Chinese Journal of Biomedical Engineering

Home

Editorial Board

Download

Quick Search

Office Online

Current Issue

Download Articles

Download

Links

2021 Vol. 40, No. 1 Published: 2021-02-20

Reviews Communications Regular Papers

Regular Papers

	1	T-Wave Morphological Classification Based on CNN and Modified Frequency Slice Wavelet Transform
		Xie Jiajing, Wei Shoushui, Jiang Xinge, Wang Chunyuan, Cui Huaijie, Liu Chengyu
		DOI: 10.3969/j.issn.0258-8021.2021.01.01
		Real-time monitoring of ECG is one important means of cardiovascular disease prevention. T-wave is an important characteristic of diseases such as myocardial ischemia and sudden cardiac death. The automatic identification of T-wave is a challenging taskin ECG remote monitoring. Due to the influence of high noise background of real time monitoring ECG, the conventional T-wave feature extraction and classification algorithm encounters a bottleneck. In this paper, a T-wave morphological recognition algorithm combining slice frequency wavelet transform and convolutional neural network was proposed. The algorithm included: locating automatically the R-wave peak's position and the T-wave ends' position to identify a segment containing the T-wave; the frequency slice wavelet transform was performed, and the generated time-frequency image was input into the convolutional neural network to complete the classification of the T-wave. The frequency slice wavelet transform transformed the signal to the time-frequency domain, which accurately presented the time-frequency energy distribution characteristics of the ECG signal. The hidden layers of the convolutional neural network completed the three features' extraction of the time-frequency image by convolving, activating and pooling the time-frequency image three times. These features have translation and scaling invariance. In this paper, 12 830 fragments in European ST-T database was used. The convolutional neural network model was trained and tested by the 3-fold cross validation method. The classification accuracy of experiment based on heart beats reached 97.34%, and the F1 measure reached 96.97%. The classification accuracy of experiment based on samples was 84.80%, and the F1 measure was 83.29%. The classification accuracy of the model tested in QT database was 87.83%, F1 measure was 85.38%, and the generalization performance was good. Compared with other T-wave classification algorithms (such as decision tree, support vector machine, etc.), the classification accuracy based on heart beat experiments was improved by 1～5%. The results demonstrated that the algorithm designed for the classification of six types of T-wave improved the accuracy and performed well in terms of robustness and generalization performance. In addition, the algorithm model was also applicable to the analysis of other physiological signals and has certain guiding significance in the field of medical image analysis.
		2021 Vol. 40 (1): 1-11 [Abstract] ( 615 ) HTML (1 KB) PDF (6098 KB) ( 669 )

	12	Method for Measuring Fetal Head Circumference in Ultrasound Images Based on Mask R-CNN
		Li Zonggui, Zhang Junhua, Mei Liye
		DOI: 10.3969/j.issn.0258-8021.2021.01.02
		Fetal head circumference is one of the most important biological characteristics in prenatal ultrasound evaluation of fetal growth and development. However, manual measurement is time-consuming and labor-consuming and may have errors by the operator. According to the feature of fetal head close to ellipse shape in ultrasound image, the head circumference measurement loss function was proposed in this paper. After the segmentation branch of Mask R-CNN, ElliFit algorithm was used to fit the ellipse of the segmentation mask. Ramanujan formula was used to calculate the fitting ellipse circumference as the measurement value of the head circumference. The mean square error of the real value of the head circumference and the measurement value was added into the original loss function as head circumference measurement loss function to allow the training process of the model to be closely related to the measurement task. By this way the measurement accuracy and speed was improved. One hundred and ninety ultrasound images of fetal head were tested. Dice’s coefficient was 96.89%±1.01%, and the measurement error was (0.33±1.54) mm. The average processing time of one ultrasound image was 0.33 s. Compared with the traditional manual measurement method or the current machine learning methods, the proposed method improved the speed between 1.13 seconds and 16.87 seconds, and improved the accuracy between 0.21 mm and 1.68 mm. The results showed that the improved Mask R-CNN increased the efficiency of doctors in measuring fetal head circumference, which met the clinical needs.
		2021 Vol. 40 (1): 12-18 [Abstract] ( 411 ) HTML (1 KB) PDF (7081 KB) ( 199 )

	19	Study on fMRI Data Analysis Based on Multi-Objective Optimization CICA
		Shi Yuhu, Zeng Weiming, Deng Jin, Wang Nizhuan
		DOI: 10.3969/j.issn.0258-8021.2021.01.03
		Constrained independent component analysis (CICA) greatly improves the performance of blind source signal analysis of independent component analysis (ICA) by incorporating priori information, nevertheless, the current CICA method has some problems, such as the difficulty in obtaining prior information, selecting threshold parameters of prior information constraints, and using priori information effectively, which need to be improved. Targeting to these problems, this paper established a CICA model that simultaneously integrated temporal and spatial priori information on the basis of multi-objective optimization framework, and solved the problem of selecting threshold parameters in CICA through multi-objective optimization strategy. Furthermore, an adaptive mining algorithm was proposed to extract intrinsic a priori information from the fMRI data of multiple subjects to guide the analysis of fMRI group data, thus providing a new way for CICA to obtain priori information. Finally, 10 simulated data, 5 task-state and 23 resting-state fMRI data were used to verify the effectiveness of the proposed method. The results showed that the spatio-temporal source signals obtained by multi-objective optimization based CICA (MOPCICA) were generally superior to those obtained by ICA, CICA with temporal reference (CICA-tR) and CICA with spatial reference (CICA-sR) (P<0.05) (in the simulation data, the corresponding spatial AUC and temporal correlation coefficients were 0.75±0.05, 0.62±0.02, 0.72±0.03, 0.71±0.06 and 0.81±0.13, 0.67±0.04, 0.74±0.09, 0.77±0.13, respectively); while the spatial independence was superior to CICA-tR and CICA-sR (P<0.05) (in the task-related data, the corresponding kurtosis and negentropy were 69.20±23.36, 17.60±13.22, 36.71±13.43 and 0.031 2±0.007 7,0.003 7±0.002 1,0.018 4±0.004 5, respectively), which indicated that it had a better performance for the blind source signal recovery. Meanwhile, the correlation coefficient between the group component obtained by MOPCICA through using the fMRI intrinsic priori information in the resting state data and the corresponding component of each subject in the group was on average higher than that of ICA, CICA-nR and CICA-fR (P<0.05), which were 0.46±0.08, 0.44±0.08, 0.45±0.08 and 0.44±0.08 separately, thus can better represented the commonality of the subjects in the group. Therefore, it has a great significance for the fMRI brain functional connectivity detection.
		2021 Vol. 40 (1): 19-32 [Abstract] ( 376 ) HTML (1 KB) PDF (20062 KB) ( 99 )

	33	Research on Using Neurofeedback to Improve Attention and Skills of SSVEP Brain-Computer Interface
		Sun Jinnan, Zhang Shangen, Gao Xiaorong
		DOI: 10.3969/j.issn.0258-8021.2021.01.04
		In the past 20 years, neural engineering has made unprecedented progress in the interpretation of brain information (e.g., brain-computer interfaces) and neuromodulation (e.g., electromagnetic stimulation and neurofeedback). However, the relationship between the internal mechanism of attention and behavioral performance is still worth exploring. In this paper, the data of different degrees of attention were obtained by controlling the difficulty of the rapid serial visual presentation task (30 subjects). The directed information transfer function was used to analyze the network information characteristics of the attention state. Subsequently, the neurofeedback training method was designed with reference to these features to improve attention and the using skill of SSVEP-based brain-computer interface. The results of the connection strength between the leads showed that the prefrontal lobe, parietal lobe and occipital lobe constituted a joint processing system in the visual task. In this system, the parietal lobe played the role of central control, and alpha oscillation participates in the modulation process of attention. Thus, up-regulating alpha-band power of the parietal lobe by neurofeedback training was present as a new neural modulating method to improve SSVEP-based BCI in this study. After this NFT, the signal-to-noise ratio (SNR), accuracy, and information transfer rate (ITR) of SSVEP-based BCI were (-15.23±5.91) dB and (-14.72±4.83) dB; 78.93%±0.14% and 83.65%±0.14%; 103.21±28.49 and 119.35±25.14 respectively (24 subjects). However, no improvement has been observed in the control group (22 subjects) in which the subjects do not participate in NFT. What's more, evidence from attention test further indicate that NFT improves attention via developing the control ability of the parietal lobe and then enhances the above SSVEP indicators. Up-regulating parietal alpha-amplitude using neurofeedback training significantly improves the SSVEP-based BCI performance through modulating the control network. In conclusion, this study validated an effective neuromodulation method, and possibly also contributed to explaining the function of the parietal lobe in the control network.
		2021 Vol. 40 (1): 33-43 [Abstract] ( 505 ) HTML (1 KB) PDF (4600 KB) ( 750 )

	44	The Different EEG Activity During Cognitive Reappraisal Task for Fearful and Sad Stimuli
		Wei Ling, Li Yingjie, Yao Xufeng
		DOI: 10.3969/j.issn.0258-8021.2021.01.05
		The processing of fear and sadness have different neural bases. The purpose of this study is to explore the electrophysiological mechanism by analyzing the features of event-related potentials in the process of emotion regulation for fearful and sad stimuli. Electroencephalography (EEG) signals were recorded in 21 healthy subjects during the watch and cognitive reappraisal tasks for fear and sadness stimuli. Late positive potential (LPP) of four brain regions (left fronto-central, right fronto-central, left centro-parietal, right centro-parietal) within the time range of 500~5 000 ms (9 windowsof 500 ms in length) after stimulation were selected to analyze the difference of fear and sadness emotion regulation. Repeated measures analysis of variance was used for statistical analysis. The within-group factors were time (9 levels), area (4 levels), task condition (2 levels, watch and regulate) and emotion (2 levels, fear and sadness). The experimental results showed that fearful stimuli evoked significantly larger LPP in right centro-parietal region than sad stimuli in the window 500~1 000 ms (P<0.05, (8.815±1.153) μV vs (5.834±1.317) μV for watch, (7.094±1.036) μV vs (6.643±1.158) μV for regulate. And the regulation of sadness significantly increased the LPP in left and right frontal-central regions in 1 500~2 500 ms after the stimulation (P<0.05, among them (10.100±2.205) μV vs (2.515±1.676) μV in the window1 500~2 000 ms). In 2 500~4 000 ms, the regulation of sadness and fear both significantly enhanced the LPP in left frontal-central (P<0.05, among them in the window 2 500~3 000 ms, (2.957±1.767) μV vs (1.899±2.969) μV for fear, (7.957±2.305) μV vs (-0.051±2.018)μV for sadness). These findings pointed out the different time course of fear and sadness emotion regulation and an important role of the left fronto-central region in down-regulation of negative emotion.
		2021 Vol. 40 (1): 44-52 [Abstract] ( 388 ) HTML (1 KB) PDF (4184 KB) ( 509 )

	53	Research on EEG Feature Extraction Method in Boundary Avoidance Task Based on Non-Negative CP Decomposition Model
		Fu Rongrong, Yu Bao, Sun Jiedi
		DOI: 10.3969/j.issn.0258-8021.2021.01.06
		This study aimed to improve the subjects' EEG arousal by establishing a “bowl-ball” model and performing visually guided boundary avoidance tasks. In the process of interacting with the “bowl-ball” model, the EEG data of 10 healthy subjects about the left and right-hand motor tasks were collected, and the optimized features of the EEG were classified to realize the decoding of the exercise intention. The EEG signal induced by the boundary avoidance task is subjected to 8-13Hz band-pass filtering to obtain data of a specific frequency band, and the frequency components of EEG were obtained through continuous wavelet transform to generate EEG tensor. We used the non-negative CP decomposition model to extract the time component features of the EEG tensor, then used the two-dimensional principal component analysis to optimize the features, used the support vector machine (SVM) to classify the features, and compared with the method of feature extraction and feature classification using common spatial pattern (CSP) and SVM. The results of all subjects showed that the general best component number of CP decomposition was 16, based on the feature extraction method of non-negative CP decomposition model, the accuracy of SVM classification was 95.5%±3.0%, and the AUC value was 0.978 2±0.012 1. The classification accuracy was better than CSP+SVM (93.7%±3.1%). The discriminant score of the classification result was tested by t-test, and the classification result had 95% confidence (P<0.05). In conclusion, the motor intention classification based on features extracted from the non-negative CP decomposition model reflected the differences in different states in boundary avoidance tasks and improve classification performance.
		2021 Vol. 40 (1): 53-59 [Abstract] ( 338 ) HTML (1 KB) PDF (1827 KB) ( 295 )

	60	EEG Study of Patients with Disorder of Consciousness Based on Nonlinear Spatiotemporal Dynamics
		Lei Ling, Yang Yong, Hou Na, Liu Kehong, Wu Li, Cheng Qiqi, Dong Tengfei, Hu Xiaohua
		DOI: 10.3969/j.issn.0258-8021.2021.01.07
		The recovery of consciousness in patients with disorder of consciousness has not been well understood. Most previous studies usedcross-sectional design, as consciousness recovery is not only dynamic but also involves interactions between various brain regions. Elucidating the mechanism of recovery requires tracking brain activity both in temporal and in spatial dimensions. In this study we took advantage of the high temporal resolution and good spatial resolution of EEG to examine 41 patients with disorder of consciousness, analyzing a total of 161 resting-state EEG measurements. We compared the changes in EEG nonlinear dynamic features of brain activity among the patients in different stages of consciousness recovery, including approximate entropy, sample entropy, and Lempel-Ziv complexity. In the temporal dimension, EEG nonlinear dynamic features for the whole brain showed a non-monotonic trend during recovery (LZC: 0.299±0.053, 0.295±0.060, 0.279±0.049, 0.302±0.053, 0.307±0.069, 0.326±0.049, 0.334±0.046; P<0.05). When patients progressed from vegetative state to minimally conscious state, there was an inflection point in the EEG features. In the spatial dimension, changes in EEG features in injured and uninjured areas were also non-monotonic during consciousness recovery, and the non-monotonic changes in the two areas were non-synchronized. In emergence from minimally conscious state, the difference between the two regions was extremely significant (injured vs uninjured: ApEn: 0.608±0.042 vs 0.63±0.030; LZC: 0.317±0.054 vs 0.351±0.039; SampEn: 0.581±0.058 vs 0.615±0.043; P<0.01). The consciousness recovery pattern was non-monotonic in the temporal and asynchronous in the spatial dimension. These findings provided insights into the mechanisms of consciousness recovery following brain injury and could serve as a basis for the treatment and rehabilitation of patients with disorder of consciousness.
		2021 Vol. 40 (1): 60-70 [Abstract] ( 427 ) HTML (1 KB) PDF (4245 KB) ( 316 )

	71	Benign and Malignant Diagnosis of Pulmonary Nodules Based on SE-CapsNet
		Ye Feng, Wang Luyao, Hong Wei, Ding Guojun, Che Jiarong
		DOI: 10.3969/j.issn.0258-8021.2021.01.08
		Over the past few years, lung cancer has been the leading cause of cancer-related deaths. This paper proposed a SE-CapsNet classification method for the low-dose computed tomography (CT) image refinement preprocessing conditions. Our work solved the problems of low classification accuracy and high false positives in traditional lung nodule diagnosis methods, which improved the capsule neural network classification algorithm, including improving the latest Hinton's capsule neural network, introducing new non-linear activation vectors, avoiding global vector compression, and optimizing the model at the feature channel level by feature reweight. We used the automatic threshold method to process the CT images by calibrating the region of interest, and took the samples at the central nodule to obtain data samples of the pre-processing results. The public data set LIDC-IDRI containing 1010 cases and 30 cases of desensitized tumor patients eliminated sensitive information from hospital were used to evaluate the improved SE-CapsNet algorithm. The evaluation criteria mainly included accuracy, sensitivity and specificity. In the LIDC-IDRI dataset and the hospital dataset, the average accuracy of the SE-CapsNet algorithm reached 95.83% and 94.67%, respectively, which was superior to that by CapsNet classification algorithm. In addition, the classification algorithm also had obvious advantages in terms of time consumption, and the improved capsule network converged faster to obtain stable results.
		2021 Vol. 40 (1): 71-80 [Abstract] ( 361 ) HTML (1 KB) PDF (5813 KB) ( 297 )

	81	The Study of the Blooming Effect in High-Field MRI at 7 T
		Zhu Yurong, Gao Yunyu, Han Jijun, Wang Jiajia, XinXuegang
		DOI: 10.3969/j.issn.0258-8021.2021.01.09
		Magnetic susceptibility can be regarded as an intrinsic property of matter. Different tissues tend to have different magnetic susceptibility due to the differences in composition and structure. Effectively uses of the magnetic susceptibility may provide additional information of the structure and function of the organization. Based on this new imaging contrast mechanism, Susceptibility-weighted imaging emerged. There were some studies that have found that specific tissues could be extended under certain conditions due to the difference in magnetic susceptibility, which was called the blooming effect. As the basis, a comprehensive and accurate assessment of the blooming effect was important for the further application of SWI in clinical. In this paper, the related research of blooming effect was carried out systematically. Combined the silico and the ex-vivo experiment, we built the relationship between the susceptibility, echo times and size with the blooming effect using the number of the pixels and the blooming factor as the evaluation indexes. The silico result showed that, the susceptibility and echo times were positively correlated with the blooming effect. In this experiment, the magnitude and phase blooming factor could be up to 37. But in the model with 0.5 voxel radius, the blooming factor obtained by SWI could be up to 51. In addition, the results of ex-vitro experiments showed that the blooming factor in the diameter of 0.3 mm and 0.46 mm models could be up to 13.25 and 10.75 respectively. So that the smaller the radius of the model, the more obvious the blooming effect could be. The results of this research has important reference value to promote the development of SWI and early detection of diseases.
		2021 Vol. 40 (1): 81-90 [Abstract] ( 368 ) HTML (1 KB) PDF (4802 KB) ( 279 )

Reviews

	91	Research Hotspots and Trends of Brain-Inspired Intelligence
		Liu Jie, Wu Hui
		DOI: 10.3969/j.issn.0258-8021.2021.01.10
		Inspired by the brain's neural operation and cognitive behavior mechanisms, the brain-inspired intelligence uses computational modeling as a means to achieve machine intelligence through hardware and software collaboration with characteristics of brain-like information processing mechanism, human-like cognitive behavior, and human intelligence or more, attracting more and more attention. From the point of co-citation and co-words, articles about the research of brain-inspired intelligence between 2010 and 2019 were retrieved in Web of Science and CiteSpace software was used to evaluate the global scientific output. Research hotspots and trends, that is, the construction of brain-like neural network computing models and learning methods with the help of brain science and the in-depth and cross-study of brain-computer interface and deep learning, were discussed in detail.
		2021 Vol. 40 (1): 91-98 [Abstract] ( 720 ) HTML (1 KB) PDF (3639 KB) ( 1030 )

	99	Progress of Sublingual Microcirculation Microimage Monitoring and Application
		Jiang Sheng, Li Peilun, Ning Gangmin
		DOI: 10.3969/j.issn.0258-8021.2021.01.11
		Microcirculatory lesions are the key link leading to tissue hypoperfusion. Monitoring of microcirculatory lesions is very important in severe diseases. The tongue body is rich in microvessels, in which the sublingual microcirculation presents a reticular structure, reflecting the state of microcirculation in living tissue, and is an ideal and important site for clinical microcirculation monitoring and microcirculation detection in living animals. The equipment, index system, application, clinical significance and future prospect of sublingual microcirculation microimage monitoring were reviewed in this article. Firstly, we introduced the monitoring equipment, including the composition of the equipment, the types of optical technology adopted by the probe, the image processing algorithm adopted by the host, and the fixed form of the probe. Secondly, the index system of sublingual microcirculation microimage monitoring was summarized, including the perfusion quality index, vascular density index and perfusion heterogeneity index. Next, an example was given to illustrate the clinical and experimental applications, including the study on the relationship between diseases and microcirculation in clinical practice, the study on the relationship between drugs and microcirculation, and the study on the relationship between visceras' microcirculation using sublingual microcirculation microimage monitoring technology. Finally, the significance of clinical diagnosis, treatment and research were discussed, and the technical improvement, development and application direction are prospected.
		2021 Vol. 40 (1): 99-106 [Abstract] ( 622 ) HTML (1 KB) PDF (4503 KB) ( 741 )

	107	Advances in Exhaled Breath Analysis: A Potential Screening Tool and its Application in Clinical Diagnosis
		Zhu Xingzhuo, Zhao Dongbujia, Zheng Yixuan, Ning Xuan, Yuan Hao, Wu Chunsheng
		DOI: 10.3969/j.issn.0258-8021.2021.01.12
		Exhaled breath analysis is a diagnostic tool that provides useful information for early clinical diagnosis by detecting the changes of characteristic components in human breath. Compared with traditional methods, exhaled breath analysis provides a non-invasive, rapid and easy screening tool that has promising potential applications in early diagnosis and screening of diseases. However, there are still some problems in the clinical applications of this technique. This review summarized three aspects, including exhaled breath collection, multiple detection methods and the application of this technology in the diagnosis of various diseases. At last, we discussed the limitations and prospects of exhaled breath analysis.
		2021 Vol. 40 (1): 107-117 [Abstract] ( 576 ) HTML (1 KB) PDF (834 KB) ( 1106 )

	118	Progress on the Application of Decellularization Techniques in Tissue Engineering Vascular Scaffolds
		Cheng Jin, Wang Cong, Gu Yongquan
		DOI: 10.3969/j.issn.0258-8021.2021.01.13
		The decellularized blood vessel prepared by decellularizing the natural blood vessel is considered as one important kind of tissue engineering vascular scaffolds and has broad application prospects. However, the preparation of extracellular matrix (ECM) still lacks uniform standards. The choice of decellularization methods depends on the tissue source and the application of the ECM. The decellularization protocols used for preparing ECM are very important, especially for the ECM scaffolds such as decellularized blood vessels that need to withstand the blood press for a long time. The maintenance of ECM integrity and cell removal efficiency are highly dependent on the specific approaches used for decellularization. This article reviewed decellularization methods applied to blood vessels, including physical, chemical, enzymatic and serum applications in the preparation of decellularized blood vessels. The decellularization efficiency of different decellularization methods and their effects on the tissue structure, mechanical properties and biological activity of the vascular matrix were compared, and the importance of developing reasonable decellularization protocols was emphasized.
		2021 Vol. 40 (1): 118-123 [Abstract] ( 805 ) HTML (1 KB) PDF (755 KB) ( 791 )

Communications

	124	Simulation and Experiment Research on Thermal Effects of Laser Ablation for Biological Soft Tissue
		Liu Kezhou, Ou Zeshi, Xu Kedi, Bai Ruiliang
		DOI: 10.3969/j.issn.0258-8021.2021.01.14

		2021 Vol. 40 (1): 124-128 [Abstract] ( 457 ) HTML (1 KB) PDF (2586 KB) ( 329 )

Copyright © Editorial Board of Chinese Journal of Biomedical Engineering
Supported by:Beijing Magtech