The main aim of the present study is to investigate the effects of a Motivational Interviewing-based artificial intelligence chatbot on belief updating in college students with elevated levels of depression and anhedonia by combining a randomized active-control intervention design with pre- and post-intervention behavioral belief updating task assessments.
Anhedonia represents a core characteristic of depression and is characterized by reduced experience of pleasure. It is closely related to decreased motivation, altered reward processing, changes in affective responsiveness, and alterations in intrinsic brain network function. Anhedonia is not specifically targeted by currently available pharmacological interventions. Initial evidence indicates that an increased willingness to change and implementation of change in daily life can alleviate anhedonia. The present study aims to examine whether a Motivational Interviewing-based AI chatbot can lead to changes in belief updating in college students with elevated anhedonia and depressive symptoms. Belief updating is included because anhedonia and depressive symptoms may be associated with biased or inflexible use of new information, which may influence future expectations, motivation, and engagement in rewarding activities. The belief updating task allows the study to examine how participants revise their expectations after receiving new information. To this end, eligible participants with a total score of 22 or higher on the Snaith-Hamilton Pleasure Scale and a score of 14 or higher on the Beck Depression Inventory will undergo a randomized, between-subjects, active-control intervention study. Participants will be assigned to either a Motivational Interviewing-based chatbot group or an active control chatbot group for 1 week. Pre- and post-intervention assessments will include self-report questionnaires and a behavioral belief updating task to examine psychological and behavioral effects of the intervention.
Study Type
INTERVENTIONAL
Allocation
RANDOMIZED
Purpose
BASIC_SCIENCE
Masking
DOUBLE
Enrollment
80
The experimental chatbot is designed to use principles of Motivational Interviewing to support participants in exploring their personal values, motivation for change, and daily behavioral goals related to pleasure, engagement, and reward-seeking. During the intervention period, participants will interact with the chatbot regularly through brief text-based conversations. The chatbot will provide empathic, non-judgmental responses, encourage reflection on current difficulties, and help participants identify small, feasible actions that may increase daily engagement and positive experiences. It will not provide diagnosis, crisis counseling, or medical treatment.
Participants will interact with a chatbot matched in format and frequency of use. This chatbot will provide neutral nature-related stories or general natural history content. It will be designed to maintain participant engagement while avoiding therapeutic techniques, motivational interviewing strategies, behavioral activation guidance, or personalized mental health advice. This active control condition will help control for nonspecific effects of chatbot interaction, attention, expectancy, and digital engagement.
University of Electronic Science and Technology of China
Chengdu, Sichuan, China
RECRUITINGValence-dependent belief updating bias
Belief updating magnitude will be calculated separately for positive and negative information conditions in the belief updating task before and after the intervention. Updating magnitude will be quantified based on changes in probability estimates before and after the presentation of new information. The difference in updating magnitude between positive and negative information conditions will be calculated and compared from baseline to post-intervention.
Time frame: Baseline before the first chatbot interaction and Week 1 after completion of the chatbot intervention.
Expectancy bias for life events
Before receiving base-rate information or new information, participants will provide initial probability estimates for positive and negative life events at baseline and post-intervention. Mean initial estimates will be calculated separately for the two event types, and pre-to-post changes in the difference between positive and negative event estimates will be compared.
Time frame: Baseline before the first chatbot interaction and Week 1 after completion of the chatbot intervention
This platform is for informational purposes only and does not constitute medical advice. Always consult a qualified healthcare professional.