Identification of Voice Disorders Using Long-Time Features and Support Vector Machine With Different Feature Reduction Methods

Identification of voice disorders has a fundamental role in our life nowadays. Therefore, many of these diseases must be diagnosed at early stages of occurrence before they lead to a critical condition. Acoustic analysis can be used to identify voice disorders as a complementary technique with other traditional invasive methods, such as laryngoscopy. In this article, we followed an extensive study in the diagnosis of voice disorders using the statistical pattern recognition techniques. Finally, we proposed a combined scheme of feature reduction methods followed by pattern recognition methods to classify voice disorders. Six classifiers are used to evaluate feature vectors obtained by principal component analysis or linear discriminant analysis (LDA) as feature reduction methods. Furthermore, individual, forward, backward, and branch-and-bound methods are examined as feature selection methods. The performance of each combined scheme is evaluated in terms of the accuracy, sensitivity, specificity, and area under the receiver operating characteristic curve (AUC). The experimental results denote that LDA along with support vector machine (SVM) has the best performance, with a recognition rate of 94.26% and AUC of 97.94%. Additionally, this structure has the lowest complexity in comparison with other architectures. Among feature selection methods, individual feature selection followed by SVM classifier shows the best recognition rate of 91.55% and AUC of 95.80%.     

Effects of trabecular bone on marrow relaxation in the tibia

The effects of bone on marrow relaxation in the trabecular volume of the most proximal 3 cm in the left tibia were studied with a RF-spoiled gradient echo MRI protocol on a 1.0 T MR unit. The MR measurements were performed on six healthy volunteers, and repeated within one month in order to assess the precision of the method. In the same subjects, the area bone mineral density (bmd, g/cm²) was measured at the left proximal femur using dual-energy X-ray absorptiometry. The calcaneus of the same side was examined with quantitative ultrasound. The marrow T21 relaxation deviated from a mono-exponential decay, and resembled the decay of subcutaneous fat. The shape of the relaxation curve reflected the presence of several spectral components in bone marrow, and was further influenced by the amount and structure of the surrounding trabecular bone. The bone marrow decays showed substantially reduced inter-subject variability after normalisation of the marrow data fit parameters to corresponding values for s.c. fat. This suggests the use of an internal adipose tissue reference in order to correct for diet-related variations of marrow T21 estimates. The mean relative precision of the MR measurements was between 5% and 10% depending on the data fit model. Moderate-to-strong correlations between DXA bmd indices in the proximal femur and MR parameters were found (r_max = −0.96; p < 0.01), while ultrasound-derived measures of bone strength measured on the calcaneus demonstrated significantly weaker correlations to the MR parameters (r_max = −0.78; p > 0.05). The method employed in this study showed reasonable precision and a moderate to good correlation compared to other bone parameters derived at the same extremity, and is a promising tool for the use on patients.     

Characterisation of cryopreserved cells freshly isolated from human bone marrow

Osteoblast progenitor cells (OBPCs) isolated from bone marrow have the ability to differentiate into osteoblasts and thus potential therapeutic use to tissue-engineer bone. In order for OBPCs to be available for clinical use a means of storing viable cells is necessary. The aim of this study was to determine whether a simple method of cryopreservation had an effect on osteogenic differentiation or growth of OBPCs isolated from fresh human bone marrow. Stro-1 was used to identify the isolated OBPCs. The osteoblastic potential of the marrow cells was confirmed as culture with osteogenic supplements (OS) significantly increased osteoblastic protein production (alkaline phosphatase (ALP), osteopontin and osteocalcin) compared with standard conditions (P less than 0.05). Ten further marrow aspirates were harvested; each was halved for either cryopreservation or control culture. Primary cultures from both populations formed colonies with recognised OBPC morphology. OS stimulated both cryopreserved and control populations to produce significantly more osteoblastic proteins (P less than 0.05) and there was no significant difference between the increase in osteogenic proteins when cultured with OS (P great than 0.2). The proliferation rate after 5 days in culture was not significantly affected by cryopreservation (P greater than 0.05). It has been suggested that OBPCs are immuno-privileged; so allogenic cells could be implanted into patients for tissue engineering bone without causing a hypersensitivity reaction. Our study demonstrates a method of storage, which allows OBPCs to be available for use without affecting osteoblastic potential or viability.     

Differentiation of bone metastases from prostate cancer and benign red marrow depositions of the pelvic bone with multiparametric MRI

PurposeTo investigate the diagnostic utilities of imaging parameters derived from T1-weighted imaging (T1WI), diffusion-weighted imaging (DWI) and dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) to differentiate bone metastases from prostate cancer and benign red marrow depositions of the pelvic bone.Materials and methodsThirty-six lesions from 36 patients with prostate cancer were analyzed with T1WI, DWI, and DCE-MRI. The lesions were classified in the bone metastases (n = 22) and benign red marrow depositions (n = 14). Lesion-muscle ratio (LMR), apparent diffusion coefficient (ADC), volume transfer constant (K^trans), reflux rate (Kep), and volume fraction of the extravascular extracellular matrix (Ve) values were obtained from the lesions. The imaging parameters of the both groups were compared using the Mann-Whitney U test, receiver operating characteristics (ROC) curves were analyzed. For the ROC curves, area under the curves (AUCs) were compared.ResultsThe ADC, K^trans, K_ep, and V_e values of bone metastases were significantly higher than those of benign red marrow depositions (Mann-Whitney U test, p < 0.05). However, there was no significant difference in LMR between the two groups (Mann-Whitney U test, p = 0.360). The AUCs of K^trans_, K_ep, ADC, V_e, and LMR were 0.896, 0.844, 0.812, 0.724, and 0.448, respectively. In the pairwise comparison of ROC curves, the AUCs of K^trans and K_ep was significantly higher than LMR.ConclusionsK^trans, K_ep, V_e, and ADC values can be used as imaging tools to differentiate bone metastases from prostate cancer and benign red marrow depositions of the pelvic bone.     

基于小波系数图和卷积神经网络的太赫兹光谱物质识别

许多太赫兹光谱物质识别方法依靠寻找该物质在太赫兹波段范围内不同光谱表现出的不同特征来识别特定物质。吸收峰提取法是常用的光谱特征提取算法,但当光谱无明显特征吸收峰或峰位、峰值相近或难以识别时,难以利用吸收峰特征辨别物质。将机器学习和统计学习技术用于太赫兹光谱的识别中虽减少了吸收峰的干扰,但常常需要人为定义特征而导致分类误差。深度学习法能自动提取特征,但在识别前往往需要进行复杂的预处理操作,并且在特征提取的过程中容易丢失部分特征从而导致分类误差。针对以上问题,提出了一种基于小波系数图和卷积神经网络的太赫兹光谱识别方法。利用太赫兹光谱信号进行小波变换时,由于小波系数矩阵的每一行系数与原始光谱信号存在着对应关系,因此将太赫兹光谱的吸收系数通过小波变换在频率域上展开,能得到不同的二维的频率-尺度分布图,又称小波系数图。然后构造一个卷积神经网络(CNN)对小波系数图进行分类,可得到太赫兹光谱物质的分类结果。为了验证所提出算法的有效性,将三组小波系数图数据与原始光谱数据分别输入CNN、Support Vector Machin (SVM)、Multilayer Perceptron (MLP)三种不同的分类器作对比,从实验结果可以发现本文算法在三组数据中的识别率均达到了100%,说明相比于传统方法,本文方法能准确分类没有明显特征吸收峰的光谱,证明了使用卷积神经网络识别小波系数图的有效性。为了体现本文算法的优势,与小波脊线寻峰识别算法作对比,实验结果表明本文算法几乎不受峰频、峰位、峰值的影响,无论是识别不存在吸收峰的淀粉,还是识别相似度高的蔗糖和葡萄糖,都具有较高的识别率,分类准确率达97.62%,证明了所提算法的优越性。该算法为太赫兹光谱数据识别提供了一种新思路,同时也可以推广运用到其他谱图物质的识别中。     

Permutation Entropy-Based Interpretability of Convolutional Neural Network Models for Interictal EEG Discrimination of Subjects with Epileptic Seizures vs. Psychogenic Non-Epileptic Seizures

The differential diagnosis of epileptic seizures (ES) and psychogenic non-epileptic seizures (PNES) may be difficult, due to the lack of distinctive clinical features. The interictal electroencephalographic (EEG) signal may also be normal in patients with ES. Innovative diagnostic tools that exploit non-linear EEG analysis and deep learning (DL) could provide important support to physicians for clinical diagnosis. In this work, 18 patients with new-onset ES (12 males, 6 females) and 18 patients with video-recorded PNES (2 males, 16 females) with normal interictal EEG at visual inspection were enrolled. None of them was taking psychotropic drugs. A convolutional neural network (CNN) scheme using DL classification was designed to classify the two categories of subjects (ES vs. PNES). The proposed architecture performs an EEG time-frequency transformation and a classification step with a CNN. The CNN was able to classify the EEG recordings of subjects with ES vs. subjects with PNES with 94.4% accuracy. CNN provided high performance in the assigned binary classification when compared to standard learning algorithms (multi-layer perceptron, support vector machine, linear discriminant analysis and quadratic discriminant analysis). In order to interpret how the CNN achieved this performance, information theoretical analysis was carried out. Specifically, the permutation entropy (PE) of the feature maps was evaluated and compared in the two classes. The achieved results, although preliminary, encourage the use of these innovative techniques to support neurologists in early diagnoses.     

Radiotherapy effects on vertebral bone marrow: Easily recognizable changes in T2 relaxation times

The effect of localized radiotherapy on vertebral bone marrow was demonstrated in two patients using quantitative MRI studies with pixel-by-pixel measurement of T₂ relaxation times with generation of T₂ images. Conventional T₁-weighted spin-echo images were obtained as well. Irradiated vertebral bone marrow was found to have longer T₂ relaxation times than the neighboring nonirradiated bone marrow. These changes corresponded to the increased signal intensity on T₁-weighted images and to the field of radiotherapy and were noted 2.5 to 32 mo after radiotherapy. Radiologists should be aware of the increased T₂ relaxation times in irradiated bone marrow to correctly assess spinal disorders in irradiated patients. The reported T₂ changes may reflect the abundance of adipose cells that proliferate in bone marrow after radiotherapy, or may indicate an additional histological change, such as bone marrow necrosis or edema. Conclusive histological proof remains to be obtained.     

Recognition algorithm of internal defect images of thermal battery北大核心CSCD

针对目前热电池内部装配缺陷检测效率低、准确度不高的问题,研究了一种可精准分割内部电池堆图像并能够准确识别缺陷种类的方法。首先采用水平、垂直积分投影法对目标电池堆边缘特征进行提取,利用局部自适应对比度增强算法对局部不清晰部分进行细节纹理增强;然后研究了缺陷结构的灰度特性,计算提取出缺陷特征参数;最后使用BP（back propagation）神经网络和CART（classification and regression tree）决策树对特征参数分类识别,并根据分类准确度进行权重分配,将加权融合后的结果作为检测的最终判据。实验结果表明:该方法对2 000个样本的检测准确度达98.9%,为热电池的X射线缺陷检测提供了有效的途径。     

Application of High-Dimensional Feature Selection in Near-Infrared Spectroscopy of Cigarettes' Qualitative Evaluation

In order to increase the classification accuracy, a new feature selection method, RFFIM-PCA, based on the random forest feature importance measure (RFFIM) and principal component analysis (PCA) for analyzing the near-infrared (NIR) spectra of tobacco, is presented in this paper. We applied the method to the classification of cigarettes' qualitative evaluation and also compared it with other methods. The result showed that RFFIM-PCA discriminates the high-dimensional data effectively and can be used to identify the cigarettes' quality. The feature selection filters the noises, while PCA eliminates the redundant features and reduces the dimensionalities as well. The experimental results showed that RFFIM-PCA successfully eliminated the noises and redundant features in high-dimensional data, leading to a promising improvement on the feature selection and classification accuracy.     

Quantification of lipids in human lower limbs using yellow bone marrow as the internal reference: gender-related effects

The main purpose of this study was to determine and compare extra- and intramyocellular (IMCL) lipids content in the calf and thigh muscles of normal male and female volunteers using high-spatial-resolution magnetic resonance spectroscopic imaging (MRSI). The study groups consisted of 10 females and 10 males. The subjects were healthy and normal-weight. Fat (yellow bone marrow) was used as the internal concentration reference. Total fat and IMCL content were computed for all muscles in the slice and for three muscle compartments in the thigh, whereas three muscles and three muscle compartments were evaluated in the calf. To avoid the confounding effects of physical activity and diet, measurements were performed in the same session. A common feature for both genders was that thigh muscles had approximately 2.5 times greater total fat content as compared to muscles of the calf. The mean IMCL level was, however, more than 3 times higher in the calf muscles compared with the thigh. No significant differences in lipid concentrations of correspondent regions of interest were found between genders. The high-spatial-resolution MRSI technique enables a more detailed study of muscle lipid distribution and can therefore improve understanding of muscle lipid metabolism in healthy volunteers and in studies of patients with metabolic disorders.     

Localized in vivo proton spectroscopy of the bone marrow in patients with leukemia

Volume selective magnetic resonance (MR) proton spectroscopy was used to investigate the haemopoietic (iliac bone) and fatty bone marrow (tibia) in patients with leukemia and polycythaemia vera. Selective measurements of the relaxation times T₁ and T₂ for the “water” and “fat” resonances in the bone marrow spectra were performed. Nine patients with acute leukemia and three patients with chronic leukemia were examined at diagnosis. Three patients with acute leukemia in remission were also examined. Five of the leukemic patients had follow-up examinations performed in relation to chemotherapeutic treatment. Nine patients with polycythaemia vera and 21 normal control subjects were examined with identical methods for comparison. All patients had bone marrow biopsies performed prior to every MR examination. Significant differences could be detected in the spectral patterns from iliac bone marrow in patients with leukemia at diagnosis compared to the healthy normal controls. The “relative water content” was increased in the iliac bone marrow spectra of the leukemic patients compared to the normal subjects, which indicates an increase in the amount of haemopoietic tissue and a corresponding decrease in marrow fat content. The T₁ relaxation times of the “water” resonance in the spectra from the iliac bone marrow of the leukemic patients were significantly prolonged at diagnosis, compared to the normal controls and the patients with polycythaemia vera. After chemotherapeutic induction of remission, the spectra from the iliac bone marrow in the patients with leukemia resembled normal spectra. Four leukemic patients had abnormal spectra from the tibial bone marrow and one patients showed early changes in tibial marrow during chemotherapeutic treatment, before any major changes could be detected in the iliac bone marrow.     

Detection of fatty acid composition of trabecular bone marrow by localized iDQC MRS at 3 T: A pilot study in healthy volunteers

BackgroundAlthough a growing body of research shows that the bone marrow adipose tissue (BMAT) may play an essential role in bone inflammation and energy metabolism, available noninvasive methods for distinguishing different fatty acids in BMAT are still limited, in spite of their potential to provide novel biomarkers for bone related diseases.PurposeTo assess the ability of a localized intermolecular double quantum coherence (iDQC) spectroscopy sequence to resolve more fatty acid peaks than conventional MR spectroscopy (MRS), like polyunsaturated fatty acids (PUFA), from the human BMAT in the presence of trabecular bone; To preliminarily investigate whether the fatty acids composition is different between different regions and groups.ResultsCompared with conventional MRS results, additional four fatty acids peaks were well resolved using the proposed method in human BMAT in the presence of trabecular bone. In addition, a different fat composition was found between distal femur and proximal tibia: fat was more unsaturated (vinyl, *p < 0.01; diallylic, *p < 0.01) in distal femur bone marrow than in proximal tibia, and this higher unsaturation level was caused by PUFA (r = 0.67, diallylic, *p < 0.01). No significant difference in fatty acid composition were found either between left and right legs, or between female and male in the healthy young subjects studied.ConclusionThis study demonstrated that the unsaturated fatty acids information of human BMAT in the presence of trabecular bone can be clearly identified with the localized iDQC at 3 T. The resolved peaks, especially PUFA, may serve as additional diagnostic biomarkers for BMAT related diseases in the future.     

Classification and recognition of transgenic product by terahertz spectroscopy and DSVM

The purpose of this paper is to construct a classification model that can identify class accurately and control imbalance. A novel adaptive decision support vector machine (DSVM) is proposed for the recognition of transgenic cotton seed based on terahertz spectroscopy (THz), which make the traditional support vector machine is ability of adaptive decision, and select optimal parameters by using particle swarm optimization (PSO). For the classification and recognition of the transgenic cotton seeds, firstly, the factor analysis (FA) is applied to reduce the dimension and extract the feature spectrum of original spectral information. Secondly, the feature spectrum is selected and fed into the model of DSVM to recognize the different transgenic cotton seeds. The experimental results show that the proposed method can effectively classify the different transgenic cotton seeds, and its recognition rate surpasses the comparative method evidently.     

Application of magnetic poly(styrene-glycidyl methacrylate) microspheres for immunomagnetic separation of bone marrow cells

Surface-functionalized magnetic poly(styrene-glycidyl methacrylate) (PS-GMA) microspheres were prepared and coupled with Sca-1 antibody for cell selection from murine bone marrow mononuclear cells (MNCs). Biotinylated Sca-1 antibody could be directly coupled to avidin-bound magnetic microspheres. Alternatively, oxidized goat anti-mouse antibody was covalently bound onto the amino group-containing magnetic microspheres in a site-directed manner, and the resultant conjugate was coupled with non-modified Sca-1 antibody. Using the indirect antibody-bound magnetic microspheres, the purity of isolated Sca-1⁺ cells increased with bead-to-cell ratio. Using a bead-to-cell ratio of 10 beads/cell, a purity of 85% Sca-1⁺ cells corresponding to a 17-fold enrichment was achieved.     

Distinction between normal and leukemic bone marrow by water protons nuclear magnetic resonance relaxation times

Pulsed nuclear magnetic resonance studies have been carried out on bone marrow of normal human subjects and patients with leukemia: chronic myeloid leukemia (CML) and acute myeloid leukemia (AML). It was observed that the proton spin-lattice relaxation time (T₁) value was discriminatory in the normal and leukemic cases with a statistical significance of (p < 0.01). Ouabain treatment of cells did not show any perceptible change of T₁ value when compared with the nontreated cells, indicating that the concomitant cation effluxes do not affect spin-lattice relaxation time. The water contents of normal, leukemic, and ouabain treated cells were in the range 60%–80%. Higher Fe levels were encountered in the normal than the leukemic samples, while levels of Zn, Cu, Mn, Co, and Ni were elevated in the leukemic samples compared with the normals. Despite the T₁ differences observed, the multiparameter studies do not uniquely pinpoint factors responsible for the elevation of T₁ in the malignant state.     

Classification of spontaneous EEG signals in migraine

We set up a classification system able to detect patients affected by migraine without aura, through the analysis of their spontaneous EEG patterns. First, the signals are characterized by means of wavelet-based features, than a supervised neural network is used to classify the multichannel data. For the feature extraction, scale-dependent and scale-independent methods are considered with a variety of wavelet functions. Both the approaches provide very high and almost comparable classification performances. A complete separation of the two groups is obtained when the data are plotted in the plane spanned by two suitable neural outputs.     

Proton relaxation times in human red bone marrow by volume selective magnetic resonance spectroscopy

In hematological diseases the composition of red bone marrow shows alterations. The relaxation timesT ₁ andT ₂ of water and lipids in human hemopoietic bone marrow of 14 normal volunteers and 10 patients with acute leukemia and bone marrow carcinosis are determined using a double spin echo spectroscopy sequencein vivo. The volumes of interest (VOI) of (13 mm)³ in the center of vertebral bodies are examined using different measurement parameters. ForT ₁ measurements an inversion-recovery method is used.T ₂ is evaluated from spectra with differentTE. T ₁ (water) is found in a range between 1000 and 1700 ms,T ₁ (lipids) in a range between 260 and 320 ms in healthy volunteers.T ₂ (water) is determined between 32 and 65 ms. In some cases phase distortions of the water signals occur in the spectra. Water flow within the VOI may be a possible reason.T ₂ (lipids) is evaluated between 73 and 91 ms. The patients with acute leukemia exhibit clearly reduced lipid signals in their spectra. Lipid relaxation times could not be determined in these cases.T ₂ (water) is prolonged in acute leukemia to 51–98 ms.T ₁ (water) was not significantly different from values of healthy volunteers in our measurements. Results are discussed in comparison to relaxometric data from imaging and STEAM spectroscopic methods of other authors.     

EEG-Based Emotion Recognition by Exploiting Fused Network Entropy Measures of Complex Networks across Subjects

It is well known that there may be significant individual differences in physiological signal patterns for emotional responses. Emotion recognition based on electroencephalogram (EEG) signals is still a challenging task in the context of developing an individual-independent recognition method. In our paper, from the perspective of spatial topology and temporal information of brain emotional patterns in an EEG, we exploit complex networks to characterize EEG signals to effectively extract EEG information for emotion recognition. First, we exploit visibility graphs to construct complex networks from EEG signals. Then, two kinds of network entropy measures (nodal degree entropy and clustering coefficient entropy) are calculated. By applying the AUC method, the effective features are input into the SVM classifier to perform emotion recognition across subjects. The experiment results showed that, for the EEG signals of 62 channels, the features of 18 channels selected by AUC were significant (p < 0.005). For the classification of positive and negative emotions, the average recognition rate was 87.26%; for the classification of positive, negative, and neutral emotions, the average recognition rate was 68.44%. Our method improves mean accuracy by an average of 2.28% compared with other existing methods. Our results fully demonstrate that a more accurate recognition of emotional EEG signals can be achieved relative to the available relevant studies, indicating that our method can provide more generalizability in practical use.     

Distinct patterns of active and non-active plaques using texture analysis on brain NMR images in multiple sclerosis patients: preliminary results

The benefits of texture analysis of magnetic resonance images have been assessed in multiple sclerosis (MS) patients. Out of thirty-two lesions identified in eight MS patients, nine were considered active, judging from their gadolinium uptake. Texture analysis allowed to obtain forty-two characterizing parameters for each lesion. Using discriminant analysis as a statistical method allowed to classify the lesions into two groups: active or non-active. An attempt to classify their level of activity by using only co-occurrence matrices was unsuccessful. Alternately, the same type of analysis performed on runlength analysis criteria allowed the accurate classification of 88% of active lesions and 96% of non-active lesions. Using incremental discriminate analysis can reduce the number of useful parameters. This method showed that among the 42 parameters, 8 only were highly significant and permitted an accurate classification. Five of these parameters are runlength parameters, and three others are more directly related to the global distribution. The main interest of runlength parameters is that they allowed to demonstrate that the lesion structure was different in active and non-active plaques. This preliminary work suggests that using texture analysis could be of interest in the follow-up of MS patients because it provides an opportunity to identify active lesions without frequent gadolinium injections.     

Radiomic prediction of mutation status based on MR imaging of lung cancer brain metastases

Lung cancer metastases comprise most of all brain metastases in adults and most brain metastases are diagnosed by magnetic resonance (MR) scans. The purpose of this study was to conduct an MR imaging-based radiomic analysis of brain metastatic lesions from patients with primary lung cancer to classify mutational status of the metastatic disease. We retrospectively identified lung cancer patients with brain metastases treated at our institution between 2009 and 2017 who underwent genotype testing of their primary lung cancer. Brain MR Images were used for segmentation of enhancing tumors and peritumoral edema, and for radiomic feature extraction. The most relevant radiomic features were identified and used with clinical data to train random forest classifiers to classify the mutation status. Of 110 patients in the study cohort (mean age 57.51 ± 12.32 years; M: F = 37:73), 75 had an EGFR mutation, 21 had an ALK translocation, and 15 had a KRAS mutation. One patient had both ALK translocation and EGFR mutation. Majority of radiomic features most relevant for mutation classification were textural. Model building using both radiomic features and clinical data yielded more accurate classifications than using either alone. For classification of EGFR, ALK, and KRAS mutation status, the model built with both radiomic features and clinical data resulted in area-under-the-curve (AUC) values based on cross-validation of 0.912, 0.915, and 0.985, respectively. Our study demonstrated that MR imaging-based radiomic analysis of brain metastases in patients with primary lung cancer may be used to classify mutation status. This approach may be useful for devising treatment strategies and informing prognosis.