Alzheimer's Disease Neuroimaging Initiative Database Research Articles

The SNP rs6859 in NECTIN2 gene is associated with underlying heterogeneous trajectories of cognitive changes in older adults

BackgroundFunctional decline associated with dementia, including in Alzheimer’s disease (AD), is not uniform across individuals, and respective heterogeneity is not yet fully explained. Such heterogeneity may in part be related to genetic variability among individuals. In this study, we investigated whether the SNP rs6859 in nectin cell adhesion molecule 2 (NECTIN2) gene (a major risk factor for AD) influences trajectories of cognitive decline in older participants from the Alzheimer’s Disease Neuroimaging Initiative (ADNI).MethodsWe retrospectively analyzed records on 1310 participants from the ADNI database for the multivariate analysis. We used longitudinal measures of Mini-Mental State Examination (MMSE) scores in participants, who were cognitively normal, or having AD, or other cognitive deficits to investigate the trajectories of cognitive changes. Multiple linear regression, linear mixed models and latent class analyses were conducted to investigate the association of the SNP rs6859 with MMSE.ResultsThe regression coefficient per one allele dose of the SNP rs6859 was independently associated with MMSE in both cross-sectional (-2.23, p < 0.01) and linear mixed models (-2.26, p < 0.01) analyses. The latent class model with three distinct subgroups (class 1: stable and gradual decline, class 2: intermediate and late decline, and class 3: lowest and irregular) performed best in the posterior classification, 42.67% (n = 559), 21.45% (n = 281), 35.88% (n = 470) were classified as class 1, class 2, and class 3. In the heterogeneous linear mixed model, the regression coefficient per one allele dose of rs6859 – A risk allele was significantly associated with MMSE class 1 and class 2 memberships and related decline; Class 1 (-2.28, 95% CI: -4.05, -0.50, p < 0.05), Class 2 (-5.56, 95% CI: -9.61, -1.51, p < 0.01) and Class 3 (-0.37, 95% CI: -1.62, 0.87, p = 0.55).ConclusionsThis study found statistical evidence supporting the classification of three latent subclass groups representing complex MMSE trajectories in the ADNI cohort. The SNP rs6859 can be suggested as a candidate genetic predictor of variation in modeling MMSE trajectory, as well as for identifying latent classes with higher baseline MMSE. Functional studies may help further elucidate this relationship.

Association of creatine kinase with cognition, neuroimaging, and Alzheimer’s pathology

AbstractBackgroundCreatine kinase (CK) and its related metabolites are known to imply cardiovascular system disease, whereas the cardiovascular metabolic state is associated with the state of cognitive well‐being of the brain. However, the extent to which CK levels is related with cognitive impairment has not been revealed.MethodThe Alzheimer’s Disease Neuroimaging Initiative (ADNI) database were used for this study. We first investigated the intergroup difference of participants demographic, cognitive status, Alzheimer’s disease (AD) diagnosis, and cerebrospinal fluid (CSF) AD pathology biomarkers. Next, we conducted to use liner regression model to examine the association of CK with CSF AD pathology biomarkers, cognition, and brain structure. Last, mediation analyses were utilized to explore whether the association between CK and cognition were mediated by CSF AD pathology biomarkers.ResultA total of 1,607 participants (mean age were 73.82±7.23 and 55.76% were male) from the ADNI database were involved in this cross‐sectional study. We found a significant difference through age, gender, AD diagnosis, and cognitive status subgroups in intergroup difference analysis (P&lt;0.05). Meanwhile, significant association of lower level of CK with decreased level of β‐amyloid1‐42 (Aβ; β = 0.069, P = 0.040) and increased level of phosphorylated tau (P‐tau; β = ‐0.084, P = 0.007), and total‐tau (T‐tau; β = ‐0.071, P = 0.021), respectively. Liner regression analysis also indicated that the lower level of CK was associated with worse cognition (Mini‐Mental State Examination: β = 0.059, P = 0.021; Alzheimer’s Disease Assessment Scale: β = ‐0.065, P = 0.012), however not to brain structure. Moreover, mediation analysis identified CSF AD pathology biomarkers mediated the association between level of CK and cognition.ConclusionOur findings offered detailed proof to advise that the lower level of CK was significantly associated with poor cognition and it might be mediated by CSF AD pathology biomarkers. For future research, it is necessary to identify the precise reason and process by which changed levels of CK impact cognition.

Elevated Homocysteine Levels and Hypertension Relate to Cognitive Impairment via Increased White Matter Hyperintensity Volume.

Recent studies have identified a relationship between elevated homocysteine levels and hypertension (HTN) with Alzheimer's disease (AD), but its pathogenesis remains unclear. To evaluate elevated homocysteine levels and HTN as risk factors for cognitive impairment (CI) and determine their relationship with white matter hyperintensity (WMH) volume. A total of 521 subjects were selected from the Alzheimer's Disease Neuroimaging Initiative (ADNI) database and divided into two groups according to the diagnostic criteria of the ADNI database. The CI group included 370 subjects, consisting of 122 with AD and 248 with mild CI, while the cognitively normal (CN) group contained 151 subjects. The history of HTN, homocysteine levels, WMH volume and Mini-Mental State Examination (MMSE) scores were analyzed. The study found that patients with CI had higher homocysteine levels than those with CN. Additionally, WMH volume was significantly correlated with homocysteine levels in CI patients, and MMSE scores decreased as WMH volume increased. Further analysis revealed that CI patients with HTN had significantly higher homocysteine levels than those without HTN. Furthermore, the correlation between WMH volume and homocysteine levels was significant only in CI patients with HTN and not in those without HTN. In CN patients, there was no correlation between WMH volume and homocysteine levels in either the HTN or non-HTN groups, and no difference was observed in homocysteine levels. It is indicated that elevated homocysteine levels in conjunction with HTN are associated with the increased volume of WMHs and CI.

Prediction of Conversion from Mild Cognitive Impairment to Alzheimer's Disease Using Amyloid PET and Brain MR Imaging Data: A 48-Month Follow-Up Analysis of the Alzheimer's Disease Neuroimaging Initiative Cohort.

We developed a novel quantification method named "shape feature" by combining the features of amyloid positron emission tomography (PET) and brain magnetic resonance imaging (MRI) and evaluated its significance in predicting the conversion from mild cognitive impairment (MCI) to Alzheimer's disease (AD) in the Alzheimer's Disease Neuroimaging Initiative (ADNI) cohort. From the ADNI database, 334 patients with MCI were included. The brain amyloid smoothing score (AV45_BASS) and brain atrophy index (MR_BAI) were calculated using the surface area and volume of the region of interest in AV45 PET and MRI. During the 48-month follow-up period, 108 (32.3%) patients converted from MCI to AD. Age, Mini-Mental State Examination (MMSE), cognitive subscale of the Alzheimer's Disease Assessment Scale (ADAS-cog), apolipoprotein E (APOE), standardized uptake value ratio (SUVR), AV45_BASS, MR_BAI, and shape feature were significantly different between converters and non-converters. Univariate analysis showed that age, MMSE, ADAS-cog, APOE, SUVR, AV45_BASS, MR_BAI, and shape feature were correlated with the conversion to AD. In multivariate analyses, high shape feature, SUVR, and ADAS-cog values were associated with an increased risk of conversion to AD. In patients with MCI in the ADNI cohort, our quantification method was the strongest prognostic factor for predicting their conversion to AD.

An ontology-based approach for modelling and querying Alzheimer’s disease data

BackgroundThe recent advances in biotechnology and computer science have led to an ever-increasing availability of public biomedical data distributed in large databases worldwide. However, these data collections are far from being “standardized” so to be harmonized or even integrated, making it impossible to fully exploit the latest machine learning technologies for the analysis of data themselves. Hence, facing this huge flow of biomedical data is a challenging task for researchers and clinicians due to their complexity and high heterogeneity. This is the case of neurodegenerative diseases and the Alzheimer’s Disease (AD) in whose context specialized data collections such as the one by the Alzheimer’s Disease Neuroimaging Initiative (ADNI) are maintained.MethodsOntologies are controlled vocabularies that allow the semantics of data and their relationships in a given domain to be represented. They are often exploited to aid knowledge and data management in healthcare research. Computational Ontologies are the result of the combination of data management systems and traditional ontologies. Our approach is i) to define a computational ontology representing a logic-based formal conceptual model of the ADNI data collection and ii) to provide a means for populating the ontology with the actual data in the Alzheimer Disease Neuroimaging Initiative (ADNI). These two components make it possible to semantically query the ADNI database in order to support data extraction in a more intuitive manner.ResultsWe developed: i) a detailed computational ontology for clinical multimodal datasets from the ADNI repository in order to simplify the access to these data; ii) a means for populating this ontology with the actual ADNI data. Such computational ontology immediately makes it possible to facilitate complex queries to the ADNI files, obtaining new diagnostic knowledge about Alzheimer’s disease.ConclusionsThe proposed ontology will improve the access to the ADNI dataset, allowing queries to extract multivariate datasets to perform multidimensional and longitudinal statistical analyses. Moreover, the proposed ontology can be a candidate for supporting the design and implementation of new information systems for the collection and management of AD data and metadata, and for being a reference point for harmonizing or integrating data residing in different sources.

Association Between False Memories and Delusions in Alzheimer Disease

Understanding the mechanisms of delusion formation in Alzheimer disease (AD) could inform the development of therapeutic interventions. It has been suggested that delusions arise as a consequence of false memories. To investigate whether delusions in AD are associated with false recognition, and whether higher rates of false recognition and the presence of delusions are associated with lower regional brain volumes in the same brain regions. Since the Alzheimer's Disease Neuroimaging Initiative (ADNI) launched in 2004, it has amassed an archive of longitudinal behavioral and biomarker data. This cross-sectional study used data downloaded in 2020 from ADNI participants with an AD diagnosis at baseline or follow-up. Data analysis was performed between June 24, 2020, and September 21, 2021. Enrollment in the ADNI. The main outcomes included false recognition, measured with the 13-item Alzheimer's Disease Assessment Scale-Cognitive Subscale (ADAS-Cog 13) and the Rey Auditory Verbal Learning Test (RAVLT) and volume of brain regions corrected for total intracranial volume. Behavioral data were compared for individuals with delusions in AD and those without using independent-samples t tests or Mann-Whitney nonparametric tests. Significant findings were further explored using binary logistic regression modeling. For neuroimaging data region of interest analyses using t tests, Poisson regression modeling or binary logistic regression modeling and further exploratory, whole-brain voxel-based morphometry analyses were carried out to explore the association between regional brain volume and false recognition or presence of delusions. Of the 2248 individuals in the ADNI database, 728 met the inclusion criteria and were included in this study. There were 317 (43.5%) women and 411 (56.5%) men. Their mean (SD) age was 74.8 (7.4) years. The 42 participants with delusions at baseline had higher rates of false recognition on the ADAS-Cog 13 (median score, 3; IQR, 1 to 6) compared with the 549 control participants (median score, 2; IQR, 0 to 4; U = 9398.5; P = .04). False recognition was not associated with the presence of delusions when confounding variables were included in binary logistic regression models. An ADAS-Cog 13 false recognition score was inversely associated with left hippocampal volume (odds ratio [OR], 0.91 [95% CI, 0.88-0.94], P < .001), right hippocampal volume (0.94 [0.92-0.97], P < .001), left entorhinal cortex volume (0.94 [0.91-0.97], P < .001), left parahippocampal gyrus volume (0.93 [0.91-0.96], P < .001), and left fusiform gyrus volume (0.97 [0.96-0.99], P < .001). There was no overlap between locations associated with false recognition and those associated with delusions. In this cross-sectional study, false memories were not associated with the presence of delusions after accounting for confounding variables, and no indication for overlap of neural networks for false memories and delusions was observed on volumetric neuroimaging. These findings suggest that delusions in AD do not arise as a direct consequence of misremembering, lending weight to ongoing attempts to delineate specific therapeutic targets for treatment of psychosis.

Multisite harmonization of diffusion tensor image analysis along the perivascular space using the COMBined Association Test

PurposeThis multisite study aimed to use the COMBined Association Test (COMBAT), a harmonization technique that uses regression of covariates with an empirical Bayesian framework, to harmonize diffusion tensor image analysis along the perivascular space (DTI-ALPS) variations caused by scanner, site, and protocol differences.Materials and methodsThis study included multisite diffusion magnetic resonance imaging (dMRI) data of 45 patients with Alzheimer’s disease (AD) and 82 cognitively normal (CN) participants from the AD neuroimaging initiative database. The dMRI data were obtained with two b values (0 and 1000 s/mm2) from 27 institutions and three different 3-Tesla MRI scanners (two vendors). The ALPS index was calculated from multisite dMRI data, and COMBAT was used to harmonize the factors causing site variations. Welch’s t test was used, Cohen’s d was calculated to compare the difference in the ALPS index between AD and CN before and after harmonization, and Pearson’s correlation coefficient was calculated to assess the relationships between the ALPS index and the cognitive score, [18F] fluorodeoxyglucose (FDG)-positron emission tomography (PET), and [18F] florbetapir (AV45)-PET standardized uptake value ratios (SUVRs).ResultsCOMBAT harmonized scanner differences and increased Cohen’s d of the left and right ALPS indexes between AD and CN from 0.288 to 0.438 and 0.328 to 0.480, respectively. The ALPS indexes were significantly different between AD and CN after harmonization (P < 0.05) but not before it. Moreover, Pearson’s correlation coefficients between the ALPS index and cognitive score, FDG-PET, and AV45-PET SUVRs were higher after harmonization than before it.ConclusionThis study demonstrates the application of COMBAT harmonization to eliminate between-scanner, site, and protocol variations in the ALPS index calculated from DTI-ALPS using dMRI and possibly facilitate the use of the ALPS index in multi-center studies.

A Comprehensive Investigation of the Potential Role of Lipoproteins and Metabolite Profile as Biomarkers of Alzheimer's Disease Compared to the Known CSF Biomarkers.

While cerebrospinal fluid (CSF) core biomarkers have been considered diagnostic biomarkers for a long time, special attention has been recently dedicated to lipoproteins and metabolites that could be potentially associated with Alzheimer's disease (AD) neurodegeneration. Herein, we aimed to investigate the relationship between the levels of CSF core biomarkers including Aβ-42, TAU, and P-TAU and plasma lipoproteins and metabolites of patients with AD from the baseline cohort of the Alzheimer's Disease Neuroimaging Initiative (ADNI) database. Using the ADNI database, fourteen subclasses of lipoproteins as well as a number of lipids and fatty acids and low-molecular metabolites including amino acids, ketone bodies, and glycolysis-related metabolites in blood samples were measured as potential noninvasive markers, and their association with the CSF core biomarkers was statistically investigated controlling for age and gender. A total number of 251 AD subjects were included, among whom 71 subjects were negative for the Apo-E ε4 allele and 150 were positive. There was no significant difference between the two groups regarding cognitive assessments, CSF core biomarkers, and lipoproteins and metabolites except the level of Aβ-42 (p < 0.001) and phenylalanine (p = 0.049), which were higher in the negative group. CSF TAU and P-TAU were significantly correlated with medium and small HDL in the negative group, and with extremely large VLDL in the positive group. Our results also indicated significant correlations of metabolites including unsaturated fatty acids, glycerol, and leucine with CSF core biomarkers. Based on our findings, a number of lipoproteins and metabolites were associated with CSF core biomarkers of AD. These correlations showed some differences in Apo-E ε4 positive and negative groups, which reminds the role of Apo-E gene status in the pathophysiology of AD development. However, further research is warranted to explore the exact association of lipoproteins and other metabolites with AD core biomarkers and pathology.

Genetic architecture of hippocampus subfields volumes in Alzheimer's disease.

The hippocampus is a heterogeneous structure, comprising histologically and functionally distinguishable hippocampal subfields. The volume reductions in hippocampal subfields have been demonstrated to be linked with Alzheimer's disease (AD). The aim of our study is to investigate the hippocampal subfields' genetic architecture based on the Alzheimer's Disease Neuroimaging Initiative (ADNI) data set. After preprocessing the downloaded genetic variants and imaging data from the ADNI database, a co-sparse reduced rank regression model was applied to analyze the genetic architecture of hippocampal subfields volumes. Homology modeling, docking, molecular dynamics simulations, and Co-IP experiments for protein-protein interactions were used to verify the function of target protein on hippocampal subfields successively. After that, the association analysis between the candidated genes on the hippocampal subfields volume and clinical scales were performed. The results of the association analysis revealed five unique genetic variants (e.g., ubiquitin-specific protease 10 [USP10]) changed in nine hippocampal subfields (e.g., the granule cell and molecular layer of the dentate gyrus [GC-ML-DG]). Among five genetic variants, USP10 had the strongest interaction effect with BACE1, which affected hippocampal subfields verified by MD and Co-IP experiments. The results of association analysis between the candidated genes on the hippocampal subfields volume and clinical scales showed that candidated genes influenced the volume and function of hippocampal subfields. Current evidence suggests that hippocampal subfields have partly distinct genetic architecture and may improve the sensitivity of the detection of AD.

Polymorphisms in the WWOX gene are associated with brain glucose metabolism and influence cerebrospinal fluid p‐tau181 levels

AbstractBackgroundGenome‐wide association studies (GWAS) studies can be a powerful tool for identifying new single nucleotide polymorphisms (SNPs) associated with multifactorial diseases. Previous GWAS have identified more than 20 genes as potential factors of genetic susceptibility in Alzheimer’s disease (AD). One of them is the protein coding gene WW Domain Containing Oxidoreductase (WWOX), which is highly expressed in mature astrocytes and neurons. Here, we investigated the association between WWOX SNPs, 18F‐FDG PET neuroimaging and cerebrospinal fluid (CSF) levels of tau phosphorylated at threonine 181 (p‐tau181) in elderly individuals from the Alzheimer’s Disease Neuroimaging Initiative (ADNI).MethodsData was acquired from 537 participants (CU= 195; MCI= 314; AD= 28) of the ADNI database. The brain glucose metabolism was estimated using the standard uptake value ratio (SUVr, cerebellum as reference) of FDG‐PET. The presence of the SNPs were integrated to FDG PET images using voxel‐wise generalized linear regression models adjusted for age, gender, diagnostic status, and years of education (RMINC package). Through the Wilcoxon test we determined if the SNPs presence influenced p‐tau181 levels in CSF. All statistical analyzes were performed in the R statistical computing software (version 4.1.2) (significance level of p &lt; 0.05).ResultsSixty‐seven WWOX SNPs were statistically associated with 18F‐FDG PET signal. Voxel‐wise analyses demonstrated that the top ten WWOX SNPs with the highest number of voxels altered in the FDG‐PET images were associated with cortical hypermetabolism (Figure 1, top three). Thirteen WWOX SNPs were significantly associated with lower p‐tau181 concentration in the CSF, including rs7193539 (p = 0.0004), rs7206890 (p = 0.0061), and rs11859281 (p = 0.0009), compared to the reference genotype.ConclusionSeveral SNPs of the WWOX gene were significantly correlated to brain metabolism. The presence of the SNPs can influence p‐tau181 levels, suggesting a role of WWOX in AD, as reported in other previous studies. Our findings suggest a protective effect of these WWOX gene variants. Additional studies are needed to better understand the biological implications of our findings.

Asynchronous functional linear regression models for longitudinal data in reproducing kernel Hilbert space.

Motivated by the analysis of longitudinal neuroimaging studies, we study the longitudinal functional linear regression model under asynchronous data setting for modeling the association between clinical outcomes and functional (or imaging) covariates. In the asynchronous data setting, both covariates and responses may be measured at irregular and mismatched time points, posing methodological challenges to existing statistical methods. We develop a kernel weighted loss function with roughness penalty to obtain the functional estimator and derive its representer theorem. The rate of convergence, a Bahadur representation, and the asymptotic pointwise distribution of the functional estimator are obtained under the reproducing kernel Hilbert space framework. We propose a penalized likelihood ratio test to test the nullity of the functional coefficient, derive its asymptotic distribution under the null hypothesis, and investigate the separation rate under the alternative hypotheses. Simulation studies are conducted to examine the finite-sample performance of the proposed procedure. We apply the proposed methods to the analysis of multitype data obtained from the Alzheimer's Disease Neuroimaging Initiative (ADNI) study, which reveals significant association between 21 regional brain volume density curves and the cognitive function. Data used in preparation of this paper were obtained from the ADNI database (adni.loni.usc.edu).

BPGAN: Brain PET synthesis from MRI using generative adversarial network for multi-modal Alzheimer’s disease diagnosis

Background and ObjectiveMulti-modal medical images, such as magnetic resonance imaging (MRI) and positron emission tomography (PET), have been widely used for the diagnosis of brain disorder diseases like Alzheimer’s disease (AD) since they can provide various information. PET scans can detect cellular changes in organs and tissues earlier than MRI. Unlike MRI, PET data is difficult to acquire due to cost, radiation, or other limitations. Moreover, PET data is missing for many subjects in the Alzheimer’s Disease Neuroimaging Initiative (ADNI) dataset. To solve this problem, a 3D end-to-end generative adversarial network (named BPGAN) is proposed to synthesize brain PET from MRI scans, which can be used as a potential data completion scheme for multi-modal medical image research. MethodsWe propose BPGAN, which learns an end-to-end mapping function to transform the input MRI scans to their underlying PET scans. First, we design a 3D multiple convolution U-Net (MCU) generator architecture to improve the visual quality of synthetic results while preserving the diverse brain structures of different subjects. By further employing a 3D gradient profile (GP) loss and structural similarity index measure (SSIM) loss, the synthetic PET scans have higher-similarity to the ground truth. In this study, we explore alternative data partitioning ways to study their impact on the performance of the proposed method in different medical scenarios. ResultsWe conduct experiments on a publicly available ADNI database. The proposed BPGAN is evaluated by mean absolute error (MAE), peak-signal-to-noise-ratio (PSNR) and SSIM, superior to other compared models in these quantitative evaluation metrics. Qualitative evaluations also validate the effectiveness of our approach. Additionally, combined with MRI and our synthetic PET scans, the accuracies of multi-class AD diagnosis on dataset-A and dataset-B are 85.00% and 56.47%, which have been improved by about 1% and 1%, respectively, compared to the stand-alone MRI. ConclusionsThe experimental results of quantitative measures, qualitative displays, and classification evaluation demonstrate that the synthetic PET images by BPGAN are reasonable and high-quality, which provide complementary information to improve the performance of AD diagnosis. This work provides a valuable reference for multi-modal medical image analysis.

Robust deep neural network estimation for multi-dimensional functional data

In this paper, we propose a robust estimator for the location function from multi-dimensional functional data. The proposed estimators are based on the deep neural networks with ReLU activation function. At the meanwhile, the estimators are less susceptible to outlying observations and model-misspecification. For any multi-dimensional functional data, we provide the uniform convergence rates for the proposed robust deep neural networks estimators. Simulation studies illustrate the competitive performance of the robust deep neural network estimators on regular data and their superior performance on data that contain anomalies. The proposed method is also applied to analyze 2D and 3D images of patients with Alzheimer’s disease obtained from the Alzheimer Disease Neuroimaging Initiative database.

Longitudinal effects of behavioral interventions: A comparison using a propensity score matching model

AbstractBackgroundBehavioral interventions are often recommended for persons with mild cognitive impairment (MCI). The Healthy Action to Benefit Independence &amp; Thinking® Program (HABIT) has been shown to benefit activities of daily living (ADL) and self‐efficacy in patients at 6‐ and 12‐months post intervention. However, to date, the findings have been limited by comparison to a relatively small, untreated control group. This study attempted to explore HABIT ADL outcomes compared to a large, presumably untreated Alzheimer's Disease Neuroimaging Initiative (ADNI) MCI cohort.MethodA total of 767 participants with MCI were selected from the ADNI database, whereas 478 participants with MCI were recruited from the HABIT IRB approved research registry. Functional outcomes were assessed with informant‐based measures ‐ Functional Activities Questionnaire (FAQ) and Everyday Cognition scale (ECOG) memory and total scores – at both 6‐month‐ and 12‐month‐follow‐ups. After near‐neighbor propensity score matching with a caliper of 0.2, 288 participants from each group were matched for the FAQ, and 333 were matched for the ECOG scores. Latent growth curve models were used to test the hypothesis that the HABIT program might delay functional decline in patients with MCI.ResultAfter matching HABIT group continued to show higher initial scores on FAQ (b=.849, p=.013), and ECOG memory (b=.200, p=.004) and total scores (b=1.115, p&lt;.001) compared to the ADNI group. Although an increase in scores was observed in both groups for each functional outcome, the slopes were steeper for the HABIT group for FAQ (b=1.194, p&lt;.001), ECOG memory (b=.073, p=.019) and ECOG total (b=.383, p=.007).ConclusionContrary to our hypothesis, in this non‐randomized study, partners of HABIT participants with MCI reported more rapid functional decline compared to those in an imaging study. This discrepancy could be due to the HABIT group’s higher baseline impairment level even after matching. Furthermore, patients with perceived more rapid decline may be more likely to seek interventions relative to those who elect to participate in a non‐interventional imaging study. In addition, the intervention sessions could prime the partners to be more scrupulous and aware when rating functional assessments. Additional randomized research may help add clarity to these findings.

Analysis of rate of change in quantitative traits for identifying biomarker-mediated genetic effects on MCI and AD.

The relationship between genetic variations such as single nucleotide polymorphisms (SNPs) and quantitative traits (QTs) is integral in the understanding of Alzheimer's Disease (AD). Given AD as a progressive disorder, QT-based rate of change (RoC) measures are important phenotypes to study in AD genetics. Further mediation analysis provides a promising method to help reveal causal relationship from genetic determinants to longitudinal changes in QTs and to clinical diagnosis. We analyzed the QT data from the Quantitative Templates for the Progression of Alzheimer's disease (QT-PAD, see http://www.pi4cs.org/qt-pad-challenge). QT-PAD is a single and shared data freeze from the Alzheimer's Disease Neuroimaging Initiative (ADNI), including 16 imaging, cognitive and fluid AD QTs. RoC between baseline and month-12 visits was calculated for each QT. Genotyping data were obtained from the ADNI database and went through a standard QC procedure. Genome wide association study (GWAS) was performed using PLINK to examine genetic associations between 565,373 SNPs and the RoC for each QT. For each SNP-RoC association with p<1e-6, a subsequent mediation analysis was performed using the "Mediation" R package to examine the mediation relationship from SNP to RoC to five different diagnostic comparisons (see Table 1). Figure 1 shows the GWAS findings on the RoC measures for the studied QTs. The results included 19 SNPs and 14 QTs, where each SNP or each QT had at least one association with p<10-6 . Table 1 shows the mediation analysis results. RoC in FAQ mediates the effects of rs429358-APOE and rs4420638-APOC1 on all five diagnostic comparisons, and mediates the effect of rs769449-APOE on four comparisons. RoC in Ventricle mediates the effects of both TOMM40 SNPs on four comparisons. RoC in CDRSB mediates the effects of rs5011804-KRAS on three comparisons. Mediation analysis of longitudinal AD biomarkers revealed several candidate causal relationships from genetic determinants (i.e., SNPs from the APOE/TOMM40/APOC1 region and KRAS) to RoC measures in QTs (FAQ, Ventricle, CDRSB), and to various diagnostic comparisons in MCI and AD. The identified patterns warrant further investigation in independent cohorts for deconvoluting mechanistic complexity of AD.

Comparison of plasma and CSF biomarkers across ethnoracial groups in ADNI

AbstractBackgroundDifferences in CSF and plasma Alzheimer’s disease biomarkers between African Americans (AA), Latinos (LA), and non‐Hispanic Whites (NHW) have been reported. The Alzheimer’s Disease Neuroimaging Initiative (ADNI) may offer insights into ethnoracial biomarker differences important for understanding disease trajectory variations among diverse populations.MethodWe queried the ADNI database for all participants with available plasma pTau181 and NfL (Simoa) and linked CSF biomarkers within 1 year of blood draw (Aβ42, pTau181, T‐tau, Elecsys). Baseline data for 47 AA and 43 LA with plasma biomarkers were matched to separate groups of NHW (1:3 nearest neighbor matching on propensity score), considering age, sex, education, CDR, APOE ε4 alleles and family history of dementia (Table 1). Cross‐sectional analyses compared biomarker levels between matched cohorts without (non‐parametric Mann‐Whitney U‐test) and with adjustment for demographics, clinical, biomarker, and medical characteristics (multi‐linear regression). Statistical threshold was set at two‐tailed p&lt;0.05.ResultUnadjusted comparisons between AA and NHW revealed no significant differences in plasma NfL [29.3 (IQR: 21.6‐42.5) vs 35.5 (27.0‐49.6), p=.12], plasma pTau181 [14.5 (9.4‐22.9) vs 15.5 (10.1‐22.8), p=.83], CSF Aβ42 [933.1 (649.5‐1590.8) vs 926.7 (677.7‐1697.8), p=.83], CSF T‐tau [214.6 (152.4‐290.7) vs 251.9 (194.5‐351.4), p=.17] or CSF pTau181 [19.6 (14.4‐27.3) vs 22.7 (17.0‐33.3), p=.32] (Fig. 1a‐b). No significant differences were identified in any of the biomarkers after adjustment for relevant covariates. Unadjusted comparisons between LA and NHW revealed no significant differences in plasma NfL [36.7 (24.4‐50.4) vs 35.6 (25.6‐47.1), p=1], plasma pTau181 [18.0 (11.3‐25.0) vs 15.7 (10.9‐23.4), p=.81], or CSF Aβ42 [852.7 (787.1‐1213.0) vs 951.6 (651.1‐1530.5), p=1]. A trend towards lower CSF T‐Tau [216.5 (147.1‐270.5) vs 257.2 (197.2‐360.9), p=.076] and CSF pTau181 [19.4 (13.6‐27.9) vs 24.7 (17.5‐33.4), p=.076] in LA was observed (Fig. 2a‐b). After adjustment for covariates, LA were found to have significantly lower levels of CSF T‐tau (β=0.154, p&lt;.05) and CSF pTau181 (β=0.164, p&lt;.05), while differences in plasma markers were not significant.ConclusionNo significant biomarker differences between AA and NHW were observed. Lower CSF pTau181 and CSF T‐Tau in LA compared to NHW after covariate adjustment may represent subtle differences driven by biological factors or unmeasured sociocultural determinants (income, occupation, neighborhood environment); larger studies are needed.

Evaluation of the Cost-effectiveness of Drug Treatment for Alzheimer Disease in a Simulation Model That Includes Caregiver and Societal Factors

The possibility of widespread use of a novel effective therapy for Alzheimer disease (AD) will present important clinical, policy, and financial challenges. To describe how including different patient, caregiver, and societal treatment-related factors affects estimates of the cost-effectiveness of a hypothetical disease-modifying AD treatment. In this economic evaluation, the Alzheimer Disease Archimedes Condition Event Simulator was used to simulate the prognosis of a hypothetical cohort of patients selected from the Alzheimer Disease Neuroimaging Initiative database who received the diagnosis of mild cognitive impairment (MCI). Scenario analyses that varied costs and quality of life inputs relevant to patients and caregivers were conducted. The analysis was designed and conducted from June 15, 2019, to September 30, 2020. A hypothetical drug that would delay progression to dementia in individuals with MCI compared with usual care. Incremental cost-effectiveness ratio (ICER), measured by cost per quality-adjusted life-year (QALY) gained. The model included a simulated cohort of patients who scored between 24 and 30 on the Mini-Mental State Examination and had a global Clinical Dementia Rating scale of 0.5, with a required memory box score of 0.5 or higher, at baseline. Using a health care sector perspective, which included only individual patient health care costs, the ICER for the hypothetical treatment was $192 000 per QALY gained. The result decreased to $183 000 per QALY gained in a traditional societal perspective analysis with the inclusion of patient non-health care costs. The inclusion of estimated caregiver health care costs produced almost no change in the ICER, but the inclusion of QALYs gained by caregivers led to a substantial reduction in the ICER for the hypothetical treatment, to $107 000 per QALY gained in the health sector perspective. In the societal perspective scenario, with the broadest inclusion of patient and caregiver factors, the ICER decreased to $74 000 per added QALY. The findings of this economic evaluation suggest that policy makers should be aware that efforts to estimate and include the effects of AD treatments outside those on patients themselves can affect the results of the cost-effectiveness analyses that often underpin assessments of the value of new treatments. Further research and debate on including these factors in assessments that will inform discussions on fair pricing for new treatments are needed.

Estimation of the mean function of functional data via deep neural networks

In this work, we propose a deep neural networks‐based method to perform non‐parametric regression for functional data. The proposed estimators are based on sparsely connected deep neural networks with rectifier linear unit (ReLU) activation function. We provide the convergence rate of the proposed deep neural networks estimator in terms of the empirical norm. Through Monte Carlo simulation studies, we examine the finite sample performance of the proposed method. Finally, the proposed method is applied to analyse positron emission tomography images of patients with Alzheimer's disease obtained from the Alzheimer Disease Neuroimaging Initiative database.

Parametric Estimation of Reference Signal Intensity for Semi-Quantification of Tau Deposition: A Flortaucipir and [18F]-APN-1607 Study.

BackgroundTau positron emission tomography (PET) imaging can reveal the pathophysiology and neurodegeneration that occurs in Alzheimer’s disease (AD) in vivo. The standardized uptake value ratio (SUVR) is widely used for semi-quantification of tau deposition but is susceptible to disturbance from the reference region and the partial volume effect (PVE). To overcome this problem, we applied the parametric estimation of reference signal intensity (PERSI) method—which was previously evaluated for flortaucipir imaging—to two tau tracers, flortaucipir and [18F]-APN-1607.MethodsTwo cohorts underwent tau PET scanning. Flortaucipir PET imaging data for cohort I (65 healthy controls [HCs], 60 patients with mild cognitive impairment [MCI], and 12 AD patients) were from the AD Neuroimaging Initiative database. [18F]-APN-1607 ([18F]-PM-PBB3) PET imaging data were for Cohort II, which included 21 patients with a clinical diagnosis of amyloid PET-positive AD and 15 HCs recruited at Huashan Hospital. We used white matter (WM) postprocessed by PERSI (PERSI-WM) as the reference region and compared this with the traditional semi-quantification method that uses the whole cerebellum as the reference. SUVRs were calculated for regions of interest including the frontal, parietal, temporal, and occipital lobes; anterior and posterior cingulate; precuneus; and Braak I/II (entorhinal cortex and hippocampus). Receiver operating characteristic (ROC) curve analysis and effect sizes were used to compare the two methods in terms of ability to discriminate between different clinical groups.ResultsIn both cohorts, regional SUVR determined using the PERSI-WM method was superior to using the cerebellum as reference region for measuring tau retention in AD patients (e.g., SUVR of the temporal lobe: flortaucipir, 1.08 ± 0.17 and [18F]-APN-1607, 1.57 ± 0.34); and estimates of the effect size and areas under the ROC curve (AUC) indicated that it also increased between-group differences (e.g., AUC of the temporal lobe for HC vs AD: flortaucipir, 0.893 and [18F]-APN-1607: 0.949).ConclusionThe PERSI-WM method significantly improves diagnostic discrimination compared to conventional approach of using the cerebellum as a reference region and can mitigate the PVE; it can thus enhance the efficacy of semi-quantification of multiple tau tracers in PET scanning, making it suitable for large-scale clinical application.

Early Detection of Alzheimer's Disease Based on Clinical Trials, Three-Dimensional Imaging Data, and Personal Information Using Autoencoders.

Background:A timely diagnosis of Alzheimer's disease (AD) is crucial to obtain more practical treatments. In this article, a novel approach using Auto-Encoder Neural Networks (AENN) for early detection of AD was proposed.Method:The proposed method mainly deals with the classification of multimodal data and the imputation of missing data. The data under study involve the MiniMental State Examination, magnetic resonance imaging, positron emission tomography, cerebrospinal fluid data, and personal information. Natural logarithm was used for normalizing the data. The Auto-Encoder Neural Networks was used for imputing missing data. Principal component analysis algorithm was used for reducing dimensionality of data. Support Vector Machine (SVM) was used as classifier. The proposed method was evaluated using Alzheimer's Disease Neuroimaging Initiative (ADNI) database. Then, 10fold crossvalidation was used to audit the detection accuracy of the method.Results:The effectiveness of the proposed approach was studied under several scenarios considering 705 cases of ADNI database. In three binary classification problems, that is AD vs. normal controls (NCs), mild cognitive impairment (MCI) vs. NC, and MCI vs. AD, we obtained the accuracies of 95.57%, 83.01%, and 78.67%, respectively.Conclusion:Experimental results revealed that the proposed method significantly outperformed most of the stateoftheart methods.