Developing open-source ressources for Danish speech technology

A generalizable speech emotion recognition model reveals depression and remission

This paper is the outcome of my internship at Roche. We analysed voice recordings of patients before and after remission from depression, and found that a speech emotion recognition model was able to accurately discriminate contorls from depression, and those in remission from those with depression.

The PSYchiatric clinical outcome prediction (PSYCOP) cohort: leveraging the potential of electronic health records in the treatment of mental disorders

In this paper we present the PSYCOP cohort and present our group's upcoming studies.