Ultrasound segmentation analysis via distinct and completed anatomical borders.

PURPOSE: Segmenting ultrasound images is important for precise area and/or volume calculations, ensuring reliable diagnosis and effective treatment evaluation for diseases. Recently, many segmentation methods have been proposed and shown impressive performance. However, currently, there is no deeper understanding of how networks segment target regions or how they define the boundaries. In this paper, we present a new approach that analyzes ultrasound segmentation networks in terms of learned borders because border delimitation is challenging in ultrasound. METHODS: We propose a way to split the boundaries for ultrasound images into distinct and completed. By exploiting the Grad-CAM of the split borders, we analyze the areas each network pays attention to. Further, we calculate the ratio of correct predictions for distinct and completed borders. We conducted experiments on an in-house leg ultrasound dataset (LEG-3D-US) as well as on two additional public datasets of thyroid, nerves, and one private for prostate. RESULTS: Quantitatively, the networks exhibit around 10% improvement in handling completed borders compared to distinct borders. Similar to doctors, the network struggles to define the borders in less visible areas. Additionally, the Seg-Grad-CAM analysis underscores how completion uses distinct borders and landmarks, while distinct focuses mainly on the shiny structures. We also observe variations depending on the attention mechanism of each architecture. CONCLUSION: In this work, we highlight the importance of studying ultrasound borders differently than other modalities such as MRI or CT. We split the borders into distinct and completed, similar to clinicians, and show the quality of the network-learned information for these two types of borders. Additionally, we open-source a 3D leg ultrasound dataset to the community https://github.com/Al3xand1a/segmentation-border-analysis .

Advancing Adversarial Training by Injecting Booster Signal.

Recent works have demonstrated that deep neural networks (DNNs) are highly vulnerable to adversarial attacks. To defend against adversarial attacks, many defense strategies have been proposed, among which adversarial training (AT) has been demonstrated to be the most effective strategy. However, it has been known that AT sometimes hurts natural accuracy. Then, many works focus on optimizing model parameters to handle the problem. Different from the previous approaches, in this article, we propose a new approach to improve the adversarial robustness using an external signal rather than model parameters. In the proposed method, a well-optimized universal external signal called a booster signal is injected into the outside of the image which does not overlap with the original content. Then, it boosts both adversarial robustness and natural accuracy. The booster signal is optimized in parallel to model parameters step by step collaboratively. Experimental results show that the booster signal can improve both the natural and robust accuracies over the recent state-of-the-art AT methods. Also, optimizing the booster signal is general and flexible enough to be adopted on any existing AT methods.

High-Performance γ-Al2O3 Multichannel Tube-Type Tight Ultrafiltration Membrane Using a Modified Sol-Gel Method.

We introduced a modified sol-gel method using polyvinyl alcohol (PVA) as an additive to improve the permeability of γ-Al2O3 membranes by minimizing the thickness of the selective layer and maximizing the porosity. First, the analysis revealed that the thickness of γ-Al2O3 decreased as the concentration of PVA increased in the boehmite sol. Second, the properties of the γ-Al2O3 mesoporous membranes were greatly influenced by the modified route (method B) compared to the conventional route (method A). The results showed that the porosity and surface area of the γ-Al2O3 membrane increased, and the tortuosity decreased considerably using method B. This effect was attributed to the adsorption of PVA molecules on the surface of the boehmite particles, which depended on the synthesis route. The experimentally determined pure water permeability trend and the Hagen-Poiseuille mathematical model confirmed that the modified method improved the performance of the γ-Al2O3 membrane. Finally, the γ-Al2O3 membrane fabricated via a modified sol-gel method with a pore size of 2.7 nm (MWCO = 5300 Da) exhibited a pure water permeability of over 18 LMH/bar, which is three times higher than that of the γ-Al2O3 membrane prepared using the conventional method.

Defending Person Detection Against Adversarial Patch Attack by Using Universal Defensive Frame.

Person detection has attracted great attention in the computer vision area and is an imperative element in human-centric computer vision. Although the predictive performances of person detection networks have been improved dramatically, they are vulnerable to adversarial patch attacks. Changing the pixels in a restricted region can easily fool the person detection network in safety-critical applications such as autonomous driving and security systems. Despite the necessity of countering adversarial patch attacks, very few efforts have been dedicated to defending person detection against adversarial patch attack. In this paper, we propose a novel defense strategy that defends against an adversarial patch attack by optimizing a defensive frame for person detection. The defensive frame alleviates the effect of the adversarial patch while maintaining person detection performance with clean person. The proposed defensive frame in the person detection is generated with a competitive learning algorithm which makes an iterative competition between detection threatening module and detection shielding module in person detection. Comprehensive experimental results demonstrate that the proposed method effectively defends person detection against adversarial patch attacks.

Associations between Cell-Free Mitochondrial DNA and Inflammation, and Their Clinical Implications for Patients on Hemodialysis: A Prospective Multicenter Cohort Study.

BACKGROUND: Cell-free mitochondrial DNA (cf-mtDNA) has recently been in the spotlight as an endogenously produced danger molecule that can potentially elicit inflammation. However, its clinical and prognostic implications are uncertain in patients undergoing hemodialysis. METHODS: We examined the association of baseline cf-mtDNA categorized as tertiles with health-related quality of life (HRQOL), inflammatory cytokines, and mortality in a multicenter prospective cohort of 334 patients on hemodialysis. To better understand cf-mtDNA-mediated inflammation, we measured cytokine production after in vitro stimulation of bone marrow-derived macrophages (BMDMs) with mtDNA. RESULTS: The higher cf-mtDNA tertile had a longer dialysis vintage, a greater comorbidity burden, and increased levels of inflammatory markers, including high-sensitivity-C-reactive protein, tumor necrosis factor-alpha, CXCL16, and osteoprotegerin. In particular, mtDNA augmented inflammatory cytokine release from BMDMs by lipopolysaccharide, the levels of which are reported to be increased in hemodialysis patients. Although the patients with higher levels of cf-mtDNA generally had lower (poorer) scores for HRQOL, cf-mtDNA was not associated with all-cause mortality in hemodialysis patients. CONCLUSION: cf-mtDNA was correlated with poor clinical status and modestly associated with impaired quality of life in patients on hemodialysis. In proinflammatory milieu in end-stage renal disease, these associations may be attributed to the boosting effects of cf-mtDNA on inflammation.

Gait speed and handgrip strength as predictors of all-cause mortality and cardiovascular events in hemodialysis patients.

BACKGROUND: Low physical performance in patients undergoing maintenance hemodialysis is associated with a high mortality rate. We investigated the clinical relevance of gait speed and handgrip strength, the two most commonly used methods of assessing physical performance. METHODS: We obtained data regarding gait speed and handgrip strength from 277 hemodialysis patients and evaluated their relationships with baseline parameters, mental health, plasma inflammatory markers, and major adverse clinical outcomes. Low physical performance was defined by the recommendations suggested by the Asian Working Group on Sarcopenia. RESULTS: The prevalence of low gait speed and handgrip strength was 28.2 and 44.8%, respectively. Old age, low serum albumin levels, high comorbidity index score, and impaired cognitive functions were associated with low physical performance. Patients with isolated low gait speed exhibited a general trend for worse quality of life than those with isolated low handgrip strength. Gait speed and handgrip strength showed very weak correlations with different determining factors (older age, the presence of diabetes, and lower serum albumin level for low gait speed, and lower body mass index and the presence of previous cardiovascular events for low handgrip strength). Patients with low gait speed and handgrip strength had elevated levels of plasma endocan and matrix metalloproteinase-7 and the highest risks for all-cause mortality and cardiovascular events among the groups (adjusted hazard ratio of 2.72, p = 0.024). Elderly patients with low gait speed and handgrip strength were at the highest risk for poor clinical outcomes. CONCLUSION: Gait speed and handgrip strength reflected distinctive aspects of patient characteristics and the use of both factors improved the prediction of adverse clinical outcomes in hemodialysis patients. Gait speed seems to be a better indicator of poor patient outcomes than is handgrip strength.

High dose haemodialysis and haemodiafiltration parameters and the relationship with advanced vascular calcification.

BACKGROUND: Vascular calcification (VC) is a risk factor for cardiovascular disease in end-stage renal disease (ESRD) patients undergoing maintenance haemodialysis (MHD). However, evidence is still insufficient about the association between dialysis parameters and VC. Thus, this study was to evaluate association of dialysis parameters with VC. METHODS: We enrolled 297 ESRD patients undergoing MHD at six distinct centers in Korea. Study participants were categorized into 3 groups by the scoring system of abdominal aortic calcification based on lateral lumbar radiography (no VC group: 0, mild VC group: 1-7 and advanced VC group: 8-24). We compared the features of dialysis parameters according to the severity of VC. Multivariate logistic regression analysis was used to calculate adjusted odd ratios (ORs) and 95% confidence interval (CI) for mild and advanced VC in each haemodialysis parameter (adjusted OR [95% CI]). RESULTS: Pooled Kt/V (spKt/V), equilibrated Kt/V (eKt/V), standard Kt/V (stdKt/V) and the proportion of haemodiafiltration were increased along with the severity of VC. Multivariate regression analysis indicated that advanced VC was positively associated with spKt/V (5.27 [1.51-18.41]), eKt/V (6.16 [1.45-26.10]), stdKt/V (10.67 [1.74-65.52]) and haemodiafiltration (3.27 [1.74 to 6.16]). CONCLUSION: High dose dialysis and haemodiafiltration were significantly associated with advanced VC.

Effects of online hemodiafiltration on anemia and nutritional status in chronic hemodialysis patients.

BACKGROUND: Online hemodiafiltration (OL-HDF) offers considerable advantages in clearance of molecules of various sizes. However, evidence of clinical effects of OL-HDF is scarce in Korea. In this study, we investigated changes in laboratory values over more than 12 months after switching to OL-HDF. METHODS: Adult patients with end-stage renal disease undergoing hemodialysis (HD) were prospectively enrolled in a K-cohort (CRIS no. KCT0003281) from 6 tertiary hospitals in South Korea. We recruited 435 patients, 339 of whom were on HD at enrollment. One hundred eighty-two patients were followed for more than 24 months. Among them, 44 were switched to OL-HDF for more than 12 months without conversion to HD. We used a paired t test to compare baseline and 24-month follow-up results. RESULTS: The mean age of the subjects was 61.2 ± 12.2 years, and 62.6% were male. The baseline hemoglobin level was not significantly different between HD and OL-HDF group (10.61 ± 1.15 vs. 10.46 ± 1.03 g/dL, P = 0.437). However, the baseline serum protein and albumin levels were significantly lower in the OL-HDF group (6.82 ± 0.49 vs. 6.59 ± 0.48 g/dL, P = 0.006; 3.93 ± 0.28 vs. 3.73 ± 0.29 g/dL, P < 0.001). In patients switched to OL-HDF, levels of hemoglobin and serum albumin significantly increased (10.46 ± 1.03 vs. 11.08 ± 0.82 g/dL, P = 0.001; 3.73 ± 0.29 vs. 3.87 ± 0.30 g/dL, P = 0.001). The normalized protein catabolic rate decreased after 24 months, but the change was not significant (1.07 ± 0.25 vs. 1.03 ± 0.21 g/kg/day, P = 0.433). Although the dose of erythropoiesis-stimulating agent was lower in patients who converted to HDF, it was not significantly different (-115.7 ± 189.7 vs. -170.5 ± 257.1 P = 0.206). CONCLUSION: OL-HDF treatment over more than 12 months was associated with no harmful effects on anemia and nutritional status.

Circulating PCSK9 Level and Risk of Cardiovascular Events and Death in Hemodialysis Patients.

Proprotein convertase subtilisin/kexin type 9 (PCSK9) is a promising new target for the prevention of cardiovascular (CV) events. However, the clinical significance of circulating PCSK9 is unclear in hemodialysis (HD) patients. A total of 353 HD patients were prospectively enrolled from June 2016 to August 2019 in a K-cohort. Plasma PCSK9 level was measured at the time of study enrollment. The primary endpoint was defined as a composite of CV event and death. Plasma PCSK9 level was positively correlated with total cholesterol level in patients with statin treatment. Multivariate linear regression analysis revealed that baseline serum glucose, albumin, total cholesterol, and statin treatment were independent determinants of circulating PCSK9 levels. Cumulative rates of composite and CV events were significantly higher in patients with tertile 3 PCSK9 (p = 0.017 and p = 0.010, respectively). In multivariate Cox-regression analysis, PCSK9 tertile 3 was associated with a 1.97-fold risk of composite events (95% CI, 1.13-3.45), and it was associated with a 2.31-fold risk of CV events (95% CI, 1.17-4.59). In conclusion, a higher circulating PCSK9 level was independently associated with incident CV events and death in HD patients. These results suggest the importance of future studies regarding the effect of PCSK9 inhibition.

Endometrium segmentation on transvaginal ultrasound image using key-point discriminator.

PURPOSE: Transvaginal ultrasound imaging provides useful information for diagnosing endometrial pathologies and reproductive health. Endometrium segmentation in transvaginal ultrasound (TVUS) images is very challenging due to ambiguous boundaries and heterogeneous textures. In this study, we developed a new segmentation framework which provides robust segmentation against ambiguous boundaries and heterogeneous textures of TVUS images. METHODS: To achieve endometrium segmentation from TVUS images, we propose a new segmentation framework with a discriminator guided by four key points of the endometrium (namely, the endometrium cavity tip, the internal os of the cervix, and the two thickest points between the two basal layers on the anterior and posterior uterine walls). The key points of the endometrium are defined as meaningful points that are related to the characteristics of the endometrial morphology, namely the length and thickness of the endometrium. In the proposed segmentation framework, the key-point discriminator distinguishes a predicted segmentation map from a ground-truth segmentation map according to the key-point maps. Meanwhile, the endometrium segmentation network predicts accurate segmentation results that the key-point discriminator cannot discriminate. In this adversarial way, the key-point information containing endometrial morphology characteristics is effectively incorporated in the segmentation network. The segmentation network can accurately find the segmentation boundary while the key-point discriminator learns the shape distribution of the endometrium. Moreover, the endometrium segmentation can be robust to the heterogeneous texture of the endometrium. We conducted an experiment on a TVUS dataset that contained 3,372 sagittal TVUS images and the corresponding key points. The dataset was collected by three hospitals (Ewha Woman's University School of Medicine, Asan Medical Center, and Yonsei University College of Medicine) with the approval of the three hospitals' Institutional Review Board. For verification, fivefold cross-validation was performed. RESULT: The proposed key-point discriminator improved the performance of the endometrium segmentation, achieving 82.67 % for the Dice coefficient and 70.46% for the Jaccard coefficient. In comparison, on the TVUS images UNet, showed 58.69 % for the Dice coefficient and 41.59 % for the Jaccard coefficient. The qualitative performance of the endometrium segmentation was also improved over the conventional deep learning segmentation networks. Our experimental results indicated robust segmentation by the proposed method on TVUS images with heterogeneous texture and unclear boundary. In addition, the effect of the key-point discriminator was verified by an ablation study. CONCLUSION: We proposed a key-point discriminator to train a segmentation network for robust segmentation of the endometrium with TVUS images. By utilizing the key-point information, the proposed method showed more reliable and accurate segmentation performance and outperformed the conventional segmentation networks both in qualitative and quantitative comparisons.

Micro-/nano-sized delivery systems of ginsenosides for improved systemic bioavailability.

Ginsenosides, dammarane-type triterpene saponins obtained from ginseng, have been used as a natural medicine for many years in the Orient due to their various pharmacological activities. However, the therapeutic potential of ginsenosides has been largely limited by the low bioavailability of the natural products caused mainly by low aqueous solubility, poor biomembrane permeability, instability in the gastrointestinal tract, and extensive metabolism in the body. To enhance the bioavailability of ginsenosides, diverse micro-/nano-sized delivery systems such as emulsions, polymeric particles, and vesicular systems have been investigated. The delivery systems improved the bioavailability of ginsenosides by enhancing solubility, permeability, and stability of the natural products. This mini-review aims to provide comprehensive information on the micro-/nano-sized delivery systems for increasing the bioavailability of ginsenosides, which may be helpful for designing better delivery systems to maximize the versatile therapeutic potential of ginsenosides.

Glycerol Steam Reforming Over Ni-Fe-Ce/Al2O3 Catalyst: Effect of Cerium.

In this work, hydrogen production from glycerol by steam reforming was studied using Ni-metal oxide catalysts. Ni-based catalyst becomes deactivated during steam reforming reactions because of coke deposits and sintering. Therefore, the aim of this study was to reduce carbon deposits and sintering on the catalyst surface by adding a promoter. Ni-metal oxide catalysts supported on Al2O3 were prepared via impregnation method, and the calcined catalyst was reduced under H2 flow for 2 h prior to the reaction. The characteristics of the catalysts were examined by XRD, TPR, TGA, and SEM. The Ni-Fe-Ce/Al2O3 catalyst, which contained less than 2 wt% Ce, showed the highest hydrogen selectivity and glycerol conversion. Further analysis of the catalysts revealed that the Ni-Fe-Ce/Al2O3 catalyst required a lower reduction temperature and produced minimum carbon deposit.

Exploration of Online Culture Through Network Analysis of Wikipedia.

Understanding online culture is becoming crucial in the global and connected world. Contrary to conventional attitudinal surveys used in cultural research, this study uses the approach of directly observing culture-specific behavior that emerges from online collaboration on the Internet. The editing data of Wikipedia were analyzed in 12 languages. Distinctive cultural dimensions were identified, including collectivism, extraversion, boldness, and egalitarianism. Using network analysis, the language-framed cultural factors were extracted as an emergent phenomenon in the virtual world.

Circulating Endothelial Progenitor Cell Levels Predict Cardiovascular Events in End-Stage Renal Disease Patients on Maintenance Hemodialysis.

BACKGROUND: The number of circulating endothelial progenitor cells (EPCs) has been identified as a surrogate biologic marker for vascular function and cumulative cardiovascular (CV) risk in the general population. Patients with end-stage renal disease (ESRD) on hemodialysis (HD) have markedly decreased EPC counts and function. We hypothesized that the number of circulating EPCs predicts death from all causes and CV events in patients with ESRD on HD. METHODS: We quantified the EPCs in blood samples from 70 patients with ESRD on HD. Circulating EPCs were counted by flow cytometry as the number of CD45(low)CD34(+)VEGFR2(+) cells. Death from all causes and CV events served as outcome variables over a median follow-up period of 20 months. RESULTS: It has been postulated that the number of circulating EPCs at baseline ranged from 1 to 350 cells/200 µl, with a mean of ± standard deviation (SD) of 26.0 ± 48.2 cells/200 µl. The median, lowest and highest tertiles of EPC counts were 11.0, 9.0, and 17.0 cells/200 µl, respectively. Patients with the lowest tertile EPC counts had significantly higher rates of CV events, but mortality was similar between the two groups. After adjusting for these risk factors, HbA1c and the lowest tertile EPC count remained as independent predictors of CV events. A cutoff value of 9.5 cells/200 µl maximized the power of the EPC count to predict future CV events as determined by ROC curve analysis. CONCLUSIONS: Reduced circulating EPC counts independently predicted CV events in 70 patients with ESRD on maintenance HD. Circulating EPCs may play a role in vascular repair, thereby affecting the clinical course of CV events.

Enhanced bioethanol production from yellow poplar by deacetylation and oxalic acid pretreatment without detoxification.

In order to produce ethanol from yellow poplar, deacetylation was performed using sodium hydroxide (NaOH). Optimal deacetylation conditions were determined by a response surface methodology. The highest acetic acid concentration obtained was 7.06 g/L when deacetylation was performed at 60 °C for 80 min with 0.8% NaOH. Acetic acid was recovered by electrodialysis from the deacetylated hydrolysate. The oxalic acid pretreatment of deacetylated biomass was carried out and the hydrolysate directly used for ethanol production without detoxification. Ethanol yields ranged from 0.34 to 0.47 g/g and the highest ethanol yield was obtained when pretreatment was carried out at 150 °C with 50 mM oxalic acid. The highest ethanol concentration obtained from pretreated biomass was 27.21 g/L at 170 °C, using a 50 mM of oxalic acid for the simultaneous saccharification and fermentation (SSF). Overall, 20.31 g of ethanol was obtained by hydrolysate and SSF from 100 g of deacetylated yellow poplar.

Febuxostat ameliorates diabetic renal injury in a streptozotocin-induced diabetic rat model.

BACKGROUND: Oxidative stress and inflammation are known to play central roles in the development of diabetic nephropathy (DN). Febuxostat is a novel non-purine xanthine oxidase (XO)-specific inhibitor developed to treat hyperuricemia. In this study, we investigated whether febuxostat could ameliorate DN via renoprotective mechanisms such as alleviation of oxidative stress and anti-inflammatory actions. METHODS: Male Sprague-Dawley rats were divided into three groups: a normal group, a diabetes group (DM group), and a febuxostat-treated diabetes group (DM+Fx group). We administered 5 mg/kg of febuxostat to experimental rats for 7 weeks and evaluated clinical and biochemical parameters and XO and xanthine dehydrogenase (XDH) activity in hepatic tissue. The degree of oxidative stress and extent of inflammation were evaluated from urine samples and renal tissue collected from each group. RESULTS: Diabetic rats (DM and DM+Fx groups) had higher blood glucose and kidney weight relative to body weight than normal rats. Albuminuria was significantly reduced in febuxostat-treated diabetic rats compared with untreated diabetic rats. Quantitative analysis showed that hepatic XO and XDH activities were higher in the DM groups, but decreased after treatment with febuxostat. Urinary 8-OHdG concentrations and renal cortical nitrotyrosine also indicated reduced oxidative stress in the DM+Fx group relative to the DM group. The number of ED-1-stained cells in the glomerulus and tubule of diabetic renal tissue decreased in febuxostat-treated diabetic rats relative to that of non-treated diabetic rats. Diabetic rats also expressed higher transcript levels of inflammatory genes (E-selectin and VCAM-1), an inflammation-induced enzyme (COX-2), and inflammatory mediators (ED-1 and NF-κB) than control rats; expression of these genes was significantly reduced by treatment with febuxostat. CONCLUSIONS: Febuxostat prevents diabetic renal injury such as albuminuria. This renoprotective effect appears to be due to attenuation of the inflammatory and oxidative effects of diabetes-induced renal damage through inhibition of XO and XDH activities.

An integrated detoxification process with electrodialysis and adsorption from the hemicellulose hydrolysates of yellow poplars.

An integrated detoxification process with electrodialysis (ED) followed by adsorption was performed to remove fermentation inhibitors from hemicellulose hydrolysates. The hydrolysates were prepared by oxalic acid pretreatment of yellow poplars at different temperatures. Of fermentation inhibitors, acetic acid showed high removal efficiency of about 90% and high transport rate during the ED process without membrane fouling. The integration of the detoxification processes increased up to the ethanol yield of 0.33g/g sugar, the ethanol production of about 9g/L, and the productivity of 0.12g/Lh, while the fermentation of non-detoxified hydrolysates did not produce bioethanol. The influence of inhibitor concentration on the fermentability showed that HMF had the highest inhibition effect. The results clearly showed that an integrated detoxification process with ED followed by adsorption removed fermentation inhibitors with high efficiency and increased the fermentability of the oxalic acid pretreated hemicellulose hydrolysates.

Improvement of the fermentability of oxalic acid hydrolysates by detoxification using electrodialysis and adsorption.

A two-step detoxification process consisting of electrodialysis and adsorption was performed to improve the fermentability of oxalic acid hydrolysates. The constituents of the hydrolysate differed significantly between mixed hardwood and softwood. Acetic acid and furfural concentrations were high in the mixed hardwood, whereas 5-hydroxymethylfurfural (HMF) concentration was relatively low compared with that of the mixed softwood. The removal efficiency of acetic acid reached 100% by electrodialysis (ED) process in both hydrolysates, while those of furfural and HMF showed very low, due to non-ionizable properties. Most of the remaining inhibitors were removed by XAD-4 resin. In the mixed hardwood hydrolysate without removal of the inhibitors, ethanol fermentation was not completed. Meanwhile, both ED-treated hydrolysates successfully produced ethanol with 0.08 and 0.15 g/Lh ethanol productivity, respectively. The maximum ethanol productivity was attained after fermentation with 0.27 and 0.35 g/Lh of detoxified hydrolysates, which were treated by ED, followed by XAD-4 resin.

Communication, opponents, and clan performance in online games: a social network approach.

Online gamers form clans voluntarily to play together and to discuss their real and virtual lives. Although these clans have diverse goals, they seek to increase their rank in the game community by winning more battles. Communications among clan members and battles with other clans may influence the performance of a clan. In this study, we compared the effects of communication structure inside a clan, and battle networks among clans, with the performance of the clans. We collected battle histories, posts, and comments on clan pages from a Korean online game, and measured social network indices for communication and battle networks. Communication structures in terms of density and group degree centralization index had no significant association with clan performance. However, the centrality of clans in the battle network was positively related to the performance of the clan. If a clan had many battle opponents, the performance of the clan improved.