Search

Abstract Recent evidence suggests that Alzheimer’s disease (AD) genetic risk variants (rs1582763 and rs6591561) of the MS4A locus are genome-wide significant regulators of soluble TREM2 levels such that the minor allele of the protective variant (rs1582763) is associated with higher sTREM2 and lower AD risk while the minor allele of (rs6591561) relates to lower sTREM2 and higher AD risk. Our group previously found that higher sTREM2 relates to higher Aβ40, worse blood–brain barrier (BBB) integrity (measured with the CSF/plasma albumin ratio), and higher CSF tau, suggesting strong associations with amyloid abundance and both BBB and neurodegeneration complicate interpretation. We expand on this work by leveraging these common variants as genetic tools to tune the interpretation of high CSF sTREM2, and by exploring the potential modifying role of these variants on the well-established associations between CSF sTREM2 as well as TREM2 transcript levels in the brain with AD neuropathology. Biomarker analyses leveraged data from the Vanderbilt Memory & Aging Project (n = 127, age = 72 ± 6.43) and were replicated in the Alzheimer’s Disease Neuroimaging Initiative (n = 399, age = 73 ± 7.39). Autopsy analyses were performed leveraging data from the Religious Orders Study and Rush Memory and Aging Project (n = 577, age = 89 ± 6.46). We found that the protective variant rs1582763 attenuated the association between CSF sTREM2 and Aβ40 (β = -0.44, p-value = 0.017) and replicated this interaction in ADNI (β = -0.27, p = 0.017). We did not observe this same interaction effect between TREM2 mRNA levels and Aβ peptides in brain (Aβ total β = -0.14, p = 0.629; Aβ1-38, β = 0.11, p = 0.200). In contrast to the effects on Aβ, the minor allele of this same variant seemed to enhance the association with blood–brain barrier dysfunction (β = 7.0e-4, p = 0.009), suggesting that elevated sTREM2 may carry a much different interpretation in carriers vs. non-carriers of this allele. When evaluating the risk variant (rs6591561) across datasets, we did not observe a statistically significant interaction against any outcome in VMAP and observed opposing directions of associations in ADNI and ROS/MAP on Aβ levels. Together, our results suggest that the protective effect of rs1582763 may act by decoupling the associations between sTREM2 and amyloid abundance, providing important mechanistic insight into sTREM2 changes and highlighting the need to incorporate genetic context into the analysis of sTREM2 levels, particularly if leveraged as a clinical biomarker of disease in the future.

Abstract Objectives Mild cognitive impairment presenting with an amnestic syndrome (aMCI) and amyloid positivity is considered due to AD. Many subjects, however, can show an overall very slow progression relevant for differential diagnosis, prognosis, and treatment. This study assessed PET biomarkers, including brain glucose metabolism, tau, and amyloid load, in a series of comparable aMCI at baseline, clinically evaluated at follow‐up. Methods We included 72 aMCI subjects from Geneva Memory Center (N = 31) and ADNI cohorts (N = 41), selected based on available FDG‐PET, tau‐PET, amyloid‐PET, and clinical follow‐up (2.3 years ± 1.2). A data‐driven algorithm classified brain metabolic patterns into subtypes that were then compared for clinical and PET biomarker measures and cognitive decline. Voxel‐wise comparisons were performed both with FDG‐PET and tau‐PET data. Results The algorithm classified three metabolic subtypes, namely “Hippocampal‐sparing with cortical hypometabolism” (Type1; N = 27), “Hippocampal and cortical hypometabolism” (Type 2; N = 23), and “Medial temporal hypometabolism” (Type 3; N = 22). Amyloid positivity and tau accumulation in the medial temporal and neocortical regions characterized Type 1 and Type 2, whereas Type 3 showed no significant tau pathology, variable amyloid positivity, and stability at follow‐up. All tau‐positive patients, independently of the FDG‐based subtype, showed faster cognitive decline. Interpretation aMCI subjects can differ in metabolic patterns, tau and amyloid pathology, and clinical progression. Here, we complemented with PET tau biomarker the specific brain hypometabolic patterns at the individual level in the prodromal phase, contributing to the patient's classification. Tau PET is the most accurate biomarker in supporting or excluding the AD diagnosis in aMCI across metabolic subtypes and also predicting the risk of decline.

BackgroundThe relationship between white matter hyperintensities (WMH) and the core features of Alzheimer’s disease (AD) remains controversial. Further, due to the prevalence of co-pathologies, the precise role of WMH in cognition and neurodegeneration also remains uncertain.MethodsHerein, we analyzed 1803 participants with available WMH volume data, extracted from the ADNI database, including 756 cognitively normal controls, 783 patients with mild cognitive impairment (MCI), and 264 patients with dementia. Participants were grouped according to cerebrospinal fluid (CSF) pathology (A/T profile) severity. Linear regression analysis was applied to evaluate the factors associated with WMH volume. Modeled by linear mixed-effects, the increase rates (Δ) of the WMH volume, cognition, and typical neurodegenerative markers were assessed. The predictive effectiveness of WMH volume was subsequently tested using Cox regression analysis, and the relationship between WMH/ΔWMH and other indicators such as cognition was explored through linear regression analyses. Furthermore, we explored the interrelationship among amyloid-β deposition, cognition, and WMH using mediation analysis.ResultsHigher WMH volume was associated with older age, lower CSF amyloid-β levels, hypertension, and smoking history (all p ≤ 0.001), as well as cognitive status (MCI, p

Abstract Background Polygenic effects have been proposed to account for some disease phenotypes; these effects are calculated as a polygenic risk score (PRS). This score is correlated with Alzheimer’s disease (AD)-related phenotypes, such as biomarker abnormalities and brain atrophy, and is associated with conversion from mild cognitive impairment (MCI) to AD. However, the AD PRS has been examined mainly in Europeans, and owing to differences in genetic structure and lifestyle, it is unclear whether the same relationships between the PRS and AD-related phenotypes exist in non-European populations. In this study, we calculated and evaluated the AD PRS in Japanese individuals using genome-wide association study (GWAS) statistics from Europeans. Methods In this study, we calculated the AD PRS in 504 Japanese participants (145 cognitively unimpaired (CU) participants, 220 participants with late mild cognitive impairment (MCI), and 139 patients with mild AD dementia) enrolled in the Japanese Alzheimer’s Disease Neuroimaging Initiative (J-ADNI) project. In order to evaluate the clinical value of this score, we (1) determined the polygenic effects on AD in the J-ADNI and validated it using two independent cohorts (a Japanese neuropathology (NP) cohort (n = 565) and the North American ADNI (NA-ADNI) cohort (n = 617)), (2) examined the AD-related phenotypes associated with the PRS, and (3) tested whether the PRS helps predict the conversion of MCI to AD. Results The PRS using 131 SNPs had an effect independent of APOE. The PRS differentiated between CU participants and AD patients with an area under the curve (AUC) of 0.755 when combined with the APOE variants. Similar AUC was obtained when PRS calculated by the NP and NA-ADNI cohorts was applied. In MCI patients, the PRS was associated with cerebrospinal fluid phosphorylated-tau levels (β estimate = 0.235, p value = 0.026). MCI with a high PRS showed a significantly increased conversion to AD in APOE ε4 noncarriers with a hazard rate of 2.22. In addition, we also developed a PRS model adjusted for LD and observed similar results. Conclusions We showed that the AD PRS is useful in the Japanese population, whose genetic structure is different from that of the European population. These findings suggest that the polygenicity of AD is partially common across ethnic differences.

Abstract Background Alzheimer’s Disease (AD) is a devastating disease that destroys memory and other cognitive functions. There has been an increasing research effort to prevent and treat AD. In the US, two major data sharing resources for AD research are the National Alzheimer’s Coordinating Center (NACC) and the Alzheimer’s Disease Neuroimaging Initiative (ADNI); Additionally, the National Institutes of Health (NIH) Common Data Elements (CDE) Repository has been developed to facilitate data sharing and improve the interoperability among data sets in various disease research areas. Method To better understand how AD-related data elements in these resources are interoperable with each other, we leverage different representation models to map data elements from different resources: NACC to ADNI, NACC to NIH CDE, and ADNI to NIH CDE. We explore bag-of-words based and word embeddings based models (Word2Vec and BioWordVec) to perform the data element mappings in these resources. Results The data dictionaries downloaded on November 23, 2021 contain 1,195 data elements in NACC, 13,918 in ADNI, and 27,213 in NIH CDE Repository. Data element preprocessing reduced the numbers of NACC and ADNI data elements for mapping to 1,099 and 7,584 respectively. Manual evaluation of the mapping results showed that the bag-of-words based approach achieved the best precision, while the BioWordVec based approach attained the best recall. In total, the three approaches mapped 175 out of 1,099 (15.92%) NACC data elements to ADNI; 107 out of 1,099 (9.74%) NACC data elements to NIH CDE; and 171 out of 7,584 (2.25%) ADNI data elements to NIH CDE. Conclusions The bag-of-words based and word embeddings based approaches showed promise in mapping AD-related data elements between different resources. Although the mapping approaches need further improvement, our result indicates that there is a critical need to standardize CDEs across these valuable AD research resources in order to maximize the discoveries regarding AD pathophysiology, diagnosis, and treatment that can be gleaned from them.

Abstract Background Joint analysis of multiple phenotypes in studies of biological systems such as Genome-Wide Association Studies is critical to revealing the functional interactions between various traits and genetic variants, but growth of data in dimensionality has become a very challenging problem in the widespread use of joint analysis. To handle the excessiveness of variables, we consider the sliced inverse regression (SIR) method. Specifically, we propose a novel SIR-based association test that is robust and powerful in testing the association between multiple predictors and multiple outcomes. Results We conduct simulation studies in both low- and high-dimensional settings with various numbers of Single-Nucleotide Polymorphisms and consider the correlation structure of traits. Simulation results show that the proposed method outperforms the existing methods. We also successfully apply our method to the genetic association study of ADNI dataset. Both the simulation studies and real data analysis show that the SIR-based association test is valid and achieves a higher efficiency compared with its competitors. Conclusion Several scenarios with low- and high-dimensional responses and genotypes are considered in this paper. Our SIR-based method controls the estimated type I error at the pre-specified level $$\alpha $$ α .

Abstract Various plasma biomarkers for amyloid-β (Aβ) have shown high predictability of amyloid PET positivity. However, the characteristics of discordance between amyloid PET and plasma Aβ42/40 positivity are poorly understood. Thorough interpretation of discordant cases is vital as Aβ plasma biomarker is imminent to integrate into clinical guidelines. We aimed to determine the characteristics of discordant groups between amyloid PET and plasma Aβ42/40 positivity, and inter-assays variability depending on plasma assays. We compared tau burden measured by PET, brain volume assessed by MRI, cross-sectional cognitive function, longitudinal cognitive decline and polygenic risk score (PRS) between PET/plasma groups (PET−/plasma−, PET−/plasma+, PET+/plasma−, PET+/plasma+) using Alzheimer’s Disease Neuroimaging Initiative database. Additionally, we investigated inter-assays variability between immunoprecipitation followed by mass spectrometry method developed at Washington University (IP-MS-WashU) and Elecsys immunoassay from Roche (IA-Elc). PET+/plasma+ was significantly associated with higher tau burden assessed by PET in entorhinal, Braak III/IV, and Braak V/VI regions, and with decreased volume of hippocampal and precuneus regions compared to PET−/plasma-. PET+/plasma+ showed poor performances in global cognition, memory, executive and daily-life function, and rapid cognitive decline. PET+/plasma+ was related to high PRS. The PET−/plasma+ showed intermediate changes between PET−/plasma− and PET+/plasma+ in terms of tau burden, hippocampal and precuneus volume, cross-sectional and longitudinal cognition, and PRS. PET+/plasma− represented heterogeneous characteristics with most prominent variability depending on plasma assays. Moreover, IP-MS-WashU showed more linear association between amyloid PET standardized uptake value ratio and plasma Aβ42/40 than IA-Elc. IA-Elc showed more plasma Aβ42/40 positivity in the amyloid PET-negative stage than IP-MS-WashU. Characteristics of PET−/plasma+ support plasma biomarkers as early biomarker of amyloidopathy prior to amyloid PET. Various plasma biomarker assays might be applied distinctively to detect different target subjects or disease stages.

Abstract Objective To explore the functional connectivity (FC) characteristics of the episodic memory network (EMN) in amnestic mild cognitive impairment (aMCI) patients with different levels of executive function (EF). Methods This study included 76 participants from the Alzheimer's Disease Neuroimaging Initiative database, comprising 23 healthy controls (HCs) and 53 aMCI patients. Based on EF levels, aMCI patients were categorized into aMCI‐highEF and aMCI‐lowEF groups. Cognitive function scores, pathological markers (cerebrospinal fluid β‐amyloid, total tau protein, phosphorylated tau protein, AV45‐PET, and FDG‐PET), and functional magnetic resonance imaging were collected and compared among the three groups. Seed‐based FC analysis was used to examine differences in the EMN among the groups, and partial correlation analysis was employed to investigate the relationship between changes in FC and cognitive function scores as well as pathological markers. Results Compared to the aMCI‐highEF group, the aMCI‐lowEF group exhibited more severe cognitive impairment, decreased cerebral glucose metabolism, and elevated AV45 levels. Significant FC differences in the left superior temporal gyrus (STG) of the EMN were observed among the three groups. Post hoc analysis revealed that the aMCI‐lowEF group had increased FC in the left STG compared to the HCs and aMCI‐highEF groups, with statistically significant differences. Correlation analysis showed a significant negative correlation between the differences in FC in the left STG of aMCI‐highEF and aMCI‐lowEF groups and Rey Auditory Verbal Learning Test forgetting scores. Receiver operator characteristic curve analysis indicated an area under the curve of 0.741 for distinguishing between aMCI‐highEF and aMCI‐lowEF groups based on FC of left STG, with a sensitivity of 0.808 and a specificity of 0.667. Conclusion aMCI‐lowEF exhibits characteristic changes in FC within the EMN, providing theoretical support for the role of EF in mediating EMN alterations and, consequently, impacting episodic memory function.

BackgroundResearch has shown disrupted structural network measures related to cognitive decline and future cortical atrophy during the progression of Alzheimer’s disease (AD). However, evidence regarding the individual variability of gray matter network measures and the associations with concurrent cognitive decline and cortical atrophy related to AD is still sparse.ObjectiveTo investigate whether alterations in single-subject gray matter networks are related to concurrent cognitive decline and cortical gray matter atrophy during AD progression.MethodsWe analyzed structural MRI data from 185 cognitively normal (CN), 150 mild cognitive impairment (MCI), and 153 AD participants, and calculated the global network metrics of gray matter networks for each participant. We examined the alterations of single-subject gray matter networks in patients with MCI and AD, and investigated the associations of network metrics with concurrent cognitive decline and cortical gray matter atrophy.ResultsThe small-world properties including gamma, lambda, and sigma had lower values in the MCI and AD groups than the CN group. AD patients had reduced degree, clustering coefficient, and path length than the CN and MCI groups. We observed significant associations of cognitive ability with degree in the CN group, with gamma and sigma in the MCI group, and with degree, connectivity density, clustering coefficient, and path length in the AD group. There were significant correlation patterns between sigma values and cortical gray matter volume in the CN, MCI, and AD groups.ConclusionThese findings suggest the individual variability of gray matter network metrics may be valuable to track concurrent cognitive decline and cortical atrophy during AD progression. This may contribute to a better understanding of cognitive decline and brain morphological alterations related to AD.

In this study, we address the early detection challenges of Alzheimer’s disease (AD) using explainable artificial intelligence (XAI) techniques. AD, characterized by amyloid plaques and tau tangles, leads to cognitive decline and remains hard to diagnose due to genetic and environmental factors. Utilizing deep learning models, we analyzed brain MRI scans from the ADNI database, categorizing them into normal cognition (NC), mild cognitive impairment (MCI), and AD. The ResNet-50 architecture was employed, enhanced by a channel-wise attention mechanism to improve feature extraction. To ensure model transparency, we integrated local interpretable model-agnostic explanations (LIMEs) and gradient-weighted class activation mapping (Grad-CAM), highlighting significant image regions contributing to predictions. Our model achieved 85% accuracy, effectively distinguishing between the classes. The LIME and Grad-CAM visualizations provided insights into the model’s decision-making process, particularly emphasizing changes near the hippocampus for MCI. These XAI methods enhance the interpretability of AI-driven AD diagnosis, fostering trust and aiding clinical decision-making. Our approach demonstrates the potential of combining deep learning with XAI for reliable and transparent medical applications.

Abstract Background There has been increasing interest in cortical microstructure as a complementary and earlier measure of neurodegeneration than macrostructural atrophy, but few papers have related cortical diffusion imaging to post-mortem neuropathology. This study aimed to characterise the associations between the main Alzheimer’s disease (AD) neuropathological hallmarks and multiple cortical microstructural measures from in vivo diffusion MRI. Comorbidities and co-pathologies were also investigated. Methods Forty-three autopsy cases (8 cognitively normal, 9 mild cognitive impairment, 26 AD) from the National Alzheimer’s Coordinating Center and Alzheimer’s Disease Neuroimaging Initiative databases were included. Structural and diffusion MRI scans were analysed to calculate cortical minicolumn-related measures (AngleR, PerpPD+, and ParlPD) and mean diffusivity (MD). Neuropathological hallmarks comprised Thal phase, Braak stage, neuritic plaques, and combined AD neuropathological changes (ADNC—the “ABC score” from NIA-AA recommendations). Regarding comorbidities, relationships between cortical microstructure and severity of white matter rarefaction (WMr), cerebral amyloid angiopathy (CAA), atherosclerosis of the circle of Willis (ACW), and locus coeruleus hypopigmentation (LCh) were investigated. Finally, the effect of coexistent pathologies—Lewy body disease and TAR DNA-binding protein 43 (TDP-43)—on cortical microstructure was assessed. Results Cortical diffusivity measures were significantly associated with Thal phase, Braak stage, ADNC, and LCh. Thal phase was associated with AngleR in temporal areas, while Braak stage was associated with PerpPD+ in a wide cortical pattern, involving mainly temporal and limbic areas. A similar association was found between ADNC (ABC score) and PerpPD+. LCh was associated with PerpPD+, ParlPD, and MD. Co-existent neuropathologies of Lewy body disease and TDP-43 exhibited significantly reduced AngleR and MD compared to ADNC cases without co-pathology. Conclusions Cortical microstructural diffusion MRI is sensitive to AD neuropathology. The associations with the LCh suggest that cortical diffusion measures may indirectly reflect the severity of locus coeruleus neuron loss, perhaps mediated by the severity of microglial activation and tau spreading across the brain. Recognizing the impact of co-pathologies is important for diagnostic and therapeutic decision-making. Microstructural markers of neurodegeneration, sensitive to the range of histopathological features of amyloid, tau, and monoamine pathology, offer a more complete picture of cortical changes across AD than conventional structural atrophy.

Abstract Background In the United States, the National Alzheimer’s Coordinating Center (NACC) and the Alzheimer’s Disease Neuroimaging Initiative (ADNI) are two major data sharing resources for Alzheimer’s Disease (AD) research. NACC and ADNI strive to make their data more FAIR (findable, interoperable, accessible and reusable) for the broader research community. However, there is limited work harmonizing and supporting cross-cohort interoperability of the two resources. Method In this paper, we leverage an ontology-based approach to harmonize data elements in the two resources and develop a web-based query system to search patient cohorts across the two resources. We first mapped data elements across NACC and ADNI, and performed value harmonization for the mapped data elements with inconsistent permissible values. Then we built an Alzheimer’s Disease Data Element Ontology (ADEO) to model the mapped data elements in NACC and ADNI. We further developed a prototype cross-cohort query system to search patient cohorts across NACC and ADNI. Results After manual review, we found 172 mappings between NACC and ADNI. These 172 mappings were further used to construct common concepts in ADEO. Our data element mapping and harmonization resulted in five files storing common concepts, variables in NACC and ADNI, mappings between variables and common concepts, permissible values of categorical type data elements, and coding inconsistency harmonization, respectively. Our cross-cohort query system consists of three core architectural elements: a web-based interface, an advanced query engine, and a backend MongoDB database. Conclusions In this work, ADEO has been specifically designed to facilitate data harmonization and cross-cohort query of NACC and ADNI data resources. Although our prototype cross-cohort query system was developed for exploring NACC and ADNI, its backend and frontend framework has been designed and implemented to be generally applicable to other domains for querying patient cohorts from multiple heterogeneous data sources.

Abstract Background Increasing research suggests that gait abnormalities can be a risk factor for Alzheimer’s Disease (AD). Notably, there is growing evidence highlighting this risk factor in individuals with amnestic Mild Cognitive Impairment (aMCI), however further studies are needed. The aim of this study is to analyze cognitive tests results and brain-related measures over time in aMCI and examine how the presence of gait abnormalities (neurological or orthopedic) or normal gait affects these trends. Additionally, we sought to assess the significance of gait and gait-related measures as prognostic indicators for the progression from aMCI to AD dementia, comparing those who converted to AD with those who remained with a stable aMCI diagnosis during the follow-up. Methods Four hundred two individuals with aMCI from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) database were included. Robust linear mixed-effects models were used to study the impact of gait abnormalities on a comprehensive neuropsychological battery over 36 months while controlling for relevant medical variables at baseline. The impact of gait on brain measures was also investigated. Lastly, the Cox proportional-hazards model was used to explore the prognostic relevance of abnormal gait and neuropsychological associated tests. Results While controlling for relevant covariates, we found that gait abnormalities led to a greater decline over time in attention (DSST) and global cognition (MMSE). Intriguingly, psychomotor speed (TMT-A) and divided attention (TMT-B) declined uniquely in the abnormal gait group. Conversely, specific AD global cognition tests (ADAS-13) and auditory-verbal memory (RAVLT immediate recall) declined over time independently of gait profile. All the other cognitive tests were not significantly affected by time or by gait profile. In addition, we found that ventricles size increased faster in the abnormal gait group compared to the normal gait group. In terms of prognosis, abnormal gait (HR = 1.7), MMSE (HR = 1.09), and DSST (HR = 1.03) covariates showed a higher impact on AD dementia conversion. Conclusions The importance of the link between gait and related cognitive functions in terms of diagnosis, prognosis, and rehabilitation in aMCI is critical. We showed that in aMCI gait abnormalities lead to executive functions/attention deterioration and conversion to AD dementia.

Abstract Background Identifying prediagnostic neurodegenerative disease is a critical issue in neurodegenerative disease research, and Alzheimer’s disease (AD) in particular, to identify populations suitable for preventive and early disease-modifying trials. Evidence from genetic and other studies suggests the neurodegeneration of Alzheimer’s disease measured by brain atrophy starts many years before diagnosis, but it is unclear whether these changes can be used to reliably detect prediagnostic sporadic disease. Methods We trained a Bayesian machine learning neural network model to generate a neuroimaging phenotype and AD score representing the probability of AD using structural MRI data in the Alzheimer’s Disease Neuroimaging Initiative (ADNI) Cohort (cut-off 0.5, AUC 0.92, PPV 0.90, NPV 0.93). We go on to validate the model in an independent real-world dataset of the National Alzheimer’s Coordinating Centre (AUC 0.74, PPV 0.65, NPV 0.80) and demonstrate the correlation of the AD-score with cognitive scores in those with an AD-score above 0.5. We then apply the model to a healthy population in the UK Biobank study to identify a cohort at risk for Alzheimer’s disease. Results We show that the cohort with a neuroimaging Alzheimer’s phenotype has a cognitive profile in keeping with Alzheimer’s disease, with strong evidence for poorer fluid intelligence, and some evidence of poorer numeric memory, reaction time, working memory, and prospective memory. We found some evidence in the AD-score positive cohort for modifiable risk factors of hypertension and smoking. Conclusions This approach demonstrates the feasibility of using AI methods to identify a potentially prediagnostic population at high risk for developing sporadic Alzheimer’s disease.

BackgroundIt has been well documented that atrophy of hippocampus and hippocampal subfields is closely linked to cognitive decline in normal aging and patients with mild cognitive impairment (MCI) and Alzheimer’s disease (AD). However, evidence is still sparce regarding the atrophy of hippocampus and hippocampal subfields in normal aging adults who later developed MCI or AD.ObjectiveTo examine whether atrophy of hippocampus and hippocampal subfields has occurred in normal aging before a diagnosis of MCI or AD.MethodsWe analyzed structural magnetic resonance imaging (MRI) data of cognitively normal (CN, n = 144), MCI (n = 90), and AD (n = 145) participants obtained from the Alzheimer’s Disease Neuroimaging Initiative. The CN participants were categorized into early dementia converters (CN-C) and non-converters (CN-NC) based on their scores of clinical dementia rating after an average of 36.2 months (range: 6–105 months). We extracted the whole hippocampus and hippocampal subfields for each participant using FreeSurfer, and analyzed the differences in volumes of hippocampus and hippocampal subfields between groups. We then examined the associations between volume of hippocampal subfields and delayed recall scores in each group separately.ResultsHippocampus and most of the hippocampal subfields demonstrated significant atrophy during the progression of AD. The CN-C and CN-NC groups differed in the left hippocampus–amygdala transition area (HATA). Furthermore, the volume of presubiculum was significantly correlated with delayed recall scores in the CN-NC and AD groups, but not in the CN-C and MCI groups.ConclusionHippocampal subfield atrophy (i.e., left HATA) had occurred in cognitively normal elderly individuals before clinical symptoms were recognized. Significant associations of presubiculum with delayed recall scores in the CN-NC and AD groups highlight the essential role of the hippocampal subfields in both early dementia detection and AD progression.

Disruptions in the inter-regional connective correlation within the brain are believed to contribute to memory impairment. To detect these corresponding correlation networks in Alzheimer’s disease (AD), we conducted three types of inter-regional correlation analysis, including structural covariance, functional connectivity and group-level independent component analysis (group-ICA). The analyzed data were obtained from the Alzheimer’s Disease Neuroimaging Initiative, comprising 52 cognitively normal (CN) participants without subjective memory concerns, 52 individuals with late mild cognitive impairment (LMCI) and 52 patients with AD. We firstly performed vertex-wise cortical thickness analysis to identify brain regions with cortical thinning in AD and LMCI patients using structural MRI data. These regions served as seeds to construct both structural covariance networks and functional connectivity networks for each subject. Additionally, group-ICA was performed on the functional data to identify intrinsic brain networks at the cohort level. Through a comparison of the structural covariance and functional connectivity networks with ICA networks, we identified several inter-regional correlation networks that consistently exhibited abnormal connectivity patterns among AD and LMCI patients. Our findings suggest that reduced inter-regional connectivity is predominantly observed within a subnetwork of the default mode network, which includes the posterior cingulate and precuneus regions, in both AD and LMCI patients. This disruption of connectivity between key nodes within the default mode network provides evidence supporting the hypothesis that impairments in brain networks may contribute to memory deficits in AD and LMCI.

Identifying biomarkers for Alzheimer's disease with a goal of early detection is a fundamental problem in clinical research. Both medical imaging and genetics have contributed informative biomarkers in literature. To further improve the performance, recently, there is an increasing interest in developing analytic approaches that combine data across modalities such as imaging and genetics. However, there are limited methods in literature that are able to systematically combine high-dimensional voxel-level imaging and genetic data for accurate prediction of clinical outcomes of interest. Existing prediction models that integrate imaging and genetic features often use region level imaging summaries, and they typically do not consider the spatial configurations of the voxels in the image or incorporate the dependence between genes that may compromise prediction ability. We propose a novel integrative Bayesian scalar-on-image regression model for predicting cognitive outcomes based on high-dimensional spatially distributed voxel-level imaging data, along with correlated transcriptomic features. We account for the spatial dependencies in the imaging voxels via a tensor approach that also enables massive dimension reduction to address the curse of dimensionality, and models the dependencies between the transcriptomic features via a Graph-Laplacian prior. We implement this approach via an efficient Markov chain Monte Carlo (MCMC) computation strategy. We apply the proposed method to the analysis of longitudinal ADNI data for predicting cognitive scores at different visits by integrating voxel-level cortical thickness measurements derived from T1w-MRI scans and transcriptomics data. We illustrate that the proposed imaging transcriptomics approach has significant improvements in prediction compared to prediction using a subset of features from only one modality (imaging or genetics), as well as when using imaging and transcriptomics features but ignoring the inherent dependencies between the features. Our analysis is one of the first to conclusively demonstrate the advantages of prediction based on combining voxel-level cortical thickness measurements along with transcriptomics features, while accounting for inherent structural information.

Abstract Objective To determine if plasma tau phosphorylated at threonine 181 (p‐tau181) distinguishes pathology‐confirmed Alzheimer's disease (AD) from normal cognition (NC) adults, to test if p‐tau181 predicts cognitive and functional decline, and to validate findings in an external cohort. Methods Thirty‐one neuropathology‐confirmed AD cases, participants with clinical diagnoses of mild cognitive impairment (MCI, N = 91) or AD dementia (N = 64), and NC (N = 241) had plasma collected at study entry. The clinical diagnosis groups had annual cognitive (Mini‐Mental State Examination, MMSE) and functional (Clinical Dementia Rating Scale, CDR) measures. NC (N = 70), MCI (N = 75), and AD dementia (N = 50) cases from the Alzheimer's Disease Neuroimaging Initiative (ADNI) were used as a validation cohort. Plasma p‐tau181 was measured using the Quanterix SiMoA HD‐X platform. Results Plasma p‐tau181 differentiated pathology‐confirmed AD from NC with negative amyloid PET scans with an AUC of 0.93. A cut point of 3.44 pg/mL (maximum Youden Index) had a sensitivity of 0.77, specificity of 0.96. p‐Tau181 values above the cut point were associated with the faster rate of decline in MMSE in AD dementia and MCI and a shorter time to a clinically significant functional decline in all groups. In a subset of MCI cases from ADNI, p‐tau181 values above the cut point associated with faster rate of decline in MMSE, and a shorter time to a clinically significant functional decline and conversion to dementia. Interpretation Plasma p‐tau181 differentiates AD pathology cases from NC with high accuracy. Higher levels of plasma p‐tau181 are associated with faster cognitive and functional decline.

Abstract Background In recent years, neuroimaging with deep learning (DL) algorithms have made remarkable advances in the diagnosis of neurodegenerative disorders. However, applying DL in different medical domains is usually challenged by lack of labeled data. To address this challenge, transfer learning (TL) has been applied to use state-of-the-art convolution neural networks pre-trained on natural images. Yet, there are differences in characteristics between medical and natural images, also image classification and targeted medical diagnosis tasks. The purpose of this study is to investigate the performance of specialized and TL in the classification of neurodegenerative disorders using 3D volumes of 18F-FDG-PET brain scans. Results Results show that TL models are suboptimal for classification of neurodegenerative disorders, especially when the objective is to separate more than two disorders. Additionally, specialized CNN model provides better interpretations of predicted diagnosis. Conclusions TL can indeed lead to superior performance on binary classification in timely and data efficient manner, yet for detecting more than a single disorder, TL models do not perform well. Additionally, custom 3D model performs comparably to TL models for binary classification, and interestingly perform better for diagnosis of multiple disorders. The results confirm the superiority of the custom 3D-CNN in providing better explainable model compared to TL adopted ones.

Abstract Background Tau-PET is a prognostic marker for cognitive decline in Alzheimer’s disease, and the heterogeneity of tau-PET patterns matches cognitive symptom heterogeneity. Thus, tau-PET may allow precision-medicine prediction of individual tau-related cognitive trajectories, which can be important for determining patient-specific cognitive endpoints in clinical trials. Here, we aimed to examine whether tau-PET in cognitive-domain-specific brain regions, identified via fMRI meta-analyses, allows the prediction of domain-specific cognitive decline. Further, we aimed to determine whether tau-PET-informed personalized cognitive composites capture patient-specific cognitive trajectories more sensitively than conventional cognitive measures. Methods We included Alzheimer’s Disease Neuroimaging Initiative (ADNI) participants classified as controls (i.e., amyloid-negative, cognitively normal, n = 121) or Alzheimer’s disease-spectrum (i.e., amyloid-positive, cognitively normal to dementia, n = 140), plus 111 AVID-1451-A05 participants for independent validation (controls/Alzheimer’s disease-spectrum=46/65). All participants underwent baseline 18F-flortaucipir tau-PET, amyloid-PET, and longitudinal cognitive testing to assess annual cognitive changes (i.e., episodic memory, language, executive functioning, visuospatial). Cognitive changes were calculated using linear mixed models. Independent meta-analytical task-fMRI activation maps for each included cognitive domain were obtained from the Neurosynth database and applied to tau-PET to determine tau-PET signal in cognitive-domain-specific brain regions. In bootstrapped linear regression, we assessed the strength of the relationship (i.e., partial R 2) between cognitive-domain-specific tau-PET vs. global or temporal-lobe tau-PET and cognitive changes. Further, we used tau-PET-based prediction of domain-specific decline to compose personalized cognitive composites that were tailored to capture patient-specific cognitive decline. Results In both amyloid-positive cohorts (ADNI [age = 75.99±7.69] and A05 [age = 74.03±9.03]), cognitive-domain-specific tau-PET outperformed global and temporal-lobe tau-PET for predicting future cognitive decline in episodic memory, language, executive functioning, and visuospatial abilities. Further, a tau-PET-informed personalized cognitive composite across cognitive domains enhanced the sensitivity to assess cognitive decline in amyloid-positive subjects, yielding lower sample sizes required for detecting simulated intervention effects compared to conventional cognitive endpoints (i.e., memory composite, global cognitive composite). However, the latter effect was less strong in A05 compared to the ADNI cohort. Conclusion Combining tau-PET with task-fMRI-derived maps of major cognitive domains facilitates the prediction of domain-specific cognitive decline. This approach may help to increase the sensitivity to detect Alzheimer’s disease-related cognitive decline and to determine personalized cognitive endpoints in clinical trials.

Abstract Background In Alzheimer’s disease (AD), fibrillar tau initially occurs locally and progresses preferentially between closely connected regions. However, the underlying sources of regional vulnerability to tau pathology remain unclear. Previous brain-autopsy findings suggest that the myelin levels—which differ substantially between white matter tracts in the brain—are a key modulating factor of region-specific susceptibility to tau deposition. Here, we investigated whether myelination differences between fiber tracts of the human connectome are predictive of the interregional spreading of tau pathology in AD. Methods We included two independently recruited samples consisting of amyloid-PET-positive asymptomatic and symptomatic elderly individuals, in whom tau-PET was obtained at baseline (ADNI: n = 275; BioFINDER-1: n = 102) and longitudinally in a subset (ADNI: n = 123, mean FU = 1.53 [0.69–3.95] years; BioFINDER-1: n = 39, mean FU = 1.87 [1.21–2.78] years). We constructed MRI templates of the myelin water fraction (MWF) in 200 gray matter ROIs and connecting fiber tracts obtained from adult cognitively normal participants. Using the same 200 ROI brain-parcellation atlas, we obtained tau-PET ROI values from each individual in ADNI and BioFINDER-1. In a spatial regression analysis, we first tested the association between cortical myelin and group-average tau-PET signal in the amyloid-positive and control groups. Secondly, employing a previously established approach of modeling tau-PET spreading based on functional connectivity between ROIs, we estimated in a linear regression analysis, whether the level of fiber-tract myelin modulates the association between functional connectivity and longitudinal tau-PET spreading (i.e., covariance) between ROIs. Results We found that higher myelinated cortical regions show lower tau-PET uptake (ADNI: rho = − 0.267, p

The Polygenic Risk Score Knowledge Base (PRSKB) is a web-based interface that stores data from >2,300 distinct genome-wide association studies, and can estimate polygenic risk scores for general use.

Abstract Background Alzheimer’s disease affects profoundly the quality of human behavior and cognition. The very broad distribution of its severity across various human subjects requires the quantitative diagnose of Alzheimer’s disease beyond the conventional tripartite classification of cohorts such as cognitively normal (CN), mild cognitive impairment (MCI), Alzheimer’s disease (AD). The unfolding of such broad distributions by the quantitative and continuous degree of AD severity is necessary for the precise diagnose in the cross-sectional study of different stages in AD. Results We conducted the massive reanalysis on MRI images of 761 human brains based on the accumulated bigdata of Alzheimer’s Disease Neuroimaging Initiative. The score matrix of cortical thickness profile at cortex points of subjects was constructed by statistically learning the cortical thickness data of 761 human brains. We also developed a new and simple algebraic predictor which provides the quantitative and continuous degree of AD severity of subjects along the scale from 0 for fully CN to 1 for fully AD state. The mathematical measure of a new predictor for the degree of AD severity is presented based on a covariance correlation matrix of cortical thickness profile between human subjects. One can remove the uncertainty in the determination of different stages in AD by the quantitative degree of AD severity and thus go far beyond the tripartite classification of cohorts. Conclusions We unfold the nature of broad distribution of AD severity of subjects even within a given cohort by the scale from 0 for fully CN to 1 for fully AD state. The quantitative and continuous degree of AD severity developed in this study would be a good practical measure for diagnosing the different stages in AD severity.

IntroductionAlzheimer's disease (AD) is a chronic neurodegenerative disease of the brain that has attracted wide attention in the world. The diagnosis of Alzheimer's disease is faced with the difficulties of insufficient manpower and great difficulty. With the intervention of artificial intelligence, deep learning methods are widely used to assist clinicians in the early recognition of Alzheimer's disease. And a series of methods based on data input with different dimensions have been proposed. However, traditional deep learning models rely on expensive hardware resources and consume a lot of training time, and may fall into the dilemma of local optima.MethodsIn recent years, broad learning system (BLS) has provided researchers with new research ideas. Based on the three-dimensional residual convolution module and BLS, a novel broad-deep ensemble model based on BLS is proposed for the early detection of Alzheimer's disease. The Alzheimer's Disease Neuroimaging Initiative (ADNI) MRI image dataset is used to train the model and then we compare the performance of proposed model with previous work and clinicians' diagnosis.ResultsThe result of experiments demonstrate that the broad-deep ensemble model is superior to previously proposed related works, including 3D-ResNet and VoxCNN, in accuracy, sensitivity, specificity and F1.DiscussionThe proposed broad-deep ensemble model is effective for early detection of Alzheimer's disease. In addition, the proposed model does not need the pre-training process of its depth module, which greatly reduces the training time and hardware dependence.

IntroductionBrain imaging genetics aims to explore the genetic architecture underlying brain structure and functions. Recent studies showed that the incorporation of prior knowledge, such as subject diagnosis information and brain regional correlation, can help identify significantly stronger imaging genetic associations. However, sometimes such information may be incomplete or even unavailable.MethodsIn this study, we explore a new data-driven prior knowledge that captures the subject-level similarity by fusing multi-modal similarity networks. It was incorporated into the sparse canonical correlation analysis (SCCA) model, which is aimed to identify a small set of brain imaging and genetic markers that explain the similarity matrix supported by both modalities. It was applied to amyloid and tau imaging data of the ADNI cohort, respectively.ResultsFused similarity matrix across imaging and genetic data was found to improve the association performance better or similarly well as diagnosis information, and therefore would be a potential substitute prior when the diagnosis information is not available (i.e., studies focused on healthy controls).DiscussionOur result confirmed the value of all types of prior knowledge in improving association identification. In addition, the fused network representing the subject relationship supported by multi-modal data showed consistently the best or equally best performance compared to the diagnosis network and the co-expression network.

Cortical uptake in brain amyloid positron emission tomography (PET) is increasingly used for the biological diagnosis of Alzheimer’s disease (AD); however, the clinical and biological relevance of the striatum beyond the cortex in amyloid PET scans remains unclear. A total of 513 amyloid-positive participants having 18F-AV45 amyloid PET scans available were included in the analysis. The associations between cognitive scores and striatal uptake were analyzed. The participants were categorized into three groups based on the residual from the linear fitting between 18F-AV45 uptake in the putamen and the cortex in the order of HighP > MidP > LowP group. We then examined the differences between these three groups in terms of clinical diagnosis, APOE genotype, CSF phosphorylated tau (ptau) concentration, hippocampal volume, entorhinal thickness, and cognitive decline rate to evaluate the additional insights provided by the putamen beyond the cortex. The 18F-AV45 uptake in the putamen was more strongly associated with ADAS-cog13 and MoCA scores (p < 0.001) compared to the uptake in the caudate nucleus. Despite comparable cortical uptakes, the HighP group had a two-fold higher risk of being ε4-homozygous or diagnosed with AD dementia compared to the LowP group. These three groups had significantly different CSF ptau concentration, hippocampal volume, entorhinal thickness, and cognitive decline rate. These findings suggest that the assessment of 18F-AV45 uptake in the putamen is of unique value for evaluating disease severity and predicting disease progression.

Abstract Polygenic risk scores (PRSs) can boost risk prediction in late-onset Alzheimer’s disease (LOAD) beyond apolipoprotein E (APOE) but have not been leveraged to identify genetic resilience factors. Here, we sought to identify resilience-conferring common genetic variants in (1) unaffected individuals having high PRSs for LOAD, and (2) unaffected APOE-ε4 carriers also having high PRSs for LOAD. We used genome-wide association study (GWAS) to contrast “resilient” unaffected individuals at the highest genetic risk for LOAD with LOAD cases at comparable risk. From GWAS results, we constructed polygenic resilience scores to aggregate the addictive contributions of risk-orthogonal common variants that promote resilience to LOAD. Replication of resilience scores was undertaken in eight independent studies. We successfully replicated two polygenic resilience scores that reduce genetic risk penetrance for LOAD. We also showed that polygenic resilience scores positively correlate with polygenic risk scores in unaffected individuals, perhaps aiding in staving off disease. Our findings align with the hypothesis that a combination of risk-independent common variants mediates resilience to LOAD by moderating genetic disease risk.

Abstract Background Cognitive reserve and resilience are terms used to explain interindividual variability in maintenance of cognitive health in response to adverse factors, such as brain pathology in the context of aging or neurodegenerative disorders. There is substantial interest in identifying tractable substrates of resilience to potentially leverage this phenomenon into intervention strategies. One way of operationalizing cognitive resilience that has gained popularity is the residual method: regressing cognition on an adverse factor and using the residual as a measure of resilience. This method is attractive because it provides a statistical approach that is an intuitive match to the reserve/resilience conceptual framework. However, due to statistical properties of the regression equation, the residual approach has qualities that complicate its interpretation as an index of resilience and make it statistically inappropriate in certain circumstances. Methods and results We describe statistical properties of the regression equation to illustrate why the residual is highly correlated with the cognitive score from which it was derived. Using both simulations and real data, we model common applications of the approach by creating a residual score (global cognition residualized for hippocampal volume) in individuals along the AD spectrum. We demonstrate that in most real-life scenarios, the residual measure of cognitive resilience is highly correlated with cognition, and the degree of this correlation depends on the initial relationship between the adverse factor and cognition. Subsequently, any association between this resilience metric and an external variable may actually be driven by cognition, rather than by an operationalized measure of resilience. We then assess several strategies proposed as potential solutions to this problem, such as including both the residual and original cognitive measure in a model. However, we conclude these solutions may be insufficient, and we instead recommend against “pre-regression” strategies altogether in favor of using statistical moderation (e.g., interactions) to quantify resilience. Conclusions Caution should be taken in the use and interpretation of the residual-based method of cognitive resilience. Rather than identifying resilient individuals, we encourage building more complete models of cognition to better identify the specific adverse and protective factors that influence cognitive decline.

Inglese et al. develop a predictive model that computes multi-regional statistical morpho-functional mesoscopic traits from T1-weighted magnetic resonance imaging scans. Their model reliably discriminates people with Alzheimer’s disease-related pathologies from those without.

ObjectiveTo evaluate the progression of brain glucose metabolism among participants with biological signature of Alzheimer’s disease (AD) and its relevance to cognitive decline.MethodWe studied 602 amyloid positive individuals who underwent 18F-fluorodeoxyglucose PET (FDG-PET) scan, 18F-AV-45 amyloid PET (AV45-PET) scan, structural MRI scan and neuropsychological examination, including 116 cognitively normal (CN) participants, 314 participants diagnosed as mild cognitive impairment (MCI), and 172 participants diagnosed as AD dementia. The first FDG-PET scan satisfying the inclusion criteria was considered as the baseline scan. Cross-sectional analysis were conducted with the baseline FDG-PET data to compare the regional differences between diagnostic groups after adjusting confounding factors. Among these participants, 229 participants (55 CN, 139 MCI, and 35 AD dementia) had two-year follow-up FDG-PET data available. Regional glucose metabolism was computed and the progression rates of regional glucose metabolism were derived from longitudinal FDG-PET scans. Then the group differences of regional progression rates were examined to assess whether glucose metabolism deficit accelerates or becomes stable with disease progression. The association of cognitive decline rate with baseline regional glucose metabolism, and progression rate in longitudinal data, were evaluated.ResultsParticipants with AD dementia showed substantial glucose metabolism deficit than CN and MCI at left hippocampus, in addition to the traditionally reported frontal and parietal–temporal lobe. More substantial metabolic change was observed with the contrast AD – MCI than the contrast MCI – CN, even after adjusting time duration since cognitive symptom onset. With the longitudinal data, glucose metabolism was observed to decline the most rapidly in the AD dementia group and at a slower rate in MCI. Lower regional glucose metabolism was correlated to faster cognitive decline rate with mild–moderate correlations, and the progression rate was correlated to cognitive decline rate with moderate-large correlations.Discussion and conclusionHippocampus was identified to experience hypometabolism in AD pathology. Hypometabolism accelerates with disease progression toward AD dementia. FDG-PET, particularly longitudinal scans, could potentially help predict how fast cognition declines and assess the impact of treatment in interventional trials.

Abstract Microglial activation occurs early in Alzheimer's disease (AD) and previous studies reported both detrimental and protective effects of microglia on AD progression. Here, we used CSF sTREM2 to investigate disease stage‐dependent drivers of microglial activation and to determine downstream consequences on AD progression. We included 402 patients with measures of earliest beta‐amyloid (CSF Aβ1‐42) and late‐stage fibrillary Aβ pathology (amyloid‐PET centiloid), as well as sTREM2, p‐tau181, and FDG‐PET. To determine disease stage, we stratified participants into early Aβ‐accumulators (Aβ CSF+/PET−; n = 70) or late Aβ‐accumulators (Aβ CSF+/PET+; n = 201) plus 131 controls. In early Aβ‐accumulators, higher centiloid was associated with cross‐sectional/longitudinal sTREM2 and p‐tau181 increases. Further, higher sTREM2 mediated the association between centiloid and cross‐sectional/longitudinal p‐tau181 increases and higher sTREM2 was associated with FDG‐PET hypermetabolism. In late Aβ‐accumulators, we found no association between centiloid and sTREM2 but a cross‐sectional association between higher sTREM2, higher p‐tau181 and glucose hypometabolism. Our findings suggest that a TREM2‐related microglial response follows earliest Aβ fibrillization, manifests in inflammatory glucose hypermetabolism and may facilitate subsequent p‐tau181 increases in earliest AD.

Introduction. While cerebrospinal fluid (CSF) core biomarkers have been considered diagnostic biomarkers for a long time, special attention has been recently dedicated to lipoproteins and metabolites that could be potentially associated with Alzheimer’s disease (AD) neurodegeneration. Herein, we aimed to investigate the relationship between the levels of CSF core biomarkers including Aβ-42, TAU, and P-TAU and plasma lipoproteins and metabolites of patients with AD from the baseline cohort of the Alzheimer’s Disease Neuroimaging Initiative (ADNI) database. Method. Using the ADNI database, fourteen subclasses of lipoproteins as well as a number of lipids and fatty acids and low-molecular metabolites including amino acids, ketone bodies, and glycolysis-related metabolites in blood samples were measured as potential noninvasive markers, and their association with the CSF core biomarkers was statistically investigated controlling for age and gender. Results. A total number of 251 AD subjects were included, among whom 71 subjects were negative for the Apo-E ε4 allele and 150 were positive. There was no significant difference between the two groups regarding cognitive assessments, CSF core biomarkers, and lipoproteins and metabolites except the level of Aβ-42 (p

Abstract Objective To investigate the relationship between the topography of amyloid‐β plaques, tau neurofibrillary tangles, and the overlap between the two, with cognitive dysfunction in individuals without dementia. Methods We evaluated 154 individuals who were assessed with amyloid‐β PET with [18F]AZD4694, tau‐PET with [18F]MK6240, structural MRI, and neuropsychological testing. We also evaluated an independent cohort of 240 individuals who were assessed with amyloid‐β PET with [18F]Florbetapir, tau‐PET with [18F]Flortaucipir, structural MRI, and neuropsychological testing. Using the VoxelStats toolbox, we conducted voxel‐wise linear regressions between amyloid‐PET, tau‐PET, and their interaction with cognitive function, correcting for age, sex, and years of education. Results In both cohorts, we observed that tau‐PET standardized uptake value ratio in medial temporal lobes was associated with clinical dementia rating Sum of Boxes (CDR‐SoB) scores independently of local amyloid‐PET uptake (FWE corrected at p

Abstract Alzheimer’s disease (AD) is a complex and heterogeneous disease that can be affected by various genetic factors. Although the cause of AD is not yet known and there is no treatment to cure this disease, its progression can be delayed. AD has recently been recognized as a brain-specific type of diabetes called type 3 diabetes. Several studies have shown that people with type 2 diabetes (T2D) have a higher risk of developing AD. Therefore, it is important to identify subgroups of patients with AD that may be more likely to be associated with T2D. We here describe a new approach to identify the correlation between AD and T2D at the genetic level. Subgroups of AD and T2D were each generated using a non-negative matrix factorization (NMF) approach, which generated clusters containing subsets of genes and samples. In the gene cluster that was generated by conventional gene clustering method from NMF, we selected genes with significant differences in the corresponding sample cluster by Kruskal–Wallis and Dunn-test. Subsequently, we extracted differentially expressed gene (DEG) subgroups, and candidate genes with the same regulation direction can be extracted at the intersection of two disease DEG subgroups. Finally, we identified 241 candidate genes that represent common features related to both AD and T2D, and based on pathway analysis we propose that these genes play a role in the common pathological features of AD and T2D. Moreover, in the prediction of AD using logistic regression analysis with an independent AD dataset, the candidate genes obtained better prediction performance than DEGs. In conclusion, our study revealed a subgroup of patients with AD that are associated with T2D and candidate genes associated between AD and T2D, which can help in providing personalized and suitable treatments.

Abstract Background Considering the limited accessibility of amyloid position emission tomography (PET) in patients with dementia, we proposed a deep learning (DL)-based amyloid PET positivity classification model from PET images with 2-deoxy-2-[fluorine-18]fluoro-D-glucose (2-[18F]FDG). Methods We used 2-[18F]FDG PET datasets from the Alzheimer's Disease Neuroimaging Initiative and Korean Brain Aging Study for the Early diagnosis and prediction of Alzheimer’s disease for model development. Moreover, we used an independent dataset from another hospital. A 2.5-D deep learning architecture was constructed using 291 submodules and three axes images as the input. We conducted the voxel-wise analysis to assess the regions with substantial differences in glucose metabolism between the amyloid PET-positive and PET-negative participants. This facilitated an understanding of the deep model classification. In addition, we compared these regions with the classification probability from the submodules. Results There were 686 out of 1433 (47.9%) and 50 out of 100 (50%) amyloid PET-positive participants in the training and internal validation datasets and the external validation datasets, respectively. With 50 times iterations of model training and validation, the model achieved an AUC of 0.811 (95% confidence interval (CI) of 0.803–0.819) and 0.798 (95% CI, 0.789–0.807) on the internal and external validation datasets, respectively. The area under the curve (AUC) was 0.860 when tested with the model with the highest value (0.864) on the external validation dataset. Moreover, it had 75.0% accuracy, 76.0% sensitivity, 74.0% specificity, and 75.0% F1-score. We found an overlap between the regions within the default mode network, thus generating high classification values. Conclusion The proposed model based on the 2-[18F]FDG PET imaging data and a DL framework might successfully classify amyloid PET positivity in clinical practice, without performing amyloid PET, which have limited accessibility.

Abstract Background We aimed to investigate the tau biomarker discrepancies of Alzheimer’s disease (AD) using plasma tau phosphorylated at threonine 181 (p-tau181), cerebrospinal fluid (CSF) p-tau181, and AV1451 positron emission tomography (PET). Methods In the Alzheimer’s Disease Neuroimaging Initiative, 724 non-demented participants were categorized into plasma/CSF and plasma/PET groups. Demographic and clinical variables, amyloid-β (Aβ) burden, flortaucipir-PET binding in Braak regions of interest (ROIs), longitudinal changes in clinical outcomes, and conversion risk were compared. Results Across different tau biomarker groups, the proportion of participants with a discordant profile varied (plasma+/CSF− 15.6%, plasma−/CSF+ 15.3%, plasma+/PET− 22.4%, and plasma−/PET+ 6.1%). Within the plasma/CSF categories, we found an increase from concordant-negative to discordant to concordant-positive in the frequency of Aβ pathology or cognitive impairment, rates of cognitive decline, and risk of cognitive conversion. However, the two discordant categories (plasma+/CSF− and plasma−/CSF+) showed comparable performances, resulting in similarly reduced cognitive capacities. Regarding plasma/PET categories, as expected, PET-positive individuals had increased Aβ burden, elevated flortaucipir retention in Braak ROIs, and accelerated cognitive deterioration than concordant-negative persons. Noteworthy, discordant participants with normal PET exhibited reduced flortaucipir uptake in Braak stage ROIs and slower rates of cognitive decline, relative to those PET-positive. Therefore, individuals with PET abnormality appeared to have advanced tau pathological changes and poorer cognitive function, regardless of the plasma status. Furthermore, these results were found only in individuals with Aβ pathology. Conclusions Our results indicate that plasma and CSF p-tau181 abnormalities associated with amyloidosis occur simultaneously in the progression of AD pathogenesis and related cognitive decline, before tau-PET turns positive.

BackgroundWhite matter hyperintensities (WMHs) and regional brain lobe atrophy coexist in the brain of patients with Alzheimer's disease (AD), but the association between them in patients with AD still lacks comprehensive investigation and solid imaging data support.ObjectiveWe explored whether WMHs can promote the pathological process of AD by aggravating atrophy in specific brain regions and tried to explain the regional specificity of these relationships.MethodsA sample of 240 adults including 180 normal controls (NCs) and 80 cases with AD were drawn from the ADNI database. T1-weighted magnetic resonance imaging (MRI) and T2-weighted fluid-attenuated MRI of the participants were downloaded and were analyzed using AccuBrain® to generate the quantitative ratio of WMHs (WMHr, WMH volumes corrected by intracranial volume) and regional brain atrophy. We also divided WMHr into periventricular WMHr (PVWMHr) and deep WMHr (DWMHr) for the purpose of this study. The Cholinergic Pathways Hyperintensities Scale (CHIPS) scores were conducted by two evaluators. Independent t-test, Mann–Whitney U test, or χ2 test were used to compare the demographic characteristics, and Spearman correlation coefficient values were used to determine the association between WMHs and different regions of brain atrophy.ResultsPositive association between WMHr and quantitative medial temporal lobe atrophy (QMTA) (rs = 0.281, p = 0.011), temporal lobe atrophy (rs = 0.285, p = 0.011), and insular atrophy (rs = 0.406, p < 0.001) was found in the AD group before Bonferroni correction. PVWMHr contributed to these correlations. By separately analyzing the relationship between PVWMHr and brain atrophy, we found that there were still positive correlations after correction in QMTA (rs = 0.325, p = 0.003), temporal lobe atrophy (rs = 0.298, p = 0.007), and insular atrophy (rs = 0.429, p < 0.001) in AD group.ConclusionWMH severity tends to be associated with regional brain atrophy in patients with AD, especially with medial temporal lobe, temporal lobe, and insular lobe atrophy. PVWMHs were devoted to these correlations.

Abstract Background People with neurodegenerative disorders show diverse clinical syndromes, genetic heterogeneity, and distinct brain pathological changes, but studies report overlap between these features. DNA methylation (DNAm) provides a way to explore this overlap and heterogeneity as it is determined by the combined effects of genetic variation and the environment. In this study, we aim to identify shared blood DNAm differences between controls and people with Alzheimer’s disease, amyotrophic lateral sclerosis, and Parkinson’s disease. Results We use a mixed-linear model method (MOMENT) that accounts for the effect of (un)known confounders, to test for the association of each DNAm site with each disorder. While only three probes are found to be genome-wide significant in each MOMENT association analysis of amyotrophic lateral sclerosis and Parkinson’s disease (and none with Alzheimer’s disease), a fixed-effects meta-analysis of the three disorders results in 12 genome-wide significant differentially methylated positions. Predicted immune cell-type proportions are disrupted across all neurodegenerative disorders. Protein inflammatory markers are correlated with profile sum-scores derived from disease-associated immune cell-type proportions in a healthy aging cohort. In contrast, they are not correlated with MOMENT DNAm-derived profile sum-scores, calculated using effect sizes of the 12 differentially methylated positions as weights. Conclusions We identify shared differentially methylated positions in whole blood between neurodegenerative disorders that point to shared pathogenic mechanisms. These shared differentially methylated positions may reflect causes or consequences of disease, but they are unlikely to reflect cell-type proportion differences.

Advances in applied machine learning techniques for neuroimaging have encouraged scientists to implement models to diagnose brain disorders such as Alzheimer’s disease at early stages. Predicting the exact stage of Alzheimer’s disease is challenging; however, complex deep learning techniques can precisely manage this. While successful, these complex architectures are difficult to interrogate and computationally expensive. Therefore, using novel, simpler architectures with more efficient pattern extraction capabilities, such as transformers, is of interest to neuroscientists. This study introduced an optimized vision transformer architecture to predict the group membership by separating healthy adults, mild cognitive impairment, and Alzheimer’s brains within the same age group (>75 years) using resting-state functional (rs-fMRI) and structural magnetic resonance imaging (sMRI) data aggressively preprocessed by our pipeline. Our optimized architecture, known as OViTAD is currently the sole vision transformer-based end-to-end pipeline and outperformed the existing transformer models and most state-of-the-art solutions. Our model achieved F1-scores of 97%±0.0 and 99.55%±0.39 from the testing sets for the rs-fMRI and sMRI modalities in the triple-class prediction experiments. Furthermore, our model reached these performances using 30% fewer parameters than a vanilla transformer. Furthermore, the model was robust and repeatable, producing similar estimates across three runs with random data splits (we reported the averaged evaluation metrics). Finally, to challenge the model, we observed how it handled increasing noise levels by inserting varying numbers of healthy brains into the two dementia groups. Our findings suggest that optimized vision transformers are a promising and exciting new approach for neuroimaging applications, especially for Alzheimer’s disease prediction.

Abstract Background Recent studies suggest that plasma phosphorylated tau181 (p-tau181) is a highly specific biomarker for Alzheimer’s disease (AD)-related tau pathology. It has great potential for the diagnostic and prognostic evaluation of AD, since it identifies AD with the same accuracy as tau PET and CSF p-tau181 and predicts the development of AD dementia in cognitively unimpaired (CU) individuals and in those with mild cognitive impairment (MCI). Plasma p-tau181 may also be used as a biomarker in studies exploring disease pathogenesis, such as genetic or environmental risk factors for AD-type tau pathology. The aim of the present study was to investigate the relation between polygenic risk scores (PRSs) for AD and plasma p-tau181. Methods Data from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) was used to examine the relation between AD PRSs, constructed based on findings in recent genome-wide association studies, and plasma p-tau181, using linear regression models. Analyses were performed in the total sample (n = 818), after stratification on diagnostic status (CU (n = 236), MCI (n = 434), AD dementia (n = 148)), and after stratification on Aβ pathology status (Aβ positives (n = 322), Aβ negatives (n = 409)). Results Associations between plasma p-tau181 and APOE PRSs (p = 3e−18–7e−15) and non-APOE PRSs (p = 3e−4–0.03) were seen in the total sample. The APOE PRSs were associated with plasma p-tau181 in all diagnostic groups (CU, MCI, and AD dementia), while the non-APOE PRSs were associated only in the MCI group. The APOE PRSs showed similar results in amyloid-β (Aβ)-positive and negative individuals (p = 5e−5–1e−3), while the non-APOE PRSs were associated with plasma p-tau181 in Aβ positives only (p = 0.02). Conclusions Polygenic risk for AD including APOE was found to associate with plasma p-tau181 independent of diagnostic and Aβ pathology status, while polygenic risk for AD beyond APOE was associated with plasma p-tau181 only in MCI and Aβ-positive individuals. These results extend the knowledge about the relation between genetic risk for AD and p-tau181, and further support the usefulness of plasma p-tau181 as a biomarker of AD.

Abstract Cortical thinning occurs throughout the entire life and extends to late-life neurodegeneration, yet the neurobiological substrates are poorly understood. Here, we used a virtual-histology technique and gene expression data from the Allen Human Brain Atlas to compare the regional profiles of longitudinal cortical thinning through life (4004 magnetic resonance images [MRIs]) with those of gene expression for several neuronal and non-neuronal cell types. The results were replicated in three independent datasets. We found that inter-regional profiles of cortical thinning related to expression profiles for marker genes of CA1 pyramidal cells, astrocytes and, microglia during development and in aging. During the two stages of life, the relationships went in opposite directions: greater gene expression related to less thinning in development and vice versa in aging. The association between cortical thinning and cell-specific gene expression was also present in mild cognitive impairment and Alzheimer’s Disease. These findings suggest a role of astrocytes and microglia in promoting and supporting neuronal growth and dendritic structures through life that affects cortical thickness during development, aging, and neurodegeneration. Overall, the findings contribute to our understanding of the neurobiology underlying variations in MRI-derived estimates of cortical thinning through life and late-life disease.

Background and ObjectiveEarly identification is important for timely Alzheimer’s disease (AD) treatment. Apolipoprotein E ε4 allele (APOE-ε4) is an important genetic risk factor for sporadic AD. The AD-Resemblance Atrophy Index (RAI)—a structural magnetic resonance imaging-derived composite index—was found to predict the risk of progression from mild cognitive impairment (MCI) to AD. Therefore, we investigated whether the AD-RAI can predict cognitive decline and progression to AD in patients with MCI carrying APOE ε4.MethodsWe included 733 participants with MCI from the Alzheimer’s Disease Neuroimaging Initiative Database (ADNI). Their APOE genotypes, cognitive performance, and levels of AD-RAI were assessed at baseline and follow-up. Linear regression models were used to test the correlations between the AD-RAI and baseline cognitive measures, and linear mixed models with random intercepts and slopes were applied to investigate whether AD-RAI and APOE-ε4 can predict the level of cognitive decline. Cox proportional risk regression models were used to test the association of AD-RAI and APOE status with the progression from MCI to AD.ResultsThe baseline AD-RAI was higher in the MCI converted to AD group than in the MCI stable group (P < 0.001). The AD-RAI was significantly correlated with cognition, and had a synergistic effect with APOE-ε4 to predict the rate of cognitive decline. The AD-RAI predicted the risk and timing of MCI progression to AD. Based on the MCI population carrying APOE-ε4, the median time to progression from MCI to AD was 24 months if the AD-RAI > 0.5, while the median time to progression from MCI to AD was 96 months for patients with an AD-RAI ≤ 0.5.ConclusionThe AD-RAI can predict the risk of progression to AD in people with MCI carrying APOE ε4, is strongly correlated with cognition, and can predict cognitive decline.

PurposeTo assess the pathological aging effect on caudate functional connectivity among mild cognitive impairment (MCI) participants and examine whether and how sex and amyloid contribute to this process.Materials and MethodsTwo hundred and seventy-seven functional magnetic resonance imaging (fMRI) sessions from 163 cognitive normal (CN) older adults and 309 sessions from 139 participants with MCI were included as the main sample in our analysis. Pearson's correlation was used to characterize the functional connectivity (FC) between caudate nuclei and each brain region, then caudate nodal strength was computed to quantify the overall caudate FC strength. Association analysis between caudate nodal strength and age was carried out in MCI and CN separately using linear mixed effect (LME) model with covariates (education, handedness, sex, Apolipoprotein E4, and intra-subject effect). Analysis of covariance was conducted to investigate sex, amyloid status, and their interaction effects on aging with the fMRI data subset having amyloid status available. LME model was applied to women and men separately within MCI group to evaluate aging effects on caudate nodal strength and each region's connectivity with caudate nuclei. We then evaluated the roles of sex and amyloid status in the associations of neuropsychological scores with age or caudate nodal strength. An independent cohort was used to validate the sex-dependent aging effects in MCI.ResultsThe MCI group had significantly stronger age-related increase of caudate nodal strength compared to the CN group. Analyzing women and men separately revealed that the aging effect on caudate nodal strength among MCI participants was significant only for women (left: P = 6.23 × 10−7, right: P = 3.37 × 10−8), but not for men (P > 0.3 for bilateral caudate nuclei). The aging effects on caudate nodal strength were not significantly mediated by brain amyloid burden. Caudate connectivity with ventral prefrontal cortex substantially contributed to the aging effect on caudate nodal strength in women with MCI. Higher caudate nodal strength is significantly related to worse cognitive performance in women but not in men with MCI.ConclusionSex modulates the pathological aging effects on caudate nodal strength in MCI regardless of amyloid status. Caudate nodal strength may be a sensitive biomarker of pathological aging in women with MCI.

Abstract Background FDG-PET hypermetabolism can be observed in mild cognitive impairment (MCI), but the link to primary pathologies of Alzheimer’s diseases (AD) including amyloid and tau is unclear. Methods Using voxel-based regression, we assessed local interactions between amyloid- and tau-PET on spatially matched FDG-PET in 72 MCI patients. Control groups included cerebrospinal fluid biomarker characterized cognitively normal (CN, n = 70) and AD dementia subjects (n = 95). Results In MCI, significant amyloid-PET by tau-PET interactions were found in frontal, lateral temporal, and posterior parietal regions, where higher local tau-PET was associated with higher spatially corresponding FDG-PET at low levels of local amyloid-PET. FDG-PET in brain regions with a significant local amyloid- by tau-PET interaction was higher compared to that in CN and AD dementia and associated with lower episodic memory. Conclusion Higher tau-PET in the presence of low amyloid-PET is associated with abnormally increased glucose metabolism that is accompanied by episodic memory impairment.

Abstract Objective To investigate the statistical power of plasma, imaging, and cognition biomarkers as Alzheimer’s disease (AD) clinical trial outcome measures. Methods Plasma neurofilament light, structural magnetic resonance imaging, and cognition were measured longitudinally in the Alzheimer’s Disease Neuroimaging Initiative (ADNI) in control (amyloid PET or CSF Aβ42 negative [Aβ‐] with Clinical Dementia Rating scale [CDR] = 0; n = 330), preclinical AD (Aβ + with CDR = 0; n = 218) and mild AD (Aβ + with CDR = 0.5‐1; n = 697) individuals. A statistical power analysis was performed across biomarkers and groups based on longitudinal mixed effects modeling and using several different clinical trial designs. Results For a 30‐month trial of preclinical AD, both the temporal composite and hippocampal volumes were superior to plasma neurofilament light and cognition. For an 18‐month trial of mild AD, hippocampal volume was superior to all other biomarkers. Plasma neurofilament light became more effective with increased trial duration or sampling frequency. Imaging biomarkers were characterized by high slope and low within‐subject variability, while plasma neurofilament light and cognition were characterized by higher within‐subject variability. Interpretation MRI measures had properties that made them preferable to cognition and pNFL as outcome measures in clinical trials of early AD, regardless of cognitive status. However, pNfL and cognition can still be effective depending on inclusion criteria, sampling frequency, and response to therapy. Future trials will help to understand how sensitive pNfL and MRI are to detect downstream effects on neurodegeneration of drugs targeting amyloid and tau pathology in AD.

Abstract Objective To determine the inflammatory analytes that predict clinical progression and evaluate their performance against biomarkers of neurodegeneration. Methods A longitudinal study of MCI‐AD patients in a Discovery cohort over 15 months, with replication in the Alzheimer’s Disease Neuroimaging Initiative (ADNI) MCI cohort over 36 months. Fifty‐three inflammatory analytes were measured in the CSF and plasma with a RBM multiplex analyte platform. Inflammatory analytes that predict clinical progression on Clinical Dementia Rating Scale‐Sum of Boxes (CDR‐SB) and Mini Mental State Exam scores were assessed in multivariate regression models. To provide context, key analyte results in ADNI were compared against biomarkers of neurodegeneration, hippocampal volume, and CSF neurofilament light (NfL), in receiver operating characteristic (ROC) analyses evaluating highest quartile of CDR‐SB change over two years (≥3 points). Results Cerebrospinal fluid inflammatory analytes in relation to cognitive decline were best described by gene ontology terms, natural killer cell chemotaxis, and endothelial cell apoptotic process and in plasma, extracellular matrix organization, blood coagulation, and fibrin clot formation described the analytes. CSF CCL2 was most robust in predicting rate of cognitive change and analytes that correlated to CCL2 suggest IL‐10 pathway dysregulation. The ROC curves for ≥3 points change in CDR‐SB over 2 years when comparing baseline hippocampal volume, CSF NfL, and CCL2 were not significantly different. Interpretation Baseline levels of immune cell chemotactic cytokine CCL2 in the CSF and IL‐10 pathway dysregulation impact longitudinal cognitive and functional decline in MCI‐AD. CCL2’s utility appears comparable to biomarkers of neurodegeneration in predicting rapid decline.

Abstract Background Quantifying changes in the levels of biological and cognitive markers prior to the clinical presentation of Alzheimer’s disease (AD) will provide a template for understanding the underlying aetiology of the clinical syndrome and, concomitantly, for improving early diagnosis, clinical trial recruitment and treatment assessment. This study aims to characterise continuous changes of such markers and determine their rate of change and temporal order throughout the AD continuum. Methods The methodology is founded on the development of stochastic models to estimate the expected time to reach different clinical disease states, for different risk groups, and synchronise short-term individual biomarker data onto a disease progression timeline. Twenty-seven markers are considered, including a range of cognitive scores, cerebrospinal (CSF) and plasma fluid proteins, and brain structural and molecular imaging measures. Data from 2014 participants in the Alzheimer’s Disease Neuroimaging Initiative database is utilised. Results The model suggests that detectable memory dysfunction could occur up to three decades prior to the onset of dementia due to AD (ADem). This is closely followed by changes in amyloid-β CSF levels and the first cognitive decline, as assessed by sensitive measures. Hippocampal atrophy could be observed as early as the initial amyloid-β accumulation. Brain hypometabolism starts later, about 14 years before onset, along with changes in the levels of total and phosphorylated tau proteins. Loss of functional abilities occurs rapidly around ADem onset. Neurofilament light is the only protein with notable early changes in plasma levels. The rate of change varies, with CSF, memory, amyloid PET and brain structural measures exhibiting the highest rate before the onset of ADem, followed by a decline. The probability of progressing to a more severe clinical state increases almost exponentially with age. In accordance with previous studies, the presence of apolipoprotein E4 alleles and amyloid-β accumulation can be associated with an increased risk of developing the disease, but their influence depends on age and clinical state. Conclusions Despite the limited longitudinal data at the individual level and the high variability observed in such data, the study elucidates the link between the long asynchronous pathophysiological processes and the preclinical and clinical stages of AD.

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.