Marcadores vocais que discriminam pacientes com e sem depressão

• Ano de defesa: 2019
• Autor(a) principal: Silva, Wegina Jordâna Nascimento da
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso embargado
• Idioma: por
• Instituição de defesa: Universidade Federal da Paraíba; Brasil; Medicina; Programa Associado de Pós Graduação em Fonoaudiologia (PPgFon/UFPB/UFRN/UNCISAL); UFPB
• Palavras-chave em Português: Depressão; Saúde mental; Acústica; Voz; Biomarcador; Depression; Mental health; Acoustics; Voice; Biomarker; CNPQ::CIENCIAS DA SAUDE::FONOAUDIOLOGIA
• Link de acesso: https://repositorio.ufpb.br/jspui/handle/123456789/19283

This dissertation investigated through theoretical and empirical approaches vocal markers in patients with and without depression. The manuscript is divided into two articles: An integrative review that sought to verify the use of vocal measures in the detection of depression and to discuss the methodological evidences of the published studies and an empirical study that sought to analyze acoustic parameters of the voice as predictors and vocal discriminants in patients with and without diagnosis of depression. The theoretical study was analyzed from 79 articles and showed that vocal measures, mainly extracted by acoustic analysis, can be useful to detect depression, predict the diagnosis, discriminate between patients with and without depression, and monitor the response to a treatment. The empirical study was carried out with 144 volunteers, 54 with depression and 90 without depression, from which acoustic parameters of medium, fashion and fundamental frequency standard deviation, Jitter, shimmer, Glotal to Noise Excitation-GNE, Cesptral Peak Prominence-Smoothed-CPPS and spectral decline were extracted. It was concluded that the SD of f0, jitter and shimmer had high values whereas GNE, CPPS and spectral decline reduced values discriminating between depressive and healthy. There was a significant association between BDI-II with jitter, shimmer, CPPS and spectral decline, as well as between CPPS and the class of antidepressants. The multiple linear regression model confirmed that jitter and CPPS are clinical predictors of depression through BDI-II. These results highlight the existing literature and prove that voice can be configured as a physiological marker of depression capable of discriminating patients with and without depression.