Browsing by Subject "natural language"

Now showing 1 - 1 of 1

Computational Methods in Psychotherapy: Prosodic and Semantic Analysis Framework for Clients with Major Depression
(2022-05) Cruz, John Henry; Pennebaker, James
With increased awareness of mental health disorders, the growing practice of psychotherapy was followed by the development of therapy models. These practices have shown themselves to be effective in treating mental health disorders through analysis of different markers seen throughout a therapy session. But still, most of this practice is heavily dependent on decisions made by a therapist, which, to a degree, are subjective. As further developments are seen in computer science, tools have been created with the purpose of analyzing natural language in the form of text and audio. Due to the development of models aimed to understand human language and emotion, we can bridge together text and audio tools along with a well-experienced therapist to develop a therapy analysis computer program. I conducted a literature review to propose a computer program that would focus on the objective features, specifically text and audio, seen in clients in therapy sessions. This could aid therapists in the decision-making process for the treatment of major depression alongside their expertise. Preliminary results include a set of metrics that measures indicators of depression, such as lower speaking frequencies and slower talking speeds. In addition to the literature review, I conducted a needs assessment survey to gauge opinions on potential metrics and the use of the computer program from those currently involved in therapy. Survey results reflected the sentiment that none of the proposed metrics would be useful in reflecting changes in depression while mixed opinions were seen for the computer programs used in assessing depression. Dialogue revolving around the use of technology in psychotherapy could potentially grow in response to the proposed computer program.