Comparison of Deep Learning and Traditional Machine Learning Models for Predicting Mild Cognitive Impairment Using Plasma Proteomic Biomarkers

International Journal of Molecular Sciences, Journal Year: 2025, Volume and Issue: 26(6), P. 2428 - 2428

Published: March 8, 2025

Mild cognitive impairment (MCI) is a clinical condition characterized by decline in ability and progression of impairment. It often considered transitional stage between normal aging Alzheimer’s disease (AD). This study aimed to compare deep learning (DL) traditional machine (ML) methods predicting MCI using plasma proteomic biomarkers. A total 239 adults were selected from the Disease Neuroimaging Initiative (ADNI) cohort along with pool 146 We evaluated seven ML models (support vector machines (SVMs), logistic regression (LR), naïve Bayes (NB), random forest (RF), k-nearest neighbor (KNN), gradient boosting (GBM), extreme (XGBoost)) six variations neural network (DNN) model—the DL model H2O package. Least Absolute Shrinkage Selection Operator (LASSO) 35 biomarkers pool. Based on grid search, DNN an activation function “Rectifier With Dropout” 2 layers 32 revealed best highest accuracy 0.995 F1 Score 0.996, while among methods, XGBoost was 0.986 0.985. Several correlated APOE-ε4 genotype, polygenic hazard score (PHS), three cerebrospinal fluid (Aβ42, tTau, pTau). Bioinformatics analysis Gene Ontology (GO) Kyoto Encyclopedia Genes Genomes (KEGG) several molecular functions pathways associated biomarkers, including cytokine-cytokine receptor interaction, cholesterol metabolism, regulation lipid localization. The results showed that may represent promising tool prediction MCI. These help early diagnosis, prognostic risk stratification, treatment interventions for individuals at

Language: Английский

---

Multivariate longitudinal clustering reveals neuropsychological factors as dementia predictors in an Alzheimer’s disease progression study

BioData Mining, Journal Year: 2025, Volume and Issue: 18(1)

Published: March 28, 2025

Dementia due to Alzheimer's disease (AD) is a multifaceted neurodegenerative disorder characterized by various cognitive and behavioral decline factors. In this work, we propose an extension of the traditional k-means clustering for multivariate time series data cluster joint trajectories different features describing progression over time. The algorithm here enables analysis longitudinal explore co-occurring trajectory factors among markers indicative in individuals participating AD study. By examining how multiple variables co-vary evolve together, identify distinct subgroups within cohort based on their trajectories. Our method enhances understanding individual development across dimensions provides deeper medical insights into decline. addition, proposed also able make selection most significant separating clusters considering This process, together with preliminary pre-processing OASIS-3 dataset, reveals important role some neuropsychological particular, has identified profile compatible syndrome known as Mild Behavioral Impairment (MBI), displaying manifestations that may precede symptoms typically observed patients. findings underscore importance clinical modeling, ultimately supporting more effective individualized patient management strategies.

Language: Английский

---

Hybrid of DSR-GAN and CNN for Alzheimer disease detection based on MRI images

Scientific Reports, Journal Year: 2025, Volume and Issue: 15(1)

Published: April 13, 2025

In this paper, we propose a deep super-resolution generative adversarial network (DSR-GAN) combined with convolutional neural (CNN) model designed to classify four stages of Alzheimer's disease (AD): Mild Dementia (MD), Moderate (MOD), Non-Demented (ND), and Very (VMD). The proposed DSR-GAN is implemented using PyTorch library uses dataset 6,400 MRI images. A (SR) technique applied enhance the clarity detail images, allowing refine particular image features. CNN undergoes hyperparameter optimization incorporates data augmentation strategies maximize its efficiency. normalized error matrix area under ROC curve are used experimentally evaluate CNN's performance which achieved testing accuracy 99.22%, an 100%, rate 0.0516. Also, assessed three different metrics: structural similarity index measure (SSIM), peak signal-to-noise ratio (PSNR), multi-scale (MS-SSIM). SSIM score 0.847, while PSNR MS-SSIM percentage 29.30 dB 96.39%, respectively. combination models provides rapid precise method distinguish between various disease, potentially aiding professionals in screening AD cases.

Language: Английский

---

An efficient method for early Alzheimer’s disease detection based on MRI images using deep convolutional neural networks

Frontiers in Artificial Intelligence, Journal Year: 2025, Volume and Issue: 8

Published: April 29, 2025

Alzheimer's disease (AD) is a progressive, incurable neurological disorder that leads to gradual decline in cognitive abilities. Early detection vital for alleviating symptoms and improving patient quality of life. With shortage medical experts, automated diagnostic systems are increasingly crucial healthcare, reducing the burden on providers enhancing accuracy. AD remains global health challenge, requiring effective early strategies prevent its progression facilitate timely intervention. In this study, deep convolutional neural network (CNN) architecture proposed classification. The model, consisting 6,026,324 parameters, uses three distinct branches with varying lengths kernel sizes improve feature extraction. OASIS dataset used includes 80,000 MRI images sourced from Kaggle, categorized into four classes: non-demented (67,200 images), very mild demented (13,700 (5,200 moderate (488 images). To address imbalance, data augmentation technique was applied. model achieved remarkable 99.68% accuracy distinguishing between stages Alzheimer's: Non-Dementia, Very Mild Dementia, Moderate Dementia. This high highlights model's potential real-time analysis diagnosis AD, offering promising tool healthcare professionals.

Language: Английский

---

Comparative Evaluation of Deep Learning Models in Alzheimer’s Disease Diagnosis

Procedia Computer Science, Journal Year: 2025, Volume and Issue: 258, P. 2352 - 2361

Published: Jan. 1, 2025

Language: Английский

---

Investigating Modifiable Risk Factors Across Dementia Subtypes: Insights from the UK Biobank

Biomedicines, Journal Year: 2024, Volume and Issue: 12(9), P. 1967 - 1967

Published: Aug. 31, 2024

This study investigates the relationship between modifiable risk factors and dementia subtypes using data from 460,799 participants in UK Biobank. Utilizing univariate Cox proportional hazards regression models, we examined associations 83 risks of all-cause (ACD), Alzheimer’s disease (AD), vascular (VD). Composite scores for different domains were generated by aggregating associated with ACD, AD, VD, respectively, their joint assessed multivariable models. Additionally, population attributable fractions (PAF) utilized to estimate potential impact eliminating adverse characteristics domains. Our findings revealed that an unfavorable medical history significantly increased VD (hazard ratios (HR) = 1.88, 95% confidence interval (95% CI): 1.74–2.03, p < 0.001; HR 1.80, CI: 1.54–2.10, 2.39, 2.10–2.71, 0.001, respectively). Blood markers (PAF 12.1%; 17.4%) emerged as most important domain preventing ACD while psychiatric 18.3%) AD. underscores its through targeted interventions factors. The distinct insights provided PAF emphasize importance considering both strength population-level prevention strategies. research provides valuable guidance developing effective public health aimed at reducing burden dementia, representing a significant advancement field.

Language: Английский

---

Navigating the Alzheimer’s Biomarker Landscape: A Comprehensive Analysis of Fluid-Based Diagnostics

Cells, Journal Year: 2024, Volume and Issue: 13(22), P. 1901 - 1901

Published: Nov. 18, 2024

Alzheimer's disease (AD) affects a significant portion of the aging population, presenting serious challenge due to limited availability effective therapies during its progression. The advances rapidly, underscoring need for early diagnosis and application preventative measures. Current diagnostic methods AD are often expensive invasive, restricting access general public. One potential solution is use biomarkers, which can facilitate detection treatment through objective, non-invasive, cost-effective evaluations AD. This review critically investigates function role biofluid biomarkers in detecting AD, with specific focus on cerebrospinal fluid (CSF), blood-based, saliva biomarkers.

Language: Английский

---

Introduction to Alzheimer's Disease and Biomarkers

Advances in bioinformatics and biomedical engineering book series, Journal Year: 2024, Volume and Issue: unknown, P. 95 - 122

Published: Nov. 1, 2024

Alzheimer's disease, the leading cause of dementia affecting 50-60% cases globally, manifests initially with cognitive impairments and progresses neurodegeneration, brain inflammation, atrophy. Early diagnosis treatment rely on identifying biomarkers, which can be invasive or non-invasive, categorized as diagnostic, prognostic, predictive, pharmacodynamic/response, susceptibility/risk, monitoring, safety biomarkers. They include amyloid Aβ plaques, Brain derived Neurotrophic factor (BDNF), pro-NGF, tau protein (t-protein) neurofibrillary tangles, apolipoprotein, novel markers in CSF, blood, urine, lipid profiles. Challenges encompass lumbar puncture, multifactorial progression, early biomarker inexplicability, pathophysiological understanding gaps.Advancement Theranostics approach is explained AD patients. Later this study, we analyzed these biomarkers using integrative approaches deep generative models focusing detecting anomalies structure, biological functions, abnormal metabolite concentrations, misfolded proteins.

Language: Английский

---

Dynamic changes and prognostic value of glutathione S-transferase alpha in mild cognitive impairment and Alzheimer’s disease

Frontiers in Aging Neuroscience, Journal Year: 2024, Volume and Issue: 16

Published: Dec. 23, 2024

Objectives Glutathione S-transferase alpha (GSTα) is an important antioxidant enzyme closely associated with the onset and progression of neurodegenerative diseases. The alterations in GSTα protein levels Alzheimer’s disease their impact on cognitive abilities remain unclear. Thus, investigating fluctuations mild impairment (MCI) (AD) essential. Methods DATA were enrolled from Disease Neuroimaging Initiative (ADNI) database, we studied healthy individuals (as controls, a total 54), patients (345), (96) A one-year follow-up was conducted to collect data dynamic changes plasma primary information data, analyze correlation between before after function its predictive value. Results Plasma significantly lower AD group than CN (0.94 vs1.05, p = 0.04) MCI vs1.09, &lt; 0.001). level positively correlated altered MMSE ( r 0.09, 0.04). AUC (95% CI) area under prediction curve for 0.63 (0.54–0.71), 0.02, 0.74 (0.69–0.80), 0.001. At same time, plotted ROC curves difference change 1 year follow-up. results showed that 0.76 (0.696–0.83), 0.001, 0.75 Conclusion findings study indicated notable differences among those period. Furthermore, positive observed GST αprotein decline both baseline function. This suggests could potentially act as biomarker AD, offering fresh insights early detection intervention strategies.

Language: Английский

---