A review of artificial intelligence methods for Alzheimer's disease diagnosis: Insights from neuroimaging to sensor data analysis

Biomedical Signal Processing and Control, Год журнала: 2024, Номер 92, С. 106023 - 106023

Опубликована: Фев. 14, 2024

Язык: Английский

---

A Transfer Learning Approach: Early Prediction of Alzheimer’s Disease on US Healthy Aging Dataset

Mathematics, Год журнала: 2024, Номер 12(14), С. 2204 - 2204

Опубликована: Июль 13, 2024

Alzheimer’s disease (AD) is a growing public health crisis, very global concern, and an irreversible progressive neurodegenerative disorder of the brain for which there still no cure. Globally, it accounts 60–80% dementia cases, thereby raising need accurate effective early classification. The proposed work used healthy aging dataset from USA focused on three transfer learning approaches: VGG16, VGG19, Alex Net. This leveraged how convolutional model pooling layers to improve reduce overfitting, despite challenges in training numerical dataset. VGG was preferably chosen as hidden layer has more diverse, deeper, simpler architecture with better performance when dealing larger datasets. It consumes less memory time. A comparative analysis performed using machine neural network algorithm techniques. Performance metrics such accuracy, error rate, precision, recall, F1 score, sensitivity, specificity, kappa statistics, ROC, RMSE were experimented compared. accuracy 100% VGG16 VGG19 98.20% precision 99.9% 96.6% Net; recall values all cases sensitivity metric 96.8% 97.9% 98.7% Net, outperformed compared existing approaches classification disease. research contributes advancement predictive knowledge, leading future empirical evaluation, experimentation, testing biomedical field.

Язык: Английский

---

Deep learning approaches for seizure video analysis: A review

Epilepsy & Behavior, Год журнала: 2024, Номер 154, С. 109735 - 109735

Опубликована: Март 23, 2024

Seizure events can manifest as transient disruptions in the control of movements which may be organized distinct behavioral sequences, accompanied or not by other observable features such altered facial expressions. The analysis these clinical signs, referred to semiology, is subject observer variations when specialists evaluate video-recorded setting. To enhance accuracy and consistency evaluations, computer-aided video seizures has emerged a natural avenue. In field medical applications, deep learning computer vision approaches have driven substantial advancements. Historically, been used for disease detection, classification, prediction using diagnostic data; however, there limited exploration their application evaluating video-based motion detection epileptology While vision-based technologies do aim replace expertise, they significantly contribute decision-making patient care providing quantitative evidence decision support. Behavior monitoring tools offer several advantages objective information, detecting challenging-to-observe events, reducing documentation efforts, extending assessment capabilities areas with expertise. main applications could (1) improved seizure methods; (2) refined semiology predicting type cerebral localization. this paper, we detail foundation systems videos, highlighting success analysis, focusing on work published last 7 years. We systematically present methods indicate how adoption recordings approached. Additionally, illustrate existing interconnected through an integrated system analysis. Each module customized adapting more accurate robust evolve. Finally, discuss challenges research directions future studies.

Язык: Английский

---

New Convolutional Neural Network and Graph Convolutional Network-Based Architecture for AI Applications in Alzheimer’s Disease and Dementia-Stage Classification

AI, Год журнала: 2024, Номер 5(1), С. 342 - 363

Опубликована: Фев. 1, 2024

Neuroimaging experts in biotech industries can benefit from using cutting-edge artificial intelligence techniques for Alzheimer’s disease (AD)- and dementia-stage prediction, even though it is difficult to anticipate the precise stage of dementia AD. Therefore, we propose a cutting-edge, computer-assisted method based on an advanced deep learning algorithm differentiate between people with varying degrees dementia, including healthy, very mild moderate classes. In this paper, four separate models were developed classifying different stages: convolutional neural networks (CNNs) built scratch, pre-trained VGG16 additional layers, graph (GCNs), CNN-GCN models. The CNNs implemented, then flattened layer output was fed GCN classifier, resulting proposed architecture. A total 6400 whole-brain magnetic resonance imaging scans obtained Disease Initiative database train evaluate methods. We applied 5-fold cross-validation (CV) technique all presented results best fold out five folds assessing performance study. Hence, CV, above-mentioned achieved overall accuracy 43.83%, 71.17%, 99.06%, 100%, respectively. model, particular, demonstrates excellent stages dementia. Understanding assist industry researchers uncovering molecular markers pathways connected each stage.

Язык: Английский

---

Machine Learning Accelerates De Novo Design of Antimicrobial Peptides

Interdisciplinary Sciences Computational Life Sciences, Год журнала: 2024, Номер 16(2), С. 392 - 403

Опубликована: Фев. 28, 2024

Язык: Английский

---

Early Prediction of Dementia Using Feature Extraction Battery (FEB) and Optimized Support Vector Machine (SVM) for Classification

Biomedicines, Год журнала: 2023, Номер 11(2), С. 439 - 439

Опубликована: Фев. 2, 2023

Dementia is a cognitive disorder that mainly targets older adults. At present, dementia has no cure or prevention available. Scientists found symptoms might emerge as early ten years before the onset of real disease. As result, machine learning (ML) scientists developed various techniques for prediction using symptoms. However, these methods have fundamental limitations, such low accuracy and bias in models. To resolve issue proposed ML model, we deployed adaptive synthetic sampling (ADASYN) technique, to improve accuracy, novel feature extraction techniques, namely, battery (FEB) optimized support vector (SVM) radical basis function (rbf) classification The hyperparameters SVM are calibrated by employing grid search approach. It evident from experimental results newly pr oposed model (FEB-SVM) improves conventional 6%. obtained 98.28% on training data testing 93.92%. Along with precision 91.80%, recall 86.59, F1-score 89.12%, Matthew's correlation coefficient (MCC) 0.4987. Moreover, outperforms 12 state-of-the-art models researchers recently presented prediction.

Язык: Английский

---

New Horizons in artificial intelligence in the healthcare of older people

Age and Ageing, Год журнала: 2023, Номер 52(12)

Опубликована: Дек. 1, 2023

Artificial intelligence (AI) in healthcare describes algorithm-based computational techniques which manage and analyse large datasets to make inferences predictions. There are many potential applications of AI the care older people, from clinical decision support systems that can identification delirium records wearable devices predict risk a fall. We held four meetings clinicians researchers. Three priority areas were identified for application people. These included: monitoring early diagnosis disease, stratified coordination between providers. However, also highlighted concerns may exacerbate health inequity people through bias within models, lack external validation amongst infringements on privacy autonomy, insufficient transparency models safeguarding errors. Creating effective interventions requires person-centred approach account needs as well sufficient technological governance meet standards generalisability, effectiveness. Education patients is needed ensure appropriate use technologies, with investment infrastructure required equity access.

Язык: Английский

---

Breaking barriers: a statistical and machine learning-based hybrid system for predicting dementia

Frontiers in Bioengineering and Biotechnology, Год журнала: 2024, Номер 11

Опубликована: Янв. 8, 2024

Dementia is a condition (a collection of related signs and symptoms) that causes continuing deterioration in cognitive function, millions people are impacted by dementia every year as the world population continues to rise. Conventional approaches for determining rely primarily on clinical examinations, analyzing medical records, administering neuropsychological testing. However, these methods time-consuming costly terms treatment. Therefore, this study aims present noninvasive method early prediction so preventive steps should be taken avoid dementia.

Язык: Английский

---

Modeling Soil Behavior with Machine Learning: Static and Cyclic Properties of High Plasticity Clays Treated with Lime and Fly Ash

Buildings, Год журнала: 2025, Номер 15(2), С. 288 - 288

Опубликована: Янв. 19, 2025

Soils may not always be suitable to fulfill their intended function. Soil improvement can achieved by mechanical or chemical methods, especially in transportation facilities. L and FA additives are frequently used as additives. In this study, two natural clay samples with extreme very high plasticity were improved using admixtures, properties under static repeated loads investigated ML methods. Two soil from different sites analyzed. eight datasets used. There 14 inputs, including specific gravity (Gs), void ratio (eo), sieve analysis (+No.4, −No.200), size, LL, plastic limit (PL), index (PI), linear shrinkage (Ls), (SL), cure day, agent, type, agent percentage. The outputs swelling (compressive, percent), compressive strengths, modulus of elasticity, compressibility soaked non-soaked conditions. Prediction is attempted (ML) techniques. techniques for regression (such Decision Tree Regression (DTR) K-nearest neighbors (KNN)). SHapley Additive Explanations (SHAP), the impact inputs on observed, it was generally found that PL LL had highest outputs. Different performance metrics evaluation. results showed these predict cyclic extremely clays (R2 > 0.99). These highlight general applicability models containing properties.

Язык: Английский

---

Boostering diagnosis of frontotemporal lobar degeneration with AI-driven neuroimaging – A systematic review and meta-analysis

NeuroImage Clinical, Год журнала: 2025, Номер unknown, С. 103757 - 103757

Опубликована: Фев. 1, 2025

Язык: Английский

---