
Biomedical Signal Processing and Control, Journal Year: 2024, Volume and Issue: 103, P. 107422 - 107422
Published: Dec. 25, 2024
Language: Английский
Biomedical Signal Processing and Control, Journal Year: 2024, Volume and Issue: 103, P. 107422 - 107422
Published: Dec. 25, 2024
Language: Английский
Scientific Reports, Journal Year: 2025, Volume and Issue: 15(1)
Published: Feb. 17, 2025
Alzheimer's Disease (AD) is a progressive condition of neurological brain disorder recognized by symptoms such as dementia, memory loss, alterations in behaviour, and impaired reasoning abilities. Recently, many researchers have been working to develop an effective AD recognition system using deep learning (DL) based convolutional neural network (CNN) model aiming deploy the automatic medical image diagnosis system. The existing still facing difficulties achieving satisfactory performance terms accuracy efficiency because lack feature ineffectiveness. This study proposes lightweight Stacked Convolutional Neural Network with Channel Attention (SCCAN) for MRI on classification overcome challenges. In procedure, we sequentially integrate 5 CNN modules, which form stack generate hierarchical understanding features through multi-level extraction, effectively reducing noise enhancing weight's efficacy. then fed into channel attention module select practical dimension, facilitating selection influential features. . Consequently, exhibits reduced parameters, making it suitable training smaller datasets. Addressing class imbalance Kaggle dataset, balanced distribution samples among classes emphasized. Extensive experiments proposed ADNI1 Complete 1Yr 1.5T, Kaggle, OASIS-1 datasets showed 99.58%, 99.22%, 99.70% accuracy, respectively. model's high surpassed state-of-the-art (SOTA) models proved its excellence significant advancement images.
Language: Английский
Citations
3Image and Vision Computing, Journal Year: 2024, Volume and Issue: 144, P. 104967 - 104967
Published: Feb. 27, 2024
Language: Английский
Citations
13Biomedical Signal Processing and Control, Journal Year: 2024, Volume and Issue: 96, P. 106614 - 106614
Published: July 6, 2024
Language: Английский
Citations
13Computers in Biology and Medicine, Journal Year: 2025, Volume and Issue: 187, P. 109810 - 109810
Published: Feb. 11, 2025
Language: Английский
Citations
2Frontiers in Psychiatry, Journal Year: 2024, Volume and Issue: 15
Published: June 24, 2024
This study addresses the pervasive and debilitating impact of Alzheimer’s disease (AD) on individuals society, emphasizing crucial need for timely diagnosis. We present a multistage convolutional neural network (CNN)-based framework AD detection sub-classification using brain magnetic resonance imaging (MRI). After preprocessing, 26-layer CNN model was designed to differentiate between healthy patients with dementia. detecting dementia, reutilized concept transfer learning further subclassify dementia into mild, moderate, severe Leveraging frozen weights developed correlated medical images facilitated process sub-classifying classes. An online dataset is used verify performance proposed CNN-based framework. The approach yielded noteworthy accuracy 98.24% in identifying classes, whereas it achieved 99.70% subclassification. Another validate framework, resulting 100% performance. Comparative evaluations against pre-trained models current literature were also conducted, highlighting usefulness superiority presenting as robust effective subclassification method.
Language: Английский
Citations
7Knowledge-Based Systems, Journal Year: 2023, Volume and Issue: 281, P. 111064 - 111064
Published: Oct. 6, 2023
Language: Английский
Citations
17EClinicalMedicine, Journal Year: 2023, Volume and Issue: 64, P. 102247 - 102247
Published: Sept. 28, 2023
Alzheimer's disease (AD) is a heterogeneously progressive neurodegeneration disorder with varied rates of deterioration, either between subjects or within different stages certain subject. Estimating the course AD at early has treatment implications. We aimed to analyze progression identify distinct patterns in trajectory.We proposed deep learning model underlying trajectory from cognitively normal (CN) state mild cognitive impairment (MCI) dementia, by jointly predicting time-to-conversion and clustering out subgroups characterized comprehensive features as well rates. designed validated our on ADNI dataset (1370 participants). Prediction was used validate expression identified patterns. Causality further inferred using Mendelian randomization (MR) analysis. External validation performed AIBL (233 participants).The clustered significantly biomarkers The discovered also showed strong prediction ability, indicated hazard ratio (CN→MCI, HR = 3.51, p < 0.001; MCI→AD, 8.11, 0.001), C-Index 0.618; 0.718), AUC 3 years 0.802, 5 0.876; 0.914, 0.957). In external cohort, demonstrated competitive performance conversion time 0.693; 0.752). Moreover, suggestive associations CN→MCI/MCI→AD four/three SNPs were mediated MR analysis causal link MCI→AD first three years.Our identifies biologically clinically meaningful real-world data provides promising trajectory, which could promote understanding progression, facilitate clinical trial design, provide potential for decision-making.The National Key Research Development Program China, R&D Zhejiang, Nature Science Foundation China.
Language: Английский
Citations
13IEEE Access, Journal Year: 2023, Volume and Issue: 11, P. 128018 - 128031
Published: Jan. 1, 2023
The early identification and treatment of Mild Cognitive Impairment (MCI) play a crucial role in managing the risk Alzheimer's disease (AD). However, current methods for categorizing progressive MCI stable based on brain MRI scans have proven insufficient due to subtle nature features involved. This research aims improve effectiveness classification through utilization Deep Learning (DL) network. primary objective this work is feature representation more accurate classification. proposed model hybrid system that integrates three components: Swin Transformer, Dimension Centric Proximity Aware Attention Network (DCPAN), Age Deviation Factor (ADF). network achieves better results unique fusion approach combines global, local, proximal features, dimensional dependencies. It effectively fine-grained details with broader contextual information extract discriminative features. Experimental demonstrate network, achieving an accuracy 79.8%, precision 76.6%, recall 80.2%, F1-score 78.4% when evaluated ADNI dataset.
Language: Английский
Citations
13IEEE Access, Journal Year: 2024, Volume and Issue: 12, P. 58722 - 58739
Published: Jan. 1, 2024
Alzheimer's disease is a neurodegenerative causing memory loss and brain protein accumulation. Early diagnosis crucial for clinical trials patient care. Magnetic resonance imaging (MRI) methods have improved prognosis, but doctors need to interpret images proficiently. Deep learning technology has shown potential in detecting disease, the progresses slower early phases. A new dual-attention convolutional autoencoder model presented, offering detection abilities real-time use diagnosis. The study utilized two datasets: first ADNI dataset, which includes three classes (MCI, CN, AD), second Disease Neuroimaging Dataset, distinct (AD MCI). We analyze effectiveness of our proposed by evaluating key performance metrics such as accuracy, precision, sensitivity, specificity, F1 score, AUC score. In addition, we utilize cross-validation mean absolute error validate while also fine-tuning parameters. Based on experimental data, accurately detected with an accuracy 0.9902 ± 0.0139. results, demonstrates excellent compared existing described literature. mode achieves specificity 0.9882 0.0587, 0.9898 0.0865, 0.9912 0.0872 respectively. achieved score 0.9992 MCI 0.9919 AD class. Furthermore, method can enhance affordability diagnostics increase rate facilitating remote healthcare.
Language: Английский
Citations
5Artificial Intelligence Review, Journal Year: 2024, Volume and Issue: 57(10)
Published: Sept. 5, 2024
Language: Английский
Citations
4