Original studybioRxiv, 2026Read the paper →
Look at a scene and it feels as though you take in one steady picture. Yet the brain receives far more than ever reaches awareness, and scientists still cannot say what separates the activity we experience from the activity we never notice. New recordings from monkeys suggest that part of the answer may lie in timing, with the brain switching rapidly between what is seen and what is pushed out of sight.
Researchers at the University of California, Berkeley, Princeton University, the California Institute of Technology, and the Howard Hughes Medical Institute recorded hundreds of nerve cells at once in two rhesus macaque monkeys. The study has been posted on the preprint server bioRxiv and has not yet been peer reviewed. The team focused on face-processing regions in the temporal lobe, and later on parts of the prefrontal cortex.
The experiments used a visual trick called binocular flash suppression. One face is shown to one eye, then a different face is flashed to the other eye, and the new face takes over conscious vision while the first stays on screen unseen. Tests with 14 human volunteers confirmed that people almost always reported seeing the newly flashed face.
The monkeys watched 1,500 different faces through filtered goggles while probes called Neuropixels tracked the cells. Because face cells respond to measurable features of shape and appearance, the researchers could rebuild an image of the face the brain was representing from its activity. They then checked, in steps of a few milliseconds, which of the two faces the activity most resembled.
Averaged over time, the cells looked only weakly swayed by what was consciously seen, as earlier studies had also found. Moment by moment, the picture changed. The population flipped about 15 times a second between two states, one carrying a clean copy of the visible face and the other a clean copy of the hidden one, with each state lasting roughly 60 milliseconds.
Faces rebuilt from the first state closely matched the face the volunteers would see, explaining more than 80% of the variation in some facial features. Faces rebuilt from the second state matched the unseen face instead. The same cells carried both faces in turn, rather than separate groups of cells handling each one.
The two states were not perfect mirror images. Tuning to the visible face was about a fifth stronger than tuning to the hidden face, and a simple decoder could tell the states apart 66% of the time. Recordings from the prefrontal cortex, often proposed as a seat of conscious access, showed a similar back-and-forth pattern.
The study has clear limits. It involved only two animals, and the monkeys did not report what they saw, so their perception was inferred from human volunteers shown the same displays. The switching was only loosely coordinated between neighbouring face areas, and the work has yet to pass peer review. The recordings also show a link between brain activity and what is seen, not proof that the switching causes awareness.
The finding feeds into a long debate over whether consciousness lives at the front or the back of the brain. These results suggest that location alone may not be the answer, and that timing helps decide which of the brain’s competing pictures reaches awareness. A rich trace of unseen images may also help explain how things we never notice can still sway attention and mood.
