Keynote Lectures

From Signals to Meaning: Biosignals as a Language for Health, Emotion, and Connection
Antonio Lanatà, Department of Information Engineering, University of Florence, Italy, Italy

Extended Reality in Healthcare
Kavitha Anandan, SSN College of Engineering, India, India

AI in Medicine: Progress, Perils, and Perspectives
Tim Hahn, University of Münster, Germany, Germany

Neuro-Acoustic Intelligence: A Unified Phonation-Encephalography Framework for Multi-Disorder Monitoring via Explainable AI
Pedro Gómez Vilda, Independent Researcher, Spain, Spain


He Huacheng, Oujiang Laboratory, China, China

(Cancelled)

Abstract
Biosignals are not just measurements of physiology; they could represent a language of life, a continuous dialogue between body, mind, and environment. Over the past two decades, advances in sensing technologies, wearable systems, and computational intelligence have transformed biosignals into powerful tools for monitoring health, decoding emotion, and enabling novel forms of interaction.
In this keynote, I will trace a research journey from the design of wearable and textile-integrated systems for physiological monitoring, through the development of advanced signal processing and AI-based models, to the exploration of biosignals as markers of emotional and cognitive states. I will highlight results from large European projects that bridged engineering and human sciences, addressing topics such as mental health monitoring, immersive environments, and emotion-driven human–machine interaction.
A new frontier is now emerging, the use of biosignals to study synchronization and coupling not only between humans, but also across species. Recent work on human–animal interaction and interspecies emotional transfer demonstrates how biosignals can be leveraged to investigate empathy, welfare, and cross-kingdom communication. This expands the scope of biosignal research from individual monitoring to collective and ecological perspectives.
Looking toward the future, biosignals research must confront critical challenges, including ensuring reproducibility and data quality, integrating multimodal information streams, and embedding ethical and transparent AI into decision-making processes. At the same time, new opportunities are arising through hyperscanning, living labs, and real-world validation of biosignal-based systems in health, education, and society at large.
By reframing biosignals as a universal language that is quantitative, interpretable, and integrative, we can move beyond monitoring toward creating intelligent, empathetic, and sustainable systems that connect humans, animals, and their environments.

Abstract
Extended Reality (XR), is an immersive platform that encompasses Virtual Reality (VR) and Augmented Reality (AR) techniques. It provides interactive experiences by merging physical and virtual worlds, thereby allowing interaction with digital objects, enhancing learning, training, and overall understanding. XR has been extensively used in various domains and has seen applications in healthcare. XR applications in healthcare include patient treatment, surgical training, and medical education. XR offers immersive experiences that can enhance patient engagement, aid in rehabilitation, and provide new ways to train healthcare professionals. This talk intends to unfold the possibilities of employing XR in various healthcare applications. A few case based applications on using Augmented reality for spine and liver surgical systems will be discussed. Some interesting VR based environments developed for enhancing learning in children with neurodevelopmental disorders will also be explained.

Abstract
AI is driving a paradigm shift in medicine, enabling thus-far unseen data analysis capabilities and improved patient care. Here, we will provide an overview of recent projects and discuss a high-level framework for translation into clinical practice.
In the area of MRI imaging, for example, projects such as deepmriprep demonstrate how deep learning can accelerate and standardize MRI preprocessing, reducing variability and enabling large-scale studies. Multimodal MRI analyses are being used to investigate brain–behavior associations, while Bayesian frameworks add uncertainty estimates to predictions of disease trajectories, for instance in multiple sclerosis.
Within medical image segmentation, AI methods are developed to automate the delineation of complex anatomical and pathological structures. These approaches reduce reliance on manual labeling and improve reproducibility across sites, addressing a central bottleneck for clinical integration. As an example beyond classical imaging, we use WiFi-based imaging to explore how ambient signals can be used to monitor mobility and behavior continuously in a non-invasive way. This provides ecologically valid, real-world data that enrich imaging-derived biomarkers. A further line of work investigates dynamical system identification using neural networks, enabling the extraction of underlying system dynamics from biomedical time series, with applications ranging from physiology to brain activity modeling.
Challenges remain with small and heterogeneous datasets, variability in pipelines, and the ethical and regulatory complexities of translation into practice. Looking ahead, standardized workflows, uncertainty-aware modeling, and multimodal integration will be crucial. By embedding ethical and legal safeguards into technical innovation, AI in medical bioimaging can evolve responsibly and contribute to more reliable, patient-centered care.

Abstract
Current clinical diagnostics for neurodevelopmental and neurodegenerative disorders rely heavily on subjective behavioral assessments, creating a critical gap in objective, longitudinal monitoring. This plenary lecture introduces a transformative paradigm: Phonation-Encephalography (phEG). By utilizing the laryngeal motor cortex as a high-fidelity proxy for central neural drive, brain-state dynamics might be non-invasively estimated directly from vocal motor output. Central to this framework is the ushape synchronization index (uSI), a novel mathematical metric designed to quantify spectral curvature and cortical-limbic synchronization failure. While originally validated in Parkinson’s Disease (PD) to detect atypical mu-band reactivity, the extensibility of the uSI to a broader "Neurological Atlas" is proposed, such as Amyotrophyc Lateral Schlerosis (ALS), Alzheimer’s, Disease and Dementia, Clinical Depression and Anxiety, Autism Spectrum Disorder (ASD), and Schizophrenia. To meet the demands of modern biomedical engineering, these physiological markers are integrated with Advanced AI methodologies. A Physiology-Informed Deep Learning approach is proposed, where the uSI serves as a critical feature in an Explainable AI (XAI) architecture. This allows for the automated detection of pathological synchronization while maintaining clinical interpretability—an essential requirement for medical adoption. Finally, the application of temporal modeling (LSTMs/Transformers) to longitudinal datasets collected at six-month intervals is discussed, showcasing how the "variational dispersion" of the phEG signal can predict disease trajectories. This fusion of neuromechanics and artificial intelligence paves the way for a new era of "Acoustic Biopsies," enabling scalable, remote, and objective monitoring of global brain health.

Abstract
Chronic wounds are most common diseases in daily life. Dressings have been extensively used for the treatments, yet the efficacy of these dressings is still quite limited in promoting wound repair, especially in chronic wounds (e.g., diabetic foot ulcer, pressure ulcer, burns and infected wounds), due to the lack of functionalities which fail to accommodate the complicated healing stages. Therefore, exploring novel wound dressing with advanced functions to expedite wound repair is highly demanded. Herein, we devise several facile strategies and technologies to fabricate various wound dressings with excellent bioactivities including anti-inflammation, antibacterial and antioxidative effects to accelerate wound regeneration. Both in vitro and in vivo investigations demonstrate that these dressings enhanced wound healing by promoting M2 macrophage polarization, epithelial-mesenchymal transition (EMT), angiogenesis, granulation tissue formation, collagen deposition, and re-epithelialization. Now we are optimizing the manufacturing conditions of these dressings and endeavoring to launch them onto the market soon.