Critical Analysis of the Impact of Glottal Features in the Classification of Clinical Depression in Speech

Abstract

The motivation for this work is in an attempt to rectify the current lack of objective tools for clinical analysis of emotional disorders. This study involves the examination of a large breadth of objectively measurable features for use in discriminating depressed speech. Analysis is based on features related to prosodics, the vocal tract, and parameters extracted directly from the glottal waveform. Discrimination of the depressed speech was based on a feature selection strategy utilizing the following combinations of feature domains: prosodic measures alone, prosodic and vocal tract measures, prosodic and glottal measures, and all three domains. The combination of glottal and prosodic features produced better discrimination overall than the combination of prosodic and vocal tract features. Analysis of discriminating feature sets used in the study reflect a clear indication that glottal descriptors are vital components of vocal affect analysis.

A comparative study of different classifiers for detecting depression from spontaneous speech

“…Since the durations differ for each subject, which may affect the comparison results, we only used an equal amount of speech data from each subject (92s) in this paper. We acknowledge that the amount of data used here is relatively small, but this is a common problem [5,14] in similar studies. As we continue to collect more data, future studies will be able to report on a larger dataset.…”

“…Moreover, the research found that formants are a feature significantly distinguishing depressed from non-depressed speech [4,5], with a noticeable decrease in the second formant for depressed compared to healthy controls [4]. There is convincing evidence that sadness and depression are associated with a decrease in loudness and energy [6].…”

Contribución al estudio de selección de parámetros para identificación de estrés en la voz

“…Un locutor muestra estrés en su habla de diferente forma cuando produce una aseveración que concuerda con su punto de vista real que en el caso de tener que elaborar un discurso ficticio para parecer creíble cuando expresa una opinión contraria a su punto de vista. Este estrés se manifiesta en su calidad de fonación [125].…”

“…Otras aproximaciones consideran la vibración de las cuerdas vocales incluyendo los parámetros del ciclo glotal [90], [124], como la fase de duración de la apertura y cierre de las cuerdas vocales, los ratios entre la fase de apertura y cierre del ciclo glotal completo, el jitter glotal y el shimmer glotal [125], [196].…”

“…En la sección 3.2 del capítulo 3 se detallaron los parámetros clásicos utilizados en la literatura. Sin embargo, no hay que olvidar nuevas fuentes de inspiración, como por ejemplo las ofrecidas por Airas y Alku en [1] que analizan el flujo glotal en segmentos de vocales o el trabajo de Moore et al que analizan las características glotales en pacientes diagnosticados con depresión [125]. La depresión es un estado anímico o emocional que conlleva unos síntomas asociados tales como, apatía, tristeza, etcétera.…”