IEEE Transactions on Biomedical Engineering最新文献

Neuroanatomical heterogeneity in Alzheimer's disease (AD) hinders precision diagnosis and treatment, as distinct brain phenotypes may correspond to different disease subtypes. However, MRI-based subtype classifications are often confounded by co-occurring pathologies and non-AD factors, such as genetic predisposition and environmental influences, limiting their clinical interpretability. We propose 3D-DisAD, an unsupervised deep learning framework that disentangles AD-specific neuroanatomical variations from unrelated influences and clusters patients into subtypes with homogeneous brain phenotypes. The framework comprises two synergistic networks: (1) Contrastive Disentanglement Network, which separates AD-specific variations from those shared by AD patients and healthy controls; and (2) Transformation Generation Network, which refines these disease-specific variations by transforming healthy brain representations into realistic, pathology-consistent anatomies via diffusion-based generative modeling. Evaluated on four public datasets, 3D-DisAD reveals strong correlations between the disentangled AD-specific variations and diverse clinical and biological profiles, validating their relevance. Using these variations, we identify four AD subtypes with significant differences in biomarkers, cognitive trajectories, and genetic signatures, and uncover distinct longitudinal progression patterns that suggest potential windows for early intervention. By disentangling AD-specific variations, our method enables more precise patient stratification and personalized treatments, particularly in the early stage of AD. Code is available at: https://github.com/cnuzh/3D-DisAD.

Stress detection is a widely studied field due to its significant implications for mental and physical health. While multimodal approaches show promising results, they present challenges related to hardware constraints and computational requirements that limits real time implementation in wearable devices. We propose a hybrid methodology combining feature extraction with ma chine learning (ML) for stress detection, using exclusively single-lead electrocardiogram (ECG) signals from which heart rate variability (HRV) and respiratory signals with their derived features are extracted. We evaluated our approach using the ES3 project database, testing various feature combinations with the XGBoost model. Results demonstrate that incorporating ECG-derived respiratory features significantly improves classification accuracy and computational efficiency compared to traditional HRV-based approaches and deep learning models. Feature importance analysis identified a reduced set of key features, resulting in a more efficient model with superior performance and substantially lower inference times than deep learning models. These findings support the feasibility of acute stress detection using a single-lead ECG-based multimodal approach that combines feature extraction with ML techniques, providing insights into stress-induced physiological responses and contributing to more accessible biomedical monitoring strategies.

Objective: Positron emission tomography (PET) is a commonly used imaging modality for assessment of neuroendocrine tumors (NETs), and lesion identification in PET images is a key step in the development of effective treatments. Deep neural networks have recently produced encouraging performance of automated lesion identification with PET imaging. However, most methods require a predefined region/volume of interest (ROI/VOI) or rely on a multi-stage, cascaded modeling pipeline, which often leads to low efficiency and/or high variability. In this paper, we propose a novel single-stage PET lesion detection method that does not need precomputed ROIs/VOIs, cascaded models or multimodal data.

Methods: We introduce a novel three-dimensional dual-decoder neural network, which contains a cross-decoder attention module to take as input gating signals from an auxiliary organ segmentation decoder and suppress irrelevant feature responses in the primary decoder of lesion detection. Additionally, we design and insert a new patchwise contrastive learning module into the primary decoder to enhance the network's discriminative power for lesions with varying volumes and shapes.

Results: We evaluate the proposed lesion identification method using multiple hepatic NET $^{68}$Ga-DOTATATE PET image datasets that are acquired from two different scanners. The method produces superior performance compared with the reference baseline and recent state-of-the-art approaches.

Conclusion: We propose a novel single-stage framework, where both the cross-decoder attention and the patchwise contrastive learning are beneficial to improvement of lesion identification performance in PET images.

Significance: The proposed study has the potential to significantly improve the efficiency of clinical interpretation of PET imaging data.

Objective: treatment of obstructive coronary artery disease (CAD) requires accurate planning to ensure effective revascularization and full restoration of myocardial perfusion. In this study, we introduce Virtual PCI, a novel computational tool designed to support pre-operative planning of Percutaneous Coronary Intervention (PCI) by predicting the hemodynamic consequences of selected revascularization treatments.

Methods: the tool leverages a fully personalized 3D multiscale perfusion model, calibrated using pre-intervention stress CT perfusion (CTP) imaging, to simulate the hemodynamic impact of different revascularization strategies in terms of post-intervention stress myocardial blood flow (MBF) and FFR. The computational framework is also capable of computing the FFR index. We conduct a validation study on patients treated with elective PCI and compare model predictions with dynamic stress CTP at follow-up.

Results: the validation study demonstrates high accuracy in predicting post-PCI myocardial perfusion, including potential residual ischemia and cardiac mass at ischemic risk. Through an integrated analysis with FFR, the tool shows potential for its prospective use, identifying in two patients optimal treatment strategies and, in one case, outperforming the executed revascularization in reduction of ischemic burden.

Conclusions: Virtual PCI enables the prediction of post-PCI myocardial blood flow (MBF) and FFR, offering a comprehensive assessment of treatment outcomes to identify the best revascularization option from the hemodynamic standpoint.

Significance: since it relies solely on non-invasive imaging (cCTA, stress-CTP), Virtual PCI can be integrated early in the diagnostic workflow, providing cardiologists with a powerful, patient-specific tool to optimize PCI planning.

Objective: Brain-computer interfaces (BCIs) based on event-related potentials (ERPs) are among the most accurate and reliable BCIs. However, current mainstream classification algorithms struggle to eliminate the need for calibration and rely on expensive labeled data, limiting the practical usability of ERP based BCIs. The development of fully unsupervised algorithms is essential for the advancement of practical applications of BCI systems.

Methods: In this study, we propose a novel unsupervised classification method called sliding-window distribution distance maximization (sDDM). This algorithm utilizes sliding windows to highlight important temporal features and transforms the metric of inter-class differences from absolute distances to relative distribution distances in Mahalanobis space, while incorporating information on target event similarity from the BCI paradigm. Additionally, our proposed spatial dimensionality reduction strategy ensures smaller spatial dimensions and more prominent spatial features.

Results: We compare our proposed method to other state of-the-art unsupervised classification methods and evaluate it offline on our self-collected dataset, a public dataset recorded during the use of a P300 Speller by patients with ALS, and the BCI Competition III Dataset II. Our results demonstrate that our proposed method achieves the best spelling accuracy across all datasets, surpassing other unsupervised algorithms. We further explore its improvement effectiveness through ablation experiments.

Conclusion: Our proposed method enhances the performance of unsupervised classification in ERP-based BCIs.

Objective: Upper limb loss due to traumatic injury or disease poses significant challenges to autonomy, daily function, and workforce reintegration, profoundly impacting overall quality of life. While myoelectric prosthetic hands have the potential to restore dexterity, many users discontinue use due to limited functionality and durability.This manuscript describes the design and rationale of an ongoing clinical trial aimed at addressing these gaps in real-world settings.

Methods: We searched for completed and ongoing clinical trials on ClinicalTrials.gov to study their structure and their gaps, and then we presented the protocol of our ongoing clinical trial. This protocol outlines a randomized crossover clinical trial enrolling 36 adults with upper limb loss to evaluate two multi-articulated myoelectric prosthetic hands.

Results: Our review of clinical trials revealed that the unique strength of our design is the integration of standardized laboratory tests, extended daily use, onboard usage data, and validated satisfaction surveys. We provided a detailed description of all design choices and rationale of the ongoing clinical study.

Conclusion: The comparison between our design and the design of other studies indicates that our design is unique in the integration of biomechanical assessments, real-world usage monitoring, and user-reported outcomes. This clinical trial should be capable of assessing if one specific device design can offer clinically meaningful advantages over another.

Significance: The design of our clinical trial could inform the design of clinical trials targeting the optimization of prostheses and their acceptance by prosthetic users.

Objective: transkeyhole microsurgery for spinal cord tumors requires intraoperative imaging guidance to ensure safe and effective tumor resection. Although optical endoscopy has been widely adopted in clinical settings, it's limited to visualizing superficial structures. Endoscopic ultrasound (EUS) offers a promising alternative. However, EUS transducers are typically fabricated from high-frequency arrays, which provide limited imaging depth and field of view. In addition, the high cost of the transducers and complicated sterilization further restrict their use in surgery.

Methods: this paper introduces an economical single-element US transducer that utilizes electromagnetic actuation operating in a resonant scanning mode. An image-based method is proposed to correct the resulting nonlinear scanning. Two prototypes were developed, having outer diameters of 14 (T14) and 9 (T9) mm. The imaging performance of the transducers was evaluated by wire phantoms, tissue-mimicking phantom, ex-vivo sheep spine, and in-vivo rabbits.

Results: The T14 and T9 achieved scanning angles over 70$^circ$ and approximately 60$^circ$, respectively, with the former maintaining a lateral resolution of 248 $mu$m and the latter yielding an optimal contrast-to-noise ratio of 2.43. The captured US imaging clearly visualized dural sac and unilateral nerve roots in sheep spine and enabled the accurate identification of subdural hemorrhage and key anatomy in rabbits.

Conclusion: the electromagnetically actuated transducer achieves a wide scanning range despite its compact size, showing great promise for surgery by facilitating the identification of subdural anatomy and enabling customized dural opening strategies.

Significance: cost reduction enables the feasible use of the transducer as a single sterile device in surgical settings.

Objective: We present a novel wearable T-shaped ultrasound (WTSUS) patch for simultaneous short-axis and long-axis imaging monitoring of carotid artery in situ within the same cardiac cycle to measure the carotid blood flow volume.

Methods: WTSUS patch consists of two same ultrathin ultrasound transducer arrays with a center frequency of 8.5 MHz. The B-mode imaging provides real-time measurement of the cross-section area of the carotid artery, while Doppler imaging captures velocity time integral.

Results: WTSUS patch exhibits a total thickness of 1.3 mm and a wide -6 dB bandwidth of 65%. The axial and lateral resolutions at a depth of 20 mm were 0.37 mm and 0.45 mm, respectively. In vitro flow volume experiments showed that the maximum measurement deviation using WTSUS patch was 6.3%. In vivo imaging of the human common carotid artery exhibited good agreement with a commercial ultrasound system, demonstrating the reliability of WTSUS-based wearable ultrasound system.

Conclusion: This study exhibits a wearable ultrasound imaging patch with a reliable continuous monitoring of the carotid blood flow volume that is also comfortable and easy to wear.

Significance: This work can provide a novel and reliable solution for noninvasive cardiac output estimation, with significant potential for applications in critical care and continuous monitoring of dynamic blood flow volume.

Objective: Fluorescence cell counting is vital in biomedical research, yet existing automated methods lack sufficient adaptability and accuracy, leading to persistent errors in complex microscopy images. This study aims to propose an adaptive, interactive approach to effectively overcome these limitations.

Methods: We introduce the Adaptive Interactive Cell Counting (AICC) framework, combining a coordinate-based prediction module with user-guided correction. Specifically, we develop two novel global correction algorithms, Proposal Expansion (PE) and Prediction Filtering (PF), coupled with a new RGB-Aware Structural Similarity (RGB-Aware SSIM) metric to identify visually similar regions and efficiently propagate minimal user corrections. Additionally, we release NEFCell, a new high-resolution fluorescence microscopy dataset designed explicitly for evaluating interactive cell counting methods.

Results: Extensive evaluations show that AICC significantly surpasses current state-of-the-art methods, reducing counting errors by up to 36.8% compared to non-interactive approaches and up to 65.3% compared to existing interactive methods, while improving localization accuracy by 7.3% on average and significantly minimizing interaction time.

Conclusion: The proposed AICC framework substantially enhances accuracy and reduces effort required for fluorescence cell counting, proving its effectiveness in integrating automation with user expertise.

Significance: AICC represents a valuable tool for biomedical researchers and clinicians, facilitating precise and efficient cell analyses in complex experimental and clinical contexts.

Tactile stimulation, especially electrotactile stimulation, have been a subject of interest in recent literature for machine-to-human communication (M2HC) of electronically gathered information for the purpose of augmenting and improving the human experience. Electrotactile is a direct noninvasive method for peripheral nerve stimulation that provides a pathway for communication with the brain. However, the widespread use of electrotactile as an M2HC pathway is hampered by the availability and ease of use of mainstream, visual and audio, communication methods and technological challenges with electrotactile stimulation that must be resolved, such as skin condition dependency, neural adaptation, and the lack of a framework for producing consistent electrotactile M2HC. As such, this paper (1) reviews the scientific and engineering literature associated with electrotactile stimulation and associated electronics with a goal of converging disciplinary knowledge of this topic, (2) summarizes recent advances and open challenges in electrotactile stimulation, and (3) discusses available techniques and introduces a unifying model for icon-based electrotactile communication. In contrast to prior review papers on the subject, this paper uniquely focuses on defining electrotactile stimulation as a method for robust machine-to-human communication while compiling and discussing relevant engineering, physiology, and neuroscience issues, thus providing a comprehensive understanding of electrotactile M2HC for the IEEE community.