Research On Odour Recognition And Colour Olfactory Material Basis Of Chinese Medicine Based On Electronic Nose Technology

This study proposes to solve the key problems in objectively quantifying the empirical expressions of the odour of traditional Chinese medicine(TCM)pieces by utilising the rapid detection method on the odour of TCM pieces based on Electronic nose(E-nose)and the research ideas of odour dynamics during the TCM processing.The electronic nose technology was used in Chinese medicine processing,and representative medicine Crataegi Fructus(CF)and Gardeniae Fructus(GF)were selected as the research objects.The quantitative model of odour evaluation has been systematically studied from static(odour of standard pieces)and dynamic(odour changes during processing),respectively.At the same time,the combination of macroscopic(external odour)and microscopic(internal components)factors provide a new way of thinking for constructing the quality evaluation model on the odour-components combination of TCM pieces.Firstly,for the aim of detecting and characterize the odour of TCM pieces effectively,the detection method of the odour of TCM pieces was established by the single-factor investigation on the E-nose detection parameters including sample quality,particle size,instrumental injection volume,incubation time and temperature.The results of the methodological investigation showed that both repeatability and stability of the method meet the requirements.Then,stepwise discriminant analysis,variance analysis and R-type cluster analysis were used to optimize the sensor array to reduce the data dimension,and Linear discriminant analysis(LDA)was used to compare and analyze the data before and after the optimization.The optimized sensor arrays(10 sensors for CF and 8 sensors for GF)by stepwise discriminant analysis proved to be better than the original sensor array,as it not only retained most of the original information,but was also easy to operate.Furthermore,it has higher correct classification rate by Bayesian discriminant analysis(BDA).Secondly,in order to control the quality of TCM pieces effectively and guide the production of TCM pieces,the digital standard and model standard of odour were established using the odour response data of CF and GF(raw,fried,parched and charry products)collected through electronic nose based on the optimized array.The digital standard range was constructed using the bilateral 90%reference ranges of the CF and GF pieces based on each sensor response by the percentile method.The rationality of the ranges were confirmed by nonparametric test.The model standard was constructed by establishing the odour response value library of the pieces using visual Discriminant factor analysis(DFA).The cross-validation results showed that the established model is excellent.In addition,Bayes discriminant function(BDF)and Back-propagation neural network(BPNN)were used to establish mathematical prediction model of different processed products of CF and GF,which can be used as a preliminary identification tool.Then,in order to reveal the material basis of the external odor of the pieces during processing,and finally establish a quantitative quality standard control mode of multiple indicators of representative components and odour response value,the dynamic change regularities of the odour during the processing in different process were explored through analyzing the odour response data of pieces collected by E-nose.Meanwhile,the GC-MS fingerprint was adopted to establish a semi-quantitative analysis method for multiple indicators of volatile components in pieces.Then,the Grey relation analysis(GRA)was used to analyze the correlation between the gas chromatographic peaks and odour response values of pieces in the fried process.Finally,characteristic peaks with higher correlation(r>0.9).in CF(33)and GF(52)pieces were obtained and identified.The results indicated that alcohols,aldehydes,ketones,acids,alkanes,etc.to be the main component group that may lead to odour changes of CF during processing,and five compounds such as acetic acid,acetone,furfural may be the main characteristic components;alcohols,ketones,aldehydes,esters,aromatic hydrocarbons,acids,furans,pyrrole classes,pyrazines,isocyanoids,etc.may be the main component group that may lead to odour changes of GF during processing,and 18 compounds such as methyl acetate,2-methylfuran and formazan are the main characteristic components.In addition,the preliminary exploration on the material basis of colour change of pieces was also carried out.Firstly,the colour values of pieces was collected by using the spectrophotometer during the processing.The UHPLC-Q-TOF-MS/MS analysis of the pieces was established at the same time.Then the relevant compounds in CF(66)and GF(107)pieces were obtained by the correlation analysis between the chromatographic peaks and colour values of pieces in the fried process.On this basis,multiple regression analysis(Regression)was used to fit the linear regression model of the relevant chromatographic peak(X)and colour parameters(Y)and further explore the influence level according to the normalized partial regression coefficient of each relevant chromatographic peak.The results showed that there were 11 and 31 compounds which had a great influence on the color change during the frying process of CF and GF.Among them,the compounds Pinnatifidanin CV and isovaleric acid had more influence on the colour change during the frying process of the CF;deacetyl lavanylate,1,2,4-benzenetriol,Gardenal II,behenyl methyl ester and 5-hydroxy-7,4’-dimethoxydihydroflavone had more influence on the colour change during the frying process of the GF.