A Quantitative Evaluation Method Of Parkinson’s Disease With Hypodynamic Dysarthria Based On Speech Characteristics

Study 1: Explore the rhythm regularity of Parkinson’s disease with hypodynamic dysarthriaObjective : Based on the pronunciation characteristics of Mandarin Chinese,a vocabulary was designed to collect speech samples from normal adults and Parkinson’s patients(PD)for a comparative study of tone and duration,to explore the prosodic acoustic variation of PD dysarthria,so as to explore whether there is sensitivity of prosodic indicators in PD dysarthria patients.Methods: The mean value,median value,maximum value,minimum value and standard deviation of the fundamental frequency F0 of 31 PD patients and 38 normal people were extracted.The data were compared between groups and within groups by ANOVA,and the overall distribution of tone was drawn.The target vowel duration v-dur was extracted and the data were compared with t-test of two samples.Results: Five parameter points of each syllable in the case group were less than those in the normal group(P < 0.01).According to the main tone distribution map,all syllables in the case group showed lower level tone,smaller slope of rising tone and flatter fallingrising tone.The mean duration of 6 vowels(/a/、/i/、/u/、/o/、/e/、/ü/)in the case group was higher than that in the normal group(P < 0.01).Conclusion: It shows that the prosody acoustic index is sensitive to the patients with Parkinson’s dysarthria,so the prosody damage is also a common symptom of Chinese Mandarin patients with Parkinson’s dysarthria.Therefore,we should not only pay attention to the rehabilitation of vocabulary and language,but also pay attention to the training of rhythm.Study 2: Explore the sensitivity of multiple vowel-related metrics to Parkinson’s disease with hypodynamic dysarthriaObjectives:Compare the difference of acoustic metrics of vowel articulation Vowel Space Area(t VSA),Vowel Articulation Index(VAI)and Formant Concentration Ratio(FCR)between normal people and Parkinson’s disease patients,and discuss whether acoustic indicators are sensitive among Parkinson’s disease patientsMethods: Extract Vowel formant F1 and F2 of Parkinson’s disease(31 cases)and normal people(38 cases)and then substitute Vowel formant F1 and F2 into the formula to calculate the value of t VSA,VAI and FCR of the subjects.The data were compared using two sample t-tests.Result: There were significant differences in the mean values of t VSA,VAI and FCR between the two groups.Conclusion: tVSA,VAI and FCR are sensitive to PD patients,which can provide some guidance for the rehabilitation evaluation and rehabilitation effect of PD patients with dysarthria.Study 3: Obtain weight parameters based on multiple regression analysis and correlation studyObjectives: The contribution of each acoustic parameter to speech intelligibility is compared to obtain weighted acoustic parameters.The purpose of this study is to find out whether the changes of various indicators are related to age,gender and duration of disease,and to explore the acoustic indicators that can be used to infer the severity of disease,the age of onset and monitor the therapeutic effect.Methods: Pearson correlation analysis was carried out between the acoustic parameters obtained for quantitative assessment of Parkinson’s dysarthria and speech sample intelligibility,patient’s age,gender,and disease duration.Scatter plot was drawn.At the same time,acoustic parameters related to intelligibility were extracted and a multiple linear regression analysis was carried outResult: The correlation between the gender of patients and the mean fundamenta l frequency of each syllable was statistically significant,and the course of diseas e was positively correlated with the range of F0 in the upper voice.The results of 4 evaluators were consistent with those of 31 subjects(Cronbach’s = 0.996,I CC = 0.983).In regression analysis,the standard regression coefficients(F = 41.281,P = 0.000)of VAI(beta =-0.955)and f2i(beta = 0.985)were significantl y higher than other acoustic indexes.Conclusion: VAI and F2 i are more sensitive,effective and reliable in distinguish ing between normal and dysarthria,and can be regarded as weight parameters.T hrough the correlation analysis,it is concluded that the correlation between the gender of patients and the basic frequency of each syllable,the correlation betw een the course of disease and the range of F0 are statistically significant.This s tudy believes that the above conclusions may be a possible trend,which paves t he way for further study of an objective,reliable,convenient and universal eval uation technique.