The clinical use of structural MRI in Alzheimer disease

Chemical exchange saturation transfer MRI for neurodegenerative diseases: An update on clinical and preclinical studies

Chemical exchange saturation transfer magnetic resonance imaging is an advanced imaging technique that enables the detection of compounds at low concentrations with high sensitivity and spatial resolution and has been extensively studied for diagnosing malignancy and stroke. In recent years, the emerging exploration of chemical exchange saturation transfer magnetic resonance imaging for detecting pathological changes in neurodegenerative diseases has opened up new possibilities for early detection and repetitive scans without ionizing radiation. This review serves as an overview of chemical exchange saturation transfer magnetic resonance imaging with detailed information on contrast mechanisms and processing methods and summarizes recent developments in both clinical and preclinical studies of chemical exchange saturation transfer magnetic resonance imaging for Alzheimer's disease, Parkinson's disease, multiple sclerosis, and Huntington's disease. A comprehensive literature search was conducted using databases such as PubMed and Google Scholar, focusing on peer-reviewed articles from the past 15 years relevant to clinical and preclinical applications. The findings suggest that chemical exchange saturation transfer magnetic resonance imaging has the potential to detect molecular changes and altered metabolism, which may aid in early diagnosis and assessment of the severity of neurodegenerative diseases. Although promising results have been observed in selected clinical and preclinical trials, further validations are needed to evaluate their clinical value. When combined with other imaging modalities and advanced analytical methods, chemical exchange saturation transfer magnetic resonance imaging shows potential as an in vivo biomarker, enhancing the understanding of neuropathological mechanisms in neurodegenerative diseases.

Association between diffusion MRI-based measures of neurite microstructure and risk of Alzheimer's disease

Early detection of Alzheimer's disease (AD) is crucial for intervention, but traditional MRI and cognitive assessments may miss pre-symptomatic changes. Advanced diffusion MRI (dMRI) methods, such as Neurite Orientation Dispersion and Density Imaging (NODDI), show promise in identifying early brain changes. We analyzed 65 cognitively unimpaired older adults (25 APOE-e4 carriers, 40 non-carriers) from the ADNI3 dataset. NODDI's neurite density index (NDI) and orientation dispersion index (ODI), volumetric MRI and cognition (MoCA) were analyzed in key brain regions like the hippocampus, fusiform gyrus, and entorhinal cortex. Statistical analyses included linear regression and t-tests, with FDR correction. NDI differed significantly between carriers and non-carriers and correlated with MoCA scores. ODI differed only in the CA1 hippocampal subfield. Volumetric MRI measures showed no group differences. Significant APOE-e4 group differences were observed in NDI for the left fusiform gyrus (β = 0.015, p = 0.02), right fusiform gyrus (β = 0.018, p = 0.02), left entorhinal cortex (β = 0.018, p = 0.04), right entorhinal cortex (β = 0.018, p = 0.03), left CA1 (β = 0.03, p = 0.02), and left CA2-3 (β = 0.03, p = 0.02). ODI differences were observed only in left CA1 (β = 0.037, p = 0.008). No volumetric measures differed significantly. MoCA correlated with NDI in bilateral entorhinal cortices (p = 0.001-0.05), left fusiform gyrus (p = 0.02), and right CA2-3 (p = 0.02). NODDI metrics, particularly NDI, could help detect early APOE-e4-related microstructural changes, while traditional volumetric MRI measures remain uninformative at early stages.

Integration of automatic MRI segmentation techniques with neuropsychological assessments for early diagnosis and prognosis of Alzheimer's disease. A systematic review

BACKGROUND

This systematic review investigates the integration of automatic segmentation techniques of magnetic resonance imaging (MRI) with neuropsychological assessments for early diagnosis and prognosis of Alzheimer's Disease (AD).

OBJECTIVES

Focus on studies that utilise automated MRI segmentation and neuropsychological evaluations across the AD spectrum.

DATA SOURCES

A literature search was conducted on the PubMed database on 7 November 2024, using key terms related to MRI, segmentation, brain structures, AD, and cognitive decline.

STUDY ELIGIBILITY CRITERIA

Studies including individuals with AD, mild cognitive impairment (MCI), or subjective cognitive decline (SCD), utilising structural MRI, focusing on grey matter and automatic segmentation, and reporting cognitive assessments were included.

STUDY APPRAISAL AND SYNTHESIS METHODS

Data were extracted and synthesised focusing on associations between MRI measures and cognitive tests, and discriminative values for diagnosis or prognosis.

RESULTS

Seven studies were included, showing a significant association between structural changes in the medial temporal lobe and cognitive decline. The combination of MRI volumetric measures and neuropsychological scores enhanced diagnostic accuracy. Neuropsychological measures demonstrated superiority in the identification of patients with MCI and mild AD in comparison to MRI measures alone.

LIMITATIONS

Heterogeneity across studies, selection and measurement bias, and lack of non-response data were noted.

CONCLUSIONS AND IMPLICATIONS

This review emphasises the necessity of integrating automated MRI segmentation with neuropsychological assessments for the diagnosis and prognosis of AD. While MRI is valuable, neuropsychological testing remains essential for early detection. Future studies should focus on developing integrated predictive models and refining neuroimaging techniques.

Alterations in Olfactory Cortex Volume in Mild Cognitive Impairment and Mild Alzheimer's Disease Dementia: A Study of Sex-Related Differences

BACKGROUND/OBJECTIVES

Aging is one of the greatest risk factors for neurodegenerative diseases such as Alzheimer's disease (AD). As the disease progresses, neural loss in brain regions, such as the olfactory cortex (OC), i.e., a set of areas including the mediotemporal and orbitofrontal regions, may lead to dysfunction in the sense of smell and affect other brain regions that relate to the olfactory cortex by either afferent or efferent projections.

METHODS

The objective of this study was to assess sex-related differences in olfactory cortex volume using magnetic resonance imaging in individuals with mild cognitive impairment, probable dementia of the AD type and in healthy older adults, using the Mini-Mental Statement Examination score, years of education, and total intracranial volume as correction factors.

RESULTS

Atrophy of the olfactory cortex was observed in patients of both sexes with probable AD dementia. However, at the MCI stage, significant volumetric loss in the OC was detected in females only but not in males.

CONCLUSIONS

This finding indicates greater pathological effects in this region in females at an earlier disease stage than in males. This study suggests that OC volume loss occurs differently between the sexes in older adults, with volumetric loss being greater in females.

Detecting mild cognitive impairment by applying integrated random forest to finger tapping

Early detection of dementia is essential to reduce the decline in quality of life (QoL) and the increase in medical and nursing care costs associated with dementia in an aging society. In this study, we aimed to develop a simple screening test for mild cognitive impairment (MCI), a preliminary stage of dementia, by creating an analytical method to accurately detect MCI through finger-tapping measurement. We extracted 248 characteristics from the finger-tapping waveforms of 182 MCI patients and 352 normal controls, applying five conventional classification methods along with an improved Random Forest (RF) method proposed in this study (Integrated RF). In the proposed method, the RF classification model for the MCI and normal control groups is supplementally integrated with the RF classification model for the Alzheimer's disease and normal control groups to generate a new classification model. When comparing the discrimination accuracy of each method, the proposed method achieved the highest accuracy, with an F1-score of 0.795 (recall = 0.778 and precision = 0.814). These results demonstrate the potential of finger-tapping measurement as a highly accurate screening test for MCI.

"Advances in biomarker discovery and diagnostics for alzheimer's disease"

BACKGROUND

Alzheimer's disease (AD) is a progressive neurodegenerative disorder characterized by intracellular neurofibrillary tangles with tau protein and extracellular β-amyloid plaques. Early and accurate diagnosis is crucial for effective treatment and management.

OBJECTIVE

The purpose of this review is to investigate new technologies that improve diagnostic accuracy while looking at the current diagnostic criteria for AD, such as clinical evaluations, cognitive testing, and biomarker-based techniques.

METHODS

A thorough review of the literature was done in order to assess both conventional and contemporary diagnostic methods. Multimodal strategies integrating clinical, imaging, and biochemical evaluations were emphasised. The promise of current developments in biomarker discovery was also examined, including mass spectrometry and artificial intelligence.

RESULTS

Current diagnostic approaches include cerebrospinal fluid (CSF) biomarkers, imaging tools (MRI, PET), cognitive tests, and new blood-based markers. Integrating these technologies into multimodal diagnostic procedures enhances diagnostic accuracy and distinguishes dementia from other conditions. New technologies that hold promise for improving biomarker identification and diagnostic reliability include mass spectrometry and artificial intelligence.

CONCLUSION

Advancements in AD diagnostics underscore the need for accessible, minimally invasive, and cost-effective techniques to facilitate early detection and intervention. The integration of novel technologies with traditional methods may significantly enhance the accuracy and feasibility of AD diagnosis.

Single-nucleus RNA sequencing and network pharmacology reveal the mediation of fisetin on neuroinflammation in Alzheimer's disease

BACKGROUND

Alzheimer's Disease (AD) is a neurodegenerative disorder characterized by a progressive decline in cognitive function and memory. This study explores cellular subgroups in AD using single-nucleus RNA sequencing (snRNA-seq). It integrates the pharmacological network of traditional Chinese medicine (TCM) to identify potential therapeutic targets, providing a theoretical basis for the development of clinical AD.

METHODS

We obtained data information from the Gene Expression Omnibus (GEO) for snRNA-seq analysis. Enrichment and pseudotime analysis were performed to explore the functions and differentiation pathways of cellular subgroups. Cellular communication networks were mapped to reveal subgroup interactions. Additionally, a pharmacological network for AD was constructed using the TCM pharmacology database.

RESULTS

We identified several cell subgroups associated with AD pathology, contributing to disease progression in various ways. Notably, the TNC+ CD44+ astrocyte subgroup activated the I-kappa B kinase/ NF-κB signaling pathway, leading to increased expression of inflammatory cytokines. In the pharmacological network, fisetin was identified as a promising compound with the potential to bind to the CD44 protein, mitigating the inflammatory response and preventing further neuronal damage.

CONCLUSIONS

By exploring the ecological landscape of various cellular subgroups in AD and investigating the roles and mechanisms, combined with molecular docking and pharmacological network screening, our findings provide new insights and therapeutic possibilities for AD treatment.

Zebrafish metabolomics: a comprehensive approach to understanding health and disease

Zebrafish (Danio rerio) have become a valuable model in biomedical research due to their genetic similarity to humans, rapid development, and suitability for high-throughput studies. Metabolomic analyses in zebrafish provide critical insights into the biochemical pathways underlying health and disease. This review explores the applications of metabolomics in zebrafish research, highlighting its contributions to understanding embryonic development, tuberculosis, neurodegenerative disorders such as Alzheimer's disease, obesity-related metabolic dysfunction, and drug-induced toxicity through a thorough literature review. Zebrafish metabolomics reveals dynamic metabolite shifts during vertebrate development. In tuberculosis research, zebrafish models have helped identify metabolic biomarkers with potential translational relevance. Studies on Alzheimer's disease suggest that metabolomics can elucidate neuroprotective mechanisms, while investigations into obesity have provided insights into metabolic imbalances associated with kidney dysfunction. Furthermore, toxicometabolomic studies have demonstrated the utility of zebrafish in assessing drug-induced renal injury. Despite their advantages, zebrafish metabolomics faces challenges, including differences in metabolic rates compared to mammals, the need for standardized protocols, and limitations in metabolite database annotations. Nonetheless, integrating metabolomics with other omics approaches holds great promise for advancing disease research and paving the way for personalized medicine.

Computational Risk Stratification of Preclinical Alzheimer's in Younger Adults

Background: Alzheimer's disease (AD) is a progressive neurodegenerative disorder that often begins decades before clinical symptoms manifest. Early detection remains critical for effective intervention, particularly in younger adults, where biomarker deviations may signal pre-symptomatic risk. This research presents a computational modeling framework to predict cognitive impairment progression and stratify individuals into risk zones based on age-specific biomarker thresholds. Methods: The model integrates sigmoid-based data generation to simulate non-linear biomarker trajectories reflective of real-world disease progression. Core biomarkers-including cerebrospinal fluid (CSF) amyloid-beta 42 (Aβ42), amyloid positron emission tomography (amyloid PET), cerebrospinal fluid Tau protein (CSF Tau), and magnetic resonance imaging with fluorodeoxyglucose positron emission tomography (MRI FDG-PET)-were analyzed simultaneously to compute the cognitive impairment (CI) score of instances, dynamically adjusted for age. Higher CSF Aβ42 levels consistently demonstrated a protective effect, while elevated amyloid PET and Tau levels increased cognitive risk. Age-specific CI thresholds prevented the overestimation of risk in younger individuals and the underestimation in older cohorts. To demonstrate its applicability, we applied the full four-stage framework-comprising data aggregation and cleaning, sigmoid-based synthetic biomarker simulation with descriptive analysis, parameter accumulation modeling, and correlation-driven CI classification-on a curated dataset of 307 instances (ages 10-110) from Kaggle, the Alzheimer's Disease Neuroimaging Initiative (ANDI), and the Open Access Series of Imaging Studies (OASIS) to evaluate age-specific stratification of preclinical AD risk. Results: The study highlights the model's potential to identify individuals in risk zones from a pool of 150 instances, enabling targeted early interventions. Furthermore, the framework supports retrospective disease trajectory analysis, offering clinicians insights into optimal intervention windows even after symptom onset. Conclusions: Future work aims to validate the model using longitudinal, inclusive, real-world datasets and expand its predictive capacity through machine learning techniques and integrating genetic and lifestyle factors. Ultimately, this research contributes to advancing precision medicine approaches in Alzheimer's disease by providing a scalable computational tool for early risk assessment and intervention planning.

Leveraging transformers and explainable AI for Alzheimer's disease interpretability

Alzheimer's disease (AD) is a progressive brain ailment that causes memory loss, cognitive decline, and behavioral changes. It is quite concerning that one in nine adults over the age of 65 have AD. Currently there is almost no cure for AD except very few experimental treatments. However, early detection offers chances to take part in clinical trials or other investigations looking at potential new and effective Alzheimer's treatments. To detect Alzheimer's disease, brain scans such as computed tomography (CT), magnetic resonance imaging (MRI), or positron emission tomography (PET) can be performed. Many researches have been undertaken to use computer vision on MRI images, and their accuracy ranges from 80-90%, new computer vision algorithms and cutting-edge transformers have the potential to improve this performance.We utilize advanced transformers and computer vision algorithms to enhance diagnostic accuracy, achieving an impressive 99% accuracy in categorizing Alzheimer's disease stages through translating RNA text data and brain MRI images in near-real-time. We integrate the Local Interpretable Model-agnostic Explanations (LIME) explainable AI (XAI) technique to ensure the transformers' acceptance, reliability, and human interpretability. LIME helps identify crucial features in RNA sequences or specific areas in MRI images essential for diagnosing AD.

Agreement between early-phase amyloid-PET and pulsed arterial spin labeling in a memory clinic cohort

Relative cerebral blood flow (rCBF), assessed using pulsed arterial spin labeling (pASL) MRI, and the standardized uptake value ratio (SUVr) in early-phase amyloid-PET (ePET) are used as proxies for brain perfusion. These methods have the potential to streamline clinical workflows and reduce the burden on patients by eliminating the need for additional procedures. While both techniques have shown good agreement with the gold standard for glucose metabolism assessment, F-fluorodeoxyglucose-PET, a direct comparison between them has yet to be fully clarified. This retrospective study aimed to compare perfusion-like data from pASL (rCBF) and ePET (SUVr) in a memory clinic cohort. We included 46 subjects (69 ± 8 years; 37 women) from the Geneva Memory Center (cognitively impaired-CI n = 29; cognitively unimpaired-CU n = 17), with available pASL and ePET. We evaluated the association between rCBF and SUVr values across 18 cortical and subcortical regions using linear regression and the within-subject coefficient of variation (wsCV). Regional differences between CU and CI groups were assessed using linear regression model corrected for age. We observed significant association between rCBF and SUVr in precuneus (β = 0.69, wsCV = 16.9), angular gyrus (β = 0.64, wsCV = 19.4), and hippocampus (β = 0.23, wsCV = 16.1). Additionally, significant differences in rCBF between CU and CI were also observed in the posterior cingulate, precuneus, calcarine, hippocampus, and composite (p < 0.05), while SUVr showed significant differences only in the hippocampus. Our findings indicate weak to moderate local correlations between the two techniques. However, both exhibited differing regional perfusion levels in CU and CI groups, with rCBF showing more regional differences between cognitive stages in comparison with SUVr. KEY MESSAGES: rCBF is assessed through pASL MRI and SUVr through ePET, both serving as proxies of brain perfusion. Weak to moderate associations between rCBF and SUVr were found in a number of brain regions. rCBF and SUVr differences between cognitive stages were observed mostly in cortical and subcortical regions respectively. Both techniques were able to identify AD perfusion-like differences expected in cognitively impaired vs unimpaired.

Classification of the stages of Alzheimer's disease based on three-dimensional lightweight neural networks

Alzheimer's disease is a neurodegenerative disease that seriously threatens the life and health of the elderly. This study used three-dimensional lightweight neural networks to classify the stages of Alzheimer's disease and explore the relationship between the stages and the variations of brain tissue. The study used CAT12 to preprocess magnetic resonance images of the brain and got three kinds of preprocessed images: standardized images, segmented standardized gray matter images, and segmented standardized white matter images. The three kinds of images were used to train four kinds of three-dimensional lightweight neural networks respectively, and the evaluation metrics of the neural networks are calculated. The accuracies of the neural networks for classifying the stages of Alzheimer's disease (cognitively normal, mild cognitive impairment, Alzheimer's disease) in the study are above 96%, and the precisions and recalls of classifying the three stages are above 94%. The study found that for the classification of cognitively normal, the best classification results can be obtained by training with the segmented standardized gray matter images, and for mild cognitive impairment and Alzheimer's disease, the best classification results can be obtained by training with the standardized images. The study analyzed that in the process of cognitively normal to mild cognitive impairment, variations in the segmented standardized gray matter images are more obvious at the beginning, while variations in the segmented standardized white matter images are not obvious. As the disease progresses, variations in the segmented standardized white matter images tend to become more significant, and variations in the segmented standardized gray matter images and white matter images are both significant in the development of Alzheimer's disease.

Characterizing the heterogeneity of neurodegenerative diseases through EEG normative modeling

Neurodegenerative diseases like Parkinson's (PD) and Alzheimer's (AD) exhibit considerable heterogeneity of functional brain features within patients, complicating diagnosis and treatment. Here, we use electroencephalography (EEG) and normative modeling to investigate neurophysiological mechanisms underpinning this heterogeneity. Resting-state EEG data from 14 clinical units included healthy adults (n = 499) and patients with PD (n = 237) and AD (n = 197), aged over 40. Spectral and source connectivity analyses provided features for normative modeling, revealing significant, frequency-dependent EEG deviations with high heterogeneity in PD and AD. Around 30% of patients exhibited spectral deviations, while ~80% showed functional source connectivity deviations. Notably, the spatial overlap of deviant features did not exceed 60% for spectral and 25% for connectivity analysis. Furthermore, patient-specific deviations correlated with clinical measures, with greater deviations linked to worse UPDRS for PD (⍴ = 0.24, p = 0.025) and MMSE for AD (⍴ = -0.26, p = 0.01). These results suggest that EEG deviations could enrich individualized clinical assessment in Precision Neurology.

Total and Structural Carotid Artery Stiffness Are Associated With Cognitive Decline and Structural Brain Abnormalities Related to Alzheimer Disease and Alzheimer Disease-Related Dementias Pathology: The Multi-Ethnic Study of Atherosclerosis

BACKGROUND

Arterial stiffness is associated with pathological changes underlying Alzheimer disease and related dementias. Total pulse wave velocity can be subdivided into 2 main mechanisms: structural arterial stiffness (S-PWV) due to intrinsic remodeling of the artery wall and load-dependent arterial stiffness due to increased blood pressure.

METHODS AND RESULTS

In this prospective cohort study, MESA (Multi-Ethnic Study of Atherosclerosis) participants completed B-mode carotid ultrasounds from which carotid total pulse wave velocity was calculated. S-PWV was calculated by adjusting pulse wave velocity to 120/80 mmHg using a nonlinear pressure-diameter relationship, and load-dependent arterial stiffness was derived by subtracting S-PWV from total pulse wave velocity. Participants had repeated cognitive assessments with the Cognitive Abilities Screening Instrument, Digit Symbol Coding, and Digit Span combined into a global cognitive composite (N=2489). Brain magnetic resonance imaging was used to generate total gray matter volume (N=906), white matter hyperintensity volume (N=896), and total white matter fractional anisotropy (N=810). Multivariable linear fixed and mixed effects regression models related standardized pulse wave velocity components to neuroimaging and cognitive decline parameters, respectively. Greater S-PWV was associated with greater longitudinal cognitive decline in global cognitive composite score (β=-0.05, P=0.002) and subtests, whereas greater load-dependent arterial stiffness was not associated with longitudinal cognitive decline. Greater S-PWV was associated with lower gray matter volume (β=-3183.4, P=0.013) and higher log white matter hyperintensity volume (β=0.20, P<0.001), whereas load-dependent arterial stiffness was associated with lower total white matter fractional anisotropy (β=-0.004, P≤0.001).

CONCLUSIONS

Higher structural stiffness of the carotid artery is associated with cognitive decline, whereas both structural and load-dependent stiffness are associated with brain structural abnormalities common in Alzheimer disease-related dementias.

Predicting progression from MCI to dementia using cortical disarray measurement from diffusion MRI

BACKGROUND

This study evaluates the capability of cortical microstructural measures from diffusion magnetic resonance imaging (MRI) to predict progression from mild cognitive impairment (MCI) to dementia, compared to commonly used macrostructural measures. Identification of high-risk individuals can support both clinical practice and trials.

METHODS

Structural and diffusion MRI scans of 826 participants from the National Alzheimer's Coordinating Center (NACC) were analyzed to extract macrostructural measures and three minicolumn-related diffusivity metrics: AngleR, PerpPD+, and ParlPD. Kaplan-Meier survival analysis was used to investigate time to progression to dementia, with participants stratified by biomarker metrics.

RESULTS

Cortical diffusivity (PerpPD+ in medial-temporal and connected regions) outperformed hippocampal volume, cortical volume, and cortical thickness in Kaplan-Meier survival analysis, predicting faster conversion to dementia.

DISCUSSION

Cortical microstructural measures from diffusion MRI provide powerful biomarkers for predicting progression from MCI to dementia, offering enhanced prognostic capabilities that could support earlier intervention strategies in clinical practice and improve the power of clinical trials.

HIGHLIGHTS

Cortical minicolumn-related diffusivity metrics measure neurodegeneration. We compare the predictive value of magnetic resonance imaging (MRI) measures for mild cognitive impairment to dementia progression. Microstructural cortical disarray outperforms macrostructural markers. These results support using diffusion MRI biomarkers to identify and monitor at-risk patients.

Common Variants in PLXNA4 and Correlation to Neuroimaging Phenotypes in Healthy, Mild Cognitive Impairment, and Alzheimer's Disease Cohorts

A comprehensive genome-wide association study (GWAS) has validated the identification of the Plexin-A 4 (PLXNA4) gene as a novel susceptibility factor for Alzheimer's disease (AD). Nonetheless, the precise role of PLXNA4 gene polymorphisms in the pathophysiology of AD remains to be established. Consequently, this study is aimed at exploring the relationship between PLXNA4 gene polymorphisms and neuroimaging phenotypes intimately linked to AD. This study encompassed 812 subjects with PLXNA4 genotype data, procured from the Alzheimer's Disease Neuroimaging Initiative (ADNI) database. Employing a tagging strategy, we identified five common variant sites within the PLXNA4 gene and assessed their associations with glucose metabolism, atrophy in AD-related brain regions (including the medial temporal lobe, hippocampus, and parahippocampal gyrus), and intracerebral Aβ deposition. We conducted a comprehensive analysis using a multiple linear regression model, with neuroimaging phenotypes as the dependent variable and PLXNA4 gene polymorphisms as the independent variable while incorporating APOE e4 carrier status, education level, age, and gender as covariates. The subjects were stratified into three groups based on their disease status: the Alzheimer's disease (AD) group, the mild cognitive impairment (MCI) group, and the cognitively normal healthy control (CN) group. Within each group, we examined the associations between PLXNA4 gene polymorphisms and various neuroimaging phenotypes. Our study identified significant associations between the rs156676-A and rs78036292-G alleles and the baseline volumes of the anterior cingulate and middle temporal gyrus, respectively, across the entire population. After 1 year of follow-up, a significant correlation was observed between the rs6467431-G allele and accelerated volumetric atrophy of the parahippocampal gyrus in the overall population. Additionally, at the 2-year follow-up, significant correlations were observed between three PLXNA4 loci (rs1863015, rs6467431, rs67468325) and volumetric atrophy in the anterior cingulate, middle temporal gyrus, and hippocampus across the entire population. Specifically, the rs1863015-G allele notably accelerated atrophy of the left middle temporal gyrus and bilateral hippocampus, whereas the A alleles of rs6467431 and rs67468325 markedly accelerated atrophy specifically in the bilateral hippocampus. Subgroup analysis further validated these findings. Additionally, in the baseline CN group, the rs78036292 allele showed a significant correlation with intracerebral Aβ deposition, while in the 2-year follow-up CN group, rs67468325 was significantly associated with alterations in glucose metabolism rates in the right cingulate gyrus. Our findings indicate that PLXNA4 genotypes may modulate the development of AD through their regulation of intracerebral Aβ deposition. Additionally, PLXNA4 genotypes are strongly associated with AD-related brain atrophy and glucose metabolism, suggesting that they may alter susceptibility to AD by modulating neurodegenerative biomarkers.

Smart plant disease net: Adaptive Dense Hybrid Convolution network with attention mechanism for IoT-based plant disease detection by improved optimization approach

Plant diseases are rising nowadays. Plant diseases lead to high economic losses. Internet of Things (IoT) technology has found its application in various sectors. This led to the introduction of smart farming, in which IoT has been utilized to help identify the exact spot of the diseased affected region on the leaf from the vast farmland in a well-organized and automated manner. Thus, the main focus of this task is the introduction of a novel plant disease detection model that relies on IoT technology. The collected images are given to the Image Transmission phase. Here, the encryption task is performed by employing the Advanced Encryption Standard (AES) and also the decrypted plant images are fed to the pre-processing stage. The Mask Regions with Convolutional Neural Networks (R-CNN) are used to segment the pre-processed images. Then, the segmented images are given to the detection phase in which the Adaptive Dense Hybrid Convolution Network with Attention Mechanism (ADHCN-AM) approach is utilized to perform the detection of plant disease. From the ADHCN-AM, the final detected plant disease outcomes are obtained. Throughout the entire validation, the offered model shows 95% enhancement in terms of MCC showcasing its effectiveness over the existing approaches.

Causal relationship between hippocampal subfield volume and alzheimer's disease: a mendelian randomization study

BACKGROUND AND OBJECTIVE

Numerous studies suggest that the development of Alzheimer's Disease (AD) leads to a reduction in overall hippocampal volume. However, there is limited research exploring whether pre-morbid differences in hippocampal volume impact the risk of AD. This study aims to delve into the causal relationship between hippocampal subregional volume and AD using bidirectional Mendelian Randomization (MR) methods.

METHODS

We extracted 44 instrumental variables for hippocampal subregional volume from the GWAS Catalog, involving 21,282 European individuals. Data on Alzheimer's Disease were sourced from the Psychiatric Genomics Consortium, comprising 1,126,563 European individuals. Rigorous methods were employed to select instrumental variables, with the primary analysis conducted using the Inverse Variance Weighted method. Several sensitivity analyses included tests for heterogeneity, pleiotropy, and outliers. The obtained SNPs were mapped to genes for pathway enrichment analysis to explore the potential mechanisms underlying the regulation of hippocampal volume in Alzheimer's disease.

RESULTS

The study found significant causal associations between increased volume of the 5 hippocampal subfields with increased risk of AD. Conversely, increased Left hippocampus amygdala-transition-area volume was associated with reduced risk of AD. In reverse MR, AD was found to decrease the volume of 8 hippocampal subfields, while increasing the volume of the left hippocampal-fissure region. Amyloid-beta formation, leukocyte activation, and positive regulation of immune response mediated the changes in hippocampal subregional volume due to AD.

CONCLUSION

This MR study provides evidence that AD is causally related to hippocampal subfield volume, highlighting the roles of amyloid-beta formation and immune alterations in this context.

Expert Panel on Neurological Imaging, Soderlund KA, Austin MJ, Ben-Haim S, Chu S, Ivanidze J, Joshi P, Kalnins A, Kennedy M, Kulshreshtha A, Kuo PH, Masdeu JC, Nikumbh T, Soares BP, Thaker AA, Wang LL, Yasar S, Shih RY. ACR Appropriateness Criteria® Dementia: 2024 Update

Dementia is defined by significant chronic or acquired impairment in a single domain or loss of two or more cognitive functions by brain disease or injury. It is a common chronic syndrome in adults and constitutes the fifth leading cause of death in patients >65 years of age. Multiple etiologies of dementia exist, most notably Alzheimer disease, frontotemporal dementia, and dementia with Lewy bodies, as well as other neurologic diseases such as vascular dementia and normal pressure hydrocephalus. In addition to aiding clinicians in selecting the most appropriate imaging test for patients suspected of one of these dementia syndromes, this document highlights the most appropriate initial imaging tests for patients with suspected mild cognitive impairment and rapidly progressive dementia, as well as the most appropriate pre- and posttreatment imaging tests for patients undergoing therapy with antiamyloid monoclonal antibodies. The American College of Radiology Appropriateness Criteria are evidence-based guidelines for specific clinical conditions that are reviewed annually by a multidisciplinary expert panel. The guideline development and revision process support the systematic analysis of the medical literature from peer reviewed journals. Established methodology principles such as Grading of Recommendations Assessment, Development, and Evaluation or GRADE are adapted to evaluate the evidence. The RAND/UCLA Appropriateness Method User Manual provides the methodology to determine the appropriateness of imaging and treatment procedures for specific clinical scenarios. In those instances where peer reviewed literature is lacking or equivocal, experts may be the primary evidentiary source available to formulate a recommendation.

Pattern-VEP findings in individuals with mild cognitive impairment

BACKGROUND

Visual evoked potential (VEP) is a technique used to evaluate the electrical response of the brain to visual stimuli. This study aimed to examine neural transmission in the visual pathway by VEP test in individuals with mild cognitive impairment (MCI) and compare it with age-appropriate controls and also investigate for a correlation between VEP parameters and cognitive test domains.

METHODS

The groups consisted of 56 MCI and 50 healthy volunteers, aged 60-80 years, matched for age and education. Mini-Mental State Examination (MMSE) was applied to the participants for cognitive assessment. Patients were also subjected to other dementia screening tests and other treatable causes were excluded. In addition, groups were formed from those who completed the test. The pattern-reversal VEP method was used in this study.

RESULTS

The mean MMSE score in the MCI group was 21.42 ± 1.55 points. Our findings showed that individuals with MCI had a longer left P100 latency compared to controls (p = 0.027). In addition, right N75-P100 and right P100-N145 amplitudes of VEP parameters were found to be positively correlated with the recall test, which is one of the MMSE domains (p < 0.05, rho: 0.399,0.314).

CONCLUSION

The evidence provided by this study supports the possibility of using pattern VEP with clinical parameters to evaluate MCI patients. In addition, a positive correlation between interpeak amplitudes and the recall test highlights the importance of the VEP test in these patients.

Predicting the progression of MCI and Alzheimer's disease on structural brain integrity and other features with machine learning

Machine learning (ML) on structural MRI data shows high potential for classifying Alzheimer's disease (AD) progression, but the specific contribution of brain regions, demographics, and proteinopathy remains unclear. Using Alzheimer's Disease Neuroimaging Initiative (ADNI) data, we applied an extreme gradient-boosting algorithm and SHAP (SHapley Additive exPlanations) values to classify cognitively normal (CN) older adults, those with mild cognitive impairment (MCI) and AD dementia patients. Features included structural MRI, CSF status, demographics, and genetic data. Analyses comprised one cross-sectional multi-class classification (CN vs. MCI vs. AD dementia, n = 568) and two longitudinal binary-class classifications (CN-to-MCI converters vs. CN stable, n = 92; MCI-to-AD converters vs. MCI stable, n = 378). All classifications achieved 70-77% accuracy and 61-83% precision. Key features were CSF status, hippocampal volume, entorhinal thickness, and amygdala volume, with a clear dissociation: hippocampal properties contributed to the conversion to MCI, while the entorhinal cortex characterized the conversion to AD dementia. The findings highlight explainable, trajectory-specific insights into AD progression.

Key biomarkers in Alzheimer's disease: Insights for diagnosis and treatment strategies

Alzheimer's disease (AD) remains a significant global health challenge, characterized by its progressive neurodegeneration and cognitive decline. The urgent need for early diagnosis and effective treatment necessitates the identification of reliable biomarkers that can illuminate the underlying pathophysiology of AD. This review provides a comprehensive overview of the latest advancements in biomarker research, focusing on their applications in diagnosis, prognosis, and therapeutic development. We delve into the multifaceted landscape of AD biomarkers, encompassing molecular, imaging, and fluid-based markers. The integration of these biomarkers, including amyloid-β and tau proteins, neuroimaging modalities, cerebrospinal fluid analysis, and genetic risk factors, offers a more nuanced understanding of AD's complex etiology. By leveraging the power of precision medicine, biomarker-driven approaches can enable personalized treatment strategies and enhance diagnostic accuracy. Moreover, this review highlights the potential of biomarker research to accelerate drug discovery and development. By identifying novel therapeutic targets and monitoring disease progression, biomarkers can facilitate the evaluation of experimental treatments and ultimately improve patient outcomes. In conclusion, this review underscores the critical role of biomarkers in advancing our comprehension of AD and driving the development of effective interventions. By providing a comprehensive overview of the current state-of-the-art, this work aims to inspire future research and contribute to the goal of conquering AD.

Congruency between publicly available pictorial displays of medial temporal lobe atrophy

The medial temporal lobe atrophy (MTA) score is used for visual assessment of MTA on radiological images in suspected neurodegenerative dementia. Although volumetric tools are available, many radiologists still use visual scoring and compare to reference images. Numerous such example images are found online on educational websites and in scientific articles. The aim of this study was to compare congruencies between MTA scores of publicly available sample images with normalized heights and areas of relevant brain structures, measured in the same images.

METHOD

Systematic online searches yielded 148 individual sample images. The height and area of relevant brain structures were manually delineated, normalized, and compared with regard to the displayed MTA score.

RESULTS

The normalized heights and areas showed correlation with MTA but with considerable overlap between adjacent scores, especially when comparing heights. Also, displays of the MTA score were more consistent with the area of the temporal horn than with the hippocampal area.

CONCLUSION

There is considerable overlap between adjacent scores in publicly available pictorial displays of the MTA grading system. Insufficient congruency leads to confusion and reduces inter-rater reliability. We also found that publicly available images are more consistent with temporal horn area than the hippocampus, which means that ventricular size may bias the grading. This can impede relevant differential diagnostics, especially regarding normal pressure hydrocephalus. Here, we present lectotype images selected specifically with regard to the hippocampal area.

KEY POINTS

Question Overlap between publicly available example images of medial temporal atrophy causes confusion and limits reliability. Findings Available images are more consistent with ventricular dilatation than hippocampal atrophy; this article provides lectotype images selected specifically regarding the hippocampal area. Clinical relevance Visual assessment of medial temporal atrophy is used daily and worldwide in radiological examinations regarding suspected dementia. In clinical routine, many radiologists experience uncertainty, and hydrocephalus is often overlooked. This may be caused by insufficient congruency between educational sample images.

LGG-NeXt: A Next Generation CNN and Transformer Hybrid Model for the Diagnosis of Alzheimer's Disease Using 2D Structural MRI

Incurable Alzheimer's disease (AD) plagues many elderly people and families. It is important to accurately diagnose and predict it at an early stage. However, the existing methods have shortcomings, such as inability to learn local and global information and the inability to extract effective features. In this paper, we propose a lightweight classification network Local and Global Graph ConvNeXt. This model has a hybrid architecture of convolutional neural network and Transformers. We build the Global NeXt Block and the Local NeXt Block to extract the local and global features of the structural magnetic resonance imaging (sMRI). These two blocks are optimized by adding global multilayer perceptron and locally grouped attention, respectively. Then, the features are fed into the pixel graph neural network to aggregate the valid pixel features using mask attention. In addition, we decoupled the loss by category to optimize the calculation of the loss. This method was tested on slices of the processed sMRI datasets from ADNI and achieved excellent performance. Our model achieves 95.81% accuracy with fewer parameters and floating point operations per second (FLOPS) than other classical efficient models in the diagnosis of AD.

A foundation model for enhancing magnetic resonance images and downstream segmentation, registration and diagnostic tasks

In structural magnetic resonance (MR) imaging, motion artefacts, low resolution, imaging noise and variability in acquisition protocols frequently degrade image quality and confound downstream analyses. Here we report a foundation model for the motion correction, resolution enhancement, denoising and harmonization of MR images. Specifically, we trained a tissue-classification neural network to predict tissue labels, which are then leveraged by a 'tissue-aware' enhancement network to generate high-quality MR images. We validated the model's effectiveness on a large and diverse dataset comprising 2,448 deliberately corrupted images and 10,963 images spanning a wide age range (from foetuses to elderly individuals) acquired using a variety of clinical scanners across 19 public datasets. The model consistently outperformed state-of-the-art algorithms in improving the quality of MR images, handling pathological brains with multiple sclerosis or gliomas, generating 7-T-like images from 3 T scans and harmonizing images acquired from different scanners. The high-quality, high-resolution and harmonized images generated by the model can be used to enhance the performance of models for tissue segmentation, registration, diagnosis and other downstream tasks.

Multimodal magnetic resonance imaging analysis of early mild cognitive impairment

BackgroundEarly mild cognitive impairment (EMCI) represents a prodromal stage of dementia, and early detection is crucial for delaying dementia progression. However, accurately identifying its neuroimaging features remains challenging.ObjectiveTo comprehensively evaluate structural and functional neuroimaging changes in EMCI using multimodal magnetic resonance imaging (MRI) techniques.MethodsOne hundred and eleven participants were included from the Alzheimer's Disease Neuroimaging Initiative (ADNI): 36 with cognitively normal (CN), 30 with EMCI, 32 with late mild cognitive impairment (LMCI), and 13 with Alzheimer's disease (AD). FreeSurfer software was employed to segment hippocampal and amygdala subregions. The amplitude of low-frequency fluctuation (ALFF), fractional ALFF (fALFF), regional homogeneity (ReHo), and functional connectivity were processed using Data Processing & Analysis for Brain Imaging toolbox. Graph Theoretical Network Analysis toolbox was utilized to evaluate global functional network.ResultsThe volume of most hippocampal and amygdala subregions was decreased in AD group than those of EMCI group in structural MRI. Significant differences were found between EMCI and AD group in fALFF (right insula) and ReHo (bilateral caudate regions). EMCI group exhibited stronger functional connectivity between left hippocampus and right inferior temporal gyrus (compared to CN), left inferior temporal gyrus (compared to LMCI), and cerebellum crus 8 (compared to AD). EMCI group exhibited stronger connectivity between right hippocampus and left anterior cingulate gyrus compared to AD. Network metrics showed no significant differences among these groups, but all exhibited small-world properties.ConclusionsMultimodal MRI analysis revealed the neuroimaging characteristics of EMCI and promoted the understanding of the mechanisms underlying neuroimaging changes in EMCI.

Quantifying Brain Atrophy Using a CSF-Focused Segmentation Approach

Background and Purpose

Brain atrophy, characterized by sulcal widening and ventricular enlargement, is a hallmark of neurodegenerative diseases such as Alzheimer's disease. Visual assessments are subjective and variable, while automated methods struggle with subtle intensity differences and standardization, highlighting limitations in both approaches. This study aimed to develop and evaluate a novel method focusing on cerebrospinal fluid (CSF) regions by assessing segmentation accuracy, detecting stage-specific atrophy patterns, and testing generalizability to unstandardized datasets.

Methods

We utilized T1-weighted magnetic resonance imaging data from 3,315 participants from Samsung Medical Center and 1,439 participants from other hospitals. Segmentation accuracy was evaluated using the Dice similarity coefficient (DSC), and W-scores were calculated for each region of interest (ROI) to assess stage-specific atrophy patterns.

Results

The segmentation demonstrated high accuracy, with average DSC values exceeding 0.9 for ventricular and hippocampal regions and above 0.8 for cortical regions. Significant differences in W-scores were observed across cognitive stages (cognitively unimpaired, mild cognitive impairment, dementia of Alzheimer's type) for all ROIs (all, p<0.05). Similar trends were observed in the images from other hospitals, confirming the algorithm's generalizability to datasets without prior standardization.

Conclusions

This study demonstrates the robustness and clinical applicability of a novel CSF-focused segmentation method for assessing brain atrophy. The method provides a scalable and objective framework for evaluating structural changes across cognitive stages and holds potential for broader application in neurodegenerative disease research and clinical practice.

Pioneering new paths: the role of generative modelling in neurological disease research

Recently, deep generative modelling has become an increasingly powerful tool with seminal work in a myriad of disciplines. This powerful modelling approach is supposed to not only have the potential to solve current problems in the medical field but also to enable personalised precision medicine and revolutionise healthcare through applications such as digital twins of patients. Here, the core concepts of generative modelling and popular modelling approaches are first introduced to consider the potential based on methodological concepts for the generation of synthetic data and the ability to learn a representation of observed data. These potentials will be reviewed using current applications in neuroimaging for data synthesis and disease decomposition in Alzheimer's disease and multiple sclerosis. Finally, challenges for further research and applications will be discussed, including computational and data requirements, model evaluation, and potential privacy risks.

Boosting working memory in the elderly: driving prefrontal theta-gamma coupling via repeated neuromodulation

The escalating global burden of age-related neurodegenerative diseases and associated healthcare costs necessitates innovative interventions to stabilize or enhance cognitive functions. Deficits in working memory (WM) are linked to alterations in prefrontal theta-gamma cross-frequency coupling. Low-intensity transcranial alternating current stimulation (tACS) has emerged as a non-invasive, low-cost approach capable of modulating ongoing oscillations in targeted brain areas through entrainment. This study investigates the impact of multi-session peak-coupled theta-gamma cross-frequency tACS administered to the dorsolateral prefrontal cortex (DLPFC) on WM performance in older adults. In a randomized, sham-controlled, triple-blinded design, 77 participants underwent 16 stimulation sessions over six weeks while performing n-back tasks. Signal detection measures revealed increased 2-back sensitivity and robust modulations of response bias, indicating improved WM and decision-making adaptations, respectively. No effects were observed in the 1-back condition, emphasizing dependencies on cognitive load. Repeated tACS reinforces behavioral changes, indicated by increasing effect sizes. This study supports prior research correlating prefrontal theta-gamma coupling with WM processes and provides unique insights into the neurocognitive benefits of repeated tACS intervention. The well-tolerated and highly effective multi-session tACS intervention among the elderly underscores its therapeutic potential in vulnerable populations.

Association of item-level responses to cognitive function index with tau pathology and hippocampal volume in the A4 Study

INTRODUCTION

Alzheimer's disease (AD) has a long preclinical phase in which individuals may accumulate amyloid beta (Aβ) and tau pathology without noticeable cognitive impairment. Subjective cognitive impairment reports can provide early insights into cognitive decline.

METHODS

In the A4 Study, 339 cognitively unimpaired, Aβ-positive individuals underwent tau positron emission tomography imaging. Tau status was classified based on medial temporal lobe tau standardized uptake value ratios (tauMTL). Participants and study partners assessed cognitive changes using the 15-item Cognitive Function Index (CFI) questionnaire. We explored the relationship among tauMTL, hippocampal volume (HVa), and CFI reports.

RESULTS

Higher tauMTL was associated with participant-reported concerns about memory and navigation, and with study partner-reported difficulty remembering appointments. Lower HVa showed a marginal association with participant-reported driving difficulty.

DISCUSSION

These findings support the utility of participant- and study partner-reported concerns as early indicators of preclinical AD pathology, with potential value for early detection and trial enrichment strategies.

Highlights

Higher tau in the medial temporal lobe (tauMTL) was linked to participant-reported memory and orientation decline such as needing reminders or getting lost.Higher tauMTL was associated with increased memory-related concerns, such as needing help with appointments and asking repetitive questions.Lower hippocampal volume was associated with spatial memory and navigation such as driving difficulties and greater memory decline as reported by study partners.

Baseline Brain Volumes Predict Future Brain Atrophy in Mild Cognitive Impairment: A Tensor-based Morphometry Study of the Alzheimer Continuum

OBJECTIVE

Prognostic evaluation of patients with mild cognitive impairment (MCI) is of great importance, and magnetic resonance imaging, as a readily available modality, can play a pivotal role in this field.

METHODS

Using the Alzheimer Disease Neuroimaging Initiative database, we conducted a retrospective longitudinal study of the associations between volumetric brain magnetic resonance imaging and cognitive composite scores in all domains (memory, executive function, language, and visuospatial) with annual whole-brain atrophy based on tensor-based morphometry (TBM) scores among patients with MCI and healthy controls (HCs). The Reliable Change Index was further used to categorize patients into 2 groups including (1) patients with meaningful 1-year reliable cognitive changes [reliable change (RC) group] and (2) patients without (non-RC).

RESULTS

One hundred thirty-seven patients with MCI and 132 HCs were enrolled. The 2 groups showed no significant differences in age, sex, and apolipoprotein E4 expression (P > 0.05). Based on the TBM score, patients with MCI had more significant 1-year brain volume loss than HCs (P < 0.001). After multiple comparison corrections, the 1-year TBM atrophy score was positively correlated with baseline whole brain (P = 0.03), hippocampus (P < 0.0001), entorhinal (P < 0.0001), and middle temporal (P < 0.0001) volumes among MCI patients, indicating that lower volumes in these regions were associated with greater 1-year atrophy rates. Regression analyses showed a positive correlation between baseline and 1-year memory composite scores and annual brain atrophy rate in MCI patients (P = 0.01, 0.04), demonstrating that lower cognitive scores were associated with a greater annual atrophy rate. However, the correlations no longer held significance after correction for multiple comparison (P = 0.05, 0.17). MCI participants with RCs in language composite scores initially had significantly greater brain atrophy than those without (P = 0.03, corrected P = 0.06). However, TBM scores showed no significant differences between RC and non-RC groups for other composite scores (P > 0.05).

CONCLUSIONS

Lower baseline volumes in multiple brain regions of MCI are associated with greater annual brain volume loss based on TBM, suggesting TBM as a potential imaging marker for conventional volumetric studies in MCI. Further research is needed to explore the link between cognitive scores and the application of Reliable Change Index in TBM imaging across the Alzheimer disease spectrum.

Unravelling pathological ageing with brain age gap estimation in Alzheimer's disease, diabetes and schizophrenia

Brain age gap estimation (BrainAGE), the difference between predicted brain age and chronological age, might be a putative biomarker aiming to detect the transition from healthy to pathological brain ageing. The biomarker primarily models healthy ageing with machine learning models trained with structural magnetic resonance imaging (MRI) data. BrainAGE is expected to translate the deviations in neural ageing trajectory and has been shown to be increased in multiple pathologies, such as Alzheimer's disease (AD), schizophrenia and Type 2 diabetes (T2D). Thus, accelerated ageing seems to be a general feature of neuropathological processes. However, neurobiological constraints remain to be identified to provide specificity to this biomarker. Explainability might be the key to uncovering age predictions and understanding which brain regions lead to an elevated predicted age on a given pathology compared to healthy controls. This is highly relevant to understanding the similarities and differences in neurodegeneration in AD and T2D, which remains an outstanding biological question. Sensitivity maps explain models by computing the importance of each voxel on the final prediction, thereby contributing to the interpretability of deep learning approaches. This paper assesses whether sensitivity maps yield different results across three conditions related to pathological neural ageing: AD, schizophrenia and T2D. Five deep learning models were considered, each model trained with different MRI data types: minimally processed T1-weighted brain scans, and corresponding grey matter, white matter, cerebrospinal fluid tissue segmentation and deformation fields (after spatial normalization). Our results revealed an increased BrainAGE in all pathologies, with a different mean, which is the smallest in schizophrenia; this is in line with the observation that neural loss is secondary in this early-onset condition. Importantly, our findings suggest that the sensitivity, indexing regional weights, for all models varies with age. A set of regions were shown to yield statistical differences across conditions. These sensitivity results suggest that mechanisms of neurodegeneration are quite distinct in AD and T2D. For further validation, the sensitivity and the morphometric maps were compared. The findings outlined a high congruence between the sensitivity and morphometry maps for age and clinical group conditions. Our evidence outlines that the biological explanation of model predictions is vital in adding specificity to the BrainAGE and understanding the pathophysiology of chronic conditions affecting the brain.

Explainable Artificial Intelligence in Neuroimaging of Alzheimer's Disease

Alzheimer's disease (AD) remains a significant global health challenge, affecting millions worldwide and imposing substantial burdens on healthcare systems. Advances in artificial intelligence (AI), particularly in deep learning and machine learning, have revolutionized neuroimaging-based AD diagnosis. However, the complexity and lack of interpretability of these models limit their clinical applicability. Explainable Artificial Intelligence (XAI) addresses this challenge by providing insights into model decision-making, enhancing transparency, and fostering trust in AI-driven diagnostics. This review explores the role of XAI in AD neuroimaging, highlighting key techniques such as SHAP, LIME, Grad-CAM, and Layer-wise Relevance Propagation (LRP). We examine their applications in identifying critical biomarkers, tracking disease progression, and distinguishing AD stages using various imaging modalities, including MRI and PET. Additionally, we discuss current challenges, including dataset limitations, regulatory concerns, and standardization issues, and propose future research directions to improve XAI's integration into clinical practice. By bridging the gap between AI and clinical interpretability, XAI holds the potential to refine AD diagnostics, personalize treatment strategies, and advance neuroimaging-based research.

Neuromorphic deviations associated with transcriptomic expression and specific cell type in Alzheimer's disease

Alzheimer's disease (AD) is known to be associated with cortical anatomical atrophy and neurodegeneration across various brain regions. However, the relationships between brain structural changes in AD and gene expression remain unclear. We perform the morphometric similarity network (MSN) analysis to reveal the consistent cortical structural differences in individuals with AD compared to controls, and investigate the associations between brain-wide gene expression and morphometric changes. Furthermore, we identify abnormally MSN-related genes linked to specific cell types as the major contributors to transcriptomic relationships. MSN-related structural changes are located in the lateral ventral prefrontal cortex, temporal pole and medial prefrontal lobe, which are highly associated with the AD's cognitive decline. Analysis of gene expression shows the spatial correlations between AD-related genes and MSN differences. Examination of cell type-specific signature genes indicates that changes in microglia and neuronal transcriptional profiles largely contribute to AD-specific MSN differences. The study map the disease-specific structural alterations in AD down to the cellular level, offering a novel perspective on the linking surface-level changes to molecular mechanisms.

Associations Between Episodic Memory and Hippocampal Volume in Late Adulthood

Different tasks of episodic memory (EM) are only moderately correlated with each other. Furthermore, various EM tasks exhibit disproportional relationships with the hippocampus. This study examined the covariance structure of EM tasks and assessed whether this structure relates differently to hippocampal volume (HV) in a sample of 648 cognitively unimpaired older adults (mean age = 69.88). A confirmatory factor analysis (CFA) and linear regression models were used to test the associations between the observed factors of EM and HV. A model with three first-order subfactors (immediate verbal recall, delayed verbal recall, and visuospatial) derived from a second-order EM domain factor satisfied model fit (χ2 p value ≥ 0.05, CFI > 0.90, RMSEA < 0.08, SRMR < 0.08). Total, left, and right HV explained a similar amount of variance in all EM subfactors. CA1, CA3, subiculum, and entorhinal cortex volume were associated with all subfactors, while CA2 and dentate gyrus volume were not associated with EM. These results suggest that EM tasks are measuring the same construct, but different complex processes contribute to EM. Furthermore, HV accounted for a small portion of the variance in EM, suggesting that HV might not be a useful marker of EM in cognitively unimpaired older adults. Finally, this study provides evidence that various hippocampal subfield volumes may not be purely associated with any one aspect of EM processing.

Cerebral Amyloid Angiopathy: Clinical Presentation, Sequelae and Neuroimaging Features-An Update

The prevalence of cerebral amyloid angiopathy (CAA) has been shown to increase with age, with rates reported to be around 50-60% in individuals over 80 years old who have cognitive impairment. The disease often presents as spontaneous lobar intracerebral hemorrhage (ICH), which carries a high risk of recurrence, along with transient focal neurologic episodes (TFNE) and progressive cognitive decline, potentially leading to Alzheimer's disease (AD). In addition to ICH, neuroradiologic findings of CAA include cortical and subcortical microbleeds (MB), cortical subarachnoid hemorrhage (cSAH) and cortical superficial siderosis (cSS). Non-hemorrhagic pathologies include dilated perivascular spaces in the centrum semiovale and multiple hyperintense lesions on T2-weighted magnetic resonance imaging (MRI). A definitive diagnosis of CAA still requires histological confirmation. The Boston criteria allow for the diagnosis of a probable or possible CAA by considering specific neurological and MRI findings. The recent version, 2.0, which includes additional non-hemorrhagic MRI findings, increases sensitivity while maintaining the same specificity. The characteristic MRI findings of autoantibody-related CAA-related inflammation (CAA-ri) are similar to the so-called "amyloid related imaging abnormalities" (ARIA) observed with amyloid antibody therapies, presenting in two variants: (a) vasogenic edema and leptomeningeal effusions (ARIA-E) and (b) hemorrhagic lesions (ARIA-H). Clinical and MRI findings enable the diagnosis of a probable or possible CAA-ri, with biopsy remaining the gold standard for confirmation. In contrast to spontaneous CAA-ri, only about 20% of patients treated with monoclonal antibodies who show proven ARIA on MRI also experience clinical symptoms, including headache, confusion, other psychopathological abnormalities, visual disturbances, nausea and vomiting. Recent findings indicate that treatment should be continued in cases of mild ARIA, with ongoing MRI and clinical monitoring. This review offers a concise update on CAA and its associated consequences.

Convolutional Neural Networks for the segmentation of hippocampal structures in postmortem MRI scans

BACKGROUND

The hippocampus plays a crucial role in memory and is one of the first structures affected by Alzheimer's disease. Postmortem MRI offers a way to quantify the alterations by measuring the atrophy of the inner structures of the hippocampus. Unfortunately, the manual segmentation of hippocampal subregions required to carry out these measures is very time-consuming.

NEW METHOD

In this study, we explore the use of fully automated methods relying on state-of-the-art Deep Learning approaches to produce these annotations. More specifically, we propose a new segmentation framework made of a set of encoder-decoder blocks embedding self-attention mechanisms and atrous spatial pyramidal pooling to produce better maps of the hippocampus and identify four hippocampal regions: the dentate gyrus, the hippocampal head, the hippocampal body, and the hippocampal tail.

RESULTS

Trained using slices extracted from 15 postmortem T1-weighted, T2-weighted, and susceptibility-weighted MRI scans, our new approach produces hippocampus parcellations that are better aligned with the manually delineated parcellations provided by neuroradiologists.

COMPARISON WITH EXISTING METHODS

Four standard deep learning segmentation architectures: UNet, Double UNet, Attention UNet, and Multi-resolution UNet have been utilized for the qualitative and quantitative comparison of the proposed hippocampal region segmentation model.

CONCLUSIONS

Postmortem MRI serves as a highly valuable neuroimaging technique for examining the effects of neurodegenerative diseases on the intricate structures within the hippocampus. This study opens the way to large sample-size postmortem studies of the hippocampal substructures.

UK Biobank MRI data can power the development of generalizable brain clocks: A study of standard ML/DL methodologies and performance analysis on external databases

In this study, we present a comprehensive pipeline to train and compare a broad spectrum of machine learning and deep learning brain clocks, integrating diverse preprocessing strategies and correction terms. Our analysis also includes established methodologies which have shown success in prior UK Biobank-related studies. For our analysis we used T1-weighted MRI scans and processed de novo all images via FastSurfer, transforming them into a conformed space for deep learning and extracting image-derived phenotypes for our machine learning approaches. We rigorously evaluated these approaches both as robust age predictors for healthy individuals and as potential biomarkers for various neurodegenerative conditions, leveraging data from the UK Biobank, ADNI, and NACC datasets. To this end we designed a statistical framework to assess age prediction performance, the robustness of the prediction across cohort variability (database, machine type and ethnicity) and its potential as a biomarker for neurodegenerative conditions. Results demonstrate that highly accurate brain age models, typically utilising penalised linear machine learning models adjusted with Zhang's methodology, with mean absolute errors under 1 year in external validation, can be achieved while maintaining consistent prediction performance across different age brackets and subgroups (e.g., ethnicity and MRI machine/manufacturer). Additionally, these models show strong potential as biomarkers for neurodegenerative conditions, such as dementia, where brain age prediction achieved an AUROC of up to 0.90 in distinguishing healthy individuals from those with dementia.

Brain structural alterations in young women with premature ovarian insufficiency: Implications for dementia risk

INTRODUCTION

Premature ovarian insufficiency (POI), marked by ovarian function loss before age 40, is linked to a higher risk of dementia, including Alzheimer's disease (AD). However, the associated brain structural changes remain poorly understood.

METHODS

We analyzed T1-weighted and diffusion tensor imaging in 33 idiopathic POI women and 51 healthy controls, using voxel-based, surface-based morphometry, and network analyses to assess gray matter volume (GMV), cortical thickness, and brain connectivity.

RESULTS

Women with POI showed significant GMV and cortical thickness reductions in the frontal, parietal, and temporal regions (p < 0.05), alongside impaired connectivity with key regions such as the hippocampus, thalamus, and amygdala (p < 0.05). Younger POI subgroups exhibited changes in more widespread brain regions. In additionally, notable atrophy was observed in specific hippocampal and thalamic subregions in POI (p < 0.05).

DISCUSSION

This preliminary study suggests early neurodegenerative patterns in POI, potentially contributing to dementia risk. Further research is needed to explore the underlying mechanisms and potential interventions.

HIGHLIGHTS

We evaluated brain structural changes in participants with idiopathic premature ovarian insufficiency (POI). The observed brain alterations in POI participants closely resemble those seen in early dementia, including regions specifically associated with Alzheimer's disease (AD). These findings highlight the critical need for early interventions to reduce the long-term risks of cognitive impairment and dementia in women with POI.

Influence of lung function on macro- and micro-structural brain changes in mid- and late-life

INTRODUCTION

Lung function has been associated with cognitive decline and dementia, but the extent to which lung function impacts brain structural changes remains unclear. We aimed to investigate the association of lung function with structural macro- and micro-brain changes across mid- and late-life.

METHODS

The study included a total of 37 164 neurologic disorder-free participants aged 40-70 years from the UK Biobank, who underwent brain MRI scans 9 years after baseline. After 2.5 years, a subsample (n = 3895) underwent a second MRI scan. Lung function was assessed using a composite score based on forced expiratory volume in 1 second, forced vital capacity, and peak expiratory flow, and divided into tertiles (i.e., low, moderate, and high). Structural brain volumes (including total brain, gray matter, white matter, hippocampus, and white matter hyperintensities) and diffusion markers (fractional anisotropy [FA] and mean diffusivity [MD]) were assessed. Data were analyzed using linear regression and mixed-effects models.

RESULTS

Compared to high lung function, low lung function was associated with smaller total brain, gray matter, white matter, and hippocampal volume, as well as lower white matter integrity. Over the 2.5-year follow-up, low lung function was associated with reduced white matter and hippocampal volume, reduced FA, and increased white matter hyperintensity volume and MD. After stratification by age, the associations remained significant among adults aged 40-60 years and 60+ years.

CONCLUSION

Low lung function is associated with macro- and micro-structural brain changes involving both neurodegenerative and vascular pathologies. This association is significant in both mid- and late-life.

Evaluating Linear Heuristics for Ventricular Volume in Healthy Adults Using a Fully Automated Algorithm: Implications for Defining the Normal

BACKGROUND AND OBJECTIVES

Linear metrics for ventricular volume play a large role in the rapid, approximate evaluation of ventricular volume. In this article, we automatically extract linear measures of ventricular volume to explore their correlation with lateral ventricular volume (LVV) in the healthy adult population and comprehensively define normal values.

METHODS

We automatically extract Evans' ratio (ER), Frontal-Occipital Horn Ratio (FOHR), and anteroposterior lateral ventricle index (ALVI) from an open MRI data set of healthy adults ( https://brain-development.org/ixi-dataset/ ). Indices have been correlated with corresponding LVVs and lateral ventricular volumes divided by supratentorial brain volumes. Spearman rank correlation was used to compare strength of correlation.

RESULTS

ER shows correlation with lateral ventricle volume based on sex (r = 0.58; men, r = 0.65; women P < .001), including when controlling for supratentorial volume (r = 0.57; men, r = 0.63). ER did not profoundly correlate with age (r = 0.29, men; r = 0.35, women; P < .001) and seemed normally distributed around 0.25. ALVI showed strong correlation with LVV with only slight gender differences (r = 0.83, men; r = 0.84, women) and LVV to supratentorial cortical volume ratio (r = 0.9, men; r = 0.86, women). FOHR was also normally distributed around a value of 0.37 and showed moderate correlation with LVV (r = 0.68, men; r = 0.73, women) and LVV to supratentorial cortical volume ratio (r = 0.69, men; r = 0.74, women).

CONCLUSION

ALVI is a newer index with strong correlation with LVV and has strong potential for clinical use. Both FOHR and ER show moderate correlation with LVV. Reference values for linear estimates of ventricular volume may help clinicians better identify patients with pathological ventriculomegaly.

A Lightweight Deep Convolutional Neural Network Extracting Local and Global Contextual Features for the Classification of Alzheimer's Disease Using Structural MRI

Recent advancements in the classification of Alzheimer's disease have leveraged the automatic feature generation capability of convolutional neural networks (CNNs) using neuroimaging biomarkers. However, most of the existing CNN-based methods often disregard the local features of the brain data, which leads to a loss of subtle fine-grained features in the brain imaging data. Moreover, the existing CNN architectures, which mainly rely on global features, do not pay much attention to the discriminability of the extracted features for the task of classification of Alzheimer's disease. Moreover, the existing architectures often end up using a large number of parameters to enhance the richness of the extracted features. This paper proposes a novel lightweight deep CNN, which extracts local and global contextual features from the sagittal slices of structural MRI data and uses both of these two types of features for the classification of the disease. The main idea used in designing the proposed network is to process separately the local and global features by using modules that pay a special attention to extract local and global contextual features. The fused local and global contextual features are then used for the classification of Alzheimer's disease. The proposed network is tested for the binary and multiclass classifications of the disease using the MR images taken from the ADNI database. The proposed network is shown to provide a performance that is significantly higher than that provided by other existing state-of-the-art networks, yet using a number of parameters that is a small fraction of that used by the other schemes.

Domain-specific longitudinal associations between brain volume, white matter lesions, and cognitive function changes

Objectives

We investigated the domain-specific patterns of the association of segmental brain volume and white matter signal abnormality (WMSA) volume with longitudinal changes in cognitive function.

Methods

Participants from an institutional health check-up program who were aged >50 years, did not have a confirmed central nervous system disorder and underwent baseline and follow-up evaluations for cognitive function and brain MRI with an interval of at least 1 year were included. Cognitive function was assessed using the Consortium to Establish a Registry for Alzheimer's Disease-Korean version (CERAD-K) assessment battery. Performance changes in each cognitive domain were analyzed for associations with serial data of segmental brain volume and WMSA volume.

Results

A total of 190 subjects were included (115 [60.1 %] females, mean age 68.2 ± 8.2 years [range 50-82 years]). Declines in global cognition were associated with lower baseline (P=0.001) and decreasing volumes (P=0.001) of the hippocampus and amygdala and with increasing total WMSA volumes (P=0.008). Declines in the executive function domain were associated with lower baseline volumes of the hippocampus and amygdala (P = 0.018) and with increasing total WMSA volumes (P=0.015). Declines in the language function and the verbal learning domains were associated with lower baseline (P=0.009 and P=0.002, respectively) and decreasing volumes (P=0.008 and P=0.001, respectively) of the hippocampus and amygdala. Decline in the memory recall was associated with higher total WMSA volumes at baseline (P=0.014). Declines in the recognition memory domains were associated with lower baseline hippocampus and amygdala volume (P = 0.020) and with increases in total WMSA volumes (P=0.012).

Conclusions

The segmental brain volume and the WMSA volume parameters have domain-specific associations with longitudinal cognitive changes, which might reflect the different dependence on the brain reserve according to the cognitive domains.

Neurofind: using deep learning to make individualised inferences in brain-based disorders

Within precision psychiatry, there is a growing interest in normative models given their ability to parse heterogeneity. While they are intuitive and informative, the technical expertise and resources required to develop normative models may not be accessible to most researchers. Here we present Neurofind, a new freely available tool that bridges this gap by wrapping sound and previously tested methods on data harmonisation and advanced normative models into a web-based platform that requires minimal input from the user. We explain how Neurofind was developed, how to use the Neurofind website in four simple steps ( www.neurofind.ai ), and provide exemplar applications. Neurofind takes as input structural MRI images and outputs two main metrics derived from independent normative models: (1) Outlier Index Score, a deviation score from the normative brain morphology, and (2) Brain Age, the predicted age based on an individual's brain morphometry. The tool was trained on 3362 images of healthy controls aged 20-80 from publicly available datasets. The volume of 101 cortical and subcortical regions was extracted and modelled with an adversarial autoencoder for the Outlier index model and a support vector regression for the Brain age model. To illustrate potential applications, we applied Neurofind to 364 images from three independent datasets of patients diagnosed with Alzheimer's disease and schizophrenia. In Alzheimer's disease, 55.2% of patients had very extreme Outlier Index Scores, mostly driven by larger deviations in temporal-limbic structures and ventricles. Patients were also homogeneous in how they deviated from the norm. Conversely, only 30.1% of schizophrenia patients were extreme outliers, due to deviations in the hippocampus and pallidum, and patients tended to be more heterogeneous than controls. Both groups showed signs of accelerated brain ageing.

The TIE-93: a Facial Emotion Recognition Test Adapted for Culturally, Linguistically, and Educationally Diverse Alzheimer's Dementia Patients in France

OBJECTIVE

Emotion recognition tests are essential for differential diagnostics when assessing patients with Alzheimer's disease (AD) dementia. However, there remains a lack of emotion recognition tests appropriate for culturally and educationally diverse populations. The aim of this study was to develop an emotion recognition test (the TIE-93) appropriate for these populations. We then examined whether the TIE-93 could reduce emotion recognition performance differences between populations with a native French versus a culturally and educationally diverse background (participants who had immigrated to France). This was assessed by comparing performance between controls of each cultural group. We also assessed the effect of demographic variables on TIE-93 test performance and whether performance in an AD patient group was consistent with the research literature.

METHODS

Fifty-seven patients with AD dementia and 240 healthy controls, from native French and culturally and educationally diverse backgrounds, were included in the study. The TIE-93 is composed of eight panels with photos of actors displaying six basic emotions. Participants were asked to identify which of the six facial expressions displayed matched an oral description of a context.

RESULTS

When comparing French and culturally and educationally diverse controls, Quade's ANCOVA revealed that there remained an effect of culture and education on TIE-93 test performance. Nonetheless, while controlling for years of education, age, sex, and cultural group, patients with AD dementia scored significantly more poorly than controls, specifically for most negative emotions.

CONCLUSION

The TIE-93 represents a first step toward developing appropriate emotion recognition tests for culturally and educationally diverse populations.

Leveraging a Vision-Language Model with Natural Text Supervision for MRI Retrieval, Captioning, Classification, and Visual Question Answering

Large multimodal models are now extensively used worldwide, with the most powerful ones trained on massive, general-purpose datasets. Despite their rapid deployment, concerns persist regarding the quality and domain relevance of the training data, especially in radiology, medical research, and neuroscience. Additionally, healthcare data privacy is paramount when querying models trained on medical data, as is transparency regarding service hosting and data storage. So far, most deep learning algorithms in radiologic research are designed to perform a specific task (e.g., diagnostic classification) and cannot be prompted to perform multiple tasks using natural language. In this work, we introduce a framework based on vector retrieval and contrastive learning to efficiently learn visual brain MRI concepts via natural language supervision. We show how the method learns to identify factors that affect the brain in Alzheimer's disease (AD) via joint embedding and natural language supervision. First, we pre-train separate text and image encoders using self-supervised learning, and jointly fine-tune these encoders to develop a shared embedding space. We train our model to perform multiple tasks, including MRI retrieval, MRI captioning, and MRI classification. We show its versatility by developing a retrieval and re-ranking mechanism along with a transformer decoder for visual question answering.

Clinical Relevance

By learning a cross-modal embedding of radiologic features and text, our approach can learn to perform diagnostic and prognostic assessments in AD research as well as to assist practicing clinicians. Integrating medical imaging with clinical descriptions and text prompts, we aim to provide a general, versatile tool for detecting radiologic features described by text, offering a new approach to radiologic research.

Reducing inference cost of Alzheimer's disease identification using an uncertainty-aware ensemble of uni-modal and multi-modal learners

While multi-modal deep learning approaches trained using magnetic resonance imaging (MRI) and fluorodeoxyglucose positron emission tomography (FDG PET) data have shown promise in the accurate identification of Alzheimer's disease, their clinical applicability is hindered by the assumption that both modalities are always available during model inference. In practice, clinicians adjust diagnostic tests based on available information and specific clinical contexts. We propose a novel MRI- and FDG PET-based multi-modal deep learning approach that mimics clinical decision-making by incorporating uncertainty estimates of an MRI-based model (generated using Monte Carlo dropout and evidential deep learning) to determine the necessity of an FDG PET scan, and only inputting the FDG PET to a multi-modal model when required. This approach significantly reduces the reliance on FDG PET scans, which are costly and expose patients to radiation. Our approach reduces the need for FDG PET by up to 92% without compromising model performance, thus optimizing resource use and patient safety. Furthermore, using a global model explanation technique, we provide insights into how anatomical changes in brain regions-such as the entorhinal cortex, amygdala, and ventricles-can positively or negatively influence the need for FDG PET scans in alignment with clinical understanding of Alzheimer's disease.

Neuroimaging advances in neurocognitive disorders among HIV-infected individuals

Although combination antiretroviral therapy (cART) has been widely applied and effectively extends the lifespan of patients infected with human immunodeficiency virus (HIV), these patients remain at a substantially increased risk of developing neurocognitive impairment, commonly referred to as HIV-associated neurocognitive disorders (HAND). Magnetic resonance imaging (MRI) has emerged as an indispensable tool for characterizing the brain function and structure. In this review, we focus on the applications of various MRI-based neuroimaging techniques in individuals infected with HIV. Functional MRI, structural MRI, diffusion MRI, and quantitative MRI have all contributed to advancing our comprehension of the neurological alterations caused by HIV. It is hoped that more reliable evidence can be achieved to fully determine the driving factors of cognitive impairment in HIV through the combination of multi-modal MRI and the utilization of more advanced neuroimaging analysis methods.

Efficacy of acetylcholinesterase inhibitors on reducing hippocampal atrophy rate: a systematic review and meta-analysis

BACKGROUND

Neurodegenerative diseases (NDs) are conditions characterized by irreversible progressive degeneration to the nervous tissue and are usually associated with cognitive decline and functional deficits, especially in elderly. Acetylcholinesterase inhibitors (AChEIs) like donepezil, rivastigmine, and galantamine are commonly prescribed to alleviate cognitive symptoms associated with NDs. However, their long-term impact on slowing structural brain degeneration, particularly hippocampal atrophy, remains unclear.

OBJECTIVE

This systematic review and meta-analysis assess the efficacy of AChEIs in reducing hippocampal atrophy in patients with NDs or clinical syndromes that lead to cognitive decline.

METHODS

A systematic search of PubMed, Scopus, Web of Science, and Cochrane databases, since inception till 20th August 2024, identified randomized controlled trials (RCTs) and comparative studies that measured hippocampal volume changes in elderly patients with NDs and other clinical syndromes. Random effect model was employed to estimate the pooled atrophy rates. Subgroup analysis was conducted by disease, dosage, and side of the measurement.

RESULTS

From 5,943 initially screened studies, nine were included in the review, and six were analyzed in the meta-analysis, encompassing a total of 2,179 participants. The meta-analysis showed that donepezil at a 10 mg dose significantly reduced hippocampal atrophy compared to placebo (SMD = 0.44, 95% CI [0.08 to 0.81], p = 0.01), whereas the 5 mg dose showed no significant effect on hippocampal volume. Overall, pooled results favored donepezil in reducing hippocampal atrophy (SMD = 0.33, p = 0.04), indicating that higher doses are more effective. Among patients with mild cognitive impairment (MCI), both donepezil and vitamin E were associated with a significant reduction in hippocampal atrophy compared to placebo (SMD = 0.27, p = 0.01). In contrast, galantamine did not significantly reduce hippocampal atrophy in the overall analysis, but it was associated with reduced whole brain atrophy in APOE ε4 carriers. Further analysis revealed no significant difference in the reduction of right or left hippocampal atrophy in donepezil-treated patients. These findings suggest that donepezil, particularly at higher doses, may have a protective effect against hippocampal atrophy in patients with AD and MCI, while galantamine's effect may be more limited, especially in certain genetic subgroups.

CONCLUSION

Higher doses of donepezil (10 mg) significantly reduce hippocampal atrophy in Alzheimer's disease and mild cognitive impairment, suggesting potential neuroprotective effects. In contrast, lower doses (5 mg) and galantamine showed no significant impact on hippocampal volume, though galantamine reduced whole brain atrophy in APOE ε4 carriers. Dosage and genetic factors are crucial in determining the efficacy of acetylcholinesterase inhibitors in slowing neurodegeneration.

Transfer Learning and Neural Network-Based Approach on Structural MRI Data for Prediction and Classification of Alzheimer's Disease

Background: Alzheimer's disease (AD) is a neurodegenerative condition that has no definitive treatment, and its early diagnosis can help to prevent or slow down its progress. Structural magnetic resonance imaging (sMRI) and the progress of artificial intelligence (AI) have significant attention in AD detection. This study aims to differentiate AD from NC and distinguish between LMCI and EMCI from the other two classes. Another goal is the diagnostic performance (accuracy and AUC) of sMRI for predicting AD in its early stages. Methods: In this study, 398 participants were used from the ADNI and OASIS global database of sMRI including 98 individuals with AD, 102 with early mild cognitive impairment (EMCI), 98 with late mild cognitive impairment (LMCI), and 100 normal controls (NC). Results: The proposed model achieved high area under the curve (AUC) values and an accuracy of 99.7%, which is very remarkable for all four classes: NC vs. AD: AUC = [0.985], EMCI vs. NC: AUC = [0.961], LMCI vs. NC: AUC = [0.951], LMCI vs. AD: AUC = [0.989], and EMCI vs. LMCI: AUC = [1.000]. Conclusions: The results reveal that this model incorporates DenseNet169, transfer learning, and class decomposition to classify AD stages, particularly in differentiating EMCI from LMCI. The proposed model performs well with high accuracy and area under the curve for AD diagnostics at early stages. In addition, the accurate diagnosis of EMCI and LMCI can lead to early prediction of AD or prevention and slowing down of AD before its progress.