The Research About Tonal Characteristics Of Children In Cerebral Palsy Children With Spastic Dysarthria Based On Experimental Phonetics

Part I:To analyze the intelligibility of mandarin in cerebral palsy children with spastic dysarthria.Objective:We explored the difference of intonation understanding degree between cerebral palsy children with spastic dysarthria and normal children of the same age.And whether the vocal tone of children with cerebral palsy is abnormal.Methods:The 7-15 years old cerebral palsy children with spastic dysarthria(n=28)were selected to analyze the intelligibility of mandarin tone,and they were compared with the same age matched with healthy controls(n=30).Results:(1)There was no significant difference in age and gender between the two groups(x~2=0.181,Z=-1.125,p>0.05).(2)The three evaluators had good consistency in the intelligibility of 58 subjects(Cronbach’s=0.931,ICC=0.907);(3)The overall tone intelligibility of the observation group was 75.70%±19.74%,the orther group was93.31%±7.77%,and the cerebral palsy group(mean rank was 58.78)was lower than the normal group(mean rank was 114.31)(Z=-7.319,p=0.000,p<0.001).And the result showed that the overall tone intelligibility of the observation group was significantly lower than that of healthy controls.Conclusions:We had found that tone of cerebral palsy children with spastic dysarthria was abnormal,and the articulation was significantly decreased compared with healthy controls.In the rehabilitation treatment of this group of people, rehabilitation training can be performed on the muscle movements related to the control of pronunciation tones.So as to improve the early speech learning of cerebral palsy and produce clear and effective speech patterns.PartⅡ:To explore the fundamental frequency differences in tonal acoustics of cerebral palsy children with sputum dysarthria.Objective:Based on experimental phonetics,we explored the difference of fundamental frequency in the acoustical tone between the cerebral palsy children with spastic dysarthria and healthy controls.Methods:The Praat software was used to extract the fundamental frequency data of the subjects’speech samples.The maximum,minimum,median and average values of the four parameter points of the fundamental frequency,which were statistically analyzed.Another the overall distribution of the tonal figure was drawn to intuitively comparing the differences in the fundamental frequency of the four different tonal changes between the two groups of children.Results:(1)Among the seven syllables,there were statistically significant for the four parameter points of the tonal average fundamental frequency values in the two groups of different groups(p<0.05),similarly the four parameter points of the four tonal average fundamental frequency values were all Statistically significant(p<0.05);(2)Among the seven syllables,the cerebral palsy group indicated that there were multiple abnormal values on the box plots of the four tonal average fundamental frequency values;(3)Among the/a/,/du/,/ge/,/bo/,/yu/syllables,it becomes steeper on the level tone in the overall distribution of the tonal figure,and the rising tone becomes steeper among the/ba/,/bi/,/du/,/ge/;Among the syllable/a/,/ge/,/bo/,/yu/,the falling-rising tone is changed from high to low,similar to the falling tone;Excepting for the syllable/du/,the other six syllables are flattened.Conclusions:(1)The fundamental frequency of mandarin tonal acoustics characteristics in cerebral palsy children with sputum dysarthria were different from that healthy controls,suggesting that the stability of tone fundamental frequency of cerebral palsy children with sputum dysarthria is lowly;(2)Compared with healthy controls,there were abnormalities in the distribution of four tonal subjects in seven different syllables of cerebral palsy children with sputum dysarthria.PartⅡI:The abnormal tone of Mandarin Chinese for cerebral palsy children with spastic dysarthria were evaluated automatically by artificial neural network.Objective:To explore the automatic evaluation of the mandarin tone in cerebral palsy by artificial neural network.Methods:After extracting the F0 of the speech samples of the two groups of subjects,the artificial neural network was used to intelligently identify the accuracy of in tone cerebral palsy children with sputum dysarthria,and tested the results..Results:(1)The tone was recognized by the artificial neural network,discovering the tone recognition rate of the control group was 84.46%±8.78%,compared with the observation group was 60.74%±15.44%,and the cerebral palsy group(mean rank was 16.54)was lower than the normal group(mean rank was 41.60)(Z=-5.655,p=0.000,p<0.001).The tone accuracy rate between the two groups was statistically significant;(2)In the four tones of cerebral palsy children with sputum dysarthria,the accuracy of rising tone and falling-rising tone in tone decreased significantly,and rising tone was the most easily confused into falling-rising tone,falling-rising tone and falling tone were the most easily confused;(3)Pearson correlation analysis showed a significant positive correlation between tone perception results and artificial neural network automatic recognition results(r=0.775,p=0.000,p<0.05),and they had a good consistency(ICC=0.931).Conclusions:(1)Recognized by the artificial neural network,the tone of cerebral palsy children with spastic dysarthria is abnormal,and rising tone was the most easily confused into falling-rising tone,falling-rising tone and falling tone were most easily confused;(2)The artificial neural network automatically recognizes the tone basically agrees with the results of tone perception,the former is more objective and efficient,and provides a preliminary theoretical basis for the use of artificial intelligence methods to recognize spastic dysphonia by speech recognition.