Perception and awareness

Any pattern of light on the retina has many possible sources in the world. Bistable displays, such as the Necker cube or different images shown to the two eyes (binocular rivalry), make the ambiguity explicit: perception switches between two interpretations while the input stays the same. We use these switches, together with attention, arousal and eye movements, to study how the brain settles on a percept and keeps perception and action calibrated.

Approach

What drives perceptual switches has been debated since the 19th century. Hermann von Helmholtz, who described perception as unconscious inference, tied them to attention; Ewald Hering saw a more automatic process. Imaging studies found transient activity in right frontal and parietal cortex around switches, which could reflect their cause (Brascamp et al., 2017). We design stimuli and tasks that separate a switch from its consequences, such as noticing and reporting it. We combine psychophysics with fMRI, intracranial recordings, spectroscopy, pharmacology and pupillometry. For eye movements, we use gaze-contingent displays that shift a target while the eye is in flight, which drives saccadic adaptation without the observer noticing.

What we have found

Frontoparietal activity follows switches. Transition-related right frontoparietal activity grew with the duration of a transition, and replayed transitions with realistic durations evoked as much activity as real ones (Knapen et al., 2011). When switches were so inconspicuous that observers could not report them, frontoparietal BOLD responses were similarly minimized, indicating that they reflect the switches’ consequences rather than their cause (Brascamp et al., 2015).

Inhibition in visual cortex shapes bistable perception. Higher GABA concentrations in human visual cortex, and lorazepam, which acts on GABA_A receptors, both slowed perceptual dynamics (van Loon et al., 2013). At 7T, eye-specific rivalry arose from inhibition between adjacent ocular dominance columns in superficial layers of V1, with eye-specific feedback from the intraparietal sulcus biasing these local competitions (Qian et al., 2025).

Illusory changes reverse the visual hierarchy. In intracranial recordings from two patients, responses to illusory changes during rivalry rose slowly and progressed from higher-order to lower-order visual regions, the reverse of the order for physical changes (de Jong et al., 2020).

The pupil separates perception from action. At perceptual changes, the pupil showed a dilation tied to task execution and a constriction tied to the perceptual transient (Brascamp et al., 2021). In sensory-motor decisions, large pupil responses predicted reduced decision bias (de Gee et al., 2017).

Saccadic adaptation can be global. Adapting saccades in all directions at once produced spatially generalized adaptation (Rolfs et al., 2010). Fast learning contributed to the adaptation of reactive saccades but was nearly absent for voluntary ones (van Es and Knapen, 2019).

Open questions

Which parts of the brain’s response to a switch cause it, and which follow from noticing and reporting it? How do local competition in sensory cortex and parietal feedback divide the work of resolving ambiguity? With colleagues, we have argued that integrated information theory is unscientific because its core claims cannot be tested even in principle (Klincewicz et al., 2025).