Interpreting artificial intelligence models: a systematic review on the application of LIME and SHAP in Alzheimer’s disease detection

Brain Informatics, Год журнала: 2024, Номер 11(1)

Опубликована: Апрель 5, 2024

Abstract Explainable artificial intelligence (XAI) has gained much interest in recent years for its ability to explain the complex decision-making process of machine learning (ML) and deep (DL) models. The Local Interpretable Model-agnostic Explanations (LIME) Shaply Additive exPlanation (SHAP) frameworks have grown as popular interpretive tools ML DL This article provides a systematic review application LIME SHAP interpreting detection Alzheimer’s disease (AD). Adhering PRISMA Kitchenham’s guidelines, we identified 23 relevant articles investigated these frameworks’ prospective capabilities, benefits, challenges depth. results emphasise XAI’s crucial role strengthening trustworthiness AI-based AD predictions. aims provide fundamental capabilities XAI enhancing fidelity within clinical decision support systems prognosis.

Язык: Английский

---

AI-Driven Innovations in Alzheimer's Disease: Integrating Early Diagnosis, Personalized Treatment, and Prognostic Modelling

Ageing Research Reviews, Год журнала: 2024, Номер unknown, С. 102497 - 102497

Опубликована: Сен. 1, 2024

Язык: Английский

---

ML-Powered Handwriting Analysis for Early Detection of Alzheimer’s Disease

IEEE Access, Год журнала: 2024, Номер 12, С. 69031 - 69050

Опубликована: Янв. 1, 2024

Alzheimer's disease (AD) is a progressive, incurable condition leading to decline of nerve cells and cognitive functions over time. Early detection essential for improving quality life, as treatment strategies aim decelerate its progression. AD also impacts fine motor control, including handwriting. Utilizing machine learning (ML) with efficient data analysis methods early through handwriting holds considerable promise clinical diagnosis, albeit challenging undertaking. In this study, we address complexity by employing ensemble learning, which amalgamates diverse ML algorithms enhance predictive performance. Our approach involves developing an model kinetics, utilizing the stacking technique integrate multiple base-level classifiers. The study encompasses 174 individuals, 89 diagnosed 85 healthy participants, drawn from DARWIN dataset (Diagnosis AlzheimeR WIth haNdwriting). To discern most effective base classifiers, employ both Repeated-k-fold Monte-Carlo Cross Validation techniques. Subsequently, top k features are selected each best-performing classifier using variance (ANOVA) recursive feature elimination (RFE). final step consolidating predictions classifiers ensemble, resulting in impressive achieves 97.14% accuracy, 95% sensitivity, 100% specificity, precision, 97.44% F1-score, 94.37% Matthews Correlation Coefficient (MCC), 94.21% Cohen Kappa, 97.5% Area Under Receiver Operating Characteristic Curve (AUC-ROC). Comparative performance demonstrates that our proposed surpasses all state-of-the-art models based on prediction. These findings underscore potential offer highly accurate affordable non-invasive manner, emphasizing significant utility, particularly analysis.

Язык: Английский

---

Performance Evaluation of Deep, Shallow and Ensemble Machine Learning Methods for the Automated Classification of Alzheimer’s Disease

International Journal of Neural Systems, Год журнала: 2024, Номер 34(07)

Опубликована: Март 15, 2024

Artificial intelligence (AI)-based approaches are crucial in computer-aided diagnosis (CAD) for various medical applications. Their ability to quickly and accurately learn from complex data is remarkable. Deep learning (DL) models have shown promising results classifying Alzheimer’s disease (AD) its related cognitive states, Early Mild Cognitive Impairment (EMCI) Late (LMCI), along with the healthy conditions known as Cognitively Normal (CN). This offers valuable insights into progression diagnosis. However, certain traditional machine (ML) classifiers perform equally well or even better than DL models, requiring less training data. particularly CAD situations limited labeled datasets. In this paper, we propose an ensemble classifier based on ML magnetic resonance imaging (MRI) data, which achieved impressive accuracy of 96.52%. represents a 3–5% improvement over best individual classifier. We evaluated popular AD classification under both data-scarce data-rich using Disease Neuroimaging Initiative Open Access Series Imaging Studies By comparing state-of-the-art CNN-centric algorithms, gain strengths weaknesses each approach. work will help users select most suitable algorithm availability.

Язык: Английский

---

A Novel Ensemble Framework for Multi-Classification of Brain Tumors Using Magnetic Resonance Imaging

Diagnostics, Год журнала: 2024, Номер 14(4), С. 383 - 383

Опубликована: Фев. 9, 2024

Brain tumors can have fatal consequences, affecting many body functions. For this reason, it is essential to detect brain tumor types accurately and at an early stage start the appropriate treatment process. Although convolutional neural networks (CNNs) are widely used in disease detection from medical images, they face problem of overfitting training phase on limited labeled insufficiently diverse datasets. The existing studies use transfer learning ensemble models overcome these problems. When examined, evident that there a lack weight ratios will be with technique. With framework proposed study, several CNN different architectures trained fine-tuning three A particle swarm optimization-based algorithm determined optimum weights for combining five most successful results across datasets as follows: Dataset 1, 99.35% accuracy 99.20 F1-score; 2, 98.77% 98.92 3, 99.92% 99.92 F1-score. We achieved performances datasets, showing reliable classification. As result, outperforms studies, offering clinicians enhanced decision-making support through its high-accuracy classification performance.

Язык: Английский

---

Explainable AI in Diagnostic Radiology for Neurological Disorders: A Systematic Review, and What Doctors Think About It

Diagnostics, Год журнала: 2025, Номер 15(2), С. 168 - 168

Опубликована: Янв. 13, 2025

Background: Artificial intelligence (AI) has recently made unprecedented contributions in every walk of life, but it not been able to work its way into diagnostic medicine and standard clinical practice yet. Although data scientists, researchers, medical experts have working the direction designing developing computer aided diagnosis (CAD) tools serve as assistants doctors, their large-scale adoption integration healthcare system still seems far-fetched. Diagnostic radiology is no exception. Imagining techniques like magnetic resonance imaging (MRI), computed tomography (CT), positron emission (PET) scans widely very effectively employed by radiologists neurologists for differential diagnoses neurological disorders decades, yet AI-powered systems analyze such incorporated operating procedures systems. Why? It absolutely understandable that medicine, precious human lives are on line, hence there room even tiniest mistakes. Nevertheless, with advent explainable artificial (XAI), old-school black boxes deep learning (DL) unraveled. Would XAI be turning point finally embrace AI radiology? This review a humble endeavor find answers these questions. Methods: In this review, we present journey recognize, preprocess, brain MRI various disorders, special emphasis CAD embedded explainability. A comprehensive literature from 2017 2024 was conducted using host databases. We also domain experts’ opinions summarize challenges up ahead need addressed order fully exploit tremendous potential application diagnostics humanity. Results: Forty-seven studies were summarized tabulated information about technology datasets employed, along performance accuracies. The strengths weaknesses discussed. addition, seven around world presented guide engineers scientists tools. Conclusions: Current research observed focused enhancement accuracies DL regimens, less attention being paid authenticity usefulness explanations. shortage ground truth explainability observed. Visual explanation methods found dominate; however, they might enough, more thorough professor-like explanations would required build trust professionals. Special factors legal, ethical, safety, security issues can bridge current gap between routine practice.

Язык: Английский

---

AlzONet: a deep learning optimized framework for multiclass Alzheimer’s disease diagnosis using MRI brain imaging

The Journal of Supercomputing, Год журнала: 2025, Номер 81(2)

Опубликована: Янв. 23, 2025

Язык: Английский

---

Advanced convolutional neural network with attention mechanism for Alzheimer's disease classification using MRI

Computers in Biology and Medicine, Год журнала: 2025, Номер 190, С. 110095 - 110095

Опубликована: Март 30, 2025

This paper introduces a novel convolutional neural network model with an attention mechanism to advance Alzheimer disease (AD) classification using Magnetic Resonance Imaging (MRI). The architecture is meticulously crafted enhance feature extraction and selectively focus on crucial regions within brain images, thereby improving diagnostic accuracy. A unique component, the MRI Segmentation Block (MSB), introduced manage skull stripping task effectively, highlighting ability learn from complex, multilayered information. We have incorporated detailed experimental evaluation of MSB, demonstrating its superior performance in cranial debridement tasks compared existing methods. experiments involved range scans, assessing MSB's accuracy through metrics like Dice Coefficient Jaccard Index against ground truth annotations by expert radiologists. results substantiate effectiveness, setting new benchmark for precision medical imaging diagnostics. proposed method integrates densely connected networks connection-wise extract multiscale features scans. Furthermore, fine-tuned emphasize salient significantly associated various stages Alzheimer's disease, Extensive Disease Neuroimaging Initiative (ADNI) dataset demonstrate superiority our over traditional contemporary approaches, achieving high computational efficiency. makes it suitable clinical applications where resources are limited. study represents significant advancement process AD, potential implications patient outcomes settings.

Язык: Английский

---

Understanding the Role of Self-Attention in a Transformer Model for the Discrimination of SCD From MCI Using Resting-State EEG

IEEE Journal of Biomedical and Health Informatics, Год журнала: 2024, Номер 28(6), С. 3422 - 3433

Опубликована: Апрель 18, 2024

The identification of EEG biomarkers to discriminate Subjective Cognitive Decline (SCD) from Mild Impairment (MCI) conditions is a complex task which requires great clinical effort and expertise. We exploit the self-attention component Transformer architecture obtain physiological explanations model's decisions in discrimination 56 SCD 45 MCI patients using resting-state EEG. Specifically, an interpretability workflow leveraging attention scores time-frequency analysis epochs through Continuous Wavelet Transform proposed. In classification framework, models are trained validated with 5-fold cross-validation evaluated on test set obtained by selecting 20% total subjects. Ablation studies hyperparameter tuning tests conducted identify optimal model configuration. Results show that best performing model, achieves acceptable results both epochs' patients' classification, capable finding specific patterns highlight changes brain activity between two conditions. demonstrate potential weights as tools guide experts understanding disease-relevant features could be discriminative MCI.

Язык: Английский

---

Assessing the Reliability of Machine Learning Models Applied to the Mental Health Domain Using Explainable AI

Electronics, Год журнала: 2024, Номер 13(6), С. 1025 - 1025

Опубликована: Март 8, 2024

Machine learning is increasingly and ubiquitously being used in the medical domain. Evaluation metrics like accuracy, precision, recall may indicate performance of models but not necessarily reliability their outcomes. This paper assesses effectiveness a number machine algorithms applied to an important dataset domain, specifically, mental health, by employing explainability methodologies. Using multiple model techniques, this work provides insights into models’ workings help determine algorithm predictions. The results are intuitive. It was found that were focusing significantly on less relevant features and, at times, unsound ranking make therefore argues it for research provide addition other accuracy. particularly applications critical domains such as healthcare.

Язык: Английский

---