Search

Abstract Background Genome-wide association studies (GWAS) provide a powerful means to identify associations between genetic variants and phenotypes. However, GWAS techniques for detecting epistasis, the interactions between genetic variants associated with phenotypes, are still limited. We believe that developing an efficient and effective GWAS method to detect epistasis will be a key for discovering sophisticated pathogenesis, which is especially important for complex diseases such as Alzheimer’s disease (AD). Results In this regard, this study presents GenEpi, a computational package to uncover epistasis associated with phenotypes by the proposed machine learning approach. GenEpi identifies both within-gene and cross-gene epistasis through a two-stage modeling workflow. In both stages, GenEpi adopts two-element combinatorial encoding when producing features and constructs the prediction models by L1-regularized regression with stability selection. The simulated data showed that GenEpi outperforms other widely-used methods on detecting the ground-truth epistasis. As real data is concerned, this study uses AD as an example to reveal the capability of GenEpi in finding disease-related variants and variant interactions that show both biological meanings and predictive power. Conclusions The results on simulation data and AD demonstrated that GenEpi has the ability to detect the epistasis associated with phenotypes effectively and efficiently. The released package can be generalized to largely facilitate the studies of many complex diseases in the near future.

ABSTRACT Introduction Computational brain network modeling using The Virtual Brain (TVB) simulation platform acts synergistically with machine learning (ML) and multi‐modal neuroimaging to reveal mechanisms and improve diagnostics in Alzheimer's disease (AD). Methods We enhance large‐scale whole‐brain simulation in TVB with a cause‐and‐effect model linking local amyloid beta (Aβ) positron emission tomography (PET) with altered excitability. We use PET and magnetic resonance imaging (MRI) data from 33 participants of the Alzheimer's Disease Neuroimaging Initiative (ADNI3) combined with frequency compositions of TVB‐simulated local field potentials (LFP) for ML classification. Results The combination of empirical neuroimaging features and simulated LFPs significantly outperformed the classification accuracy of empirical data alone by about 10% (weighted F1‐score empirical 64.34% vs. combined 74.28%). Informative features showed high biological plausibility regarding the AD‐typical spatial distribution. Discussion The cause‐and‐effect implementation of local hyperexcitation caused by Aβ can improve the ML–driven classification of AD and demonstrates TVB's ability to decode information in empirical data using connectivity‐based brain simulation.

Abstract Introduction This study aimed to develop and validate a 3‐year dementia risk score in individuals with mild cognitive impairment (MCI) based on variables collected in routine clinical care. Methods The prediction score was trained and developed using data from the National Alzheimer's Coordinating Center (NACC). Selection criteria included aged 55 years and older with MCI. Cox models were validated externally using two independent cohorts from the Prospective Registry of Persons with Memory Symptoms (PROMPT) registry and the Alzheimer's Disease Neuroimaging Initiative (ADNI) database. Results Our Mild Cognitive Impairment to Dementia Risk (CIDER) score predicted dementia risk with c‐indices of 0.69 (95% confidence interval [CI] 0.66–0.72), 0.61 (95% CI 0.59–0.63), and 0.72 (95% CI 0.69–0.75), for the internally validated and the external validation PROMPT, and ADNI cohorts, respectively. Discussion The CIDER score could be used to inform clinicians and patients about the relative probabilities of developing dementia in patients with MCI.

Abstract Introduction Amyloid measurement provides important confirmation of pathology for Alzheimer's disease (AD) clinical trials. However, many amyloid positive (Am+) early‐stage subjects do not worsen clinically during a clinical trial, and a neurodegenerative measure predictive of decline could provide critical information. Studies have shown correspondence between perfusion measured by early amyloid frames post‐tracer injection and fluorodeoxyglucose (FDG) positron emission tomography (PET), but with limitations in sensitivity. Multivariate machine learning approaches may offer a more sensitive means for detection of disease related changes as we have demonstrated with FDG. Methods Using summed dynamic florbetapir image frames acquired during the first 6 minutes post‐injection for 107 Alzheimer's Disease Neuroimaging Initiative subjects, we applied optimized machine learning to develop and test image classifiers aimed at measuring AD progression. Early frame amyloid (EFA) classification was compared to that of an independently developed FDG PET AD progression classifier by scoring the FDG scans of the same subjects at the same time point. Score distributions and correlation with clinical endpoints were compared to those obtained from FDG. Region of interest measures were compared between EFA and FDG to further understand discrimination performance. Results The EFA classifier produced a primary pattern similar to that of the FDG classifier whose expression correlated highly with the FDG pattern (R‐squared 0.71), discriminated cognitively normal (NL) amyloid negative (Am–) subjects from all Am+ groups, and that correlated in Am+ subjects with Mini‐Mental State Examination, Clinical Dementia Rating Sum of Boxes, and Alzheimer's Disease Assessment Scale–13‐item Cognitive subscale (R = 0.59, 0.63, 0.73) and with subsequent 24‐month changes in these measures (R = 0.67, 0.73, 0.50). Discussion Our results support the ability to use EFA with a multivariate machine learning–derived classifier to obtain a sensitive measure of AD‐related loss in neuronal function that correlates with FDG PET in preclinical and early prodromal stages as well as in late mild cognitive impairment and dementia. Highlights The summed initial post‐injection minutes of florbetapir positron emission tomography correlate with fluorodeoxyglucose. A machine learning classifier enabled sensitive detection of early prodromal Alzheimer's disease. Early frame amyloid (EFA) classifier scores correlate with subsequent change in Mini‐Mental State Examination, Clinical Dementia Rating Sum of Boxes, and Alzheimer's Disease Assessment Scale–13‐item Cognitive subscale. EFA classifier effect sizes and clinical prediction outperformed region of interest standardized uptake value ratio. EFA classification may aid in stratifying patients to assess treatment effect.

Hippocampal subfield atrophy is a prime structural change in the brain, associated with cognitive aging and neurodegenerative diseases such as Alzheimer’s disease. Recent developments in genome-wide association studies (GWAS) have identified genetic loci that characterize the risk of hippocampal volume loss based on the processes of normal and abnormal aging. Polygenic risk scores are the genetic proxies mimicking the genetic role of the pre-existing vulnerabilities of the underlying mechanisms influencing these changes. Discriminating the genetic predispositions of hippocampal subfield atrophy between cognitive aging and neurodegenerative diseases will be helpful in understanding the disease etiology. In this study, we evaluated the polygenic risk of Alzheimer’s disease (AD PGRS) for hippocampal subfield atrophy in 1,086 individuals (319 cognitively normal (CN), 591 mild cognitively impaired (MCI), and 176 Alzheimer’s disease dementia (ADD)). Our results showed a stronger association of AD PGRS effect on the left hemisphere than on the right hemisphere for all the hippocampal subfield volumes in a mixed clinical population (CN+MCI+ADD). The subfields CA1, CA4, hippocampal tail, subiculum, presubiculum, molecular layer, GC-ML-DG, and HATA showed stronger AD PGRS associations with the MCI+ADD group than with the CN group. The subfields CA3, parasubiculum, and fimbria showed moderately higher AD PGRS associations with the MCI+ADD group than with the CN group. Our findings suggest that the eight subfield regions, which were strongly associated with AD PGRS are likely involved in the early stage ADD and a specific focus on the left hemisphere could enhance the early prediction of ADD.

Abstract Background We aim to provide a systematic study of the impact of white matter (WM) spill-in on the calculation of standardized uptake value ratios (SUVRs) on Aβ-negative subjects, and we study the effect of including WM in the reference region as a compensation. In addition, different partial volume correction (PVC) methods are applied and evaluated. Methods We evaluated magnetic resonance imaging and 18F-AV-45 positron emission tomography data from 122 cognitively normal (CN) patients recruited at the Alzheimer’s Disease Neuroimaging Initiative (ADNI). Cortex SUVRs were obtained by using the cerebellar grey matter (CGM) (SUVRCGM) and the whole cerebellum (SUVRWC) as reference regions. The correlations between the different SUVRs and the WM uptake (WM-SUVRCGM) were studied in patients, and in a well-controlled framework based on Monte Carlo (MC) simulation. Activity maps for the MC simulation were derived from ADNI patients by using a voxel-wise iterative process (BrainViset). Ten WM uptakes covering the spectrum of WM values obtained from patient data were simulated for different patients. Three different PVC methods were tested (a) the regional voxel-based (RBV), (b) the iterative Yang (iY), and (c) a simplified analytical correction derived from our MC simulation. Results WM-SUVRCGM followed a normal distribution with an average of 1.79 and a standard deviation of 0.243 (13.6%). SUVRCGM was linearly correlated to WM-SUVRCGM (r = 0.82, linear fit slope = 0.28). SUVRWC was linearly correlated to WM-SUVRCGM (r = 0.64, linear fit slope = 0.13). Our MC results showed that these correlations are compatible with those produced by isolated spill-in effect (slopes of 0.23 and 0.11). The impact of the spill-in was mitigated by using PVC for SUVRCGM (slopes of 0.06 and 0.07 for iY and RBV), while SUVRWC showed a negative correlation with SUVRCGM after PVC. The proposed analytical correction also reduced the observed correlations when applied to patient data (r = 0.27 for SUVRCGM, r = 0.18 for SUVRWC). Conclusions There is a high correlation between WM uptake and the measured SUVR due to spill-in effect, and that this effect is reduced when including WM in the reference region. We also evaluated the performance of PVC, and we proposed an analytical correction that can be applied to preprocessed data.

The growth-associated protein 43 (GAP-43) is a presynaptic phosphoprotein in cerebrospinal fluid (CSF). The ε4 allele of apolipoprotein E (APOE) is an important genetic risk factor for Alzheimer’s disease (AD). We aimed to evaluate the association of CSF GAP-43 with cognition and whether this correlation was related to the APOE ε4 status. We recruited participants from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) database, and they were divided into cognitively normal (CN) ε4 negative (CN ε4−), CN ε4 positive (CN ε4+), mild cognitive impairment (MCI) ε4 negative (MCI ε4−), MCI ε4 positive (MCI ε4+), AD ε4 negative (AD ε4−), and AD ε4 positive (AD ε4+) groups. Spearman’s correlation was utilized to evaluate the relationship between CSF GAP-43 and core AD biomarkers at the baseline. We performed receiver-operating characteristic (ROC) curve analyses to investigate the diagnostic accuracy of CSF GAP-43. The correlations between CSF GAP-43 and the Mini-Mental State Examination (MMSE) scores and brain atrophy at baseline were assessed by using multiple linear regression, while the association between CSF GAP-43 and MMSE scores at the follow-up was tested by performing the generalized estimating equation (GEE). The role of CSF GAP-43 in the conversion from MCI to AD was evaluated using the Cox proportional hazard model. We found that the CSF GAP-43 level was significantly increased in MCI ε4+, AD ε4− and AD ε4+ groups compared with CN ε4− or MCI ε4− group. The negative associations between the CSF GAP-43 and MMSE scores at the baseline and follow-up were found in MCI ε4− and MCI ε4+ groups. In addition, baseline CSF GAP-43 was able to predict the clinical progression from MCI to AD. CSF GAP-43 may be a promising biomarker to screen cognition for AD. The effects of CSF GAP-43 on cognition were suspected to be relevant to APOE ε4 status.

Subtype and Stage Inference (SuStaIn) is an unsupervised learning algorithm that uniquely enables the identification of subgroups of individuals with distinct pseudo-temporal disease progression patterns from cross-sectional datasets. SuStaIn has been used to identify data-driven subgroups and perform patient stratification in neurodegenerative diseases and in lung diseases from continuous biomarker measurements predominantly obtained from imaging. However, the SuStaIn algorithm is not currently applicable to discrete ordinal data, such as visual ratings of images, neuropathological ratings, and clinical and neuropsychological test scores, restricting the applicability of SuStaIn to a narrower range of settings. Here we propose ‘Ordinal SuStaIn’, an ordinal version of the SuStaIn algorithm that uses a scored events model of disease progression to enable the application of SuStaIn to ordinal data. We demonstrate the validity of Ordinal SuStaIn by benchmarking the performance of the algorithm on simulated data. We further demonstrate that Ordinal SuStaIn out-performs the existing continuous version of SuStaIn (Z-score SuStaIn) on discrete scored data, providing much more accurate subtype progression patterns, better subtyping and staging of individuals, and accurate uncertainty estimates. We then apply Ordinal SuStaIn to six different sub-scales of the Clinical Dementia Rating scale (CDR) using data from the Alzheimer’s disease Neuroimaging Initiative (ADNI) study to identify individuals with distinct patterns of functional decline. Using data from 819 ADNI1 participants we identified three distinct CDR subtype progression patterns, which were independently verified using data from 790 ADNI2 participants. Our results provide insight into patterns of decline in daily activities in Alzheimer’s disease and a mechanism for stratifying individuals into groups with difficulties in different domains. Ordinal SuStaIn is broadly applicable across different types of ratings data, including visual ratings from imaging, neuropathological ratings and clinical or behavioural ratings data.

Abstract Background Preclinical Alzheimer’s disease (AD) clinical trials require participants to enroll with a study partner, a person who can attend visits and report changes in the participant’s cognitive ability. Whether study partners, compared to participants themselves, provide added information about participant cognition in preclinical AD trials is an open question. We tested the hypothesis that study partners provide meaningful information related to participant cognition cross-sectionally and longitudinally, and assessed whether amyloid status modified observed effects. Methods We assessed participant and study partner Everyday Cognition (ECog) scores and participant Alzheimer’s Disease Assessment Scale 13-item cognitive subscale (ADAS13) data from 335 cognitively normal participant-partner dyads in the AD Neuroimaging Initiative. We used random forest and linear mixed effects (LME) models to predict ADAS13 scores as a function of participant and/or study partner ECog scores over time. LME models were adjusted for potential confounding factors, including APOE4 status, amyloid status, baseline age, years of education, and sex. Random forest models were split into the above factors, as well as race/ethnicity and other available neuropsychological battery test scores. Results In random forest models predicting ADAS13 12 months from baseline, we observed no difference in the estimated mean variable importance (eMVI) associated with baseline study partner ECog compared to the baseline participant ECog (eMVI = 0.15, 95%CB 0.13, 0.16 for partner; eMVI = 0.15, 95%CB 0.14, 0.16 for participant). In models predicting ADAS13 48 months after baseline, the eMVI associated with baseline study partner ECog was slightly lower than that associated with baseline participant ECog (eMVI = 0.21, 95%CB 0.20, 0.22 for partner; eMVI = 0.24, 95%CB 0.22, 0.25 for participant). In cross-sectional models, study partner eMVI was twice as large as participant eMVI at 12 months (eMVI = 0.20, 95%CB 0.19, 0.21 for partner; eMVI = 0.09, 95%CB 0.09, 0.10 for participant) and three times as large at 48 months (eMVI = 0.38, 95%CB 0.36, 0.39 for partner; eMVI = 0.13, 95%CB 0.12, 0.14 for participant). We did not observe qualitative differences by amyloid status. Conclusions While baseline participant reports reasonably predict subsequent cognitive change, informants perform better at cross-sectionally recognizing cognitive status as observation time grows. The study partner requirement may be essential to ensure trial data integrity, especially in longer trials.

Abstract Background Magnetic resonance imaging (MRI) has unveiled specific alterations at different stages of Alzheimer’s disease (AD) pathophysiologic continuum constituting what has been established as “AD signature”. To what extent MRI can detect amyloid-related cerebral changes from structural MRI in cognitively unimpaired individuals is still an area open for exploration. Method Longitudinal 3D-T1 MRI scans were acquired from a subset of the ADNI cohort comprising 403 subjects: 79 controls (Ctrls), 50 preclinical AD (PreAD), and 274 MCI and dementia due to AD (MCI/AD). Amyloid CSF was used as gold-standard measure with established cutoffs ( 2.5 years, and hence, only subjects within this temporal span are used for evaluation (15 Ctrls, 10 PreAD). The longitudinal voxel-based classifier achieved an AUC = 0.87 (95%CI 0.72–0.97). The brain regions that showed the highest discriminative power to detect amyloid abnormalities were the medial, inferior, and lateral temporal lobes; precuneus; caudate heads; basal forebrain; and lateral ventricles. Conclusions Our work supports that machine learning applied to longitudinal brain volumetric changes can be used to predict, with high precision, the presence of amyloid abnormalities in cognitively unimpaired subjects. Used as a triaging method to identify a fixed number of amyloid-positive individuals, this longitudinal voxel-wise classifier is expected to avoid 55% of unnecessary CSF and/or PET scans and reduce economic cost by 40%.

Abstract Objective To determine the key inflammatory pathways that are activated in the peripheral and CNS compartments at the mild cognitive impairment (MCI) stage of Alzheimer’s disease (AD). Methods A cross‐sectional study of patients with clinical and biomarker characteristics consistent with MCI‐AD in a discovery cohort, with replication in the Alzheimer’s Disease Neuroimaging Initiative (ADNI) cohort. Inflammatory analytes were measured in the CSF and plasma with the same validated multiplex analyte platform in both cohorts and correlated with AD biomarkers (CSF Aβ42, total tau (t‐tau), phosphorylated tau (p‐tau) to identify key inflammatory pathway activations. The pathways were additionally validated by evaluating genes related to all analytes in coexpression networks of brain tissue transcriptome from an autopsy confirmed AD cohort to interrogate if the same pathway activations were conserved in the brain tissue gene modules. Results Analytes of the tumor necrosis factor (TNF) signaling pathway (KEGG ID:4668) in the CSF and plasma best correlated with CSF t‐tau and p‐tau levels, and analytes of the complement and coagulation pathway (KEGG ID:4610) best correlated with CSF Aβ42 levels. The top inflammatory signaling pathways of significance were conserved in the peripheral and the CNS compartments. They were also confirmed to be enriched in AD brain transcriptome gene clusters. Interpretation A cell‐protective rather than a proinflammatory analyte profile predominates in the CSF in relation to neurodegeneration markers among MCI‐AD patients. Analytes from the TNF signaling and the complement and coagulation pathways are relevant in evaluating disease severity at the MCI stage of AD.

Abstract Background Late-onset Alzheimer’s disease (AD) is characterized by primary memory impairment, which then progresses towards severe deficits across cognitive domains. Here, we report how performance in cognitive domains relates to patterns of tau deposition and cortical thickness. Methods We analyzed data from 131 amyloid-β positive participants (55 cognitively normal, 46 mild cognitive impairment, 30 AD) of the Alzheimer’s Disease Neuroimaging Initiative who underwent magnetic resonance imaging (MRI), flortaucipir (FTP) positron emission tomography, and neuropsychological testing. Surface-based vertex-wise and region-of-interest analyses were conducted between FTP and cognitive test scores, and between cortical thickness and cognitive test scores. Results FTP and thickness were differentially related to cognitive performance in several domains. FTP-cognition associations were more widespread than thickness-cognition associations. Further, FTP-cognition patterns reflected cortical systems that underlie different aspects of cognition. Conclusions Our findings indicate that AD-related decline in domain-specific cognitive performance reflects underlying progression of tau and atrophy into associated brain circuits. They also suggest that tau-PET may have better sensitivity to this decline than MRI-derived measures of cortical thickness.

Abstract Background Amyloid β (Aβ) and tau proteins are considered as critical factors that affect Alzheimer’s disease (AD) and mild cognitive impairment (MCI). Although many studies have conducted on these two proteins, little study has investigated the relationship between their spatial distributions. This study aims to explore the associations of spatial patterns between Aβ deposition and tau deposition in patients with MCI and normal control (NC). Methods We used multimodality positron emission tomography (PET) data from a clinically heterogeneous population of patients with MCI and NC. All data were obtained from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) database containing information of 65 patients with MCI and 75 NC who both had undergone AV45 (Aβ) and AV1451 (tau) PET. To assess the spatial distribution of Aβ and tau deposition, we employed parallel independent component analysis (pICA), which enabled the joint analysis of multimodal imaging data. pICA was conducted to identify the significant difference and correlation relationship of brain networks between Aβ PET and tau PET in MCI and NC groups. Results Our results revealed the strongly correlated network between Aβ PET and tau PET were colocalized with the default-mode network (DMN). Simultaneously, in comparison of the spatial distribution between Aβ PET and tau PET, it was found that the significant differences between MCI and NC were mainly distributed in DMN, cognitive control network and visual networks. The altered brain networks obtained from pICA analysis are consistent with the abnormalities of brain network in MCI patients. Conclusions Findings suggested the abnormal spatial distribution regions of tau PET were correlated with the abnormal spatial distribution regions of Aβ PET, and both of which were located in DMN network. This study revealed that combining pICA with multimodal imaging data is an effective approach for distinguishing MCI patients from NC group.

Abstract Alzheimer’s disease is the most common neurodegenerative disease and is characterized by the accumulation of amyloid-beta peptides leading to the formation of plaques and tau protein tangles in brain. These neuropathological features precede cognitive impairment and Alzheimer’s dementia by many years. To better understand and predict the course of disease from early-stage asymptomatic to late-stage dementia, it is critical to study the patterns of progression of multiple markers. In particular, we aim to predict the likely future course of progression for individuals given only a single observation of their markers. Improved individual-level prediction may lead to improved clinical care and clinical trials. We propose a two-stage approach to modeling and predicting measures of cognition, function, brain imaging, fluid biomarkers, and diagnosis of individuals using multiple domains simultaneously. In the first stage, joint (or multivariate) mixed-effects models are used to simultaneously model multiple markers over time. In the second stage, random forests are used to predict categorical diagnoses (cognitively normal, mild cognitive impairment, or dementia) from predictions of continuous markers based on the first-stage model. The combination of the two models allows one to leverage their key strengths in order to obtain improved accuracy. We characterize the predictive accuracy of this two-stage approach using data from the Alzheimer’s Disease Neuroimaging Initiative. The two-stage approach using a single joint mixed-effects model for all continuous outcomes yields better diagnostic classification accuracy compared to using separate univariate mixed-effects models for each of the continuous outcomes. Overall prediction accuracy above 80% was achieved over a period of 2.5 years. The results further indicate that overall accuracy is improved when markers from multiple assessment domains, such as cognition, function, and brain imaging, are used in the prediction algorithm as compared to the use of markers from a single domain only.

Abstract Background Biomarkers such as cerebrospinal fluid (CSF) and magnetic resonance imaging (MRI) have predictive value for progression to dementia in patients with mild cognitive impairment (MCI). The pre-dementia stage takes far longer, and the interpretation of biomarker findings is particular relevant for individuals who present at a memory clinic, but are deemed cognitively normal. The objective of the current study is to construct biomarker-based prognostic models for personalized risk of clinical progression in cognitively normal individuals presenting at a memory clinic. Methods We included 481 individuals with subjective cognitive decline (SCD) from the Amsterdam Dementia Cohort. Prognostic models were developed by Cox regression with patient characteristics, MRI, and/or CSF biomarkers to predict clinical progression to MCI or dementia. We estimated 5- and 3-year individualized risks based on patient-specific values. External validation was performed on Alzheimer’s Disease Neuroimaging Initiative (ADNI) and an European dataset. Results Based on demographics only (Harrell’s C = 0.70), 5- and 3-year progression risks varied from 6% [3–11] and 4% [2–8] (age 55, MMSE 30) to 38% [29–49] and 28% [21–37] (age 70, MMSE 27). Normal CSF biomarkers strongly decreased progression probabilities (Harrell’s C = 0.82). By contrast, abnormal CSF markedly increased risk (5 years, 96% [56–100]; 3 years, 89% [44–99]). The CSF model could reclassify 58% of the individuals with an “intermediate” risk (35–65%) based on the demographic model. MRI measures were not retained in the models. Conclusion The current study takes the first steps in a personalized approach for cognitively normal individuals by providing biomarker-based prognostic models.

Abstract Background The Centiloid scale has been developed to standardize measurements of amyloid PET imaging. Reference cut-off values of this continuous measurement enable the consistent operationalization of decision-making for multicentre research studies and clinical trials. In this study, we aimed at deriving reference Centiloid thresholds that maximize the agreement against core Alzheimer’s disease (AD) cerebrospinal fluid (CSF) biomarkers in two large independent cohorts. Methods A total of 516 participants of the ALFA+ Study (N = 205) and ADNI (N = 311) underwent amyloid PET imaging ([18F]flutemetamol and [18F]florbetapir, respectively) and core AD CSF biomarker determination using Elecsys® tests. Tracer uptake was quantified in Centiloid units (CL). Optimal Centiloid cut-offs were sought that maximize the agreement between PET and dichotomous determinations based on CSF levels of Aβ42, tTau, pTau, and their ratios, using pre-established reference cut-off values. To this end, a receiver operating characteristic analysis (ROC) was conducted, and Centiloid cut-offs were calculated as those that maximized the Youden’s J Index or the overall percentage agreement recorded. Results All Centiloid cut-offs fell within the range of 25–35, except for CSF Aβ42 that rendered an optimal cut-off value of 12 CL. As expected, the agreement of tau/Aβ42 ratios was higher than that of CSF Aβ42. Centiloid cut-off robustness was confirmed even when established in an independent cohort and against variations of CSF cut-offs. Conclusions A cut-off of 12 CL matches previously reported values derived against postmortem measures of AD neuropathology. Together with these previous findings, our results flag two relevant inflection points that would serve as boundary of different stages of amyloid pathology: one around 12 CL that marks the transition from the absence of pathology to subtle pathology and another one around 30 CL indicating the presence of established pathology. The derivation of robust and generalizable cut-offs for core AD biomarkers requires cohorts with adequate representation of intermediate levels. Trial registration ALFA+ Study, NCT02485730 ALFA PET Sub-study, NCT02685969

Abstract Background TREM2 is a transmembrane receptor that is predominantly expressed by microglia in the central nervous system. Rare variants in the TREM2 gene increase the risk for late-onset Alzheimer’s disease (AD). Soluble TREM2 (sTREM2) resulting from shedding of the TREM2 ectodomain can be detected in the cerebrospinal fluid (CSF) and is a surrogate measure of TREM2-mediated microglia function. CSF sTREM2 has been previously reported to increase at different clinical stages of AD, however, alterations in relation to Amyloid β-peptide (Aβ) deposition or additional pathological processes in the amyloid cascade (such as tau pathology or neurodegeneration) remain unclear. In the current cross-sectional study, we employed the biomarker-based classification framework recently proposed by the NIA-AA consensus guidelines, in combination with clinical staging, in order to examine the CSF sTREM2 alterations at early asymptomatic and symptomatic stages of AD. Methods A cross-sectional study of 1027 participants of the Alzheimer’s Disease Imaging Initiative (ADNI) cohort, including 43 subjects carrying TREM2 rare genetic variants, was conducted to measure CSF sTREM2 using a previously validated enzyme-linked immunosorbent assay (ELISA). ADNI participants were classified following the A/T/N framework, which we implemented based on the CSF levels of Aβ1-42 (A), phosphorylated tau (T) and total tau as a marker of neurodegeneration (N), at different clinical stages defined by the clinical dementia rating (CDR) score. Results CSF sTREM2 differed between TREM2 variants, whereas the p.R47H variant had higher CSF sTREM2, p.L211P had lower CSF sTREM2 than non-carriers. We found that CSF sTREM2 increased in early symptomatic stages of late-onset AD but, unexpectedly, we observed decreased CSF sTREM2 levels at the earliest asymptomatic phase when only abnormal Aβ pathology (A+) but no tau pathology or neurodegeneration (TN-), is present. Conclusions Aβ pathology (A) and tau pathology/neurodegeneration (TN) have differing associations with CSF sTREM2. While tau-related neurodegeneration is associated with an increase in CSF sTREM2, Aβ pathology in the absence of downstream tau-related neurodegeneration is associated with a decrease in CSF sTREM2.

Abstract Background Predicting progression from Mild Cognitive Impairment (MCI) to Alzheimer’s Disease (AD) is an utmost open issue in AD-related research. Neuropsychological assessment has proven to be useful in identifying MCI patients who are likely to convert to dementia. However, the large battery of neuropsychological tests (NPTs) performed in clinical practice and the limited number of training examples are challenge to machine learning when learning prognostic models. In this context, it is paramount to pursue approaches that effectively seek for reduced sets of relevant features. Subsets of NPTs from which prognostic models can be learnt should not only be good predictors, but also stable, promoting generalizable and explainable models. Methods We propose a feature selection (FS) ensemble combining stability and predictability to choose the most relevant NPTs for prognostic prediction in AD. First, we combine the outcome of multiple (filter and embedded) FS methods. Then, we use a wrapper-based approach optimizing both stability and predictability to compute the number of selected features. We use two large prospective studies (ADNI and the Portuguese Cognitive Complaints Cohort, CCC) to evaluate the approach and assess the predictive value of a large number of NPTs. Results The best subsets of features include approximately 30 and 20 (from the original 79 and 40) features, for ADNI and CCC data, respectively, yielding stability above 0.89 and 0.95, and AUC above 0.87 and 0.82. Most NPTs learnt using the proposed feature selection ensemble have been identified in the literature as strong predictors of conversion from MCI to AD. Conclusions The FS ensemble approach was able to 1) identify subsets of stable and relevant predictors from a consensus of multiple FS methods using baseline NPTs and 2) learn reliable prognostic models of conversion from MCI to AD using these subsets of features. The machine learning models learnt from these features outperformed the models trained without FS and achieved competitive results when compared to commonly used FS algorithms. Furthermore, the selected features are derived from a consensus of methods thus being more robust, while releasing users from choosing the most appropriate FS method to be used in their classification task.

BackgroundApolipoprotein E (APOE) ε2 is a protective genetic factor for Alzheimer’s disease (AD). However, the potential interaction effects between the APOE ε2 allele and disease status on the intrinsic brain activity remain elusive.MethodsWe identified 73 healthy control (HC) with APOE ε3/ε3, 61 mild cognitive impairment (MCI) subjects with APOE ε3/ε3, 24 HC with APOE ε2/ε3, and 10 MCI subjects with APOE ε2/ε3 from the ADNI database. All subjects underwent a resting-state functional MRI and Fluoro-deoxy-glucose positron emission tomography (FDG-PET). We used a fractional amplitude of low-frequency fluctuation (fALFF) to explore the spontaneous brain activity. Based on the mixed-effects analysis, we explored the interaction effects between the APOE ε2 allele versus disease status on brain activity and metabolism in a voxel-wise fashion (GRF corrected, p < 0.01), followed by post hoc two-sample t-tests (Bonferroni corrected, p < 0.05). We then investigated the relationship between the mean imaging metrics and cognitive abilities.ResultsThere are no significant differences in gender, age, or education among the four groups. The interaction effect on brain activity was located in the inferior parietal lobule (IPL). Post hoc analysis showed that APOE ε2/ε3 MCI had an increased IPL fALFF than APOE ε3/ε3 MCI. Regarding the APOE ε2 allele effects, we found that ε2 carriers had a decreased fALFF in the transverse temporal gyrus than non-carriers. Also, FDG-PET results showed a lower SUVR of the frontal lobe in APOE ε2 carriers than non-carriers. Furthermore, fALFF of IPL was correlated with the visuospatial function (r = −0.16, p < 0.05).ConclusionAPOE ε2 carriers might have a better brain reservation when coping with AD-related pathologies.

Abstract Introduction Using brain network and graph theory methods to analyze the Alzheimer's disease (AD) and mild cognitive impairment (MCI) abnormal brain function is more and more popular. Plenty of potential methods have been proposed, but the representative signal of each brain region in these methods remains poor performance. Methods We propose a highly‐available nodes approach for constructing brain network of patients with MCI and AD. With resting‐state functional magnetic resonance imaging (rs‐fMRI) data of 84 AD subjects, 81 MCI subjects, and 82 normal control (NC) subjects from the Alzheimer's Disease Neuroimaging Initiative Database, we construct connected weighted brain networks based on the different sparsity and minimum spanning tree. Support Vector Machine of Radial Basis Function kernel was selected as classifier. Results Accuracies of 74.09% and 77.58% in classification of MCI and AD from NC, respectively. We also performed a hub node analysis and found 18 significant brain regions were identified as hub nodes. Conclusions The findings of this study provide insights for helping understanding the progress of the AD. The proposed method highlights the effectively representative time series of brain regions of rs‐fMRI data for construction and topology analysis brain network.

Abstract Introduction Analyses of off‐label use of acetylcholinesterase inhibitors (AChEIs) in mild cognitive impairment (MCI) has produced mixed results. Post hoc analyses of observational cohorts, such as the Alzheimer's Disease Neuroimaging Initiative (ADNI), have reported deleterious effects in AChEI‐treated subjects (AChEI+). Here, we used neuroimaging biomarkers to determine whether AChEI+ subjects had a greater rate of neurodegeneration than untreated (AChEI–) subjects while accounting for baseline differences. Methods We selected 121 ADNI MCI AChEI+ subjects and 151 AChEI– subjects with a magnetic resonance imaging (MRI) scan; 82 AChEI+ and 110 AChEI– also had a fluorodeoxyglucose (FDG) scan. A subset (83 AChEI+ and 98 AChEI–) had cerebrospinal fluid (CSF) or amyloid positron emission tomography (PET) assessment for amyloid positivity. Linear regression models were used to compare the effect of treatment on changes in Mini‐Mental State Examination and Clinical Dementia Rating‐Sum of Boxes scores. We used standard regression in SPM (for baseline) and the SPM toolbox sandwich estimator, SwE (for longitudinal) for comparisons of AChEI+ and AChEI– FDG PET and MRI data. Results At baseline, the AChEI+ group had significantly reduced cortical gray matter density (GMD) and more hypometabolism than AChEI– subjects. The greater rate of atrophy and hypometabolic changes over time in AChEI+ compared to AChEI– subjects did not survive correction for baseline differences. AChEI+ participants were more likely to be amyloid‐positive and have lower GMD and FDG standardized uptake value ratio than AChEI– at baseline. AChEI+ subjects showed greater atrophy over time, which remained significant after controlling for amyloid status. Discussion Our data suggest that the observed differences in rates of cognitive decline, atrophy, and hypometabolism are likely the result of significant baseline differences between the groups. Furthermore, the data indicate no treatment effect of AChEI (positive of negative), rather that physicians prescribe AChEI to subjects who present with more severe clinical impairment. This alone may account for the negative effect seen previously in the ADNI population of AChEI use among MCI subjects.

Abstract Introduction We examine whether distinct brain atrophy patterns (using brain parenchymal fraction [BPF]) differentially predict functional performance and decline in Alzheimer's disease (AD), and are independently moderated by (1) a key AD genetic risk marker (apolipoprotein E [APOE]), (2) sex, and (3) high‐risk group (women APOE ɛ4 carriers). Methods We used a 2‐year longitudinal sample of AD patients (baseline N = 170; mean age = 71.3 [9.1] years) from the Sunnybrook Dementia Study. We applied latent class analysis, latent growth modeling, and path analysis. We aimed to replicate our findings (N = 184) in the Alzheimer's Disease Neuroimaging Initiative. Results We observed that high brain atrophy class predicted lower functional performance and steeper decline. This association was moderated by APOE, sex, and high‐risk group. Baseline findings as moderated by APOE and high‐risk group were replicated. Discussion Women APOE ɛ4 carriers may selectively be at a greater risk of functional impairment with higher brain atrophy.

Abstract Introduction This study sought to determine whether adding cognition to a model with Alzheimer's disease biomarkers based on the amyloid, tau, and neurodegeneration/neuronal injury—AT(N)—biomarker framework predicts rates of cognitive and functional decline in older adults without dementia. Methods The study included 465 participants who completed amyloid positron emission tomography, cerebrospinal fluid phosphorylated tau, structural magnetic resonance imaging, and serial neuropsychological testing. Using the AT(N) framework and a newly validated cognitive metric as the independent variables, we used linear mixed effects models to examine a 4‐year rate of change in cognitive and functional measures. Results The inclusion of baseline cognitive status improved model fit in predicting rate of decline in outcomes above and beyond biomarker variables. Specifically, those with worse cognitive functioning at baseline had faster rates of memory and functional decline over a 4‐year period, even when accounting for AT(N). Discussion Including a newly validated measure of baseline cognition may improve clinical prognosis in non‐demented older adults beyond the use of AT(N) biomarkers alone.

Abstract Introduction Positron emission tomography (PET) imaging targeting neurofibrillary tau tangles is increasingly used in the study of Alzheimer's disease (AD), but its utility may be limited by conventional quantitative or qualitative evaluation techniques in earlier disease states. Convolutional neural networks (CNNs) are effective in learning spatial patterns for image classification. Methods 18F‐MK6240 (n = 320) and AV‐1451 (n = 446) PET images were pooled from multiple studies. We performed iterations with differing permutations of radioligands, heuristics, and architectures. Performance was compared to a standard region of interest (ROI)‐based approach on prediction of memory impairment. We visualized attention of the network to illustrate decision making. Results Overall, models had high accuracy (> 80%) with good average sensitivity and specificity (75% and 82%, respectively), and had comparable or higher accuracy to the ROI standard. Visualizations of model attention highlight known characteristics of tau radioligand binding. Discussion CNNs could improve tau PET's role in early disease and extend the utility of tau PET across generations of radioligands.

Abstract Introduction Objectively‐defined subtle cognitive decline (Obj‐SCD) and plasma phosphorylated‐tau181 (p‐tau181) are promising early Alzheimer's disease (AD) markers. However, associations between Obj‐SCD and p‐tau181, and their combined prognostic potential, are unknown. Methods Baseline and 4‐year longitudinal p‐tau181 changes were compared across cognitively unimpaired (CU; n = 402), Obj‐SCD (n = 199), and mild cognitive impairment (MCI; n = 346) groups. CU and Obj‐SCD participants were further classified as p‐tau181‐positive or negative. Results CU and Obj‐SCD has lower baseline p‐tau181 than MCI and did not differ from one another. Longitudinally, Obj‐SCD had the steepest p‐tau181 increase. Obj‐SCD/p‐tau181‐positive participants had the fastest rates of amyloid accumulation, cognitive decline, and functional decline. Conclusions Despite assumptions that cognitive changes invariably follow biomarker changes, early neuropsychological difficulties may emerge before/concurrently with plasma p‐tau181 changes. Combining Obj‐SCD and p‐tau181, two potentially accessible early markers, was associated with the faster declines in AD‐related outcomes.

Abstract Introduction Neuropsychological test scores are limited and standard outcomes may mask the heterogeneity of cognitive impairment. This article presents the calculation and evaluation of six composite scores that quantify domain‐specific impairment. Methods Parameters for composite scores calculation were learned by performing confirmatory factor analysis in a sample of participants from the Alzheimer's Disease Neuroimaging Initiative database. The obtained scores were evaluated with a separate sample of mild cognitive impairment (MCI) in two automated tasks: unsupervised partition in different subgroups and prediction of progression to dementia for different time windows. Results MCI subgroups with distinctive cognitive profiles and risk of progression emerged from cluster analysis. Composite scores outperform standard neuropsychological tests when automatically predicting progression within time windows up to 5 years. Conclusions Domain‐specific composite scores are useful to delineate profiles of impairment, stratify the MCI risk, and predict progression to dementia.

Abstract Introduction Our goal was to determine whether cognitive and cerebrospinal fluid (CSF) markers of tau and amyloid beta 1‐42 (Aβ42) differ between Vietnam‐era veterans with and without history of traumatic brain injury (TBI) and whether TBI moderates the association between CSF markers and neurocognitive functioning. Methods A total of 102 male participants (52 TBI, 50 military controls [MCs]; mean age = 68) were included. Levels of CSF Aβ42, tau phosphorylated at the threonine 181 position (p‐tau), and total tau (t‐tau) were quantified. Group differences in CSF markers and cognition as well as the moderating effect of TBI on CSF and cognition associations were explored. Results Relative to MCs, the TBI group showed significantly higher p‐tau (P = .01) and t‐tau (P = .02), but no differences in amyloid (P = .09). TBI history moderated the association between CSF tau and performance on a measure of processing speed (t‐tau: P = .04; p‐tau: P = .02). Discussion Tau accumulation may represent a mechanism of dementia risk in older veterans with remote TBI.

Background: Suspected non-Alzheimer disease pathophysiology (SNAP) refers to the subjects who feature negative β-amyloid (Aβ) but positive tau or neurodegeneration biomarkers. It accounts for a quarter of the elderly population and is associated with cognitive decline. However, the underlying pathophysiology is still unclear.Methods: We included 111 non-demented subjects, then classified them into three groups using cerebrospinal fluid (CSF) Aβ 1–42 (A), phosphorylated tau 181 (T), and total tau (N). Specifically, we identified the normal control (NC; subjects with normal biomarkers, A-T-N-), SNAP (subjects with normal amyloid but abnormal tau, A−T+), and predementia Alzheimer’s disease (AD; subjects with abnormal amyloid and tau, A+T+). Then, we used the static amplitude of low-frequency fluctuation (sALFF) and dynamic ALFF (dALFF) variance to reflect the intrinsic functional network strength and stability, respectively. Further, we performed a correlation analysis to explore the possible relationship between intrinsic brain activity changes and cognition.Results: SNAP showed decreased sALFF in left superior frontal gyrus (SFG) while increased sALFF in left insula as compared to NC. Regarding the dynamic metric, SNAP showed a similarly decreased dALFF in the left SFG and left paracentral lobule as compared to NC. By contrast, when compared to NC, predementia AD showed decreased sALFF in left inferior parietal gyrus (IPG) and right precuneus, while increased sALFF in the left insula, with more widely distributed decreased dALFF variance across the frontal, parietal and occipital lobe. When directly compared to SNAP, predementia AD showed decreased sALFF in left middle occipital gyrus and IPG, while showing decreased dALFF variance in the left temporal pole. Further correlation analysis showed that increased sALFF in the insula had a negative correlation with the general cognition in the SNAP group. Besides, sALFF and dALFF variance in the right precuneus negatively correlated with attention in the predementia AD group.Conclusion: SNAP and predementia AD show distinct functional impairment patterns. Specifically, SNAP has functional impairments that are confined to the frontal region, which is usually spared in early-stage AD, while predementia AD exhibits widely distributed functional damage involving the frontal, parietal and occipital cortex.

Alzheimer’s disease (AD) is a progressive neurodegenerative disease, for which aging remains the major risk factor. Aging is under a consistent pressure of increasing brain entropy (BEN) due to the progressive brain deteriorations. Noticeably, the brain constantly consumes a large amount of energy to maintain its functional integrity, likely creating or maintaining a big “reserve” to counteract the high entropy. Malfunctions of this latent reserve may indicate a critical point of disease progression. The purpose of this study was to characterize BEN in aging and AD and to test an inverse-U-shape BEN model: BEN increases with age and AD pathology in normal aging but decreases in the AD continuum. BEN was measured with resting state fMRI and compared across aging and the AD continuum. Associations of BEN with age, education, clinical symptoms, and pathology were examined by multiple regression. The analysis results highlighted resting BEN in the default mode network, medial temporal lobe, and prefrontal cortex and showed that: (1) BEN increased with age and pathological deposition in normal aging but decreased with age and pathological deposition in the AD continuum; (2) AD showed catastrophic BEN reduction, which was related to more severe cognitive impairment and daily function disability; and (3) BEN decreased with education years in normal aging, but not in the AD continuum. BEN evolution follows an inverse-U trajectory when AD progresses from normal aging to AD dementia. Education is beneficial for suppressing the entropy increase potency in normal aging.

Current researches on default mode network (DMN) in normal elderly have mainly focused on finding some dysfunctional areas with decreased or increased connectivity. The global network dynamics of apolipoprotein E (APOE) e4 allele group is rarely studied. In our previous brain network study, we have demonstrated the advantage of persistent homology. It can distinguish robust and noisy topological features over multiscale nested networks, and the derived properties are more stable. In this study, for the first time we applied persistent homology to analyze APOE-related effects on whole-brain functional network. In our experiments, the risk allele group exhibited lower network radius and modularity in whole brain DMN based on graph theory, suggesting the abnormal organization structure. Moreover, two suggested measures from persistent homology detected significant differences between groups within the left hemisphere and in the whole brain in two datasets. They were more statistically sensitive to APOE genotypic differences than standard graph-based measures. In summary, we provide evidence that the e4 genotype leads to distinct DMN functional alterations in the early phases of Alzheimer’s disease using persistent homology approach. Our study offers a novel insight to explore potential biomarkers in healthy elderly populations carrying APOE e4 allele.

Abstract Introduction Triggering receptor expressed on myeloid cells‐like transcript 2 gene (TREML2) is a newly identified AD susceptibility gene. Its missense variant rs3747742‐C substantially decreases AD risk in both Caucasians and Han Chinese, but the underlying mechanisms remain elusive. In the present study, to uncover the possible mechanisms by which TREML2 rs3747742‐C reduces AD risk, we investigated the possible relation of this variant with AD‐related brain structures using a cognitively normal elderly population from Alzheimer's Disease Neuroimaging Initiative (ADNI) database. Methods In total, 158 cognitively normal elders from ADNI database with complete data for brain structures and TREML2 rs3747742 genotype were included in this study. The association of TREML2 rs3747742 genotype with the structures of three cerebral cortices (entorhinal cortex, middle temporal gyrus, and parahippocampal gyrus), two subcortical regions (amygdala and hippocampus), and three subfields of hippocampus (CA1, CA2 + CA3, and CA4 + dentate gyrus) was investigated. Results A significant difference was noted in the volume of right CA1 subfield among three genotypes of TREML2 rs3747742 (p = .0364). In the multivariate analysis, TREML2 rs3747742‐C significantly increased right CA1 subfield volume after adjusting for age, gender, education years, APOE ε4 status, and intracranial volume under the recessive genetic model (Bonferroni corrected p = .003586). Conclusion The present study provides the first evidence that TREML2 rs3747742‐C carriers have larger volumes of hippocampal CA1 subfield in a cognitively normal elderly population. These findings imply that enhancement of brain reserve may contribute to the protection of TREML2 rs3747742‐C in AD susceptibility.

A hierarchical clustering algorithm was applied to magnetic resonance images (MRI) of a cohort of 751 subjects having a mild cognitive impairment (MCI), 282 subjects having received Alzheimer’s disease (AD) diagnosis, and 428 normal controls (NC). MRIs were preprocessed to gray matter density maps and registered to a stereotactic space. By first rendering the gray matter density maps comparable by regressing out age, gender, and years of education, and then performing the hierarchical clustering, we found clusters displaying structural features of typical AD, cortically-driven atypical AD, limbic-predominant AD, and early-onset AD (EOAD). Among these clusters, EOAD subjects displayed marked cortical gray matter atrophy and atrophy of the precuneus. Furthermore, EOAD subjects had the highest progression rates as measured with ADAS slopes during the longitudinal follow-up of 36 months. Striking heterogeneities in brain atrophy patterns were observed with MCI subjects. We found clusters of stable MCI, clusters of diffuse brain atrophy with fast progression, and MCI subjects displaying similar atrophy patterns as the typical or atypical AD subjects. Bidirectional differences in structural phenotypes were found with MCI subjects involving the anterior cerebellum and the frontal cortex. The diversity of the MCI subjects suggests that the structural phenotypes of MCI subjects would deserve a more detailed investigation with a significantly larger cohort. Our results demonstrate that the hierarchical agglomerative clustering method is an efficient tool in dividing a cohort of subjects with gray matter atrophy into coherent clusters manifesting different structural phenotypes.

Abstract Introduction Glucose hypometabolism and tau formation are key features of Alzheimer's disease (AD). Less is known about the relationship between fasting glucose and regional tau accumulation. Methods Cerebrospinal fluid (CSF) glucose was linearly regressed on regional tau (flortaucipir) among 169 Alzheimer's Disease Neuroimaging Initiative (ADNI3) participants. Flortaucipir uptake was examined by Braak stages and regions of interest (ROIs). Interactions were explored between CSF glucose and AD risk factors including regional amyloid beta (Aβ), sex, Apolipoprotein E ε4 (APOEε4) status, AD parental family history (AD FH), and cognitive impairment (CI). Results Interactions found higher CSF glucose tracked less tau in ROIs or Braak stages I/II (women, APOE ε4+, regional Aβ), III/IV (AD FH+, regional Aβ), and V/VI (AD FH+). CI drove Braak III‐VI associations. Discussion Among women and APOE ε4 carriers, higher CSF glucose tracked less early‐stage tau. Higher CSF glucose may reflect compensation against tau spreading in CI, Aβ+, or AD FH+.

Abstract Introduction Composite scores may be useful to summarize overall language or visuospatial functioning in studies of older adults. Methods We used item response theory to derive composite measures for language (ADNI‐Lan) and visuospatial functioning (ADNI‐VS) from the cognitive battery administered in the Alzheimer's Disease Neuroimaging Initiative (ADNI). We evaluated the scores among groups of people with normal cognition, mild cognitive impairment (MCI), and Alzheimer's disease (AD) in terms of responsiveness to change, association with imaging findings, and ability to differentiate between MCI participants who progressed to AD dementia and those who did not progress. Results ADNI‐Lan and ADNI‐VS were able to detect change over time and predict conversion from MCI to AD. They were associated with most of the pre‐specified magnetic resonance imaging measures. ADNI‐Lan had strong associations with a cerebrospinal fluid biomarker pattern. Discussion ADNI‐Lan and ADNI‐VS may be useful composites for language and visuospatial functioning in ADNI.

Abstract Introduction Use of cognitive composites as primary outcome measures is increasingly common in clinical trials of preclinical and prodromal Alzheimer's disease (AD). Composite outcomes can decrease intra‐individual variability, resulting in improved sensitivity to detect longitudinal change and increased statistical power. We developed a novel composite outcome, the ADAS‐Cog‐Exec, for use in the EXERT trial—a Phase 3 randomized, controlled, 12‐month exercise intervention in mild cognitive impairment (MCI). Methods Three combinations of cognitive measures selected from the Alzheimer's Disease Assessment Scale‐Cognitive Subscale version 13 (ADAS‐Cog13), tests of executive function, and the Clinical Dementia Rating (CDR) were created based on previously documented sensitivity to longitudinal change in MCI and to the effects of exercise. Optimally weighted composites of each combination were modeled using data from the ADNI‐1 MCI cohort. Ten‐fold cross‐validation was performed to obtain a bias‐corrected mean to standard deviation ratio (MSDR). The cognitive composites were assessed for their sensitivity to detect 12‐month change in MCI. Results The MSDR of 12‐month change for each of the composite outcomes tested exceeded that of the ADAS‐Cog13 total score. The composite with the highest MSDR (MSDR = 0.48) and associated statistical power included scores on ADAS‐Cog13 Word Recall, Delayed Word Recall, Orientation, and Number Cancellation subtests; Trail‐Making Tests A & B, Digit Symbol Substitution and Category Fluency; and cognitive components of the CDR (Memory, Orientation, Judgement & Problem Solving). Discussion An optimally weighted cognitive composite measure was identified and validated for use in EXERT. This composite contained selected subtests from the ADAS‐Cog13, additional measures of executive function, and box scores for cognitive components of the CDR. Because this composite score demonstrated high sensitivity to longitudinal change in MCI it will be used as the primary outcome measure for the EXERT trial.

Abstract Introduction Selecting individuals at high risk of Alzheimer's disease (AD) dementia and using the most sensitive outcome measures are important aspects of trial design. Methods We divided participants from Alzheimer's Disease Neuroimaging Initiative at the 50th percentile of the predicted absolute risk of the polygenic hazard score (PHS). Outcome measures were the Alzheimer's Disease Assessment Schedule‐Cognitive Subscale (ADAS‐Cog), ADNI‐Mem, Clinical Dementia Rating‐Sum of Boxes (CDR SB), and Cognitive Function Composite 2 (CFC2). In addition to modeling, we use a power analysis compare numbers needed with each technique. Results Data from 188 cognitively normal and 319 mild cognitively impaired (MCI) participants were analyzed. Using the ADAS‐Cog to estimate sample sizes, without stratification over 24 months, would require 930 participants with MCI, while using the CFC2 and restricting participants to those in the upper 50th percentile would require only 284 participants. Discussion Combining stratification by PHS and selection of a sensitive combined outcome measure in a cohort of patients with MCI can allow trial design that is more efficient, potentially less burdensome on participants, and more cost effective.

Abstract Introduction Hippocampal volume (HV) and cortical thickness are commonly used imaging biomarkers in Alzheimer's disease (AD) trials, and may have utility as selection criteria for enrichment strategies. Atrophy rates of these measures, in the high‐risk apolipoprotein E (APOE) ε4/ε4 homozygous AD subjects are unknown. Methods Data from Alzheimer's Disease Neuroimaging Initiative (ADNI‐1) and a tramiprosate trial were analyzed in APOE ε4/ε4 and APOE ε3/ε3 subjects with mild cognitive impairment (MCI) or mild AD. Magnetic resonance imaging (MRI) data were centrally processed using FreeSurfer; total HV and composite average cortical thickness were derived and adjusted for age, head size, and education. Volumetric changes from baseline were assessed using Boundary Shift Integral, and correlated with cognitive changes. Results APOE ε4/ε4 MCI subjects showed significantly higher % HV atrophy and cortical thinning at 12 months (4.4%, 3.1%, n = 29) compared to APOE ε3/ε3 subjects (2.8%, 1.8%, n = 93) and similarly in mild AD (7.4%, 4.7% n = 21 vs 5.4%, 3.3% n = 29). Differences were all significant at 24 months. Over 24 months, HV atrophy and cortical thinning correlated significantly with Alzheimer's Disease Assessment Scale‐Cognitive subscale worsening in APOE ε4/ε4 MCI subjects, but not in mild AD. Discussion Correlation of volumetric measures to cognitive change in APOE ε4/ε4 subjects with early AD supports their role as efficacy biomarkers. If confirmed in a Phase 3 trial with ALZ‐801 (pro‐drug of tramiprosate) in APOE ε4/ε4 early AD subjects, it may allow their use as surrogate outcomes in future treatment or prevention trials in AD.

Abstract Introduction Abnormal brain amyloid beta (Aβ) is typically assessed in vivo using global concentrations from cerebrospinal fluid and positron emission tomography (PET). However, it is unknown whether the assessment of the topographical distribution of Aβ pathology can provide additional information to identify, among global Aβ positive individuals, those destined for dementia. Methods We studied 260 amnestic mild cognitive impairment (MCI) subjects who were Aβ‐PET positive with [18F]florbetapir. Using [18F]florbetapir, we assessed the percentage of voxels sowing Aβ abnormality as well as the standardized uptake value ratio (SUVR) values across brain regions. Regressions tested the predictive effect of Aβ on progression to dementia over 2 years. Results Neither global nor regional [18F]florbetapir SUVR concentrations predicted progression to dementia. In contrast, the spatial extent of Aβ pathology in regions comprising the default mode network was highly associated with the development of dementia over 2 years. Discussion These results highlight that the regional distribution of Aβ abnormality may provide important complementary information at an individual level regarding the likelihood of Aβ positive MCI to progress to dementia.

Abstract Introduction: The female advantage in brain metabolic function may confer cognitive resilience against Alzheimer's disease (AD). Methods: A total of 1259 participants (44% women; 52% mild cognitive impairment; 18% AD) aged 55 to 90 from the Alzheimer's Disease Neuroimaging Initiative (ANDI) completed tests of global cognition, verbal memory, and executive function, and neuroimaging assessments of regional glucose metabolism, hippocampal volume (HV), and amyloid beta (Aβ). We examined sex differences in brain metabolism and cognition by AD biomarker quartiles (Aβ, HV). We then examined if metabolism mediates sex differences in cognition. Results: Metabolism was higher in women versus men when pathology was mild‐to‐moderate (quartiles 2 to 3). Women outperformed men on all cognitive outcomes at ≥1 biomarker quartile, reflecting minimal‐to‐moderate pathology; however, these differences were eliminated/attenuated after adjusting for metabolism. The female advantage in verbal memory was also observed at minimal pathology quartiles but was unchanged after metabolism adjustment. Discussion: Women's greater brain metabolism may confer cognitive resilience against early AD.

Abstract Introduction Loss of entorhinal cortex (EC) layer II neurons represents the earliest Alzheimer's disease (AD) lesion in the brain. Research suggests differing functional roles between two EC subregions, the anterolateral EC (aLEC) and the posteromedial EC (pMEC). Methods We use joint label fusion to obtain aLEC and pMEC cortical thickness measurements from serial magnetic resonance imaging scans of 775 ADNI‐1 participants (219 healthy; 380 mild cognitive impairment; 176 AD) and use linear mixed‐effects models to analyze longitudinal associations among cortical thickness, disease status, and cognitive measures. Results Group status is reliably predicted by aLEC thickness, which also exhibits greater associations with cognitive outcomes than does pMEC thickness. Change in aLEC thickness is also associated with cerebrospinal fluid amyloid and tau levels. Discussion Thinning of aLEC is a sensitive structural biomarker that changes over short durations in the course of AD and tracks disease severity—it is a strong candidate biomarker for detection of early AD.

Abstract Introduction Positron emission tomography targeting tau (tau‐PET) is a promising diagnostic tool for the identification of Alzheimer's disease (AD). Currently available data rely on quantitative measures, and a visual interpretation method, critical for clinical translation, is needed. Methods We developed a visual interpretation method for 18F‐flortaucipir tau‐PET and tested it on 274 individuals (cognitively normal controls, patients with mild cognitive impairment [MCI], AD dementia, and non‐AD diagnoses). Two readers interpreted 18F‐flortaucipir PET using two complementary indices: a global visual score and a visual distribution pattern. Results Global visual scores were reliable, correlated with global cortical 18F‐flortaucipir standardized uptake value ratio (SUVR) and were associated with clinical diagnosis and amyloid status. The AD‐like 18F‐flortaucipir pattern had good sensitivity and specificity to identify amyloid‐positive patients with AD dementia or MCI. Discussion This 18F‐flortaucipir visual rating scheme is associated with SUVR quantification, clinical diagnosis, and amyloid status, and constitutes a promising approach to tau measurement in clinical settings.

Abstract Introduction Cognitive decline in Alzheimer's disease is associated with amyloid beta (Aβ) accumulation, neurodegeneration, and cerebral small vessel disease, but the temporal relationships among these factors is not well established. Methods Data included white matter hyperintensity (WMH) load, gray matter (GM) atrophy and Alzheimer's Disease Assessment Scale‐Cognitive‐Plus (ADAS13) scores for 720 participants and cerebrospinal fluid amyloid (Aβ1–42) for 461 participants from the Alzheimer's Disease Neuroimaging Initiative. Linear regressions were used to assess the relationships among baseline WMH, GM, and Aβ1–42 to changes in WMH, GM, Aβ1–42, and cognition at 1‐year follow‐up. Results Baseline WMHs and Aβ1–42 predicted WMH increase and GM atrophy. Baseline WMHs and Aβ1–42 predicted worsening cognition. Only baseline Aβ1–42 predicted change in Aβ1–42. Discussion Baseline WMHs lead to greater future GM atrophy and cognitive decline, suggesting that WM damage precedes neurodegeneration and cognitive decline. Baseline Aβ1–42 predicted WMH increase, suggesting a potential role of amyloid in WM damage.

Background There is a need for outcome measures with improved responsiveness to changes in pre-dementia populations. Both cognitive and motor function play important roles in neurodegeneration; motor function decline is detectable at early stages of cognitive decline. This proof of principle study used a Pooled Index approach to evaluate improved responsiveness of the predominant outcome measure (ADAS-Cog: Alzheimer’s Disease Assessment Scale-Cognitive Subscale) when assessment of motor function is added. Methods Candidate Pooled Index variables were selected based on theoretical importance and pairwise correlation coefficients. Kruskal-Wallis and Mann-Whitney U tests assessed baseline discrimination. Standardized response means assessed responsiveness to longitudinal change. Results Final selected variables for the Pooled Index include gait velocity, dual-task cost of gait velocity, and an ADAS-Cog-Proxy (statistical approximation of the ADAS-Cog using similar cognitive tests). The Pooled Index and ADAS-Cog-Proxy scores had similar ability to discriminate between pre-dementia syndromes. The Pooled Index demonstrated trends of similar or greater responsiveness to longitudinal decline than ADAS-Cog-Proxy scores. Conclusion Adding motor function assessments to the ADAS-Cog may improve responsiveness in pre-dementia populations.

An amendment to this paper has been published and can be accessed via the original article.

Abstract Imaging genetics deals with relationships between genetic variation and imaging variables, often in a disease context. The complex relationships between brain volumes and genetic variants have been explored with both dimension reduction methods and model-based approaches. However, these models usually do not make use of the extensive knowledge of the spatio-anatomical patterns of gene activity. We present a method for integrating genetic markers (single nucleotide polymorphisms) and imaging features, which is based on a causal model and, at the same time, uses the power of dimension reduction. We use structural equation models to find latent variables that explain brain volume changes in a disease context, and which are in turn affected by genetic variants. We make use of publicly available spatial transcriptome data from the Allen Human Brain Atlas to specify the model structure, which reduces noise and improves interpretability. The model is tested in a simulation setting and applied on a case study of the Alzheimer’s Disease Neuroimaging Initiative.

Abstract Background Individuals with subjective memory complaints (SMC) feature a higher risk of cognitive decline and clinical progression of Alzheimer’s disease (AD). However, the pathological mechanism underlying SMC remains unclear. We aimed to assess the intrinsic connectivity network and its relationship with AD-related pathologies in SMC individuals. Methods We included 44 SMC individuals and 40 normal controls who underwent both resting-state functional MRI and positron emission tomography (PET). Based on graph theory approaches, we detected local and global functional connectivity across the whole brain by using degree centrality (DC) and eigenvector centrality (EC) respectively. Additionally, we analyzed amyloid deposition and tauopathy via florbetapir-PET imaging and cerebrospinal fluid (CSF) data. The voxel-wise two-sample T-test analysis was used to examine between-group differences in the intrinsic functional network and cerebral amyloid deposition. Then, we correlated these network metrics with pathological results. Results The SMC individuals showed higher DC in the bilateral hippocampus (HP) and left fusiform gyrus and lower DC in the inferior parietal region than controls. Across all subjects, the DC of the bilateral HP and left fusiform gyrus was positively associated with total tau and phosphorylated tau181. However, no significant between-group difference existed in EC and cerebral amyloid deposition. Conclusion We found impaired local, but not global, intrinsic connectivity networks in SMC individuals. Given the relationships between DC value and tau level, we hypothesized that functional changes in SMC individuals might relate to pathological biomarkers.

Abstract Background Alzheimer’s disease (AD) is one of the most common neurodegenerative diseases that causes problems related to brain function. To some extent it is understood on a molecular level how AD arises, however there are a lack of biomarkers that can be used for early diagnosis. Two popular methods to identify AD-related biomarkers use genetics and neuroimaging. Genes and neuroimaging phenotypes have provided some insights as to the potential for AD biomarkers. While the field of imaging-genomics has identified genetic features associated with structural and functional neuroimaging phenotypes, it remains unclear how variants that affect splicing could be important for understanding the genetic etiology of AD. Methods In this study, rare variants (minor allele frequency

Abstract Background Multi-site neuroimaging offer several benefits and poses tough challenges in the drug development process. Although MRI protocol and clinical guidelines developed to address these challenges recommend the use of good quality images, reliable assessment of image quality is hampered by the several shortcomings of existing techniques. Methods Given a test image two feature images are extracted. They are grayscale and contrast feature images. Four binary images are generated by setting four different global thresholds on the feature images. Image quality is predicted by measuring the structural similarity between appropriate pairs of binary images. The lower and upper limits of the quality index are 0 and 1. Quality prediction is based on four quality attributes; luminance contrast, texture, texture contrast and lightness. Results Performance evaluation on test data from three multi-site clinical trials show good objective quality evaluation across MRI sequences, levels of distortion and quality attributes. Correlation with subjective evaluation by human observers is ≥ 0.6. Conclusion The results are promising for the evaluation of MRI protocols, specifically the standardization of quality index, designed to overcome the challenges encountered in multi-site clinical trials.

The spatial distribution of atrophy in neurodegenerative diseases suggests that brain connectivity mediates disease propagation. Different descriptors of the connectivity graph potentially relate to different underlying mechanisms of propagation. Previous approaches for evaluating the influence of connectivity on neurodegeneration consider each descriptor in isolation and match predictions against late-stage atrophy patterns. We introduce the notion of a topological profile — a characteristic combination of topological descriptors that best describes the propagation of pathology in a particular disease. By drawing on recent advances in disease progression modeling, we estimate topological profiles from the full course of pathology accumulation, at both cohort and individual levels. Experimental results comparing topological profiles for Alzheimer’s disease, multiple sclerosis and normal ageing show that topological profiles explain the observed data better than single descriptors. Within each condition, most individual profiles cluster around the cohort-level profile, and individuals whose profiles align more closely with other cohort-level profiles show features of that cohort. The cohort-level profiles suggest new insights into the biological mechanisms underlying pathology propagation in each disease.