Using a conversational agent for thought recording as a cognitive therapy task: Feasibility, content, and feedback.

Area of Science:

• Digital mental health
• Human-computer interaction
• Clinical psychology

Background:

• E-mental health interventions for depression show promise but struggle with patient adherence due to decreased therapist involvement.
• Existing autonomous systems often lack sophisticated input interpretation and feedback capabilities, hindering user engagement.
• Thought recording is a key cognitive therapy technique for understanding the link between thoughts, feelings, and behaviors.

Purpose of the Study:

• To assess the feasibility of using a conversational agent for collecting thought record data through dialogue.
• To analyze the content of thought records generated by individuals with subclinical depression.
• To investigate whether content-based feedback enhances motivation for thought recording.

Main Methods:

• 308 participants with subclinical depression symptoms were recruited via Prolific.
• Participants completed two thought recording sessions with a conversational agent: one practice and one open session.
• Randomization into three conditions with varying feedback richness was employed.

Main Results:

• Conversational agent use for thought record completion was feasible, with machine learning interpretability.
• Interpersonal situations were reported nearly three times more often than achievement-related situations.
• No significant motivational effect was found for providing richer content-based feedback.

Conclusions:

• Conversational agents represent a feasible method for gathering thought record data in digital mental health applications.
• The study highlights the prevalence of interpersonal issues in thought records of individuals with depressive symptoms.
• Future research should explore alternative motivational strategies beyond feedback richness for thought recording.