DI-UMONS : Dépôt institutionnel de l’université de Mons

Recherche transversale
Rechercher
(titres de publication, de périodique et noms de colloque inclus)
2018-11-29 - Colloque/Présentation - communication orale - Anglais - page(s)

Christodoulides George , Romoaldo Gilson , Delvaux Véronique , Huet Kathy , Piccaluga Myriam , Harmegnies Bernard , Blankert Bertrand , Colet Jean-Marie , "Combining Prosodic Profile and Biomarker Analysis to Study the Effects of Stress and Cognitive Load on Speech Production" in Réunion du Groupe de Contact FNRS en Psycholinguistique et Neurolinguistique, Mons, Belgium, 2018

  • Codes CREF : Psycholinguistique (DI421B), Linguistique générale (DI5310)
  • Unités de recherche UMONS : Métrologie et Sciences du langage (P362), Biologie humaine et Toxicologie (M125), Analyse pharmaceutique (M130)
  • Instituts UMONS : Institut de recherche en sciences et technologies du langage (Langage), Institut des Sciences et Technologies de la Santé (Santé)

Abstract(s) :

(Anglais) The BioVoc project aims at studying the effects of stress, mental fatigue and cognitive load on the speech signal and on biological markers. While some of these effects have been described independently, our approach is to assess the relationships between them, including by objectifying the emotional and cognitive state of the speaker using physiological observations. The project focuses particularly on high-stress and high-load environments and communicative situations, such as those arising in aviation. In this presentation, we describe the preliminary results of a study based on the Trier Social Stress Test (TSST; Kirschbaum et al. 2011). Seventeen young male students have been submitted to a modified version of the TSST protocol in order to induce psychosocial stress. Speech and electroglottographic data were recorded throughout the test, and saliva samples were collected at key time points (before and during the test, and after a resting period). We attempt to correlate observations relating to the subjects’ speech production with biomarkers for each phase of the test. With respect to speech analysis, we use a fully automated procedure to extract profiles for the prosodic and temporal organisation of speech production, including the distribution of silent pauses, speech rate and its variability, and pitch range variability, in line with previous studies (e.g. Jameson et al. 2009, Tet Fei Yap 2012, Christodoulides 2016). With respect to biomarkers, saliva samples are processed using H-NMR spectra and a multivariate analysis of the salivary metabonomic profile is performed in order to identify potential biomarkers; additionally, an analytical quantification of salivary biomarkers known to be associated with emotional states (such as 3-Methoxy-4-hydroxyphenylglycol; cf. Yang et al 1997) is performed. We discuss the merits of combining biomarker data with speech production measures in studying the effects of stress and cognitive load in speech, and we present the design of future studies in this direction.

(Anglais) The BioVoc project aims at studying the effects of stress, mental fatigue and cognitive load on the speech signal and on biological markers. While some of these effects have been described independently, our approach is to assess the relationships between them, including by objectifying the emotional and cognitive state of the speaker using physiological observations. The project focuses particularly on high-stress and high-load environments and communicative situations, such as those arising in aviation. In this presentation, we describe the preliminary results of a study based on the Trier Social Stress Test (TSST; Kirschbaum et al. 2011). Seventeen young male students have been submitted to a modified version of the TSST protocol in order to induce psychosocial stress. Speech and electroglottographic data were recorded throughout the test, and saliva samples were collected at key time points (before and during the test, and after a resting period). We attempt to correlate observations relating to the subjects’ speech production with biomarkers for each phase of the test. With respect to speech analysis, we use a fully automated procedure to extract profiles for the prosodic and temporal organisation of speech production, including the distribution of silent pauses, speech rate and its variability, and pitch range variability, in line with previous studies (e.g. Jameson et al. 2009, Tet Fei Yap 2012, -12- Christodoulides 2016). With respect to biomarkers, saliva samples are processed using H-NMR spectra and a multivariate analysis of the salivary metabonomic profile is performed in order to identify potential biomarkers; additionally, an analytical quantification of salivary biomarkers known to be associated with emotional states (such as 3-Methoxy-4-hydroxyphenylglycol; cf. Yang et al 1997) is performed. We discuss the merits of combining biomarker data with speech production measures in studying the effects of stress and cognitive load in speech, and we present the design of future studies in this direction.