| Objective: Through clinical collection of traditional Chinese medicine(TCM)inquiries and voice information from patients with coronary heart disease(CHD)and normal people,the study applies integrated information processing methods of the modern speech signal processing and pattern recognition,analyses the basic features of phonetic changes in CHD patients,exploring changes of the speech features in CHD patients from different TCM syndromes and viscera syndromes,initially builds a recognition model of TCM syndromes of CHD patients.The research findings will supply an objective basis to clinical differentiate diseases and syndromes according to voice characteristics,and have a certain significance to deeply understand the theory of acoustic diagnosis in TCM and clinical application and promotion.Methods:(1)Based on the "Information Collection Form for TCM inquiry of CHD" from research group,the study collects TCM inquiry information from CHD patients,including general information,key consultation,general consultation,tongue,surface,pulse and other information.(2)Using Logitech H340 headset type microphone and the acoustic diagnosis module of Smart TCM-1 TCM life information analysis system to collect voice information.(3)Assign values of "1" and "0" respectively according to "yes" and "no" of TCM consultation information.The speech information is preprocessed by Python3.7.The wavelet packet analysis method is used to analyze the wavelet packet energy entropy and wavelet packet Shannon entropy in different frequency bands of CHD patients.Epidata software is used to construct the TCM clinical information database of CHD patients.(4)Based on the SPSS21.0 statistical software,frequency analysis method is used to analyze the distribution characteristics of CHD patients with different TCM syndromes and different viscera syndromes.Using the rank-sum test to describe the degree of dispersion with the quartile interval M(Q1,Q3),and compare the differences of speech characteristics between CHD patients and normal people,CHD patients with different TCM syndromes and CHD patients with different viscera syndromes.The CHAID decision tree method is used to explore the establishment of TCM syndrome identification model for patients with CHD.Results:(1)According to TCM syndrome differentiation and classification,among 489 patients with CHD,193 cases(39.47%)with deficiency of both qi and yin,accounting for the largest proportion;141 cases of qi deficiency and blood stasis syndrome(28.83%);57 cases of qi stagnation and blood stasis syndrome(11.66%).The rest are from high to low in order: 47 cases(9.61%)of heart blood stasis,32 cases(6.54%)of heart and kidney yin deficiency,15 cases(3.07%)of phlegm obstruction of heart pulse,4 cases(0.82%)of yang qi deficiency and failure.(2)According to the differentiation of TCM zang-fu syndrome,only 124 cases(25.36%)are in heart diseases,and 158 cases(32.31%)are in CHD combined with liver diseases.The rest are from high to low: 112cases(22.90%)of CHD combined with renal diseases,55 cases(11.25%)of CHD combined with splenic diseases,and 40 cases(8.18%)of CHD combined with pulmonary diseases.(3)Compared with the normal control group,the CHD patients show wavelet packet energy entropy are in the frequency band of 1750-4750Hz(wavelet energy 08-wavelet energy 19)and 6250-8000Hz(wavelet energy 26-wavelet energy32).The wavelet packet Shannon entropy features are in the frequency band of7750-8000Hz(Shannon entropy 32),the difference is statistically significant.(4)The comparison of speech features of different syndromes of CHD patients shows that the wavelet packet energy entropy features are in the two frequency bands of 5250-6500Hz(wavelet energy 22-26)and 6750-8000Hz(wavelet energy 28-32),and the differences between the seven groups of different TCM syndromes and between the two groups are statistically significant.The wavelet packet shannon entropy characteristics in 0-250Hz(shannon entropy 01),750-1000Hz(shannon entropy 04),2250-3000Hz(shannon entropy10,11,12),3750-4000Hz(shannon entropy 16),5250-5500Hz(shannon entropy22),6000-6250Hz(shannon entropy 25),7000-7250Hz(shannon entropy 29),7500-7750Hz(shannon entropy 31),seven groups of different TCM syndrome types and two more statistically significant differences between groups.There is no significant difference in wavelet packet energy characteristics among male patients with different syndromes in CHD.In the frequency band of 2250-2500Hz(shannon entropy 10)and 2500-2750Hz(shannon entropy 11),the difference between five groups of different TCM syndrome types and pair comparison among each group is statistically significant.In female patients with different syndromes in CHD,the wavelet packet energy characteristics are within the frequency band of 2500-2750Hz(wavelet energy 11),5250-5500Hz(wavelet energy 22)and 5750-8000Hz(wavelet energy 24-wavelet energy 32)and the wavelet packet shannon entropy features in the 0-250Hz(shannon entropy 01),6000-6250Hz(shannon entropy 25)frequency band,the five groups of different TCM syndrome types and pair comparison of differences between each group are statistically significant.(5)Comparison of speech features of different viscera and syndromes in patients with CHD shows that the energy features of wavelet packet are in the frequency band of2000-2250Hz(wavelet energy 09),and the differences among different viscera and syndromes are statistically significant.The wavelet packet shannon entropy characteristic is in the frequency band of 3000-3250Hz(shannon entropy 13),between different viscera and syndrome and pair comparison,the shannon entropy value of CHD combined with liver diseases is lower than that of the CHD combined with splenic diseases in this frequency band,which is statistically significant.(6)Using the CHAID decision tree model,the accuracy rate of inspection-interrogation-palpation and acoustic diagnosis for qi deficiency and blood stasis syndrome is 63.4% higher than that of inspection-interrogation-palpation diagnosis,56.4%.The sensitivity,specificity and accuracy are 64.6%,69.8% and 67.3% for deficiency of both qi and yin.The accuracy rate of inspection-interrogation-palpation and acoustic diagnosis for syndrome of qi stagnation and bloodstasis is 84.2%,lower than that diagnosis,85.1%.The identification accuracy of inspection-interrogation-palpation and acoustic diagnosis is 91.1% lower than that of inspection-interrogation-palpation diagnosis,92.1%.Conclusion:(1)The distribution of TCM syndromes in CHD patients is mainly characterized by deficiency of both qi and Yin,deficiency of qi and blood stasis,stagnation of qi and blood stasis,which reflects the pathogenesis characteristics of CHD with the deficiency and deficiency as well as the mixture of deficiency and excess.The distribution of zang-fu syndromes is more common with liver syndromes,simple heart syndromes and kidney syndromes,indicating the dysfunction of the heart,liver and kidney and the obstruction of the heart arteries.(2)In some frequency bands between CHD patients and normal control group,the wavelet packet energy entropy may provide certain reference value for clinical diagnosis of CHD.(3)In some frequency bands,the characteristics of speech energy may be specific for the clinical discrimination of the heart and kidney yin deficiency syndrome and blood stasis syndrome of CHD.It may be of significance for the determination of heart and kidney yin deficiency syndrome,qi stagnation and blood stasis syndrome and blood stasis syndrome in female patients with CHD.The phonetic characteristics of shannon entropy can reflect the heart blood stasis syndrome and yang qi deficiency syndrome of CHD.Etiology and pathogenesis characteristics of the deficiency of heart and kidney yin and qi and yin in male CHD patients and the deficiency of heart and kidney yin and blood stasis in female CHD patients,which provide objective basis for clinical syndrome differentiation and classification.(4)In some frequency bands,the characteristics of speech energy have a certain specificity for the identification of viscera syndromes of CHD combined with pulmonary diseases.The characteristics of phonetic Shannon entropy in the CHD combined with splenic diseases and viscera are more uncertain and less regular than those in other organs.In addition to the CHD combined with liver and splenic diseases,the CHD and CHD combined with splenic diseases also have a certain trend of difference,which may provide reference value for the further study and discussion of the theory of "five viscera sound" in CHD in the future.The packet shannon entropy may become one of the auxiliary indicators for the diagnosis of CHD complicated with complications.(5)To construct a TCM syndrome identification model for CHD patients,TCM acoustic diagnosis has certain guiding significance for the clinical discrimination of CHD syndrome of qi deficiency and blood stasis.Some phonetic energy and shannon entropy characteristics may provide objective indicators for TCM acoustic diagnosis for its clinical syndrome differentiation and classification. |
PDF Full Text Request