Search

Objective. The data scarcity problem in emotion recognition from electroencephalography (EEG) leads to difficulty in building an affective model with high accuracy using machine learning algorithms or deep neural networks. Inspired by emerging deep generative models, we propose three methods for augmenting EEG training data to enhance the performance of emotion recognition models. Approach. Our proposed methods are based on two deep generative models, variational autoencoder (VAE) and generative adversarial network (GAN), and two data augmentation ways, full and partial usage strategies. For the full usage strategy, all of the generated data are augmented to the training dataset without judging the quality of the generated data, while for the partial usage, only high-quality data are selected and appended to the training dataset. These three methods are called conditional Wasserstein GAN (cWGAN), selective VAE (sVAE), and selective WGAN (sWGAN). Main results. To evaluate the effectiveness of these proposed methods, we perform a systematic experimental study on two public EEG datasets for emotion recognition, namely, SEED and DEAP. We first generate realistic-like EEG training data in two forms: power spectral density and differential entropy. Then, we augment the original training datasets with a different number of generated realistic-like EEG data. Finally, we train support vector machines and deep neural networks with shortcut layers to build affective models using the original and augmented training datasets. The experimental results demonstrate that our proposed data augmentation methods based on generative models outperform the existing data augmentation approaches such as conditional VAE, Gaussian noise, and rotational data augmentation. We also observe that the number of generated data should be less than 10 times of the original training dataset to achieve the best performance. Significance. The augmented training datasets produced by our proposed sWGAN method significantly enhance the performance of EEG-based emotion recognition models.

The lack of training data is an obstacle to build satisfactory multimodal emotion recognition models. Generative adversarial network (GAN) has recently shown great successes in generating realistic-like data. In this paper, we propose a GAN-based data augmentation method for enhancing the performance of multimodal emotion recognition models. We adopt conditional Boundary Equilibrium GAN (cBEGAN) to generate artificial differential entropy features of electroencephalography signal, eye movement data and their direct concatenations. The main advantage of cBEGAN is that it can overcome the instability of conventional GAN and has very quick converge speed. We evaluate our proposed method on two multimodal emotion datasets. The experimental results demonstrate that our proposed method achieves 4.6% and 8.9% improvements of mean accuracies on classifying three and five emotions, respectively.

To explore selection and efficacy assessment for membranous urethral trauma caused by pelvic fracture.From June 2000 to August 2010, 72 patients with membranous urethral trauma caused by pelvic fracture were selected. There were 46 males and 26 females,ranging age from 26 to 62 years (averaged 35.2 years). The time from injury to hospitalization time was 1 to 3 hours. According to Tile pelvic fracture classification, there were 8 patients with type A, 45 patients with type B, 19 patients with type C. Thirty of the 35 patients with partial rupture of posterior urethral were treated by catheterization,5 patients treated by rupture anastomosis on the stage I combined with cystostomy; 25 of the 37 patients with complete rupture of posterior urethra were treated by early realignment, and 12 patients were treated by cystostomy. Urinary incontinence, impotence and urethrostenosis were evaluated.All patients were followed up for 5 to 10 years (mean 7.7 years). Incidence of urethrostenosis, impotence and urinary incontinence in patients treated by cystostomy were significantly higher than rupture anastomosis on the stage I and early realignment (P0.05); while incidence in patients treated by catheterization was significantly lower than other groups (P0.05).For patients with partial rupture of posterior urethral, catheterization and rupture anastomosis on the stage I are preferred methods; while patients with complete rupture of posterior urethra, early realignment is a preferred method with advantages of simple operation and less complications.

Background More than 1 million tuberculosis (TB) patients are receiving the standard anti-TB treatment provided by China National Tuberculosis Prevention and Control Scheme (CNTS) in China every year. Adverse reactions (ADRs) induced by anti-TB drugs could both do harm to patients and lead to anti-TB treatment failure. The ADACS aimed to explore ADRs' incidences, prognoses, economical and public health impacts for TB patients and TB control, and build a DNA bank of TB patients. Methods/Design Multiple study designs were adopted. Firstly, a prospective cohort with 4488 sputum smears positive pulmonary tuberculosis patients was established. Patients were followed up for 6-9 months in 52 counties of four regions. Those suspected ADRs should be checked and confirmed by Chinese State Food and Drug Administration (SFDA). Secondly, if the suspected ADR was anti-TB drug induced liver injury (ATLI), a nested case-control study would be performed which comprised choosing a matched control and doing a plus questionnaire inquiry. Thirdly, health economical data of ADRs would be collected to analyze financial burdens brought by ADRs and cost-effectiveness of ADRs' treatments. Fourthly, a drop of intravenous blood for each patient was taken and saved in FTA card for DNA banking and genotyping. Finally, the demographic, clinical, environmental, administrative and genetic data would be merged for the comprehensive analysis. Discussion ADACS will give an overview of anti-TB drugs induced ADRs' incidences, risk factors, treatments, prognoses, and clinical, economical and public health impacts for TB patients applying CNTS regimen in China, and provide suggestions for individualized health care and TB control policy.

To compare enzyme-linked immunospot assay (ELISPOT) and tuberculin skin test (TST) and explore their roles in the auxiliary diagnosis of initial pulmonary tuberculosis.Totally 123 patients with initial pulmonary tuberculosis (tuberculosis group) and 102 patients with non-tuberculosis pulmonary disease (control group) were enrolled. The peripheral blood mononuclear cells of all participants were co-cultured with early secretiny antigen target-6/culture filtrate protein-10 fusion protein (ESAT-6/CFP-10), and spot forming cells (SFCs) were enumerated by ELISPOT (ESAT-6/CFP-10-ELISPOT). TST was also performed simultaneously.ESAT-6/CFP-10-ELISPOT showed significantly higher numbers of SFCs after stimulation in tuberculosis group than in control group (P = 0.000). The sensitivity, specificity, positive likelihood ratio, negative likelihood ratio, positive predictive value, and negative predictive value of ESAT-6/CFP-10-ELISPOT were 91.1% (111/123), 81.4% (82/102), 4.60, 0.12, 0.85, and 0.87 respectively, while the above values of TST were 65.6% (59/90), 45.1% (46/102), 1.31, 0.76, 0.51, and 0.60, respectively. The sensitivity and specificity of ESAT-6/CFP-10-ELISPOT were significantly higher than those of TST (all P = 0.000). The number of SFCs were not significantly different between smear-positive tuberculosis subgroup and smear-negative tuberculosis subgroup (P = 0.166). The sensitivities were 91.8% (67/73) and 88.0% (44/50) in these two subgroups, respectively, (P = 0.448).ESAT-6/CFP-10-ELISPOT may be a more accurate approach for the auxiliary diagnosis of initial pulmonary tuberculosis; meanwhile, it offers certain diagnostic evidences for smear-negative tuberculosis. However, its specificity may be affected by latent tuberculosis infection. On the contrary, TST has poor value in the auxiliary diagnosis of initial pulmonary tuberculosis.

To evaluate the effect of interventional therapy with antituberculous drug instillation to the lesions in the treatment of multi-drug resistant pulmonary tuberculosis (MDR-PTB) on conventional therapy.Sixty-one cases of MDR TB were included from January 2001 to October 2002 in five hospitals. Pasiniazide, rifapentine levofloxacin, ethambutol, ethionamide, amikacin and clarithromycin were used as the basic chemotherapy regimen. In addition, M. vaccac and interventional therapy were used, and chemotherapy was continued for a total of 18 months.The sputum negative conversion rate (including sputum smear and culture) was 50.8% (31/61) after 3 months of interventional therapy. The rate increased to 83.6% (51/61) after 18 months of therapy. Chest X-ray showed that, the foci were markedly absorbed in 50.8% (31/61), and the effective rate was 93.4% (57/61) after 3 months of therapy. The foci were markedly absorbed in 78.7% (48/61) after 18 months of treatment. The effective rate was 96.7%. The rate of cavity closure was 21.3% (13/61) after 3 months of interventional therapy and it increased to 49.2% (30/61) after 18 months of treatment. The rate of symptom disappearance was 73.2%-94.4%, including fever, hemoptysis and dyspnea.For the treatment of MDR-TB, interventional therapy is effective in improving sputum negative conversion, lesion absorption and cavity closure.

To establish in vivo and in vitro models of persistent Mycobacterium tuberculosis infection, and therefore to study the persisters in different conditions and periods of chemotherapy, and to explore the relationship between the persisters and chemotherapy.The persisters in the two models were examined by culturing Mycobacterium tuberculosis in oxygen-starved condition and determining the mRNA expression of isocitrate lyase (ICL), alpha-crystallin chaperone (Acr) and 85B through quantitative PCR.The bacteria which could be cultured in oxygen-starved condition were discovered in both models. The mRNA expression of ICL and Acr increased gradually and dramatically after culture for 4 days in the in vitro model, their values being (5.3 +/- 0.9) and (6.4 +/- 1.6) log copy/ml respectively. While the mRNA expression of 85B showed no significant change in oxygen-starved condition, it increased significantly in the standard condition, the values being (6.1 +/- 0.9) log copy/ml at 10th day. In the mice infected with Mycobacterium tuberculosis, the mRNA expression of ICL was detected in 2 and 4 weeks post-infection, and decreased 4 weeks after treatment. The Acr mRNA showed no or very low expression 4 weeks post-infection or 4 weeks after treatment, but it increased significantly 8 and 10 weeks after treatment, with a value of (6.2 +/- 1.7) log copy/ml at 10 weeks, and its expression was still detected 12 weeks after treatment and 4 weeks after the cessation of treatment [(3.0 +/- 1.6) log copy/ml]. The expression of 85B mRNA was high before the treatment [(6.4 +/- 1.1) log copy/ml], and decreased gradually during the therapy.The models of in vitro and in vivo persistence of Mycobacterium tuberculosis were established. ICL and Acr mRNAs were highly expressed, which may be the markers of persisters. Persisters can be detected by culture in oxygen-starved conditions and the measurement of mRNA expression.

To evaluate the curative effect and safety of a long course regimen containing Chinese-made rifabutin as compared to the regimen containing rifapentine in the treatment of multi-drug resistant pulmonary tuberculosis.During 18 month treatment, 130 patients with multi-drug resistant pulmonary tuberculosis were divided into a treatment group (rifabutin, pasiniazide, levofloxacin, ethambutol, ethionamide, amikacin for 3 months, rifabutin, pasiniazide, levofloxacin, ethambutol, ethionamide for 6 months, rifabutin, pasiniazide, levofloxacin, ethambutol for 9 months), and a control group (rifapentine, pasiniazide, levofloxacin, ethambutol, ethionamide, amikacin for 3 months, rifabutin, pasiniazide, levofloxacin, ethambutol, ethionamide for 6 months, rifabutin, pasiniazide, levofloxacin, ethambutol for 9 months) with proportion 1:1 random, and parallel compared method.After intensive phase, the sputum negative conversion rates (smear negative, culture negative) of the treatment group and the control group were 41.54% (27/65) and 35.94% (23/65), chi(2) = 2.42, P0.05, respectively. The remarkable effective rates in chest X-ray of the two groups were all 10.77% (7/65), chi(2) = 0.01, P0.05, and the effective rates were 67.69% (44/65) and 56.92% (37/65), chi(2) = 1.44, P0.05, respectively. At the end of the treatment, the sputum negative conversion rate (smear negative, culture negative) of the treatment group was 75.0% (48/65), and of the control group was 65.08% (41/65), chi(2) = 1.88, P0.05. The remarkable effective rates in chest X-ray of the two groups were 46.15% (30/65) and 44.62% (29/65), chi(2) = 0.02, P0.05, and the effective rates were 76.92% (50/65) and 73.85% (48/65), chi(2) = 0.19, P0.05, respectively. The cavity closure rates were 23.64% (13/55) and 33.33% (17/51), chi(2) = 0.00, P0.05, respectively.Regimens containing rifabutin or rifapentine. are very effective in sputum negative conversion rate, lesion absorption and cavity closing for the treatment of multi-drug resistant pulmonary tuberculosis, with good safety and tolerance.

To evaluate the significance of quantitative analysis of Mycobacterium tuberculosis mRNA in the test of susceptibility of Mycobacterium tuberculosis strains to rifampin.The susceptibility to rifampin of fifty-three clinical isolated strains was test by the percentage of Mycobacterium tuberculosis 85B mRNA copies before and after the use of rifampin, and 1% and 10% were used as the standards for drug resistance, which was compared with the absolute concentration method. Among them, 29 were rifampin resistant strains and 24 rifampin sensitive strains.When the concentration of rifampin was 1 micro g/ml, and 1% and 10% were used as the standards, the accuracy of quantitative analysis of Mycobacterium tuberculosis mRNA with the absolute concentration method was 70% and 81% respectively, and the sensitivity of quantitative analysis of Mycobacterium tuberculosis mRNA was 100% and 100% respectively, while the specificity of quantitative analysis of Mycobacterium tuberculosis mRNA was 33% and 58%. When the concentration of rifampin was 2 micro g/ml, the accuracy of quantitative analysis of mRNA with the absolute concentration method was 85% and 93% respectively, and the sensitivity was 100% and 97% respectively, while the specificity was 67% and 88% respectively. When the concentration of rifampin was 4 micro g/ml, the accuracy was 93% and 93%, and the sensitivity was 93% and 93%, while the specificity was 92% and 92%.Quantitative analysis of Mycobacterium tuberculosis mRNA was a rapid, sensitive method in rifampin resistance screening. We suggest that the critical concentration of rifampin be 2 micro g/ml, and the critical proportion of mRNA copy be 10%.

To study and evaluate the efficacy and safety of recombinant human interleukin-2 (IL-2) in the treatment of pulmonary tuberculosis.Two hundred and nine cases with re-treated Mycobacterium tuberculosis-positive pulmonary tuberculosis were randomly divided into a trial group (106 cases, treated with 3PaZ (TH)L(2)VE(AK) + IL-2/4PaL(2)V) and a control group (103 cases, treated with 3PaZ(TH)L(2)VE(AK)/4PaL(2)V). The efficacy of 203 cases was available for evaluation when the course was completed (trial group 103 cases, control group 100 cases).The sputum smear-negative conversion rates at the 1st and the 2nd month of therapy were 33.3% and 69.4% in the trial group, and 7.2% and 44.9% in the control group (P0.01). At the completion of the therapy, the X-ray resolution rates were 64.1% and 36.0% respectively for the trial and the control groups, the difference being significant (P0.001). There were significant differences in CD(4) T cells, the ratio of CD(4)/CD(8) and NK cells between the two groups (P0.01). The level of soluble interleukin-2 receptor (sIL-2R) was significantly different between the two groups after treatment for 3 months (P0.05). IL-2 associated side effects were rare and mild.As an effective and relatively safe biological agent, IL-2 can be added to the standard chemotherapy for pulmonary tuberculosis.

To study whether quantitative analysis of M. tuberculosis mRNA could be used to assess bacterial viability.The levels of M. tuberculosis mRNA were compared 24, 48, and 72 hours after M. tuberculosis H(37)R(V) was treated with no drug, 1.0 micro g/ml and 10.0 micro g/ml rifampicin, 0.2 micro g/ml and 1.0 micro g/ml isoniazid, 2 micro g/ml and 4 micro g/ml ethambutol, 2 micro g/ml and 4 micro g/ml streptomycin, 1.25 micro g/ml and 5.00 micro g/ml ofloxacin. The levels of mRNA were determined by quantitative PCR.After exposure to isoniazid, rifampicin, and streptomycin for 24 h, the level of M. tuberculosis H(37)R(V) was reduced markedly. Similarly, after exposure to isoniazid, streptomycin, ethambutol, and ofloxacin for 72 h, the level of M. tuberculosis H(37)R(V) was identical, but higher than that after exposure to rifampicin. Exposure of M. tuberculosis H(37)R(V) to rifampicin for 24 h reduced the levels of 85B mRNA to 0.02% of the control; exposure to other drugs for 24 h reduced the levels to 1% approximately 10% of the control, and for 72 h to less than 1% of the control.The levels of mRNA can show the difference between M. tuberculosis H(37)R(V) exposed and unexposed to drugs, and are consistent with colony forming unit. mRNA could be used as a marker for assessing the viability of M. tuberculosis.