Speech technology in psychotherapy: exploring transcription tools and their potential impact

Authors: Martha Schubert, Matthias Busch, Julia Krüger, Ingo Siegert

Abstract:

Psychotherapy is an inherently complex process, where keeping track of the multitude of different conversations a therapist engages in daily presents a major challenge. As a result, documenting psychotherapy by recording and transcribing sessions can provide substantial assistance to therapists by allowing them to revisit and examine what was said and experienced during the session. However, transcribing therapy sessions manually is highly time-consuming and requires a lot of resources. This is why using the current advancements of machine learning to automatically transcribe and evaluate the therapy sessions seems promising. While transcription in itself is already very helpful, more advanced analysis tools that go beyond the literal conversation contents could provide further valuable insights. In these ways, it could be possible to support the prediction of crucial therapeutic constructs, such as therapeutic alliance, motivation or symptoms, by analyzing patterns in speech and language.