Authors:
Shohei Kato
1
;
Yuta Suzuki
1
;
Akiko Kobayashi
2
;
Toshiaki Kojima
2
;
Hidenori Itoh
1
and
Akira Homma
3
Affiliations:
1
Nagoya Institute of Technology, Japan
;
2
Ifcom. Co. and Ltd., Japan
;
3
Tokyo Dementia Care Research and Training Center, Japan
Keyword(s):
Early detection of dementia, Speech prosody, Acoustic analysis, Feature selection, Speech prosody-based cognitive impairment rating.
Related
Ontology
Subjects/Areas/Topics:
Acoustic Signal Processing
;
Applications and Services
;
Biomedical Engineering
;
Biomedical Signal Processing
;
Computer Vision, Visualization and Computer Graphics
;
Medical Image Detection, Acquisition, Analysis and Processing
;
Speech Recognition
Abstract:
This paper presents a novel approach for early detection of cognitive impairment in the elderly. Our approach incorporates the use of speech sound analysis and multivariate statistical techniques. In this paper, we focus on the prosodic features of speech. Fifty six Japanese subjects (22 males and 34 females between the ages of 64 and 90 years) participated in this study. Blind to clinical groups, we collected speech sounds from segments of dialogue during an HDS-R examination. The segments corresponds to speech sounds from answers to questions about time orientation and number backward counting. Ninety eight prosodic features were extracted from each of the speech sounds. These prosodic features consisted of spectral and pitch features (13), formant features (61), intensity features (22), and speech rate and response time (2). These features were refined by principal component analysis and/or feature selection. In addition, we calculated speech prosody-based cognitive impairment rat
ing (SPCIR) by multiple linear regression analysis. The results indicate that a moderately significant correlation exists between the HDS-R score and the synthesis of several selected prosodic features. Consequently, the adjusted coefficient of determination (R2 = 0.57) suggests that prosody-based speech sound analysis could potentially be used to detect cognitive impairment in elderly subjects.
(More)