IEEE Access

I.AO_SCPLOWBSTRACTC_SCPLOWCoronary Artery Disease (CAD) is a leading cause of cardiovascular-related mortality and affects 20.5 million people in the United States and approximately 315 million people worldwide in 2022. The asymptomatic and progressive nature of CAD presents challenges for early diagnosis and timely intervention. Traditional diagnostic methods such angiography and stress tests are known to be resource-intensive and prone to human error. This calls for a need for automated and time-effective detection methods. In this paper, this paper introduces a novel approach to the diagnosis of CAD based on a Convolutional Neural Network (CNN) with a temporal attention mechanism. The model will be developed on an architecture that will automatically extract and emphasize critical features from sequential medical imaging data from coronary angiograms, allowing subtle signs of CAD to be easily spotted, which could not have been detected by convention. The temporal attention mechanism strengthens the ability of a model to focus on relevant temporal patterns, thus improving sensitivity and robustness in detecting CAD for various stages of the disease. Experimental validation on a large and diverse dataset demonstrates the efficacy of the proposed method, with significant improvements in both detection accuracy and processing time compared to traditional CNN architectures. The results of this study propose a scalable solution system for the diagnosis of CAD. This proposed system can be integrated into clinical workflows to assist healthcare professionals. Ultimately, this research contributes to the field of AI-driven healthcare solutions and has the potential to reduce the global burden of CAD through early automated detection.

1Accurate assessment of patient-ventilator interaction is critical for optimizing respiratory support and detecting harmful dyssynchronies linked to adverse outcomes, including ventilator-induced lung injury and prolonged ICU stays. This requires precise, breath-by-breath segmentation and phase delineation of ventilator waveforms, specifically pressure, flow, and volume. Current reliance on manual annotation limits scalability and consistency, particularly given the variability of waveforms across diverse patient conditions and ventilator settings. To address this challenge, we present a fully automated, two-stage hybrid pipeline that integrates a rule-based algorithm with a Deep Learning (DL) model. The rule-based module generates pseudo-labels by detecting steep rises in the pressure derivative for breath segmentation and analyzing zero-crossings in the flow signal for phase delineation. These labels train a modified 1D U-Net enhanced with Bidirectional Long Short-Term Memory (Bi-LSTM), which captures temporal dependencies and improves adaptability to complex waveform morphologies, such as double-triggered ventilator dyssynchrony breaths. The framework was developed using data from adult ICU patients and evaluated on an independently annotated test set. The Bi-LSTM U-Net model achieved a Dice score of 0.9611, surpassing both the rule-based method, which scored 0.9321, and baseline U-Net architectures, which scored 0.9587. The model demonstrated high temporal precision, with inspiration offset and onset errors of 0.004{+/-} 0.013 seconds and 0.013 {+/-}0.028 seconds, respectively. The Bi-LSTM architecture proved particularly effective, reducing inspiration offset errors by 43% and onset errors by 28% compared to the rule-based method and baseline U-Net, while also maintaining low error variability. This hybrid approach provides a scalable, accurate, and fully automated solution for ventilator waveform analysis, enabling enhanced assessment of patient-ventilator synchrony without manual intervention.

This paper presents a comprehensive comparative study of five state-of-the-art CNN architectures, VGG19, ResNet50, InceptionV3, DenseNet121, and EfficientNetB0 for multi-class classification of Chest X-ray images (CXR) into four categories: Edema, Normal, Pneumonia, and Tuberculosis (TB). The models were trained, validated, and tested on a dataset comprising 6,092 training and 325 testing images across four distinct classes. Each architecture was initialized with ImageNet weights, augmented with a custom classifier, and fine-tuned under identical conditions to ensure a fair comparison. The models are evaluated on a comprehensive set of metrics, including accuracy, per-class recall, training time, and model complexity. Experimental results indicate that VGG19 achieved the highest classification accuracy of 98.15%, followed closely by ResNet50 at 97.54%. This study provides empirical evidence to guide the selection of appropriate deep learning models for chest X-ray diagnosis, balancing performance with operational constraints

Freezing of gait (FoG) prediction is clinically meaningful only when warnings arrive sufficiently early for subsequent action. Therefore, we adopt a Freezing Prediction Horizon (FPH) evaluation that reports prediction performance as a function of the warning horizon before onset, making the lead-time versus reliability trade-off explicit. Within this protocol, we develop a Transformer-based predictor with a progressive self-paced learning strategy and evaluate it on a 55-patient clinical dataset and two public datasets. The horizon-performance curves show that Macro-F1 remains stable up to approximately 2.5 seconds before FoG onset in our dataset, after which a gradual decline is observed. This horizon-based characterization replaces single, fixed ahead-of-onset windows with a continuous method that summarizes achievable advanced time at specified accuracy levels. In this way, it offers a principled basis for setting targets in real-time implementationslinking algorithmic early-warning capacity to the lead times that practical systems may require-while remaining compatible with conventional metrics. By centering evaluation on FPH, this study clarifies how far in advance FoG can be predicted with confidence, and it positions horizon-based assessment as a reproducible complement to standard reporting for future work on deployable FoG prediction. Ultimately, quantifying advance warning is a prerequisite for prevention-oriented use, by indicating whether sufficient time can be reserved for cueing prior to onset.

Causal discovery algorithms are often leveraged for inferring causal relationships and recovering a causal model from data. However, causal discovery from data alone is limited by the structural constraints of the used dataset, the lack of causal logic, and the lack of external knowledge. Thus, data-driven causal discovery can only suggest possible causal relationships at best. To overcome these limitations, Large Language Models (LLMs) and knowledge systems, such as Retrieval-Augmented Generation (RAG), have been proposed as alternatives to data-driven causal discovery and as a method to augment causal discovery algorithms. Using an expert-defined causal graph of chronic lower back pain, we further propose knowledge graph based RAG systems, such as GraphRAG, as an improvement over RAG systems for augmenting causal discovery (F1 0.745), benchmarking its performance against augmenting causal discovery with an LLM (F1 0.636), augmenting causal discovery with RAG (F1 0.714), and causal discovery alone (F1 0.396). We also explore the impact of different prompting methods for causality, such as querying for the plausibility of causal relationships, the presence of statistical associations, and the existence of temporal causal relationships, as inspired by the methodology of the domain experts constructing our ground truth. Lastly, we discuss how applications of LLMs, RAG, and graph-based RAG systems can impact and accelerate the causal modeling of chronic lower back pain by bridging the gap between domain knowledge and data driven approaches to causal modeling. Graphical Abstract O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=93 SRC="FIGDIR/small/26346255v1_ufig1.gif" ALT="Figure 1"> View larger version (31K): org.highwire.dtl.DTLVardef@f3387org.highwire.dtl.DTLVardef@2dforg.highwire.dtl.DTLVardef@bc839aorg.highwire.dtl.DTLVardef@63f6ea_HPS_FORMAT_FIGEXP M_FIG C_FIG

This research presents multimodal deep learning for structural heart disease prediction. We evaluated multiple deep learning architectures, including TCN, Simple CNN, ResNet1d18, Light transformer and Hybrid model. The models were examined across the three seeds to ensure robustness, and bootstrap confidence interval is used to measure performance differences. TCN consistently outperforms other competing architectures, achieving statistically significant improvements with stable performance across runs. Similarly in predictive analysis, TCN has efficient computation and stable training compared to all competing architectures. Our results show that TCN emphasizes fairness evaluation when developing deep learning models for healthcare applications.

Handheld ultrasound devices have revolutionized point-of-care diagnostics, but their effectiveness remains limited by operator dependency and the need for specialized training. This paper presents an intelligent guidance and diagnostic assistance system for the handheld wireless ultrasound device, enabling automated carotid artery and thyroid examinations through handheld operation. Drawing inspiration from the Actor-Critic framework, we implement a simulation-based reinforcement learning approach for real-time probe navigation toward standard anatomical views. The system integrates YOLOv8n-based detection networks for carotid plaque and thyroid nodule identification, achieving real-time inference at 30 frames per second. Furthermore, we propose a hybrid measurement approach combining UNet segmentation with the Snake algorithm for precise biometric quantification, including carotid intima-media thickness (IMT), lumen diameter, and lesion dimensions. Experimental validation on clinical datasets demonstrates that the proposed system achieves 91.2% accuracy in standard plane acquisition, 87.5% mean average precision (mAP) for plaque detection, and 89.3% mAP for nodule identification. Measurement results show excellent agreement with expert sonographers, with IMT measurements exhibiting a mean absolute difference of 0.08 mm. These findings demonstrate the feasibility of intelligent handheld ultrasound examination, significantly reducing operator dependency while maintaining diagnostic accuracy comparable to experienced clinicians.

BackgroundSegmentation is a routine, yet time-consuming and subjective step in the analysis of positron emission tomography (PET) images. Automatic methods to do it have been suggested, but recent method development has focused on supervised approaches. The previously published unsupervised segmentation methods for PET images are outdated for the arising dynamic human total-body PET images now enabled by the evolving scanner technology. MethodsIn this study, we introduce an unsupervised general purpose automatic segmentation method for modern PET images consisting of tens of millions of voxels. We provide its implementation in an easy-to-use format and demonstrate its performance on two datasets of real human total-body images scanned using different radiotracers. Results and conclusionsOur results show that the suggested method can identify functionally distinct areas within the anatomical organs. Combined with anatomical segments obtained from other imaging modalities, this enables great potential to improve clinically meaningful segmentation and reduce time-consuming manual work.

BackgroundDeep learning algorithms for tuberculosis (TB) screening frequently achieve radiologist-level performance during internal evaluation, yet their reliability often degrades when deployed to populations differing from the training domain. Such degradation is clinically consequential for screening tools, where the World Health Organization (WHO) emphasizes high sensitivity to minimize missed infectious cases. MethodsA DenseNet-121 convolutional neural network was trained using transfer learning on the Shenzhen chest X-ray dataset (China; total n=662). To prevent anatomically implausible augmentation, horizontal flipping was excluded during training. The model was trained in two stages (head training followed by fine-tuning) and evaluated on: (i) an internal test set from China, (ii) an external balanced cohort from Montgomery County (USA; n=138), and (iii) an external TB-positive cohort from India (n=155). The India dataset served as a sensitivity stress test; specificity and ROC-AUC were not computed for this cohort due to the absence of negative controls. Model attention was explored using Grad-CAM. ResultsInternal validation yielded an Area Under the Curve (AUC) of 0.889 and accuracy of 85.6%. External testing revealed divergent failure modes. On the USA cohort, sensitivity was high (94.8%) but specificity decreased significantly (43.7%), indicating false-positive inflation. Conversely, on the India TB-only cohort, sensitivity collapsed to 52.3%, implying that 47.7% of confirmed TB cases were missed under domain shift. All metrics are reported as point estimates ConclusionGeographic domain shift produced non-uniform degradation false-positive surges in a low-burden setting and sensitivity collapse in a high-burden setting. These findings highlight the safety risks of deploying single-source TB screening AI without local validation and calibration.

Parkinsons Disease (PD) is the most prevalent neurodegenerative disorder after Alzheimers, yet its diagnosis largely relies on subjective clinical assessments. Thus, this study proposes a systematic, data-driven approach to accurately classify PD patients using heterogeneous risk factors along with efficient machine learning. Six machine learning algorithms, Support Vector Machine(SVM), Random Forest(RF), Extreme Gradient Boosting(XGBoost), Logistic Regression(LR), K-Nearest Neighbour (KNN), and Decision Tree(DT) were utilized and evaluated their performances to identify the most robust and efficient model with high discrimination power. SVM model outper-formed all other machine learning models, and it has been identified as the highest-quality model to classify PD patients from others with at least 96% accuracy. Further-more, Feature importance was analyzed using SHAP to enhance the interpretability of the proposed model. This study contributes to the integration of artificial intelligence in the healthcare domain, emphasizing the value of data-driven classification modeling techniques in supporting healthcare professionals with accurate, personalized, and actionable insights for high-risk patients. Together, these approaches enhance the precision of early detection of PD, paving the way for more informed clinical decision-making and improved patient care.

BackgroundDistinguishing individuals with cognitive decline (CD), including early Alzheimers disease, from cognitively normal (CN) individuals is essential for improving diagnostic accuracy and enabling timely intervention. Positron emission tomography (PET) captures functional brain alterations associated with CD, but its broader application is often limited by cost and radiation exposure. To enhance the clinical utility of PET while addressing data limitations, we propose a multi-representational learning framework that leverages both imaging data and region-level quantification in a data-efficient manner. MethodsVoxel-level features were extracted using convolutional neural networks (CNN) or principal component analysis networks (PCANet) from [{superscript 1}F]FDG PET imaging. Region-level features were derived from standardized uptake value ratio measurements across predefined brain regions and processed using a deep neural network (DNN). These voxel- and region-level information are integrated through direct concatenation. For final prediction, different machine learning models and ensemble technique were applied. The models were trained and validated using 5-fold cross-validation on PET scans from 252 participants in the Alzheimers Disease Neuroimaging Initiative (ADNI), comprising 118 CN and 134 CD subjects. Additional correlation analysis and disease classification comparison with the Mini-Mental State Examination (MMSE) were also performed. ResultsIn 5-fold cross-validation, CNN, PCANet, and DNN models achieved classification accuracies of 0.69 {+/-} 0.04, 0.69 {+/-} 0.06, and 0.82 {+/-} 0.06, respectively. The integrated DNN-CNN model using direct concatenation yielded the highest accuracy (0.87 {+/-} 0.05), with a 6.10% improvement in accuracy and reduced standard deviation relative to the DNN-only model. Moreover, there were an increase of 14.29% in Recall (0.77 to 0.88) and an increase of 7.32% in F1-Score (0.82 to 0.88). Moreover, the model output showed a significant level of relation with MMSE, and it outperformed the MMSE-based classification in accuracy, recall, and f1, except precision. ConclusionCombining PET imaging with region-level quantification and deep learning improves diagnostic performance over single-feature based models. Notably, fusion-based approaches enhanced sensitivity to cognitive decline. This multimodal strategy offers a more data-efficient and accurate approach for classifying cognitive decline and supports broader PET application in clinical settings.

BackgroundThe deployment of high-fidelity diagnostic Artificial Intelligence (AI) in resource-constrained environments is hindered by the stochastic nature of network latency and bandwidth limitations. Traditional tele-radiology relies on static cloud offloading, which introduces unacceptable latency for critical care scenarios. Zero-Touch Network and Service Management (ZSM) offers a paradigm for automated network orchestration, yet current frameworks lack application-layer awareness regarding clinical urgency and image complexity. MethodologyThis study proposes a novel Cross-Modal Latent Transformer (CMLT) integrated within a Zero-Touch Network Orchestration architecture. The system utilizes a lightweight Edge-Gating mechanism to dynamically partition inference tasks between edge nodes and cloud resources based on feature entropy. The model was trained and validated on the MIMIC-CXR (v2.0.0) (n = 377, 110) and CheXpert (n = 224, 316) datasets, employing a 70/10/20 split. ResultsThe proposed orchestration framework achieved an AUC-ROC of 0.962 [95% CI: 0.941-0.983] for Atelectasis detection, comparable to full-cloud inference, while reducing network bandwidth consumption by 64.3%. McNemars test indicated no statistically significant difference in diagnostic accuracy between the orchestrated hybrid approach and the full-precision cloud baseline (p > 0.05), despite a 120 ms reduction in mean inference latency. Clinical SignificanceBy embedding clinical feature extraction directly into the network orchestration logic, this framework enables real-time, zero-touch provisioning of diagnostic resources, facilitating reliable AI deployment in rural and bandwidth-limited clinical settings.

INTRODUCTIONConnected speech analyses can help characterize linguistic impairments in primary progressive aphasia (PPA) and classify variants, however, manual transcription of speech samples is time-consuming and expensive. Automated speech recognition (ASR) may be efficacious for transcribing PPA speech. METHODSTranscripts of picture descriptions (109 PPA, 32 healthy controls (HC)) were generated using a manual, automated (Whisper) or semi-automated approach including a quality control (QC) step. We evaluated transcript accuracy, the reliability of ASR-derived linguistic features, and classification performance. RESULTSWhisper demonstrated lowest error rates for HC, followed by semantic, logopenic and non-fluent PPA variants. Errors correlated with overall disease severity for semantic and logopenic variants. QC of Whisper outputs reduced errors and improved the reliability of linguistic features. Overall, ASR-derived features achieved better classification performance than manual transcription features. DISCUSSIONResults support the use of off-the-shelf ASR for scalable, cost-efficient transcription of PPA speech and classification.

This study presents RED-RHD, a machine learning methodology for early detection and classification of Rheumatic Heart Disease (RHD) using heart sound recordings. By leveraging OpenL3 deep acoustic embeddings, cloud-based workflows, and an ensemble of SVM and XGBoost classifiers, RED-RHD achieves an average precision of 95.62% for murmur detection (Normal vs. Abnormal) and 99.00% precision for systolic vs. diastolic murmur classification, demonstrating marked improvements over prior methods with poor cross-dataset generalization (e.g., specificity as low as 4.3% in ResNet-based approaches). These results confirm the systems robustness across diverse, noisy clinical datasets. Additionally, we introduce a novel dynamic adaptive model selection mechanism that enables the framework to automatically select the most appropriate pretrained machine learning model based on extracted heart sound features, optimizing prediction accuracy for different regional or demographic populations. By incorporating this adaptive intelligence, RED-RHD addresses population variability and supports precision diagnostics in globally diverse patient groups, advancing the potential for scalable, AI-driven auscultation in low-resource environments.

ObjectiveThis study investigates whether incorporating physiological coupling concepts into neural network design can support stable and interpretable feature learning for histopathological image classification under limited data conditions. MethodsA physiologically inspired architecture, termed CardioPulmoNet, is introduced to model interacting feature streams analogous to pulmonary ventilation and cardiac perfusion. Local and global tissue features are integrated through bidirectional multi-head attention, while a homeostatic regularization term encourages balanced information exchange between streams. The model was evaluated on three histopathological datasets involving oral squamous cell carcinoma, oral submucous fibrosis, and heart failure. In addition to end-to-end training, learned representations were assessed using linear support vector machines to examine feature separability. ResultsCardioPulmoNet achieved performance comparable to several pretrained convolutional neural networks across the evaluated datasets. When combined with a linear classifier, improved classification performance and higher area under the receiver operating characteristic curve were observed, suggesting that the learned feature embeddings are well structured for downstream discrimination. ConclusionThese results indicate that physiologically motivated architectural constraints may contribute to stable and discriminative representation learning in computational pathology, particularly when training data are limited. The proposed framework provides a step toward integrating physiological modeling principles into medical image analysis and may support future development of transferable and interpretable learning systems for histopathological diagnosis.

Burn injuries are a significant concern in developing countries due to limited infrastructure, and treating them remains a major challenge. The manual assessment of burn severity is subjective and depends, to a large extent, on individual expertise. Artificial intelligence can automate this task with greater accuracy and improved predictions, which can assist healthcare professionals in making more informed decisions while triaging burn injuries. This study established a model pipeline for detecting burn injuries in images using multiple deep learning models, including U-Net, DenseNet, ResNet, VGG, EfficientNet, and transfer learning with the Segment Anything Model2 (SAM2). The problem statement was divided into two stages: 1) removing the background and 2) burn skin segmentation. ResNet50, used as an encoder with a U-Net decoder, performs better for the background removal task, achieving an accuracy of 0.9757 and an intersection over union (Jaccard index) of 0.9480. DenseNet169, used as an encoder with a U-Net decoder, performs well in burn skin segmentation, achieving an accuracy of 0.9662 and an intersection over Union of 0.8504. The dataset collected during the project is available for download to facilitate further research and advancements (Link to dataset: https://geninfo.iitr.ac.in/projects). TBSA was estimated from predicted burn masks using scale-based calibration

Can self-supervised speech foundation models (SFMs) be used for automatic patient metadata extraction, even when no prior demographic information is available and speech is affected by pathology? SFMs show strong cross-task generalization, yet it remains unclear to what extent demographic attributes such as age and gender are intrinsically encoded, particularly in pathological speech. This study evaluated the capability of a pretrained SFM Wav2Vec 2.0 to estimate age and gender across healthy controls (HC), Parkinsons disease (PD) subjects, and related parkinsonian syndromes (multiple system atrophy, progressive supranuclear palsy), without exposing the model to any data from the evaluated datasets. A frozen, publicly available Wav2Vec 2.0 model was used to extract speech representations from three independent multilingual datasets. No machine learning model was trained or fine-tuned on the target data. The analysis solely assessed information already present in the pretrained embeddings. Multiple speech tasks (read text, diadochokinesis, sustained vowels) and diagnostic groups were evaluated using gender accuracy, correlations with true age, chi-square tests, and group-level analyses. Gender estimation achieved consistently high accuracy (min. 94%, up to 100%) across datasets and tasks. Age estimation showed significant correlations with true age for read speech, including PD speakers. Analyses of vowel data demonstrated preserved gender distributions but systematic age bias across all diagnostic groups. Without task-specific training, pretrained Wav2Vec 2.0 embeddings robustly encode gender and preserve age-related structure in connected, including pathological, speech, whereas age estimation from isolated vowel phonation remains unreliable. These findings highlight the demographic robustness of SFMs and the need to monitor age-related bias in clinical applications.

One of the objectives of digital neuropsychology is to apply computational methods to improve the accuracy of traditional assessments. We created and evaluated a computerized TMT (cTMT) that preserves its original structure and records high-resolution mouse trajectories. Seventy-four older adults (41 with mild cognitive impairment and 33 healthy controls) completed the cTMT and a standard diagnostic battery. We also developed NeuroTask, a Python library that extracts features from cursor time series, including reaction times, speed and acceleration metrics, trajectory deviations, and state-based measures. We compared demographic, digital, and non-digital models using nested cross-validation and non-parametric permutation tests. Demographic models provided only modest discrimination (AUC = 0.56). Digital hand features improved performance (AUC = 0.65), and combining them with demographics reached an AUC of 0.70, which approached the performance of the neuropsychological battery used to define the diagnosis (AUC = 0.76). In complementary regression analyses with digital plus demographic features, we obtained significant predictions for five of seven target scores: MMSE, Digit Symbol, TMT-A, TMT-B, and Forward Digit Span. These results indicate that fine-grained hand-movement features from the cTMT provide useful information for classifying mild cognitive impairment and for predicting multiple neuropsychological scores.

The electrocardiogram (ECG) is a critical tool in the diagnosis and monitoring of cardiovascular disease. Although traditional 12-lead ECGs offer comprehensive in-sights into the electrical activity of the heart, they typically require clinical settings and expert interpretation, which limits their accessibility. In contrast, smartwatch 1-lead ECGs can be recorded at home, allowing more frequent and rapid monitoring. This opens opportunities not only for early detection but also for enhancing patient autonomy. This study investigates whether 1-lead ECGs can provide information beyond heart rhythm, specifically whether they can be used to assess left ventricular function (LVF) using explainable deep learning models. Our findings show that LVF can be accurately predicted from 1-lead ECGs (AUC = 0.883), nearly matching the performance of 12-lead ECGs (AUC = 0.897). These results suggest that 1-lead ECGs, when combined with interpretable AI, could support broader clinical applications and empower patients, particularly in resource-limited or remote settings.

BackgroundSpeech abnormalities are common in Parkinsons disease (PD). Machine learning (ML) offers potential for objective and scalable speech-based diagnostics. This study introduces an explainable ML pipeline that leverages a novel vowel articulation-based composite input to detect PD and identify phoneme-level biomarkers. MethodsTwo publicly available datasets of PD speech recordings were analyzed. Sustained vowel articulations were converted into log-mel spectrograms either individually or as a composite image by vertically concatenating a set of vowels per subject. Processed spectrograms were used to train ML models, with performance assessed using five-fold cross-validation and bootstrapped area under the receiver operating characteristics curve (AUROC). Gradient-weighted Class Activation Mapping (Grad-CAM) was applied to quantify model attention across vowel regions. ResultsA total of 150 patients (49.3% PD) were included. Acoustic analysis revealed significant group differences in cepstral peak prominence and harmonics-to-noise ratio, particularly for vowel /u/ (p < 0.05). The ML model achieved an average AUROC of 0.805 using individual vowels and improved to 0.928 with the composite input (p < 0.001). Grad-CAM demonstrated the highest activation for vowel /u/ (p < 0.001), consistent with acoustic findings. ConclusionThe proposed explainable composite spectrogram approach dually enabled high classification performance and identification of a vowel biomarker. Concordance between ML and acoustic analyses highlights the translational potential of explainable ML in PD speech assessment and its ability to reveal underlying pathophysiological insights.