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>>
2019 Vol. 38, No. 3
Published: 2019-06-20
Reviews
Communications
Regular Paper
256
catalogue
2019 Vol. 38 (3): 256-256 [
Abstract
] (
163
)
HTML
(1 KB)
PDF
(267 KB) (
245
)
Regular Paper
257
Segmentation of Organs at Risk on Head and Neck CT for Radiotherapy Based on 3D Deep Residual Fully Convolutional Neural Network
Tian Juanxiu, Liu Guocai, Gu Shanshan, Gu Dongdong, Gong Junhui
DOI: 10.3969/j.issn.0258-8021.2019.03.001
Segmentation of organs at risk (OARs) is a crucial process during the planning of radiation therapy for head and neck cancer treatment. However, accurate OAR segmentation in CT images is a challenging task. Manual delineation of OARs is tedious, time-consuming and inconsistent. To tackle these challenges, we proposed an automatic deep-learning-based method for head and neck OARs segmentation. A modified V-Net structure was constructed to extract deep and shallow features of OARs by specialized end-to-end supervised learning. To address the extremely class imbalances of small organs, a positional prior knowledge restricted sampling strategy was proposed, and Dice loss function was used to train the network. The strategy could not only accelerate the training process and improve the segmentation performance, but also ensure the accuracy of small organ segmentation. The performance of the proposed method was validated on PDDCA dataset, which was used in Head and Neck Auto-Segmentation Challenge of MICCAI 2015. The mean Dice coefficient of each organ was 0.945 of mandible, 0.884 of left parotid gland, 0.882 of right parotid gland, 0.863 of brainstem, 0.825 of leftsubmandibular gland, 0.842 of rightsubmandibular gland, 0.807 of left optic nerve, 0.847 of right optic nerve and 0.583 of optic chiasm. The 95% of Hausdorff distances of mandible, parotid glands, brainstem and submandibular glands was all within 3 mm. The contour mean distance of all organs was less than 1.2 mm. The experimental results demonstrated that the performance of the proposed method was superior to the compared state-of-the-art algorithms on segmentation of most of OARs.
2019 Vol. 38 (3): 257-265 [
Abstract
] (
465
)
HTML
(1 KB)
PDF
(9582 KB) (
475
)
266
Automated Detection and Quantification of Coronary Artery Stenoses Based on Vessel Tree Segmentation in X-Ray Angiography
Chen Jianhui, Zhao Lei, Li Deyu, Wan Tao
DOI: 10.3969/j.issn.0258-8021.2019.03.002
Automated identification and quantification of the vascular stenoses in coronary angiographical images are essential in a computer-aided diagnosis system, which can improve the diagnosis accuracy, while reducing the labor intensity of doctors in daily clinical practice. We presented a computerized method for automatic detection and grading of vascular stenoses on X-ray angiography in this work, which included two main parts. In the vessel segmentation part, image enhancement was first performed by an improved Frangi Hessian based method, and then the blood vessel regions were segmented using a statistical region merging approach, which could provide good partition of the vascular tree with a complex structure. In the stenosis assessment part, a vessel skeleton was first obtained from a skeletonization method, then vessel diameters were computed based on the boundary points from the vessel tree segmentation, and finally stenosis degree were calculated using ratio between minimum and maximum of the vessel diameters in the vessel stenosis segment. The method was tested on 153 patient studies, and a total of 208 vessel segments were identified, including 84 mild, 42 moderate, and 82 severe stenoses. The method achieved detection accuracy 93.59%, sensitivity 88.76%, specificity 95.58%, and a positive predictive value 90.51%, which suggested the method was able to effectively detect and quantify the artery vessel stenoses, and provided supplement opinion for clinical diagnosis of cardiovascular disease.
2019 Vol. 38 (3): 266-272 [
Abstract
] (
524
)
HTML
(1 KB)
PDF
(4637 KB) (
516
)
273
Epileptic Seizure Detection Based on theSum of Degree and Entropy of Weighted Complex Network
Zhang Hanyong, Meng Qingfang, Du Lei, Liu Mingmin
DOI: 10.3969/j.issn.0258-8021.2019.03.003
Epileptic seizure detection has always been a challenging task. With the increasing of epilepsy, high-performance epileptic automatic detection algorithm can reduce the workload of medical workers and has important clinical research significance. In this paper we proposed a new seizure detection method based on weighted horizontal visibility graph (WHVG). Firstly, the single channel electroencephalogram (EEG) signal was transformed into complex network by using WHVG. Then, the square of degree and weighted degree entropy of the complex network was extracted. Finally, the sum of this two extracted features was used as a single feature. The single feature was inputted into a linear classifier to identify interictal and ictal signals. The experiment evaluating the performance of proposed method was conducted on the epileptic EEG dataset of the University of Bonn. This experiment used 100 samples in interictal and 100 samples in ictal and each sample contains 1024 points. Experimental results showed that the proposed method had high classification accuracy, which was up to 98.5%. In addition, the feature used in the method was a single feature that was more simple and efficient. In conclusion, the proposed method was promising for uses in online automatic epileptic seizure detection.
2019 Vol. 38 (3): 273-280 [
Abstract
] (
411
)
HTML
(1 KB)
PDF
(3345 KB) (
287
)
281
Research on ERP for the Ways of Payment on Consumer Product Preferences
Song Zhijie, Zhang Liping, Wang Dandan, Li Tianjiao, Shi Rui
DOI: 10.3969/j.issn.0258-8021.2019.03.004
In order to explore the influence of the ways for payment on product purchasing preference and its brain neural mechanisms, event-related potential (ERP) experiments with two-factor internal design, 2 (product type: hedonic products and practical products)×2 (payment methods: cash payment and non-cash payment), were used based on mental accounting theory. Participants were asked to make purchasing choices under four different experimental scenarios. And the neurophysiological processes were recorded at the same time for consumption choice. The results showed that in early cognitive stage, when the consumers were stimulated by practical products and hedonic products, they automatically allocated more attention to hedonic products than that to practical products and caused more volatile P2 components ((5.809±0.811)μV vs (4.878±0.944)μV ,
P
<0.05); In cognitive process stage, when the consumers faced different payments, it was easy to create conflicts of decision-making, increasing the difficulty of decision-making, cash payments caused less volatile P3 component than that for non-cash payments ((3.267±0.907)μV vs (3.913±0.833)μV,
P
<0.05). In cognitive decision-making stage, when consumers paid for hedonic products, pain of payment and guilt of shopping resulted in more volatile LPP components compared with that for paying practical products in cash((3.825±0.843)μV vs (2.929±0.769)μV,
P
<0.05). And when consumers paid for hedonic products, pleasure of payments and shopping triggered strong emotional arousal resulted in more volatile LPP components compared with that for paying practical products in non-cash ((3.375±0.887)μV vs (2.030±0.768)μV,
P
<0.05).Therefore, consumers preferred practical products when using cash payments and hedonic products when using non-cash payments. Analysis from brain electrophysiological perspective has importance for consumers, merchants and payment institutions.
2019 Vol. 38 (3): 281-290 [
Abstract
] (
299
)
HTML
(1 KB)
PDF
(12446 KB) (
55
)
291
The Distribution of Short-Term Heart Rate Variability in Long-Term Series and the Influence of Aging
Pan Yue, Wang Zhigang, Zhang Zhengguo, Peng Yi
DOI: 10.3969/j.issn.0258-8021.2019.03.005
The autonomic nervous system state reflected by non-invasive heart rate variability (HRV) can be affected by physiological, pathological and psychological factors. In this paper, we proposed to study the distribution of short-term HRV indices in long-term series and explore the possible changes of autonomic nervous system with age in normal people. The data were provided by the database Normal in THEW (http://www.thew-project.org). The 24-hour Holter data of normal people (
n
=177) were divided into 5 age groups (18≤
y
≤25,
n
=35; 25<
y
≤35,
n
=44; 35<
y
≤45,
n
=41; 45<
y
≤55,
n
=34;
y
>55,
n
=23). Linear and non-linear measures of short-term HRV indices (LF/HF and
α
1
) were performed along the 24 h RR interval (RRI) series using a 5 min sliding window with 2.5 min overlap. Then, mean RRI (MRRI) in each sliding window were calculated. For each Holter record, Spearman correlation coefficients (Spearman CC) between MRRI and LF/HF, as well as that between MRRI and
α
1
were calculated. And the percentage of people with good correlation in each age group was counted. Then, 93 subjects (25<
y
≤65) were selected from 177 normal persons and divided into 4 age groups (at intervals of 10 years old) according to the standard of normal working time and sufficient data length. The mean values of sliding windows (EM_MRRI、EM_LF/HF和EM_
α
1
) were calculated in each 2 h period for each person. The results showed that, for Spearman CC, the proportion of people with good correlation remained high (94%~100%) in the 4 age groups with the age≤55. But the percentage of the persons with good correlation decreased sharply in the group with the age > 55 (78.26% for MRRI
vs
LF/HF, 65.22% for MRRI
vs
α
1
). In the morning minimum EMRRI episodes, there were no significant differences in EM_MRRI, EM_LF/HF and EM_
α
1
among the 4 groups, but there might be significant differences in other periods. With the development of wearable technology, the availability of long-term RRI series has been greatly improved. The results of this study provide a new idea for HRV analysis.
2019 Vol. 38 (3): 291-297 [
Abstract
] (
657
)
HTML
(1 KB)
PDF
(834 KB) (
709
)
298
Feature Recognition of Steady Somatosensory Evoked Potential Based on Convolutional Neural Network
Xu Guizhi, Hu Zhongtao, Wang Lei, Qi Zhiguang, Guo Miaomiao
DOI: 10.3969/j.issn.0258-8021.2019.03.006
Brain-computer interface (BCI) system provides a new way of rehabilitation for patients with aphasia. Previous studies have shown that applying a certain frequency of somatosensory stimulation to the finger or median nerve triggers a spatially-specific steady-state somatosensory evoked potential (SSSEP) with the same frequency. In order to enhance the overall performance of a steady-state somatosensory evoked potentials based BCI system, this work aimed at finding the person-specific resonance-like frequencies of left-hand of 12 healthy subjects by using fast fourier transform (FFT), and the significant time and frequency range for signal feature were selected by event related spectral perturbation (ERSP) for detecting steady-state somatosensory evoked potential signals. The experimentally induced EEG signals were filtered by 1 Hz band-pass based on person-specific resonance-like frequencies to obtain the data of the specific frequency band, then the convolutional neural network (CNN) learning algorithm is used to classify the data. The methods of feature extraction and classification using common spatial pattern (CSP) and support vector machine (SVM) was ompared. The results of all the subjects showed that the accuracy of offline classification obtained by CNN learning algorithm was higher than 85% based on the person-specific resonance-like frequencies filtering method, and the classification accuracy of CNN learning algorithm was higher than using of CSP with SVM (91.8%±5.9%
vs
77.4%±8.5%,
P
<0.05). Therefore, compared with the traditional machine learning classification algorithm (such as common spatial pattern with support vector machine), the CNN learning algorithm can significantly improveed the classification accuracy for feature recognition of brain-computer interface based on steady-state somatosensory evoked potential, and improved the overall performance of the brain-computer interface
2019 Vol. 38 (3): 298-305 [
Abstract
] (
425
)
HTML
(1 KB)
PDF
(2820 KB) (
361
)
306
Research on Methods of Semantic Concept Detection from Medical Images
Wang Xuwen, Zhang Yu Guo, Zhen Li Jiao
DOI: 10.3969/j.issn.0258-8021.2019.03.007
Identifying useful concepts from large scale medical images is an important technology for image knowledge representation. Developing semantic concept detection algorithms is helpful to machine understanding and learning latent knowledge from medical images, and plays an important role in image-assisted diagnosis and intelligent image reading. In this study, the problem of detecting high-frequency concepts from medical images was transformed into a multi-label classification task. The deep transfer learning method based on convolutional neural network (CNNs) was used to recognize high-frequency medical concepts. The image retrieval-based topic modeling method was used to obtain the semantically related concepts from the similar images of given medical images. The CLEF cross language image retrieval track (ImageCLEF) launched the ImageCLEFcaption 2018 evaluation task on May, 2018, in which the Concept Detection subtask identified 111,156 semantic concepts from 222,314 training images and 9,938 test images. Our research group presented experimental results of both methods. The CNNs-based deep transfer learning method achieved the F1 score of 0.0928, which ranked second in all the submission teams. The retrieval-based topic model could recall some low-frequency concepts and achieved the F1 score of 0.0907, but dependent heavily on the image retrieval results. The results proved that the CNNs-based deep transfer learning method showed preferable robustness on high-frequency concept detection, but still a lot of room for improvement in the research of large-scale open semantic concept detection.
2019 Vol. 38 (3): 306-314 [
Abstract
] (
336
)
HTML
(1 KB)
PDF
(2728 KB) (
325
)
315
HOMPPD:A Comprehensive Protein Sequence Database for Human Oral Metaproteomic Studies
Song Tingting, Shao Chen, Du Peng, Zhang Benyu, Zhu Weimin, Jiang Jizhi
DOI: 10.3969/j.issn.0258-8021.2019.03.008
Compared with metagenomics, metaproteomics has the advantage of characterizing not only the composition of environmental microbiome, but also their functional dynamics. Metaproteomic exploration of oral microbiome is an emerging approach for the studies of oral diseases. A comprehensive protein database with sufficient coverage of oral microbiome is an essential tool for the success of these studies. The Human Oral Microbiome Database (HOMD) that collects protein sequences from 117 genera and 367 species is a database used most frequently for oral metaproteomic analysis. It has been recently reported. However, a large collection of microorganisms are not yet included in HOMD, which will certainly limit the correct identification. In this study, protein sequences from 95 more genera reported by published oral metagenomic studies were collected and filtered. Redundant sequences from the same taxon were removed. The resulting sequences were then integrated with HOMD to construct the Human Oral MetaProteome Plus Database (HOMPPD). Due to the high individualization of oral microbiota, we proposed an improved two-step search method, that is, a two-step search for each sample. In order to evaluate the efficient identification of these newly included oral microorganisms, we then used our database to reanalyze public oral mass spectrometry raw data. Our results demonstrated that the constructed HOMPPD collected proteins sequence of 184 genera and 2793 species, 39 new oral genera and 124 new oral species were confidentially identified when searching HOMPPD, and that HOMPPD is a better tool for oral metaproteomics researches for its significantly more comprehensive oral microbiome coverage. HOMPPD is publicly available at ftp://111.198.139.72:4000//pub//metaproteomics//homppd.fasta.
2019 Vol. 38 (3): 315-323 [
Abstract
] (
580
)
HTML
(1 KB)
PDF
(5237 KB) (
422
)
324
Research on Heterogeneity and Compatibility of Biomedical Field Metadata Supported by Ontology
Zhang Lulu, Yang Sheng, Shi Furen, Pan Hongjie, Wang Zhigang, Yang Xiaolin
DOI: 10.3969/j.issn.0258-8021.2019.03.009
Using ontologies to support the representation of data elements is an important means to improve the machine′s understanding of metadata. In this paper, we evaluated the semantic heterogeneity of data elements in caDSR and assessed two related data elements integration ability. First, 60 pairs of common data elements were selected from caDSR, covering demography, lifestyle, medical history and laboratory measurements. Next, the essential components of data elements were extracted according to the ISO/IEC11179 standard and the similarity of these essential components between every pair of data elements with the support of NCIT was calculated. At last, the compatibility between related data elements was predicted by using SVM based on the semantic similarity between corresponding CDE components. The overall accuracy was above 80%. The results showed that there was currently considerable heterogeneity in the definition of metadata in the caDSR database, especially in the conceptual domain of data elements. Nevertheless, our method still could realize the automatic judgement of data compatibility based on the definition of existing data elements by the help of machine learning. The method established in this study has a certain value for optimizing data element construction process and enriching data standardization tools.
2019 Vol. 38 (3): 324-331 [
Abstract
] (
445
)
HTML
(1 KB)
PDF
(2554 KB) (
388
)
332
Experimental Mechanism Study on Killing of Chemoresistant Tumor Cells Exposed to High Voltage Nanosecond Pulse Field
Liu Hongmei, Wang Li, Dong Shoulong, Ma Jianhao, Wang Yilin, Yao Chenguo
DOI: 10.3969/j.issn.0258-8021.2019.03.010
Tumor seriously endanger the lives and health of human people. Chemoresistance is one of the main barriers to successful tumor treatment. In order to explore whether high voltage nanosecond pulsed electric fields can be equally sensitive to chemotherapy-resistant tumor cells compared with its homology tumor cells, lung cancer cell line A549 and cisplatin chemotherapy-resistant A549/R were used as the research object in this paper. After exposing to the 80 nanosecond pulses with a pulse width of 200 ns and a field strength of 5-15 kV/cm, the activity of cell, apoptosis and necrosis and ablation effect were analyzed and compared in groups. The results showed that the survival rate of A549/R was 5.65%, 7.78% and 2.80% respectively when exposed topulses electric field with strength of 8kV/cm, 10 kV/cm and 15 kV/cm, which was significantly lower than that of A549 cells (48.31%, 26.8%, and 5.96%, respectively for 8 kV/cm, 10 and 15 kV/cm). The induced apoptosis and necrosis rates were also significantly higher than A549 (
P
<0.05); In addition, when exposed to 80 pulses(200 ns, 15 kV/cm), the ablation area of single-layer cells for A549/R was 1.59 times that of A549 cells, and the ablation threshold for A549/R (8 kV/cm) is significantly lower than that of A549 (13 kV/cm). Therefore, the nanosecond pulsed electric field with a pulse width of 200 ns could preferentially kill A549/R, inducing a higher rate of apoptosis, ablation area and a lower ablation threshold. The experimental results showed that high-voltage nanosecond pulse could preferentially act on chemotherapy-resistant tumor cells.
2019 Vol. 38 (3): 332-338 [
Abstract
] (
369
)
HTML
(1 KB)
PDF
(3550 KB) (
324
)
339
Finite Element Analysis of Mechanical Properties of Graded Sphere Porosity Structure with Different Pore Size Distributions
Shi Zhiliang, Huang Chen, Lu Xiaolong, Li Feng, Sun Yunlong
DOI: 10.3969/j.issn.0258-8021.2019.03.011
To establish a three-dimensional finite element model of titanium alloy porous structure with the different pore size distributions and to implement the finite element mechanical analysis. The sphere, sphere_line, sphere_plane, sphere_point three-dimensional models of porous structure was created by software Rhino 5.0. The finite element analysis was implemented by Abaqus 6.1 to calculate the Von-Mises stress and maximum principal stress. Before calculated, The 200, 400, 600, 800, and 1000 N loading forces were applied to the upper surface, and the direction was vertically downward. The lower surface was selected as a fixed constraint. Under the same loading force, thesphere plane has the maximum principal stress and the maximum equivalent stress,followed were the sphere point gradient pore model and sphere line gradient pore model, the sphere without gradient pore model was the smallest. Among the four porous structures with different gradient pore size distributions, sphere has the lowest actual load and the best carrying capacity, the next was sphere line gradient hole structure and then sphere point. Sphere plane gradient pore structure had the maximum actual load, and its carrying capacity was worse than the three formers. Finally, the four models that were designed in this work were prepared by rapid prototyping to obtain the porous titanium alloy body, and the mechanical property test on these bodies was carried out. The conclusion verifies the reliability of numerical simulation results. The results of this research could be used to provide relevant reference and theoretical basis for porous titanium alloy implants and clinical applications.
2019 Vol. 38 (3): 339-347 [
Abstract
] (
467
)
HTML
(1 KB)
PDF
(10677 KB) (
85
)
Reviews
348
Advances in the Extraction and Classification of EEG Dynamic Features in Patients with Mild Cognitive Impairment
Li Hengzhi, Wen Dong, Wei Zhenhao, Zhou Yanhong
DOI: 10.3969/j.issn.0258-8021.2019.03.012
Electroencephalography (EEG) can be used to assess pathological changes in mild cognitive impairment (MCI) disorders. In recent years, the feature extraction and classification methods in the field of EEG have been widely applied to the diagnosis of mild cognitive impairment diseases. In order to understand the current development of MCI EEG signal analysis, this work first analyzed in depth the application of EEG signal from MCI patients in the field of feature extraction, its advantages and disadvantages from two aspects of local coupling and global synchronization. Then the method of classification of EEG signal from MCI patients was summarized and analyzed, such as support vector machines, K-means, and convolutional neural networks that have been widely used in recent years. Finally, the future development trend of dynamic feature extraction and classification methods was prospected
2019 Vol. 38 (3): 348-354 [
Abstract
] (
441
)
HTML
(1 KB)
PDF
(787 KB) (
555
)
355
A Review on the Objective Evaluation of Cochlear Implant Speech Perception with Mismatch Negativity
Xu Danying, Chen Fei
DOI: 10.3969/j.issn.0258-8021.2019.03.013
Speech perception by cochlear implants is one of the research hotspots in speech communication and hearing rehabilitation. Because of the limitations of traditional subjective behavioral assessment methods, the mismatch negativity from electroencephalography has been used to objectively assess the performance of speech communication and hearing rehabilitation. The mismatch negativity does not require listeners′ attention modulation, and can provide more objective and accurate research findings when compared with other behavioral experiments. Nowadays, mismatch negativity has been widely used in the field of cochlear implants. This paper reviewed the advances of this field from several aspects, including the principle of cochlear implant, mismatch negativity evaluation method and the state of the art in objectively evaluating cochlear implant speech perception with mismatch negativity. Meanwhile, it also introduced the performance of cochlear implants perception in speech and musical sound, and hearing recovery performance after implantation. The limitations of these methods were discussed, and research directions were summarized.
2019 Vol. 38 (3): 355-366 [
Abstract
] (
346
)
HTML
(1 KB)
PDF
(1383 KB) (
814
)
367
Research Progress on the Structure Design and Optimization of Biodegradable Stents
Peng Kun, Li Jing, Wang Sirui, Xia Jun, Qiao Aike
DOI: 10.3969/j.issn.0258-8021.2019.03.014
Due to the advantages of degradability and excellent biocompatibility, the biodegradable alloy stents, are expected to be the promising solution for the problems such as the restenosis of the permanent stents, the late stents thrombosis of drug eluting stents, leading the fourth generation of innovation in the field of the cardiovascular interventional therapy. However, insufficient scaffold performance and too short degradation time are the main limitations for the biodegradable stents in clinical application currently. To deal with these problems, stent design and optimization can be an effective method to improve scaffold performance of biodegradable stents and prolong their service time. Structure design and optimization of biodegradable stents is a complex process involving balance of different mechanical performances, hemodynamics and the changing of scaffold performance in corrosion environment. Therefore, in this article, the state-of-the-art design and optimization of the biodegradable stents were reviewed and the challenges and future research directions were pointed out.
2019 Vol. 38 (3): 367-374 [
Abstract
] (
507
)
HTML
(1 KB)
PDF
(4197 KB) (
490
)
Communications
375
Study on Magnetic Resonance Electrical Characteristic Imaging Method of Permanent Magnet
Zhao Shilong, Li Xiaonan, Liu Guoqiang, Hu Lili, Chen Haiyan
DOI: 10.3969/j.issn.0258-8021.2019.03.015
2019 Vol. 38 (3): 375-379 [
Abstract
] (
299
)
HTML
(1 KB)
PDF
(7908 KB) (
127
)
380
Design of Blood Cell Damage Monitoring Device for Cardiopulmonary Bypass System
Wang Taotao, Gu Xuelian, Li Yueru, Zhang Xiaoying
DOI: 10.3969/j.issn.0258-8021.2019.03.016
2019 Vol. 38 (3): 380-384 [
Abstract
] (
275
)
HTML
(1 KB)
PDF
(10868 KB) (
63
)
Copyright © Editorial Board of Chinese Journal of Biomedical Engineering
Supported by:
Beijing Magtech