A New Brain Network Construction Paradigm for Brain Disorder via Diffusion-Based Graph Contrastive Learning

IEEE Transactions on Pattern Analysis and Machine Intelligence, Год журнала: 2024, Номер 46(12), С. 10389 - 10403

Опубликована: Авг. 13, 2024

Brain network analysis plays an increasingly important role in studying brain function and the exploring of disease mechanisms. However, existing construction tools have some limitations, including dependency on empirical users, weak consistency repeated experiments time-consuming processes. In this work, a diffusion-based pipeline, DGCL is designed for end-to-end networks. Initially, region-aware module (BRAM) precisely determines spatial locations regions by diffusion process, avoiding subjective parameter selection. Subsequently, employs graph contrastive learning to optimize connections eliminating individual differences redundant unrelated diseases, thereby enhancing networks within same group. Finally, node-graph loss classification jointly constrain process model obtain reconstructed network, which then used analyze connections. Validation two datasets, ADNI ABIDE, demonstrates that surpasses traditional methods other deep models predicting development stages. Significantly, proposed improves efficiency generalization construction. summary, can be served as universal scheme, effectively identify through generative paradigms has potential provide interpretability support neuroscience research.

Язык: Английский

---

A-GCL: Adversarial graph contrastive learning for fMRI analysis to diagnose neurodevelopmental disorders

Medical Image Analysis, Год журнала: 2023, Номер 90, С. 102932 - 102932

Опубликована: Авг. 21, 2023

Язык: Английский

---

A dual-stream spatio-temporal fusion network with multi-sensor signals for remaining useful life prediction

Journal of Manufacturing Systems, Год журнала: 2024, Номер 76, С. 43 - 58

Опубликована: Июль 25, 2024

Язык: Английский

---

DSAM: A deep learning framework for analyzing temporal and spatial dynamics in brain networks

Medical Image Analysis, Год журнала: 2025, Номер 101, С. 103462 - 103462

Опубликована: Янв. 29, 2025

Resting-state functional magnetic resonance imaging (rs-fMRI) is a noninvasive technique pivotal for understanding human neural mechanisms of intricate cognitive processes. Most rs-fMRI studies compute single static connectivity matrix across brain regions interest, or dynamic matrices with sliding window approach. These approaches are at risk oversimplifying dynamics and lack proper consideration the goal hand. While deep learning has gained substantial popularity modeling complex relational data, its application to uncovering spatiotemporal still limited. In this study we propose novel interpretable framework that learns goal-specific directly from time series employs specialized graph network final classification. Our model, DSAM, leverages temporal causal convolutional networks capture in both low- high-level feature representations, attention unit identify important points, self-attention construct matrix, variant spatial downstream To validate our approach, conducted experiments on Human Connectome Project dataset 1075 samples build interpret model classification sex group, Adolescent Brain Cognitive Development Dataset 8520 independent testing. Compared proposed other state-of-art models, results suggested approach goes beyond assumption fixed provides evidence patterns, which opens up potential gain deeper insights into how adapts specific task implementation can be found https://github.com/bishalth01/DSAM.

Язык: Английский

---

Efficient Clustering Method for Graph Images Using Two-Stage Clustering Technique

Electronics, Год журнала: 2025, Номер 14(6), С. 1232 - 1232

Опубликована: Март 20, 2025

Graphimages, which represent data structures through nodes and edges, present significant challenges for clustering due to their intricate topological properties. Traditional algorithms, such as K-means Density-Based Spatial Clustering of Applications with Noise (DBSCAN), often struggle effectively capture both spatial structural relationships within graph images. To overcome these limitations, we propose a novel two-stage approach that integrates conventional techniques graph-based methodologies enhance accuracy efficiency. In the first stage, distance- or density-based algorithm (e.g., DBSCAN) is applied generate initial cluster formations. second clusters are refined using spectral community detection better preserve exploit features. We evaluate our dataset 8118 images derived from depth measurements taken at various angles. The experimental results demonstrate method surpasses single-method approaches in terms silhouette score, Calinski-Harabasz index (CHI), modularity. score measures how similar an object its own compared other clusters, while CHI, also known Variance Ratio Criterion, evaluates quality based on ratio between-cluster dispersion within-cluster dispersion. Modularity, metric commonly used clustering, assesses strength division network into communities. Furthermore, qualitative analysis visualization confirms proposed more differentiates similarities These findings underscore potential hybrid applications, including three-dimensional (3D) measurement analysis, medical imaging, social analysis.

Язык: Английский

---

Graph neural network and machine learning analysis of functional neuroimaging for understanding schizophrenia

BMC Neuroscience, Год журнала: 2024, Номер 25(1)

Опубликована: Янв. 2, 2024

Graph representational learning can detect topological patterns by leveraging both the network structure as well nodal features. The basis of our exploration involves application graph neural architectures and machine to resting-state functional Magnetic Resonance Imaging (rs-fMRI) data for purpose detecting schizophrenia. Our study uses single-site avoid shortcomings in generalizability neuroimaging obtained from multiple sites.

Язык: Английский

---

Neuroimage analysis using artificial intelligence approaches: a systematic review

Medical & Biological Engineering & Computing, Год журнала: 2024, Номер 62(9), С. 2599 - 2627

Опубликована: Апрель 26, 2024

Язык: Английский

---

A comprehensive survey of complex brain network representation

Meta-Radiology, Год журнала: 2023, Номер 1(3), С. 100046 - 100046

Опубликована: Ноя. 1, 2023

Recent years have shown great merits in utilizing neuroimaging data to understand brain structural and functional changes, as well its relationship different neurodegenerative diseases other clinical phenotypes. Brain networks, derived from modalities, attracted increasing attention due their potential gain system-level insights characterize dynamics abnormalities neurological conditions. Traditional methods aim pre-define multiple topological features of networks relate these measures or demographical variables. With the enormous successes deep learning techniques, graph played significant roles network analysis. In this survey, we first provide a brief overview neuroimaging-derived networks. Then, focus on presenting comprehensive both traditional state-of-the-art deep-learning for mining. Major models, objectives are reviewed within paper. Finally, discuss several promising research directions field.

Язык: Английский

---

Spectral Graph Neural Network-Based Multi-Atlas Brain Network Fusion for Major Depressive Disorder Diagnosis

IEEE Journal of Biomedical and Health Informatics, Год журнала: 2024, Номер 28(5), С. 2967 - 2978

Опубликована: Фев. 16, 2024

Major Depressive Disorder (MDD) imposes a substantial burden within the healthcare domain, impacting millions of individuals worldwide. Functional Magnetic Resonance Imaging (fMRI) has emerged as promising tool for objective diagnosis MDD, enabling investigation functional connectivity patterns in brain associated with this disorder. However, most existing methods focus on single atlas, which limits their ability to capture complex, multi-scale nature networks. To address these limitations, we propose novel multi-atlas fusion method that incorporates early and late unified framework. Our introduces concept holistic Connectivity Network (FCN), captures both intra-atlas relationships individual atlases inter-regional between different parcellation scales. This comprehensive representation enables identification potential disease-related MDD stage our Moreover, by decoding FCN from various perspectives through multiple spectral Graph Convolutional Neural Networks fusing results decision-level ensembles, further improve performance diagnosis. approach is easily implemented minimal modifications model structures demonstrates robust across baseline models. method, evaluated public resting-state fMRI datasets, surpasses current methods, enhancing accuracy The proposed framework provides more reliable diagnostic technique. Experimental show outperforms single- multi-atlas-based demonstrating its effectiveness advancing

Язык: Английский

---

Multiview hyperedge-aware hypergraph embedding learning for multisite, multiatlas fMRI based functional connectivity network analysis

Medical Image Analysis, Год журнала: 2024, Номер 94, С. 103144 - 103144

Опубликована: Март 19, 2024

Язык: Английский

---