BACKGROUND:To have maximal benefit, Alzheimer's disease-modifying treatments might need to be started before the onset of clinical symptoms. Mutations of the PSEN1 gene are inherited as fully penetrant, autosomal-dominant traits, which almost always result in the clinical onset of Alzheimer's disease before the age of 65 years. We aimed to evaluate the efficacy, including possible delayed emergence of cognitive impairment, and safety of crenezumab, an anti-amyloid monoclonal antibody, in cognitively unimpaired carriers of the PSEN1Glu280Ala mutation at high imminent risk of developing symptoms due to Alzheimer's disease. METHODS:This 5-8-year common-close, double-blind, placebo-controlled, single-centre trial screened kindred members aged 30-60 years from the main health-care site in Medellín, Colombia. Participants who were cognitively unimpaired and carried the PSEN1Glu280Ala autosomal-dominant mutation were randomly assigned 1:1 to receive placebo or subcutaneous crenezumab (investigators and participants were masked to treatment allocation), with an initial 300 mg dose every 2 weeks that increased to 720 mg every 2 weeks, and a later optional increase to 60 mg/kg intravenously every 4 weeks. Randomisation was stratified by age, education, APOE ɛ4 carrier status, and baseline Clinical Dementia Rating. Mutation non-carriers received placebo and were included in a 1:2 ratio of non-carriers to carriers to maintain genotype masking and include a genetic kindred control. Dual primary outcomes were the annualised rates of change in the Alzheimer's Prevention Initiative (API) preclinical autosomal-dominant Alzheimer's disease (ADAD) composite test total score and Free and Cued Selective Reminding Test-Cueing Index (FCSRT-CI) assessed in randomised participants who received at least one dose of the study drug, according to treatment assignment. Primary endpoints were assessed with a random coefficient regression model with a missing-at-random assumption adjusting for randomisation factors. Safety endpoints for mutation carriers were assessed in randomised participants who received at least one dose of the study drug. This trial is registered with ClinicalTrials.gov (NCT01998841) and is completed. FINDINGS:619 Colombian API registrants were prescreened, 315 were assessed for eligibility, and 252 were enrolled (crenezumab-carrier, n=85; placebo-carrier, n=84; placebo-non-carrier, n=83; 160 [63%] women and 92 [37%] men) between Dec 20, 2013, and Feb 27, 2017. 237 (94%) completed the trial, with final data collection on March 22, 2022. The annualised rate of change in the API ADAD composite was -1·10 (SE 0·29) in the crenezumab group and -1·43 (0·29) in the placebo group (between-group difference 0·33 [95% CI -0·48 to 1·13]; p=0·43). The annualised rate of change in FCSRT-CI was -0·03 (0·00) in the crenezumab group and -0·04 (0·00) in the placebo group (between-group difference 0·01 [0·00 to 0·02]; p=0·16). All participants had at least one adverse event; serious adverse events occurred in 23 (27%) of 84 in the crenezumab group and 21 (25%) of 84 in the placebo group. No fatalities occurred. INTERPRETATION:Crenezumab therapy administered for 5-8 years did not result in significant benefits on our primary clinical outcomes in cognitively unimpaired participants predisposed to developing ADAD dementia; secondary and exploratory outcomes also showed no significant effect on removal of amyloid plaques or other clinical or biomarker outcomes. Together with the results of other anti-amyloid β trials, robust fibrillar amyloid removal appears necessary for clinical efficacy in people with elevated brain amyloid. This study will further inform the biomarker, cognitive, and clinical trajectory of preclinical ADAD, the risk of clinical progression in amyloid-positive and amyloid-negative mutation carriers, and the size and design of future secondary and primary prevention trials. FUNDING:US National Institute on Aging (NIA), Banner Alzheimer's Institute, Genentech, F Hoffmann-La Roche.

INTRODUCTION:Chronic traumatic encephalopathy (CTE) is a tauopathy linked to repetitive head impacts. Factors influencing brain regional susceptibility to tau deposition and spreading remain unclear. METHODS:We used three datasets: [18F]flortaucipir positron emission tomography (PET) in 157 former professional American football players and 53 controls (DIAGNOSE CTE); cortical myelin water fractions (MWF) in 50 healthy individuals (Myelin Water Atlas); and white matter (WM) tract MWF and functional connectivity (FC) in 100 healthy individuals (Human Connectome Project). We tested associations between tau-PET uptake and covariance in football players and typical cortical gray matter (GM) MWF, WM tract MWF, and FC. RESULTS:Cortical regions with lower typical GM MWF showed higher tau-PET uptake (β = -0.399, p = 0.001). WM tracts with lower typical MWF were associated with higher tau-PET covariance (β = -0.238, p < 0.001). Higher typical FC was associated with higher tau-PET covariance (β = 0.447, p < 0.001). DISCUSSION:In former football players at risk for CTE, regional susceptibility to tau deposition may be driven by low myelin and high FC.

INTRODUCTION:Diffusion-weighted imaging derived mean diffusivity (MD) correlates with Alzheimer's disease (AD) biomarkers, yet its neuropathological correlates remain unclear. METHODS:Diffusion-weighted imaging, post mortem neuropathology, and cognitive performance data were obtained from the National Alzheimer's Coordinating Center (N = 97), Alzheimer's Disease Neuroimaging Initiative (N = 21), and Arizona Study of Aging and Neurodegenerative Disorders (N = 15). We examined MD associations with neuropathology, cognitive decline, and expression profiles of AD-implicated genes. RESULTS:Results revealed two latent variables-one linked to amyloid/tau, the other to vascular pathology-explaining between 70% and 16% of MD-pathology covariance, respectively. Higher MD correlated with worse cognitive performance, both cross-sectionally and up to 16 years prior to death. MD was regionally associated with Thal phase, neuritic plaque density, Braak stage (temporal/limbic), and infarcts (thalamus), and reflected gene expression patterns related to AD. DISCUSSION:In vivo MD captures distinct AD-related pathologies across brain regions and relates to cognitive trajectories and gene expression. HIGHLIGHTS:Diffusion-weighted imaging (DWI) can detect early gray matter microstructural differences in the Alzheimer's disease (AD) continuum. Mean diffusivity (MD) is associated with tau, amyloid and vascular neuropathologies. Thal amyloid phase and Braak tau score correlate with MD in temporal and limbic regions. Multivariate MD scores differentially relate to proteinopathies vs. vascular damage. MD can serve as a non-invasive biomarker to predict post mortem AD neuropathology.

Objectives To determine whether the rare APOE3 Christchurch (APOE3 Ch ) variant, known to delay clinical onset in PSEN1 E280A -associated autosomal-dominant Alzheimer's disease (ADAD), also attenuates age-related changes in plasma biomarkers of neurodegeneration and neuroinflammation. Methods In this cross-sectional study, plasma Aβ42/40, phosphorylated tau 181 (p-tau181), glial fibrillary acidic protein (GFAP), and neurofilament light (NfL) were quantified in 134 members of the Colombian PSEN1 E280A kindred, including eight heterozygous APOE3 Ch -PSEN1 E280A carriers and 14 APOE3 Ch -only carriers. Results APOE3 Ch heterozygosity was associated with significantly attenuated age-related increases in GFAP (f²=0.247), p-tau181 (f²=0.133), and NfL (f²=0.133) among PSEN1 E280A carriers. No significant effect was observed for Aβ42/40, nor among non-carriers. Elevated GFAP and NfL concentrations correlated with poorer memory performance in mutation carriers. Discussion APOE3 Ch -associated resilience is measurably reflected in plasma biomarker trajectories, indicating attenuation of both neuroinflammatory and neurodegenerative processes. Plasma GFAP, p-tau181, and NfL emerge as candidate pharmacodynamic markers for clinical trials targeting APOE3 Ch -mediated mechanisms. These findings open a new avenue for monitoring resilience-based therapeutic strategies in ADAD.

Abstract INTRODUCTION Positron emission tomography (PET) without usable or accompanying magnetic resonance imaging (MRI) is typically excluded in quantitative analyses of Alzheimer's disease, potentially limiting study generalizability. We investigated participant features predicting data exclusion in magnetic resonance (MR)‐dependent analyses and evaluated an existing MR‐free PET pipeline to quantify these missing data. METHODS Imaging, clinical, cognitive, and sociodemographic data were analyzed for 2119 individuals in a multi‐site cohort. Agreement between MR‐dependent and MR‐free Centiloids (CL) assessed using intra‐class correlations and features predicting data exclusion were examined using logistic regressions. RESULTS MR‐free and MR‐dependent CLs generally agreed, but MR‐free CLs underestimated MR‐dependent cross‐sectionally and longitudinally. Approximately 19.5% (n = 405) of our cohort would have been excluded in MR‐dependent analyses. Age and cerebrovascular comorbidities were consistent exclusion features across multiple sites. DISCUSSION Data exclusion in imaging studies is not entirely random. Flexible quantification methods like MR‐free PET could supplement traditional methods to improve generalizability in large, multi‐site studies.

Glucose hypometabolism is observed in early Alzheimer's disease. However, there are regional discrepancies in hypometabolism and Alzheimer's pathological markers. We examined the local and global contributions of amyloid-β and tau pathology to glucose metabolism and their interplay in memory decline in Presenilin-1 E280A mutation carriers and non-carriers from the largest autosomal-dominant Alzheimer's disease kindred. This cross-sectional study included 43 mutation carriers (6 cognitively impaired) and 39 non-carriers from the Colombia-Boston Biomarker Study. Glucose metabolism was assessed with [18F]fluorodeoxyglucose PET, and memory performance with the Consortium to Establish a Registry for Alzheimer's Disease word list learning. A subgroup of 22 carriers and 26 non-carriers additionally had measures of amyloid-β and tau using 11C-Pittsburgh compound B and 18F-flortaucipir PET, respectively. First, we compared regional glucose metabolism between groups using the Wilcoxon rank-sum test. Then, we studied regional glucose metabolism associations with age, co-localized amyloid-β and tau pathology, and memory using Spearman correlation. Local specificity was assessed by partial correlations controlling for global amyloid-β and tau burden. Finally, we studied whether the link between Alzheimer's pathology and memory was mediated by regional glucose hypometabolism. Mutation carriers exhibited lower glucose metabolism in the precuneus and isthmus cingulate compared to non-carriers. Hypometabolism correlated locally with greater tau accumulation in the medial temporal lobe, inferior temporal gyrus and prefrontal cortex, and with greater amyloid-β accumulation in the inferior temporal gyrus in carriers. These associations were no longer significant when controlled for global pathology, except for the frontal tau-hypometabolism correlation, which was independent of global tau burden, suggesting local specificity. Additionally, lower memory performance in carriers was associated with hypometabolism in regions typically affected by tau. The mediation analysis revealed a region-specific interplay in pathology, with the associations of amyloid-β and tau pathology with memory decline being mediated by hypometabolism in the inferior temporal. Our findings highlight the metabolic vulnerability of the precuneus in early stages, supporting a common pathophysiology between autosomal-dominant and sporadic Alzheimer's disease. The lack of local correlations between amyloid-β, tau and hypometabolism suggests that distant effects may explain the regional discrepancies between pathology accumulation and metabolic alterations. This study describes a model where pathology advances and interacts in a region-specific manner to impact clinical outcomes, underscoring the importance of regional [18F]fluorodeoxyglucose PET as an independent predictor of cognitive decline. Overall, our findings improve understanding of the spatial progression of pathology, which could have important implications in disease management.

INTRODUCTION:Plasma phosphorylated tau (p-tau), particularly p-tau217, is a highly specific biomarker of Alzheimer's disease (AD) pathology. However, plasma p-tau217 can be elevated in rare non-AD conditions. Brain-derived (BD) p-tau217 may reduce these off-target effects, but its performance against neuropathology has not been evaluated. METHODS:We compared p-tau217, BD p-tau217, their amyloid beta 42 (Aβ42) ratios, BD p-tau217/p-tau217, and BD p-tau217/BD tau in end-of-life plasma from 288 neuropathologically characterized participants using a fully automated immunoassay. Biomarkers were assessed against National Institute on Aging-Alzheimer's Association (NIA-AA) classification, Thal phase, Braak stage, cognitive decline, and tau-PET (positron emission tomography). RESULTS:All markers tracked neuropathological severity, with BD p-tau217 having larger fold-changes than p-tau217 but BD p-tau217/Aβ42 enhancing this further. BD p-tau217/BD tau achieved the highest area under the curve (AUC) for distinguishing Intermediate/High from Not/Low AD neuropathological change (ADNC) (0.89 vs 0.82 for p-tau217). Although BD p-tau217/p-tau217 showed smaller fold-changes, it had the strongest association with continuous tangle burden in AD (R2 = 0.68) and best predicted Clinical Dementia Rating Sum of Boxes (CDR-SB decline) (R2 = 0.26). DISCUSSION:BD p-tau217 and BD-based ratios enhance dynamic range and prognostic performance while maintaining diagnostic accuracy, supporting further clinical evaluation.

Accurate quantification of tau pathology via tau positron emission tomography (PET) scan is crucial for diagnosing and monitoring Alzheimer's disease (AD). However, the high cost and limited availability of tau PET restrict its widespread use. In contrast, structural magnetic resonance imaging (MRI) and plasma-based biomarkers provide non-invasive and widely available complementary information related to brain anatomy and disease progression. In this work, we propose a text-guided 3D diffusion model for 3D tau PET image synthesis, leveraging multimodal conditions from both structural MRI and plasma measurement. Specifically, the textual prompt is from the plasma p-tau217 measurement, which is a key indicator of AD progression, while MRI provides anatomical structure constraints. The proposed framework is trained and evaluated using clinical AV1451 tau PET data from the Alzheimer's Disease Neuroimaging Initiative (ADNI) database. Experimental results demonstrate that our approach can generate realistic, clinically meaningful 3D tau PET across a range of disease stages. The proposed framework can help perform tau PET data augmentation under different settings, provide a non-invasive, cost-effective alternative for visualizing tau pathology, and support the simulation of disease progression under varying plasma biomarker levels and cognitive conditions.

Background:Plasma brain-derived pTau217 (BD-pTau217) may provide a Alzheimer's disease-specific plasma tau measure than total pTau217, but its prognostic value is unclear. We compared BD-pTau217 and total plasma pTau217 for predicting clinical and amyloid PET progression in cognitively unimpaired (CU) ADNI participants. Methods:Plasma NULISAseq biomarkers were measured in 1,427 ADNI participants, including 529 CU individuals. Amyloid PET progression was assessed in baseline CU amyloid-negative participants (Centiloid ≥24.1) with longitudinal PET imaging; clinical progression was assessed in all baseline CU participants. Associations were evaluated using Cox models and time-dependent AUC. Results:BD-pTau217 did not clearly outperform total pTau217 for predicting progression to mild cognitive impairment or dementia. However, among baseline amyloid-negative participants (N=175), BD-pTau217 better predicted amyloid PET positivity at 2.5 years (tdAUC 0.82 vs 0.69; HR=10.54, p=0.00015) and 4 years (tdAUC 0.77 vs 0.64; HR=7.03, p=0.00055). Conclusion:BD-pTau217 improved prediction of near-term amyloid PET progression, with less clear advantage for clinical progression.

INTRODUCTION:The basal forebrain (BF), a key cholinergic hub, undergoes atrophy in Alzheimer's disease (AD), contributing to cognitive decline. However, its age-related differences and early vulnerability in autosomal dominant AD (ADAD) remain unclear. METHODS:We studied 158 individuals from the Colombian Presenilin-1 (PSEN1) E280A kindred, including 80 carriers (60 cognitively unimpaired, 20 cognitively impaired). Participants underwent structural magnetic resonance imaging, blood sampling, and neuropsychological testing. Analysis of covariance and false discovery rate-corrected t tests assessed group differences. Correlations evaluated associations among BF volume, age, and cognitive scores. Hamiltonian Markov chain Monte Carlo modeling estimated the age at which BF volume diverged between carriers and non-carriers. RESULTS:BF volume was comparable between cognitively unimpaired carriers and non-carriers but declined more rapidly in carriers, with divergence at ≈ 37.8 years, 6 years prior to the median age at onset of mild cognitive impairment. DISCUSSION:BF volume changes precede the onset of clinical symptoms in ADAD, supporting its potential as an early biomarker of cholinergic degeneration and therapeutic target. HIGHLIGHTS:There were not basal forebrain (BF) volume differences between PSEN1 E280A unimpaired carriers and non-carriers. Age-related modeling revealed a faster BF volume decline in carriers vs non-carriers. Age-related differences first emerged at 37.8 years, about 6 years before clinical onset. BF volume was related to age, cognition, and plasma phosphorylated tau 217 levels. BF changes may be early indicators of Alzheimer's disease-related neurodegeneration.

BACKGROUND:The Alzheimer's Prevention Initiative Autosomal Dominant Alzheimer's Disease (ADAD) Colombia Trial evaluated the biological, cognitive, and clinical effects of crenezumab, an anti-oligomeric and monomeric amyloid-beta (Aβ) monoclonal antibody, in 30-60-year-old PSEN1 E280A mutation carriers without cognitive impairment from the world's largest ADAD kindred, finding no significant treatment effects on Alzheimer's disease progression. This article describes baseline biomarker, cognitive, and clinical measurements and placebo-related longitudinal changes in the randomised prevention trial's mutation carrier and non-carrier groups. METHODS:Crenezumab and placebo-treated mutation carriers and placebo-treated non-carriers were assessed using amyloid and fluorodeoxyglucose positron emission tomography (PET), magnetic resonance imaging, plasma, and optional tau PET and cerebrospinal fluid (CSF) biomarker, cognitive, and clinical measurements over 5-8 years. FINDINGS:94% of the 252 kindred members (85 crenezumab-treated mutation carriers, 84 placebo-treated carriers, and 83 placebo-treated non-carriers) completed the trial. 55% of the carriers had baseline PET evidence of substantial Aβ plaques. 32.9% and 6.8% of amyloid PET-positive and PET-negative carriers, respectively, 36.5%, 13.2%, and 0% of pTau217-positive, pTau217-intermediate, and pTau217-negative carriers, respectively, and no non-carriers became cognitively impaired over the next 5 years. Carriers were distinguished from non-carriers by several baseline and longitudinal Aβ, tau, neurodegenerative, and inflammatory biomarker measures, but not by CSF oligomeric Aβ measurements. INTERPRETATION:Despite the absence of significant treatment effects, these findings and the trial itself continue to inform the course of preclinical ADAD, advance Alzheimer's disease prevention research, and provide a shared resource of data and samples for the field (ClinicalTrials.gov ID: NCT01998841; trial completed). FUNDING:National Institute on Aging, Banner Alzheimer's Institute, Genentech, Inc., and F. Hoffmann-La Roche Ltd.

Background Autosomal-dominant Alzheimer’s disease (ADAD) offers a model to define early biological changes in Alzheimer’s disease due to its predictable age at symptom onset. Although ultrasensitive plasma assays are available, their associations with age in ADAD remain incompletely characterized. Objectives To characterize age-related changes in plasma biomarkers and examine associations with cognition in PSEN1 E280A ADAD. Design and setting Cross-sectional observational study in members of the Colombian PSEN1 E280A kindred. Participants A total of 164 individuals were included, comprising 83 mutation carriers (mean age 34.36±9.82 years; 54% female) and 81 non-carriers (mean age 33.75±9.84 years; 52% female). Measurements Plasma Aβ42/Aβ40, phospho-tau217 (p-tau217), brain-derived tau (BD-tau), glial fibrillary acidic protein (GFAP), and neurofilament light (NfL) were quantified. Sex-adjusted associations with age, divergence ages between groups, classification performance (ROC curves), and associations with cognition (MMSE and CERAD delayed recall) were assessed. Results All plasma biomarkers were associated with age (p < .01). Divergence between carriers and non-carriers began with Aβ42/Aβ40 before age 18, followed by p-tau217 (26.0 years), GFAP (26.1 years), BD-tau (27.9 years), and NfL (38.7 years). Aβ42/Aβ40 showed the highest discrimination of mutation status (AUC=0.99), followed by p-tau217 (AUC=0.87) and GFAP (AUC=0.84). Among carriers, p-tau217, GFAP, BD-tau, and NfL were associated with MMSE, while p-tau217, GFAP, and NfL predicted CERAD delayed recall. Conclusion Plasma biomarkers exhibit a temporal cascade in PSEN1 E280A ADAD. P-tau217 and GFAP show the strongest associations with early cognitive decline, suggesting their potential utility for tracking disease progression and monitoring treatment effects in E280A carriers.

Representation learning on large-scale unstructured volumetric and surface meshes poses significant challenges in neuroimaging, especially when models must incorporate diverse vertex-level morphometric descriptors, such as cortical thickness, curvature, sulcal depth, and myelin content, which carry subtle disease-related signals. Current approaches either ignore these clinically informative features or support only a single mesh topology, restricting their use across imaging pipelines. We introduce a hierarchical transformer framework designed for heterogeneous mesh analysis that operates on spatially adaptive tree partitions constructed from simplicial complexes of arbitrary order. This design accommodates both volumetric and surface discretizations within a single architecture, enabling efficient multi-scale attention without topology-specific modifications. A feature projection module maps variable-length per-vertex clinical descriptors into the spatial hierarchy, separating geometric structure from feature dimensionality and allowing seamless integration of different neuroimaging feature sets. Self-supervised pretraining via masked reconstruction of both coordinates and morphometric channels on large unlabeled cohorts yields a transferable encoder backbone applicable to diverse downstream tasks and mesh modalities. We validate our approach on Alzheimer's disease classification and amyloid burden prediction using volumetric brain meshes from ADNI, as well as focal cortical dysplasia detection on cortical surface meshes from the MELD dataset, achieving state-of-the-art results across all benchmarks.

INTRODUCTION:Repetitive head impacts (RHI) from contact sports may cause a unique pattern of white matter hyperintensities (WMH) on T2-weighted fluid-attenuated inversion recovery (FLAIR) magnetic resonance imaging (MRI), termed RHI-associated WMH (RHI-WMH). These lesions are punctate, circular, and located at the gray-white matter boundary, an area vulnerable to trauma-related damage. METHODS:We investigated the association of RHI with these lesions in two aging cohorts: (1) former American football players versus asymptomatic unexposed men and (2) individuals with RHI from various contact sports versus non-RHI participants. RHI-WMH were assessed using visual ratings and a novel automated quantification pipeline. RESULTS:Individuals with RHI had greater RHI-WMH by both detection methods in both cohorts. RHI-WMH were associated with plasma neurofilament light and p-tau231, and flortaucipir positron emission tomography (PET) uptake. DISCUSSION:RHI-WMH may represent a new supportive biomarker for the detection of RHI-related neuropathologies later in life.

Artificial Intelligence (AI) has been increasingly applied to investigate genetic irregularities associated with Alzheimer's Disease (AD). However, its potential to uncover deeper, more detailed molecular and cellular mechanisms remains underexplored, primarily due to limitations in integrating large-scale data and capturing the complex, cell-type-specific dynamics involved in AD pathology. Single-cell RNA sequencing (scRNA-seq) has emerged as a powerful tool in transcriptomics, offering high-resolution, cell-specific insights into complex biological systems. Despite this advancement, a significant gap remains in identifying both common and cell-type-specific transcriptomic signatures that define AD-related cellular and molecular processes. To address this, we propose a deep learning framework leveraging a Multi-Layer Perceptron (MLP) to classify AD versus control nuclei using scRNA-seq data from the Religious Orders Study/Memory and Aging Project (ROSMAP). We focus on microglial subclusters, particularly those representing homeostatic and activated states, to train the MLP model for optimal classification performance. We utilize the predicted embeddings from the MLP to model a disease progression trajectory for each of the datasets. Our model demonstrates strong performance in both classification and disease trajectory inference. To enhance interpretability, SHapley Additive exPlanations (SHAP) are applied to identify key AD-associated genes. Based on the most salient genes implicated in AD, we built transcription gene regulatory networks, revealing novel transcription factors and regulons for AD pathogenesis. These regulons highlight profound impacts of dysregulations of proteostasis, endoplasmic reticulum (ER) stress responses, and circadian rhythm on synaptic plasticity and neuronal survival in AD, offering a more holistic approach to drug target discovery compared to conventional single-target strategies, potentially leading to greater efficacy in slowing or reversing disease progression. This work demonstrates the transformative potential of AI in elucidating the molecular mechanisms of AD, offering improvements over traditional methods and uncovering novel insights into disease pathogenesis and potential therapeutic targets.

, we propose a Cycle-GAN based harmonization model which eliminates the need of paired data for model training. Only a small fraction of paired data are required for model selection. By utilizing a larger number of unpaired data, it is expected to have better generality compared to using only a small number of paired data. Cycle-GAN model is adopted for unpaired harmonization in our study due to its promising performance in various unpaired tasks. Even though it is designed for image data, we modify the generator and discriminator by using multilayer perceptron (MLP) with skip-connection for our tabular data. The model is trained on unpaired data. During the training, we utilize a small fraction of paired validation data for model selection. we first generate the translated PiB ROIs from the corresponding FBP ROIs. Then, we calculate mean-cortical SUVR (mcSUVR) for both translated PiB ROIs and the corresponding paired PiB ROIs for Pearson correlation. The model with highest Pearson correlation is selected for further testing. We summarize our data in Tab. 1. The selected model achieved testing Pearson correlation between the mcSUVR calculated from translated PiB data and the corresponding testing PiB data (85 ROIs) compared to baseline calculated from paired FBP and PiB data. According to the Steiger's Z test, our method shows statistically significant improvement ( p  <0.0001) compared to the baseline. Besides, we also tried to include extra CL feature and two demographic features (age and sex) and we achieved and respectively. An unpaired harmonization model based on Cycle-GAN was developed. It only requires unpaired data for training and a small fraction of paired data for model selection. We achieved promising harmonization results on PiB and FBP measurements of cortical A.

BACKGROUND:The 677T and 1298C alleles of MTHFR (methylenetetrahydrofolate reductase) are both involved in homocysteine processing and are associated with increased risks for both Alzheimer's disease (AD) and cardiovascular disease. Given the role of vascular factors in AD, we investigated whether MTHFR allele interactions with APOE e4 and white matter hyperintensity volume (WMHV) are associated with amyloid-positivity in cognitively unimpaired (CU) older adults. METHOD:Data from 341 CU subjects (Age: 74.34±6.94, 58% Female, 32% e4 carriers, 33% amyloid-positive) from the Alzheimer's Disease Neuroimaging Initiative (ADNI) were analyzed. Amyloid-PET SUVR of >=1.11 was used to classify amyloid positivity. WMHV was quantified using the Lesion Segmentation Toolbox. Separate logistic regression models that adjusted for age and sex were used to investigate the contribution of the MTHFR 1298C and MTHFR 677T alleles individually as well as their interactions with APOE e4 carrier status and with WMHV on amyloid positivity. RESULT:Main effects for both MTHFR alleles were not significant (1298C: OR = 1.45, 95% CI = (0.56, 3.74), p =  0.44); (677T: OR = 0.59, 95% CI = (0.23, 1.54), p =  0.28). 1298C interactions with APOE e4 carrier status (OR = 0.35, 95% CI = (0.08, 1.43), p =  0.14) and WMHV (OR = 0.99, 95% CI = (0.90, 1.08), p =  0.84) were not associated with amyloid positivity. There was a significant interaction for 677T and APOE e4 carrier status (OR = 4.53, 95% CI = (1.01, 19.69), p =  0.04), but not with WMHV (OR = 1.08, 95% CI = (0.99, 1.19), p =  0.06). CONCLUSION:CU individuals who carry the APOE e4 and 677T allele of MTHFR are more likely to be classified as amyloid-positive. Treatments that target vascular-related APOE e4 and MTHFR 677T pathways may help ameliorate downstream amyloid pathology in AD.

Over the past decade, generative models have achieved significant success in enhancement fundus images.However, the evaluation of these models still presents a considerable challenge. A comprehensive evaluation benchmark for fundus image enhancement is indispensable for three main reasons: 1) The existing denoising metrics (e.g., PSNR, SSIM) are hardly to extend to downstream real-world clinical research (e.g., Vessel morphology consistency). 2) There is a lack of comprehensive evaluation for both paired and unpaired enhancement methods, along with the need for expert protocols to accurately assess clinical value. 3) An ideal evaluation system should provide insights to inform future developments of fundus image enhancement. To this end, we propose a novel comprehensive benchmark, EyeBench, to provide insights that align enhancement models with clinical needs, offering a foundation for future work to improve the clinical relevance and applicability of generative models for fundus image enhancement. EyeBench has three appealing properties: 1) multi-dimensional clinical alignment downstream evaluation: In addition to evaluating the enhancement task, we provide several clinically significant downstream tasks for fundus images, including vessel segmentation, DR grading, denoising generalization, and lesion segmentation. 2) Medical expert-guided evaluation design: We introduce a novel dataset that promote comprehensive and fair comparisons between paired and unpaired methods and includes a manual evaluation protocol by medical experts. 3) Valuable insights: Our benchmark study provides a comprehensive and rigorous evaluation of existing methods across different downstream tasks, assisting medical experts in making informed choices. Additionally, we offer further analysis of the challenges faced by existing methods. The code is available at

The utilization of artificial intelligence in studying the dysregulation of gene expression in Alzheimer's disease (AD) affected brain tissues remains underexplored, particularly in delineating common and specific transcriptomic signatures across different brain regions implicated in AD-related cellular and molecular processes, which could help illuminate novel disease biology for biomarker and target discovery. Herein we developed a deep learning framework, which consisted of multi-layer perceptron (MLP) models to classify neuropathologically confirmed AD versus controls, using bulk tissue RNA-seq data from the RNAseq Harmonization Study of the Accelerating Medicines Project for Alzheimer's Disease (AMP-AD) consortium. The models were trained based on data from three distinct brain regions, including dorsolateral prefrontal cortex (DLPFC), posterior cingulate cortex (PCC), and head of the caudate nucleus (HCN), obtained from the Religious Orders Study/Memory and Aging Project (ROSMAP). Subsequently, we inferred a disease progression trajectory for each brain region by applying unsupervised dimensionality transformation to the distribution of the subjects' expression profiles. To interpret the MLP models, we employed an interpretable method for deep neural network models, obtaining SHapley Additive exPlanations (SHAP) values and identified the most significantly AD-implicated genes for gene co-expression network analysis. Our models demonstrated robust performance in classification and prediction across two other external datasets from the Mayo RNA-seq (MAYO) cohort and the Mount Sinai Brain Bank (MSBB) cohort of AMP-AD. By interpreting the models both mechanistically and biologically, our study elucidated subtle molecular alterations in various brain regions, uncovering shared transcriptomic signatures activated in microglia and sex-specific modules in neurons relevant to AD. Notably, we identified, for the first time, a sex-linked transcription factor pair (ZFX/ZFY) associated with more pronounced neuronal loss in AD females, shedding light on a novel mechanism for sex dimorphism in AD. This study lays the groundwork for leveraging artificial intelligence methodologies to investigate AD at the molecular level, which is not readily achievable from conventional analysis approaches such as differential gene expression (DGE) analysis. The transcription factor implicated in sex difference also underpins a new molecular mechanistic basis of women's greater neurodegeneration in AD warranting further study.