Objective:To explore the construction and efficacy of cranial magnetic resonance imaging(MRI)classification model based on attention mechanism in type 2 diabetes patients with cognitive impairment.Methods:The case data of 100 patients with type 2 diabetes who were treated in the General Hospital of North China Petroleum Administration Bureau from June 2022 to January 2024 were retrospectively selected.A total of 100 MRI images of cranial FLAIR_LongTR sequence with cognitive impairment(32 cases)and those without cognitive impairment(68 cases)were respectively collected.The images of the above two kinds of samples were horizontally and vertically translated to expand to 1000 samples,respectively.The samples were randomly divided into training samples(n=700)and test samples(n=300)as the ratio of 7:3 according to affine transformation data augmentation method.Then,the attention mechanism model was established to test the images with full scan of the test samples.The ability of the attention mechanism system in screening cognitive impairment was analyzed according to the method of setting threshold value.The 100 MRI images of cranial FLAIR_LongTR sequence of patients in our hospital from January 2024.From January to May 2024 were used as a verification set to verify the diagnostic value of attention mechanism.Results:With the increasing of iteration times,the sample loss of training and verification of attention mechanism model gradually decreased and tended toward stability,and the accuracy of training set and verification set gradually increased and tended toward stability.In the attention mechanism model,the average loss rate of training samples was 10.024%,and that of test samples was 15.247%.In the attention mechanism model,the average accuracy of training samples was 99.078%,and the average accuracy of test samples was 99.753%.Receiver operating characteristic(ROC)curves showed that the area under curve(AUC)of attention mechanism model was 0.998,which can better diagnose cognitive impairment of patients with type 2 diabetes than the resNET model(AUC=0.656)(Z=3.437,P<0.001).Conclusion:The constructed cranial MRI classification model by using attention mechanism has favorable diagnostic value for cognitive impairment in patients with type 2 diabetes.

Objective:To explore the construction and efficacy of cranial magnetic resonance imaging(MRI)classification model based on attention mechanism in type 2 diabetes patients with cognitive impairment.Methods:The case data of 100 patients with type 2 diabetes who were treated in the General Hospital of North China Petroleum Administration Bureau from June 2022 to January 2024 were retrospectively selected.A total of 100 MRI images of cranial FLAIR_LongTR sequence with cognitive impairment(32 cases)and those without cognitive impairment(68 cases)were respectively collected.The images of the above two kinds of samples were horizontally and vertically translated to expand to 1000 samples,respectively.The samples were randomly divided into training samples(n=700)and test samples(n=300)as the ratio of 7:3 according to affine transformation data augmentation method.Then,the attention mechanism model was established to test the images with full scan of the test samples.The ability of the attention mechanism system in screening cognitive impairment was analyzed according to the method of setting threshold value.The 100 MRI images of cranial FLAIR_LongTR sequence of patients in our hospital from January 2024.From January to May 2024 were used as a verification set to verify the diagnostic value of attention mechanism.Results:With the increasing of iteration times,the sample loss of training and verification of attention mechanism model gradually decreased and tended toward stability,and the accuracy of training set and verification set gradually increased and tended toward stability.In the attention mechanism model,the average loss rate of training samples was 10.024%,and that of test samples was 15.247%.In the attention mechanism model,the average accuracy of training samples was 99.078%,and the average accuracy of test samples was 99.753%.Receiver operating characteristic(ROC)curves showed that the area under curve(AUC)of attention mechanism model was 0.998,which can better diagnose cognitive impairment of patients with type 2 diabetes than the resNET model(AUC=0.656)(Z=3.437,P<0.001).Conclusion:The constructed cranial MRI classification model by using attention mechanism has favorable diagnostic value for cognitive impairment in patients with type 2 diabetes.

Objective To prepare monoclonal antibodies (mAb) against the type Ⅳ secretion system protein VirB5 of Brucella melitensis and to provide a basis for pathgenic diagnosis and research of brucellosis.Methods Four SPF female BALB/c mice were subcutaneously immunized with purified VirB5 protein at a dose of 60 μg/mice,and immunization was strengthened every 2 weeks at a dose of 30 μg/mice,three times in total.Two weeks later,the orbital venous blood of mice was taken to determine the antibody titer,and then intraperitoneally injected for the fourth time to strengthen immunization.Three days later,mouse spleen cells were fused with mouse myeloma SP2/O cells in a ratio of 5:1.After 3 times of cell screening and monoclonal cloning,the hybridoma cell lines with stable secretion of VirB5 antibody were established;one BALB/c mouse was intraperitoneally injected with hybridoma cells,and ascites were collected and antibody was purified when the mouse abdomen was significantly enlarged.The immunological characteristics of mAbs were identified by indirect enzyme-linked immunosorbent assay (ELISA) and Western blotting.Results A total of 6 monoclonal cell lines (2-2,2-12,2-19,2-25,2-31 and 2-40) capable of secreting VirB5 antibody were established.Among them,the cell line 2-19 can stably secrete an antibody that specifically recognized the VirB5 protein,and the VirB5 antibody secreted by the cell line was identified as an IgG1 subtype,a kappa light chain,a mAb affinity constant of 1.6 × 108.The titer of ascites antibody of mouse intraperitoneally injected with hybridoma cell 2-19 was 1:51 200.Conclusion The high-affinity mAb of type Ⅳ secretion system protein VirB5 is successfully prepared,and the antibody can rapidly bind specifically to pathogens,providing an alternative material for establishment of brucellosis pathogen diagnostic method.

Iron is involved in the virulence and pathogenic effects of certain intracellular parasites.In the pathogenic process of Brucella,the uptaking and metabolism of host iron are closely related to intracellular parasitism and immunity escape of Brucella.In this paper,we elucidated the iron transport system,iron response regulators and nutrient immunity of iron based on the latest report and data about Brucella.