Melbourne School of Psychological Sciences - Theses
Permanent URI for this collection
Search Results
1 - 5 of 5
-
ItemVolumetric Measures of the Medial Temporal Structures and their Neuropsychological Correlates in Alzheimer’s Disease( 2020)Diagnosing Alzheimer’s disease (AD) in its earliest phases is difficult. A combination of standardised cognitive testing with AD-related biomarkers such as medial temporal atrophy observed on neuroimaging scans has been proposed to improve the diagnostic process of AD detection. The availability of automated volumetric segmentation software presents as a promising avenue for the incorporation of quantitative volumetric measures into routine clinical workup. However, the performance of these automated protocols in a clinical context remains to be clarified. This thesis aimed to examine the use of automated segmentation in a “real-world” clinical setting against the gold standard of manual segmentation, assess the difficulties surrounding reliable automated measurements, and evaluate the relationship between volume, diagnosis, and memory function. Methods utilised to address the research questions included a review of the diagnostic accuracy of various diagnostic tools, in-depth examination of structural variations that could undermine automated segmentation efforts, a series of validation studies, and assessment of medial temporal volumes and memory function in samples of clinical and healthy controls. The results indicate that a multimodal approach to diagnosis that includes a combination of cognitive assessment and other biomarkers would likely yield a more accurate diagnosis. Although automated segmentation of the medial temporal structures was promising for the hippocampus, the entorhinal and transentorhinal cortices were inherently more challenging to delineate and produced much poorer reliability measures. Medial temporal structural volumes contributed little beyond what was already provided by comprehensive memory assessment for distinguishing between the clinical and healthy control samples. Taken altogether, these findings highlight the need for additional refinement of automated algorithms and further investigation into the ideal combination of cognitive testing and other biomarkers.
-
ItemArbitrary associative memory: a candidate cognitive endophenotype for probable Alzheimer's disease( 2015)The genetic basis of sporadic, senile onset Alzheimer’s disease (AD) is not as well documented as familial, younger-onset AD. At present, there is no evidence to support monogenic inheritance in senile onset AD or “probable AD”. As far as genetic risk factors are concerned, the apolipoprotein E epsilon-4 allele is the only reliable genetic susceptibility factor for probable AD. The investigation of cognitive endophenotypes for probable AD, or cognitive markers of liability to the condition that are theoretically closer to its genetic mechanisms than its clinical phenotype, might assist in the identification of susceptibility genes for AD. This is especially the case if these cognitive differences reflect genetically predetermined differences in brain structure and function. The traditional view of probable AD is that cognitive dysfunction emerges many years, and possibly decades, after underlying brain neuropathological, metabolic, and structural changes have begun. According to this theory, while brain functional and structural changes might be present in middle age in people at risk of developing the disease, cognitive abnormalities emerge only when a “threshold” of neurodegeneration has been reached and unaffected brain regions can no longer compensate functionally for affected areas. An alternative viewpoint is that cognitive deficits are present in the earliest stages of probable AD, but that two issues have limited our ability to detect such changes: First, previous studies have utilised tests which are not sensitive to the earliest neurocognitive changes in the condition. Second, the heterogeneous clinical phenotype of probable AD has meant that the conventional approach of comparing mean scores on cognitive tests between low-risk and high-risk groups might be masking subsets of affected people within high-risk groups. The use of more sensitive tests, in conjunction with cognitive discrepancy analyses, which compare performances within high-risk groups on combinations of tasks (e.g., Test A versus Test B) that are related to the neuroanatomical changes in probable AD, might reveal cognitive markers of liability to the condition. The current research focused on middle-aged people with a parental history of probable AD. These people are at a greater risk of developing the condition themselves, than any other first-degree relative, and this risk might be independent of the apolipoprotein E epsilon-4 allele. People with a maternal history of probable AD might be particularly vulnerable to developing the condition themselves, with studies suggesting that this group is more likely than people with a paternal history to demonstrate patterns of brain structural and metabolic change in middle-age that are similar to those observed in probable AD. Thus, the identification of cognitive markers of liability to probable AD in offspring of probable AD patients, and in particular, those with a maternal history of the condition, has the potential to assist in the early identification of the disease and, if these markers turn out to be endophenotypic in probable AD, it might also lead to the identification of putative susceptibility genes for the condition. The research had two major aims. The first aim was to summarise the limited data on the cognitive functioning of people with and without a parental history of probable AD, and determine whether conventional between-group analyses could reveal differences in cognitive functioning between these two groups. To this end, the investigator conducted a meta-analysis of 17 relevant studies in literature, and examined the magnitude of cognitive impairment across several domains, with a particular focus on episodic memory (Study 1). As expected, effect sizes for all cognitive domains were small, and none were significantly different to zero. The largest difference between groups was found for visuospatial ability (d = 0.16), with offspring of probable AD patients outperforming controls. This was followed by verbal ability (d = -0.10) and episodic memory (d = -0.09), with offspring performing worse than controls. Effect sizes for all other cognitive domains, including global cognition (d = 0), executive function (d = 0), attention and concentration (d = 0), and perceptual speed (d = -0.04) were essentially zero. The results suggested that differences in cognitive function between high-and low-risk groups were present, but subtle, and raised the possibility that the use of more sensitive memory tasks and analytical techniques, could reveal more pronounced deficits in memory function in the offspring group. The second aim of the research was to determine whether differences in arbitrary associative memory function were detectable in middle-aged people with a parental history of probable AD, using conventional between-group analyses and novel cognitive discrepancy analyses. Arbitrary associative memory function was examined for four reasons: First, the arbitrary associative memory system is responsible for the rapid uptake of novel relationships. This function is considered crucial for all new learning and memory in humans. Second, arbitrary associative memory is heavily reliant on the medial temporal lobe – a region that is affected in the earliest stages of probable AD. Third, measures of arbitrary associative cued-recall are sensitive to the earliest changes in cognition in amnestic mild cognitive impairment and probable AD, but have not been utilised in parental AD. Finally, the arbitrary associative recognition literature suggests that the ability to recognise that two similar items belong together, or “within-domain arbitrary associative recognition”, is heavily dependent on the rhinal cortex. The rhinal cortex, in turn, is the site of the earliest neuropathological change in AD. The ability to recognise that two dissimilar items belong together, or “between-domain arbitrary associative recognition”, is more dependent on the hippocampus. The hippocampus is affected at a later stage in AD. The investigator reasoned that within-domain and between-domain arbitrary associative recognition function could be differentially affected in the parental AD group and conducted two studies to investigate this hypothesis. Based on the staging of neuropathological and structural change in probable AD, it was hypothesised that a subset of people with a parental history of the condition would perform more poorly on within-domain arbitrary associative recognition tasks, than on between-domain recognition tasks. In Study 2, 25 cognitively and neurologically normal people, aged between 45 and 55 years, completed four novel arbitrary associative recognition tasks, developed by the investigator in a series of pilot studies. They had no first-degree family history of any dementia. Two within-domain arbitrary associative recognition tasks (i.e., word-word, face-face), and two between-domain arbitrary associative recognition tasks (i.e., word-face, pattern-word) were developed. The purpose of the study was to examine the difficulty of the recognition tasks. The results demonstrated that the tasks did not suffer from floor or ceiling effects. Discrimination scores for the word-word (d’ = 1.81 ± 0.62) and pattern-word tasks (d’ = 1.58 ± 0.56) were not significantly different, but both were significantly higher than scores for the face-face (d’ = 0.64 ± 0.39) and word-face (d’ = 0.82 ± 0.41) tasks. Discrimination scores for the latter two tasks were also not significantly different. The results of Study 2 meant that discrepancies in within-domain and between-domain arbitrary associative recognition could be examined at “easy” and “hard” difficulty levels in people with a parental history of probable AD in Study 3. The word-word and pattern-word tasks comprised the “easy” cognitive discrepancy, while the face-face and word-face tasks comprised the “hard” cognitive discrepancy. Twenty three people with only a parental history of probable AD, and 30 people without a first-degree family history of any dementia, completed a range of clinical memory tests (e.g., California Verbal Learning Test-II), as well as arbitrary associative cued-recall (e.g., Paired Associate Learning Test of the Cambridge Neuropsychological Test Automated Battery) and arbitrary associative recognition tasks (e.g., novel word-word face-face, word-face, and pattern-word tasks). As predicted, mean scores on the clinical memory tasks were compared between groups, and no significant differences were found; however, contrary to predictions, offspring did not perform worse than controls on measures of arbitrary associative cued-recall and recognition. The hypothesis that group membership (offspring, control) and apolipoprotein epsilon 4 status (positive, negative) would not interact to affect memory test scores was partially supported, with significant interactions found only for average discrimination scores on the Face-Face and Word-Face tasks. No between-group differences were found on memory test scores when the parental AD group was broken down into maternal AD and paternal AD subgroups. Cognitive discrepancy analyses were conducted on z-scores for the arbitrary associative recognition tasks. It was hypothesised that, when z-scores on each within-domain and between-domain task were compared within groups (e.g., word-word minus pattern-word; face-face minus word-face), greater-than-expected percentages of offspring would perform more poorly on the within-domain tasks, than on the between-domain tasks – i.e., display negative cognitive discrepancy scores or negative cognitive profiles. Controls were expected to perform similarly on the within-domain and between-domain arbitrary associative recognition tasks – i.e., display neutral discrepancy scores. The hypotheses were partially supported: while associations between group (offspring and control) and profile type were not statistically significant, the percentages of offspring who displayed negative cognitive profiles were greater than expected by chance alone. This pattern was found for most indices of within-domain and between-domain arbitrary associative recognition function. Moreover, as predicted, the results were largely driven by the maternal AD group, with greater-than-expected percentages of these people displaying negative discrepancy scores. When task difficulty was ignored and z-scores for composite within-domain (e.g., word-word, face-face) and between-domain (e.g., pattern-word, word-face) variables were compared, the association between parental history type (maternal AD, paternal AD, control) and profile type (negative, positive) was statistically significant. By contrast, controls were consistently more likely to display neutral or positive discrepancy scores. The current research challenges the notion that differences in cognitive function are not detectable in middle-aged people at risk of probable AD and raises the possibility that deficits in arbitrary associative memory are endophenotypic in the condition. It is hoped that the knowledge gained from the present research will add to the growing body of literature attempting to assist in the early detection of probable AD and will encourage other researchers to search for cognitive endophenotypes for the condition, as it might provide some clues about its underlying neurobiological and genetic mechanisms.
-
ItemThe effect of vascular risk factors in dementia of the Alzheimer's type( 2016)The influence of vascular risk factors on Alzheimer’s disease and its clinical presentation is an area of intense scientific debate. There is now growing evidence to suggest that the presence of vascular risk factors in early-stage Alzheimer’s disease might accelerate its clinical presentation. Previous studies have also demonstrated that clinically healthy individuals with vascular risk factors perform below expectations on cognitive assessment. Further, the burden of vascular risk factors might be a predictor of future cognitive decline. However, the relationship between vascular risk factors and cognitive deterioration in healthy older adults has not been clearly established. The aim of this study was to investigate the relationship between vascular risk factors and cognitive function in a large cohort of 768 cognitively healthy older individuals drawn from the Australian Imaging, Biomarkers and Lifestyle (AIBL) Study of Ageing. This thesis addresses four main questions: (1) is there a difference in cognitive performance between individuals with and without vascular risk factors?, (2) does the presence of vascular risk factors affect the rate of cognitive change? (3) does the number of vascular risk factors influence cognition? and (4) does the presence of vascular risk factors increase the likelihood of developing Mild Cognitive Impairment? Participants were aged between 60–95 years, (M= 69.9 ± 6.9). All participants underwent a comprehensive baseline assessment and three additional serial assessments were repeated at 18-month intervals over a 54-month period. By the end of the study period, 172 individuals had withdrawn from the cohort. Of the remaining 596 participants, 43 had attracted a classification of Mild Cognitive Impairment. The sample was divided into subgroups of those with and without vascular risk factors. Participants in the vascular risk factors sub-group had at least one of the following risk factors: (1) hypertension, (2) diabetes, (3) dyslipidaemia, (4) Body Mass Index > 30 kg/m², (5) chronic kidney disease, (6) smoking history of more than 20 cigarettes per day for more than one year, and (7) elevated homocysteine levels (males>16.2 μmol/L; females>13.6 μmol/L). This thesis comprises a series of analyses. For the first analysis, demographic characteristics of individuals with and without vascular risk factors was compared. The data suggested that vascular risk factors were associated with older age (F (1, 712) = 8.1, p = 0.005), lower level of education (χ2 (1, n =711) = 7.8, p = 0.005) and apolipoprotein E ε4 carriage (χ2 (1, n = 714) =6.9, p = 0.009). In the second analysis, a statistically significant difference in cognitive performance between individuals with and without vascular risk factors was identified (F (4, 669) = 4.28, p = 0.002; partial eta square = 0.03). The results revealed that participants with vascular risk factors demonstrated poorer performance on variables that assessed visual cognition and executive functions than participants without vascular risk factors. In the third analysis, from a longitudinal perspective, five specific neuropsychological tests were subjected to a linear mixed model adjusted for age and apolipoprotein E ε4 carriage. The presence of vascular risk factors was associated with increased rates of cognitive decline on measures of verbal fluency (F (2, 1096.8) = 3.03, p < 0.05) and visuospatial skills (F (1204.6, 1) = 1.055; p < 0.05). The relationship between vascular risk factors and cognitive decline became more evident in the fourth analysis when the neuropsychological test battery was reduced into cognitive composite measures. Participants with vascular risk factors showed significantly greater decline on measures of verbal (F (624.016, 1) = 7.2; p = 0.01) and visual (F (1775.533, 1) = 6.89; p = 0.01) memory, as well as on visuospatial skills (F (1204.6, 1) = 1.055; p = 0.03) and executive functions (F (1821.035, 1) = 4.48; p = 0.03). No difference was found in the language composite measure. In the fifth analysis, higher vascular risk factor burden was associated with increased rates of cognitive decline, suggesting the presence of a dose-response relationship. In individuals with three or more vascular risk factors, a higher magnitude of decline on tasks that measure executive functions (Cohen's d = 0.35), visuospatial skills (Cohen's d = 0.35) and visual memory (Cohen's d = 0.47) was identified. No dose-response relationship was found in either the language or verbal memory cognitive domains. The sixth analysis focused on the group of participants who attracted a diagnosis of MCI over the 54 months. The data showed that there was a significant difference in the vascular risk factor burden distribution between individuals who remain cognitively stable and those who transition to Mild Cognitive Impairment (X² (2) = 44.88, p < 0.001), with greater risk factor burden observed in the latter. The interaction effect between clinical classification and presence/absence of vascular risk factors on the cognitive composite factors was further investigated. The only cognitive variable that showed a statistically significant difference was the visuospatial skill composite (1, 513) = 5.52; p = 0.019; partial eta square = 0.011). Finally, the analysis showed that vascular risk factors increased the likelihood of developing Mild Cognitive Impairment by 39%. In conclusion, the data suggest that vascular risk factors confer a degree of cognitive vulnerability on cognitively healthy older adults. This effect appears to be selective, impacting executive functions and visuospatial cognition but not language and verbal memory. Executive functions and visuospatial cognition are complex cognitive domains and share extensive distributed networks, which may increase their susceptibility to vascular risk factors over time. Based on the overall findings of this research, we propose that vascular risk factors – in combination with genetic and demographic factors – contribute to a reduction of brain reserve, precipitating an earlier onset of cognitive decline in the transition to Mild Cognitive Impairment. This implies that individuals who suffer from several risk factors may have less brain reserve and may therefore be more susceptible to developing dementia. Vascular risk factors are preventable for at least 95% of people with relatively simple changes in diet and lifestyle. The findings from this thesis contribute to a developing literature suggesting that decreases in vascular risk may reduce susceptibility for dementia and cognitive decline in later life.
-
ItemCognitively relevant networks in Alzheimer's disease: the preferential contribution of amyloid-β and hyperphosphorylated tau( 2015)Alzheimer's disease (AD) is characterised by two cardinal pathologies, namely the extracellular accumulation amyloid-related aggregates, and the intracellular formation of tau-related neurofibrillary tangles. Both of these pathologies disrupt normal neuronal function, but develop according to different anatomical trajectories. While cerebral tau-pathology first forms in mesial temporal structures, amyloid-pathology first accumulates in neocortex. The eventual distribution of these pathologies reflects their origins. The mesial temporal structures are the most heavily affected by tau-pathology while amyloid-pathology is most prominently deposited in neocortical regions. This raises the possibility that the cardinal pathologies preferentially disrupt different brain networks, resulting in differential contributions to the cognitive expression of AD. The present inquiry aimed to clarify the relationship between the cardinal pathologies and the disruption of cognitively relevant networks. It was predicted that tau-pathology would preferentially disrupt mesial temporal networks associated with fundamental memory function. In contrast, it was expected that amyloid-pathology would be preferentially associated with neocortical, particularly anterior, networks associated with the executive functions. The first study investigated the preferential association hypothesis in 191 patients with mild cognitive impairment. As expected, CSF biomarkers of tau-pathology were most strongly associated with measures of memory function. The relationship between amyloid-pathology and memory was indirect, that is, mediated by tau-pathology. Amyloid-pathology had a separate, tau-independent, relationship with information processing speed. In the second study, the hypothesis was investigated in 39 patients with dementia of the Alzheimer's type using CSF biomarkers and functional connectivity analyses obtained from resting state functional magnetic resonance imaging. Tau-pathology was preferentially associated with functional connectivity in the mesial temporal regions. Amyloid-pathology was preferentially associated with functional connectivity in the dorsal anterior cingulate cortex, a region intimately connected to anterior neocortical structures. In the third study, the hypothesis was investigated in the same cohort using CSF biomarkers, structural magnetic resonance imaging, and fluorodeoxyglucose positron emission tomography (18F-FDG-PET). Amyloid-related pathology was preferentially related to the posterior cingulate and lateral temporal cortices, while no preferential relationships were found for tau-pathology. Taken together, these findings provide early support for the preferential effects of the cardinal pathologies. The results support the link between tau-pathology, mesial temporal structures, and fundamental memory impairment in AD. Less support was found for the association between amyloid-pathology, neocortical networks, and executive function. While amyloid-pathology was preferentially associated with structurofunctional changes in a number of neocortical regions, the predicted relationship with executive function was not strongly supported. The emerging view is that abnormal tau is a destructive, intracellular pathology with a proximal relationship to cognitive impairment in AD. In contrast, abnormal amyloid is a pan-cerebral, extracellular pathology that subtly disrupts diffusely represented networks. The present findings have implications for disease-modifying therapeutic trials in AD, arguing for the need to match neurocognitive and neuroimaging endpoints with the cardinal pathology being targeted by the treatment. Overall, this thesis opens up a new, and more complex, view of the cardinal pathologies and their relationship to cognitively relevant brain networks.
-
ItemMemory changes in community-based older adults: subjective memory complaints, personal memory and their relationship with Alzheimer’s disease biomarkers( 2015)Subjective concerns of memory decline (or subjective memory complaints) are a common phenomenon in the normal population, particularly in older adults. Complaints arise from a negative self-appraisal of memory performance, and form the bridge that connects an individual to clinical attention. The experience of memory failure, whether momentary, gradually progressive, or sudden and acute, is universal and inherently subjective. Within this experiential realm lies rich and complex subjective detail. A threshold exists at which point self-appraised memory dysfunction becomes intolerable, and most individuals seek diagnostic resolution for their symptoms. Alzheimer’s disease (AD) is the most common form of dementia, which is characterised by early impairment in objective memory performance, and gradual accrual of in vivo biological markers. The current trend in the literature is to assume a monotonic relationship between memory complaint severity and level of objective memory impairment, but in general studies have been unable to find a consistent association. The relationship between subjective memory complaining and biomarkers of AD, such as neocortical β-amyloid (Aβ) burden, measured via positron emission tomography (PET) imaging, and brain atrophy, measured via magnetic resonance imaging (MRI), is gathering pace. Brain atrophy in the medial temporal lobes (MTL) is consistently related to memory complaining, but the literature relating to neocortical Aβ burden is less clear; recent studies have shown a relationship, but the effect sizes are small. Autobiographical (ABM) and personal semantic memories (PSM), along with their autonoetic re-experiencing, form the foundation of everyday personal memory. An assessment of personal memory impairment does not feature in the mainstream diagnostic approach to mild cognitive impairment (MCI), the clinical transitional stage prior to a diagnosis of AD. ABM, or the highly contextualised details related to a personal event, are consistently found to be impaired in individuals with MCI, but the pattern of impairment of PSM is less understood. As the most ecologically valid form of memory, it is possible that a breakdown in ABM and PSM is the main driver for subsequent memory complaining. To the candidate’s knowledge, there is no literature which has yet examined a possible connection between memory complaining and personal memory function. It is also unclear how markers of AD pathology, such as Aβ burden and brain atrophy, may be related to ABM and PSM, but functional MRI studies of these personal memory systems suggest they are supported by differing functional networks. ABM is related to a core functional network that primarily involves the MTL, frontal, and posterior cingulate regions, while PSM has been attributed to networks in the lateral temporal lobe. The current inquiry is aimed at describing the phenomenological experience of memory change in non-demented older adults in order to elucidate complaints in those at greater risk of AD, such as individuals with evidence of biological markers of AD, or those with MCI. The specific objectives of this these were: 1) to determine cognitive, affective and AD biological marker correlates of SMCs in healthy older adults and individuals with MCI examined via a commonly-used subjective memory complaint questionnaire (SMCQ), 2) to characterise the phenomenological experiences of memory loss in the earliest stages of Alzheimer’s disease, such as healthy older adults with evidence of AD biological markers and those with MCI, and 3) to examine the pattern of ABM and PSM impairment and its relationship with AD biomarkers in healthy memory complainers and individuals with MCI. The first study (Chapter 4), investigated memory, affective and AD biomarker correlates of SMC severity as measured via a commonly-used SMCQ, the Memory Assessment Clinics Questionnaire (MAC-Q), in healthy older adults and individuals with MCI. Participants were healthy controls (n = 674) and individuals with MCI (n = 66) in the Australian Imaging, Biomarker and Lifestyle (AIBL) study of ageing, who responded to the MAC-Q. Scores from measures of new learning and retention, as well as mood (anxiety and depression), age, education and diagnostic category, were regressed onto the total MAC-Q score, to determine whether memory would relate to SMC severity over and above the other variables. Diagnostic category, depressive symptomatology, and age, were the strongest correlates. In healthy older adults, the strongest driver of greater MAC-Q score was depressive symptomatology, while in individuals with MCI, the sole correlate was age. Aβ deposition, grey matter volumes, and APOE ε4 carrier status were regressed onto MAC-Q total score in both diagnostic groups, but no correlates were found. A strong relationship between depression and memory complaint severity in healthy older adults aligns with other studies using SMCQs. The relationship with age in MCI supports the notion of an underlying organic driver of memory complaining, and a transition from affectively-driven phenomena. The second study (Chapter 5), addressed the characterisation of SMCs in different AD-risk groups, such as healthy older adults with high Aβ load, healthy memory complainers (identified by asking “do you have difficulties with your memory, yes or no?”), and individuals with MCI. Healthy control (n = 80) and MCI (n = 43) participants were randomly recruited from the AIBL study for the next three studies. A semi-structured interview was developed by the candidate, in conjunction with supervisor, MS to probe the phenomenological experience of memory loss. A thematic analysis was conducted and meaningful phrases were extracted, and grouped similar phrases into what ultimately became twelve themes. A comparison between healthy individuals with high and low Aβ load showed that greater Aβ deposition was related to greater acknowledgement of a progressive memory decline. Complaint themes in healthy memory complainers closely aligned to those expressed by MCI individuals, suggesting that a single SMC question has the potential to uncover individuals with subjectively similar experiences of memory loss as those with a diagnosed memory dysfunction. Individuals with MCI diverged from healthy memory complainers in their greater acknowledgement of dependency on a significant other, and the implementation of burdensome coping strategies. Thus, an increasing acknowledgement of dysfunction in daily activities may well signify an interim early stage outcome of accumulating memory problems. The pattern of ABM and PSM impairment in healthy memory complainers and those with MCI was examined in the third study (Chapter 6). The Episodic Autobiographical Memory Interview, a validated semi-structured interview, measured the spatiotemporal and emotional details of a recent personal event (ABM) and personally factual knowledge from the recent past (PSM). Individuals with MCI showed a deficit in both ABM and PSM, supporting conceptual notions of a dynamic and interactive overlap between these two memory systems. Healthy memory complainers did not show a deficit compared to non-complaining healthy older adults, raising the question as to whether current measures of memory complaining are sensitive to the subjective experience of memory decline. In the fourth and final study, AD biomarkers were considered in relation to personal memory performance, namely ABM, PSM and the conscious re-experiencing connected with ABM (Chapter 7). Aβ deposition, grey matter volumes and APOE ε4 carrier status were regressed onto each form of personal memory. PSM was negatively associated with neocortical Aβ burden in healthy older adults, but neither ABM nor autonoetic consciousness were related to any AD biomarkers. ABM impairment exerted a large magnitude of effect on MCI unlike PSM, so there are potentially other underlying neurodegenerative processes affecting ABM performance. The prevailing assumption in the literature is to treat subjective memory complaining as a unidimensional proxy for neuropsychological measures of memory impairment. The problem with this assumption is that it does not recognise the complex, and often counter-intuitive, nature of subjective experiences. SMCQs, such as the MAC-Q, are structured as screening tools in that they are non-interactive, non-specific and designed for quick self-administration. This approach provides no diagnostic information over and above a quantified score, and neglects the possibility that phenomenological experiences contain important clinical information. The impairment to both ABM and PSM in MCI, highlighting that both personal memory systems are affected by underlying AD-related neuropathology. Specifically, findings from this thesis suggest that ABM and PSM are affected by differential disease mechanisms, with PSM strongly influenced by neocortical Aβ burden in healthy older adults, and ABM showing a trend towards association with hippocampal volume. Thus, elevated AD biomarkers were found to have an effect on the subjective experience (via complaints) and expression (via autobiographical narratives) of memory, supporting the notion that these phenomena are affected in the early stages of AD. This thesis highlights the subjective experience of memory change in non-demented older adults, which will help to inform future approaches to the clinical evaluation of memory complaints prior to diagnosis.