Explainable Artificial Intelligence in Neuroimaging of Alzheimer’s Disease DOI Creative Commons

Mahdieh Taiyeb Khosroshahi,

Soroush Morsali, Sohrab Gharakhanlou

et al.

Diagnostics, Journal Year: 2025, Volume and Issue: 15(5), P. 612 - 612

Published: March 4, 2025

Alzheimer's disease (AD) remains a significant global health challenge, affecting millions worldwide and imposing substantial burdens on healthcare systems. Advances in artificial intelligence (AI), particularly deep learning machine learning, have revolutionized neuroimaging-based AD diagnosis. However, the complexity lack of interpretability these models limit their clinical applicability. Explainable Artificial Intelligence (XAI) addresses this challenge by providing insights into model decision-making, enhancing transparency, fostering trust AI-driven diagnostics. This review explores role XAI neuroimaging, highlighting key techniques such as SHAP, LIME, Grad-CAM, Layer-wise Relevance Propagation (LRP). We examine applications identifying critical biomarkers, tracking progression, distinguishing stages using various imaging modalities, including MRI PET. Additionally, we discuss current challenges, dataset limitations, regulatory concerns, standardization issues, propose future research directions to improve XAI's integration practice. By bridging gap between AI interpretability, holds potential refine diagnostics, personalize treatment strategies, advance research.

Language: Английский

FC-HGNN: A heterogeneous graph neural network based on brain functional connectivity for mental disorder identification DOI
Yuheng Gu, Shoubo Peng, Yaqin Li

et al.

Information Fusion, Journal Year: 2024, Volume and Issue: 113, P. 102619 - 102619

Published: Aug. 6, 2024

Language: Английский

Citations

7

A novel transfer learning-based model for diagnosing malaria from parasitized and uninfected red blood cell images DOI Creative Commons

Azam Mehmood Qadri,

Ali Raza, Fatma Eid

et al.

Decision Analytics Journal, Journal Year: 2023, Volume and Issue: 9, P. 100352 - 100352

Published: Nov. 4, 2023

Malaria represents a potentially fatal communicable illness triggered by the Plasmodium parasite. This disease is transmitted to humans through bites of Anopheles mosquitoes that carry infection. has significant and devastating consequences on health systems fragile countries, particularly in sub-Saharan Africa. affects red blood cells invading replicating within them, destroying releasing toxic byproducts into bloodstream. The parasite's ability stick modify surface can cause them become sticky, obstructing flow vital organs such as brain spleen. Therefore, efficient approaches for early detection malaria are critical saving patients' lives. main aim this study develop an model diagnosis. We used images based parasitized uninfected experiments. applied neural network-based Neural Search Architecture Network (NASNet) compared its performance with machine learning techniques. Moreover, we proposed novel NNR (NASNet Random forest) method feature engineering. approach first extracts spatial features from input images, then class prediction probability extracted these features. set obtained data extraction trains models. Our comprehensive experiments show support vector outperformed state-of-the-art models, achieving high-performance score 99% having inference time near 0.025 s. validated using k-fold cross-validation optimized hyperparameters tuning. research improved diagnosis assist medical specialists reducing mortality rate.

Language: Английский

Citations

16

Time-series visual explainability for Alzheimer’s disease progression detection for smart healthcare DOI Creative Commons
Nasir Rahim, Tamer Abuhmed, Seyedali Mirjalili

et al.

Alexandria Engineering Journal, Journal Year: 2023, Volume and Issue: 82, P. 484 - 502

Published: Oct. 20, 2023

Artificial intelligence (AI)-based diagnostic systems provide less error-prone and safer support to clinicians, enhancing the medical decision-making process. This study presents a smart reliable healthcare framework for detecting Alzheimer's disease (AD) progression. Early detection of AD before onset clinical symptoms is most crucial step in starting timely treatment. To predict conversion cognitively normal patients those with AD, three-dimensional 3D magnetic resonance imaging (MRI) whole-brain neuroimaging methods have been extensively studied. However, depending on volume, this method computationally expensive. solve problem, we used an approximate rank pooling originally designed video action recognition MRI volume obtain compressed representation multiple two-dimensional (2D) slices. proposes hybrid multimodal CNN-BiLSTM deep model progression detection, which resulting dynamic 2D images are fused cognitive features. Moreover, novel explainable AI approach proposed visual explanations using longitudinal images. Temporal were provided by visualizing affected brain regions captured MRIs. By utilizing sample 1,692 subjects data from Disease Neuroimaging Initiative dataset, our was assessed 10-fold cross-validation The achieved area under receiver operating characteristics curve (AUC) 94% three-time-step image data. fusion features enhanced performance 2% terms AUC. Patients who gradually develop show changes various regions. For such patients, system highlights critical role hippocampus, medial amygdala, caudal lateral amygdala at initial time steps. In late stages detects abnormalities extra as temporal gyrus, superior fusiform hippocampus; indicating that completely progressed AD.

Language: Английский

Citations

15

Information fusion-based Bayesian optimized heterogeneous deep ensemble model based on longitudinal neuroimaging data DOI
Nasir Rahim, Shaker El–Sappagh, Haytham Rizk

et al.

Applied Soft Computing, Journal Year: 2024, Volume and Issue: 162, P. 111749 - 111749

Published: May 15, 2024

Language: Английский

Citations

6

Explaining graph convolutional network predictions for clinicians—An explainable AI approach to Alzheimer's disease classification DOI Creative Commons
Sule Tekkesinoglu, Sara Pudas

Frontiers in Artificial Intelligence, Journal Year: 2024, Volume and Issue: 6

Published: Jan. 8, 2024

Graph-based representations are becoming more common in the medical domain, where each node defines a patient, and edges signify associations between patients, relating individuals with disease symptoms classification task. In this study, Graph Convolutional Networks (GCN) model was utilized to capture differences neurocognitive, genetic, brain atrophy patterns that can predict cognitive status, ranging from Normal Cognition (NC) Mild Cognitive Impairment (MCI) Alzheimer's Disease (AD), on Neuroimaging Initiative (ADNI) database. Elucidating predictions is vital applications promote clinical adoption establish physician trust. Therefore, we introduce decomposition-based explanation method for individual patient classification.

Language: Английский

Citations

5

Predicting Progression From Mild Cognitive Impairment to Alzheimer's Dementia With Adversarial Attacks DOI
İnci M. Baytaş

IEEE Journal of Biomedical and Health Informatics, Journal Year: 2024, Volume and Issue: 28(6), P. 3750 - 3761

Published: March 20, 2024

Early diagnosisof Alzheimer's disease plays a crucial role in treatment planning that might slow down the disease's progression. This problem is commonly posed as classification task performed by machine learning and deep techniques. Although data-driven techniques set state-of-the-art many domains, scale of available datasets research not sufficient to learn complex models from patient data. study proposes simple yet promising framework predict conversion Mild Cognitive Impairment (MCI) Disease (AD). The proposed comprises shallow neural network for binary single-step gradient-based adversarial attack find an progression direction input space. step size required change patient's diagnosis MCI AD indicates distance decision boundary. at next visit predicted employing this notion We also present potential application subtyping. Experiments with two publicly imply can MCI-to-AD conversions assist subtyping only training network.

Language: Английский

Citations

5

Revolutionizing tumor detection and classification in multimodality imaging based on deep learning approaches: methods, applications and limitations DOI
Dildar Hussain, Mohammed A. Al‐masni, Muhammad Aslam

et al.

Journal of X-Ray Science and Technology, Journal Year: 2024, Volume and Issue: 32(4), P. 857 - 911

Published: April 30, 2024

The emergence of deep learning (DL) techniques has revolutionized tumor detection and classification in medical imaging, with multimodal imaging (MMI) gaining recognition for its precision diagnosis, treatment, progression tracking.

Language: Английский

Citations

5

MACFNet: Detection of Alzheimer's disease via multiscale attention and cross-enhancement fusion network DOI Creative Commons
Chaosheng Tang, Mengbo Xi, Junding Sun

et al.

Computer Methods and Programs in Biomedicine, Journal Year: 2024, Volume and Issue: 254, P. 108259 - 108259

Published: June 6, 2024

Alzheimer's disease (AD) is a dreaded degenerative that results in profound decline human cognition and memory. Due to its intricate pathogenesis the lack of effective therapeutic interventions, early diagnosis plays paramount role AD. Recent research based on neuroimaging has shown application deep learning methods by multimodal neural images can effectively detect However, these only concatenate fuse high-level features extracted from different modalities, ignoring fusion interaction low-level across modalities. It consequently leads unsatisfactory classification performance. In this paper, we propose novel multi-scale attention cross-enhanced network, MACFNet, which enables multi-stage between inputs learn shared feature representations. We first construct Cross-Enhanced Fusion Module (CEFM), fuses modalities through cross-structure. addition, an Efficient Spatial Channel Attention (ECSA) module proposed, able focus important AD-related more efficiently achieve enhancement two-stage residual concatenation. Finally, also multiscale guiding block (MSAG) dilated convolution, obtain rich receptive fields without increasing model parameters computation, improve efficiency extraction. Experiments Disease Neuroimaging Initiative (ADNI) dataset demonstrate our MACFNet better performance than existing methods, with accuracies 99.59%, 98.85%, 99.61%, 98.23% for AD vs. CN, MCI, CN MCI respectively, specificity 98.92%, 97.07%, 99.58% 99.04%, sensitivity 99.91%, 99.89%, 99.63% 97.75%, respectively. The proposed high-accuracy diagnostic framework. Through cross mechanism efficient attention, make full use modal medical pay local global information images. This work provides valuable reference multi-mode diagnosis.

Language: Английский

Citations

4

TDF-Net: Trusted Dynamic Feature Fusion Network for breast cancer diagnosis using incomplete multimodal ultrasound DOI
Pengfei Yan,

Wushuang Gong,

Minglei Li

et al.

Information Fusion, Journal Year: 2024, Volume and Issue: 112, P. 102592 - 102592

Published: July 20, 2024

Language: Английский

Citations

4

Advancing healthcare through multimodal data fusion: a comprehensive review of techniques and applications DOI Creative Commons

Jing Ru Teoh,

Jian Dong, Xi‐Nian Zuo

et al.

PeerJ Computer Science, Journal Year: 2024, Volume and Issue: 10, P. e2298 - e2298

Published: Oct. 30, 2024

With the increasing availability of diverse healthcare data sources, such as medical images and electronic health records, there is a growing need to effectively integrate fuse this multimodal for comprehensive analysis decision-making. However, despite its potential, fusion in remains limited. This review paper provides an overview existing literature on healthcare, covering 69 relevant works published between 2018 2024. It focuses methodologies that different types enhance analysis, including techniques integrating with structured unstructured data, combining multiple image modalities, other features. Additionally, reviews various approaches fusion, early, intermediate, late methods, examines challenges limitations associated these techniques. The potential benefits applications diseases are highlighted, illustrating specific strategies employed artificial intelligence (AI) model development. research synthesizes information facilitate progress using improved diagnosis treatment planning.

Language: Английский

Citations

4