Home    About Journal    Editorial Board    Instruction    Subscribe    Download    Messages Board    Contact Us    中文     Quick Search

Office Online     Current Issue Accepted Current Issue Archive Adv Search Read Articles Download Articles Email Alert     Download More>>     Links More>>

2024 Vol. 43, No. 2 Published: 2024-04-20 Reviews Communications Regular Papers CONTENTS        CONTENTS 0 CONTENTS 2024 Vol. 43 (2): 0-0 [Abstract] ( 44 ) HTML (1 KB)  PDF (235 KB)  ( 5 )        Regular Papers 129 EEG Emotion Recognition Based on Source-Free Domain Adaptation Zhao Hongyu, Li Chang, Liu Yu, Cheng Juan, Song Rencheng, Chen Xun DOI: 10.3969/j.issn.0258-8021.2024.02.001 Existing domain adaptation methods in EEG emotion recognition utilize source domain data and feature distribution to train the model, which inevitably requires frequent access to the source domain and thus may lead to leakage of private information of the source domain subjects. To address this problem, this paper proposed a source-free domain adaptation EEG emotion recognition method based on the Gaussian mixture model, nuclear-norm maximization, and Tsallis entropy (GNTSFDA). First, based on the source domain data and the CNN andtransformer feature mixture (CTFM) network, the source domain model was trained to obtain the source domain model using the cross-entropy loss. Then, the pseudo-labels of the target domain data were generated by clustering with the Gaussian mixture model to construct the classification loss. Finally, based on the pseudo-labels and the classification loss, the source domain model was re-trained on the target domain data to update its parameters to obtain the target domain model, and the nuclear-norm maximization loss was also utilized during the training process to enhance the class discriminative property and the diversity of the model predictions, and Tsallis entropy loss was utilized to reduce the model predictions' uncertainty. The GNTSFDA method was experimented on the SEED (14 subjects in the source domain, 1 subject in the target domain), SEED-IV (14 subjects in the source domain, 1 subject in the target domain), and DEAP (31 subjects in the source domain, 1 subject in the target domain) public datasets, using a leave-one-subject cross-validation experimental paradigm. The results showed that on the three datasets, the accuracies of emotion recognition of the target domain model was 80.20%, 61.20%, and 58.89%, respectively, which was an improvement of 8.98%, 7.72%, and 6.54%, respectively, compared with thatobtained from the source domain model. The GNTSFDA method only needs to access the source domain model parameters, instead of the source domain, therefore, effectively protected the privacy information of source domain subjects and is of great significance in the practical application of EEG-based emotion recognition. 2024 Vol. 43 (2): 129-142 [Abstract] ( 127 ) HTML (1 KB)  PDF (5713 KB)  ( 269 ) 143 ERP-based Study on the Influence of Cognitive Load on Time-on-task Effect Ma Ke, Ke Yufeng, Wang Tao, Ming Dong DOI: 10.3969/j.issn.0258-8021.2024.02.002 The aim of this work is to study the influence of time-on-task effects on event related potential characteristics under different cognitive loads and explore the impact of mental workload on time-on-task effects. A cognitive function experimental paradigm based on the N-back task was designed.Twenty healthy subjects were recruited to participate in the experiment adopting this paradigm. Throughout the experiment, the subjective scale results and EEG signals of the subjects were collected for behavioral statistical analysis and event related potential analysis. With the enhancement of time on task, the reaction time (RT) of the subjects was significantly decreased when accomplishing the 3-back task, coefficient of variation of reaction time was increased both in 0-back task and 3-back task. During the N-back experiment, N1 amplitude increased significantly in occipital regions, while P300 amplitude in the prefrontal and central frontal regions decreased significantly. In both the 0-back and 3-back tasks, changes in P2 amplitude in subjects′ prefrontal regions were linearly correlated with reaction time (r=-0.44, P<0.05;r=-0.59, P<0.05), while P2 amplitude was negatively correlated with the coefficient of variation of reaction time (r=-0.39, P<0.05;r=-0.42, P<0.05). In the 3-back task, the mean P300 amplitude in frontal, central and parietal regions was significantly correlated with reaction time (r=-0.49, P<0.05), whereas the P300 amplitude in the 0-back task was significantly negatively correlated with reaction time and the coefficient of variation of reaction time (r=-0.69, P<0.05;r=-0.51, P<0.05). These results suggested that P2 and P300 in prefrontal and midfrontal regions were influenced by time-on-task effect, while the P300 of occipital region was more likely to be affected by cognitive load and time-on-task effect. 2024 Vol. 43 (2): 143-152 [Abstract] ( 123 ) HTML (1 KB)  PDF (8977 KB)  ( 153 ) 153 A Skin Cancer Detection Framework Based on Double-Branch Attention Neural Networks Wang Yufeng, Cheng Haoyuan, Wan Chengbei, Zhang Bo, Shi Aiju DOI: 10.3969/j.issn.0258-8021.2024.02.003 Skin cancer is a major cancerand has increased rapidly in the past decades. Early detection can significantly increase the cure rate. Recently, deep learning models, especially various convolutional Neural Networks using dermatoscope images (i.e., dermoscopy) have been widely adopted to classify skin lesions. Different from traditional image classification, several challenges in detecting and classifying skin cancers still exist, including imbalanced training data in each skin cancer category, small visual differences between categories, and small area of skin lesion. To solve these challenges, this paper proposed a skin cancer classification framework based on double-branch attention convolutional neural networks (DACNN). First, in data pre-processing, the whole dataset was divided into finer-grained categories according to the natural sub-classes in each category to alleviate the imbalanced data. Next, from the viewpoint of neural network structure, attention residual learning (ARL) modules were used as basic blocks in upper-branch, which was able to effectively extract the features of potential sick area, then thelesion location network (LLN) was designed to localize, cut out and zoom-in the sick sub-area, followed by being sent to down-branch with the same neural structure as the upper-branch, for extracting the locally detailed features. Then, the inferred features from both branches were integrated for effective detection and classification. Moreover, to further alleviate the impact of imbalanced categorical data, weighted loss function was utilized in the model training. The proposed DACNN model was implemented in the real dataset consisting of 10015 dermatoscope images and compared with several typical deep learning based skin lesion detection methods. Experimental results showed that the performance metrics of sensitivity, accuracy and F1_score reached 0.922, 0.942 and 0.933, respectively. Compared with recurrent attention convolutional neural network (RACNN) detection methods, these three metrics were improved by 3.48%, 2.95% and 3.44% respectively. In summary, our work significantly improved the accuracy of dermoscopy based skin cancer detection through appropriate division of dermatoscope image classes, used the double-branch attention neural networks to firstly localize and enlarge the features of potential sick area, and then further extracted the locally detailed features, which solved the intrinsic issues of dermatoscope images, including imbalanced samples in each skin cancer category, vague visual differences between categories, and small areaof skin lesion. 2024 Vol. 43 (2): 153-161 [Abstract] ( 126 ) HTML (1 KB)  PDF (5618 KB)  ( 184 ) 162 An Improved AdaBoost Cascade Classifier for Identifying Breath Signals of Liver Cancer Hao Lijun, Zhu Geng, Huang Gang, Yan Jiayong DOI: 10.3969/j.issn.0258-8021.2024.02.004 To reduce false negative rate of breath detection techniques in liver cancer screening, an improved AdaBoost cascade classifier was designed and applied to discriminate breath signals from healthy volunteers and liver cancer patients. First, a set of training subsets was obtained by self-help division of training samples. Based on the training subset, multiple sub-classifiers were successively obtained using different machine learning algorithms with K-fold cross-training and voting method. Next, multiple sub-classifiers were weighted and combined to obtain an improved AdaBoost classifier. Then, the training samples were self-subdivided and trained again with a new training subset to obtain another AdaBoost classifier. Finally, the two AdaBoost classifiers were concatenated in tandem to form a cascade classifier. After the test samples were fed into this cascade classifier, potentially anomalous samples were repeatedly screened according to the cascade rule. In this study, therelief-optimized feature set of the breath signals of 120 volunteers collected by the electronic nose (eNose) was used as the training sample to construct an improved AdaBoost cascade classifier and to discriminate the 40 test samples. The results showed that the classifier effectively distinguished the exhaled breath signals of liver cancer patients and healthy people in the test group, and the average sensitivity reached 93.42%, which was significantly better than the traditional AdaBoost cascade classifier, and the false negative rate was significantly reduced. In addition, the stability of this cascade classifier was good, and the coefficient of variation of the precision was only 3.95%. In conclusion, the improved AdaBoost cascade classifier effectively improved the classifier′s discrimination accuracy of liver cancer breath signals, which was important for the study ofbreath-based noninvasive universal screening for liver cancers. 2024 Vol. 43 (2): 162-172 [Abstract] ( 80 ) HTML (1 KB)  PDF (9200 KB)  ( 105 ) 173 Construction of a New Length Feature for Evaluating the Morphological Changes of BrainWhite Matter Fibers Dong Weihong, Qin Jiaolong, Ni Huangjing, Luo Dandan, Wu Ye, Yao Zhijian, Lu Qing DOI: 10.3969/j.issn.0258-8021.2024.02.005 Based on diffusion tensor imaging (DTI) data, whole brain white matter fibers can be presented in three dimensions. At present, morphological analysis of brain white matter fibers has been used to study the changes of fiber morphology in the development process or pathological conditions. In this study, we proposed a new feature to characterize the length of brain white matter fibers by calculating the euclidean space distance accumulated by points along the fiber, and further explored the stability of the feature from two aspects using 50 samples from the human connectome project (HCP) dataset. First, the influence of white matter fibers reconstructed by different fiber tracking algorithms on the feature. Second, under the same fiber tracking algorithm, the influence of different fiber numbers on the feature. Finally, 254 subjects from the HCP dataset were included, and we used the feature to preliminarily explore the influence of gender on the morphology of brain white matter fibers by voxel-based analysis (VBA). Through the calculation of intra-class correlation coefficients (ICC) model, it was found that when the overall lengths of white matter fibers reconstructed by the two fiber tracking algorithms were relatively close, the ICCs of the corresponding feature values in most intracranial voxels were above 0.4. In addition, under the same fiber tracking algorithm, different numbers of white matter fibers had little influence on the feature, and the ICCs corresponding to the feature values of intracranial voxels were concentrated above 0.8. The analysis of the influence of gender on the morphology of white matter fibers in the brain showed that compared with males, the length feature values of brain white matter fibers in females were significantly higher in the thalamus, fornix, middle cerebellar peduncle and pallidum. Males had significantly higher feature values in the right rectus and the right pallidum than females. The length feature of brain white matter fibers proposed in this paper is expected to enrich analysis methods of brain white matter fiber morphology for uses in the studies of brain development and brain related diseases. 2024 Vol. 43 (2): 173-182 [Abstract] ( 85 ) HTML (1 KB)  PDF (8759 KB)  ( 121 ) 183 An Optical Biosensor Based on Few-Layer MoS2 Nanosheets for Sensitive Detection of TumorMarker ctDNA Wei Huyue, Li Dujuan, Cui Zhilian, Fan Kai, Yang Weihuang, Liu Hongying, Li Lili, Wu Wei, Wang Gaofeng DOI: 10.3969/j.issn.0258-8021.2024.02.006 Cancer has become one of the major diseases threatening human health and life. Circulating tumor DNA (ctDNA) testing, as a practical liquid biopsy technique, is a promising method for cancer diagnosis, targeted therapy and prognosis. In this work, an optical biosensor based on few-layer MoS2 nanosheets was studied for the highly sensitive detection of tumor marker ctDNA. Firstly, The few-layer MoS2 nanosheets were prepared by shear exfoliation method, and the preparation conditions were optimized to prepare the few-layer MoS2 nanosheets with large transverse size and uniform thickness. Secondly, the accelerated aggregation and precipitation behavior of MoS2 nanosheets in high concentration salt solution was studied by detecting the absorbance change of dispersion. Then, based on the adsorption capacity of MoS2 nanosheets on single-stranded DNA, the inhibition effect of single-stranded DNA on salt-induced precipitation of few-layer MoS2 nanosheets was studied. Finally, the optical biosensor was constructed by using the change of absorbance of dispersion caused by the fall off of double-stranded DNA formed by hybridization from the few-layer MoS2 nanosheets, and the performance of the sensor was tested by hybridizing cpDNA with different concentrations of ctDNA. Results showed that the absorbance of ctDNA in the concentration range of 25 nM - 100 nM was inversely linear with the dispersion of MoS2 nanosheets when cpDNA concentration was 100 nM. The linear regression equations between the absorbance of the dispersion of few-layer MoS2 nanosheets and the ctDNA concentration at wavelengths of 401 nm and 448 nm were Y=0.235 57-0.000 70X, R2=0.942 22 and Y=0.212 53-0.000 50X, R2=0.951 41, respectively. This study demonstrated that the optical biosensor was able to detect tumor marker ctDNA and provided an effective sensing strategy for future in vitro cancer detection and diagnosis. 2024 Vol. 43 (2): 183-195 [Abstract] ( 82 ) HTML (1 KB)  PDF (13133 KB)  ( 143 ) 196 A Study of the Vibration Characteristics of Human Whole Lumbar Spine Based on 3D Finite Element Model Fan Wei, Zhang Chi, Zhang Dongxiang, Wang Qingdong, Guo Lixin DOI: 10.3969/j.issn.0258-8021.2024.02.007 Exposure to vibration has been considered a major cause of the lumbar degenerative disease and low back pain. The aim of this study was to explore the effects of vibration load on lumbar spine biomechanics. Based on CT scan data of human lumbar spine from L1 segment to pelvis (L1-pelvis), a 3-D geometric model of the L1-pelvis was reconstructed. Then, mesh generation and material properties assignment were performed to the geometric model to develop a 3-D finite element model of the L1-pelvis, which was validated according to several available experimental data. Based on this finite-element model, the biomechanical responses of each lumbar spinal segment under an axial sinusoidal vibration load of 40 N with a frequency of 5 Hz were computed through transient dynamic analysis and compared with the corresponding results under -40 N and +40 N axial static loads. The response parameters included axial displacement of vertebral center, disc bulge, and von-Mises stress in annulus ground substance. The results showed that compared with the static loads, the amplitudes of vertebral axial displacement, disc bulge, and annulus stress for each lumbar spinal segment under the vibration load increased by 0.550~1.020 mm, 0.124~0.251 mm, and 0.043~0.099 MPa, respectively, and the maximum increasing effect reached to 195.0%, 175.7%, and 151.4%, respectively. It implies that the lumbar spine might face a higher likelihood of injury under the vibration load. 2024 Vol. 43 (2): 196-203 [Abstract] ( 86 ) HTML (1 KB)  PDF (6850 KB)  ( 124 )        Reviews 204 Research Progress of Optogenetic Brain-Computer Interface Meng Zhaoyang, Pu Jiangbo, Li Xiangning DOI: 10.3969/j.issn.0258-8021.2024.02.008 As bidirectional communication systems between human brains and external devices, brain-computer interfaces (BCIs) have been widely employed in brain function enhancement, human-computer interaction and nerve rehabilitation. BCIs based on optogenetics remedy for the deficits of electrode stimulation in biological compatibility, stimulation accuracy and cell type specificity, thus becoming a hotspot in the research of neural engineering. In this review, we first described the application of optogenetic interfaces in animal experiments such as closed-loop control of brain activity, virtual sensation feedback, or brain-brain information channel. Subsequently, frontiers of novel integrated and miniaturized optogenetic interfaces were summarized. At last, we recapitulated the current challenges of optogenetic brain-computer interfaces, along with the prospects of optogenetic interface in multimodal brain activity monitoring and brain-computer intelligence. 2024 Vol. 43 (2): 204-213 [Abstract] ( 134 ) HTML (1 KB)  PDF (2017 KB)  ( 515 ) 214 Research Progress of Flexible Wearable Sensors for Cardiovascular Disease Monitoring Huang Qiufan, Ma Ye, Xu Zhiyun DOI: 10.3969/j.issn.0258-8021.2024.02.009 As cardiovascular diseases ranking first in terms of morbidity and mortality in China and showing a tendency of continuous increase and rejuvenation, they have become a major public health concern. Recently, flexible electronic technology has been widely used in the healthcare industry, offering new approaches for timely detection and precise management of various diseases. Notably, the flexible electronic technology can greatly improve the wearing comfort and monitoring accuracy of wearable sensors for cardiovascular diseases, which are characterized by acute onset, frequent recurrence and insidiousness. In this review, classifications and principles of wearable sensor technologies are briefly introduced in combination with the development trend of flexible electronics and wearable sensors. Recent advances in flexible wearable sensors are discussed, with a primary focus on monitoring important parameters related to cardiovascular disease, including heart rate, blood pressure, blood flow and oxygen saturation. Finally, this article discusses the challenges and issues that need to be addressed for flexible wearable sensors in the field of cardiovascular diagnosis and treatment, and suggests the corresponding development directions. 2024 Vol. 43 (2): 214-226 [Abstract] ( 304 ) HTML (1 KB)  PDF (9879 KB)  ( 324 ) 227 Research Progress of Transcranial Random Noise Stimulation Zou Huiru, Zhang Zhiguo, Huang Gan, Li Linling, Liang Zhen, Zhang Li, Wei Jinwen DOI: 10.3969/j.issn.0258-8021.2024.02.010 Transcranial random noise stimulation (tRNS) is a specific type of transcranial electrical stimulation technique that uses current of random frequency and amplitude to modulate neural activity. It affects brain activity and cognitive behavior through mechanisms such as stochastic resonance, and has demonstrated remarkable modulation effects in the fields of neuroscience and neuropathology, thus gaining increasing attention and application. This reviewmainly introduced tRNS’s physiological effects, implementation, and the status and development trend in modulating perception, motion, learning and memory, and psychiatric symptoms. 2024 Vol. 43 (2): 227-239 [Abstract] ( 148 ) HTML (1 KB)  PDF (1212 KB)  ( 206 ) 240 A Review of Brain Network Research Methods Based on Canonical Correlation Analysis Yin Shunjie, Chen Kai, Xue Kaiqing, Yao Dezhong, Xu Peng, Zhang Tao DOI: 10.3969/j.issn.0258-8021.2024.02.011 Brain network analysis plays important roles in studying the cognitive activity of brain, including exploring the information processing mode of the brain and assisting the diagnosis of mental diseases. In recent years, brain network research methods based on multivariate datasets have attracted great attention. Canonical correlation analysis (CCA), as a data-driven multivariate statistical method, can effectively capture the implicit relationship between multivariate data and is widely used in brain network research. This article reviewed the roles of CCA in the brain network research, specific application modes, and advantages and limitations. Firstly, the algorithm principles of traditional CCA and its common variants were summarized. Next, the research status of CCA-based analysis methods in the brain network construction, brain network analysis, and brain network marker identification were described. At last, the methods of brain network research based on CCA were summarized and the future research directions were discussed. 2024 Vol. 43 (2): 240-251 [Abstract] ( 152 ) HTML (1 KB)  PDF (4581 KB)  ( 479 )        Communications 252 Study and Analysis of Thermal Damage Area of Radio Frequency Ablation Needle Yu Liyang, Chen Yaoying, Tao Jiawei, Rao Xin, Yang Yong DOI: 10.3969/j.issn.0258-8021.2024.02.012 2024 Vol. 43 (2): 252-256 [Abstract] ( 96 ) HTML (1 KB)  PDF (7606 KB)  ( 170 )

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