This study investigates speech and voice features as classifiers for depression, schizophrenia, and healthy controls. Speech and voice features were calculated from picture descriptions of 240 speech samples (20 participants with SSD, 20 with MDD, and 20 HC each with 4 samples). Support vector machine (SVM) models achieved high accuracy (0.947 for HC vs. SSD, 0.920 for HC vs. MDD, and 0.932 for SSD vs. MDD). The most important features were articulation coordination, pauses per minute, and speech variability, showing moderate correlations with positive symptoms for SSD. These findings suggest differences in speech characteristics related to psychomotor slowing, alogia, and flat affect among the groups.
Publisher
Translational Psychiatry
Published On
Sep 19, 2023
Authors
Mark Berardi, Katharina Brosch, Julia-Katharina Pfarr, Katharina Schneider, Angela Sültmann, Florian Thomas-Odenthal, Adrian Wroblewski, Paula Usemann, Alexandra Philipsen, Udo Dannlowski, Igor Nenadić, Tilo Kircher, Axel Krug, Frederike Stein, Maria Dietrich
Tags
speech features
voice analysis
mental health
depression
schizophrenia
classification
SVM models
Related Publications
Explore these studies to deepen your understanding of the subject.
