Monkey Brains Alternated Between Seen and Hidden Faces 15 Times a Second
Science

Monkey Brains Alternated Between Seen and Hidden Faces 15 Times a Second

Recordings from two macaques suggest that face-processing neurons alternated about 15 times per second between an image inferred to be consciously seen and one suppressed from awareness. The striking result comes from an unreviewed preprint, and the animals could not report their perception directly.

NewTqnia Science Desk Updated 4 min read
Monkey Brains Alternated Between Seen and Hidden Faces 15 Times a Second

Neural recordings from two rhesus macaques suggest that the brain may represent a consciously seen face and a suppressed face in rapidly alternating time windows, rather than holding only the visible image continuously. The team reports switching about 15 times per second in a bioRxiv preprint posted on October 16, 2026.

The 30-second summary

  • What happened? Researchers recorded hundreds of neurons while two macaques viewed different faces through red-and-cyan glasses. Neural patterns alternated between the face inferred to be visible and the face suppressed from awareness.
  • Why does it matter? The finding suggests that timing, not only brain location, may help separate conscious from unconscious visual information.
  • What is the catch? The work is not peer reviewed, involved only two macaques, and inferred their perception from a separate experiment with 14 humans.
Key Number: The two neural representations alternated roughly 15 times per second while the visual input and presumed percept remained stable.

Two faces reached the eyes, but only one was expected to be seen

The researchers used binocular flash suppression. First, one face was shown to one eye. A second face then appeared to the other eye while the first image remained physically present. Human observers usually report seeing the newer image while the earlier one disappears from awareness.

Fourteen human participants confirmed that this setup generally produced the expected percept. The macaques could not describe what they saw, so the researchers assumed that the newer face dominated their awareness in the same way. That assumption is central to interpreting the neural states as conscious and suppressed.

The same neurons appeared to time-share two representations

Neuropixels probes recorded activity in face-selective regions of the inferotemporal cortex and in prefrontal cortex. The team first measured the pattern associated with each face when shown alone. During suppression trials, it then classified short periods of activity according to which face pattern they resembled.

The population signal did not settle permanently on the presumed visible face. It alternated between brief periods dominated by that face and periods dominated by the suppressed one. During each type of period, the researchers could reconstruct features of the corresponding image from the recorded activity. Feature tuning was stronger during the periods associated with conscious perception.

This is not a system for reading arbitrary thoughts. A related NewTqnia report on decoding typed sentences from brain signals explains why reconstruction in a constrained experiment should not be confused with unrestricted access to a person’s mind.

Why timing changes the consciousness question

Many experiments ask which brain regions respond differently when a stimulus is consciously perceived. The new result proposes an additional dimension: the same neural population may carry both visible and suppressed content at different moments. Averaging activity over longer intervals could blur those alternating states into one mixed signal.

That interpretation does not explain why one representation reaches awareness while the other does not. It instead suggests that conscious access may depend partly on how neural states are selected over time. Outside researchers quoted by Nature described the methods as rigorous and the result as interesting, while cautioning that parts of the finding were less surprising than the online attention implied.

Before we overstate the result

  • The preprint has not completed peer review, and independent groups have not replicated the 15-hertz pattern.
  • Only two macaques supplied the direct neural recordings. Their presumed conscious percept was inferred rather than reported.
  • The probes sampled selected face-processing and prefrontal regions, not the entire brain or consciousness in general.
  • The analysis concerns visual competition between two controlled face images. It does not establish that ordinary experience repeatedly switches between complete versions of reality.

What should happen next

The claim will become stronger if other laboratories reproduce the alternation with larger samples, different visual categories and methods that provide an independent behavioral indication of the animal’s percept. Researchers must also test whether the timing pattern survives alternative ways of defining the short neural states, and whether it occurs beyond the face-processing network.

Image: Illustration of the binocular-suppression experiment. Credit: Hesse et al., bioRxiv (2026), CC BY 4.0.

Verified topics and entities

Sources and citations3 sources

Published by

N

NewTqnia Science Desk

An institutional editorial team within NewTqnia

A new version of NewTqnia is ready.