Biocybernetics and Biomedical Engineering最新文献

Objectives

Vascular depression hypothesis (VDH) bases on co-occurrence of vascular and mental dysfunctions in advanced age; however, there may be still a controversy about whether there is some direct association between vascular and mental properties or the co-occurrence is only a statistical artifact caused by commonness of these dysfunctions in the elderly. COVID-19 gave opportunity to test VDH under conditions different from aging.

Methods

25 patients were examined 3–6 month after SARS-CoV-2 infection. Subjective worsening of mental functions, presumably caused by the disease, was quantified with three psychometric tests. Blood flow waveforms were obtained for the left brachial and common carotid arteries. The waveform shape changes continuously with age; therefore, an individual shape can be characterized by the index WA being the calendar age (CA) of the average healthy rested subject having the most similar shape (consequently, in healthy rested subjects WA-CA = 0, in average). The mathematical functional analysis was used to calculate WA.

Results

Brachial WA-CA = 13 yrs, in average (p < 0.00005; Cohen’s d = 0.99), and was correlated with tests scores (r = 0.55, 0.65, 0.46). Mean carotid WA-CA were smaller (7.2 and 1.6) but they were also correlated with the scores (right: r = 0.44, 0.55, 0.32; left: r = 0.49, 0.51, 0.38). Scores of two tests were inversely correlated with the systolic (r = -0.54, −0.58) and diastolic (r = -0.46, −0.56) pressures.

Conclusions

Since neither vascular nor mental problems are common after COVID-19, these relatively high correlations indicate that vascular and mental properties are not independent, i.e., they support VDH. Note that this not only concerns cerebral vasculature.

This paper explores the parallel collaboration of multimodal physiological signals, combining eye tracker output signals, motor imagery, and error-related potentials to control a computer mouse. Specifically, a parallel working mechanism is implemented in the decision layer, where the eye tracker manages cursor movements, and motor imagery manages click functions. Meanwhile, the eye tracker output signals are integrated with electroencephalography data to detect the idle state for asynchronous control. Additionally, error-related potentials evoked by visual feedback, are detected to reduce the cost of error corrections. To efficiently collect data and provide continuous evaluations, we performed offline training and online testing in the designed paradigm. To further validate the practicability, we conducted online experiments on the real-world computer, focusing on a scenario of opening and closing files. The experiments involved seventeen subjects. The results showed that the stability of the eye tracker was optimized from 67.6% to 95.2% by the designed filter, providing the support for parallel control. The accuracy of motor imagery conducted simultaneously with fixations reached 93.41 ± 2.91%, proving the feasibility of parallel control. Furthermore, the real-world experiments took 45.86 ± 14.94 s to complete three movements and clicks, and showed a significant improvement compared to the baseline experiment without automatic error correction, validating the practicability of the system and the efficacy of error-related potentials detection. Moreover, this system freed users from the stimulus paradigm, enabling a more natural interaction. To sum up, the parallel collaboration of multimodal physiological signals is novel and feasible, the designed mouse is practical and promising.

For individuals with Type-1 diabetes mellitus, accurate prediction of future blood glucose values is crucial to aid its regulation with insulin administration, tailored to the individual’s specific needs. The authors propose a novel approach for the integration of a neural architecture search framework with deep reinforcement learning to autonomously generate and train architectures, optimized for each subject over model size and analytical prediction performance, for the blood glucose prediction task in individuals with Type-1 diabetes. The authors evaluate the proposed approach on the OhioT1DM dataset, which includes blood glucose monitoring records at 5-min intervals over 8 weeks for 12 patients with Type-1 diabetes mellitus. Prior work focused on predicting blood glucose levels in 30 and 45-min prediction horizons, equivalent to 6 and 9 data points, respectively. Compared to the previously achieved best error, the proposed method demonstrates improvements of 18.4 % and 22.5 % on average for mean absolute error in the 30-min and 45-min prediction horizons, respectively, through the proposed deep reinforcement learning framework. Using the deep reinforcement learning framework, the best-case and worst-case analytical performance measured over root mean square error and mean absolute error was obtained for subject ID 570 and subject ID 584, respectively. Models optimized for performance on the prediction task and model size were obtained after implementing neural architecture search in conjunction with deep reinforcement learning on these two extreme cases. The authors demonstrate improvements of 4.8 % using Long Short Term Memory-based architectures and 5.7 % with Gated Recurrent Units-based architectures for patient ID 570 on the analytical prediction performance by integrating neural architecture search with deep reinforcement learning framework. The patient with the lowest performance (ID 584) on the deep reinforcement learning method had an even greater performance boost, with improvements of 10.0 % and 12.6 % observed for the Long Short-Term Memory and Gated Recurrent Units, respectively. The subject-specific optimized models over performance and model size from the neural architecture search in conjunction with deep reinforcement learning had a reduction in model size which ranged from 20 to 150 times compared to the model obtained using only the deep reinforcement learning method. The smaller size, indicating a reduction in model complexity in terms of the number of trainable network parameters, was achieved without a loss in the prediction performance.

Background and motivation:

The application of artificial intelligence in medical research, particularly unsupervised learning techniques, has shown promising potential. Medical time series data poses a unique challenge for analysis due to its complexity. Existing unsupervised learning methods often fail to effectively classify these variations, highlighting a gap in current approaches. We introduce a methodological clustering classification framework designed to accurately handle such data, aiming for improved classification tasks in biomedical signals.

Methods:

To address these challenges, we introduce a novel approach for the analysis and classification of medical time series data. Our method integrates agglomerative hierarchical clustering with Hilbert vector space representations of medical signals and biological sequences. We rigorously define the mathematical principles and conduct evaluations using simulations of cardiac signals, real-world neural signal datasets, open-source protein sequences, and the MNIST dataset for illustrative purposes.

Results:

The proposed method exhibited a 96% success rate in classifying protein sequences by function and effectively identifying families within a large protein set. In cardiac signal analysis, it retained 0.996 variance in a condensed 6-dimensional space, accurately classifying 87.4% of simulated atrial flutter groups and 99.91% of main groups when excluding conduction direction. For neural signals, it demonstrated near-perfect tracking accuracy of neural activity in mouse brain recordings, as confirmed by expert evaluations.

Conclusion:

Our proposed method offers a novel, translational approach for the treatment and classification of medical and biological time series, addressing some of the prevalent challenges in the field and paving the way for more reliable and effective biomedical signal analysis.

The transition from fetal to newborn condition involves complex physiological adaptations for extrauterine life. A crucial event in this process is the clamping of the umbilical cord, which can be categorized as immediate or delayed. The type of clamping significantly influences the hemodynamics of the newborn. In this study, we developed a simulator based on existing cardiovascular models to better understand this practice. The simulator covers the period from late gestation to 24 h after birth and faithfully reproduces flow patterns observed in real-life situations (as evaluated by clinical specialists), considering factors such as the timing of cord clamping and the altitude of the birth location. It also reproduces blood pressure values reported in clinical data. Under similar conditions, the simulation results indicate that delayed cord clamping leads to increased oxygen concentration and improved blood volume compared to immediate cord clamping. Delayed cord clamping also had a positive impact on sustained placental respiration. Furthermore, this study provides further evidence that umbilical cord clamping should be based on physiological criteria rather than predefined time intervals.

Breast cancer is a prevalent malignant tumour with high global incidence. Its diagnosis relies primarily on the analysis of pathological breast images. Owing to the complex organisation of the tumour microenvironment, neural network models are essential as efficient classification tools in the field of pathological image analysis. This study introduced spatially-aware attention swift parallel convolution network (SPA-SPCNet), a lightweight and low-latency model for classifying breast pathologies. A novel module for multi-scale feature extraction was constructed using a depthwise separable convolution method. It focuses on the multi-scale features of pathological images to alleviate recognition problems caused by similar local features in breast cancer tissues. The module concatenates the convolutions of different kernels from three branches. Second, a lightweight dynamic spatially-aware attention module was introduced to integrate the visual graph convolutional architecture in a branch. This allowed the model to capture the spatial structure and relationships in image, enabling better handling of the unique spatial distribution relationship between breast cancer tissue structures. The other branch utilises a self-attention mechanism in the transformer. The module can dynamically adjust the attention of the model to different regions in the image, allowing it to focus on the key features of the complex spatial distribution of breast cancer tissue. This feature fusion method enabled the model to capture both global semantics and local details. Compared with existing lightweight models, the proposed model has advantages in terms of tissue structure classification accuracy, parameter quantity, floating-point operations, and real-time inference speed, providing a powerful tool for computer-aided breast pathological image classification.

Breast cancer is a prevalent global disease where early detection is crucial for effective treatment and reducing mortality rates. To address this challenge, a novel Computer-Aided Diagnosis (CAD) framework leveraging Artificial Intelligence (AI) techniques has been developed. This framework integrates capabilities for the simultaneous detection and classification of breast lesions. The AI-based CAD framework is meticulously structured into two pipelines (Stage 1 and Stage 2). The first pipeline (Stage 1) focuses on detectable cases where lesions are identified during the detection task. The second pipeline (Stage 2) is dedicated to cases where lesions are not initially detected. Various experimental scenarios, including binary (benign vs. malignant) and multi-class classifications based on BI-RADS scores, were conducted for training and evaluation. Additionally, a verification and validation (V&V) scenario was implemented to assess the reliability of the framework using unseen multimodal datasets for both binary and multi-class tasks. For the detection tasks, the recent AI detectors like YOLO (You Only Look Once) variants were fine-tuned and optimized to localize breast lesions. For classification tasks, hybrid AI models incorporating ensemble convolutional neural networks (CNNs) and the attention mechanism of Vision Transformers were proposed to enhance prediction performance. The proposed AI-based CAD framework was trained and evaluated using various multimodal ultrasound datasets (BUSI and US2) and mammogram datasets (MIAS, INbreast, real private mammograms, KAU-BCMD, and CBIS-DDSM), either individually or in merged forms. Visual t-SNE techniques were applied to visually harmonize data distributions across ultrasound and mammogram datasets for effective various datasets merging. To generate visually explainable heatmaps in both pipelines (stages 1 and 2), Grad-CAM was utilized. These heatmaps assisted in finalizing detected boxes, especially in stage 2 when the AI detector failed to automatically detect breast lesions. The highest evaluation metrics achieved for merged dataset (BUSI, INbreast, and MIAS) were 97.73% accuracy and 97.27% mAP50 in the first pipeline. In the second pipeline, the proposed CAD achieved 91.66% accuracy with 95.65% mAP50 on MIAS and 95.65% accuracy with 96.10% mAP50 on the merged dataset (INbreast and MIAS). Meanwhile, exceptional performance was demonstrated using BI-RADS scores, achieving 87.29% accuracy, 91.68% AUC, 86.72% mAP50, and 64.75% mAP50-95 on a combined dataset of INbreast and CBIS-DDSM. These results underscore the practical significance of the proposed CAD framework in automatically annotating suspected lesions for radiologists.

Radial flow packed-bed bioreactors (rPBBs) overcome the transport limitations of static and axial-flow perfusion bioreactors and enable development of clinical-scale bioengineered tissues. We developed criteria to design rPBBs with uniform medium radial flux distribution along bioreactor length ensuring uniform construct perfusion. We report a model-based analysis of the effect of non-uniform axial distribution of medium radial flux on pericellular concentration of oxygen and glucose. Albeit pseudo-homogeneous, the model predicts how medium flux, solutes transport and cellular consumption interact and determine the pericellular oxygen and glucose concentrations in the presence of pore transport resistance to design optimal axisymmetric rPBBs and enable control of pericellular environment. Thus, oxygen and glucose supply may match cell requirements as tissue matures. Flow and solute transport in bioreactor empty spaces and construct was described with Navier-Stokes and Darcy-Brinkman equations, and with convection–diffusion and convection–diffusion-reaction equations, respectively. Solute transport in construct accounted for Michaelian cellular consumption and bulk medium-to-cell surface oxygen transport resistance in terms of a transport-equivalent bed of Raschig rings. The effect of relevant dimensionless groups on pericellular and bulk solute concentrations was predicted under typical tissue engineering operation and evaluated against literature data for bone tissue engineering. Axial distribution of medium radial flux influenced the distribution of pericellular solutes concentration, more so at high cell metabolic activity. Increasing medium feed flow rates relieved non-uniform solute concentration distribution and decayed at cell surface for metabolic consumption, also starting from axially non-uniform radial flux distribution. Model predictions were obtained in runtimes compatible with on-line control strategies.

Sleep disorders have increasingly impacted healthy lifestyles. Accurate scoring of sleep stages is crucial for diagnosing patients with sleep disorders. The precision of sleep staging differs notably between healthy individuals and those with sleep apnea (SA). SA disrupts the regularity of sleep stages, affecting the performance of sleep stage detection and influencing the accuracy of sleep staging, thereby impacting sleep quality assessment results. The study compares the accuracy of sleep staging between healthy individuals and SA patients using the same algorithm, revealing variations in performance based on different severities of sleep apnea. This suggests limitations in the generalization ability of current sleep staging methods. Accordingly, researchers are working to develop sleep staging methods that can diminish the impact of sleep apnea and exhibit better generalization capabilities. Furthermore, the study emphasizes the advantages of automated methods over manual scoring due to being less subjective and resource-intensive, making them more suitable for practical applications. The emphasis is on recent research findings on automatic sleep stage classification based on electroencephalography (EEG). The study outlines potential applications and distinctions of various algorithm models rooted in machine learning and deep learning within the context of sleep staging. These methods are applied to the well-known public EEG dataset Sleep-EDF. The study applies four widely studied algorithms to the single-channel EEG of 20 subjects, comparing the results of the models’ automatic sleep staging with the manual sleep staging annotations by clinical experts.

Artificial pancreas systems should be designed considering different patient profiles, which is challenging from a control theory perspective. In this paper, a flexible-hybrid dual-hormone control algorithm for an artificial pancreas is proposed. The algorithm handles announced/unannounced meals by means of a non-interacting feedforward scheme that safely incorporates prandial boluses. Also, a coordination strategy is employed to distribute the counter-regulatory actions, which can be delivered as a continuous glucagon infusion via an automated pump, as an oral rescue carbohydrate recommendation, or as a rescue glucagon dose recommendation to be administrated through a glucagon pen. The different configurations of the proposed controller were evaluated in silico using a 14-day virtual scenario with random meal intakes and exercise sessions, achieving above 80% time-in-range and low time spent in hypoglycemia.