The existing epilepsy seizure detection algorithms have problems such as overfitting and poor generalization ability due to high reliance on manual labeling of electroencephalogram's data and data imbalance between seizure and interictal periods. An unsupervised learning detection method for epileptic seizure that jointed graph attention network (GAT) and Transformer framework (GAT-T) was proposed. In this method, channel correlations were adaptively learned by GAT encoder. Temporal information was captured by one-dimensional convolution decoder. Combining outputs of the two mentioned above, predicted values for electroencephalogram were generated. The collective anomaly score was calculated and the detection threshold was determined. The results demonstrated that GAT-T achieved the average performance exceeding 90% (or 99%) with a 0.25 s (or 2 s) time segment length, which could effectively detect epileptic seizures. Moreover, the channel association probability matrix was expected to assist clinicians in the initial screening of the epileptogenic zone, and ablation experiments also reflected the significance of each module in GAT-T. This study may assist clinicians in making more accurate diagnostic and therapeutic decisions for epilepsy patients.
Humans ; Electroencephalography/methods* ; Epilepsy/physiopathology* ; Algorithms ; Seizures/physiopathology* ; Neural Networks, Computer ; Signal Processing, Computer-Assisted

OBJECTIVES: We propose a multimodal model integrating social media text and image data for automated assessment of psychological stress in college students to support the development of intelligent mental health services in higher education institutions. METHODS: Based on deep learning technology, we designed an evaluation framework comprising a text sentiment modeling module, an image sentiment modeling module, and a multimodal fusion prediction module. Text sentiment features were extracted using Bi-LSTM, and image semantic cues were extracted via U-Net. A feature concatenation strategy was used to enable cross-modal semantic collaboration to achieve automatic identification of 3 psychological stress levels: mild, moderate, and severe. We constructed a multimodal annotated dataset using social platform data from 1577 students across multiple universities in Guangdong Province. After data cleaning, 252 samples were randomly selected for model training and testing. RESULTS: In the 3-classification task, the model demonstrated outstanding performance on the test set, and achieved an accuracy of 92.86% and an F1 score of 0.9276, exhibiting excellent stability and consistency. Confusion matrix analysis further revealed the model's ability to effectively distinguish between different pressure levels. CONCLUSIONS The multimodal psychological stress assessment model developed in this study effectively integrates unstructured social behavior data to enhance the scientific rigor and practical applicability of psychological state recognition, and thus provides support for developing intelligent psychological service systems.
Humans ; Stress, Psychological/diagnosis* ; Students/psychology* ; Universities ; Social Media ; Deep Learning

Objective:To explore the value of the Nomogram model established by CT radiomics combined with clinical indicators for prediction of the severity of early acute pancreatitis (AP).Methods:From January 2016 to March 2023, the AP patients in the Second Affiliated Hospital of Soochow University were retrospectively collected. According to the revised Atlanta classification and definition of acute pancreatitis in 2012, all patients were divided into the severe group and the non-severe group. All patients were first diagnosed, and abdominal CT plain scan and enhanced scan were completed within 1 week. Patients were randomly (random number) divided into training and validation groups at a ratio of 7:3. The pancreatic parenchyma was delineated as the region of interest on each phase CT images, and the radiomics features were extracted by python software. LASSO regression and 10-fold cross-validation were used to reduce the dimension and select the optimal features to establish the radiomics signature. Multivariate Logistic regression was used to select the independent predictors of severe acute pancreatitis (SAP), and a clinical model was established. A Nomogram model was established by combining CT radiomics signature and clinical independent predictors. Receiver operating characteristic (ROC) curve and decision curve analysis (DCA) were used to evaluate the predictive efficacy of each model.Results:Total of 205 AP patients were included (59 cases in severe group, 146 cases in non-severe group). 3, 5, 5 and 5 optimal radiomics features were selected from the plain CT scan, arterial phase, venous phase and delayed phase images of all patients, and the radiomics models were established. Among them, the arterial phase radiomics model had relatively better performance in predicting SAP, with an area under curve (AUC) of 0.937 in the training group and 0.913 in the validation group. Multivariate Logistic regression showed that C-reactive protein (CRP) and lactate dehydrogenase (LDH) were independent predictors of SAP, and they were used to establish a clinical model. The AUC in the training and validation groups were 0.879 and 0.889, respectively. The Nomogram model based on arterial phase CT radiomics signature, CRP and LDH was established, and the AUC was 0.956 and 0.947 in the training group and validation group, respectively. DCA showed that the net benefit of Nomogram model was higher than that of clinical model or radiomics model alone.Conclusions:The Nomogram model established by CT radiomics combined with clinical indicators has high application value for early prediction of the severity of AP, which is conducive to the formulation of clinical treatment plans and prognosis evaluation.

Objective:To investigate the value of dual-detector spectral CTA in distinguishing infarct core from penumbra in patients with acute ischemic stroke(AIS), and to further explore the risk factors associated with infarct core and their predictive value.Methods:The imaging and clinical data of 163 patients with AIS who met the inclusion criteria admitted to the Second Affiliated Hospital of Soochow University from March 2022 to May 2023 were retrospectively analyzed. Patients from March 2022 to December 2022 were used as the training group, and patients from January 2023 to May 2023 were used as the validation group for internal validation. The head and neck spectral CTA and brain CT perfusion imaging with dual-layer detector spectral CT were all carried out on all patients. Using CTP as reference, the patients were divided into infarct core group and non-infarct core group according to whether an infarct core occurred in the hypoperfusion regions of brain tissue. Multivariate logistic regression analysis was used to screen predictors related to the infarct core. The receiver operating characteristic (ROC) curve was used to evaluate the predictive efficacy.Results:A total of 163 patients were included in the study, including 112 in the training group and 51 in the validation group. There were significant differences in iodine density, effective atomic number, hypertension, triglyceride and neutrophils between the two groups ( P< 0.05). The cutoff values for iodine density values and effective atomic number values were 0.215 mg/mL and 7.405, respectively. Multivariate logistic regression analysis showed that iodine density and hypertension were independent risk factors for infarct core in AIS, and triglyceride was an independent protective factor. The area under the ROC curve (AUC) of iodine density value was the largest (0.859), with a sensitivity of 70.27%, and a specificity of 90.67%, which had a good predictive value. The ROC curve analysis results for the validation group were consistent with the training group. Conclusions:Spectral CT parameters iodine density values and effective atomic number values have the potential to distinguish the infarct core area from the penumbra area in patients with AIS. Iodine density and hypertension were independent risk factors of infarct core in AIS, triglyceride was an independent protective factor, and iodine density values obtained by dual-layer spectral detector CT had a high predictive value.

Objective To quantitatively study the correlation between T1ρ and T2 values of glenohumeral articular cartilage and the degree of rotator cuff injury.Methods A total of 149 patients with rotator cuff injury and healthy volunteers were prospectively selected.All of them underwent MRI routine scanning and T1 ρ and T2 mapping sequences.The degree of rotator cuff injury was graded,and the T1ρ and T2 values of glenohumeral articular cartilage were measured to analyze their relationship.Results With the development of rotator cuff injury grading,the T1 ρ and T2 values of glenohumeral articular cartilage increased.There were statistically significant differences in T1ρ and T2 values of articular cartilage between the different grades of rotator cuff injury(P＜0.001).Conclusion The injury of glenohumeral articular cartilage is aggravated with the severity of rotator cuff injury.The severity of rotator cuff injury can be evaluated by analyzing the T1 ρ and T2 values of glenohumeral articular cartilage.

Objective To explore the feasibility of predicting expression level of Ki-67 in lung adenocarcinoma via multi-parameter of spectral CT.Methods The data of 226 patients with lung adenocarcinoma confirmed by pathology were analyzed retrospectively.The conventional and spectral CT parameters of the lesions were analyzed.According to Ki-67 expression level,all patients were divided into low expression group and high expression group.The parameters with statistical significance were identified as independent variables for multivariate logistic regression analysis to establish a logistic regression model for predicting the expression level of Ki-67.Receiver operating characteristic(ROC)curve was used to assess the diagnostic performance for each model,respectively.Results There were significant differences in the clinical factors of gender,smoking and chest pain between high and low Ki-67 expression groups.In spectral CT parameters,CT40 keV,CT100 keV,Z-effective(Zeff)and iodine concentration(IC)in the high expression group in arterial phase were significantly higher than those in the low expression group.Logistic regression analysis showed that CT100 keV was the independent risk factor for Ki-67 expression level.Both the spectral CT model and the combined model had high value in predicting the expression level of Ki-67 in lung adenocarcinoma,and the combined model had better diagnostic efficacy.Conclusion Spectral CT parameters combined with clinical factors have a certain value in predicting the expression level of Ki-67 in lung adenocarcinoma.

Humans ; Aged ; Lung Diseases ; Pneumonia/drug therapy* ; Adrenal Cortex Hormones/therapeutic use* ; Pulmonary Disease, Chronic Obstructive/drug therapy* ; Administration, Inhalation ; Community-Acquired Infections/drug therapy*

Objective:To investigate the predictive value of clinical factor model, deep learning model based on baseline plain CT images, and combination of both for predicting hematoma expansion in cerebral hemorrhage.Methods:The study was cross-sectional. Totally 471 cerebral hemorrhage patients who were firstly diagnosed in the Second Affiliated Hospital of Soochow University from January 2017 to December 2021 were collected retrospectively. These patients were randomly divided into a training dataset ( n=330) and a validation dataset ( n=141) at a ratio of 7∶3 by using the random function. All patients underwent two noncontrast CT examinations within 24 h and an increase in hematoma volume of >33% or an absolute increase in hematoma volume of >6 ml was considered hematoma enlargement. According to the presence or absence of hematoma enlargement, all patients were divided into hematoma enlargement group and hematoma non-enlargement group.Two-sample t test, Mann-Whitney U test or χ2 test were used for univariate analysis. The factors with statistically significant differences were included in multivariate logistic regression analysis, and independent influences related to hematoma enlargement were screened out to establish a clinical factor model. ITK-SNAP software was applied to manually label and segment the cerebral hemorrhage lesions on plain CT images to train and build a deep learning model based on ResNet50 architecture. A combination model for predicting hematoma expansion in cerebral hemorrhage was established by combining independent clinical influences with deep learning scores. The value of the clinical factor model, the deep learning model, and the combination model for predicting hematoma expansion in cerebral hemorrhage was evaluated using receiver operating characteristic (ROC) curves and decision curves in the training and validation datasets. Results:Among 471 cerebral hemorrhage patients, 136 cases were in the hematoma enlargement group and 335 cases were in the hematoma non-enlargement group. Regression analyses showed that male ( OR=1.790, 95% CI 1.136-2.819, P=0.012), time of occurrence ( OR=0.812, 95% CI 0.702-0.939, P=0.005), history of oral anticoagulants ( OR=2.157, 95% CI 1.100-4.229, P=0.025), admission Glasgow Coma Scale score ( OR=0.866, 95% CI 0.807-0.929, P<0.001) and red blood cell distribution width ( OR=1.045, 95% CI 1.010-1.081, P=0.011) were the independent factors for predicting hematoma expansion in cerebral hemorrhage. ROC curve analysis showed that in the training dataset, the area under the curve (AUC) of clinical factor model, deep learning model and combination model were 0.688 (95% CI 0.635-0.738), 0.695 (95% CI 0.642-0.744) and 0.747 (95% CI 0.697-0.793) respectively. The AUC of the combination model was better than that of the clinical model ( Z=0.54, P=0.011) and the deep learning model ( Z=2.44, P=0.015). In the validation dataset, the AUC of clinical factor model, deep learning model and combination model were 0.687 (95% CI 0.604-0.763), 0.683 (95% CI 0.599-0.759) and 0.736 (95% CI 0.655-0.806) respectively, with no statistical significance. Decision curves showed that the combination model had the highest net benefit rate and strong clinical practicability. Conclusions:Both the deep learning model and the clinical factor model established in this study have some predictive value for hematoma expansion in cerebral hemorrhage; the combination model established by the two together has the highest predictive value and can be applied to predict hematoma expansion.

Objective To investigate the prevalence of Cryptosporidium infection and the distribution of parasite species and genotypes among HIV-positive individuals in Jiangxi Province. Methods HIV-positive individuals' sociodemographic and clinical data were collected from three AIDS designated hospitals in Jiangxi Province from January 2022 to March 2023. Subjects' stool samples were collected, and genomic DNA was extracted from stool samples. Nested PCR assay was performed based on the small subunit ribosomal RNA (SSU rRNA) gene of Cryptosporidium, and Cryptosporidium gp60 gene was amplified in stool samples positive for the SSU rRNA gene. The second-round PCR amplification product was checked with 1.5% agarose gel electrophoresis, and the products of suspected positive amplifications were sequenced, followed by sequence alignment. The phylogenetic tree was created using the Neighbor-Joining method with the software MEGA 11.0, to characterize the species, genotypes and sub-genotypes of Cryptosporidium. Results A total of 382 HIV-positive individuals were enrolled, with two cases identified with Cryptosporidium infection (0.52% prevalence), and both cases had no abdominal pain or diarrhea. Following sequencing and sequence alignment, the gene sequences of these two Cryptosporidium isolates shared 99.76% and 99.88% similarity with the gene sequence of C. meleagridis isolates. Phylogenetic analysis based on the Cryptosporidium SSU rRNA gene sequence identified the species of these two Cryptosporidium-positive stool samples as C. meleagridis. Following nested PCR amplification of the Cryptosporidium gp60 gene, sequencing and sequence alignment, the two C. meleagridis isolates were characterized as III eA17G2R1 and III bA25G1R1a sub-genotypes, and the sub-genotype III bA25G1R1a was firstly described in humans. Conclusion The prevalence of Cryptosporidium is low among HIV-positive individuals in Jiangxi Province. The likelihood of Cryptosporidium infection cannot be neglected among HIV-positive individuals without diarrhea.

Objective:To explore the pathogenesis of flunarizine-induced parkinsonism from gut-brain axis perspective.Methods:Thirty male C57BL/6 mice were randomly divided into control group and flunarizine group ( n=15). Mice in the control group were given 0.1 mL 50% polyethylene glycol 400+50% saline by gavage once/d for 2 weeks, while mice in the flunarizine group were given 6 mg/mL flunarizine+50% polyethylene glycol 400+50% saline by gavage at a daily dose of 30 mg/kg for 2 weeks. Body mass was recorded 1, 3, 5, 7, 10 and 14 d after drug administration, and motor function was assessed by rotarod test 14 d after drug administration; 16s RNA sequencing was performed in the feces to observe the intestinal flora; intestinal transit function was detected by Evans blue by gavage; and then, the mice were sacrificed and homogenate or frozen sections (brain and intestinal tissues) were prepared; dopamine-ergic neuron expression was detected by Western blotting; RT-qPCR was applied to detect the expressions of inflammatory factors in the substantia nigra, and immunofluorescent staining was used to detect the expressions of ZO-1 and Claudin-5 in the intestinal epithelial tissues. Results:Compared with the control group, the flunarizine group had lower body mass ratio 1, 3, 5, 7, 10 and 14 d after drug administration (ratio to body mass before drug administration). Compared with the control group, the flunarizine group had significantly shortened residence time in rod rotating and lower rotational speed when falling ( P<0.05). Compared with the control group, the flunarizine group had decreased tyrosine hydroxylase protein in the substantia nigra without significant difference ( P>0.05). Compared with the control group, the flunarizine group had significantly increased interleukin-6 and tumor necrosis factor-α in the substantia nigra (1.00±0.00 vs. 2.79±0.83; 1.00±0.00 vs. 3.39±1.37), significantly lower intestinal Evans blue propulsion rate (80.67%±4.51% vs. 50.67%±6.03%), and statistically decreased ZO-1 and Claudin-5 expressions in the colonic epithelial tissues (27.01±1.41 vs. 16.32±2.83; 37.00±2.80 vs. 24.52±2.12, P<0.05). Totally, 576 microorganisms were noted in both control group and flunarizine group, 744 in the control group alone, and 634 in the flunarizine group alone. The intestinal flora β diversity indices in the 2 groups were significantly different based on weighted Unifrac-principle coordinates analysis (PCoA, PCoA1: 39.88%; PCoA2: 30.69%). Compared with the control group, the microbial colony structure of mice in flunarizine group was dominated by phylum thick-walled bacteria and phylum warty microbacteria, and by families Muribaculaceae, Lachnospiraceae and Akkermansiaceae. Compared with the control group, the flunarizine group had significantly decreased relative abundance of Ackermannia spp. and Lactobacillus spp. in the intestinal flora ( P<0.05). Conclusion:Flunarizine may contribute to the pathogenesis of DIP by causing structural disturbances in the intestinal flora and inducing neuroinflammation based on the gut-brain axis.