| ObjectiveMetabolic disease is a major chronic non-communicable disease in clinical practice.This study will introduce the concept of“Glucolipid Metabolic Disorders(GLMD)" first proposed by Professor Guo Jiao.Objectively collect all kinds of tongue images of glycolipid metabolic disorders by means of objective computer tongue image acquisition platform image analysis technology,explore the common tongue images of glycolipid metabolism disorders,analyze the differences of various tongue images,and explore the tongue image of glycolipid metabolic disorders.Relationship with etiology,pathogenesis characteristics and objective indicators of glycolipid metabolic disorders.At the same time,we established a TCM syndrome artificial neural network model that integrates the high-frequency tongue image of glycolipid metabolic disorders,in order to provide a reliable objective basis for the individualized syndrome differentiation and treatment of glycolipid metabolic disorders,and to establish an objective criteria for the evaluation of tongue image in the future.Methods1.Using clinical epidemiological research methods,according to strict diagnosis,inclusion,and exclusion criteria,screening patients with glycolipid metabolism disorders who meet the research conditions,and collecting general information and glucose and lipid metabolism indicators.The Daosheng DSO1-A tongue-and-groove information acquisition physique identification system is used to collect the patient’ s tongue image information according to the unified SOP to ensure the objectification of tongue image information collection,and at the same time realize the acquisition of tongue-like objective information data.Comprehensive analysis of tongue image information by frequency analysis,Pearsonχ2 test,one-way analysis of variance and other statistical methods.2.Based on the research results of the pre-existing syndrome research results and the clinical epidemiological investigation of the tongue image of glycolipid metabolism disease,the artificial neural network was used to construct the identification model of glycolipid metabolic disorders syndrome based on artificial neural network.The general information of patients and the four diagnostic information of Chinese medicine were collected using the"Glucolipid Metabolic Disorders Clinical Four Diagnostic Information Questionnaire" to train and verify the system.At the same time,the artificial neural network model inherent in SPSS software was used to predict the TCM syndromes of glycolipid metabolism disorders.Resu ts1.The results of the research on the clinical epidemiology of tongue of glycolipid metabolic disorders show that:(1)The general epidemiological characteristics of glycolipid metabolic disorders:the ratio of male to female is 1:1.12,more women than men.359 cases were aged 30-87 years with an average age of 63.15±11.86 years.Among them,16 cases were<40 years old,accounting for 4.45%;115 cases were 40-59 years old,accounting for 32.04%;178 cases were 60-75 years old,accounting for 49.58%;50 cases were>75 years old,accounting for 13.93%.There were 46 cases in stage I of glycolipid metabolic disorders,accounting for 12.81%;259 cases in stage II of glycolipid metabolic disorders,accounting for 72.15%;54 cases in stage III of glycolipid metabolic disorders,accounting for 15.04%.TCM syndrome differentiation was 77 cases of liver stagnation and spleen deficiency,accounting for 21.45%;121 cases of dampness block syndrome,accounting for 33.71%;26 cases of dampness and heat internal syndrome,accounting for 7.24%;57 cases of qi and yin deficiency syndrome,accounting for 15.88%;61 cases of spleen and kidney yang deficiency,accounting for 16.99%;16 cases of phlegm and blood stasis syndrome,accounting for 4.46%;1 case of yang deficiency and turbidity syndrome,accounting for 0.27%.In 359 cases,the constitution of TCM was 41 cases,accounting for 11.43%;61 cases of yang deficiency,accounting for 16.99%;25 cases of qi deficiency,accounting for 6.96%;119 cases of phlegm dampness,accounting for 33.15%;52 cases of yin deficiency,Accounted for 14.48%;moist heat quality in 18 cases,accounting for 5.01%;blood stasis in 16 cases,accounting for 4.46%;qi stagnation in 22 cases,accounting for 6.13%;special enamel in 5 cases,accounting for 1.39%.(2)The distribution of tongue in patients with glycolipid metabolic disorders:the tongue shape was the most common in the tongue(200 cases),followed by the scallop tongue(166 cases),the crack tongue(156 cases)and the fat tongue(149 cases)also accounted for A high proportion,while a small number of thin tongues and point tongues.Tongue color:the dark red tongue is the most,a total of 210 cases(58.50%);followed by light red tongue,a total of 80 cases(22.28%)and 67 cases of light tongue(18.66%),red tongue,pale purple tongue The proportion is small.(3)The distribution of tongue coating in patients with glycolipid metabolic disorders:the moss is mostly white,a total of 210 cases,followed by yellow and white moss(61 cases);gray black moss(3 cases)and yellow moss(18 cases)are less.In terms of moss,there were 67 cases of no moss(18.66%)and 3 cases of exfoliation(0.84%).The appearance of no moss and exfoliation suggested that the condition was severe and the stomach was injured.Thin moss(190 cases)accounted for the largest proportion,accounting for 52.92%.In addition,the ratio of thick moss to greasy moss is also quite large,accounting for 24.23%and 28.97%,respectively.The appearance of thick and greasy moss indicates that there are phlegm,dampness,heat and dampness,suggesting that dampness,heat and dampness are the main pathological products of glycolipid metabolic disorders.(4)There was a statistically significant difference in the frequency of dentate tongue between patients with different genders of glycolipid metabolic disorders.There was no statistical difference in the frequency of other tongue markers.There was no statistical difference in the frequency of occurrence of tongue coating between patients with different genders of glycolipid metabolic disorders.(5)The frequency of appearance of fat tongue,cracked tongue,thick moss and thin moss was statistically different among patients with glycolipid metabolic disorders at different ages.(6)The frequency of occurrence of scalloped tongue,dark red tongue,no moss,white moss,thin moss,and greasy moss was statistically different among patients with different stages of glycolipid metabolic disorders.(7)There is a difference between the peeling of moss and gray-black moss in patients with different syndromes of glycolipid metabolic disorders.The tongue image of each TCM syndrome type is basically consistent with its syndrome type on the pathogenesis of traditional Chinese medicine.(8)There is a difference in the peeling of moss in patients with different constitutions of glycolipid metabolic disorders.The tongue image of patients with different constitution types is roughly consistent with the pathology of traditional Chinese medicine.(9)Objective tongue image parameters are different in patients with different stages,different constitutions and different syndromes of glycolipid metabolic disorders,or can be used as a reference for judging different stages,different constitutions and different syndrome types of patients with glycolipid metabolic disorders.(10)The objective tongue image parameters of patients with glycolipid metabolic disorders have a certain correlation with the glucose and lipid metabolism index.The objective parameters of tongue image may be one of the reference indicators for the severity,prognosis and efficacy evaluation of patients with glycolipid metabolic disorders.2.Based on artificial neural network,the glycolipid metabolic disorders syndrome identification system was used to classify and identify TCM syndromes of glycolipid metabolic disorders in 264 patients with glycolipid metabolic disorders.The results show that the identification system has achieved a high accuracy rate for each syndrome(average above 60%),indicating that the neural network can effectively deal with the existence of strong subjective,non-linear TCM data.For the syndrome identification accuracy of the four syndrome types of dampness block,damp heat intrinsic,qi and yin deficiency,and phlegm and blood stasis,the artificial neural network with multiple outputs of a single network(ACON)structure has higher accuracy.However,the artificial neural network of the multi-network(OCON)structure with single output of liver stagnation and spleen deficiency syndrome is higher.It is suggested that the identification of different syndromes of TCM in glycolipid metabolic disorders,the neural network of ACON structure and the recognition rate of neural network of OCON structure are different.Based on the artificial neural network algorithm inherent in SPSS20.0 software,the artificial neural network of MLP and RBF architecture can also effectively identify the TCM syndromes of glycolipid metabolic disorders.Both research results suggest that artificial neural networks can be used.Glycolipid metabolic disorders,a complex TCM syndrome identification of chronic diseases,and the improvement of the TCM syndrome identification system for glycolipid metabolic disorders,still need more data for learning and verification,in order to find a higher recognition accuracy Algorithm.ConclusionThis study preliminarily clarified the common tongue image of glycolipid metabolic disorders(moderate tongue shape,dark red tongue,white moss,yellow and white moss,fat tongue accompanied by cracks and tooth marks).The relationship between tongue image and gender,age,stage of glycolipid metabolic disorders,syndrome type,constitution,and glucose and lipid metabolism index was preliminarily defined.The analysis of the etiology and pathogenesis of glycolipid metabolic disorders from the perspective of tongue image was completed.The main pathogenesis machine,the spleen loses health and travels throughout,and the phlegm,dampness,phlegm,heat and poison are the main pathological products.It preliminarily clarifies the relationship between tongue image of glycolipid metabolic disorders and its objective index and glucose and lipid metabolism index,which not only finds potential evaluation index for glycolipid metabolism disease,but also lays a foundation for objective evaluation criteria of tongue and throat of glycolipid metabolic disorders.This study preliminarily demonstrated that artificial neural networks can be used in the identification of TCM syndromes of complex chronic diseases such as glycolipid metabolic disorders.The construction of TCM syndrome identification model based on artificial neural network for glycolipid metabolic disorders provides a new idea for the objective study of TCM syndromes of glycolipid metabolic disorders,and also provides a reliable basis for TCM objective differentiation and treatment of glycolipid metabolic disorders.Innovation1.Break through the previous study of tongue diagnosis of metabolic diseases under the framework of single disease,and carry out research on tongue diagnosis of metabolic diseases in the macroscopic view of comprehensive integration.2.For the first time,the relationship between objective tongue-like parameters of glycolipid metabolic disorders and syndromes and glucose and lipid metabolism indexes was discussed from the aspects of mathematical statistics.The pathology of TCM pathogenesis of glycolipid metabolic disorders was explained from the tongue diagnosis information.3.The TCM syndrome identification model based on artificial neural network for glycolipid metabolic disorders was constructed for the first time,which is conducive to perfecting the syndrome identification system of glycolipid metabolic disorders and improving the objectification level of syndrome identification of glycolipid metabolic disorders. |
PDF Full Text Request