Pupil dilation responses (PDRs) are reliable physiological indicators of arousal in response to unexpected events in different sensory modalities. However, a direct comparison of PDRs between modalities is currently lacking. We investigated PDRs in response to transitions between regular and random sequences of visual dots (Experiments 1-3) and auditory tones (Experiment 3). In Experiment 1, we observed PDRs when a regular visual sequence transitioned to a random one, but not for the reverse transition. Transitions between two regular sequences elicited weaker PDRs on average. Experiment 2 replicated these results, confirming the robustness of the observed patterns of PDRs. In Experiment 3, we compared PDRs to visual and auditory versions of the same sequences and transitions. We found strong cross-modal similarities in pupil dynamics, particularly for transitions between regular and random sequences. Remarkably, these similarities were also evident during baseline trials without explicit transitions. To further characterize these effects, we decomposed the pupil time series to extract phasic pupil dilation events. dilation event patterns mirrored overall pupil time series, with modality-specific differences in event amplitude emerging for the transitions between distinct regular sequences. Taken together, our findings suggest that pupil-linked arousal reflects inference of statistical structure and its violations, demonstrating substantial correspondence across sensory modalities.

The study was supported by the German Research Foundation (DFG): SFB 1233, Robust Vision: Inference Principles and Neural Mechanisms, TP C1 (previously TP 05), No 276693517, Max Planck Society and Humboldt Foundation (PD), and Machine Learning Cluster of Excellence, EXC 2064/1 No 39072764 (VF).