Attention-guided electroencephalography fusion with fractal and nonlinear dynamics decodes early Alzheimer’s biomarkers

Alzheimer’s disease (AD), the most common cause of dementia, is characterized by progressive memory and cognitive decline. Conventional genome-wide association studies (GWAS) comparing AD cases and controls may miss genetic influences that act along a continuum of cognitive function. Using data from 3007 participants in the National Institute on Aging Late-Onset Alzheimer’s Disease Family Study (NIA-LOAD GWAS), we conducted a family-based GWAS of eight quantitative cognitive phenotypes encompassing episodic memory (Logical Memory IA and IIA), working memory (Digit Span Forward, Backward, and Ordering), and semantic fluency (Animal, Fruit and Vegetable, and Vegetable Fluency). Family-based association testing in PLINK v1.9 identified numerous single nucleotide polymorphisms (SNPs) associated with cognitive phenotypes at genome-wide significant (p < 5 × 10−8) levels. Notably, genome-wide significant variants with cognatic functions were localized to genes implicated in synaptic function, neurodevelopment, and neurodegeneration, including TOMM40 (rs2075650), ERBB4 (rs1521543), APLP2 (rs12281267, rs959354), PTPRD (rs1353983, rs970347, rs1392511), NCAM2 (rs2826728), GRM7 (rs6788201), PAX5 (rs2988003, rs2381595), NRG1 (rs16875655), and NRG3 (rs1937957). Furthermore, the TOMM40 (rs2075650) was significantly associated with AD as a binary outcome (p = 4.60 × 10−24) and APLP2 (rs12281267, rs959354), APOE (rs405509), PTPRD (rs1353983, rs970347, rs1392511) were associated with AD (p < 0.001). Additionally, several pathways including the ERBB4 signaling pathway (adjusted p = 2.82 × 10−3), driven by ERBB4, NRG1, and NRG3 may contribute to cognitive impairments. This study provides a comprehensive resource of cognitive endophenotype associations in AD families, advancing understanding of the genetic architecture underlying memory, executive function, and cognitive aging, and highlights new therapeutic targets for replication and functional follow-up.

Diagnostic models using primary care routine clinical variables have been limited in their ability to identify Alzheimer's disease (AD) patients. In this study, we sought to better understand the effect of mild cognitive impairment (MCI) on the predictive performance of AD diagnostic models. We sourced data from the Alzheimer's Disease Neuroimaging Initiative (ADNI) cohort. CatBoost was used to assess the utility of routine clinical variables that are accessible to primary care physicians, such as hematological and blood tests and medical history, in multiclass classification between healthy controls, MCI, and AD. Our results indicated that MCI indeed affected the predictive performance of AD diagnostic models. Of the three subgroups of MCI that we found, this finding was driven by a subgroup of MCI patients who likely have prodromal AD. This work highlights the importance of continuing to focus on better classification of the different types of MCI to improve diagnostic models of AD, rather than focusing on binary classifications between AD and control cases. Future research should focus on distinguishing MCI from prodromal AD as the utmost priority for improving translational AD diagnostic models for primary care physicians.

Immunohistochemically (IHC) measured transactive response DNA-binding protein 43 (TDP-43) inclusions are observed in Alzheimer's disease (AD) and are associated with medial temporal lobe atrophy. Accumulation of cryptic exons occurs in AD in response to TDP-43 pathology. We aimed to assess relationships between IHC and biochemically measured insoluble TDP-43 and cryptic exons and assess associations with hippocampal and amygdala volume loss and atrophy rates on magnetic resonance imaging (MRI). Eighty-one neuropathologically diagnosed AD cases were analyzed. For biochemistry, insoluble TDP-43 was quantified using a Meso-scale discovery (MSD) immunoassay. IHC-TDP burden was quantified with digital histopathology. Cryptic RNAs were assessed via quantitative real-time polymerase chain reaction (qRT-PCR). Thirty-eight cases had serial brain MRI. Hippocampal and amygdala volumes were calculated using FreeSurfer. Regression models were used to investigate associations among IHC-TDP-43 status/burden, MSD-TDP status/levels, cryptic RNAs, and hippocampal and amygdala volumes and atrophy rates. IHC-TDP(+) cases exhibited elevated levels of MSD-TDP and cryptic RNAs (KCNQ2, STMN2, and UNC13A) and increased MSD-TDP levels were associated with increased cryptic RNA levels, in the hippocampus and amygdala. IHC-TDP(+) cases had smaller hippocampal and amygdala volumes compared to IHC-TDP(-) cases. MSD-TDP(+) cases had smaller hippocampal volumes and faster amygdala rates of atrophy compared with MSD-TDP(-) cases. Higher KCNQ2 and UNC13A levels were associated with smaller amygdala volumes. MSD-TDP level is a reliable surrogate for IHC-based TDP-43 status. Both TDP-43 and cryptic RNA levels are associated with reduced medial temporal volumes, suggesting cryptic exons may be playing a role in brain volume loss in AD. ANN NEUROL 2026.

INTRODUCTIONThe olfactory system is an early target in Alzheimer's disease (AD), yet regional glial pathology interactions remain poorly defined. We examined how glial activation and pathological burden differ between the olfactory cortex (OC) and olfactory bulb (OB) across disease stages.METHODSPost mortem OC and OB samples from cognitively normal (CN), mild cognitive impairment, and AD cases were analyzed using immunohistochemistry and immunofluorescence for amyloid beta (Aβ), phosphorylated tau (pTau), Iba1 (microglia), GFAP (astrocyte), and apolipoprotein E (apoE).RESULTSBoth regions showed stage‐dependent increases in Aβ and pTau, with regionally distinct glial responses. ApoE signal varied with clinical stage rather than genotype. Co‐expression analyses revealed astrocyte‐linked networks in the OC and microglia‐linked relationships in the OB.DISCUSSIONFindings demonstrate spatially heterogenous glial pathology architectures in the human olfactory system, supporting its role as an early and regionally diverse site of AD vulnerability.

Apolipoprotein E ( APOE ) genotype contributes significantly to Alzheimer's disease (AD) risk and pathogenesis. Cell-type specific effects of APOE alleles have been studied. However, due to the variable prevalence of APOE genotypes within human populations, characterization of cell-type specific transcriptomes across APOE genotypes has been challenging. Here, we integrated previous and newly generated single-nuclei sequencing (snRNA-seq) data in the prefrontal cortex (PFC) from individuals across APOE genotypes ( 2/2 , 2/3 , 3/3 , 3/4 , 4/4 ). Clustering analysis revealed distinct excitatory and microglial subpopulations that were uniquely enriched or depleted for APOE4/4 AD. Notably, an excitatory neuronal cluster exhibited neurofibrillary tangle (NFT) signatures and was selectively depleted in APOE4/4 AD cases. In addition, several microglial subpopulations were influenced by both APOE4 dosage and disease status. Among these, the putative AD risk gene FRMD4A emerged as APOE4 dose and AD-dependent. These findings were validated by RNAscope in an extended cohort. Together, our findings provide insights into how APOE4 reshapes cellular states and contributes to cell-type-specific vulnerability in AD.

Neuroinflammation plays a crucial role in neurodegenerative disorders such as Alzheimer's disease. The P2X7 receptor (P2X7R), expressed on microglia, is involved in neuroinflammatory responses. Despite evidence of P2X7R upregulation in Alzheimer's disease models, its role in human Alzheimer's disease remains unclear. The PET radioligand [18F]JNJ-64413739 enables the assessment of P2X7R distribution in post-mortem Alzheimer's disease brain tissue. Post-mortem brain tissue from Alzheimer's disease and control subjects was obtained. [18F]JNJ-64413739 was synthesised and applied to tissue sections from the temporal and parietal cortex. Autoradiography was conducted with and without the P2X7R antagonist JNJ54173717. [18F]JNJ-64413739 binding was observed across all brain regions, with effective blocking confirming specificity. No significant differences were found between Alzheimer's disease and controls in the temporal (P = 0.84) or parietal cortex (P = 0.90) in the first experiment. The second experiment, using a modified protocol also did not reveal a significant difference between controls and Alzheimer's disease in either temporal (P = 0.66) or parietal cortex (P = 0.38). White matter exhibited significantly higher binding than grey matter (P < 0.01), but no disease-specific differences were noted. This study demonstrates P2X7 receptor-specific binding of [18F]JNJ-64413739 but finds no significant differences between post-mortem tissue of Alzheimer's disease cases and controls. These findings suggest that while the tracer shows promising in vitro characteristics, the role of P2X7R in Alzheimer's disease pathology and its utility as a biomarker require further validation through in vivo imaging studies across disease stages.

Proteomic studies have generated robust assessments of protein abundance changes in Alzheimer's disease (AD); however, identifying how the protein abundance changes affect specific biological processes remains a challenge. To address these hurdles, we used a multi-network computational analysis approach that integrated dendritic spine morphometry data with mass spectrometry-based proteomics from the same individuals. The samples exhibited a range of AD neuropathology and were categorized into three groups: controls, asymptomatic AD, and AD cases. Multiplex tandem mass tag mass spectrometry proteomic data (N = 8,212 proteins) was generated on Brodmann area 46 (BA46) dorsolateral prefrontal cortex (DLPFC) human samples (N = 41, 23 males and 18 females), from which dendritic spine morphometry analysis existed. To integrate the multi-scale data types, two computational network analysis methods were performed, including WeiGhted co-expression network analysis (WGCNA) and SpeakEasy2 (SE2). Both WGCNA and SE2 revealed that the mitochondria protein modules were decreased in AsymAD and AD cases compared to controls, whereas the DNA repair modules were increased in AsymAD and AD compared to controls. Synaptic protein modules that correlated to multiple spine morphology traits were identified in both WGCNA and SE2. Pearson correlation analyses identified over a dozen individual proteins linked to multiple dendritic spine density and morphology traits. Collectively, these findings demonstrate how integration of spine morphometry data with proteomics can contextualize proteins for functional validation and identify synaptic alterations in AD progression.Significance Statement Cognitive decline in Alzheimer's disease associates more strongly with synapse and dendritic spine loss than amyloid-beta or tau pathology. However, one in three individuals harbor Alzheimer's disease neuropathology at death but were cognitively indistinguishable from baseline in life. Preservation of spines and synapses is hypothesized to prevent cognitive decline in these individuals. Identifying the molecular drivers of synaptic changes in Alzheimer's disease could yield deeper understanding of disease progression. Here, we utilized two computational network approaches that integrated multi-scale data, including proteomics and dendritic spine morphometry from the same humans, to identify proteins relevant to synapses in Alzheimer's disease. Hundreds of proteins related to mitochondria, DNA repair, and synaptic signaling were associated with alterations in synapse structure and function in Alzheimer's disease.

SGK1 inhibition restores excitability of cortical neurons derived from Alzheimer's disease patients.

Lecanemab binds "protofibrils," which are poorly characterized in human brain. It is unknown why lecanemab caused fewer amyloid-related imaging abnormalities (ARIAs) than other antibodies in trials. The apolipoprotein E (APOE) ε4 allele increases ARIA risk through unknown mechanisms. Equilibrium binding constants (KD) and total amyloid beta (Aβ) binding (Bmax) of aducanumab, lecanemab, and donanemab equivalents to soluble and insoluble amyloid plaque-enriched and cerebral amyloid angiopathy (CAA)-enriched Aβ were compared across 17 Alzheimer's disease (AD) cases by mixed models. Titrated immunofluorescence (IF) staining compared antibody binding. Lecanemab and aducanumab had indistinguishable preference for "protofibrils." Antibody preference for plaque-enriched versus CAA-enriched Aβ did not differ in soluble extracts or by IF staining but differed slightly in insoluble extracts. The APOE ε4 allele was associated with more soluble antibody-accessible Aβ. Lecanemab's binding target is similar to other antibodies'. Differences in antibody preference for plaque versus CAA Aβ may not explain differences in ARIA with edema rates.

An effective alzheimer disease diagnosis using resting state fmri images and broad learning system.

Whether healthy lifestyle behaviors are associated with Alzheimer's disease (AD) risk among individuals with hyperlipidemia remains unclear. We analyzed 241,642 dementia-free participants from the UK Biobank. A weighted lifestyle score (0-7) was derived from seven factors and categorized into five tiers. Hyperlipidemia was defined as lipid-lowering medication use or LDL-cholesterol ≥ 4.0 mmol/L. Cox regression estimated hazard ratios (HRs) and 95% confidence intervals (CIs). Over a median follow-up of 14.5 years, 1728 AD cases occurred, including 977 cases among 104,082 individuals with hyperlipidemia. Compared with the intermediate tier, unhealthy lifestyle was associated with elevated AD risk (HR: 1.17; 95% CI: 1.02-1.35), while healtshy and very healthy tiers were associated with progressively lower risk (HR=0.85 and 0.74, respectively). These associations were evident among individuals with hyperlipidemia, but not statistically significant among those without hyperlipidemia. Healthy lifestyle patterns were associated with lower AD risk among individuals with hyperlipidemia, with greater risk reductions observed for healthier lifestyle tiers.

To address the limited accuracy of speech-based Alzheimer's Disease (AD) screening and the shortage of paired multimodal data, this paper proposes a detection framework based on feature alignment and Rectified Flow-driven latent representation generation. The EEG dataset consists of 36 AD patients and 29 Healthy Controls (HC). The speech dataset contains 399 samples, which include 114 AD cases, 132 Mild Cognitive Impairment (MCI) cases, and 153 HC cases. We extracted multidimensional features of EEG signals, such as time-domain and frequency-domain characteristics, alongside behavioral representations of speech. A heterogeneous alignment network was used to map these features into a common semantic subspace, where an adaptive interpolation strategy reconstructed the missing pathological trajectories of MCI within the latent space. On this basis, a conditional Rectified Flow model was introduced to learn the optimal transport mapping from speech to EEG. This model generated physiological-information-rich latent representations to compensate for semantic gaps. Experimental results showed that the fused features from speech and latent representations achieved a three-class classification accuracy of 89.08%, a precision of 88.77%, and a recall of 88.71%. This performance represented an accuracy improvement of 9.28% compared with the speech-based baseline system. Our method combines the convenience of speech screening with the high reliability of neurophysiological signals, and it provides a new approach for low-cost early detection of AD.

ADAM10 and ADAM17, members of the membrane-bound disintegrin metalloproteinase (ADAM) family, act as α-secretases in the non-amyloidogenic processing of the amyloid precursor protein. While alterations in ADAM10, the primary α-secretase, have been reported in Alzheimer’s disease (AD), with certain controversy between studies, ADAM17 changes remain less understood. Both enzymes display a complex pattern of expression that includes immature zymogen and mature full-length, assigned as the active form, and soluble cleaved species. We reported for the first time that ADAM10 is present in human cerebrospinal fluid (CSF), and levels of the full-length and soluble fragments decrease in samples from AD patients. ADAM10 and ADAM17 levels were determined in human samples from frontal cortex (Brodmann areas 9/10) from histopathologically-confirmed AD cases (n = 16) and non-dementia controls (n = 13), and in CSF from individuals with cognitive symptoms and an AD-consistent biomarker profile (n = 20) and from controls with normal biomarker parameters (n = 18). Levels were also assessed in SH-SY5Y cells differentiated to neurons, after 48 h of 3 µM Aβ42 treatment. Protein species were resolved by electrophoresis followed by quantitative fluorescent western blotting. Significant differences between AD and control groups were evaluated using Student’s unpaired t-test or the Mann–Whitney U test, as appropriate. We confirm the significant reduction in levels of mature and soluble ADAM10 species in lumbar CSF of AD patients. We also validate in postmortem AD brain the decrease in mature species of ADAM10 with no change in the levels of its zymogen nor mRNA transcript, suggesting a post-translational dysregulation in the production of active protein. In contrast, ADAM17 levels remained unchanged in both brain and CSF samples, highlighting a differential regulation between these two α-secretases in AD pathology. In SH-SY5Y-differentiated neurons, treatment with Aβ42 also led to a specific reduction in mature ADAM10, implicating Aβ in the disruption of ADAM10 maturation. Our study links a compromise in α-secretase to AD pathology. It demonstrates a selective reduction of the mature forms of ADAM10 in AD that may contribute to an impaired non-amyloidogenic APP processing. Aβ42 appears to be a key modulator of this dysfunction, suggesting that ADAM10 downregulation may be an early event in AD pathogenesis. The results don’t evidence any major role for ADAM17 in AD pathogenesis.

BackgroundIncreasing evidence implicates herpes simplex virus type 1 (HSV1) in the pathogenesis of Alzheimer's disease (AD). The genome-wide association studies have not detected any genes regulating antibody responses to HSV1.ObjectiveTo determine whether the magnitude of antibody responses to HSV1 and to its glycoprotein D (gD) differed between AD patients and controls. Using a candidate gene approach, determine if the antibody responses were associated with immunoglobulin GM (γ marker) and KM (κ marker) allotypes-hereditary antigenic determinants of γ and κ chains, respectively. We also aimed to determine whether GM and KM allotypes epistatically interacted with an AD risk gene-reelin-encoding RELN (rs2299356)-and contributed to immunity to HSV1 and HSV1-gD.MethodsGenotyping was done by polymerase chain reaction-restriction fragment length polymorphism, TaqMan®, and rhAMP® SNP genotyping assays. IgG antibodies to HSV1 and HSV1-gD were measured by an enzyme-linked immunosorbent assay.ResultsAnti-HSV1 antibody levels were not significantly different between AD cases and controls; however, anti-HSV1-gD antibody levels were over two-fold higher (p < 0.0001) in AD cases compared to controls. GM 23 allele was associated with higher anti-HSV1 antibody levels in both cases and controls. Potential interaction between KM and RELN rs2299356 alleles on antibody responses to HSV1-gD was detected in AD cases, but not in controls.ConclusionsIn view of the fact that HSV1-gD is a vaccine candidate, the findings of higher anti-HSV1-gD antibody levels in AD patients and potential interaction of KM and RELN rs2299356 alleles in this study warrant additional large-scale multiethnic studies on this issue.

BackgroundCyclin-dependent kinase 5 (CDK5) plays a critical role in neuronal development, synaptic plasticity, and cytoskeletal regulation. Its dysregulation has been implicated in Alzheimer's disease (AD); however, supporting genetic evidence remains limited in Middle Eastern populations.ObjectiveThis study aimed to identify CDK5 genetic variants in a Saudi cohort with AD and evaluate their association with disease susceptibility, without addressing downstream functional mechanisms.MethodsPeripheral blood samples were obtained from 52 Saudi patients with AD and 60 age- and sex-matched healthy controls. Genomic DNA was extracted and quantified using a NanoDrop One spectrophotometer. CDK5 exons 1, 2, 6, 7, and 12 were amplified and sequenced by Sanger methodology on an ABI 3500 Genetic Analyzer. Variant screening and frequency comparisons were performed between dementia and control groups.ResultsEight CDK5 variants were identified in exonic (2, 12) and intronic (1-2, 7-8, 11-12) regions, including three novel variants (c.105delG, c.858C > T, and one intronic variant). No variants were detected in exons 1, 6, or 7. Frameshift variants c.103del and c.105delG were significantly more frequent in AD cases (38%) than controls (5%) (p < 0.001). Conversely, the synonymous c.855C > T variant (Exon 12) was more common in controls (23%) than cases (5%) (p < 0.01), indicating a significant association between specific CDK5 variants and AD.ConclusionsIdentification of novel CDK5 variants associated with AD in a Saudi cohort highlights population-specific genetic susceptibility. Further functional and biomarker studies are needed to clarify their biological relevance.

Alzheimer's disease (AD) and Down Syndrome (DS) are characterized by the aggregation of tau tangles. As a novel tau PET tracer in AD, [18F]MK-6240 has the potential in Down syndrome-associated Alzheimer’s disease (DSAD) to elucidate pathophysiology and advance diagnostic strategies. Autoradiography of frontal cortex and temporal cortex postmortem brain slices of DSAD (n = 5), AD (n = 5), and cognitively normal (CN) (n = 5) cases indicated similarly high [18F]MK-6240 binding in DSAD and AD cases. [18F]MK-6240 binding in CN was substantially less than DSAD and AD. Anti-tau immunostains confirmed total tau presence that aligned with the localization of [18F]MK-6240 binding. DSAD and AD cases exhibited higher gray matter (GM)/white matter ratios of 2.8 and 2.5 respectively. For drug effects on [18F]MK-6240 binding, self-displacement by MK-6240 (10 µM) reduced binding by 88% among DSAD cases and 85% among AD cases while IPPI (10 µM) displaced [18F]MK-6240 by 81% and 74% in DSAD and AD cases respectively. KuFal194 (10 µM), a specific phosphokinase inhibitor, minimally displaced [18F]MK-6240 binding. Harmine competed with [18F]MK-6240 with an IC50 value of 290 nM and 92 nM for DSAD and AD cases, respectively. High meninges off-target (non-tau) binding of [18F]MK-6240 was observed in a CN case, comparable to the GM in DSAD and AD. MK-6240 (10 µM) blocked 44% and T807 (10 µM) blocked 30% of meninges binding. Incubation of meninges in the presence of 0.2% polyethyleneimine reduced 70% of [18F]MK-6240 binding. The tau radioligand, [125I]IPPI, an analog of [18F]MK-6240, exhibited minimal binding to CN meninges compared to the DSAD and AD cases. Our findings suggest [18F]MK-6240 to be selective tau imaging agent in DSAD and AD, harmine to uniquely compete with [18F]MK-6240 binding in DSAD and provide preliminary insights to the off-target nonspecific binding of [18F]MK-6240 in postmortem meninges.

Millions of individuals worldwide are affected by Alzheimer's disease dementia (ADD) and frontotemporal dementia (FTD), with FTD characterized by degeneration of the frontal and temporal lobes leading to cognitive and behavioral impairments. A subset of Alzheimer's cases exhibits familial inheritance, with the PAISA mutation, a glutamic acid to alanine substitution at codon 280 (E280A) in the PSEN1 gene, being a primary cause of early-onset dementia. PSEN1 encodes a key component of the γ-secretase complex, which cleaves amyloid precursor protein (APP) to generate beta-amyloid (Aβ) peptides. The PAISA mutation disrupts normal Aβ processing, leading to overproduction or accumulation of Aβ, formation of amyloid plaques, and accelerated progression of dementia. Its prevalence is particularly high in Colombian families, giving rise to the term "PAISA mutation." The APOE genotype further modulates the clinical manifestation in PAISA carriers, with APOE2 potentially delaying disease onset, whereas APOE4 is associated with earlier onset. Recent research highlights TAF2N (also known as RBP56, encoded by TAF15) as a promising therapeutic target, as its modulation may regulate AD-associated genes, reduce toxic Aβ isoforms, modulate tau and APP pathways, protect neurons, and enhance synaptic function. Overall, understanding the molecular effects of PAISA mutations and exploring TAF2N-targeted therapies offers novel avenues for addressing early-onset familial AD, providing insights into broader mechanisms of disease pathogenesis.

GDF-11, GDF-15, Jag-1, and leptin in neuronal-derived extracellular vesicles as aging-related biomarkers to identify individuals at risk of Alzheimer's dementia: A pilot study.

BackgroundThe Monitoring drug Efficacy through Multi-Omics Research Initiative in Alzheimer's Disease (MEMORI-AD) study aims to characterize drug response to donepezil and memantine in a Filipino cohort with late-onset Alzheimer's disease (AD). To enhance enrollment, community-based dementia screening and recruitment were conducted across barangays in Manila.ObjectiveThis study describes the experiences and challenges in community-based recruitment for multi-omics research in vulnerable populations.MethodsAfter coordinating with local officials for approval, recruitment targeted residents aged 65 years and above. A public health lecture on dementia was held, followed by rapid screening using the “10 Warning Signs of Dementia” questionnaire by the Alzheimer's Disease Association of the Philippines and the AD-8 Dementia Screening Interview. Eligible participants were invited for clinical screening, including laboratory work, neuropsychological tests, and neuroimaging. Reasons for non-consent and withdrawal were recorded.ResultsOf 301 eligible participants, 42 (14.0%) were suspected of dementia and eligible for work-up. Among them, 14 (33.3%) did not consent, 15 (53.6%) initially consented but withdrew, and 13 (46.4%) completed the assessment. Only 2 patients were enrolled. The main cause for non-consent was disinterest (n = 8, 57.1%), followed by unavailability of a legally authorized representative (LAR) (n = 3, 21.4%), which was also the primary reason for withdrawal (n = 9, 60%).​ConclusionsRecruitment of vulnerable Filipinos in multi-omics research within a lower-middle income country faced significant barriers, encompassing economic, psychosocial, and health constraints, largely due to disinterest and unavailability of LAR. Further studies on family dynamics, socioeconomic, and cultural factors could offer insights to overcome these recruitment barriers.Trial Registration NumberPhilippine Health Research Registry ID PHRR230220-0054116; ClinicalTrials.gov ID NCT05801380.