Nausea reshapes visual exploration and pupillary responses to affective images

Perception is not a passive recording of external events. It is a continuous, active, and dynamic process of inference shaped by current goals, past experience, and hard biological constraints. The physiological state of the body is arguably the most fundamental constraint. Continuous inference about interoceptive states colors perception with emotions, which in turn affect basic perceptual and attentional processes. How, then, do we really look at the world when experiencing malaise? Here we leverage Galvanic Vestibular Stimulation (GVS) - a non-invasive brain stimulation technique that induces mild dizziness, motion sickness, and nausea - in conjunction with high-resolution eye tracking, pupillometry, and computational modelling. We demonstrate that mild vestibular perturbations disrupt the coupling between gaze and informative low-level visual features. Crucially, however, this effect depends on individual susceptibility and on stimulus valence. Particularly susceptible participants experiencing stronger motion sickness increased their reliance on visually salient features, and particularly so for threatening (negatively valenced) images. Pupillometry revealed larger tonic pupil size in these participants, consistent with heightened arousal, as well as altered phasic responses to affective pictures. There was a close relationship between subjective reports of nausea, oculomotor behavior, and pupillary dynamics, highlighting an important role of inter-individual differences and related modulators (e.g., interoceptive accuracy). Altogether, these results unveil the continuous and dynamic interplay between interoceptive and exteroceptive processes, showing that how we perceive the environment fundamentally resonates with how we feel. HighlightsO_LIGVS caused variable degrees of dizziness and nausea C_LIO_LIGVS led to spatial biases in visual exploration which were, however, further dependent on image valence and content C_LIO_LIDuring GVS eye fixations landed, on average, at locations that were less salient based on their mere physical properties, provided participants did not experience strong interoceptive perturbations C_LIO_LIHighly susceptible participants showed the opposite pattern: their fixations landed on visually more salient and informative locations, especially for negative images C_LIO_LIParticipants who felt stronger effects, including motion sickness, also showed the largest increase in tonic pupil size; concurrently, the signature phasic dilation associated with negative emotions was disrupted C_LIO_LIOculomotor behavior and perception adjust dynamically to ongoing interoceptive processes; GVS can be an effective tool to study this dialogue, provided that inter-individual differences are accounted for C_LI