Dense convolution-based attention network for Alzheimer’s disease classification

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

Published: Feb. 17, 2025

Recently, deep learning-based medical image classification models have made substantial advancements. However, many existing prioritize performance at the cost of efficiency, limiting their practicality in clinical use. Traditional Convolutional Neural Network (CNN)-based methods, Transformer-based and hybrid approaches combining these two struggle to balance model complexity. To achieve efficient predictions with a low parameter count, we propose DenseAttentionNetwork (DANet), lightweight for Alzheimer's disease detection 3D MRI images. DANet leverages dense connections linear attention mechanism enhance feature extraction capture long-range dependencies. Its architecture integrates convolutional layers localized global context, enabling multi-scale reuse across network. By replacing traditional self-attention parameter-efficient mechanism, overcomes some limitations standard self-attention. Extensive experiments multi-institutional datasets demonstrate that achieves best area under receiver operating characteristic curve (AUC), which underscores model's robustness effectiveness capturing relevant features while also attaining strong accuracy structure fewer parameters. Visualizations based on activation maps further verify ability highlight AD-relevant regions images, providing clinically interpretable insights into progression.

Language: Английский

---

Enhancing multi-class neurodegenerative disease classification using deep learning and explainable local interpretable model-agnostic explanations

Frontiers in Medicine, Journal Year: 2025, Volume and Issue: 12

Published: April 1, 2025

Alzheimer's disease (AD) and Parkinson's (PD) are two of the most prevalent neurodegenerative disorders, necessitating accurate diagnostic approaches for early detection effective management. This study introduces deep learning architectures, Residual-based Attention Convolutional Neural Network (RbACNN) Inverted (IRbACNN), designed to enhance medical image classification AD PD diagnosis. By integrating self-attention mechanisms, these models improve feature extraction, interpretability, address limitations traditional methods. Additionally, explainable AI (XAI) techniques incorporated provide model transparency clinical trust in automated diagnoses. Preprocessing steps such as histogram equalization batch creation applied optimize quality balance dataset. The proposed achieved an outstanding accuracy 99.92%. results demonstrate that combination with XAI, facilitate precise diagnosis, thereby contributing reducing global burden diseases.

Language: Английский

---

A Custom Convolutional Neural Network Model-Based Bioimaging Technique for Enhanced Accuracy of Alzheimer’s Disease Detection

Published: April 14, 2025

Language: Английский

---

Deep generative models for physiological signals: A systematic literature review

Artificial Intelligence in Medicine, Journal Year: 2025, Volume and Issue: unknown, P. 103127 - 103127

Published: April 1, 2025

Language: Английский

---

A Hybrid Deep Learning Approach for Improved Detection and Prediction of Brain Stroke

Applied Sciences, Journal Year: 2025, Volume and Issue: 15(9), P. 4639 - 4639

Published: April 22, 2025

Brain stroke is the leading cause of death and disability globally; hence, early identification prediction are critical for better patient outcomes. Traditional diagnostic procedures, such as manually interpreting clinical images, time consuming error prone. This research investigates use hybrid deep learning models, recurrent neural networks (RNNs), long short-term memory (LSTM), convolutional (CNNs), to improve accuracy. The current study compared performance these individual models with developed model on brain dataset. By merging we reached an overall accuracy 96% in identifying risk low, medium, or high. categorization may offer healthcare practitioners actionable insights by assisting them allowing make decisions. technique represents a substantial improvement preventive practices. model’s can further be tested more complicated demographic data that will help generalize real-world applications. Furthermore, combining this electronic health records (EHR) systems also assist identification, tailored therapies, improved management, enhancing outcomes lowering costs.

Language: Английский

---

Optimized Hybrid Deep Learning Model for Accurate Classification of Alzheimer’s Stages

Lecture notes on data engineering and communications technologies, Journal Year: 2025, Volume and Issue: unknown, P. 138 - 150

Published: Jan. 1, 2025

Language: Английский

---

A novel deep learning technique for multi classify Alzheimer disease: hyperparameter optimization technique

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

Published: April 24, 2025

A progressive brain disease that affects memory and cognitive function is Alzheimer’s (AD). To put therapies in place potentially slow the progression of AD, early diagnosis detection are essential. Early these phases enables activities, which essential for controlling disease. address issues with limited data computing resources, this work presents a novel deep-learning method based on using newly proposed hyperparameter optimization to identify hyperparameters ResNet152V2 model classifying AD more accurately. The compared state-of-the-art models divided into two categories: transfer learning classical showcase its effectiveness efficiency. This comparison four performance metrics: recall, precision, F1 score, accuracy. According experimental results, efficient effective various phases.

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

---

Advancements in deep learning for early diagnosis of Alzheimer’s disease using multimodal neuroimaging: challenges and future directions

Frontiers in Neuroinformatics, Journal Year: 2025, Volume and Issue: 19

Published: May 2, 2025

Introduction Alzheimer’s disease is a progressive neurodegenerative disorder challenging early diagnosis and treatment. Recent advancements in deep learning algorithms applied to multimodal brain imaging offer promising solutions for improving diagnostic accuracy predicting progression. Method This narrative review synthesizes current literature on applications using neuroimaging. The process involved comprehensive search of relevant databases (PubMed, Embase, Google Scholar ClinicalTrials.gov ), selection pertinent studies, critical analysis findings. We employed best-evidence approach, prioritizing high-quality studies identifying consistent patterns across the literature. Results Deep architectures, including convolutional neural networks, recurrent transformer-based models, have shown remarkable potential analyzing neuroimaging data. These models can effectively structural functional modalities, extracting features associated with pathology. Integration multiple modalities has demonstrated improved compared single-modality approaches. also promise predictive modeling, biomarkers forecasting Discussion While approaches show great potential, several challenges remain. Data heterogeneity, small sample sizes, limited generalizability diverse populations are significant hurdles. clinical translation these requires careful consideration interpretability, transparency, ethical implications. future AI neurodiagnostics looks promising, personalized treatment strategies.

Language: Английский

---

Refined causal graph structure learning via curvature for brain disease classification

Artificial Intelligence Review, Journal Year: 2025, Volume and Issue: 58(8)

Published: May 3, 2025

Language: Английский

---