A Bayesian semiparametric approach for inference on the population partly conditional mean from longitudinal data with dropout.

Studies of memory trajectories using longitudinal data often result in highly nonrepresentative samples due to selective study enrollment and attrition. An additional bias comes from practice effects that result in improved or maintained performance due to familiarity with test content or context. These challenges may bias study findings and severely distort the ability to generalize to the target population. In this study, we propose an approach for estimating the finite population mean of a longitudinal outcome conditioning on being alive at a specific time point. We develop a flexible Bayesian semiparametric predictive estimator for population inference when longitudinal auxiliary information is known for the target population. We evaluate the sensitivity of the results to untestable assumptions and further compare our approach to other methods used for population inference in a simulation study. The proposed approach is motivated by 15-year longitudinal data from the Betula longitudinal cohort study. We apply our approach to estimate lifespan trajectories in episodic memory, with the aim to generalize findings to a target population.

Individual differences in brain aging: heterogeneity in cortico-hippocampal but not caudate atrophy rates.

It is well documented that some brain regions, such as association cortices, caudate, and hippocampus, are particularly prone to age-related atrophy, but it has been hypothesized that there are individual differences in atrophy profiles. Here, we document heterogeneity in regional-atrophy patterns using latent-profile analysis of 1,482 longitudinal magnetic resonance imaging observations. The results supported a 2-group solution reflecting differences in atrophy rates in cortical regions and hippocampus along with comparable caudate atrophy. The higher-atrophy group had the most marked atrophy in hippocampus and also lower episodic memory, and their normal caudate atrophy rate was accompanied by larger baseline volumes. Our findings support and refine models of heterogeneity in brain aging and suggest distinct mechanisms of atrophy in striatal versus hippocampal-cortical systems.

Educational attainment does not influence brain aging.

Education has been related to various advantageous lifetime outcomes. Here, using longitudinal structural MRI data (4,422 observations), we tested the influential hypothesis that higher education translates into slower rates of brain aging. Cross-sectionally, education was modestly associated with regional cortical volume. However, despite marked mean atrophy in the cortex and hippocampus, education did not influence rates of change. The results were replicated across two independent samples. Our findings challenge the view that higher education slows brain aging.

Education and Income Show Heterogeneous Relationships to Lifespan Brain and Cognitive Differences Across European and US Cohorts.

Higher socio-economic status (SES) has been proposed to have facilitating and protective effects on brain and cognition. We ask whether relationships between SES, brain volumes and cognitive ability differ across cohorts, by age and national origin. European and US cohorts covering the lifespan were studied (4-97 years, N = 500 000; 54 000 w/brain imaging). There was substantial heterogeneity across cohorts for all associations. Education was positively related to intracranial (ICV) and total gray matter (GM) volume. Income was related to ICV, but not GM. We did not observe reliable differences in associations as a function of age. SES was more strongly related to brain and cognition in US than European cohorts. Sample representativity varies, and this study cannot identify mechanisms underlying differences in associations across cohorts. Differences in neuroanatomical volumes partially explained SES-cognition relationships. SES was more strongly related to ICV than to GM, implying that SES-cognition relations in adulthood are less likely grounded in neuroprotective effects on GM volume in aging. The relatively stronger SES-ICV associations rather are compatible with SES-brain volume relationships being established early in life, as ICV stabilizes in childhood. The findings underscore that SES has no uniform association with, or impact on, brain and cognition.

Memory profiles predict dementia over 23-28 years in normal but not successful aging.

OBJECTIVES: Prospective studies suggest that memory deficits are detectable decades before clinical symptoms of dementia emerge. However, individual differences in long-term memory trajectories prior to diagnosis need to be further elucidated. The aim of the current study was to investigate long-term dementia and mortality risk for individuals with different memory trajectory profiles in a well-characterized population-based sample. METHODS: 1062 adults (aged 45-80 years) who were non-demented at baseline were followed over 23-28 years. Dementia and mortality risk were studied for three previously classified episodic memory trajectory groups: maintained high performance (Maintainers; 26%), average decline (Averages; 64%), and accelerated decline (Decliners; 12%), using multistate modeling to characterize individuals' transitions from an initial non-demented state, possibly to a state of dementia and/or death. RESULTS: The memory groups showed considerable intergroup variability in memory profiles, starting 10-15 years prior to dementia diagnosis, and prior to death. A strong relationship between memory trajectory group and dementia risk was found. Specifically, Decliners had more than a fourfold risk of developing dementia compared to Averages. In contrast, Maintainers had a 2.6 times decreased dementia risk compared to Averages, and in addition showed no detectable memory decline prior to dementia diagnosis. A similar pattern of association was found for the memory groups and mortality risk, although only among non-demented. CONCLUSION: There was a strong relationship between accelerated memory decline and dementia, further supporting the prognostic value of memory decline. The intergroup differences, however, suggest that mechanisms involved in successful memory aging may delay symptom onset.

Poor Self-Reported Sleep is Related to Regional Cortical Thinning in Aging but not Memory Decline-Results From the Lifebrain Consortium.

We examined whether sleep quality and quantity are associated with cortical and memory changes in cognitively healthy participants across the adult lifespan. Associations between self-reported sleep parameters (Pittsburgh Sleep Quality Index, PSQI) and longitudinal cortical change were tested using five samples from the Lifebrain consortium (n = 2205, 4363 MRIs, 18-92 years). In additional analyses, we tested coherence with cell-specific gene expression maps from the Allen Human Brain Atlas, and relations to changes in memory performance. "PSQI # 1 Subjective sleep quality" and "PSQI #5 Sleep disturbances" were related to thinning of the right lateral temporal cortex, with lower quality and more disturbances being associated with faster thinning. The association with "PSQI #5 Sleep disturbances" emerged after 60 years, especially in regions with high expression of genes related to oligodendrocytes and S1 pyramidal neurons. None of the sleep scales were related to a longitudinal change in episodic memory function, suggesting that sleep-related cortical changes were independent of cognitive decline. The relationship to cortical brain change suggests that self-reported sleep parameters are relevant in lifespan studies, but small effect sizes indicate that self-reported sleep is not a good biomarker of general cortical degeneration in healthy older adults.

Successful Memory Aging.

For more than 50 years, psychologists, gerontologists, and, more recently, neuroscientists have considered the possibility of successful aging. How to define successful aging remains debated, but well-preserved age-sensitive cognitive functions, like episodic memory, is an often-suggested criterion. Evidence for successful memory aging comes from cross-sectional and longitudinal studies showing that some older individuals display high and stable levels of performance. Successful memory aging may be accomplished via multiple paths. One path is through brain maintenance, or relative lack of age-related brain pathology. Through another path, successful memory aging can be accomplished despite brain pathology by means of efficient compensatory and strategic processes. Genetic, epigenetic, and lifestyle factors influence memory aging via both paths. Some of these factors can be promoted throughout the life course, which, at the individual as well as the societal level, can positively impact successful memory aging.

Longitudinal Evidence for Increased Functional Response in Frontal Cortex for Older Adults with Hippocampal Atrophy and Memory Decline.

The functional organization of the frontal cortex is dynamic. Age-related increases in frontal functional responses have been shown during various cognitive tasks, but the cross-sectional nature of most past studies makes it unclear whether these increases reflect reorganization or stable individual differences. Here, we followed 130 older individuals' cognitive trajectories over 20-25 years with repeated neuropsychological assessments every 5th year, and identified individuals with stable or declining episodic memory. Both groups displayed significant gray matter atrophy over 2 successive magnetic resonance imaging sessions 4 years apart, but the decline group also had a smaller volume of the right hippocampus. Only individuals with declining memory demonstrated increased prefrontal functional responses during memory encoding and retrieval over the 4-year interval. Regions with increased functional recruitment were located outside, or on the borders of core task-related networks, indicating an expansion of these over time. These longitudinal findings offer novel insight into the mechanisms behind age-associated memory loss, and are consistent with a theoretical model in which hippocampus atrophy, past a critical threshold, induces episodic-memory decline and altered prefrontal functional organization.

Elevated hippocampal resting-state connectivity underlies deficient neurocognitive function in aging.

The brain is not idle during rest. Functional MRI (fMRI) studies have identified several resting-state networks, including the default mode network (DMN), which contains a set of cortical regions that interact with a hippocampus (HC) subsystem. Age-related alterations in the functional architecture of the DMN and HC may influence memory functions and possibly constitute a sensitive biomarker of forthcoming memory deficits. However, the exact form of DMN-HC alterations in aging and concomitant memory deficits is largely unknown. Here, using both task and resting data from 339 participants (25-80 y old), we have demonstrated age-related decrements in resting-state functional connectivity across most parts of the DMN, except for the HC network for which age-related elevation of connectivity between left and right HC was found along with attenuated HC-cortical connectivity. Elevated HC connectivity at rest, which was partly accounted for by age-related decline in white matter integrity of the fornix, was associated with lower cross-sectional episodic memory performance and declining longitudinal memory performance over 20 y. Additionally, elevated HC connectivity at rest was associated with reduced HC neural recruitment and HC-cortical connectivity during active memory encoding, which suggests that strong HC connectivity restricts the degree to which the HC interacts with other brain regions during active memory processing revealed by task fMRI. Collectively, our findings suggest a model in which age-related disruption in cortico-hippocampal functional connectivity leads to a more functionally isolated HC at rest, which translates into aberrant hippocampal decoupling and deficits during mnemonic processing.

Brain characteristics of individuals resisting age-related cognitive decline over two decades.

Some elderly appear to resist age-related decline in cognitive functions, but the neural correlates of successful cognitive aging are not well known. Here, older human participants from a longitudinal study were classified as successful or average relative to the mean attrition-corrected cognitive development across 15-20 years in a population-based sample (n = 1561). Fifty-one successful elderly and 51 age-matched average elderly (mean age: 68.8 years) underwent functional magnetic resonance imaging while performing an episodic memory face-name paired-associates task. Successful older participants had higher BOLD signal during encoding than average participants, notably in the bilateral PFC and the left hippocampus (HC). The HC activation of the average, but not the successful, older group was lower than that of a young reference group (n = 45, mean age: 35.3 years). HC activation was correlated with task performance, thus likely contributing to the superior memory performance of successful older participants. The frontal BOLD response pattern might reflect individual differences present from young age. Additional analyses confirmed that both the initial cognitive level and the slope of cognitive change across the longitudinal measurement period contributed to the observed group differences in BOLD signal. Further, the differences between the older groups could not be accounted for by differences in brain structure. The current results suggest that one mechanism behind successful cognitive aging might be preservation of HC function combined with a high frontal responsivity. These findings highlight sources for heterogeneity in cognitive aging and may hold useful information for cognitive intervention studies.

Age-related and genetic modulation of frontal cortex efficiency.

The dorsolateral pFC (DLPFC) is a key region for working memory. It has been proposed that the DLPFC is dynamically recruited depending on task demands. By this view, high DLPFC recruitment for low-demanding tasks along with weak DLPFC upregulation at higher task demands reflects low efficiency. Here, the fMRI BOLD signal during working memory maintenance and manipulation was examined in relation to aging and catechol-O-methyltransferase (COMT) Val(158)Met status in a large representative sample (n = 287). The efficiency hypothesis predicts a weaker DLPFC response during manipulation, along with a stronger response during maintenance for older adults and COMT Val carriers compared with younger adults and COMT Met carriers. Consistent with the hypothesis, younger adults and met carriers showed maximal DLPFC BOLD response during manipulation, whereas older adults and val carriers displayed elevated DLPFC responses during the less demanding maintenance condition. The observed inverted relations support a link between dopamine and DLPFC efficiency.

Revised Temperament and Character Inventory factors predict neuropsychiatric symptoms and aging-related cognitive decline across 25 years.

Introduction: Personality traits and neuropsychiatric symptoms such as neuroticism and depression share genetic overlap and have both been identified as risks factors for development of aging-related neurocognitive decline and Alzheimer's disease (AD). This study aimed to examine revised personality factors derived from the Temperament and Character Inventory, previously shown to be associated with psychiatric disorders, as predictors of neuropsychiatric, cognitive, and brain trajectories of participants from a population-based aging study. Methods: Mixed-effect linear regression analyses were conducted on data for the full sample (Nmax = 1,286), and a healthy subsample not converting to AD-dementia during 25-year follow-up (Nmax = 1,145), complemented with Cox proportional regression models to determine risk factors for conversion to clinical AD. Results: Two personality factors, Closeness to Experience (CE: avoidance of new stimuli, high anxiety, pessimistic anticipation, low reward seeking) and Tendence to Liabilities (TL: inability to change, low autonomy, unaware of the value of their existence) were associated with higher levels of depressive symptoms, stress (CE), sleep disturbance (TL), as well as greater decline in memory, vocabulary and verbal fluency in the full sample. Higher CE was additionally associated with greater memory decline across 25 years in the healthy subsample, and faster right hippocampal volume reduction across 8 years in a neuroimaging subsample (N = 216). Most, but not all, personality-cognition associations persisted after controlling for diabetes, hypertension and cardiovascular disease. Concerning risks for conversion to AD, higher age, and APOE-ε4, but none of the personality measures, were significant predictors. Conclusion: The results indicate that personality traits associated with psychiatric symptoms predict accelerated age-related neurocognitive declines even in the absence of neurodegenerative disease. The attenuation of some personality effects on cognition after adjustment for health indicators suggests that those effects may be partly mediated by somatic health. Taken together, the results further emphasize the importance of personality traits in neurocognitive aging and underscore the need for an integrative (biopsychosocial) perspective of normal and pathological age-related cognitive decline.

Longitudinal structure-function correlates in elderly reveal MTL dysfunction with cognitive decline.

By integrating behavioral measures and imaging data, previous investigations have explored the relationship between biological markers of aging and cognitive functions. Evidence from functional and structural neuroimaging has revealed that hippocampal volume and activation patterns in the medial temporal lobe (MTL) may predict cognitive performance in old age. Most past demonstrations of age-related differences in brain structure-function were based on cross-sectional comparisons. Here, the relationship between 6-year intraindividual change in functional magnetic resonance imaging (fMRI) signal and change in memory performance over 2 decades was examined. Correlations between intraindividual change in fMRI signal during episodic encoding and change in memory performance measured outside of scanning were used as an estimate for relating brain-behavior changes. The results revealed a positive relationship between activation change in the hippocampus (HC) and change in memory performance, reflecting reduced hippocampal activation in participants with declining performance. Using a similar analytic approach as for the functional data, we found that individuals with declining performance had reduced HC volume compared with individuals with intact performance. These observations provide a strong link between cognitive change in older adults and MTL structure and function and thus provide insights into brain correlates of individual variability in aging trajectories.

Longitudinal evidence for diminished frontal cortex function in aging.

Cross-sectional estimates of age-related changes in brain structure and function were compared with 6-y longitudinal estimates. The results indicated increased sensitivity of the longitudinal approach as well as qualitative differences. Critically, the cross-sectional analyses were suggestive of age-related frontal overrecruitment, whereas the longitudinal analyses revealed frontal underrecruitment with advancing age. The cross-sectional observation of overrecruitment reflected a select elderly sample. However, when followed over time, this sample showed reduced frontal recruitment. These findings dispute inferences of true age changes on the basis of age differences, hence challenging some contemporary models of neurocognitive aging, and demonstrate age-related decline in frontal brain volume as well as functional response.

Explaining graph convolutional network predictions for clinicians-An explainable AI approach to Alzheimer's disease classification.

Introduction: Graph-based representations are becoming more common in the medical domain, where each node defines a patient, and the edges signify associations between patients, relating individuals with disease and symptoms in a node classification task. In this study, a Graph Convolutional Networks (GCN) model was utilized to capture differences in neurocognitive, genetic, and brain atrophy patterns that can predict cognitive status, ranging from Normal Cognition (NC) to Mild Cognitive Impairment (MCI) and Alzheimer's Disease (AD), on the Alzheimer's Disease Neuroimaging Initiative (ADNI) database. Elucidating model predictions is vital in medical applications to promote clinical adoption and establish physician trust. Therefore, we introduce a decomposition-based explanation method for individual patient classification. Methods: Our method involves analyzing the output variations resulting from decomposing input values, which allows us to determine the degree of impact on the prediction. Through this process, we gain insight into how each feature from various modalities, both at the individual and group levels, contributes to the diagnostic result. Given that graph data contains critical information in edges, we studied relational data by silencing all the edges of a particular class, thereby obtaining explanations at the neighborhood level. Results: Our functional evaluation showed that the explanations remain stable with minor changes in input values, specifically for edge weights exceeding 0.80. Additionally, our comparative analysis against SHAP values yielded comparable results with significantly reduced computational time. To further validate the model's explanations, we conducted a survey study with 11 domain experts. The majority (71%) of the responses confirmed the correctness of the explanations, with a rating of above six on a 10-point scale for the understandability of the explanations. Discussion: Strategies to overcome perceived limitations, such as the GCN's overreliance on demographic information, were discussed to facilitate future adoption into clinical practice and gain clinicians' trust as a diagnostic decision support system.

Sixteen-Year Longitudinal Evaluation of Blood-Based DNA Methylation Biomarkers for Early Prediction of Alzheimer's Disease.

BACKGROUND: DNA methylation (DNAm), an epigenetic mark reflecting both inherited and environmental influences, has shown promise for Alzheimer's disease (AD) prediction. OBJECTIVE: Testing long-term predictive ability (>15 years) of existing DNAm-based epigenetic age acceleration (EAA) measures and identifying novel early blood-based DNAm AD-prediction biomarkers. METHODS: EAA measures calculated from Illumina EPIC data from blood were tested with linear mixed-effects models (LMMs) in a longitudinal case-control sample (50 late-onset AD cases; 51 matched controls) with prospective data up to 16 years before clinical onset, and post-onset follow-up. Novel DNAm biomarkers were generated with epigenome-wide LMMs, and Sparse Partial Least Squares Discriminant Analysis applied at pre- (10-16 years), and post-AD-onset time-points. RESULTS: EAA did not differentiate cases from controls during the follow-up time (p > 0.05). Three new DNA biomarkers showed in-sample predictive ability on average 8 years pre-onset, after adjustment for age, sex, and white blood cell proportions (p-values: 0.022-<0.00001). Our longitudinally-derived panel replicated nominally (p = 0.012) in an external cohort (n = 146 cases, 324 controls). However, its effect size and discriminatory accuracy were limited compared to APOEɛ4-carriership (OR = 1.38 per 1 SD DNAm score increase versus OR = 13.58 for ɛ4-allele carriage; AUCs = 77.2% versus 87.0%). Literature review showed low overlap (n = 4) across 3275 AD-associated CpGs from 8 published studies, and no overlap with our identified CpGs.

No Association Between Loneliness, Episodic Memory and Hippocampal Volume Change in Young and Healthy Older Adults: A Longitudinal European Multicenter Study.

Background: Loneliness is most prevalent during adolescence and late life and has been associated with mental health disorders as well as with cognitive decline during aging. Associations between longitudinal measures of loneliness and verbal episodic memory and brain structure should thus be investigated. Methods: We sought to determine associations between loneliness and verbal episodic memory as well as loneliness and hippocampal volume trajectories across three longitudinal cohorts within the Lifebrain Consortium, including children, adolescents (N = 69, age range 10-15 at baseline examination) and older adults (N = 1468 over 60). We also explored putative loneliness correlates of cortical thinning across the entire cortical mantle. Results: Loneliness was associated with worsening of verbal episodic memory in one cohort of older adults. Specifically, reporting medium to high levels of loneliness over time was related to significantly increased memory loss at follow-up examinations. The significance of the loneliness-memory change association was lost when eight participants were excluded after having developed dementia in any of the subsequent follow-up assessments. No significant structural brain correlates of loneliness were found, neither hippocampal volume change nor cortical thinning. Conclusion: In the present longitudinal European multicenter study, the association between loneliness and episodic memory was mainly driven by individuals exhibiting progressive cognitive decline, which reinforces previous findings associating loneliness with cognitive impairment and dementia.

Healthy Middle-Aged Adults Have Preserved Mnemonic Discrimination and Integration, While Showing No Detectable Memory Benefits.

Declarative memory abilities change across adulthood. Semantic memory and autobiographic episodic knowledge can remain stable or even increase from mid- to late adulthood, while episodic memory abilities decline in later adulthood. Although it is well known that prior knowledge influences new learning, it is unclear whether the experiential growth of knowledge and memory traces across the lifespan may drive favorable adaptations in some basic memory processes. We hypothesized that an increased reliance on memory integration may be an adaptive mechanism to handle increased interference from accumulating memory traces and knowledge across adulthood. In turn, this may confer an improved ability for integration, observable in middle-age, before the onset of major aging-related declines. We further tested whether the hypothesized increase would be associated with previously observed reductions in memory discrimination performance in midlife. Data from a sample of healthy middle-aged (40-50 years, n = 40) and younger adults (20-28 years, n = 41) did not support the hypothesis of improved integration, as assessed by an associative inference paradigm. Instead, age-equivalent performance on both integration and discrimination measures were observed [Bayes factors (BFs)10 = 0.19-0.25], along with expected higher verbal knowledge and slower perceptual speed for middle-aged [(BFs)10 = 8.52-73.52]. The results contribute to an increased understanding of memory processing in midlife, an understudied portion of the lifespan, and suggest that two core episodic memory processes, integration and discrimination, can be maintained in healthy middle-aged adults.

Sex-specific effects of polygenic risk for schizophrenia on lifespan cognitive functioning in healthy individuals.

Polygenic risk for schizophrenia has been associated with lower cognitive ability and age-related cognitive change in healthy individuals. Despite well-established neuropsychological sex differences in schizophrenia patients, genetic studies on sex differences in schizophrenia in relation to cognitive phenotypes are scarce. Here, we investigated whether the effect of a polygenic risk score (PRS) for schizophrenia on childhood, midlife, and late-life cognitive function in healthy individuals is modified by sex, and if PRS is linked to accelerated cognitive decline. Using a longitudinal data set from healthy individuals aged 25-100 years (N = 1459) spanning a 25-year period, we found that PRS was associated with lower cognitive ability (episodic memory, semantic memory, visuospatial ability), but not with accelerated cognitive decline. A significant interaction effect between sex and PRS was seen on cognitive task performance, and sex-stratified analyses showed that the effect of PRS was male-specific. In a sub-sample, we observed a male-specific effect of the PRS on school performance at age 12 (N = 496). Our findings of sex-specific effects of schizophrenia genetics on cognitive functioning across the lifespan indicate that the effects of underlying disease genetics on cognitive functioning is dependent on biological processes that differ between the sexes.

Short Leukocyte Telomeres, But Not Telomere Attrition Rates, Predict Memory Decline in the 20-Year Longitudinal Betula Study.

Leukocyte telomere length (LTL) is a proposed biomarker for aging-related disorders, including cognitive decline and dementia. Long-term longitudinal studies measuring intra-individual changes in both LTL and cognitive outcomes are scarce, precluding strong conclusions about a potential aging-related relationship between LTL shortening and cognitive decline. This study investigated associations between baseline levels and longitudinal changes in LTL and memory performance across an up to 20-year follow-up in 880 dementia-free participants from a population-based study (mean baseline age: 56.8 years, range: 40-80; 52% female). Shorter baseline LTL significantly predicted subsequent memory decline (r = .34, 95% confidence interval: 0.06, 0.82), controlling for age, sex, and other relevant covariates. No significant associations were however observed between intra-individual changes in LTL and memory, neither concurrently nor with a 5-year time-lag between LTL shortening and memory decline. These results support the notion of short LTL as a predictive factor for aging-related memory decline, but suggest that LTL dynamics in adulthood and older age may be less informative of cognitive outcomes in aging. Furthermore, the results highlight the importance of long-term longitudinal evaluation of outcomes in biomarker research.