Abstract Emotion regulation is a crucial skill for managing emotions in everyday life, yet finding a constructive and accessible method to support these processes remains challenging due to their cognitive demands. In this study, we explore how regular interactions with a social robot, conducted in a structured yet familiar environment within university halls and departments, can provide effective support for emotion regulation through cognitive reappraisal. Twenty-one students participated in a five-session study at a university hall or department, where the robot facilitated structured conversations, encouraging the students to reinterpret emotionally charged situations that they shared with the robot. Quantitative and qualitative results indicate significant improvements in emotion self-regulation, with participants reporting better understanding and control of their emotions. The intervention led to significant changes in constructive emotion regulation tendencies and positive effects on mood and sentiment after each session. The findings also demonstrate that repeated interactions with the robot encouraged greater emotional expressiveness, including longer speech disclosures, increased use of affective language, and heightened facial arousal. Notably, expressiveness followed structured patterns aligned with the reappraisal process, with expression peaking during key reappraisal moments, particularly when participants were prompted to reinterpret negative experiences. The qualitative feedback further highlighted how the robot fostered introspection and provided a supportive space for discussing emotions, enabling participants to confront long-avoided emotional challenges. These findings demonstrate the potential of robots to effectively assist in emotion regulation in familiar environments, offering both emotional support and cognitive guidance.
Introduction
University students frequently face high-stress environments that tax their cognitive resources, making it difficult to apply constructive emotion regulation strategies like cognitive reappraisal. Reappraisal involves reframing a negative experience to change its emotional impact, a process that is often cognitively demanding during a crisis. Traditional human-led interventions, while effective, can be hindered by social stigma, availability, and the anxiety of opening up to another person. This study investigates whether a social robot can serve as a scalable, non-judgmental mediator to guide students through the reappraisal process in a familiar setting.
Methods
The researchers conducted a longitudinal study with 21 students over five sessions. The intervention utilized the QTrobot (LuxAI), a humanoid social robot. The robot's conversational capabilities were powered by GPT-3.5, integrated via an open-source framework, using AWS Polly for speech synthesis and Google Cloud Speech-to-Text for input.
Intervention Logic
The sessions were structured around the PERMA model (Positive Emotions, Engagement, Relationships, Meaning, and Accomplishments). The interaction logic followed a specific sequence designed to build a "resource bank" before tackling negative experiences.
As shown in the logic flow, the robot first asked "Positive Connotation Questions" (PCQn). It then transitioned to a "Negative Connotation Question" (NCQ) regarding a current challenge. Finally, the LLM generated a Reappraisal Prompt (CR), specifically instructed to "bridge" the negative experience with the positive elements shared earlier in the session to encourage reframing.
Key Results
Improved Self-Regulation and Mood
Quantitative analysis using Linear Mixed-Effects Models (LME) revealed significant improvements in students' emotional self-regulation. Specifically, participants reported a significant increase in their "Understanding of the Situation" and "Feeling in Control" over the five sessions. Additionally, participants reported a significant improvement in mood immediately following each session.
The "Reappraisal Wave"
Using Generalized Additive Mixed Models (GAMM), the researchers identified a non-linear pattern in emotional expression. Expressiveness, measured by speech duration and adjective density, peaked precisely when the robot prompted the student to reframe a negative experience. This suggests the robot successfully triggered deep cognitive processing during the reappraisal phase.
Qualitative Feedback and the "Anonymity Advantage"
Participants highlighted the robot's lack of judgment as a primary benefit. Many reported that the robot provided a "safe space" for introspection, allowing them to confront emotional challenges they had previously avoided. The robot was perceived as a supportive resource rather than a social evaluator, which reduced the social anxiety often associated with human-led therapy.
Discussion and Limitations
The study demonstrates that LLMs are effective at "affective labeling" and synthesis—weaving disparate positive thoughts into a coherent counter-narrative for negative events. However, while the intervention increased constructive habits (reappraisal, acceptance), it did not significantly decrease maladaptive habits like rumination.
Future work should include control groups to isolate the effects of the specific reappraisal logic and explore whether these skills persist long-term. Additionally, tailoring LLM prompts to specifically target and reduce maladaptive strategies could further enhance the intervention's efficacy.
Summary updated