计算机语音分析系统对全口义齿初戴前后辅音的声学分析(英文)

Computer analysis of acoustic features of consonants before and after restoration with complete denture

背景:传统的音声检测方法是对异常语音特征的主观描述,或对元音进行定量分析,对敏感音辅音只作定性的观察。目的:对全牙列缺失患者全口义齿初戴前后敏感的辅音进行记录,定量分析其辅音的音声特征。设计:随机对照实验。单位:上海第二医科大学附属第九人民医院口腔修复科,口腔颌面外科。对象:选择上海第二医科大学附属第九人民医院口腔修复科2001-11/2003-03收治的全牙列缺失患者30例,按义齿基托材料的不同分为塑料基托组、金属基托组,各15例,两组患者性别、年龄等基线资料无明显差异。纳入标准:无全口义齿修复史,年龄45～75岁,有一定文化程度,能讲普通话且无明显方言,无明显听力障碍(≤20dB),未接受过语音治疗训练。方法:塑料基托组:塑料基托厚2.0～2.5mm,腭前区磨光面光滑;金属基托组:金属基托厚1.0～1.5mm,覆盖整个腭部,并形成腭皱形态,在牙槽嵴腭侧移形为网状基托,颊侧为塑料基托。基托范围、义齿后堤区、前牙覆牙合、覆盖度、牙合平面、面下1/3垂直距离均按常规制作。两组全牙列缺失患者在全口义齿初戴前、初戴时、初戴后1,2,4,8周分别进行测试被检音/zi/、/ji/,使用计算机语音分析系统,利用线形预测技术和快速傅立叶转换,记录敏感音的第1,3共振峰,计算机语音分析采样频率为10000Hz,线形预测技术的阶数为12。主要观察指标:被检音为/zi/、/ji/时两组的声学参数差异。结果:按意向处理分析,实验纳入30例患者均进入结果分析。①被检音为/zi/时,与初戴前比较,塑料基托组和金属基托组第1共振峰值大多在义齿初戴后逐渐降低,以戴后8周最为明显[(353.67±35.65),(322.60±44.03);(332.27±37.10),(302.20±40.39);P均<0.05];而第3共振峰在两组变化趋势不同,在塑料基托组逐渐上升,在金属基托组逐渐下降。②被检音为/ci/时,两组第1共振峰值的变化与被检音为/zi/时基本相同,第3共振峰在两组的变化亦呈现不同趋势。结论:减少全口义齿基托厚度、重建腭皱形态有助于牙列缺失患者语音功能的恢复,计算机语音分析系统可对全口义齿初戴前后患者的语音变化特征进行定量分析。

BACKGROUND： Traditional acoustic examination is to make subjective descriptions of abnormal acoustic features, or carry out quantitative analysis of vowels with sensitive consonant utterance only subjected to qualitative analysis. OBJECTIVE： To make quantitative analysis of the acoustic features of sensitive consonants before and after restoration with complete denture. DESIGN： Randomized controlled experiment. SETTING： Department of Prosthodontics; Department of Oral ＆ Maxillofacial Surgery, Ninth People＇s Hospital of Shanghai Second Medical University. PARTICIPANTS： Totally 30 edentulous patients who received treatmentin the Department of Prosthedontics, Ninth People＇s Hospital of Shanghai Second Medical University, between November 2001 and March 2003-03 were enrolled in this study. They were divided into plastic base group and metal base group with 15 patients in each group according to the denture base; the two groups of patients were matched in baseline data such as sex and age. Inclusion criteria： had no history of denture repairing, aged 45-75 years, had received certain years of education, could speak Mandarin without obvious dialect, had no obvious dysaudia （≤ 20dB）, and had not received acoustic therapeutic training. METHODS： Plastic base group： the plastic base was 2.0-2.5 mm thick, with smooth anterior palate polished surface. Metal base group： the metal base was 1.0-1.5 nun thick, covering entire palate and forming palatal rugae, which transformed to net base at alveolar ridge side, and cheek base was plastic. The scope of base, the posterior bank of denture, overbite of anterior teeth, coverage degree, occlusal surface and 1/3 vertical range beneath surface were designed conventionally. Utterance of/zi/and/ji/by edentulous patients was detected before restoration, at the beginning and after restoration 1, 2, 4, 8 weeks respectively, with the results analyzed with the assistance of computer acousticanalysis system, the linear forecast technology and fast Fourier transformation were used to record the 1^st and 3^rd formants, the computer acoustic analytic sampling frequency was 10 000 Hz and the step number of linear forecast technology was 12. MAIN OUTCOME MEASURES： The differences in acoustic parameters between the detected utterances/zi/and/ji/. RESULTS： According to the intention-to-treat analysis, 30 patients were enrolled in this experiment and all of them entered the result analysis. ① When utterance/zi/was detected, most of 1st formants were found gradually reduced in plastic base group and metal base group after restoration compared to that before restoration, most obviously at week 8 [（353.67±35.65）, （322.60±44.03）; （332.27±37.10）, （302.20±40.39）; P 〈 0.05]. However, the 3^rd formant variance was different between the two groups, presented by gradually ascending in plastic base group and gradually descending in metal base group. ② When the detected utterance was/ci/, changes of the 1^st formants in two groups were basically the same as /zi/; moreover the changes of the 3^rd formants also displayed different tendency in the two groups. CONCLUSION： Reducing the basal thickness of complete denture and reconstructing palatal rugae will be helpful for the acoustic functional restoration in patients with dentition defects. The computer acoustic analysis system can be used for quantitative analysis of acoustic features of patients before and after restoration with complete denture.