TY - JOUR
T1 - Individual differences of emotional expression in speaker's behavioral and autonomic responses
AU - Arimoto, Yoshiko
AU - Okanoya, Kazuo
PY - 2013
Y1 - 2013
N2 - The goal of this study is to elucidate differences in speak- ers' emotional expressions in behavioral and autonomic re- sponses. Verbal and non-verbal emotional behaviors of inter- locutors were recorded during two types of dialogs (competi- Tive and cooperative). Autonomic nervous system (ANS) activ- ity (heart rate and skin conductance level) was also recorded as an internal measure of emotional reactions toward an interlocu- Tor. To annotate the emotional states of speakers, the speakers who participated in the recording evaluated their own emotional states (arousal, valence and positivity) and their interlocutor's states along with the time course of the dialogs. The behavioral and autonomic emotional reactions were used as independent variables for speaker-independent and speaker-specific models to predict a speaker's emotion. The results demonstrate that speaker-independent models could explain each emotional state in a certain degree; in contrast, some speaker-specific models could explain each emotional state with moderate or high accu- racy. Moreover, a comparison of the absolute standard partial regression coefficients of each variable of the models revealed that there are two types of emotional expression styles, one in which emotional behavioral expression is dominant and another in which emotional autonomic reaction is dominant.
AB - The goal of this study is to elucidate differences in speak- ers' emotional expressions in behavioral and autonomic re- sponses. Verbal and non-verbal emotional behaviors of inter- locutors were recorded during two types of dialogs (competi- Tive and cooperative). Autonomic nervous system (ANS) activ- ity (heart rate and skin conductance level) was also recorded as an internal measure of emotional reactions toward an interlocu- Tor. To annotate the emotional states of speakers, the speakers who participated in the recording evaluated their own emotional states (arousal, valence and positivity) and their interlocutor's states along with the time course of the dialogs. The behavioral and autonomic emotional reactions were used as independent variables for speaker-independent and speaker-specific models to predict a speaker's emotion. The results demonstrate that speaker-independent models could explain each emotional state in a certain degree; in contrast, some speaker-specific models could explain each emotional state with moderate or high accu- racy. Moreover, a comparison of the absolute standard partial regression coefficients of each variable of the models revealed that there are two types of emotional expression styles, one in which emotional behavioral expression is dominant and another in which emotional autonomic reaction is dominant.
KW - Autonomic nervous system activity
KW - Emotional speech
KW - Facial expressions
KW - Spontaneous dialog
UR - http://www.scopus.com/inward/record.url?scp=84906221352&partnerID=8YFLogxK
U2 - 10.21437/interspeech.2013-297
DO - 10.21437/interspeech.2013-297
M3 - 会議記事
AN - SCOPUS:84906221352
SN - 2308-457X
SP - 1101
EP - 1105
JO - Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH
JF - Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH
T2 - 14th Annual Conference of the International Speech Communication Association, INTERSPEECH 2013
Y2 - 25 August 2013 through 29 August 2013
ER -