Computer methods and programs in biomedicine最新文献

{"title":"DG-MSGAT: A Biologically-informed Differential Gene Multi-Scale Graph Attention Network for predicting neoadjuvant therapy response in rectal cancer.","authors":"Xu Luo, Pei Shu, Ning Liu, Dong Miao, Xiuding Cai, Yu Yao, Xin Wang","doi":"10.1016/j.cmpb.2025.108974","DOIUrl":"10.1016/j.cmpb.2025.108974","url":null,"abstract":"<p><strong>Background and objective: </strong>Accurate prediction of the efficacy of neoadjuvant therapy - particularly the likelihood of achieving a pathological complete response (pCR) - is critical to improving outcomes in patients with rectal cancer. The anticipation of therapeutic response prior to surgery enables the development of personalized treatment strategies and reduces unnecessary interventions for non-responders. While genetic profiling has been integrated into predictive models to enhance response estimation, many existing approaches overlook gene-gene interactions. Furthermore, they often struggle with the high dimensionality, noise, and sparsity inherent in gene expression data. To address these limitations, we propose a biologically informed model, the Differential Gene Multi-Scale Graph Attention Network (DG-MSGAT). This model integrates differential expression signals with multi-scale gene interaction patterns to improve the accuracy of treatment response prediction.</p><p><strong>Methods: </strong>By integrating gene expression profiles with differential expression signals, we construct a patient-specific gene graph whose edges are defined based on curated protein-protein interaction data. This graph is then processed by DG-MSGAT, a multi-scale graph attention network that utilizes stacked attention layers and residual connections to model hierarchical gene dependencies and preserve feature integrity. The resulting representation is subsequently used to estimate the probability of achieving a pathological complete response.</p><p><strong>Results: </strong>In patients with locally advanced rectal cancer, the DG-MSGAT model substantially outperformed conventional algorithms - including support vector machines, decision trees, and random forests - in predicting neoadjuvant therapy efficacy. Network analysis identified key genes (e.g., TP53, EGFR, CTNNB1) and immune-related pathways that are consistent with clinically established determinants of therapeutic response.</p><p><strong>Conclusion: </strong>The DG-MSGAT model offers a promising advancement in the prediction of neoadjuvant therapy outcomes in rectal cancer. By effectively modeling gene interactions and mitigating the limitations associated with high-dimensional gene expression data, it provides a clinically relevant tool to support personalized treatment decision-making.</p>","PeriodicalId":10624,"journal":{"name":"Computer methods and programs in biomedicine","volume":"271 ","pages":"108974"},"PeriodicalIF":4.8,"publicationDate":"2025-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144803827","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Prediction of breast cancer HER2 status changes based on ultrasound radiomics attention network.","authors":"Jian Liu, Xinzheng Xue, Yuqi Yan, Qian Song, Yuhu Cheng, Liping Wang, Xuesong Wang, Dong Xu","doi":"10.1016/j.cmpb.2025.108987","DOIUrl":"10.1016/j.cmpb.2025.108987","url":null,"abstract":"<p><strong>Background and objective: </strong>Following Neoadjuvant Chemotherapy (NAC), there exists a probability of changes occurring in the Human Epidermal Growth Factor Receptor 2 (HER2) status. If these changes are not promptly addressed, it could hinder the timely adjustment of treatment plans, thereby affecting the optimal management of breast cancer. Consequently, the accurate prediction of HER2 status changes holds significant clinical value, underscoring the need for a model capable of precisely forecasting these alterations.</p><p><strong>Methods: </strong>In this paper, we elucidate the intricacies surrounding HER2 status changes, and propose a deep learning architecture combined with radiomics techniques, named as Ultrasound Radiomics Attention Network (URAN), to predict HER2 status changes. Firstly, radiomics technology is used to extract ultrasound image features to provide rich and comprehensive medical information. Secondly, HER2 Key Feature Selection (HKFS) network is constructed for retain crucial features relevant to HER2 status change. Thirdly, we design Max and Average Attention and Excitation (MAAE) network to adjust the model's focus on different key features. Finally, a fully connected neural network is utilized to predict HER2 status changes. The code to reproduce our experiments can be found at https://github.com/didadiuouo/URAN.</p><p><strong>Results: </strong>Our research was carried out using genuine ultrasound images sourced from hospitals. On this dataset, URAN outperformed both state-of-the-art and traditional methods in predicting HER2 status changes, achieving an accuracy of 0.8679 and an AUC of 0.8328 (95% CI: 0.77-0.90). Comparative experiments on the public BUS_UCLM dataset further demonstrated URAN's superiority, attaining an accuracy of 0.9283 and an AUC of 0.9161 (95% CI: 0.91-0.92). Additionally, we undertook rigorously crafted ablation studies, which validated the logicality and effectiveness of the radiomics techniques, as well as the HKFS and MAAE modules integrated within the URAN model. The results pertaining to specific HER2 statuses indicate that URAN exhibits superior accuracy in predicting changes in HER2 status characterized by low expression and IHC scores of 2+ or below. Furthermore, we examined the radiomics attributes of ultrasound images and discovered that various wavelet transform features significantly impacted the changes in HER2 status.</p><p><strong>Conclusions: </strong>We have developed a URAN method for predicting HER2 status changes that combines radiomics techniques and deep learning. URAN model have better predictive performance compared to other competing algorithms, and can mine key radiomics features related to HER2 status changes.</p>","PeriodicalId":10624,"journal":{"name":"Computer methods and programs in biomedicine","volume":"271 ","pages":"108987"},"PeriodicalIF":4.8,"publicationDate":"2025-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144803828","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Quantum LBP-driven heart sound analysis with quality assessment in real-world noisy environments","authors":"Subhashree Sahoo,&nbsp;Puneet Kumar Jain","doi":"10.1016/j.cmpb.2025.109082","DOIUrl":"10.1016/j.cmpb.2025.109082","url":null,"abstract":"<div><h3>Background and Objective:</h3><div>Cardiovascular disease (CVD) is a major global health concern with increasing prevalence. While electrocardiography (ECG) and echocardiography offer high accuracy, their reliance on specialized equipment and trained personnel makes them costly and less accessible in rural areas. In contrast, the Phonocardiography (PCG) provides a more affordable alternative via capturing heart sound signals using a stethoscope. However, PCG analysis is often compromised by environmental noise. Most existing methods address this issue with denoising techniques, which can inadvertently suppress vital heart sound components. To address this challenge, the objective of this work is to develop a computationally efficient and noise-resilient method for PCG classification.</div></div><div><h3>Methods:</h3><div>The proposed approach introduces a quality assessment metric (<span><math><mrow><mi>P</mi><mi>C</mi><msub><mrow><mi>G</mi></mrow><mrow><mi>Q</mi><mi>A</mi></mrow></msub></mrow></math></span>) to select the least noisy subsequences for further processing. For noise-robust feature extraction, an Enhanced Quantum Local Binary Pattern (EQLBP) method is employed, which adaptively selects reference pixels and extracts uniform patterns from the signal spectrogram to mitigate noise effects. In addition, Discrete Wavelet Transform (DWT) features are extracted to capture multi-resolution time–frequency characteristics, complementing the local texture features obtained from EQLBP. The combined feature set is then used to train conventional machine learning models.</div></div><div><h3>Results:</h3><div>The proposed method was evaluated using 10-fold cross-validation on two publicly available datasets: CinC-2016 and HSM-2018. On CinC-2016, it achieved an accuracy of 97.22%, precision of 98.29%, recall of 98.63%, and an F1-score of 98.46%. On HSM-2018, the method obtained an accuracy of 98.70%, precision of 99.05%, recall of 99.00%, and an F1-score of 99.00%. These results highlight the superior performance of the proposed approach compared to existing methods.</div></div><div><h3>Conclusions:</h3><div>With its computational efficiency and robust performance, the proposed method is well-suited for out-of-clinic applications, particularly in rural and remote areas where access to advanced diagnostic tools is limited.</div></div>","PeriodicalId":10624,"journal":{"name":"Computer methods and programs in biomedicine","volume":"273 ","pages":"Article 109082"},"PeriodicalIF":4.8,"publicationDate":"2025-09-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145181373","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"MitraClip device automated localization in 3D transesophageal echocardiography via deep learning.","authors":"Riccardo Munafò, Simone Saitta, Luca Vicentini, Davide Tondi, Veronica Ruozzi, Francesco Sturla, Giacomo Ingallina, Andrea Guidotti, Eustachio Agricola, Emiliano Votta","doi":"10.1016/j.cmpb.2025.109083","DOIUrl":"https://doi.org/10.1016/j.cmpb.2025.109083","url":null,"abstract":"<p><strong>Background and objective: </strong>The MitraClip is the most widely used percutaneous treatment for mitral regurgitation, typically performed under the real-time guidance of 3D transesophageal echocardiography (TEE). However, artifacts and low image contrast in echocardiography hinder accurate clip visualization. This study presents a proof-of-concept of an automated pipeline for clip detection from 3D TEE images acquired in a controlled in vitro simulation environment.</p><p><strong>Methods: </strong>An Attention UNet was employed to segment the device, while a DenseNet classifier predicted its configuration among ten possible states, ranging from fully closed to fully open. Based on the predicted configuration, a template model derived from computer-aided design (CAD) was automatically registered to refine the segmentation and enable quantitative characterization of the device. The pipeline was trained and validated on 196 3D TEE images acquired using a heart simulator, with ground-truth annotations refined through CAD-based templates.</p><p><strong>Results: </strong>The Attention UNet achieved an average surface distance of 0.76 mm and a 95% Hausdorff distance of 2.44 mm for segmentation, while the DenseNet achieved an average weighted F1-score of 0.80 for classification. Post-refinement, segmentation accuracy improved, with average surface distance and 95% Hausdorff distance reduced to 0.69 mm and 1.83 mm, respectively.</p><p><strong>Conclusion: </strong>This pipeline enhanced clip visualization, providing fast and accurate detection with quantitative feedback, potentially improving procedural efficiency and reducing adverse outcomes.</p>","PeriodicalId":10624,"journal":{"name":"Computer methods and programs in biomedicine","volume":"272 ","pages":"109083"},"PeriodicalIF":4.8,"publicationDate":"2025-09-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145211974","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Morphogeometric variation analysis of cortical surfaces in Alzheimer's Disease. A novel approach for prognostic biomarker discovery.","authors":"José González-Cabrero, Carmelo Gómez, Francisco Cavas","doi":"10.1016/j.cmpb.2025.109089","DOIUrl":"https://doi.org/10.1016/j.cmpb.2025.109089","url":null,"abstract":"<p><strong>Background and objective: </strong>Longitudinal morphogeometric analysis is essential to understand neurodegenerative progression in Alzheimer's disease (AD). This research evaluates external and internal cortical surfaces' morphology extracted from MRI scans to characterize structural changes in AD.</p><p><strong>Methods: </strong>MRI scans from 22 patients with AD across multiple imaging sessions were segmented to generate 3D cortical reconstructions. A novel sectional analysis method was applied to sweep coronal planes at 1mm. intervals along the posterior-anterior axis. Morphogeometric indices were calculated for each section to generate sectional curves, and mathematical curve descriptors were computed as potential biomarkers. Statistical evaluation of these descriptors included both an effectiveness metric and a Linear Mixed Model (LMM) analysis to assess longitudinal trends and determine statistical significance.</p><p><strong>Results: </strong>Curve descriptors showed greater effectiveness to detect morphological changes than traditional whole-brain geometric metrics. The external cortical surface volume curve achieved 87.88 % effectiveness, surpassing whole-brain volume (84.85 %). The internal cortical surface area sectional curve reached 81.82 %, outperforming traditional measures (75.78 %). The novel IECN index achieved 72.73 %, highlighting its biomarker potential.</p><p><strong>Conclusions: </strong>Novel morphogeometric indices and sectional curve descriptors complement traditional biomarkers, improving AD detection and monitoring. The employed methodology is sensitive to local cortical changes that may be overlooked in whole-brain assessments.</p>","PeriodicalId":10624,"journal":{"name":"Computer methods and programs in biomedicine","volume":"272 ","pages":"109089"},"PeriodicalIF":4.8,"publicationDate":"2025-09-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145212006","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"MultiExCam: A multi approach and explainable artificial intelligence architecture for skin lesion classification","authors":"Tommaso Ruga ,&nbsp;Luciano Caroprese ,&nbsp;Eugenio Vocaturo ,&nbsp;Ester Zumpano","doi":"10.1016/j.cmpb.2025.109081","DOIUrl":"10.1016/j.cmpb.2025.109081","url":null,"abstract":"<div><h3>Background and Objective:</h3><div>Cutaneous melanoma remains the most lethal form of skin cancer. Although incurable at advanced stages, if diagnosed at an early, localized stage, the five-year survival rate is remarkably high. Recent advancements in artificial intelligence have paved the way for early skin lesion diagnosis, leveraging digital imaging processes into effective solutions. Most of these, however, use Machine Learning and Deep Learning techniques compartmentalized, without combining the produced predictions.</div></div><div><h3>Methods:</h3><div>This paper introduces MultiExCam, a novel multi approach and explainable architecture for skin cancer detection that integrates both machine and deep learning. Three heterogeneous data from three different techniques are used: dermatoscopic images, features extracted from deep learning techniques, and hand-crafted statistical features. A convolutional neural network is used for both deep feature extraction and initial classification, with the extracted features being combined with handcrafted ones to train four additional machine learning models. An advanced ensemble model, implemented as a Feed Forward Neural Network with gating and attention mechanism, produces the final classification. To enhance interpretability, the architecture employs GradCAM for visualizing critical regions in input images and SHAP for evaluating the contribution of individual features to predictions.</div></div><div><h3>Results:</h3><div>MultiExCam demonstrates robust performance across three diverse datasets (HAM10000, ISIC, MED-NODE), achieving AUC scores of 97%, 91%, and 98% respectively, with corresponding F1-scores of 92%, 87%, and 94%. Comprehensive ablation studies validate the importance of the preprocessing pipeline and ensemble integration, with the hybrid approach consistently outperforming baseline deep learning models by 1–3 percentage points. Unlike existing compartmentalized hybrid solutions, MultiExCam’s adaptive ensemble architecture learns personalized decision strategies for individual lesions, mimicking expert dermatological workflows that integrate multiple evidence sources. The explainability analysis reveals clinically meaningful activation patterns corresponding to established diagnostic criteria including asymmetry, border irregularity, and color variation.</div></div><div><h3>Conclusion:</h3><div>MultiExCam establishes a new paradigm for AI-assisted dermatological diagnosis by demonstrating that true hybrid integration of deep learning and machine learning, combined with comprehensive explainability techniques, can achieve both superior diagnostic performance and clinical interpretability. The architecture’s ability to provide accurate classifications while explaining prediction rationale addresses critical requirements for medical AI adoption, offering a promising foundation for clinical decision support systems in melanoma detection.</div></div>","PeriodicalId":10624,"journal":{"name":"Computer methods and programs in biomedicine","volume":"273 ","pages":"Article 109081"},"PeriodicalIF":4.8,"publicationDate":"2025-09-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145181374","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Beyond predictive accuracy: Statistical validation of feature importance in biomedical machine learning","authors":"Souichi Oka ,&nbsp;Nobuko Inoue ,&nbsp;Yoshiyasu Takefuji","doi":"10.1016/j.cmpb.2025.109085","DOIUrl":"10.1016/j.cmpb.2025.109085","url":null,"abstract":"<div><div>In medical machine learning (ML), a fundamental methodological distinction exists between optimizing model performance for predictive tasks and pursuing causal inference for mechanistic interpretation. Achieving high predictive accuracy does not necessarily imply that a model can uncover the true physiological mechanisms underlying the data. This letter addresses a critical interpretational challenge in medical machine learning, building upon Yuyang Yan et al.’s valuable work on exacerbation classification in asthma and COPD. While their multi-feature fusion model, particularly comprising models such as K-Nearest Neighbors (KNN), Support Vector Machines (SVM), Random Forest (RF), and Bidirectional Long Short-Term Memory (BiLSTM) demonstrates high predictive accuracy for respiratory exacerbations, we highlight that such performance alone does not guarantee reliable insights into feature importance. Complex tree-based models like RF, when interpreted via methods like SHapley Additive exPlanations (SHAP), can exhibit inherent biases, overemphasizing features used in early splits and reflecting what is important for their specific prediction rather than the true underlying physiological drivers. Validating feature importance remains challenging without ground truth, as different models often yield varying rankings. We argue that solely relying on model-dependent interpretations risks misrepresenting the actual mechanisms of complex medical phenomena. Therefore, we advocate for a robust analytical strategy that transcends mere predictive metrics. This involves a synergistic approach combining the predictive power of ML with impartial, complementary statistical methodologies—such as non-parametric correlation and mutual information—to ensure genuinely trustworthy scientific insights into the true drivers of respiratory exacerbations.</div></div>","PeriodicalId":10624,"journal":{"name":"Computer methods and programs in biomedicine","volume":"272 ","pages":"Article 109085"},"PeriodicalIF":4.8,"publicationDate":"2025-09-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145155302","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"High-order mesoscale modeling with geometrically conforming gray/white matter interface for traumatic brain injury","authors":"Denis Molchanov ,&nbsp;Chaokai Zhang ,&nbsp;Nan Lin ,&nbsp;Songbai Ji ,&nbsp;Zhangxian Yuan","doi":"10.1016/j.cmpb.2025.109074","DOIUrl":"10.1016/j.cmpb.2025.109074","url":null,"abstract":"<div><h3>Background and Objective:</h3><div>Brain injury models with a supramillimeter resolution are not feasible to provide spatially detailed strains at or below millimeter scales, especially in regions of convoluted geometry such as at the gray/white matter interface. Furthermore, non-conforming mesh boundaries resulting from discretization errors can lead to inaccurate strain and stress distributions near interfaces, areas typically associated with elevated vulnerabilities in traumatic brain injury (TBI). Conventional approaches using extremely small linear elements are not effective to address the issue because of challenges in generating boundary-conforming meshes and slow convergence.</div></div><div><h3>Methods:</h3><div>In this study, we adapt the Non-Uniform Rational B-Splines (NURBS) and isogeometric analysis (IGA) to develop high-order mesoscale models that smoothly represent complex tissue boundaries with highly resolved strain distributions. We address key challenges for applications to the brain, including the construction of smooth tissue boundaries from voxelized image segmentation and overcoming numerical difficulties arising from near-incompressibility.</div></div><div><h3>Results:</h3><div>Compared to the conventional model using linear elements, the high-order mesoscale model demonstrates superior efficiency by achieving the same accuracy but with two orders of magnitude fewer degrees of freedom and at least one order of magnitude reduction in computational cost. Two-dimensional mesoscale models are constructed at gray/white matter interface to simulate realistic impact loading. The high-order mesoscale models discover strain concentration at the convoluted tissue boundary missing from the global model (e.g., up to 20% difference in magnitude). Notable differences in strain distribution also exist, with a normalized root mean squared error of up to 7.7% for strains sampled near the interface. These strain differences have major implications on downstream axonal injury model simulations.</div></div><div><h3>Conclusion:</h3><div>This study demonstrates the unique potential of leveraging IGA to develop mesoscale brain models with conforming tissue boundaries, and is important for filling a critical gap between global and cellular brain injury models in a multiscale modeling framework. The technique is general and scalable as it is applicable to diverse two- and three-dimensional biomechanical problems, including and beyond brain biomechanics.</div></div>","PeriodicalId":10624,"journal":{"name":"Computer methods and programs in biomedicine","volume":"272 ","pages":"Article 109074"},"PeriodicalIF":4.8,"publicationDate":"2025-09-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145155376","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Whole-brain computational modeling reveals disruption of microscale brain dynamics in Parkinson’s disease","authors":"Wenqian Zhao ,&nbsp;Xueao Li ,&nbsp;Jing Wang ,&nbsp;Chunyan Zhang ,&nbsp;Yuchuan Zhuang ,&nbsp;Yanbo Dong ,&nbsp;Andrey Tulupov ,&nbsp;Jing Li ,&nbsp;Fengshou Zhang ,&nbsp;Jianfeng Bao","doi":"10.1016/j.cmpb.2025.109076","DOIUrl":"10.1016/j.cmpb.2025.109076","url":null,"abstract":"<div><h3>Background and Objective</h3><div>Parkinson’s disease (PD) alters the brain’s neurodynamic properties, contributing to both motor and non-motor symptoms. Although advances in neuroimaging techniques—such as resting-state functional MRI (rsfMRI), diffusion tensor imaging (DTI), and structural MRI (sMRI)—have enhanced our understanding of brain structure and function, they remain limited in detecting subtle, region-specific dynamic alterations associated with functional deficits. This study aims to apply the relaxed mean field dynamic modeling (rMFM) to identify microscale dynamic abnormalities in PD and to link these changes with network topology and clinical characteristics.</div></div><div><h3>Methods</h3><div>We employed the rMFM, a biophysically informed computational framework that integrates structural and functional imaging data with microstructural features to simulate local dynamics of brain regions. Unlike traditional models, rMFM allows the optimization of regional recurrent connection strength w and subcortical input I, thereby capturing inter-regional heterogeneity more effectively. Separate rMFM models were constructed for the PD and healthy control (HC) groups. Group differences in model parameters were assessed, followed by graph-theoretical analysis to examine alterations in brain network topology. Correlation analyses were also performed to investigate the relationships between model parameters, network metrics, and clinical variables.</div></div><div><h3>Results</h3><div>Significant alterations in w and I were observed in regions such as the middle temporal gyrus and banks of the superior temporal sulcus (bankssts) in the PD group, suggesting localized dynamic disruptions related to language, memory, and cognitive impairments. Corresponding alterations in brain network topology accompanied these parameter changes. At the same time, the results of graph theory analysis suggest that in early PD, functional disorders may appear before obvious structural changes.</div></div><div><h3>Conclusions</h3><div>This study introduces rMFM as an innovative approach for modeling local brain dynamics by integrating multimodal MRI data with microscale neural features. The findings highlight distinctive microscale dynamic abnormalities in PD and their linkage to large-scale network changes. This approach enhances our understanding of PD pathophysiology and provids a basis for identifying potential disease-specific biomarkers.</div></div>","PeriodicalId":10624,"journal":{"name":"Computer methods and programs in biomedicine","volume":"272 ","pages":"Article 109076"},"PeriodicalIF":4.8,"publicationDate":"2025-09-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145155377","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Measuring phase-amplitude coupling using dispersion fuzzy mutual information","authors":"Hao Zhang ,&nbsp;Zhijie Bian ,&nbsp;Xiaonan Guo ,&nbsp;Xiaoli Li ,&nbsp;Shimin Yin ,&nbsp;Dong Cui","doi":"10.1016/j.cmpb.2025.109075","DOIUrl":"10.1016/j.cmpb.2025.109075","url":null,"abstract":"<div><h3>Background</h3><div>Mild Cognitive Impairment (MCI) is a preliminary stage of Alzheimer’s disease (AD), and early diagnosis of MCI electroencephalography (EEG) signals using the Phase-Amplitude Coupling (PAC) phenomenon in neural oscillations as an EEG marker has become a promising technique. Nonetheless, the PAC estimators, which are frequently employed in clinical practice, exhibit considerable limitations with regard to their application conditions. In order to explore a PAC estimator with strong applicability, the Dispersion Fuzzy Mutual Information (DFMI) method is proposed.</div></div><div><h3>Methods</h3><div>The DFMI method employs the symbolization principle of dispersion entropy and mutual information theory to process time series. This approach addresses the challenges posed by ambiguous quantitative fluctuations in the number of patterns in fuzzy entropy and upgrades the single-channel fuzzy entropy algorithm to a dual-channel DFMI algorithm. Subsequently, through simulation analysis, it was compared with the commonly used PAC estimator in clinical practice in terms of coupling strength sensitivity, data length dependency, noise resistance, coupling frequency band sensitivity, and artifact resistance.</div></div><div><h3>Results</h3><div>The simulation results indicate that the DFMI method can effectively obtain PAC strength, is less dependent on data length, produces stable calculation results, and is less affected by pseudo-trace signals. The MCI-EEG data results demonstrated that MCI patients significantly enhanced whole-brain theta-gamma coupling activity, while alpha-gamma coupling activity shifted to the low-frequency band.</div></div><div><h3>Conclusion</h3><div>The DFMI can be utilized as a PAC estimator to assess PAC phenomena in neural signals, and the coupling of neural oscillations in the MCI brain may manifest as coupling band attenuation.</div></div>","PeriodicalId":10624,"journal":{"name":"Computer methods and programs in biomedicine","volume":"272 ","pages":"Article 109075"},"PeriodicalIF":4.8,"publicationDate":"2025-09-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145148137","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}