients and Control Subjects' Performance on Picture Recognition and Picture Fragment Completion Tests. *Mild, Moderate, and Severe indicates level of Functional Impairment. AD Patients Controls^All^Mild*^Moderate* Severe* (n=40)^(n=20) (n=6)^(n=7)^(n=7) EXPLICIT MEMORY/ Picture Recognition (don't know and false negative responses are not reported) Corrected Score X  11.70 (0.61) 97.50  5.90 (3.78) 49.20  8.50 (3.45) 70.80  4.15 (3.76) 34.50  4.40 (3.26) 36.67  11.85 (0.36) 98.75  9.45 (2.96) 78.75  11.33 (1.21) 94.42  8.29 (4.11) 69.08  7.97 (2.08) 66.40  % Score  11.70 (1.43) 97.50  7.80 (4.13) 65.00  8.50 (3.39) 70.83  7.57 (5.03) 63.08  7.43 (4.31) 61.92  False Positive M = % Score  0.15 (0.53) 1.25  3.55 (3.69) 29.58  2.83 (2.92) 23.58  4.14 (4.91) 34.50  3.57 (3.31) 29.75  Z2  % Score  Target X  W.  % Score  Distractor X  22  ^  continued  311 IMPLICIT MEMORY/ Picture Fragment Completion Test (max. score = 8) Priming or Corrected Score N^1.05 0.49 0.46 0.32 (0.65) (0.64) (0.41) (0.53) aa % Score 6.10 13.12 5.75 4.00  Target  0.48 (0.51) 6.00  M 5_2  4.71 (0.66)  5.92 (0.91)  5.91 (0.97)  6.23 (0.34)  6.32 (0.67)  baseline M 5..12  5.76 (0.57)  6.41 (0.73)  6.37 (1.03)  6.55 (0.57)  6.80 (1.23)  312 Table G -10b Two Factor Repeated Measures MANCOVA* with Subjects as the Between Factor (AD Patients, Control Subjects) and Test Type (Explicit, Implicit) as the Within Factor for Picture Materials.  Source of Variation  df  Mean Square  Homogeneity Tests 3,34728 Box M Bartlett-Box(Explicit) 1,^7249 (Implicit) 1,^7249 Between Factor Effects Within cells Regression Constant Group  57 1 1 1  201.78 .10 15429.02 18233.71  Within Factor Effects Within cells Test Type Group by Test Type  58 1** 1**  196.68 105139.41 11913.92  Main Effects Tests Explicit Test by Group 1,58** Implicit Test by Group 1,58  31148.15 489.55  26.59 80.14 .27  .001 .001 .602  0.01 76.46 90.36  .983 .001 .001  534.56 .001 60.57 .001  90.90 9.36  .001 .003  Note. *Subjects' level of education is the covariate, Picture  Recognition is the explicit, and Picture Fragment Completion is the implicit memory test. ** Box-adjusted values reported in the results section.  313  Table G-lla AD patients and Control Subjects' Mean Performance on the Tactile Recognition Test. AD Patients Controls^All^Mild^Moderate Severe  (n=40)^(n=16) (n=6)  (n=5)*^(n=5)*  EXPLICIT MEMORY/ Tactile Recognition (don't know and false negative responses are not reported). Corrected Score M SD % Score  Target M SD % Score  11.67 (0.94) 97.25  11.82 (0.59) 98.50  5.90 (3.96) 49.17  8.25 (3.71) 68.75  8.83 (2.93) 73.58  11.00 (1.09) 91.67  6.14 (4.29) 51.17  8.00 (4.08) 66.67  3.14 (2.54) 26.17  6.14 (3.62) 51.17  Distractor M SD % Score  11.82 (0.59) 98.50  9.15 (2.93) 76.25  9.17 (3.90) 76.42  9.86 (1.95) 82.17  8.43 (3.05) 70.25  False Positive M SD % Score  0.15 (0.43) 1.25  2.35 (2.25) 19.58  2.17 (2.78) 18.08  1.86 (1.57) 15.50  3.00 (2.52) 25.00  ^continued  314 Table G -11a continued AD patients and Control Subjects' Mean Performance on the Tactile Identification Test. AD Patients Controls^All^Mild^Moderate Severe  (n=40)^(n=16) (n=6)^(n=5)*^(n=5)*  IMPLICIT MEMORY/Tactile Identification test (median time savings in milliseconds % score as a proportion of baseline performance) 014 Materials Priming or Corrected Score M SD % $core Target SD Baseline  SD  119.65 288.31 159.33 175.80 (103.04) (454.41) (65.53) (202.40) 59.29 74.62** 50.05 64.87** 304.07 (127.11) 184.42 (68.47)  304.40 (823.82) 32.00  774.56 372.83 527.00 1253.00 (812.87)(112.49) (202.20) (1366.93) 486.25 213.50 (691.34) (68.82)  351.20 948.60 (74.84) (1174.39)  New Materials Priming or Corrected Score M^136.40 64.13 158.83 104.20 -89.60 an (84.80) (577.84) (89.93) (133.84) (1085.87) 73.96** 13.19 % Score 74.39** 29.67 -9.45 Target  5D Baseline  $D  320.83 (127.11) 184.43 (68.47)  550.38 372.33 455.40 (412.63)(125.55) (151.68) 486.25 213.50 (691.34) (68.82)  859.00 (646.44)  351.20 948.60 (74.84) (1174.39)  Note. *Four patients could not perform the tactile identification test and so their performance was excluded. **Patients took longer to perform this test and so there was a greater margin available for improvement.  315 Table G-lib Two Factor Repeated Measures MANCOVA* with Subjects*** as the Between Factor (AD Patients, Control Subjects) and Test Type (Explicit, Implicit) as the Within Factor for Old Materials.  Source of Variation  df  Mean Square  p  Homogeneity Tests Box M 3,15033 Bartlett-Box(Explicit) 1,^4959 (Implicit) 1,^4959  17.44 54.52 .01  .001 .001 .990  Between Factor Effects Within cells Regression Constant Group  53 1 1 1  1989.20 1829.05 64667.98 7908.01  .92 32.51 3.98  .342 .001 .051  Within Factor Effects Within cells Test Type Group by Test Type  54 1** 1**  1776.48 44.24 16976.41  .02 9.56  .875 .003  77.55 .36  .001 .549  Main Effects Tests Explicit Test by Group 1,54** 30880.21 Implicit Test by Group 1,54 1221.67  Note. *Subjects' level of education is the covariate, Tactile Recognition is the explicit, and Tactile Identification is the implicit memory test. **Box-adjusted values reported in the results section. ***Four subjects dropped from analysis.  316  Table G -11c Two Factor Repeated Measures MANCOVA* with Subjects as the Between Factor (AD Patients, Control Subjects) and Test Type (Explicit, Implicit) as the Within Factor for New Materials.  Source of Variation^df^Mean Square  F  p  Homogeneity Tests Box M^ 3,15033 Bartlett-Box(Explicit) 1, 4959 (Implicit) 1, 4959  18.75 60.14 .03  .001 .001 .844  Between Factor Effects Within cells^53^2454.61 Regression^1^1669.13 Constant^ 1^85935.89 Group^ 1^35444.57  .68 35.01 14.44  .542 .001 .001  Within Factor Effects Within cells^54^28971.84 Test Type^1** 2337448.05 Group by Test Type^1** 1029079.76  80.68 35.52  .001 .001  Main Effects Tests Explicit Test by Group 1,54** ^30880.21 Implicit Test by Group 1,54^8470.10  77.55 9.36  .001 .003  Note. *Subjects' level of education is the covariate, Object Recognition is the explicit, and Object Identification is the implicit memory test. **Box-adjusted values reported in the results section.  317 Table G-12a AD Patients and Controls' Performance in the Levels of Processing Experiment.  Controls All (n=401 (n=20)  AD Patients Mild^Moderate Severe (n=6) (n=7) (n=7)  Score (maximum=6) Free Recall Meaning Encode M % Corrected  3.88 (1.31) 64.67  1.65 (1.69) 27.50  2.33 (1.86) 38.83  1.86 (1.95) 31.00  0.86 (1.07) 14.33  Letter Encode M ail % Corrected  2.45 (1.32) 40.83  0.70 (1.26) 11.67  0.16 (0.41) 2.67  1.43 (1.90) 23.83  0.43 (0.54) 7.17  20.75  27,42  10.75  a2  Average Free Recall Score (%) 52.75^19.58 Recognition Meaning Encode M SD % Score % Corrected  2.08 (1.27) 34.67 98.11  2.90 (1.80) 48.33 66.67  3.33 (1.63) 55.50 90.70  2.29 (2.63) 38.17 55.31  3.57 (1.81) 59.50 69.46  Letter Encode M SD % Score % Corrected  3.08 (1.47) 51.33 86.76  3.30 (2.36) 55.00 62.26  4.50 (1.64) 75.00 77.05  1.89 (1.68) 31.50 41.36  3.29 (2.43) 54.83 59.06  83.88  48.34  64.26  Average Cued Recall Score (%) 92.43  64.46  Note. Average Free Recall and Recognition scores are derived from  different unit values and therefore do not sum to 100%.  318 Table G-12b^Levels of Processing Experiment Three Factor Repeated Measures MANCOVA* with Subjects as the Between Factor (AD Patients, Control Subjects) and Two Within Subject Factors (Study Task, Test Type) each with Two Levels (Meaning & Letter Encode, Free Recall and Recognition). Source of Variation^df^Mean Square  F  Homogeneity Tests Box M^  10,6902  7.98  .001  Bartlett-Box meaning/free recall ^1,7249 meaning/recognition^1,7249 letter/free recall^1,7249 letter/recognition^1,7249  1.80 70.74 .05 8.05  .180 .001 .821 .005  Between Factor Effects Within Cells^57^1060.73 Regression^1^38.98 Constant^ 1^62688.97 1^42093.20 Group^  .04 59.10 39.68  .849 .001 .001  Within Factor Effects-Study Task Within Cells^58^341.06 Study Task^1^13546.87 Group by Study Task^1^175.21  39.72 .51  .001 .476  Within Factor Effects-Test Type Within Cells^58^627.29 Test Type^ 1** 103840.83 Group by Test Type^1**^925.93  165.54 1.48  .001 .229  Within Factor Effects-Study Task by Test Type Within Cells^58^364.72 Study Task by Test Type^1**^792.25 2.17 Group by Study Task by^1**^245.58 .67 Test Type  .146 .415  Note. *Subjects' level of education is the covariate and one degree of freedom is lost from the within cells source of variations as a result. **Box-adjusted values are reported in the results section.  319  Table G-13a AD Patients and Controls' Performance in the SPT-EPT Memory Strateay Experiment. AD Patients Controls^All^Mild^Moderate Severe Ln=40)^(n=20)^(n=6)^(n=7)^(n=7)  Score (maximum=6) Free Recall SPT.Performed M^4.20^0.75^1.50^0.43^0.43 aa^(1.29)^(0.91)^(1.05)^(0.54)^(0.79) % Corrected 70.00^12.50^25.00^7.15^7.16 EPT.Performed M^3.70^0.75^2.17^0.14^0.14 aa^(1.52)^(1.21)^(1.33)^(0.38)^(0.38) % Corrected 61.67^12,50^36.17^2.38^2.33 Average Free Recall Score (%) ^ ^ 65.83 12.50^30.58 4.75^4.75 Cued Recall SPT.Performed M SD % Score % Corrected EPT.Performed M  SD  % Score Corrected  1.43 (1.08) 23.83 79.44  2.95 (1.61) 49.17 56.19  3.33 (1.03) 55.50 74.00  2.86 (2.12) 47.67 51.35  2.71 (1.60) 45.17 48.65  1.58 (1.24) 26.33 68.69  2.45 (1.47) 40.83 46.67  1.83 (1.17) 30.50 47.78  2.43 (1.51) 40.50 41.47  3.00 (1.63) 50.00 51.19  Average Free Recall Score (%) ^ ^ 60.89 74.06^51.43 46.41^49.92 Note. SPT = subject performed, EPT = experimenter performed minitask. Average Free and Cued Recall scores are derived from different unit values and therefore do not sum to 100%.  320  Table G-13b^SPT-EPT Experiment Three Factor Repeated Measures MANCOVA* with Subjects as the Between Factor (AD Patients, Control Subjects) and Two Within Subject Factors (Study Task, Test Type) each with Two Levels (Subject-Performed & Experimenter-Performed, Free and Cued Recall). Source of Variation^df  Mean Square  F  Homogeneity Tests 10,6902  1.19  .286  Bartlett-Box SPT/free recall^1,7249 SPT/cued recall^1,7249 EPT/free recall^1,7249 EPT/cued recall^1,7249  2.71 .01 1.26 .90  .100 .922 .262 .342  .10 32.67 58.79  .748 .001 .001  8.34 .88  .005 .352  550.01 38940.02 8755.21  70.80 15.92  .003.  AD Patients SPT vs EPT Cued Recall^1,19 31733.89 568.89 SPT vs EPT Free Recall^1,19  63.15 1.26  .001 .275  SPT Free vs. Cued Recall 1,19 20400.28 EPT Free vs. Cued Recall 1,19 11902.50  49.19 22.09  .001 .001  Box M^  between Factor Effects Within Cells^57 Regression^1 Constant^ 1 1 Group^  1347.75 140.10 44031.65 79227.64  Within Factor Effects-Study Task Within Cells^58^398.64 3325.02 Study Task^1 350.21 Group by Study Task^1  Within Factor Effects Test Type -  Within Cells^58 Test Type^ 1 Group by Test Type^1  .001  Main Effects Tests  ^  continued  321  Table G-13b continued^SPT-EPT Experiment Three Factor Repeated Measures MANCOVA* with Subjects as the Between Factor (AD Patients, Control Subjects) and Two Within Subject Factors (Study Task, Test Type) each with Two Levels (Subject-Performed & Experimenter-Performed, Free and Cued Recall). Source of Variation  ^  df^Mean Square F  Main Effects Tests Control Subjects 1,39 1,39  8075.07 4375.07  14.09 9.01  .001 .005  SPT Free vs. Cued Recall 1,39 EPT Free vs. Cued Recall 1,39  5335.56 2920.14  12.47 4.45  .001 .041  Within Factor Effects-Study Task by Test Type Within Cells^58^491.08 741.69 1 Study Task by Test Type 1.51 Group by Study Task by Test Type 137.25 .28 1  .224  SPT vs EPT Cued Recall SPT vs EPT Free Recall  .599  322  Table G-14a AD Patients and Controls' Performance in the Multisensory Experiment (i.e., Do, Say, and See conditions).  AD  Patients  Controls^All^Mild^Moderate Severe (n=20) (n=40) (n=6) (n=7) (n=7) Score^(maximum=4) Free Recall Do 2.90 M (0.96) aa 72.50 % Corrected Say M 1.68 (1.21) an 42.00 % Corrected See 1.38 M (1.13) SD 34.50 % Corrected  0.65 (0.99) 16.25  1.50 (1.05) 37.50  0.29 (0.76) 7.25  0.29 (0.76) 7.25  0.25 (0.72) 6.25  0.67 (1.21) 16.75  0.14 (0.38) 3.50  0.00 (0.00) 0.00  0.20 (0.52) 5.00  0.33 (0.82) 8.25  0.14 (0.38) 3.50  0.14 (0.38) 3.50  Average Free Recall Score (%) 49.67 Cued Recall Do M  SD  % Score % Corrected  Say., M SD % Score % Corrected See M  SD  % Score % Corrected  9.17  20.83  4.75  3.58  0.75 (0.74) 18.75 68.18  1.50 (1.28) 37.50 44.77  1.33 (0.82) 33.25 53.20  1.86 (1.57) 46.50 50.13  1.29 (1.38) 32.25 34.77  0.85 (0.92) 21.25 36.63  0.45 (0.76) 11.25 12.00  1.00 (1.09) 25.00 30.03  0.29 (0.49) 7.25 30.03  0.14 (0.38) 3.50 7.51  1.08 (0.79) 27.00 41.22  0.45 (0.89) 11.25 11.84  0.67 (1.21) 16.68 18.26  0.29 (0.76) 7.25 7.51  0.43 (0.79) 10.75 11.14  31.58  21.35  Average Cued Recall Score (%) 44.37^22.02  4.24  Note. Average Free and Cued Recall scores are derived from different unit values and therefore do not sum to 100%.  323 Table G-14b^Multisensory Experiment Three Factor Repeated Measures MANCOVA* with Subjects as the Between Factor (AD Patients, Control Subjects) and Two Within Subject Factors (Study Task, Test Type).^Study Task has Three Levels (Do, Say, and See) and Test Type has two Levels (Free and Cued Recall). df^Mean Square  Source of Variation  F  Homogeneity Tests 21,^5551  3.05  .001  1,7249 1,7249 1,7249 1,7249 1,7249 1,7249  11.67 4.47 5.87 2.08 .03 .50  .001 .035 .015 .149 .866 .478  57^1927.50 1^2314.78 1^17218.98 1^91084.47  1.20 8.93 47.26  .278 .004 .001  Within Factor Effects-Study Task Within Cells^116^856.86 2**^24390.05 Study Task Group by Study Task 2**^1720.29  28.46 2.01  .001 .139  Box M Bartlett-Box see/free recall see/cued recall say/free recall say/cued recall do/free recall do/cued recall Between Factor Effects Within Cells Regression Constant Group  Within Factor Effects-Test Type Within Cells^58^886.28 1**^8336.81 Test Type Group by Test Type* 1^1167.05  9.41 1.32  .003 .256  Within Factor _Effects-Study Task by Test Type Within Cells Study Task by Test Type Group by Study Task by Test Type*  116 2** 2**  560.40 542.82 1266.59  .97 2.26  .383 .109  Note. * Subjects' level of education is the covariate and one degree of freedom is lost from the within cells source of variation. **Box adjusted values reported in results section.  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0098809/manifest

Comment

Related Items