Search

Objective To investigate the consistency between thyrotropin receptor antibody (TRAb) and thyroid-stimulating antibody (TSAb) levels in patients with Graves disease (GD). Methods We performed a cross-sectional observational study to recruit eligible patients with GD who visited the outpatient endocrinology clinic for the purpose of evaluating the consistency between their TRAb and TSAb levels. Our cohort included 28 men and 99 women. Results The median levels of TRAb and TSAb were 5.65 IU/L and 3.76 IU/L, respectively, in the enrolled patients with GD. The levels of TRAb (5.03 vs 8.42 IU/L; P = .008) and TSAb (2.69 vs 5.37 IU/L; P = .008) in patients with adequate thyroid regulation were all lower than those in patients with inadequate thyroid regulation. Conclusions Although TRAb is closely related to TSAb, we observed high heterogeneity of TRAb due to relatively low consistency between the levels of the 2 antibodies.

Background Pseudohypoparathyroidism is a rare genetic disease characterized by hypocalcaemia and hyperphosphataemia due to the defect to the guanine nucleotide-binding protein alpha subunit (GNAS) gene. Patients with pseudoparathyroidism type 1a and 1c could manifest Albright’s hereditary osteodystrophy and multiple hormone resistance including gonadotropin and thyroid stimulating hormone. Case presentation Here we report a Chinese man who presented with fatigue, recurrent seizure and Albright’s hereditary osteodystrophy. His genetic study revealed a heterozygote mutation in the GNAS gene [NM_000516.4(GNAS): c2787_2788del (p.Val930AspfsTer12)]. After calcium and calcitriol supplement, his seizures achieved partially remission. Conclusions We report a case of PHP1a or 1c with a novel frameshift mutation in GNAS gene in a patient presenting with AHO, as well as TSH and partial gonadotropin resistance. This mutation in this case has not been reported in literature and adds to the spectrum of genetic mutations related to PHP.

Background This study aimed to investigate the effects of blood sampling after calorie intake on thyroid stimulating hormone (TSH) level, compared with blood sampling in fasting state. Methods This was a prospective, randomized, controlled study. Subjects from the outpatients in the department of endocrinology without evidence of thyroid diseases were included and then randomized into the fasting group, diet intake group, and glucose intake group, respectively. Fasting blood was collected from all subjects at 7:00 am for the measurement of TSH and free thyroxine (FT4) concentrations. Afterwards, the subjects were maintained at fasting state (fasting group), had an intake of the mixed diet with 400 kcal calories (diet intake group), and had an intake of 75 g glucose (glucose intake group), respectively, and blood was collected again 2 h later (9:00 am on the same day) for TSH and FT4 level measurement and comparison. Results A total of 150 subjects were enrolled, of whom 146 met the inclusion criteria, with 48, 48, and 50 in the diet intake group, glucose intake group, and fasting group, respectively. The TSH in the diet intake group was significantly lower at 9:00 am (TSH9am) than the level at 7:00 am (TSH7am) (P 0.001), with a median variation of -0.71 mU/L, and a median variation rate of -32.4%. In the glucose intake group, TSH9am was also significantly lower than TSH7am (P 0.001), with a median variation of -0.73 mU/L, and a median variation rate of -31.5%. For the fasting group, TSH9am decreased slightly but was significantly lower than TSH7am (P 0.001), with a median variation of -0.1 mU/L, and a median variation rate of -5.2%. According to TSH7am measurements, 9 subjects (3 subjects in each group) met the diagnostic criteria of subclinical hypothyroidism. However, according toTSH9am measurements, only 2 patients in the fasting group met the diagnostic criteria of subclinical hypothyroidism. Conclusion Compared with the fasting state, the TSH level at 2 h after the calorie intake was decreased by about 30%, which might influence the diagnosis of subclinical hypothyroidism. Trial registration ChiCTR2100047454 (18/06/2021).

Deep learning is a branch of machine learning, and its methods are now being used to solve all kinds of problems. Deep learning algorithms can learn advanced features from massive data and automatically extract features, which makes deep learning surpass traditional machine learning algorithms. However, as deep learning algorithms rely on large amounts of data and run too slowly, transfer learning arises in response to this disadvantage. Transfer learning allows the use of existing knowledge in the relevant domain to solve a learning problem with only a small number of sample data in the target domain. Combining the two technologies of deep learning and transfer learning, on the one hand, advanced features of data samples can be automatically learned, and on the other hand, it can get rid of the dependence on sample data capacity. In this paper, the electrocardiogram (ECG) signal into spectrogram, and the model is trained with the ImageNet dataset, and then the trained model is transferred, because AlexNet model needs to be fixed image size, so the last pool layer is replaced by a spatial pyramid pooling layer, finally use Softmax classifier for PhysioNet challenge 2017 electrocardiogram data sets are classified, get a 92.84% accuracy and 83.26% F1.

EEG signal classification has been a research hotspot recently. The combination of EEG signal classification with machine learning technology is very popular. Traditional machine leaning methods for EEG signal classification assume that the EEG signals are drawn from the same distribution. However, the assumption is not always satisfied with the practical applications. In practical applications, the training dataset and the testing dataset are from different but related domains. How to make best use of the training dataset knowledge to improve the testing dataset is critical for these circumstances. In this paper, a novel method combining the non-negative matrix factorization technology and the transfer learning (NMF-TL) is proposed for EEG signal classification. Specifically, the shared subspace is extracted from the testing dataset and training dataset using non-negative matrix factorization firstly and then the shared subspace and the original feature space are combined to obtain the final EEG signal classification results. On the one hand, the non-negative matrix factorization can assure to obtain essential information between the testing and the training dataset; on the other hand, the combination of shared subspace and the original feature space can fully use all the signals including the testing and the training dataset. Extensive experiments on Bonn EEG confirmed the effectiveness of the proposed method.

Pulsatile intravenous insulin therapy (PIVIT) is a means of imitating naturally occurring insulin pulses artificially. It is thought to improve carbohydrate metabolism, which can be assessed using the respiratory quotient (RQ). The aim of this present study was to assess the efficacy and safety of PIVIT for the improvement of RQ in Chinese patients with diabetes mellitus (DM). This 12-week, multi-center, prospective, randomized, open-label, parallel-group study involved 110 DM patients (both type 1 and type 2) whose RQ was

Introduction To investigate the effects of pioglitazone (PIO) on insulin resistance and first phase insulin secretion among obese and lean Chinese people with type 2 diabetes mellitus (T2DM). Methods Sixty-eight drug-naïve patients with T2DM were treated with PIO for 16 weeks. Before and after the treatment, insulin sensitivity was evaluated by the euglycemic hyperinsulinemic clamp test. Plasma insulin levels at 0, 3, 5, 7, and 10 min during intravenous glucose tolerance test were determined to calculate the first phase insulin secretion and pancreatic β-cell function. Circulating adiponectin levels were quantified. Results In both the lean and the obese patients with T2DM, the reduction of HbA1c following the PIO treatment was more than 1% (P

Human beings have experienced agricultural, industrial and information society, followed by the era of artificial intelligence. The remarkable features of the era are big data and machine learning. It is an important task to cultivate high-tech innovative talents for colleges and universities in era of artificial intelligence background. So how to combine machine learning methods with training of innovation ability in the teaching process to improve the quality of talent training is a key problem in college teaching. To this end, a new teaching process framework based on machine learning and innovation ability cultivation is proposed, which lays a solid foundation for improving teaching quality.

Telecom customer churn data is not publicly available because involving users' personal privacy. In 2009, the French telecommunications company Orange for knowledge discovery and data mining (KDD) competition provides a telecom customer churn data set KDD Cup 09. In order to solve the high dimensional problem of KDD Cup 09, a new feature reduction method is used to explore the influence of different features on the prediction of classification model. In this paper, a new K- local maximum margin feature extraction algorithm (KLMM) is proposed. Through researching on the diversification subspace partition rules, the corresponding potential field structure is constructed. According to the data source in the dimension of scalability, the intrinsic link between data attributes and classification results is revealed. The extracted features can reduce the dimension of the churn prediction in telecom data. The KLMM method adapts auto selection sigma factor to reflect the anisotropy of features. The potential function is used to assess the weights of attributes and find the potential important weight. Experiments and analysis show that the extracted features by KLMM are more likely to find a classification hyperplane which can separate data points of the different classes.

Semi-supervised classification methods attempt to improve classification performance based on a small amount of labeled data through full use of abundant unlabeled data. Although existing semi-supervised classification methods have exhibited promising results in many applications, they still have drawbacks, including performance degeneration, due to the introduction of unlabeled data and partially false labels in a small amount of labeled data. To circumvent such drawbacks, a new semi-supervised classification method OCHS-SSC through oversampling and a common hidden space is proposed in the paper. The primary characteristics of the proposed method include two aspects. One is that unlabeled data are only used to generate new synthetic data to extend the minimal amount of labeled data. The other is that the final classifier is learned in the extended feature space, which is composed of the original feature space and the common hidden space found between labeled data and the synthetic data instead of the original feature space. Extensive experiments on 23 datasets indicate the effectiveness of the proposed method.

Early detection of Alzheimer's disease is the key to treatment. Neuropsychological testing has the advantages of being non-invasive and low-cost, but the need for manual selection of features and expert diagnosis is not conducive to the popularity of this method. This paper proposes an approach for automatically extracting and selecting features from texts. First, it uses the bag-of-words model of natural language processing technology to extract all the vocabulary features in the texts. Secondly, unlike the manual selection of features by t-test, it uses the genetic algorithm to select lexical features automatically. We tested the new approach with the DementiaBank database. Its classification accuracy for Alzheimer's disease is 79%, close to the best value of the hand-crafted-feature-based method. The new approach also has the ability to process data quickly and automatically, which can greatly help clinicians improve their work.