IEEE Journal of Biomedical and Health Informatics最新文献

With the rapid development of Large Language Model (LLM) technology, it has become an indispensable force in biomedical data analysis research. However, biomedical researchers currently have limited knowledge about LLM. Therefore, there is an urgent need for a summary of LLM applications in biomedical data analysis. Herein, we propose this review by summarizing the latest research work on LLM in biomedicine. In this review, LLM techniques are first outlined. We then discuss biomedical datasets and frameworks for biomedical data analysis, followed by a detailed analysis of LLM applications in genomics, proteomics, transcriptomics, radiomics, single-cell analysis, medical texts and drug discovery. Finally, the challenges of LLM in biomedical data analysis are discussed. In summary, this review is intended for researchers interested in LLM technology and aims to help them understand and apply LLM in biomedical data analysis research.

Despite achieving impressive results in general-purpose semantic segmentation with strong generalization on natural images, the Segment Anything Model (SAM) has shown less precision and stability in medical image segmentation. In particular, the SAM architecture is designed for 2D natural images and is therefore not support to handle three-dimensional information, which is particularly important for medical imaging modalities that are often volumetric or video data. In this paper, we introduce MediViSTA, a parameter-efficient fine-tuning method designed to adapt the vision foundation model for medical video, with a specific focus on echocardiography segmentation. To achieve spatial adaptation, we propose a frequency feature fusion technique that injects spatial frequency information from a CNN branch. For temporal adaptation, we integrate temporal adapters within the transformer blocks of the image encoder. Using a fine-tuning strategy, only a small subset of pre-trained parameters is updated, allowing efficient adaptation to echocardiography data. The effectiveness of our method has been comprehensively evaluated on three datasets, comprising two public datasets and one multi-center in-house dataset. Our method consistently outperforms various state-of-the-art approaches without using any prompts. Furthermore, our model exhibits strong generalization capabilities on unseen datasets, surpassing the second-best approach by 2.15% in Dice and 0.09 in temporal consistency. The results demonstrate the potential of MediViSTA to significantly advance echocardiography video segmentation, offering improved accuracy and robustness in cardiac assessment applications.

This paper introduces a dual-modal early cognitive impairment detection system based on autobiographical memory (AM) tests, and our approach is to automatically extract pre-defined acoustic features and self-designed embeddings to enhance linguistic representation of the spontaneous speech data. By integrating dual-modal data, we effectively enrich the features that aid in model learning, especially addressing the subtle symptoms exhibited by individuals with mild cognitive impairment (MCI), an intermediate stage between healthy individuals and those with Alzheimer's disease (AD). To account for spontaneous speech's unstructured and implicit nature, two additional embeddings, namely, speaker embedding and conversation embedding, are introduced to augment the information available for model learning, thus enriching the feature set for improving the model accuracy. The proposed dual-modal approach is tested on a self-collected Chinese spontaneous speech dataset due to the limited unstructured speech open-access dataset for MCI detection. The system's effectiveness is evaluated through a series of experiments, including ablation studies, to determine the impact of each module on overall performance. The proposed system achieved an average accuracy of 78% in detecting MCI, demonstrating its comparative effectiveness. Enhancements in our system are achieved by integrating a directional encoder tailored to capture temporal information across sequential visits. This addition leads to a 3% increase in detection accuracy within a subset of participants who have undergone multiple AM test sessions. Implementing such a longitudinal approach in analyzing unstructured speech data for MCI detection taps into a relatively underexplored area of research, offering new insights.

Remote photoplethysmography (rPPG) aims to measure non-contact physiological signals from facial videos, which has shown great potential in many applications. Most existing methods directly extract video-based rPPG features by designing neural networks for heart rate estimation. Although they can achieve acceptable results, the recovery of rPPG signal faces intractable challenges when interference from real-world scenarios takes place on facial video. Specifically, facial videos are inevitably affected by non-physiological factors (e.g., camera device noise, defocus, and motion blur), leading to the distortion of extracted rPPG signals. Recent rPPG extraction methods are easily affected by interference and degradation, resulting in noisy rPPG signals. In this paper, we propose a novel method named CodePhys, which innovatively treats rPPG measurement as a code query task in a noise-free proxy space (i.e., codebook) constructed by ground-truth PPG signals. We consider noisy rPPG features as queries and generate high-fidelity rPPG features by matching them with noise-free PPG features from the codebook. Our approach also incorporates a spatial-aware encoder network with a spatial attention mechanism to highlight physiologically active areas and uses a distillation loss to reduce the influence of non-periodic visual interference. Experimental results on four benchmark datasets demonstrate that CodePhys outperforms state-of-the-art methods in both intra-dataset and cross-dataset settings.

Intrinsically disordered regions (IDRs) of proteins are crucial for a wide range of biological functions, with molecular recognition features (MoRFs) being of particular significance in protein interactions and cellular regulation. However, the identification of MoRFs has been a significant challenge in computational biology owing to their disorder-to-order transition properties. Currently, only a limited number of experimentally validated MoRFs are known, which has prompted the development of computational methods for predicting MoRFs from protein chains. Considering the limitations of existing MoRF predictors regarding prediction accuracy and adaptability to diverse protein sequence lengths, this study introduces Trans-MoRFs, a novel MoRF predictor based on the transformer architecture, for identifying MoRFs within IDRs of proteins. Trans-MoRFs employ the self-attention mechanism of the transformer to efficiently capture the interactions of distant residues in protein sequences. They demonstrate stability and high efficiency in dealing with protein sequences of different lengths and performs well on both short and long sequences. On multiple benchmark datasets, the model attained a mean area under the curve score of 0.94, which is higher than those of all existing models, and significantly outperformed existing combined and single MoRF prediction tools on multiple performance metrics. Trans-MoRFs have excellent accuracy and a wide range of applications for predicting MoRFs and other functionally important fragments in the disordered regions of proteins. They offer significant assistance in comprehending protein functions, precisely pinpointing functional segments within disordered protein regions and facilitating the discovery of novel drug targets. We also offer a web server for related research: http://112.124.26.17:8007.

The identification of protein-peptide interacting pairs and their corresponding binding residues is fundamentally crucial and can greatly facilitate the peptide therapeutics designing and understanding the mechanisms of protein functions. Recently, several computational approaches have been proposed to solve the protein-peptide interaction prediction problem. However, most existing prediction methods cannot directly predict the protein-peptide interacting pairs as well as the binding residues from protein and peptide sequences simultaneously. Here, we developed a Comprehensive Protein-Peptide Interaction prediction Framework (CPPIF), to predict both binary protein-peptide interaction and their binding residues. We also constructed a benchmark dataset containing more than 8,900 protein-peptide interacting pairs with non-covalent interactions and their corresponding binding residues to systematically evaluate the performances of existing models. Comprehensive evaluation on the benchmark datasets demonstrated that CPPIF can successfully predict the non-covalent protein-peptide interactions that cannot be effectively captured by previous prediction methods. Moreover, CPPIF outperformed other state-of-the-art methods in predicting binding residues in the peptides and achieved good performance in the identification of important binding residues in the proteins.

Multimodal MRIs play a crucial role in clinical diagnosis and treatment. Feature disentanglement (FD)-based methods, aiming at learning superior feature representations for multimodal data analysis, have achieved significant success in multimodal learning (MML). Typically, existing FD-based methods separate multimodal data into modality-shared and modality-specific features, and employ concatenation or attention mechanisms to integrate these features. However, our preliminary experiments indicate that these methods could lead to a loss of shared information among subsets of modalities when the inputs contain more than two modalities, and such information is critical for prediction accuracy. Furthermore, these methods do not adequately interpret the relationships between the decoupled features at the fusion stage. To address these limitations, we propose a novel Complete Feature Disentanglement (CFD) strategy that recovers the lost information during feature decoupling. Specifically, the CFD strategy not only identifies modality-shared and modality-specific features, but also decouples shared features among subsets of multimodal inputs, termed as modality-partial-shared features. We further introduce a new Dynamic Mixture-of-Experts Fusion (DMF) module that dynamically integrates these decoupled features, by explicitly learning the local-global relationships among the features. The effectiveness of our approach is validated through classification tasks on three multimodal MRI datasets. Extensive experimental results demonstrate that our approach outperforms other state-of-the-art MML methods with obvious margins, showcasing its superior performance.