Behavior and Emotion Analysis through wearable Technology

Yusuke Yokota is a Neuroscientist at VIE Inc., Japan. He received his Ph.D. in Engineering from Toyohashi University of Technology. Following his doctoral studies, he joined the National Institute of Information and Communications Technology (NICT), where he specialized in the development of electroencephalography (EEG) systems and conducted brainwave measurement experiments in real-world environments. His research has led to significant findings in Error-Related Potentials (ErrP) and Auditory Steady-State Responses (ASSR). Currently, at VIE Inc., he leads research and development initiatives focused on Gamma Music and its diverse neuro-applications.

Title: Wearable Neurotechnology for Real-World Sensing: Introducing the VIE EEG Headphone and Gamma Music Intervention

Abstract: While computer vision excels at capturing observable behaviors, understanding a user’s internal state requires directly monitoring their neural dynamics. Electroencephalography (EEG) captures neural processing with high temporal resolution. For instance, analyzing Event-Related Potentials (ERPs) like the N170 allows us to objectively observe how the brain perceives faces. Historically, capturing such high-fidelity ERPs was strictly confined to controlled laboratories. To overcome this limitation, we developed the VIE EEG Headphone—a proprietary, wearable device designed to bring research-grade brain monitoring into real-world environments. This keynote will feature a live demonstration of its real-time neural feature extraction capabilities. Furthermore, we will present our recent findings on Gamma Music. By shifting from merely “reading” the brain to actively “enhancing” cognitive performance, we aim to demonstrate how wearable neurotechnology will shape the future of human-computer interaction and mental health care.

Zilu Liang, Kyoto University of Advanced Science (KUAS), Japan, is an Associate Professor at KUAS, where she leads the Ubiquitous and Personal Computing Lab. She received her Ph.D. and M.Sc. in Electrical Engineering and Information Systems from the University of Tokyo. Before joining KUAS, she held research and academic positions at the University of Tokyo, the University of Melbourne, the University of Oxford, and Imperial College London. Her research focuses on human-centered computing, wearable and ubiquitous technologies, and AI-driven methods for understanding and supporting human behavior.

Title: Frictionless by Design, Bounded by Ethics: Rethinking In-the-Wild Data Collection with Wearable Sensing

Abstract: Wearable sensing research has long struggled with the trade-off between experimental control and ecological validity. As wearable technologies become increasingly integrated into everyday life, we now have an opportunity to rethink how behavioral data can be collected in ways that better reflect people’s natural experiences and behaviors in the real world. This talk first introduces a frictionless-first philosophy for in-the-wild wearable sensing. Instead of relying on rigid protocols and researcher-dependent sessions, emerging systems increasingly leverage consumer smartwatches not only for passive physiological sensing, but also for lightweight self-reporting directly on the wrist. Our research further draws inspiration from behavioral nudging and serious games to reduce reporting burden while improving motivation and long-term participation. Although longitudinal evidence is still emerging, these approaches have strong potential for improving ecological validity, participant authenticity, and sustained engagement in real-world behavioral research. Alongside these advances, however, wearable sensing faces another challenge that is fundamentally ethical rather than technological. Many target events are inherently rare in natural settings, such as emotional crises, thermal stress, wound formation, or acute health deterioration. Researchers may therefore attempt to induce these events in semi-controlled environments to obtain labeled data for AI training and validation. Yet in many safety-critical contexts, doing so would expose participants to the very harm the sensing system is intended to prevent. The second part of the talk introduces the concept of “data non-collectable” — ground-truth labels that remain ethically inaccessible because the target condition itself cannot be responsibly induced. This creates an emerging ethical ceiling for AI validation in wearable and behavioral sensing research. The talk concludes by discussing alternative pathways including opportunistic clinical collaboration, longitudinal passive sensing, and precursor-focused annotation strategies.

The field of wearable and behavioral technology has become remarkably effective at collecting more data, more continuously, and more unobtrusively than ever before. Yet its next defining challenge may not be sensing itself, but learning how to build intelligent systems around phenomena we cannot ethically induce.

We welcome submissions to two tracks:

• Main Track (Original Research): This track accepts original and unpublished work of 4 to 8-pages (excluding references). Submissions should follow the template of the main FG conference: the formatting guidelines and templates are available on the IEEE FG 2026 website in the format section: https://fg2026.ieee-biometrics.org/cfp/. All submissions will undergo a double-blind review process, with three reviewers assigned to each paper. Accepted papers will be included in the workshop proceedings, and presented as posters or oral presentations.

In the case of main track submissions, we do not consider a paper on arXiv.org and other open repositories as a dual submission. However, the papers deposited to paid-access repositories (such as ResearchGate, Academia) will not be accepted.

• Non-Archival Track (Published Work): Authors of papers previously published in other conferences or journals are invited to submit a 1-page summary to present their work, which will not be included in the proceedings. No specific template is required, but authors must properly cite their original publication. These submissions will be evaluated and selected by the organizers based on quality and relevance to the workshop topics.

Topics of interest include, but are not limited to:

• Machine Learning and computational models for movement and physiological wearables
• Resource efficient and lightweight models
• Multimodal fusion and synchronization strategies
• Methods for irregularly sampled or missing data
• Individual differences, personalization and context-awareness
• Ethical and privacy-preserving AI in wearable systems
• Novel wearables and applications
• Experimental methods for validation of wearable systems
• Lab-controlled experiments and In-the-wild deployment
• Datasets and Benchmarks
• Responsible data management and user consent
• Applications in Affective Computing / Mobile Health / Action Recognition / Social Interaction / HRI