Statistical approach for the emotional speeches of Parkinson Dysarthria
Dysarthria is a condition where the speech muscles become weak and difficult to control. It leads to major deficiencies in human speech. The dysarthria speakers with Parkinson disease (PD) suffer from slurred and slow voice. It becomes difficult to understand by others. In this study we conduct a statistical analysis on the six basic emotions (anger, disgust, fear, sadness, happiness, surprise) from the speech records of various dysarthria speakers. As a method to extract the features, we use short term spectral analysis (STSA) to extract the major prosodic features from the speech samples. This method is widely used in the field of speech recognition and feature extraction. We derive hidden Marcov models (HMMs) for each prosodic feature. Using the HMMs, we determine the coefficients in order to construct a multivariate linear regression model. The aim of this statistical analysis is to validate the para-linguistic information of every speech samples of PD with the speech samples of healthy controls (HC). We also discuss the role of the basal ganglia in the processing of emotional speech samples.

footer 著作権について 倫理綱領 プライバシーポリシー セキュリティ 情報処理学会