Early autism diagnosis based on path signature and Siamese unsupervised feature compressor.

Autism spectrum disorder has been emerging as a growing public health threat. Early diagnosis of autism spectrum disorder is crucial for timely, effective intervention and treatment. However, conventional diagnosis methods based on communications and behavioral patterns are unreliable for children younger than 2 years of age. Given evidences of neurodevelopmental abnormalities in autism spectrum disorder infants, we resort to a novel deep learning-based method to extract key features from the inherently scarce, class-imbalanced, and heterogeneous structural MR images for early autism diagnosis. Specifically, we propose a Siamese verification framework to extend the scarce data, and an unsupervised compressor to alleviate data imbalance by extracting key features. We also proposed weight constraints to cope with sample heterogeneity by giving different samples different voting weights during validation, and used Path Signature to unravel meaningful developmental features from the two-time point data longitudinally. We further extracted machine learning focused brain regions for autism diagnosis. Extensive experiments have shown that our method performed well under practical scenarios, transcending existing machine learning methods and providing anatomical insights for autism early diagnosis.

Identification of prognostic factors and construction of nomogram to predict cancer-specific survival for patients with ovarian granulosa cell tumors.

BACKGROUND: Ovarian granulosa cell tumors (OGCTs) feature low incidence, indolent growth and late recurrence. Treatment for recurrent OGCTs is challenging. METHODS: The present study was designed to explore the prognostic factors and establish a nomogram to predict cancer-specific survival (CSS) for OGCTs patients. Enrolled in the study were 1459 eligible patients in the Surveillance, Epidemiology, and End Results (SEER) database, who were randomized to the training (n = 1021) or testing set (n = 438) at a ratio of 7:3. Univariate and multivariate Cox regression analyses were employed to screen the prognostic factors. The predictors were determined by using the Least absolute shrinkage and selection operator (LASSO) regression analysis. The model was constructed via the Cox proportional hazards risk regression analysis. The performance and clinical value of the nomograms was assessed with C-index, calibration plots, and decision curve analysis. RESULTS: Age, pTNM stage, tumor size, surgery of the primary tumor, surgery of regional lymph nodes (LNs), residual disease after surgery, and chemotherapy were considered as significant predictive factors for CSS in OGCTs patients. After screening, the prognostic factors except surgery of regional LNs and chemotherapy were employed to build the nomogram. With desirable discrimination and calibration, the nomogram was more powerful in predicting CSS than the American Joint Committee on Cancer staging system in clinical use. CONCLUSION: This novel prognostic nomogram, which comprises a stationary nomogram and a web-based calculator, offers convenience for clinicians in personalized decision-making including optimal treatment plans and prognosis assessments for OGCTs patients.

In situ growth of Prussian blue analogue-derived Fe-doped NiS on Ni(OH)2 for efficient hydrogen evolution reaction.

The energy industry is placing more and more emphasis on the need for effective and affordable electrocatalysts for hydrogen evolution reactions (HER). In this work, an iron-doped NiS/Ni(OH)2/CC composite material was synthesized by simple hydrothermal sulfurization processes of bimetallic Prussian blue analogue (PBAs) precursors grown in situ on three-dimensional (3D) Ni(OH)2 nanosheets. The overpotential can be 103 mV to attain current densities of 10 mA cm-2. The excellent catalytic activity of Fe-NiS/Ni(OH)2/CC is because of the unique 3D structure and the uniform doping of iron caused by the in situ growth of PBA, as well as the high conductivity of the self-supported electrode carbon cloth.

One-Dimensional CoMoP Nanostructures as Bifunctional Electrodes for Overall Water Splitting.

As high-quality substitutes for conventional catalysts, the bifunctional catalytic properties of the coating of transition-metal-based materials are pivotal for improving water-splitting efficiency. Herein, cobalt-molybdenum bimetallic phosphide nanofibers (CoMoP NFs) were synthesized via a series of facile strategies, which are divided into pyrolysis electrospun PAN and metal salts, to obtain one-dimensional morphology and a gas-solid phosphating precursor. The obtained CoMoP NFs catalyst has superior catalytic activity performance in 1M KOH. Serving as an oxygen evolution reaction (OER) catalyst, the electrode of the CoMoP NFs affords different kinds of current densities at 50 mA cm-2 and 100 mA cm-2, with low overpotentials of 362 and 391 mV, respectively. In addition, the hydrogen evolution reaction (HER) performance of the CoMoP NFs mainly shows when under a low overpotential of 126 mV, which can deliver a current density of 10 mA cm-2. In order to further detect the stability of the catalyst, we used multiple cyclic voltammetry and chronopotentiometry tests for OERs and HERs, which maintain performance and carry a current density of 10 mA cm-2 for longer. As an integrated high-performance bifunctional electrode for overall water splitting, the CoMoP NFs only require 1.75 V@10 mA cm-2 for 40 h. This work highlights a facile method to create an electrocatalyst with fiber nanostructures which possesses excellent activity as an alkaline electrolyte.

One-dimensional Ni2P/Mn2O3 nanostructures with enhanced oxygen evolution reaction activity.

Nickel-based transition metal phosphides have been widely flourishing as oxygen evolution reaction (OER) electrocatalysts. In this study, a new catalyst Ni2P/Mn2O3 nanofibers with the advantages of ultra-high electrochemical active area were successfully synthesized. We explored effective strategies that are constructing one-dimensional nanostructures and composite oxides to promote the electrocatalytic performance of transition metal phosphides. The Ni2P/Mn2O3 nanofibers only require a low overpotential of 280 mV to deliver a current density of 10 mA cm-2 in 1 M KOH for oxygen production. As a result, it is worthily mentioned that the activity of Ni2P/Mn2O3 nanofibers is virtually unchanged after 2000 cycles of voltammetry measurements with a stable nanostructure. This research provides a feasible solution for the design and realization of nanostructured electrocatalysts for the enhancing performance of the OER process.

One-dimensional CoP/MnO hollow nanostructures with enhanced oxygen evolution reaction activity.

Cobalt-based transition metal phosphides are flourishing as OER electrocatalysts. In this study, the CoP/MnO hollow nanofibers with the advantages of a more extensive contact interface were successfully synthesized. We found that the construction of hollow nanostructures and the composite of oxides are effective strategies to optimize the OER catalytic performance of transition metal phosphides. The template of the precursor can adjust the hollow nanostructure and keep it stable during the phosphating process. It is worth noting that the CoP/MnO composite material only needs an overpotential of 230 mV at a current density of 10 mA cm-2. In addition, it maintains the overpotential 263.5 mV after 5000 cycles of voltammetry measurements. In short, this research provides a simple solution for the design and realization of nanostructured electrocatalysts with excellent electrochemical performance.

SIAMESE VERIFICATION FRAMEWORK FOR AUTISM IDENTIFICATION DURING INFANCY USING CORTICAL PATH SIGNATURE FEATURES.

Autism spectrum disorder (ASD) is a complex neurodevelopmental disability, which is lack of biologic diagnostic markers. Therefore, exploring the ASD Identification directly from brain imaging data has been an important topic. In this work, we propose the Siamese verification model to identify ASD using 6 and 12 months cortical features. Rather than directly classifying a testing subject is ASD or not, we determine whether it has the same or different label with the reference subject who has been successfully diagnosed. Then, based on the comparison to all the reference subjects, we can predict the label of the testing subject. The advantage of modeling the classification problem as a verification framework is that it can greatly enlarge the training data size and enable us to train a more accurate and reliable model in an end-to-end manner. In addition, to further improve the classification performance, we introduce the path signature (PS) features, which can capture the dynamic longitudinal information of the brain development for the ASD Identification. Experiments showed that our proposed method reaches the best result, i.e., 87% accuracy, 83% sensitivity and 90% specificity comparing to the state-of-the-art methods.