AI-based tool for early detection of Alzheimer's disease

In the context of Alzheimer's disease (AD), timely identification is paramount for effective management, acknowledging its chronic and irreversible nature, where medications can only impede its progression. Our study introduces a holistic solution, leveraging the hippocampus and the VGG16 model with transfer learning for early AD detection. The hippocampus, a pivotal early affected region linked to memory, plays a central role in classifying patients into three categories: cognitively normal (CN), representing individuals without cognitive impairment; mild cognitive impairment (MCI), indicative of a subtle decline in cognitive abilities; and AD, denoting Alzheimer's disease. Employing the Alzheimer's Disease Neuroimaging Initiative (ADNI) dataset, our model undergoes training enriched by advanced image preprocessing techniques, achieving outstanding accuracy (testing 98.17 %, validation 97.52 %, training 99.62 %). The strategic use of transfer learning fortifies our competitive edge, incorporating the hippocampus approach and, notably, a progressive data augmentation technique. This innovative augmentation strategy gradually introduces augmentation factors during training, significantly elevating accuracy and enhancing the model's generalization ability. The study emphasizes practical application with a user-friendly website, empowering radiologists to predict class probabilities, track disease progression, and visualize patient images in both 2D and 3D formats, contributing significantly to the advancement of early AD detection.

Differential atrophy along the longitudinal hippocampal axis in Alzheimer's disease for the Alzheimer's Disease Neuroimaging Initiative

Alzheimer's disease (AD) is a progressive neurodegenerative disorder that primarily affects the hippocampus. Since hippocampal studies have highlighted a differential subregional regulation along its longitudinal axis, a more detailed analysis addressing subregional changes along the longitudinal hippocampal axis has the potential to provide new relevant biomarkers. This study included structural brain MRI data of 583 participants from the Alzheimer's Disease Neuroimaging Initiative (ADNI). Cognitively normal (CN) subjects, mild cognitively impaired (MCI) subjects and AD patients were conveniently selected considering the age and sex match between clinical groups. Structural MRI acquisitions were pre-processed and analysed with a new longitudinal axis segmentation method, dividing the hippocampus in three subdivisions (anterior, intermediate, and posterior). When normalizing the volume of hippocampal sub-divisions to total hippocampus, the posterior hippocampus negatively correlates with age only in CN subjects (r = -.31). The longitudinal ratio of hippocampal atrophy (anterior sub-division divided by the posterior one) shows a significant increase with age only in CN (r = .25). Overall, in AD, the posterior hippocampus is predominantly atrophied early on. Consequently, the anterior/posterior hippocampal ratio is an AD differentiating metric at early disease stages with potential for diagnostic and prognostic applications.

Analysing heterogeneity in Alzheimer's Disease using multimodal normative modelling on ATN biomarkers

Background and Objectives

Previous approaches pursuing normative modelling for analyzing heterogeneity in Alzheimer's Disease (AD) have relied on a single neuroimaging modality. However, AD is a multi-faceted disorder, with each modality providing unique and complementary info about AD. In this study, we used a deep-learning based multimodal normative model to assess the heterogeneity in regional brain patterns for ATN (amyloid-tau-neurodegeneration) biomarkers.

Methods

We selected discovery (n = 665) and replication (n = 430) cohorts with simultaneous availability of ATN biomarkers: Florbetapir amyloid, Flortaucipir tau and T1-weighted MRI (magnetic resonance imaging) imaging. A multimodal variational autoencoder (conditioned on age and sex) was used as a normative model to learn the multimodal regional brain patterns of a cognitively unimpaired (CU) control group. The trained model was applied on individuals on the ADS (AD Spectrum) to estimate their deviations (Z-scores) from the normative distribution, resulting in a Z-score regional deviation map per ADS individual per modality. Regions with Z-scores < -1.96 for MRI and Z-scores > 1.96 for amyloid and tau were labelled as outliers. Hamming distance was used to quantify the dissimilarity between individual based on their outlier deviations across each modality. We also calculated a disease severity index (DSI) for each ADS individual which was estimated by averaging the deviations across all outlier regions corresponding to each modality.

Results

ADS individuals with moderate or severe dementia showed higher proportion of regional outliers for each modality as well as more dissimilarity in modality-specific regional outlier patterns compared to ADS individuals with early or mild dementia. DSI was associated with the progressive stages of dementia, (ii) showed significant associations with neuropsychological composite scores and (iii) related to the longitudinal risk of CDR progression. Findings were reproducible in both discovery and replication cohorts.

Discussion

Our is the first study to examine the heterogeneity in AD through the lens of multiple neuroimaging modalities (ATN), based on distinct or overlapping patterns of regional outlier deviations. Regional MRI and tau outliers were more heterogenous than regional amyloid outliers. DSI has the potential to be an individual patient metric of neurodegeneration that can help in clinical decision making and monitoring patient response for anti-amyloid treatments.

Identifying potential genetic epistasis implicated in Alzheimer's disease via detection of SNP-SNP interaction on quantitative trait CSF Aβ42

Although genome-wide association studies have identified multiple Alzheimer's disease (AD)-associated loci by selecting the main effects of individual single-nucleotide polymorphisms (SNPs), the interpretation of genetic variance in AD is limited. Based on the linear regression method, we performed genome-wide SNP-SNP interaction on cerebrospinal fluid Aβ42 to identify potential genetic epistasis implicated in AD, with age, gender, and diagnosis as covariates. A GPU-based method was used to address the computational challenges posed by the analysis of epistasis. We found 368 SNP pairs to be statistically significant, and highly significant SNP-SNP interactions were identified between the marginal main effects of SNP pairs, which explained a relatively high variance at the Aβ42 level. Our results replicated 100 previously reported AD-related genes and 5 gene-gene interaction pairs of the protein-protein interaction network. Our bioinformatics analyses provided preliminary evidence that the 5-overlapping gene-gene interaction pairs play critical roles in inducing synaptic loss and dysfunction, thereby leading to memory decline and cognitive impairment in AD-affected brains.

Alzheimer's disease diagnostic accuracy by fluid and neuroimaging ATN framework

OBJECTIVES

The ATN's different modalities (fluids and neuroimaging) for each of the Aβ (A), tau (T), and neurodegeneration (N) elements are used for the biological diagnosis of Alzheimer's disease (AD). We aim to identify which ATN category achieves the highest potential for diagnosis and predictive accuracy of longitudinal cognitive decline.

METHODS

Based on the availability of plasma ATN biomarkers (plasma-derived Aβ42/40 , p-tau181, NFL, respectively), CSF ATN biomarkers (CSF-derived Aβ42 /Aβ40 , p-tau181, NFL), and neuroimaging ATN biomarkers (18F-florbetapir (FBP) amyloid-PET, 18F-flortaucipir (FTP) tau-PET, and fluorodeoxyglucose (FDG)-PET), a total of 2340 participants were selected from ADNI.

RESULTS

Our data analysis indicates that the area under curves (AUCs) of CSF-A, neuroimaging-T, and neuroimaging-N were ranked the top three ATN candidates for accurate diagnosis of AD. Moreover, neuroimaging ATN biomarkers display the best predictive ability for longitudinal cognitive decline among the three categories. To note, neuroimaging-T correlates well with cognitive performances in a negative correlation manner. Meanwhile, participants in the "N" element positive group, especially the CSF-N positive group, experience the fastest cognitive decline compared with other groups defined by ATN biomarkers. In addition, the voxel-wise analysis showed that CSF-A related to tau accumulation and FDG-PET indexes more strongly in subjects with MCI stage. According to our analysis of the data, the best three ATN candidates for a precise diagnosis of AD are CSF-A, neuroimaging-T, and neuroimaging-N.

CONCLUSIONS

Collectively, our findings suggest that plasma, CSF, and neuroimaging biomarkers differ considerably within the ATN framework; the most accurate target biomarkers for diagnosing AD were the CSF-A, neuroimaging-T, and neuroimaging-N within each ATN modality. Moreover, neuroimaging-T and CSF-N both show excellent ability in the prediction of cognitive decline in two different dimensions.

DAT1 and BDNF polymorphisms interact to predict Aβ and tau pathology

Previous work has associated polymorphisms in the dopamine transporter gene (rs6347 in DAT1/SLC6A3) and brain derived neurotrophic factor gene (Val66Met in BDNF) with atrophy and memory decline. However, it is unclear whether these polymorphisms relate to atrophy and cognition through associations with Alzheimer's disease pathology. We tested for effects of DAT1 and BDNF polymorphisms on cross-sectional and longitudinal β-amyloid (Aβ) and tau pathology (measured with positron emission tomography (PET)), hippocampal volume, and cognition. We analyzed a sample of cognitively normal older adults (cross-sectional n = 321) from the Alzheimer's Disease Neuroimaging Initiative (ADNI). DAT1 and BDNF interacted to predict Aβ-PET, tau-PET, and hippocampal atrophy. Carriers of both "non-boptimal" DAT1 C and BDNF Met alleles demonstrated greater pathology and atrophy. Our findings provide novel links between dopamine and neurotrophic factor genes and AD pathology, consistent with previous research implicating these variants in greater risk for developing AD.

Precision Brain Morphometry Using Cluster Scanning

Measurement error limits the statistical power to detect group differences and longitudinal change in structural MRI morphometric measures (e.g., hippocampal volume, prefrontal thickness). Recent advances in scan acceleration enable extremely fast T1-weighted scans (~1 minute) to achieve morphometric errors that are close to the errors in longer traditional scans. As acceleration allows multiple scans to be acquired in rapid succession, it becomes possible to pool estimates to increase measurement precision, a strategy known as "cluster scanning." Here we explored brain morphometry using cluster scanning in a test-retest study of 40 individuals (12 younger adults, 18 cognitively unimpaired older adults, and 10 adults diagnosed with mild cognitive impairment or Alzheimer's Dementia). Morphometric errors from a single compressed sensing (CS) 1.0mm scan with 6x acceleration (CSx6) were, on average, 12% larger than a traditional scan using the Alzheimer's Disease Neuroimaging Initiative (ADNI) protocol. Pooled estimates from four clustered CSx6 acquisitions led to errors that were 34% smaller than ADNI despite having a shorter total acquisition time. Given a fixed amount of time, a gain in measurement precision can thus be achieved by acquiring multiple rapid scans instead of a single traditional scan. Errors were further reduced when estimates were pooled from eight CSx6 scans (51% smaller than ADNI). Neither pooling across a break nor pooling across multiple scan resolutions boosted this benefit. We discuss the potential of cluster scanning to improve morphometric precision, boost statistical power, and produce more sensitive disease progression biomarkers.

The effects of CNS atrophy and ICVD on tests of executive function and functional status are mediated by intelligence

Background

Impairments in executive function (EF) are often attributed to ischemic cerebrovascular disease (ICVD) and frontal circuit pathology. However, EF can be distinguished from general intelligence and the latter is likely to manifest in "executive" measures. We aimed to distinguish the effects of imaging biomarkers on these constructs.

Methods

We tested neuroimaging biomarkers as independent predictors of observed 12 month-prospective cognitive performance by a Multiple Indicators Multiple Causes (MIMIC) model in the Alzheimer's Disease Neuroimaging Initiative (ADNI) (N ≅ 1750).

Results

ICVD was associated with ''Organization" (ORG) and "Planning" (PLAN) domain scores from the test of Every Day Cognition. Left anterior cingulate (LAC) atrophy was independently associated with Trail-Making part B and Animal Naming. The MIMIC model had excellent fit and tests additional latent variables i.e., EF and dEF (a latent δ homolog derived from Spearman's general intelligence factor, g). Only dEF was associated with instrumental activities of daily living (IADL). ICVD and LAC were both associated with observed executive measures through dEF. ICVD was independently associated with those same measures through EF.

Conclusions

Observed EF is independently determined by multiple factors. The effects of EF-associated MRI biomarkers can be related to disability and dementia only via their effects on g. Because g /δ are unlikely to be located within the frontal lobes, the dementia-specific variance in executive measures may have little to do with either frontal structure or function. Conversely, domain-specific variance in EF may have little to do with either IADL-impairment or dementia.

Fine-grained age-matching improves atrophy-based detection of mild cognitive impairment more than amyloid-negative reference subjects

INTRODUCTION

In clinical practice, differentiating between age-related gray matter (GM) atrophy and neurodegeneration-related atrophy at early disease stages, such as mild cognitive impairment (MCI), remains challenging. We hypothesized that fined-grained adjustment for age effects and using amyloid-negative reference subjects could increase classification accuracy.

METHODS

T1-weighted magnetic resonance imaging (MRI) data of 131 cognitively normal (CN) individuals and 91 patients with MCI from the Alzheimer's disease neuroimaging initiative (ADNI) characterized concerning amyloid status, as well as 19 CN individuals and 19 MCI patients from an independent validation sample were segmented, spatially normalized and analyzed in the framework of voxel-based morphometry (VBM). For each participant, statistical maps of GM atrophy were computed as the deviation from the GM of CN reference groups at the voxel level. CN reference groups composed with different degrees of age-matching, and mixed and strictly amyloid-negative CN reference groups were examined regarding their effect on the accuracy in distinguishing between CN and MCI. Furthermore, the effects of spatial smoothing and atrophy threshold were assessed.

RESULTS

Approaches with a specific reference group for each age significantly outperformed all other age-adjustment strategies with a maximum area under the curve of 1.0 in the ADNI sample and 0.985 in the validation sample. Accounting for age in a regression-based approach improved classification accuracy over that of a single CN reference group in the age range of the patient sample. Using strictly amyloid-negative reference groups improved classification accuracy only when age was not considered.

CONCLUSION

Our results demonstrate that VBM can differentiate between age-related and MCI-associated atrophy with high accuracy. Crucially, age-specific reference groups significantly increased accuracy, more so than regression-based approaches and using amyloid-negative reference groups.

Multinomial Logistic Factor Regression for Multi-source Functional Block-wise Missing Data

Multi-source functional block-wise missing data arise more commonly in medical care recently with the rapid development of big data and medical technology, hence there is an urgent need to develop efficient dimension reduction to extract important information for classification under such data. However, most existing methods for classification problems consider high-dimensional data as covariates. In the paper, we propose a novel multinomial imputed-factor Logistic regression model with multi-source functional block-wise missing data as covariates. Our main contribution is to establishing two multinomial factor regression models by using the imputed multi-source functional principal component scores and imputed canonical scores as covariates, respectively, where the missing factors are imputed by both the conditional mean imputation and the multiple block-wise imputation approaches. Specifically, the univariate FPCA is carried out for the observable data of each data source firstly to obtain the univariate principal component scores and the eigenfunctions. Then, the block-wise missing univariate principal component scores instead of the block-wise missing functional data are imputed by the conditional mean imputation method and the multiple block-wise imputation method, respectively. After that, based on the imputed univariate factors, the multi-source principal component scores are constructed by using the relationship between the multi-source principal component scores and the univariate principal component scores; and at the same time, the canonical scores are obtained by the multiple-set canonial correlation analysis. Finally, the multinomial imputed-factor Logistic regression model is established with the multi-source principal component scores or the canonical scores as factors. Numerical simulations and real data analysis on ADNI data show the proposed method works well.

Brain morphometry in older adults with and without dementia using extremely rapid structural scans

T1-weighted structural MRI is widely used to measure brain morphometry (e.g., cortical thickness and subcortical volumes). Accelerated scans as fast as one minute or less are now available but it is unclear if they are adequate for quantitative morphometry. Here we compared the measurement properties of a widely adopted 1.0 mm resolution scan from the Alzheimer's Disease Neuroimaging Initiative (ADNI = 5'12'') with two variants of highly accelerated 1.0 mm scans (compressed-sensing, CSx6 = 1'12''; and wave-controlled aliasing in parallel imaging, WAVEx9 = 1'09'') in a test-retest study of 37 older adults aged 54 to 86 (including 19 individuals diagnosed with a neurodegenerative dementia). Rapid scans produced highly reliable morphometric measures that largely matched the quality of morphometrics derived from the ADNI scan. Regions of lower reliability and relative divergence between ADNI and rapid scan alternatives tended to occur in midline regions and regions with susceptibility-induced artifacts. Critically, the rapid scans yielded morphometric measures similar to the ADNI scan in regions of high atrophy. The results converge to suggest that, for many current uses, extremely rapid scans can replace longer scans. As a final test, we explored the possibility of a 0'49'' 1.2 mm CSx6 structural scan, which also showed promise. Rapid structural scans may benefit MRI studies by shortening the scan session and reducing cost, minimizing opportunity for movement, creating room for additional scan sequences, and allowing for the repetition of structural scans to increase precision of the estimates.

ComBat Harmonization: Empirical Bayes versus fully Bayes approaches

Studying small effects or subtle neuroanatomical variation requires large-scale sample size data. As a result, combining neuroimaging data from multiple datasets is necessary. Variation in acquisition protocols, magnetic field strength, scanner build, and many other non-biologically related factors can introduce undesirable bias into studies. Hence, harmonization is required to remove the bias-inducing factors from the data. ComBat is one of the most common methods applied to features from structural images. ComBat models the data using a hierarchical Bayesian model and uses the empirical Bayes approach to infer the distribution of the unknown factors. The empirical Bayes harmonization method is computationally efficient and provides valid point estimates. However, it tends to underestimate uncertainty. This paper investigates a new approach, fully Bayesian ComBat, where Monte Carlo sampling is used for statistical inference. When comparing fully Bayesian and empirical Bayesian ComBat, we found Empirical Bayesian ComBat more effectively removed scanner strength information and was much more computationally efficient. Conversely, fully Bayesian ComBat better preserved biological disease and age-related information while performing more accurate harmonization on traveling subjects. The fully Bayesian approach generates a rich posterior distribution, which is useful for generating simulated imaging features for improving classifier performance in a limited data setting. We show the generative capacity of our model for augmenting and improving the detection of patients with Alzheimer's disease. Posterior distributions for harmonized imaging measures can also be used for brain-wide uncertainty comparison and more principled downstream statistical analysis.Code for our new fully Bayesian ComBat extension is available at https://github.com/batmanlab/BayesComBat.

The effect of plasma cortisol on hippocampal atrophy and clinical progression in mild cognitive impairment

Introduction

Both elevated cortisol and hippocampal volume have been linked to an increased risk for the development of Alzheimer's disease (AD). This longitudinal study assessed the effects of plasma cortisol on hippocampal atrophy and clinical progression rates in patients with mild cognitive impairment (MCI).

Methods

Patients with amnestic MCI (n = 304) were selected from the Alzheimer's Disease Neuroimaging Initiative (ADNI) based on availability of baseline plasma cortisol and hippocampal volume measures, assessed at baseline and during follow-ups. We investigated associations between plasma cortisol, hippocampal volume, and risk of clinical progression to AD over a study period of up to 100 months (mean follow-up time 36.8 months) using linear mixed models, Cox proportional hazards models, and Kaplan-Meier estimators.

Results

Plasma cortisol predicted greater hippocampal atrophy, such that participants with higher cortisol showed faster decline in hippocampal volume over time (interaction: β = -0.15, p = 0.004). Small hippocampal volume predicted a higher risk of clinical progression to AD (haard ratio [HR] = 2.15; confidence in terval [CI], 1.64-2.80; p < 0.001). A similar effect was not found for cortisol (HR = 1.206; CI, 0.82-1.37; p = 0.670) and there was no statistical evidence for an interaction between hippocampal volume and cortisol on clinical progression (HR = 0.81; CI, 0.57-0.17; p = 0.260).

Discussion

Our findings suggest that higher cortisol predicts higher hippocampal atrophy, which in turn is a risk factor for progression to AD. Regulation of the hypothalamic-pituitary-adrenal axis through stress-reducing lifestyle interventions might be a protective factor against hippocampal degeneration at the prodromal stage of AD.

Genome-Wide Epistasis Study of Cerebrospinal Fluid Hyperphosphorylated Tau in ADNI Cohort

Alzheimer's disease (AD) is the main cause of dementia worldwide, and the genetic mechanism of which is not yet fully understood. Much evidence has accumulated over the past decade to suggest that after the first large-scale genome-wide association studies (GWAS) were conducted, the problem of "missing heritability" in AD is still a great challenge. Epistasis has been considered as one of the main causes of "missing heritability" in AD, which has been largely ignored in human genetics. The focus of current genome-wide epistasis studies is usually on single nucleotide polymorphisms (SNPs) that have significant individual effects, and the amount of heritability explained by which was very low. Moreover, AD is characterized by progressive cognitive decline and neuronal damage, and some studies have suggested that hyperphosphorylated tau (P-tau) mediates neuronal death by inducing necroptosis and inflammation in AD. Therefore, this study focused on identifying epistasis between two-marker interactions at marginal main effects across the whole genome using cerebrospinal fluid (CSF) P-tau as quantitative trait (QT). We sought to detect interactions between SNPs in a multi-GPU based linear regression method by using age, gender, and clinical diagnostic status (cds) as covariates. We then used the STRING online tool to perform the PPI network and identify two-marker epistasis at the level of gene-gene interaction. A total of 758 SNP pairs were found to be statistically significant. Particularly, between the marginal main effect SNP pairs, highly significant SNP-SNP interactions were identified, which explained a relatively high variance at the P-tau level. In addition, 331 AD-related genes were identified, 10 gene-gene interaction pairs were replicated in the PPI network. The identified gene-gene interactions and genes showed associations with AD in terms of neuroinflammation and neurodegeneration, neuronal cells activation and brain development, thereby leading to cognitive decline in AD, which is indirectly associated with the P-tau pathological feature of AD and in turn supports the results of this study. Thus, the results of our study might be beneficial for explaining part of the "missing heritability" of AD.

Association of rs2062323 in the TREM1 gene with Alzheimer's disease and cerebrospinal fluid-soluble TREM2

INTRODUCTION AND AIMS

Genetic variations play a significant role in determining an individual's AD susceptibility. Research on the connection between AD and TREM1 gene polymorphisms (SNPs) remained lacking. We sought to examine the associations between TREM1 SNPs and AD.

METHODS

Based on the 1000 Genomes Project data, linkage disequilibrium (LD) analyses were utilized to screen for candidate SNPs in the TREM1 gene. AD cases (1081) and healthy control subjects (870) were collected and genotyped, and the associations between candidate SNPs and AD risk were analyzed. We explored the associations between target SNP and AD biomarkers. Moreover, 842 individuals from ADNI were selected to verify these results. Linear mixed models were used to estimate associations between the target SNP and longitudinal cognitive changes.

RESULTS

The rs2062323 was identified to be associated with AD risk in the Han population, and rs2062323T carriers had a lower AD risk (co-dominant model: OR, 0.67, 95% CI, 0.51-0.88, p = 0.0037; additive model: OR, 0.82, 95% CI, 0.72-0.94, p = 0.0032). Cerebrospinal fluid (CSF) sTREM2 levels were significantly increased in middle-aged rs2062323T carriers (additive model: β = 0.18, p = 0.0348). We also found significantly elevated levels of CSF sTREM2 in the ADNI. The rate of cognitive decline slowed down in rs2062323T carriers.

CONCLUSIONS

This study is the first to identify significant associations between TREM1 rs2062323 and AD risk. The rs2062323T may be involved in AD by regulating the expression of TREM1, TREML1, TREM2, and sTREM2. The TREM family is expected to be a potential therapeutic target for AD.

Brain structure and allelic associations in Alzheimer's disease

BACKGROUND

Alzheimer's disease (AD), the most prevalent form of dementia, affects 6.5 million Americans and over 50 million people globally. Clinical, genetic, and phenotypic studies of dementia provide some insights of the observed progressive neurodegenerative processes, however, the mechanisms underlying AD onset remain enigmatic.

AIMS

This paper examines late-onset dementia-related cognitive impairment utilizing neuroimaging-genetics biomarker associations.

MATERIALS AND METHODS

The participants, ages 65-85, included 266 healthy controls (HC), 572 volunteers with mild cognitive impairment (MCI), and 188 Alzheimer's disease (AD) patients. Genotype dosage data for AD-associated single nucleotide polymorphisms (SNPs) were extracted from the imputed ADNI genetics archive using sample-major additive coding. Such 29 SNPs were selected, representing a subset of independent SNPs reported to be highly associated with AD in a recent AD meta-GWAS study by Jansen and colleagues.

RESULTS

We identified the significant correlations between the 29 genomic markers (GMs) and the 200 neuroimaging markers (NIMs). The odds ratios and relative risks for AD and MCI (relative to HC) were predicted using multinomial linear models.

DISCUSSION

In the HC and MCI cohorts, mainly cortical thickness measures were associated with GMs, whereas the AD cohort exhibited different GM-NIM relations. Network patterns within the HC and AD groups were distinct in cortical thickness, volume, and proportion of White to Gray Matter (pct), but not in the MCI cohort. Multinomial linear models of clinical diagnosis showed precisely the specific NIMs and GMs that were most impactful in discriminating between AD and HC, and between MCI and HC.

CONCLUSION

This study suggests that advanced analytics provide mechanisms for exploring the interrelations between morphometric indicators and GMs. The findings may facilitate further clinical investigations of phenotypic associations that support deep systematic understanding of AD pathogenesis.

Individualized network analysis: A novel approach to investigate tau PET using graph theory in the Alzheimer's disease continuum

Introduction

Tau PET imaging has emerged as an important tool to detect and monitor tangle burden in vivo in the study of Alzheimer's disease (AD). Previous studies demonstrated the association of tau burden with cognitive decline in probable AD cohorts. This study introduces a novel approach to analyze tau PET data by constructing individualized tau network structure and deriving its graph theory-based measures. We hypothesize that the network- based measures are a measure of the total tau load and the stage through disease.

Methods

Using tau PET data from the AD Neuroimaging Initiative from 369 participants, we determine the network measures, global efficiency, global strength, and limbic strength, and compare with two regional measures entorhinal and tau composite SUVR, in the ability to differentiate, cognitively unimpaired (CU), MCI and AD. We also investigate the correlation of these network and regional measures and a measure of memory performance, auditory verbal learning test for long-term recall memory (AVLT-LTM). Finally, we determine the stages based on global efficiency and limbic strength using conditional inference trees and compare with Braak staging.

Results

We demonstrate that the derived network measures are able to differentiate three clinical stages of AD, CU, MCI, and AD. We also demonstrate that these network measures are strongly correlated with memory performance overall. Unlike regional tau measurements, the tau network measures were significantly associated with AVLT-LTM even in cognitively unimpaired individuals. Stages determined from global efficiency and limbic strength, visually resembled Braak staging.

Discussion

The strong correlations with memory particularly in CU suggest the proposed technique may be used to characterize subtle early tau accumulation. Further investigation is ongoing to examine this technique in a longitudinal setting.

Association of rs2072446 in the NGFR gene with the risk of Alzheimer's disease and amyloid-β deposition in the brain

INTRODUCTION AND AIMS

Alzheimer's disease (AD) is the most common form of dementia with a complex genetic background. The cause of sporadic AD (sAD) remains largely unknown. Increasing evidence shows that genetic variations play a crucial role in sAD. P75 neurotrophin receptor (p75NTR, encoded by NGFR) plays a critical role in the pathogenesis of AD. Yet, the relationship between NGFR gene polymorphisms and AD was less studied. This study aims to analyze the relationship of NGFR gene polymorphism with the risk of AD in the Chinese Han population and amyloid-β deposition in the ADNI cohort.

METHODS

This case-control association study was conducted in a Chinese Han cohort consisting of 366 sporadic AD (sAD) patients and 390 age- and sex-matched controls. Twelve tag-SNPs were selected and genotyped with a multiplex polymerase chain reaction-ligase detection reaction (PCR-LDR) method. The associations between tag-SNPs and the risk of AD were analyzed by logistic regression. Moreover, another cohort from the Alzheimer's Disease Neuroimaging Initiative (ADNI) database was included to examine the association of one tag-SNP (rs2072446) with indicators of amyloid deposition. Kaplan-Meier survival analysis and Cox proportional hazards models were used to test the predictive abilities of rs2072446 genotypes for AD progression. The mediation effects of Aβ deposition on this association were subsequently tested by mediation analyses.

RESULTS

After multiple testing corrections, one tag-SNP, rs2072446, was associated with an increased risk of sAD (additive model, OR = 1.79, P_adjustment  = 0.0144). Analyses of the ADNI cohort showed that the minor allele (T) of rs2072446 was significantly associated with the heavier Aβ burden, which further contributed to an increased risk of AD progression in APOE ε4 non-carrier.

CONCLUSION

Our study found that rs2072446 in NGFR is associated with both the risk of sAD in the Chinese Han population and the amyloid burden in the ADNI cohort, which reveals the role of p75NTR in AD from a genetic perspective.

Predicting conversion from mild cognitive impairment to Alzheimer's disease with multimodal latent factors

INTRODUCTION

We studied the ability of latent factor scores to predict conversion from mild cognitive impairment (MCI) to Alzheimer's disease (AD) and investigated whether multimodal factor scores improve predictive power, relative to single-modal factor scores.

METHOD

We conducted exploratory factor analyses (EFAs) and confirmatory factor analyses (CFAs) of the baseline data of MCI subjects in the Alzheimer's Disease Neuroimaging Initiative (ADNI) to generate factor scores for three data modalities: neuropsychological (NP), magnetic resonance imaging (MRI), and cerebrospinal fluid (CSF). Factor scores from single or multiple modalities were entered in logistic regression models to predict MCI to AD conversion for 160 ADNI subjects over a 2-year interval.

RESULTS

NP factors attained an area under the curve (AUC) of .80, with a sensitivity of .66 and a specificity of .77. MRI factors reached a comparable level of performance (AUC = .80, sensitivity = .66, specificity = .78), whereas CSF factors produced weaker prediction (AUC = .70, sensitivity = .56, specificity = .79). Combining NP factors with MRI or CSF factors produced better prediction than either MRI or CSF factors alone. Similarly, adding MRI factors to NP or CSF factors produced improvements in prediction relative to NP or CSF factors alone. However, adding CSF factors to either NP or MRI factors produced no improvement in prediction.

CONCLUSIONS

Latent factor scores provided good accuracy for predicting MCI to AD conversion. Adding NP or MRI factors to factors from other modalities enhanced predictive power but adding CSF factors did not.

Transfer learning for cognitive reserve quantification

Cognitive reserve (CR) has been introduced to explain individual differences in susceptibility to cognitive or functional impairment in the presence of age or pathology. We developed a deep learning model to quantify the CR as residual variance in memory performance using the Structural Magnetic Resonance Imaging (sMRI) data from a lifespan healthy cohort. The generalizability of the sMRI-based deep learning model was tested in two independent healthy and Alzheimer's cohorts using transfer learning framework. Structural MRIs were collected from three cohorts: 495 healthy adults (age: 20-80) from RANN, 620 healthy adults (age: 36-100) from lifespan Human Connectome Project Aging (HCPA), and 941 adults (age: 55-92) from Alzheimer's Disease Neuroimaging Initiative (ADNI). Region of interest (ROI)-specific cortical thickness and volume measures were extracted using the Desikan-Killiany Atlas. CR was quantified by residuals which subtract the predicted memory from the true memory. Cascade neural network (CNN) models were used to train RANN dataset for memory prediction. Transfer learning was applied to transfer the T1 imaging-based model from source domain (RANN) to the target domains (HCPA or ADNI). The CNN model trained on the RANN dataset exhibited strong linear correlation between true and predicted memory based on the T1 cortical thickness and volume predictors. In addition, the model generated from healthy lifespan data (RANN) was able to generalize to an independent healthy lifespan data (HCPA) and older demented participants (ADNI) across different scanner types. The estimated CR was correlated with CR proxies such education and IQ across all three datasets. The current findings suggest that the transfer learning approach is an effective way to generalize the residual-based CR estimation. It is applicable to various diseases and may flexibly incorporate different imaging modalities such as fMRI and PET, making it a promising tool for scientific and clinical purposes.

Comparison of plasma and CSF biomarkers across ethnoracial groups in the ADNI

Introduction

Ethnoracial differences in cerebrospinal fluid (CSF; amyloid beta 42 [Aβ42], total tau [t-tau], phosphorylated tau 181 [p-tau181], and plasma (p-tau181, neurofilament light [NfL]) biomarkers of Alzheimer's disease (AD) are incompletely understood.

Methods

We performed cross-sectional analyses with and without adjustment for covariates comparing baseline CSF (Aβ42, t-tau, p-tau181) and plasma (p-tau181, NfL) values in 47 African Americans (AAs) matched to 141 non-Hispanic Whites (NHWs) and 43 Latinos (LAs) matched to 129 NHWs from the Alzheimer's Disease Neuroimaging Initiative (ADNI).

Results

Unadjusted comparisons revealed no significant differences in plasma or CSF biomarkers between AAs and NHWs. A trend toward a lower CSF t-tau and p-tau181 in LAs compared to NHWs was observed, without significant differences in plasma biomarkers. After adjusting for covariates, there were no significant differences in CSF or plasma biomarkers between AAs and NHWs or between LAs and NHWs.

Discussion

Plasma and CSF AD biomarkers may perform similarly across diverse populations but future studies in large, diverse cohorts are needed.

Applications of Bayesian shrinkage prior models in clinical research with categorical responses

BACKGROUND

Prediction and classification algorithms are commonly used in clinical research for identifying patients susceptible to clinical conditions such as diabetes, colon cancer, and Alzheimer's disease. Developing accurate prediction and classification methods benefits personalized medicine. Building an excellent predictive model involves selecting the features that are most significantly associated with the outcome. These features can include several biological and demographic characteristics, such as genomic biomarkers and health history. Such variable selection becomes challenging when the number of potential predictors is large. Bayesian shrinkage models have emerged as popular and flexible methods of variable selection in regression settings. This work discusses variable selection with three shrinkage priors and illustrates its application to clinical data such as Pima Indians Diabetes, Colon cancer, ADNI, and OASIS Alzheimer's real-world data.

METHODS

A unified Bayesian hierarchical framework that implements and compares shrinkage priors in binary and multinomial logistic regression models is presented. The key feature is the representation of the likelihood by a Polya-Gamma data augmentation, which admits a natural integration with a family of shrinkage priors, specifically focusing on Horseshoe, Dirichlet Laplace, and Double Pareto priors. Extensive simulation studies are conducted to assess the performances under different data dimensions and parameter settings. Measures of accuracy, AUC, brier score, L1 error, cross-entropy, and ROC surface plots are used as evaluation criteria comparing the priors with frequentist methods as Lasso, Elastic-Net, and Ridge regression.

RESULTS

All three priors can be used for robust prediction on significant metrics, irrespective of their categorical response model choices. Simulation studies could achieve the mean prediction accuracy of 91.6% (95% CI: 88.5, 94.7) and 76.5% (95% CI: 69.3, 83.8) for logistic regression and multinomial logistic models, respectively. The model can identify significant variables for disease risk prediction and is computationally efficient.

CONCLUSIONS

The models are robust enough to conduct both variable selection and prediction because of their high shrinkage properties and applicability to a broad range of classification problems.

Assessing and validating reliable change across ADNI protocols

OBJECTIVE

Reliable change methods can aid in determining whether changes in cognitive performance over time are meaningful. The current study sought to develop and cross-validate 12-month standardized regression-based (SRB) equations for the neuropsychological measures commonly administered in the Alzheimer's Disease Neuroimaging Initiative (ADNI) longitudinal study.

METHOD

Prediction algorithms were developed using baseline score, retest interval, the presence/absence of a 6-month evaluation, age, education, sex, and ethnicity in two different samples (n = 192 each) of robustly cognitively intact community-dwelling older adults from ADNI - matched for demographic and testing factors. The developed formulae for each sample were then applied to one of the samples to determine goodness-of-fit and appropriateness of combining samples for a single set of SRB equations.

RESULTS

Minimal differences were seen between Observed 12-month and Predicted 12-month scores on most neuropsychological tests from ADNI, and when compared across samples the resultant Predicted 12-month scores were highly correlated. As a result, samples were combined and SRB prediction equations were successfully developed for each of the measures.

CONCLUSIONS

Establishing cross-validation for these SRB prediction equations provides initial support of their use to detect meaningful change in the ADNI sample, and provides the basis for future research with clinical samples to evaluate potential clinical utility. While some caution should be considered for measuring true cognitive change over time - particularly in clinical samples - when using these prediction equations given the relatively lower coefficients of stability observed, use of these SRBs reflects an improvement over current practice in ADNI.

Examining the role of repeated test exposure over 12 months across ADNI protocols

Objective: Changes to study protocols during longitudinal research may alter cognitive testing schedules over time. Unlike in prior Alzheimer's Disease Neuroimaging Initiative (ADNI) protocols, where testing occurred twice annually, participants enrolled in the ADNI-3 are no longer exposed to cognitive materials at 6 months. This may affect their 12-month performance relative to earlier ADNI cohorts, and potentially confounds data harmonization attempts between earlier and later ADNI protocols. Method: Using data from participants enrolled across multiple ADNI protocols, this study investigated whether test exposure during 6-month cognitive evaluation influenced scores on subsequent 12-month evaluation. Results: No interaction effects were observed between test exposure group and time at 12 months on cognitive performance. No improvements, and limited declines, were seen between baseline and 12-month follow-up scores on most measures. Conclusions: The 6-month testing session had minimal impact on 12-month performance in ADNI. Collapsing longitudinal data across ADNI protocols in future research appears appropriate.

Caspase-1 variant influencing CSF tau and FDG PET levels in non-demented elders from the ADNI cohort

BACKGROUND

Genetic variations in the inflammatory Caspase-1 gene have been shown associated with cognitive function in elderly individuals and in predisposition to Alzheimer's disease (AD), but its detailed mechanism before the typical AD onset was still unclear. Our current study evaluated the impact of Caspase-1 common variant rs554344 on the pathological processes of brain amyloidosis, tauopathy, and neurodegeneration.

METHODS

Data used in our study were obtained from the Alzheimer's Disease Neuroimaging Initiative (ADNI) cohort. We examined the relationship between Caspase-1 rs554344 allele carrier status with AD-related cerebrospinal fluid (CSF), PET, and MRI measures at baseline by using a multiple linear regression model. We also analyzed the longitudinal effects of this variant on the change rates of CSF biomarkers and imaging data using a mixed effect model.

RESULTS

We found that Caspase-1 variant was significantly associated with FDG PET levels and CSF t-tau levels at baseline in total non-demented elderly group, and especially in mild cognitive impairment (MCI) subgroup. In addition, this variant was also detected associated with CSF p-tau levels in MCI subgroup. The mediation analysis showed that CSF p-tau partially mediated the association between Caspase-1 variant and CSF t-tau levels, accounting for 80% of the total effect.

CONCLUSIONS

Our study indicated a potential role of Caspase-1 variant in influencing cognitive function might through changing tau related-neurodegeneration process.

The accuracy and robustness of plasma biomarker models for amyloid PET positivity

Background

Plasma biomarkers for Alzheimer’s disease (AD) have broad potential as screening tools in primary care and disease-modifying trials. Detecting elevated amyloid-β (Aβ) pathology to support trial recruitment or initiating Aβ-targeting treatments would be of critical value. In this study, we aimed to examine the robustness of plasma biomarkers to detect elevated Aβ pathology at different stages of the AD continuum. Beyond determining the best biomarker—or biomarker combination—for detecting this outcome, we also simulated increases in inter-assay coefficient of variability (CV) to account for external factors not considered by intra-assay variability. With this, we aimed to determine whether plasma biomarkers would maintain their accuracy if applied in a setting which anticipates higher variability (i.e., clinical routine).

Methods

We included 118 participants (cognitively unimpaired [CU, n = 50], cognitively impaired [CI, n = 68]) from the ADNI study with a full plasma biomarker profile (Aβ42/40, GFAP, p-tau181, NfL) and matched amyloid imaging. Initially, we investigated how simulated CV variations impacted single-biomarker discriminative performance of amyloid status. Then, we evaluated the predictive performance of models containing different biomarker combinations, based both on original and simulated measurements. Plasma Aβ42/40 was represented by both immunoprecipitation mass spectrometry (IP-MS) and single molecule array (Simoa) methods in separate analyses. Model selection was based on a decision tree which incorporated Akaike information criterion value, likelihood ratio tests between the best-fitting models and, finally, and Schwartz’s Bayesian information criterion.

Results

Increasing variation greatly impacted the performance of plasma Aβ42/40 in discriminating Aβ status. In contrast, the performance of plasma GFAP and p-tau181 remained stable with variations >20%. When biomarker models were compared, the models “AG” (Aβ42/40 + GFAP; AUC = 86.5), “A” (Aβ42/40; AUC = 82.3), and “AGP” (Aβ42/40 + GFAP + p-tau181; AUC = 93.5) were superior in determining Aβ burden in all participants, within-CU, and within-CI groups, respectively. In the robustness analyses, when repeating model selection based on simulated measurements, models including IP-MS Aβ42/40 were also most often selected. Simoa Aβ42/40 did not contribute to any selected model when used as an immunoanalytical alternative to IP-MS Aβ42/40.

Conclusions

Plasma Aβ42/40, as quantified by IP-MS, shows high performance in determining Aβ positivity at all stages of the AD continuum, with GFAP and p-tau181 further contributing at CI stage. However, between-assay variations greatly impacted the performance of Aβ42/40 but not that of GFAP and p-tau181. Therefore, when dealing with between-assay CVs that exceed 5%, plasma GFAP and p-tau181 should be considered for a more robust determination of Aβ burden in CU and CI participants, respectively.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13195-021-00942-0.

Toward a Multimodal Computer-Aided Diagnostic Tool for Alzheimer's Disease Conversion

Alzheimer’s disease (AD) is a progressive neurodegenerative disorder. It is one of the leading sources of morbidity and mortality in the aging population AD cardinal symptoms include memory and executive function impairment that profoundly alters a patient’s ability to perform activities of daily living. People with mild cognitive impairment (MCI) exhibit many of the early clinical symptoms of patients with AD and have a high chance of converting to AD in their lifetime. Diagnostic criteria rely on clinical assessment and brain magnetic resonance imaging (MRI). Many groups are working to help automate this process to improve the clinical workflow. Current computational approaches are focused on predicting whether or not a subject with MCI will convert to AD in the future. To our knowledge, limited attention has been given to the development of automated computer-assisted diagnosis (CAD) systems able to provide an AD conversion diagnosis in MCI patient cohorts followed longitudinally. This is important as these CAD systems could be used by primary care providers to monitor patients with MCI. The method outlined in this paper addresses this gap and presents a computationally efficient pre-processing and prediction pipeline, and is designed for recognizing patterns associated with AD conversion. We propose a new approach that leverages longitudinal data that can be easily acquired in a clinical setting (e.g., T1-weighted magnetic resonance images, cognitive tests, and demographic information) to identify the AD conversion point in MCI subjects with AUC = 84.7. In contrast, cognitive tests and demographics alone achieved AUC = 80.6, a statistically significant difference (n = 669, p < 0.05). We designed a convolutional neural network that is computationally efficient and requires only linear registration between imaging time points. The model architecture combines Attention and Inception architectures while utilizing both cross-sectional and longitudinal imaging and clinical information. Additionally, the top brain regions and clinical features that drove the model’s decision were investigated. These included the thalamus, caudate, planum temporale, and the Rey Auditory Verbal Learning Test. We believe our method could be easily translated into the healthcare setting as an objective AD diagnostic tool for patients with MCI.

Pairwise Correlation Analysis of the Alzheimer's Disease Neuroimaging Initiative (ADNI) Dataset Reveals Significant Feature Correlation

The Alzheimer's Disease Neuroimaging Initiative (ADNI) contains extensive patient measurements (e.g., magnetic resonance imaging [MRI], biometrics, RNA expression, etc.) from Alzheimer's disease (AD) cases and controls that have recently been used by machine learning algorithms to evaluate AD onset and progression. While using a variety of biomarkers is essential to AD research, highly correlated input features can significantly decrease machine learning model generalizability and performance. Additionally, redundant features unnecessarily increase computational time and resources necessary to train predictive models. Therefore, we used 49,288 biomarkers and 793,600 extracted MRI features to assess feature correlation within the ADNI dataset to determine the extent to which this issue might impact large scale analyses using these data. We found that 93.457% of biomarkers, 92.549% of the gene expression values, and 100% of MRI features were strongly correlated with at least one other feature in ADNI based on our Bonferroni corrected α (p-value ≤ 1.40754 × 10-13). We provide a comprehensive mapping of all ADNI biomarkers to highly correlated features within the dataset. Additionally, we show that significant correlation within the ADNI dataset should be resolved before performing bulk data analyses, and we provide recommendations to address these issues. We anticipate that these recommendations and resources will help guide researchers utilizing the ADNI dataset to increase model performance and reduce the cost and complexity of their analyses.

Machine learning of brain structural biomarkers for Alzheimer's disease (AD) diagnosis, prediction of disease progression, and amyloid beta deposition in the Japanese population

INTRODUCTION

We developed machine learning (ML) designed to analyze structural brain magnetic resonance imaging (MRI), and trained it on the Alzheimer's Disease Neuroimaging Initiative (ADNI) database. In this study, we verified its utility in the Japanese population.

METHODS

A total of 535 participants were enrolled from the Japanese ADNI database, including 148 AD, 152 normal, and 235 mild cognitive impairment (MCI). Probability of AD was expressed as AD likelihood scores (ADLS).

RESULTS

The accuracy of AD diagnosis was 88.0% to 91.2%. The accuracy of predicting the disease progression in non-dementia participants over a 3-year observation was 76.0% to 79.3%. More than 90% of the participants with low ADLS did not progress to AD within 3 years. In the amyloid positron emission tomography (PET)-positive MCI, the hazard ratio of progression was 2.39 with low ADLS, and 5.77 with high ADLS. When high ADLS was defined as N+ and Pittsburgh compound B (PiB) PET positivity was defined as A+, the time to disease progression for 50% of MCI participants was 23.7 months in A+N+, whereas it was 52.3 months in A+N-.

CONCLUSION

These results support the feasibility of our ML for the diagnosis of AD and prediction of the disease progression.

Glucose hypometabolism in the Auditory Pathway in Age Related Hearing Loss in the ADNI cohort

PURPOSE

Hearing loss (HL) is one of the most common age-related diseases. Here, we investigate the central auditory correlates of HL in people with normal cognition and mild cognitive impairment (MCI) and test their association with genetic markers with the aim of revealing pathogenic mechanisms.

METHODS

Brain glucose metabolism based on FDG-PET, self-reported HL status, and genetic data were obtained from the Alzheimer's Disease Neuroimaging Initiative (ADNI) cohort. FDG-PET data was analysed from 742 control subjects (non-HL with normal cognition or MCI) and 162 cases (HL with normal cognition or MCI) with age ranges of 72.2 ± 7.1 and 77.4 ± 6.4, respectively. Voxel-wise statistics of FDG uptake differences between cases and controls were computed using the generalised linear model in SPM12. An additional 1515 FDG-PET scans of 618 participants were analysed using linear mixed effect models to assess longitudinal HL effects. Furthermore, a quantitative trait genome-wide association study (GWAS) was conducted on the glucose uptake within regions of interest (ROIs), which were defined by the voxel-wise comparison, using genotyping data with 5,082,878 variants available for HL cases and HL controls (N = 817).

RESULTS

The HL group exhibited hypometabolism in the bilateral Heschl's gyrus (kleft = 323; kright = 151; Tleft = 4.55; Tright = 4.14; peak Puncorr < 0.001), the inferior colliculus (k = 219;T = 3.53; peak Puncorr < 0.001) and cochlear nucleus (k = 18;T = 3.55; peak Puncorr < 0.001) after age correction and using a cluster forming height threshold P < 0.005 (FWE-uncorrected). Moreover, in an age-matched subset, the cluster comprising the left Heschl's gyrus survived the FWE-correction (kleft = 1903; Tleft = 4.39; cluster PFWE-corr = 0.001). The quantitative trait GWAS identified no genome-wide significant locus in the three HL ROIs. However, various loci were associated at the suggestive threshold (p < 1e-05).

CONCLUSION

Compared to the non-HL group, glucose metabolism in the HL group was lower in the auditory cortex, the inferior colliculus, and the cochlear nucleus although the effect sizes were small. The GWAS identified candidate genes that might influence FDG uptake in these regions. However, the specific biological pathway(s) underlying the role of these genes in FDG-hypometabolism in the auditory pathway requires further investigation.

Data analysis with Shapley values for automatic subject selection in Alzheimer's disease data sets using interpretable machine learning

BACKGROUND

For the recruitment and monitoring of subjects for therapy studies, it is important to predict whether mild cognitive impaired (MCI) subjects will prospectively develop Alzheimer's disease (AD). Machine learning (ML) is suitable to improve early AD prediction. The etiology of AD is heterogeneous, which leads to high variability in disease patterns. Further variability originates from multicentric study designs, varying acquisition protocols, and errors in the preprocessing of magnetic resonance imaging (MRI) scans. The high variability makes the differentiation between signal and noise difficult and may lead to overfitting. This article examines whether an automatic and fair data valuation method based on Shapley values can identify the most informative subjects to improve ML classification.

METHODS

An ML workflow was developed and trained for a subset of the Alzheimer's Disease Neuroimaging Initiative (ADNI) cohort. The validation was executed for an independent ADNI test set and for the Australian Imaging, Biomarker and Lifestyle Flagship Study of Ageing (AIBL) cohort. The workflow included volumetric MRI feature extraction, feature selection, sample selection using Data Shapley, random forest (RF), and eXtreme Gradient Boosting (XGBoost) for model training as well as Kernel SHapley Additive exPlanations (SHAP) values for model interpretation.

RESULTS

The RF models, which excluded 134 of the 467 training subjects based on their RF Data Shapley values, outperformed the base models that reached a mean accuracy of 62.64% by 5.76% (3.61 percentage points) for the independent ADNI test set. The XGBoost base models reached a mean accuracy of 60.00% for the AIBL data set. The exclusion of those 133 subjects with the smallest RF Data Shapley values could improve the classification accuracy by 2.98% (1.79 percentage points). The cutoff values were calculated using an independent validation set.

CONCLUSION

The Data Shapley method was able to improve the mean accuracies for the test sets. The most informative subjects were associated with the number of ApolipoproteinE ε4 (ApoE ε4) alleles, cognitive test results, and volumetric MRI measurements.

Signal and Texture Features from T2 Maps for the Prediction of Mild Cognitive Impairment to Alzheimer's Disease Progression

Early detection of Alzheimer's disease (AD) is crucial to preserve cognitive functions and provide the opportunity for patients to enter clinical trials. In recent years, some studies have reported that features related to the signal and texture of MRI images can be an effective biomarker of AD. To test these claims, a study was conducted using T2 maps, a sequence not previously studied, of 40 patients with mild cognitive impairment (MCI) from the Alzheimer's Disease Neuroimaging Initiative database, who either progressed to AD (18) or remained stable (22). From these maps, the mean value and absolute difference of 37 signal and texture imaging features for 40 contralateral pairs of regions were measured. We used seven machine learning methods to analyze whether, by adding these imaging features to the neuropsychological studies currently used for diagnosis, we could more accurately identify patients who will progress to AD. The predictive models improved with the addition of signal and texture features. Additionally, features related to the signal and texture of the images were much more relevant than volumetric ones. Our results suggest that contralateral signal and texture features should be further investigated as potential biomarkers for the prediction of AD.

Right entorhinal cortical thickness is associated with Mini-Mental State Examination scores from multi-country datasets using MRI

PURPOSE

To discover common biomarkers correlating with the Mini-Mental State Examination (MMSE) scores from multi-country MRI datasets.

METHODS

The first dataset comprised 112 subjects (49 men, 63 women; range, 46-94 years) at the National Hospital Organization Kyushu Medical Center. A second dataset comprised 300 subjects from the Alzheimer's Disease Neuroimaging Initiative (ADNI) database (177 men, 123 women; range, 57-91 years). Three-dimensional T1-weighted MR images were collected from both datasets. In total, 14 deep gray matter volumes and 70 cortical thicknesses were obtained from MR images using FreeSurfer software. Total hippocampal volume and the ratio of hippocampus to cerebral volume were also calculated. Correlations between each variable and MMSE scores were assessed using Pearson's correlation coefficient. Parameters with moderate correlation coefficients (r > 0.3) from each dataset were determined as independent variables and evaluated using general linear model (GLM) analyses.

RESULTS

In Pearson's correlation coefficient, total and bilateral hippocampal volumes, right amygdala volume, and right entorhinal cortex (ERC) thickness showed moderate correlation coefficients (r > 0.3) with MMSE scores from the first dataset. The ADNI dataset showed moderate correlations with MMSE scores in more variables, including bilateral ERC thickness and hippocampal volume. GLM analysis revealed that right ERC thickness correlated significantly with MMSE score in both datasets. Cortical thicknesses of the left parahippocampal gyrus, left inferior parietal lobe, and right fusiform gyrus also significantly correlated with MMSE score in the ADNI dataset (p < 0.05).

CONCLUSION

A positive correlation between right ERC thickness and MMSE score was identified from multi-country datasets.

Machine learning identifies novel markers predicting functional decline in older adults

The ability to carry out instrumental activities of daily living, such as paying bills, remembering appointments and shopping alone decreases with age, yet there are remarkable individual differences in the rate of decline among older adults. Understanding variables associated with a decline in instrumental activities of daily living is critical to providing appropriate intervention to prolong independence. Prior research suggests that cognitive measures, neuroimaging and fluid-based biomarkers predict functional decline. However, a priori selection of variables can lead to the over-valuation of certain variables and exclusion of others that may be predictive. In this study, we used machine learning techniques to select a wide range of baseline variables that best predicted functional decline in two years in individuals from the Alzheimer's Disease Neuroimaging Initiative dataset. The sample included 398 individuals characterized as cognitively normal or mild cognitive impairment. Support vector machine classification algorithms were used to identify the most predictive modality from five different data modality types (demographics, structural MRI, fluorodeoxyglucose-PET, neurocognitive and genetic/fluid-based biomarkers). In addition, variable selection identified individual variables across all modalities that best predicted functional decline in a testing sample. Of the five modalities examined, neurocognitive measures demonstrated the best accuracy in predicting functional decline (accuracy = 74.2%; area under the curve = 0.77), followed by fluorodeoxyglucose-PET (accuracy = 70.8%; area under the curve = 0.66). The individual variables with the greatest discriminatory ability for predicting functional decline included partner report of language in the Everyday Cognition questionnaire, the ADAS13, and activity of the left angular gyrus using fluorodeoxyglucose-PET. These three variables collectively explained 32% of the total variance in functional decline. Taken together, the machine learning model identified novel biomarkers that may be involved in the processing, retrieval, and conceptual integration of semantic information and which predict functional decline two years after assessment. These findings may be used to explore the clinical utility of the Everyday Cognition as a non-invasive, cost and time effective tool to predict future functional decline.

The topology of ventricle surfaces and its application in the analysis of hydrocephalic ventricles: a proof-of-concept study

PURPOSE

The cerebral ventricles deform in a non-uniform fashion in response to increased CSF volume and/or pressure in hydrocephalic syndromes. Current research is focused on volumetric analyses, while topological analysis of ventricular surfaces remains understudied. We developed a method of quantitatively modeling the curvature of ventricular surfaces to analyze changes in ventricular surfaces in normal pressure hydrocephalus (NPH) and Alzheimer's disease (AD), using the left frontal horn as an example.

METHODS

Twenty-one patients with NPH were recruited from our institution, and 21 healthy controls (HC) and patients with Alzheimer's disease (AD) were identified from the Alzheimer's Disease Neuroimaging Initiative (ADNI) database. On T1-weighted fine-cut magnetic resonance sequences, 3D Slicer was used to segment the left frontal horn. Next, the mean curvatures at a set of points on the ventricular surface were determined. The frontal horns were scaled and centered into normalized volumes, allowing for pooling across the study subjects. The frontal horn was divided into superolateral, superomedial, inferolateral, and inferomedial surfaces, and locoregional mean curvatures were analyzed. Statistical comparisons were made between NPH, AD, and HC groups.

RESULTS

Significant differences in the mean curvature of lateral surfaces of the ventricles distinguished patterns of distortion between all three cohorts. Significant flattening of the superomedial surface discriminated NPH from HC and AD. However, significant rounding of the inferomedial surface compared to controls was a distinguishing feature of NPH alone.

CONCLUSION

NPH ventricles deform non-uniformly. The pattern of surface distortion may be used as an additional tool to differentiate between these hydrocephalic conditions.

Structural Brain Differences Between Cognitively Impaired Patients With and Without Apathy

OBJECTIVE

Since apathy increases in prevalence with severity of dementia pathology, we sought to distinguish concomitant neurodegenerative processes from brain differences associated with apathy in persons with mild cognitive impairment (MCI) and Alzheimer's Disease (AD). We examined relative structural brain differences between case-control matched cognitively impaired patients with and without apathy.

DESIGN

Cross-sectional case-control study.

SETTING

Fifty-eight clinical sites in phase 2 of the AD Neuroimaging Initiative across the United States and Canada.

PARTICIPANTS

The ≥ 55 years of age with MCI or AD dementia and no major neurological disorders aside from suspected incipient AD dementia. Participants with apathy (n=69) were age-, sex-, apolipoprotein E ε4 allele carrier status-, Mini-Mental State Exam score-, and MCI or AD dementia diagnosis-matched to participants without apathy (n=149).

INTERVENTIONS

The 3-tesla T1-weighted MRI scan and neurocognitive assessments. Using the Neuropsychiatric Inventory apathy domain scores, participants were dichotomized into a with-apathy group (score ≥ 1) and a without-apathy group (score = 0).

MEASUREMENTS

Cortical thicknesses from 24 a priori regions of interest involved in frontostriatal circuits and frontotemporal association areas.

RESULTS

False-discovery rate adjusted within-group comparisons between participants with apathy and participants without apathy showed thinner right medial orbitofrontal (mOFC; meandifference(MD)±standarderrorofMD(SE)=-0.0879±0.0257mm; standardizedMD(d)=-0.4456) and left rostral anterior cingulate (rACC; MD±SE=-0.0905±0.0325mm; d=-0.3574) cortices and thicker left middle temporal cortices (MTC; MD±SE=0.0688±0.0239mm; d=0.3311) in those with apathy.

CONCLUSION

Atrophy of the right mOFC and left rACC and sparing of atrophy in the left MTC are associated with apathy in cognitively impaired persons.

Verbal memory and hippocampal volume predict subsequent fornix microstructure in those at risk for Alzheimer's disease

While strong cross-sectional evidence supported the use of fornix microstructure as a marker for detecting Alzheimer’s disease (AD), longitudinal data remains inconclusive on the sequential nature of fornix microstructure abnormalities and AD progression. An unequivocal longitudinal relationship between fornix microstructure and markers of AD progression –memory impairment and hippocampal atrophy, must be established to validate fornix microstructure as a marker of AD progression. We included 115 participants from the Alzheimer’s Disease Neuroimaging Initiative across the non-demented AD spectrum— defined as those who had at least one AD risk marker at baseline (e.g., mild cognitive impairment (MCI) due to AD diagnosis, amyloid or ApoE4 positivity) and/or ‘cognitively normal individuals who converted to MCI due to AD or AD, with structural and diffusion tensor imaging scans at baseline and two years follow-up. Hippocampal volumes (HV), fractional anisotropy (FA) and mean diffusivity (MD) in the fornix were extracted. Memory was indexed via composite scores of verbal memory tests. Structural equation models tested the bidirectional cross-lagged effects of fornix microstructure, memory, and HV. Impaired memory and smaller HV at baseline significantly predicted worse fornix microstructure (decreased FA and increased MD) two years later. Baseline fornix microstructure was not associated with subsequent changes in memory and HV. Fornix microstructure is compromised likely at a later stage, where significant decline in memory and hippocampal atrophy have occurred. This limits the utility of fornix microstructure in the early detection of AD. Our findings inform the possible pathophysiology and refined the use of AD neural markers.

Set-Based Rare Variant Expression Quantitative Trait Loci in Blood and Brain from Alzheimer Disease Study Participants

Because studies of rare variant effects on gene expression have limited power, we investigated set-based methods to identify rare expression quantitative trait loci (eQTL) related to Alzheimer disease (AD). Gene-level and pathway-level cis rare-eQTL mapping was performed genome-wide using gene expression data derived from blood donated by 713 Alzheimer's Disease Neuroimaging Initiative participants and from brain tissues donated by 475 Religious Orders Study/Memory and Aging Project participants. The association of gene or pathway expression with a set of all cis potentially regulatory low-frequency and rare variants within 1 Mb of genes was evaluated using SKAT-O. A total of 65 genes expressed in the brain were significant targets for rare expression single nucleotide polymorphisms (eSNPs) among which 17% (11/65) included established AD genes HLA-DRB1 and HLA-DRB5. In the blood, 307 genes were significant targets for rare eSNPs. In the blood and the brain, GNMT, LDHC, RBPMS2, DUS2, and HP were targets for significant eSNPs. Pathway enrichment analysis revealed significant pathways in the brain (n = 9) and blood (n = 16). Pathways for apoptosis signaling, cholecystokinin receptor (CCKR) signaling, and inflammation mediated by chemokine and cytokine signaling were common to both tissues. Significant rare eQTLs in inflammation pathways included five genes in the blood (ALOX5AP, CXCR2, FPR2, GRB2, IFNAR1) that were previously linked to AD. This study identified several significant gene- and pathway-level rare eQTLs, which further confirmed the importance of the immune system and inflammation in AD and highlighted the advantages of using a set-based eQTL approach for evaluating the effect of low-frequency and rare variants on gene expression.

The Associations of Cerebrospinal Fluid ApoE and Biomarkers of Alzheimer's Disease: Exploring Interactions With Sex

BACKGROUND

Sex-related difference in Alzheimer's disease (AD) has been proposed, and apolipoprotein E (ApoE) isoforms have been suggested to be involved in the pathogenesis of AD.

OBJECTIVE

We aimed to explore whether cerebrospinal fluid (CSF) ApoE is associated with AD biomarkers and whether the associations are different (between sexes).

METHODS

Data of 309 participants [92 with normal cognition, 148 with mild cognitive impairment (MCI), and 69 with AD dementia] from the Alzheimer's Disease Neuroimaging Initiative (ADNI) were cross-sectionally evaluated with the multiple linear regression model and longitudinally with the multivariate linear mixed-effects model for the associations of CSF ApoE with AD biomarkers. Sex-ApoE interaction was used to estimate whether sex moderates the associations of CSF ApoE and AD biomarkers.

RESULTS

Significant interactions between CSF ApoE and sex on AD biomarkers were observed [amyloid-β (Aβ): p = 0.0169 and phosphorylated-tau (p-tau): p = 0.0453]. In women, baseline CSF ApoE levels were significantly associated with baseline Aβ (p = 0.0135) and total-tau (t-tau) (p < 0.0001) as well as longitudinal changes of the biomarkers (Aβ: p = 0.0104; t-tau: p = 0.0110). In men, baseline CSF ApoE levels were only correlated with baseline p-tau (p < 0.0001) and t-tau (p < 0.0001) and did not aggravate AD biomarkers longitudinally.

CONCLUSION

The associations between CSF ApoE and AD biomarkers were sex-specific. Elevated CSF ApoE was associated with longitudinal changes of AD biomarkers in women, which indicates that CSF ApoE might be involved in the pathogenesis of AD pathology in a sex-specific way.

Machine Learning for Diagnosis of AD and Prediction of MCI Progression From Brain MRI Using Brain Anatomical Analysis Using Diffeomorphic Deformation

Background: With the growing momentum for the adoption of machine learning (ML) in medical field, it is likely that reliance on ML for imaging will become routine over the next few years. We have developed a software named BAAD, which uses ML algorithms for the diagnosis of Alzheimer's disease (AD) and prediction of mild cognitive impairment (MCI) progression. Methods: We constructed an algorithm by combining a support vector machine (SVM) to classify and a voxel-based morphometry (VBM) to reduce concerned variables. We grouped progressive MCI and AD as an AD spectrum and trained SVM according to this classification. We randomly selected half from the total 1,314 subjects of AD neuroimaging Initiative (ADNI) from North America for SVM training, and the remaining half were used for validation to fine-tune the model hyperparameters. We created two types of SVMs, one based solely on the brain structure (SVMst), and the other based on both the brain structure and Mini-Mental State Examination score (SVMcog). We compared the model performance with two expert neuroradiologists, and further evaluated it in test datasets involving 519, 592, 69, and 128 subjects from the Australian Imaging, Biomarker & Lifestyle Flagship Study of Aging (AIBL), Japanese ADNI, the Minimal Interval Resonance Imaging in AD (MIDIAD) and the Open Access Series of Imaging Studies (OASIS), respectively. Results: BAAD's SVMs outperformed radiologists for AD diagnosis in a structural magnetic resonance imaging review. The accuracy of the two radiologists was 57.5 and 70.0%, respectively, whereas, that of the SVMst was 90.5%. The diagnostic accuracy of the SVMst and SVMcog in the test datasets ranged from 88.0 to 97.1% and 92.5 to 100%, respectively. The prediction accuracy for MCI progression was 83.0% in SVMst and 85.0% in SVMcog. In the AD spectrum classified by SVMst, 87.1% of the subjects were Aβ positive according to an AV-45 positron emission tomography. Similarly, among MCI patients classified for the AD spectrum, 89.5% of the subjects progressed to AD. Conclusion: Our ML has shown high performance in AD diagnosis and prediction of MCI progression. It outperformed expert radiologists, and is expected to provide support in clinical practice.

Modelling Decline in Cognition to Decline in Function in Alzheimer's Disease

OBJECTIVES

The study aimed to evaluate and quantify the temporal link between cognitive and functional decline, and assess the impact of the apolipoprotein E4 (APOE-e4) genotype on Alzheimer's disease (AD) progression.

METHODS

A nonlinear mixed-effects Emax model was developed using longitudinal data from 659 patients with dementia due to AD sourced from the Alzheimer's disease neuroimaging initiative (ADNI) database. A cognitive decline model was first built using a cognitive subscale of the AD assessment scale (delayed word recall) as the endpoint, followed by a functional decline model, using the functional assessment questionnaire (FAQ) as the endpoint. Individual and population cognitive decline from the first model drove a functional decline in the second model. The impact of the APOE-e4 genotype status on the dynamics of AD progression was evaluated using the model.

RESULTS

Mixed-effects Emax models adequately quantified population average and individual disease trajectories. The model captured a higher initial cognitive impairment and final functional impairment in APOE-e4 carriers than non-carriers. The age at cognitive decline and diagnosis of dementia due to AD was significantly lower in APOE-e4 carriers than that of non-carriers. The average [standard deviation] time shift between cognitive and functional decline, i.e. the time span between half of the maximum cognitive decline and half of the maximum functional decline, was estimated as 1.5 [1.6] years.

CONCLUSION

The present analysis quantifies the temporal link between a cognitive and functional decline in AD progression at the population and individual level, and provides information about the potential benefits of pre-clinical AD treatments on both cognition and function.

Bayesian sparse heritability analysis with high-dimensional neuroimaging phenotypes

Heritability analysis plays a central role in quantitative genetics to describe genetic contribution to human complex traits and prioritize downstream analyses under large-scale phenotypes. Existing works largely focus on modeling single phenotype and currently available multivariate phenotypic methods often suffer from scaling and interpretation. In this article, motivated by understanding how genetic underpinning impacts human brain variation, we develop an integrative Bayesian heritability analysis to jointly estimate heritabilities for high-dimensional neuroimaging traits. To induce sparsity and incorporate brain anatomical configuration, we impose hierarchical selection among both regional and local measurements based on brain structural network and voxel dependence. We also use a nonparametric Dirichlet process mixture model to realize grouping among single nucleotide polymorphism-associated phenotypic variations, providing biological plausibility. Through extensive simulations, we show the proposed method outperforms existing ones in heritability estimation and heritable traits selection under various scenarios. We finally apply the method to two large-scale imaging genetics datasets: the Alzheimer's Disease Neuroimaging Initiative and United Kingdom Biobank and show biologically meaningful results.

Impacts of CR1 genetic variants on cerebrospinal fluid and neuroimaging biomarkers in alzheimer's disease

BACKGROUND

The complement component (3b/4b) receptor 1 gene (CR1) gene has been proved to affect the susceptibility of Alzheimer's disease (AD) in different ethnic and districts groups. However, the effect of CR1 genetic variants on amyloid β (Aβ) metabolism of AD human is still unclear. Hence, the aim of this study was to investigate genetic influences of CR1 gene on Aβ metabolism.

METHODS

All data of AD patients and normal controls (NC) were obtained from alzheimer's disease neuroimaging initiative database (ADNI) database. In order to assess the effect of each single nucleotide polymorphism (SNP) of CR1 on Aβ metabolism, the PLINK software was used to conduct the quality control procedures to enroll appropriate SNPs. Moreover, the correlation between CR1 genotypes and Aβ metabolism in all participants were estimated with multiple linear regression models.

RESULTS

After quality control procedures, a total of 329 samples and 83 SNPs were enrolled in our study. Moreover, our results identified five SNPs (rs10494884, rs11118322, rs1323721, rs17259045 and rs41308433), which were linked to Aβ accumulation in brain. In further analyses, rs17259045 was found to decrease Aβ accumulation among AD patients. Additionally, our study revealed the genetic variants in rs12567945 could increase CSF Aβ42 in NC population.

CONCLUSIONS

Our study had revealed several novel SNPs in CR1 genes which might be involved in the progression of AD via regulating Aβ accumulation. These findings will provide a new basis for the diagnosis and treatment AD.

Genome-wide transcriptome analysis identifies novel dysregulated genes implicated in Alzheimer's pathology

INTRODUCTION

Abnormal gene expression patterns may contribute to the onset and progression of late-onset Alzheimer's disease (LOAD).

METHODS

We performed transcriptome-wide meta-analysis (N = 1440) of blood-based microarray gene expression profiles as well as neuroimaging and cerebrospinal fluid (CSF) endophenotype analysis.

RESULTS

We identified and replicated five genes (CREB5, CD46, TMBIM6, IRAK3, and RPAIN) as significantly dysregulated in LOAD. The most significantly altered gene, CREB5, was also associated with brain atrophy and increased amyloid beta (Aβ) accumulation, especially in the entorhinal cortex region. cis-expression quantitative trait loci mapping analysis of CREB5 detected five significant associations (P < 5 × 10^-8 ), where rs56388170 (most significant) was also significantly associated with global cortical Aβ deposition measured by [¹⁸ F]Florbetapir positron emission tomography and CSF Aβ_1-42 .

DISCUSSION

RNA from peripheral blood indicated a differential gene expression pattern in LOAD. Genes identified have been implicated in biological processes relevant to Alzheimer's disease. CREB, in particular, plays a key role in nervous system development, cell survival, plasticity, and learning and memory.

Semantic Segmentation of White Matter in FDG-PET Using Generative Adversarial Network

In the diagnosis of neurodegenerative disorders, F-18 fluorodeoxyglucose positron emission tomography/computed tomography (18F-FDG PET/CT) is used for its ability to detect functional changes at early stages of disease process. However, anatomical information from another modality (CT or MRI) is still needed to properly interpret and localize the radiotracer uptake due to its low spatial resolution. Lack of structural information limits segmentation and accurate quantification of the 18F-FDG PET/CT. The correct segmentation of the brain compartment in 18F-FDG PET/CT will enable the quantitative analysis of the 18F-FDG PET/CT scan alone. In this paper, we propose a method to segment white matter in 18F-FDG PET/CT images using generative adversarial network (GAN). The segmentation result of GAN model was evaluated using evaluation parameters such as dice, AUC-PR, precision, and recall. It was also compared with other deep learning methods. As a result, the proposed method achieves superior segmentation accuracy and reliability compared with other deep learning methods.

The correlation of everyday cognition test scores and the progression of Alzheimer's disease: a data analytics study

The process of diagnosing dementia conditions, especially Alzheimer's disease, and the cognitive tests that are involved in this process, are important areas of study. Everyday Cognition (ECog) is one test that can be used as part of Alzheimer's disease diagnosis to measure cognitive decline in different areas. In this study, we investigate two versions of the ECog test: the study partner reported version (ECogSP), and the patient reported version (ECogPT). We compare these, using statistical analysis and machine learning techniques, to create classification models to demonstrate the progression in ECog scores over time by using the Alzheimer's Disease Neuroimaging Initiative longitudinal data repository (ADNI); participants are classed with having normal cognition, mild cognitive impairment, or Alzheimer's disease. We found that participants who are diagnosed with Alzheimer's disease at baseline, or during a subsequent visit, tend to self-report consistent ECogPT scores over time indicating no change in cognitive ability. However, study partners tend to report higher and increasing ECogSP scores on behalf of participants in the same diagnosis category; this would indicate a degradation in the participant's cognitive ability over time, consistent with the progress of Alzheimer's disease.

Longitudinal ComBat: A method for harmonizing longitudinal multi-scanner imaging data

While aggregation of neuroimaging datasets from multiple sites and scanners can yield increased statistical power, it also presents challenges due to systematic scanner effects. This unwanted technical variability can introduce noise and bias into estimation of biological variability of interest. We propose a method for harmonizing longitudinal multi-scanner imaging data based on ComBat, a method originally developed for genomics and later adapted to cross-sectional neuroimaging data. Using longitudinal cortical thickness measurements from 663 participants in the Alzheimer's Disease Neuroimaging Initiative (ADNI) study, we demonstrate the presence of additive and multiplicative scanner effects in various brain regions. We compare estimates of the association between diagnosis and change in cortical thickness over time using three versions of the ADNI data: unharmonized data, data harmonized using cross-sectional ComBat, and data harmonized using longitudinal ComBat. In simulation studies, we show that longitudinal ComBat is more powerful for detecting longitudinal change than cross-sectional ComBat and controls the type I error rate better than unharmonized data with scanner included as a covariate. The proposed method would be useful for other types of longitudinal data requiring harmonization, such as genomic data, or neuroimaging studies of neurodevelopment, psychiatric disorders, or other neurological diseases.

Multi-atlas image registration of clinical data with automated quality assessment using ventricle segmentation

Registration is a core component of many imaging pipelines. In case of clinical scans, with lower resolution and sometimes substantial motion artifacts, registration can produce poor results. Visual assessment of registration quality in large clinical datasets is inefficient. In this work, we propose to automatically assess the quality of registration to an atlas in clinical FLAIR MRI scans of the brain. The method consists of automatically segmenting the ventricles of a given scan using a neural network, and comparing the segmentation to the atlas ventricles propagated to image space. We used the proposed method to improve clinical image registration to a general atlas by computing multiple registrations - one directly to the general atlas and others via different age-specific atlases - and then selecting the registration that yielded the highest ventricle overlap. Finally, as an example application of the complete pipeline, a voxelwise map of white matter hyperintensity burden was computed using only the scans with registration quality above a predefined threshold. Methods were evaluated in a single-site dataset of more than 1000 scans, as well as a multi-center dataset comprising 142 clinical scans from 12 sites. The automated ventricle segmentation reached a Dice coefficient with manual annotations of 0.89 in the single-site dataset, and 0.83 in the multi-center dataset. Registration via age-specific atlases could improve ventricle overlap compared to a direct registration to the general atlas (Dice similarity coefficient increase up to 0.15). Experiments also showed that selecting scans with the registration quality assessment method could improve the quality of average maps of white matter hyperintensity burden, instead of using all scans for the computation of the white matter hyperintensity map. In this work, we demonstrated the utility of an automated tool for assessing image registration quality in clinical scans. This image quality assessment step could ultimately assist in the translation of automated neuroimaging pipelines to the clinic.

Machine Learning Models Reveal the Importance of Clinical Biomarkers for the Diagnosis of Alzheimer's Disease

Alzheimer's Disease (AD) is a neurodegenerative disease that causes complications with thinking capability, memory and behavior. AD is a major public health problem among the elderly in developed and developing countries. With the growth of AD around the world, there is a need to further expand our understanding of the roles different clinical measurements can have in the diagnosis of AD. In this work, we propose a machine learning-based technique to distinguish control subjects with no cognitive impairments, AD subjects, and subjects with mild cognitive impairment (MCI), often seen as precursors of AD. We utilized several machine learning (ML) techniques and found that Gradient Boosting Decision Trees achieved the highest performance above 84% classification accuracy. Also, we determined the importance of the features (clinical biomarkers) contributing to the proposed multi-class classification system. Further investigation on the biomarkers will pave the way to introduce better treatment plan for AD patients.

Harnessing peripheral DNA methylation differences in the Alzheimer's Disease Neuroimaging Initiative (ADNI) to reveal novel biomarkers of disease

BACKGROUND

Alzheimer's disease (AD) is a chronic progressive neurodegenerative disease impacting an estimated 44 million adults worldwide. The causal pathology of AD (accumulation of amyloid-beta and tau), precedes hallmark symptoms of dementia by more than a decade, necessitating development of early diagnostic markers of disease onset, particularly for new drugs that aim to modify disease processes. To evaluate differentially methylated positions (DMPs) as novel blood-based biomarkers of AD, we used a subset of 653 individuals with peripheral blood (PB) samples in the Alzheimer's disease Neuroimaging Initiative (ADNI) consortium. The selected cohort of AD, mild cognitive impairment (MCI), and age-matched healthy controls (CN) all had imaging, genetics, transcriptomics, cerebrospinal protein markers, and comprehensive clinical records, providing a rich resource of concurrent multi-omics and phenotypic information on a well-phenotyped subset of ADNI participants.

RESULTS

In this manuscript, we report cross-diagnosis differential peripheral DNA methylation in a cohort of AD, MCI, and age-matched CN individuals with longitudinal DNA methylation measurements. Epigenome-wide association studies (EWAS) were performed using a mixed model with repeated measures over time with a P value cutoff of 1 × 10-5 to test contrasts of pairwise differential peripheral methylation in AD vs CN, AD vs MCI, and MCI vs CN. The most highly significant differentially methylated loci also tracked with Mini Mental State Examination (MMSE) scores. Differentially methylated loci were enriched near brain and neurodegeneration-related genes (e.g., BDNF, BIN1, APOC1) validated using the genotype tissue expression project portal (GTex).

CONCLUSIONS

Our work shows that peripheral differential methylation between age-matched subjects with AD relative to healthy controls will provide opportunities to further investigate and validate differential methylation as a surrogate of disease. Given the inaccessibility of brain tissue, the PB-associated methylation marks may help identify the stage of disease and progression phenotype, information that would be central to bringing forward successful drugs for AD.