Eye Tracking in the Age of AI: Explainability, Prediction, and Applications
the chair
1st Journal of Eye Movement Research (JEMR) Webinar
Eye Tracking in the Age of AI: Explainability, Prediction, and Applications
Welcome to the first of three webinars accompanying the Special Issue Eye Tracking in the Age of AI: Explainability, Prediction, and Applications, published in the Journal of Eye Movement Research.
Eye tracking has changed its role. For most of its history it was an instrument of description, recording where attention had already gone. With the arrival of AI-based modelling it has become an instrument of anticipation, estimating from an image alone what a person will look at, how much effort a scene will demand, and what is likely to remain once they have looked away. That shift has been rapid, and adoption has moved faster than the methodological questions it raises.
Two of those questions are put to this session. The first concerns the measure itself. If a model is trained on a metric that is not sufficiently valid, sensitive, specific, and reliable, it will predict confidently and be wrong, and the error will be invisible in the accuracy score. The opening presentation (Dr. Thomas Zoëga Ramsøy) addresses what it takes to treat AI-driven measurement as a scientific instrument rather than as software trusted because it is fast.
The second concerns what is being predicted in the first place. Attention and memory are routinely reported as a single indicator of effectiveness, although they are separable constructs, and the models that made both predictable were trained almost entirely on flat images viewed on screens, while the decisions they inform increasingly concern physical space. The second presentation (Dr. Hedda Martina Šola) asks which construct predictive models are actually answering for, and what a benchmark would have to specify before their predictions can be relied upon beyond the screen.
Taken together, the two talks address the same problem from either end: what a measure must satisfy before it is modelled, and what construct the model is answering for.
This Special Issue was opened to bring exactly these questions into the literature, alongside work on robustness, benchmarking, transparency, ethics, and the generalizability of predictive models across contexts and populations. Submissions remain open until 20 November 2026, and contributions are welcome from every domain represented in this field. I look forward to the discussion.
Dr. Hedda Martina Šola
Guest Editor and Chair
Date: September 14, 2026
Time: 3:00 pm CEST | 9:00 am EDT
Webinar Secretariat: journal.webinar@mdpi.com
Flyers
- Dr_Sola_PPT for webinar.pdf
Meet Our Speakers
Institute for Neuromarketing and Intellectual Property, Independent Research Institute, Croatia;
Dr. Hedda Martina Šola is Director of the Institute for Neuromarketing and Intellectual Property in Zagreb, a research institution that measures and predicts consumer behaviour using high-end neuroscientific technology and its own proprietary protocols. She holds a PhD in Economics with a specialisation in neuromarketing and brings more than thirty years in strategic marketing and over fifteen in consumer neuroscience, research and higher education, including Vice Dean and doctoral teaching appointments and five years at Oxford Business College. For over eleven years she has served as a registered permanent court expert in marketing and in the valuation of intellectual property rights across the European Union. She is the author of more than forty publications, the holder of numerous awards and recognitions for her work and her contribution to science, and Guest Editor of this Special Issue.
Dr. Thomas Zoëga Ramsøy
Neurons (Founder and CEO), Denmark, International Center for Applied Neuroscience (Founder & CEO), Denmark, IULM, Faculty Member, Italy;
Dr. Thomas Zoëga Ramsøy is a world-renowned neuroscientist, entrepreneur, and trailblazer in using AI and deep neuroscience to predict consumer responses—attention, emotion, cognition, and memory—in seconds from images and videos. As the founder, CEO, and Chief Science Officer of Neurons, he leads the charge in revolutionizing marketing through cutting-edge neuro-AI solutions that empower businesses to connect with their audiences like never before. With over 25 years of groundbreaking work, including inventing tools like the Perceptual Awareness Scale (PAS) and NeuroPrototyping earlier in his career, Thomas pushes the boundaries of what’s possible at the nexus of science, technology, and human behavior.
