A02 Natural Language

Quantifying words and unravelling the process of psychotherapy

Communications in a medical setting involve multi-layered information is shared conveyed language such as empathic responding, sharing and learning experts’ knowledge.
Our research unit aims to develop the field of quantifying quality medical communication in terms of how words are used.

Yoshitake Takebayashi
Associate professor, Fukushima Medical University, School of medicine
Link to members

Our Project

In our unit, we apply natural language processing to linguistic data obtained from various communications in medical settings. These include roleplay transcripts from a seminar and individual or group psychotherapy transcripts and words written in worksheets.
Applying the state-of-the-art natural language processing to these data, we will extract the markers associated with active listening, empathy, and other foundational factors of medical communications, as well as markers that can inform the predicted symptom improvements.
Subsequently, we are going to use these markers to create a computer program that can automatically assess whether psychotherapy is carried out appropriately and in the intended manner based on the used languages.
In our final year, we aim to integrate our products with the findings from other research units such as the Acoustic Analysis Unit to build a communication predictive model with high definition.