Experiments in Replicating Science

Early diagnosis of Alzheimer's Disease (AD) is crucial for delaying its progression. While AI-based speech detection is non-invasive and cost-effective, it faces a critical data efficiency dilemma due to medical data scarcity and privacy barriers. Therefore, we propose FAL-AD, a novel framework that synergistically integrates federated learning with data augmentation to systematically optimize data efficiency. Our approach delivers three key breakthroughs: First, absolute efficiency improvement through voice conversion-based augmentation, which generates diverse pathological speech samples via cross-category voice-content recombination. Second, collaborative efficiency breakthrough via an adaptive federated learning paradigm, maximizing cross-institutional benefits under privacy constraints. Finally, representational efficiency optimization by an attentive cross-modal fusion model, which achieves fine-grained word-level alignment and acoustic-textual interaction. Evaluated on ADReSSo, FAL-AD achieves a state-of-the-art multi-modal accuracy of 91.52%, outperforming all centralized baselines and demonstrating a practical solution to the data efficiency dilemma. Our source code is publicly available at https://github.com/smileix/fal-ad.

Large language models (LLMs) have shown great potential for healthcare applications. However, existing evaluation benchmarks provide minimal coverage of Alzheimer's Disease and Related Dementias (ADRD). To address this gap, we introduce ADRD-Bench, the first ADRD-specific benchmark dataset designed for rigorous evaluation of LLMs. ADRD-Bench has two components: 1) ADRD Unified QA, a synthesis of 1,352 questions consolidated from seven established medical benchmarks, providing a unified assessment of clinical knowledge; and 2) ADRD Caregiving QA, a novel set of 149 questions derived from the Aging Brain Care (ABC) program, a widely used, evidence-based brain health management program. Guided by a program with national expertise in comprehensive ADRD care, this new set was designed to mitigate the lack of practical caregiving context in existing benchmarks. We evaluated 33 state-of-the-art LLMs on the proposed ADRD-Bench. Results showed that the accuracy of open-weight general models ranged from 0.63 to 0.93 (mean: 0.78; std: 0.09). The accuracy of open-weight medical models ranged from 0.48 to 0.93 (mean: 0.82; std: 0.13). The accuracy of closed-source general models ranged from 0.83 to 0.91 (mean: 0.89; std: 0.03). While top-tier models achieved high accuracies (>0.9), case studies revealed that inconsistent reasoning quality and stability limit their reliability, highlighting a critical need for domain-specific improvement to enhance LLMs' knowledge and reasoning grounded in daily caregiving data. The entire dataset is available at https://github.com/IIRL-ND/ADRD-Bench.

Portable, ultra-low-field (ULF) magnetic resonance imaging has the potential to expand access to neuroimaging but currently suffers from coarse spatial and angular resolutions and low signal-to-noise ratios. Diffusion tensor imaging (DTI), a sequence tailored to detect and reconstruct white matter tracts within the brain, is particularly prone to such imaging degradation due to inherent sequence design coupled with prolonged scan times. In addition, ULF DTI scans exhibit artifacting that spans both the space and angular domains, requiring a custom modelling algorithm for subsequent correction. We introduce a nine-direction, single-shell ULF DTI sequence, as well as a companion Bayesian bias field correction algorithm that possesses angular dependence and convolutional neural network-based superresolution algorithm that is generalizable across DTI datasets and does not require re-training (''DiffSR''). We show through a synthetic downsampling experiment and white matter assessment in real, matched ULF and high-field DTI scans that these algorithms can recover microstructural and volumetric white matter information at ULF. We also show that DiffSR can be directly applied to white matter-based Alzheimers disease classification in synthetically degraded scans, with notable improvements in agreement between DTI metrics, as compared to un-degraded scans. We freely disseminate the Bayesian bias correction algorithm and DiffSR with the goal of furthering progress on both ULF reconstruction methods and general DTI sequence harmonization. We release all code related to DiffSR for $\href{https://github.com/markolchanyi/DiffSR}{public \space use}$.

Alzheimer's disease (AD) is characterized by the gradual accumulation of amyloid pathology before the onset of cognitive impairment, yet effective strategies to alter disease progression remain limited. Here, we tested whether induction of a hibernation-like physiological state can slow amyloid pathology in an AD mouse model. To achieve sustained effects, we developed a protocol for repeated induction of a hypometabolic/hypothermic state and found that it significantly slowed plaque accumulation in a duration-dependent manner. This intervention also attenuated neuroinflammation and the formation of dystrophic neurites associated with plaques. Notably, this approach does not target specific molecular pathways but instead modulates global brain state. Our findings establish a proof-of-concept for a brain-state-based therapeutic paradigm for modifying disease progression in AD.

Epigenetic aging clocks based on DNA methylation patterns across the genome have emerged as a potential biomarker for risk of age-related diseases, like Alzheimer's disease (AD), and environmental and social stressors. However, methylation clocks have not been comprehensively validated in genetically diverse individuals. Here we evaluate a set of first-, second-, and third-generation methylation clocks in 621 AD patients and matched controls from African American, Hispanic, and White cohorts. The clocks are less accurate at predicting age in genetically admixed cohorts compared to the White cohort, especially for those with substantial African ancestry. This decreased accuracy holds in >2,500 individuals of European and African ancestry from three additional datasets. The clocks also fail to consistently identify age acceleration in admixed AD cases compared to controls. To explore potential causes for the lack of generalization of the clocks, we intersected clock CpGs with methylation, germline genetic variants, and methylation QTL (meQTL) data from global populations. We find differential methylation between African and European ancestry individuals is common for clock CpGs. Genetic variants rarely disrupt clock CpGs between populations, but a substantial fraction of clock CpGs have meQTL with significantly higher frequencies in African genetic ancestries. Our results demonstrate that methylation clocks often fail to predict age and AD risk when applied across populations and suggest avenues for improving their portability by considering differences in genetic and epigenetic patterns across human populations.

Mitochondrial contact site and cristae organizing system (MICOS) complexes are critical for maintaining the mitochondrial architecture, cristae integrity, and organelle communication in neurons. MICOS disruption has been implicated in neurodegenerative disorders, including Alzheimers disease (AD), yet the spatiotemporal dynamics of MICOS-associated neuronal alterations during aging remain unclear. Using three-dimensional reconstructions of hypothalamic and cortical neurons, we observed age-dependent fragmentation of mitochondrial cristae, reduced intermitochondrial connectivity, and compartment-specific changes in mitochondrial size and morphology. Notably, these structural deficits were most pronounced in neurons vulnerable to AD-related pathology, suggesting a mechanistic link between MICOS disruption and the early mitochondrial dysfunction observed in patients with AD. Our findings indicate that the loss of MICOS integrity is a progressive feature of neuronal aging, contributing to impaired bioenergetics and reduced resilience to metabolic stress and potentially facilitating neurodegenerative processes. MICOS disruption reduced neuronal firing and synaptic responsiveness, with miclxin treatment decreasing mitochondrial connectivity and inducing cristae disorganization. These changes link MICOS structural deficits directly to impaired neuronal excitability, highlighting vulnerability to AD-related neurodegeneration. These results underscore the importance of MICOS as a critical determinant of neuronal mitochondrial health and as a potential target for interventions aimed at mitigating AD-related mitochondrial dysfunction.

Dementia is a global public health challenge, with obesity emerging as an important modifiable risk factor. Here, we examined whether lifestyle interventions can mitigate the effects of diet-induced obesity on body weight, behaviour, and brain anatomy in mouse models. Using a longitudinal design, wild-type and triple-transgenic (3xTgAD) mouse models of Alzheimers disease were exposed to a high-fat diet and assigned to one of three interventions: voluntary physical activity, a low-fat diet, and their combination. A high-fat diet significantly increased body weight and induced widespread neuroanatomical changes, with effects modulated by sex and genotype. The combined intervention led to significant weight loss in males of both genotypes. Neuroanatomical analyses revealed that a high-fat diet significantly reduced hippocampal and cerebellar volumes in wild-type mice but had a less pronounced effect on 3xTgAD mice; nevertheless, interventions, particularly the combined approach, increased localized brain volumes in these regions regardless of genotype. Multivariate integration of behavioural and neuroanatomical measures identified a brain pattern linking hippocampal and cerebellar volumes to intervention and behavioural performance. Spatial gene enrichment analysis of this pattern identified biological processes, including glucose homeostasis, as potential biological mechanisms underlying intervention effects. Overall, these findings suggest that voluntary physical activity and a low-fat diet can modulate brain structure and behaviour, partially counteracting the effects of a high-fat diet, and potentially recruiting biological processes that may support brain health.

Much research into the aetiology of Alzheimers disease (AD) has focused on neuronal cell types, while studies on the contribution of glial cells, particularly oligodendrocytes (OLGs), are only starting to emerge. Altered brain DNA methylation, an epigenetic modification that provides the interplay between genetics and environmental cues to tightly regulate gene expression, is well documented in AD. Yet, cell-type-specific investigations remain limited. Here, we examine the role of DNA methylation and OLGs in AD, and how such changes may impact gene expression. We performed weighted-gene correlation network analysis (WGCNA) on multiple brain omics AD datasets across species: human DNA methylation data from 4 brain regions, human brain single-nuclei RNA sequencing data and mouse brain RNA sequencing data. We compared AD-associated network modules enriched for OLG genes across AD brain regions, as well as with other neurodegenerative disease DNA methylation datasets. We identified a DNA methylation signature associated with AD, enriched for OLGs, and preserved across brain regions representing early and late AD pathology stages. Genes within this signature showed altered expression in AD OLGs, confirming cell-type specificity and relevance to AD. This OLG signature was also preserved in transgenic mice with early A{beta} pathology and in other neurodegenerative diseases without A{beta} pathology. We reveal a consistent pattern of OLG dysfunction spanning early to late stages of AD, across DNA methylation and gene expression. Our findings highlight OLG-associated DNA methylation changes as important in AD pathogenesis, and possibly in other neurodegenerative diseases, opening new avenues for therapeutic development.

Tractometry enables detailed mapping of white matter microstructure along individual tracts and is widely used to study disease effects such as those seen in Alzheimers disease (AD). However, how different tractography algorithms influence tractometry outcomes remains unclear. Here, we compared whole-brain deterministic and probabilistic tractography using the BUndle ANalytics (BUAN) framework in the Alzheimers Disease Neuroimaging Initiative (ADNI) dataset, including 118 AD and 728 cognitively normal (CN) participants. Both approaches revealed the expected pattern of lower fractional anisotropy (FA) and higher mean, radial, and axial diffusivity (MD, RD, AxD) in AD, consistent with white matter degeneration. Despite broadly similar global trends, substantial bundle-level differences emerged between the two tractography methods. Probabilistic tracking produced stronger and more spatially extended effects in the fornix, a small and highly curved limbic pathway vulnerable to AD-related degeneration, whereas deterministic tracking showed greater sensitivity in the posterior segments of the right superior longitudinal fasciculus (SLF R). These discrepancies highlight that the choice of tractography algorithm can alter detecting disease effects, emphasizing the need for cross-method validation to ensure the robustness and interpretability of along-tract measures.

Alzheimer's disease (AD) alters brain electrophysiology and disrupts multichannel EEG dynamics, making accurate and clinically useful EEG-based diagnosis increasingly important for screening and disease monitoring. However, many existing approaches rely on black-box classifiers and do not explicitly model the underlying dynamics that generate observed signals. To address these limitations, we propose LERD, an end-to-end Bayesian electrophysiological neural dynamical system that infers latent neural events and their relational structure directly from multichannel EEG without event or interaction annotations. LERD combines a continuous-time event inference module with a stochastic event-generation process to capture flexible temporal patterns, while incorporating an electrophysiology-inspired dynamical prior to guide learning in a principled way. We further provide theoretical analysis that yields a tractable bound for training and stability guarantees for the inferred relational dynamics. Extensive experiments on synthetic benchmarks and two real-world AD EEG cohorts demonstrate that LERD consistently outperforms strong baselines and yields physiology-aligned latent summaries that help characterize group-level dynamical differences.

Alzheimer's disease (AD) and Lewy body dementia (LBD) present overlapping clinical features yet require distinct diagnostic strategies. While neuroimaging-based brain network analysis is promising, atlas-based representations may obscure individualized anatomy. Gyral folding-based networks using three-hinge gyri provide a biologically grounded alternative, but inter-individual variability in cortical folding results in inconsistent landmark correspondence and highly irregular network sizes, violating the fixed-topology and node-alignment assumptions of most existing graph learning methods, particularly in clinical datasets where pathological changes further amplify anatomical heterogeneity. We therefore propose a probability-invariant random-walk-based framework that classifies individualized gyral folding networks without explicit node alignment. Cortical similarity networks are built from local morphometric features and represented by distributions of anonymized random walks, with an anatomy-aware encoding that preserves permutation invariance. Experiments on a large clinical cohort of AD and LBD subjects show consistent improvements over existing gyral folding and atlas-based models, demonstrating robustness and potential for dementia diagnosis.

Alzheimer's disease (AD) has become a prevalent neurodegenerative disease worldwide. Traditional diagnosis still relies heavily on medical imaging and clinical assessment by physicians, which is often time-consuming and resource-intensive in terms of both human expertise and healthcare resources. In recent years, large language models (LLMs) have been increasingly applied to the medical field using electronic health records (EHRs), yet their application in Alzheimer's disease assessment remains limited, particularly given that AD involves complex multifactorial etiologies that are difficult to observe directly through imaging modalities. In this work, we propose leveraging LLMs to perform Chain-of-Thought (CoT) reasoning on patients' clinical EHRs. Unlike direct fine-tuning of LLMs on EHR data for AD classification, our approach utilizes LLM-generated CoT reasoning paths to provide the model with explicit diagnostic rationale for AD assessment, followed by structured CoT-based predictions. This pipeline not only enhances the model's ability to diagnose intrinsically complex factors but also improves the interpretability of the prediction process across different stages of AD progression. Experimental results demonstrate that the proposed CoT-based diagnostic framework significantly enhances stability and diagnostic performance across multiple CDR grading tasks, achieving up to a 15% improvement in F1 score compared to the zero-shot baseline method.

Predicting time-to-event outcomes when event times are interval censored is challenging because the exact event time is unobserved. Many existing survival analysis approaches for interval-censored data rely on strong model assumptions or cannot handle high-dimensional predictors. We develop ICODEN, an ordinary differential equation-based neural network for interval-censored data that models the hazard function through deep neural networks and obtains the cumulative hazard by solving an ordinary differential equation. ICODEN does not require the proportional hazards assumption or a prespecified parametric form for the hazard function, thereby permitting flexible survival modeling. Across simulation settings with proportional or non-proportional hazards and both linear and nonlinear covariate effects, ICODEN consistently achieves satisfactory predictive accuracy and remains stable as the number of predictors increases. Applications to data from multiple phases of the Alzheimer's Disease Neuroimaging Initiative (ADNI) and to two Age-Related Eye Disease Studies (AREDS and AREDS2) for age-related macular degeneration (AMD) demonstrate ICODEN's robust prediction performance. In both applications, predicting time-to-AD or time-to-late AMD, ICODEN effectively uses hundreds to more than 1,000 SNPs and supports data-driven subgroup identification with differential progression risk profiles. These results establish ICODEN as a practical assumption-lean tool for prediction with interval-censored survival data in high-dimensional biomedical settings.

Normative modeling learns a healthy reference distribution and quantifies subject-specific deviations to capture heterogeneous disease effects. In Alzheimers disease (AD), multimodal neuroimaging offers complementary signals but VAE-based normative models often (i) fit the healthy reference distribution imperfectly, inflating false positives, and (ii) use posterior aggregation (e.g., PoE/MoE) that can yield weak multimodal fusion in the shared latent space. We propose mmSIVAE, a multimodal soft-introspective variational autoencoder combined with Mixture-of-Product-of-Experts (MOPOE) aggregation to improve reference fidelity and multimodal integration. We compute deviation scores in latent space and feature space as distances from the learned healthy distributions, and map statistically significant latent deviations to regional abnormalities for interpretability. On ADNI MRI regional volumes and amyloid PET SUVR, mmSIVAE improves reconstruction on held-out controls and produces more discriminative deviation scores for outlier detection than VAE baselines, with higher likelihood ratios and clearer separation between control and AD-spectrum cohorts. Deviation maps highlight region-level patterns aligned with established AD-related changes. More broadly, our results highlight the importance of training objectives that prioritize reference-distribution fidelity and robust multimodal posterior aggregation for normative modeling, with implications for deviation-based analysis across multimodal clinical data.