Thesis

Computerized speech recognition technology as a tool to differentiate depression and anxiety symptomatology

Anxiety and depression are often misdiagnosed and mistreated as many sufferers, unaware of the real nature of their afflictions, seek medical assistance from health practitioners who may not recognize the underlying problem. One way to deal with this problem is to utilize state-of-the-art technology that can reliably and objectively assess these disorders. Major depression has been previously examined using computerized speech recognition (CSR), however, to our knowledge generalized anxiety disorder (GAD) has not been examined utilizing CSR in English-and Spanish-speaking participants. The current project was designed to examine a newly created measure for anxiety symptomatology to be administered via CSR in order to enhance and improve our understanding of affective disorders. Two studies were conducted to explore the psychometric properties of a voice interactive depression/anxiety assessment system (VIDAAS). The findings suggested that VIDAAS-IV in general and the new Anxiety DSM Scale (ADS) specifically demonstrated good reliability and validity properties for both English-and Spanishspeaking populations. In addition, analyses exploring the relationship between anxiety and depression revealed the importance of assessing each disorder in conjunction with the other. Finally, the analyses conducted on voice characteristics illustrate that physiological differences are not only present but assessable, making CSR a significant tool for understanding and detecting anxiety and depression.

Relationships

Items