How do humans select and process information when the environment is dynamic, multisensory, interactive, and full of distractions? My research programme — Naturalistic Attention Dynamics — builds experimental and computational accounts of attention that survive contact with the messy, real-world conditions that classical laboratory paradigms are deliberately stripped of.
What the programme tackles
1. Complexity and interaction
I study attention in immersive Virtual Reality and in haptic foraging — paradigms in which observers physically search, pick up, and rearrange real objects rather than just looking at displays. This lets us measure how 3D spatial depth, body movement, and physical effort cost reshape attentional priorities and gaze dynamics. Recent work shows that statistical learning reweights attention along the depth axis in immersive VR (Stefani & Sauter, 2025, PsyArXiv), building on the foundational interactive-search paradigm (Sauter, Stefani & Mack, 2020, Brain Sciences).
2. Adaptive cognitive control under change
How do observers detect, time, and gate updates to learned attentional priorities when environmental regularities drift? My published work establishes region-based distractor suppression that transfers across feature dimensions (Sauter, Liesefeld & Müller, 2019, JEP:LMC), shows that learnt priority is updated post-capture rather than purely pre-attentively (Sauter et al., 2021, Cortex), and decomposes distractor dwell times into capture and disengagement components (Stefani & Sauter, 2023, Scientific Reports).
3. Applied cognition at the computer-science interface
Attention diagnostics belong in the interactive, AI-mediated environments people actually inhabit. Since 2021 I have built this thread through DFG SPP 2199 Scalable Interaction Paradigms for Pervasive Computing Environments and joint work with the Rukzio HCI group — yielding the Best Paper Award at ACM Mensch und Computer 2023 for eye-movement visualisation in online teaching, and the Best Paper Award at the COGAIN Symposium at ACM ETRA 2022 for visualising the attention of many observers simultaneously.
Cognimize Classroom: live cognitive experiments for international teaching
The platform tying these threads together is Cognimize Classroom (CogClass) — an open-source web platform (currently under development) funded in 2026 by Stiftung Innovation in der Hochschullehre via UGo! Global Teaching Labs.
Cognimize hosts 33 classical cognitive paradigms (Stroop, Flanker, visual search, change-blindness, …) that students run on their own devices during regular lectures. Results aggregate in real time on the projection, and a Global Highscore System lets cohorts at different universities compare their group-level effects. A Collaborative Online International Learning (COIL) pilot with the University of Oxford (Prof. Dejan Draschkow) extends the platform into cross-cultural empirical teaching. The platform will be released as open source on GitHub at the end of the funding period.
Methods
- Behavioural experiments (in-lab and online)
- Stationary and mobile eye tracking
- Pupillometry — incl. Locus-Coeruleus / Norepinephrine indices of arousal and effort
- Immersive Virtual Reality (3D / depth paradigms)
- EEG and TMS (selective applications)
- Large-N analytics and replication science — contributing role in the Many-Labs Nature 2026 study; cross-cultural regression on N = 13,464 participants from 109 countries (Sauter, Braun & Mack, 2021, Cyberpsychology, Behavior, and Social Networking)
Open science
All confirmatory studies are preregistered. Data, code, and materials are shared via OSF and GitHub. Manuscripts are deposited on PsyArXiv prior to peer review. I contributed to the Many-Labs replication consortium (Tyner et al., 2026, Nature) — a direct demonstration of openly conducted, large-scale, transparent science.