Paramètres acoustiques vocaux comme marqueurs du stress de parents d’enfants traités pour un trouble de stress post-traumatique
Year: 2023
Type: dissertation
Study design: Observational Study
Abstract: Background: recent literature suggests that artificial intelligence may be a promising tool for the automated assessment of an individual's psychological state, particularly through the analysis of their vocal acoustic parameters. Its application in detecting parental stress may improve tertiary prevention in the care of a child's post-traumatic stress disorder (PTSD). Method: based on an ongoing cross-sectional observational study, our research explored the relationships between various aspects of stress and variations in vocal acoustic parameters in parents of PTSD child. Parents’ stress was assessed using selfreport questionnaires (State Trait Anxiety Inventory (STAI)-state, STAI-trait, Parenting Stress Index (PSI)-SF, and Parental Sense of Competence( PSOC)), considered as independent variables. Four acoustic parameters were measured (Fundamental frequency (FO) mean, FOstandard deviation, jitter, shimmer) during three spontaneous speech vocal tasks (Neutral Vocal Task (NVT), Stressful Vocal Task (SVT), and Positive Vocal Task ( PVT)). A fourth vocal task was considered, averaging the three vocal tasks for each acoustic parameter, the Mean Vocal Task (MVT). Results and discussion: for each of the nine included subjects, we extracted 16 dependent variables corresponding to the acoustic parameters of our four vocal tasks. We calculated correlations with independent variables. There is a tendency for important parental stress to be related, in the MVT, to a decrease of FO mean and an increase of FO standard deviation, jitter and shimmer. Our results suggest that the variations in acoustic parameters of subjects with heightened parental stress were similar to the variations found in depression in literature. Comparisons of acoustic parameter variations between NVT, SVT, and PVT suggest a decrease of emotional reactivity during affectively valenced vocal tasks, decrease also described in depression in literature. Limitations and conclusion: due in part to the small sample size and a lack of standardization in the ongoing protocol, we were unable to demonstrate any significant result after p-value corrections. It seems necessary, in addition to continuing with inclusions, to establish future scientific protocols that allow the comparison of our results with those from other studies addressing vocal acoustic parameters as markers of psychological states. Subsequent research on parental stress could verify the trends described by our results and investigate potential mechanismsshared with depression.
Source: HAL
Author: Cunéo, Paul-Camille
Institution: Université Côte d'Azur
Language: English
Topic: Emotion and Mood Recognition
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7051628534
