Beyond Microexpressions: Subtle and Dynamic Affective Signals in Embodied Human-Robot Interaction

BackgroundEmbodied agents are moving from controlled laboratories into homes, schools, clinics, workplaces, and public spaces, where affect is often communicated through signals that are brief, low-intensity, ambiguous, and distributed across the body. Microexpressions are a prominent example, but subtle affect also appears in gaze shifts, facial action dynamics, vocal prosody, speech disfluency, posture, gesture, interpersonal distance, touch, physiology, and interpersonal synchrony. These cues unfold at different timescales and acquire meaning from the task, environment, culture, individual, and interaction history. Systems trained on posed expressions or isolated modalities therefore transfer poorly to natural interaction. Embodiment creates both a challenge and an opportunity: an agent must perceive faint signals while its own appearance and behavior continuously shape the signals it observes.GoalThis Research Topic will establish a broader science of subtle and dynamic affective signaling for embodied interaction, using microexpression research as a methodological foundation while moving beyond the face and beyond recognition alone. It will connect reliable signal capture, temporal and multimodal modeling, contextual interpretation, and adaptive embodied response. Central questions include how weak signals can be separated from noise and ordinary movement; how evidence across modalities and timescales should be fused; how uncertainty, individual variability, culture, neurodiversity, and masking should be represented; and how an agent can respond without overclaiming access to a person’s internal state. We seek work that links computational performance to rigorous human-subject evidence and full perception-action-feedback loops. The collection will promote ecologically valid datasets, reproducible protocols, transparent reporting, and evaluation of downstream effects on trust, comfort, task performance, agency, and well-being, enabling affect-sensitive embodied systems that are useful, cautious, and socially appropriate.Scope and Information for AuthorsWe welcome Original Research, Methods, Technology and Code, Data Reports, Reviews, Systematic Reviews, Perspectives, and Hypothesis and Theory articles on: spontaneous microexpressions and facial action dynamics; subtle cues in gaze, voice, language, posture, gesture, proximity, touch, physiology, and interpersonal synchrony; multimodal fusion and temporal modeling across micro-to-longitudinal timescales; uncertainty-aware, explainable, few-shot, and personalized inference; active perception and sensing shaped by robot motion or morphology; affect-sensitive behavior generation, feedback, and co-regulation; datasets, elicitation protocols, benchmarks, and reproducibility standards; real-world, longitudinal, cross-cultural, neurodiversity-aware, and participatory evaluation; and privacy, consent, bias, deception, manipulation risk, and user control. Applications may include education, health support, care, work, public service, and collaborative robotics. Virtual agents are in scope when embodiment or sensorimotor contingencies are central. Papers limited to posed-expression classification without an embodied interaction question are outside the focus. Topic editors: Jingting Li, Su-Jing Wang, Xinqi Fan Source: https://www.frontiersin.org/research-topics/86615/beyond-microexpressions-subtle-and-dynamic-affective-signals-in-embodied-human-robot-interaction