Telling the story and re-living the past: How speech analysis can reveal emotions in post-traumatic stress disorder (PTSD) patients


Share/Save/Bookmark

Broek, Egon L. van den and Sluis, Frans van der and Dijkstra, Ton (2011) Telling the story and re-living the past: How speech analysis can reveal emotions in post-traumatic stress disorder (PTSD) patients. In: Sensing Emotions: The impact of context on experience measurements. Philips Research Book Series, 12 . Springer Science+Business Media B.V., Dordrecht, The Netherlands, pp. 153-180. ISBN 9789048132577

[img] PDF
Restricted to UT campus only

230kB
Abstract:A post-traumatic stress disorder (PTSD) is a severe stress disorder and, as such, a severe handicap in daily life. To this date, its treatment is still a big endeavor for therapists. This chapter discusses an exploration towards automatic assistance in treating patients suffering from PTSD. Such assistance should enable objective and unobtrusive stress measurement, provide decision support on whether or not the level of stress is excessive, and, consequently, be able to aid in its treatment. Speech was chosen as an objective, unobtrusive stress indicator, considering that most therapy sessions are already recorded anyway. Two studies were conducted: a (controlled) stress-provoking story telling (SPS) and a(n ecologically valid) re-living (RL) study, each consisting of a “happy” and an “anxiety triggering” session. In both studies the same 25 PTSD patients participated. The Subjective Unit of Distress (SUD) was determined as a subjective measure, which enabled the validation of derived speech features. For both studies, a Linear Regression Model (LRM) was developed, founded on patients’ average acoustic profile. It used five speech features: amplitude, zero crossings, power, high-frequency power, and pitch. From each feature, 13 parameters were derived; hence, in total 65 parameters were calculated. Using LRMs, respectively 83 and 69% of the variance was explained for the SPS and RL study. Moreover, a set of generic speech signal parameters was presented. Together, the models created and parameters identified can serve as the foundation for future artificial therapy assistants.
Item Type:Book Section
Faculty:
Electrical Engineering, Mathematics and Computer Science (EEMCS)
Research Group:
Link to this item:http://purl.utwente.nl/publications/78442
Official URL:http://dx.doi.org/10.1007/978-90-481-3258-4_10
Export this item as:BibTeX
EndNote
HTML Citation
Reference Manager

 

Repository Staff Only: item control page

Metis ID: 279682