The post-hoc montage of perception: deep layers of primary visual cortex encode postdictive percepts

Our eyes are constantly bombarded with light, and it seems natural to assume that what we see is a live broadcast of the world happening right now. We look at a scene, and our brain instantly translates the light hitting our retinas into a clear picture of reality. However, there is a strange glitch in this system called postdiction. This is a phenomenon where information that arrives after a visual event actually changes how we remember or perceive that event. It is as if the brain waits a fraction of a second, gathers more clues from the future, and then edits the past to make sense of the whole story. For decades, scientists have debated how the brain manages this editing. One idea is that the brain simply holds onto early signals in higher, more complex regions until later information catches up. Another idea is that the brain sends a message back down to the very first stop of visual processing to rewrite the initial record.

A team of researchers at University College London and Maastricht University has now looked inside the human brain to see which of these ideas is correct. They focused on the primary visual cortex, the thin strip of tissue at the back of the brain that acts as the first gateway for everything we see. Specifically, they wanted to know if this area, which was once thought to be a passive camera, actually gets involved in rewriting our perception after the fact. To find out, they used a powerful magnetic scanner to watch the brain of twenty-three volunteers as they experienced two specific visual tricks. In these tricks, the presence or absence of a sound would cause the brain to either invent a flash of light that wasn't there, or to erase a flash of light that actually was there. By watching the brain during these moments, the researchers could see if the primary visual cortex was merely recording the light that hit the eye, or if it was reflecting what the person actually saw in their mind.

The experiment relied on a clever setup involving flashes of light and beeping tones. In one version of the test, participants saw two flashes of light with a sound in between them. Even though there was no light in the middle, the sound tricked their brains into seeing a third flash right where the sound occurred. In the other version, they saw three flashes of light, but if the middle sound was missing, their brains would suppress the middle flash, making it disappear from their perception. The researchers used a 7 Tesla MRI scanner, a machine with extreme sensitivity, to take pictures of the brain every 68 milliseconds. They did not just look at the brain as a whole; they looked at the different layers of the primary visual cortex. Think of these layers like floors in a building: the bottom floors receive signals directly from the eye, while the top floors send signals back down from other parts of the brain.

When the researchers analyzed the data, they found a clear difference between what the eye saw and what the brain perceived. In the layers of the visual cortex that receive direct input from the eye, the brain activity matched the actual light hitting the retina. If a flash was physically there, the signal was there; if it was missing, the signal was missing. However, in the deep layers of the same area, the story was different. When a participant saw an illusory flash that never physically existed, the deep layers of the visual cortex lit up in a pattern that looked exactly like they were seeing a real flash. Conversely, when a real flash was physically present but the participant's brain decided to ignore it because of a missing sound, the deep layers showed a pattern that looked like the flash was gone. This means that the deep layers of the visual cortex were not just recording the raw data from the eye; they were recording the final, edited version of reality that the person experienced.

To understand where this editing message was coming from, the researchers looked at the connection between the visual cortex and the superior temporal cortex, a region higher up in the brain known for combining sound and sight. They found that the activity patterns in this higher region were closely linked to the activity in the deep layers of the visual cortex, but not to the middle layers. This suggests that the higher region is sending a message back down to the deep layers, telling them to update the picture based on the new information. The study did not find this same connection for the middle layers, which supports the idea that these layers are mostly for receiving new input, while the deep layers are for receiving updates.

The findings challenge the old view that the primary visual cortex is just a passive receiver of information. Instead, it appears to be an active participant in constructing our experience of the world, even after the event has passed. The brain does not just stream the present; it constantly revises the past to create a coherent story. In these experiments, the revision process happened in the deep layers of the visual cortex, driven by feedback from higher brain areas that had integrated the sound and sight. While the researchers could not see the exact timing of these events with enough precision to know if the brain was editing the past or just predicting the future, the evidence clearly shows that the primary visual cortex is involved in this post-hoc reconstruction. The study suggests that our perception is a montage, a collection of moments stitched together by the brain, and that the very first stop on the visual highway is where the final cut is made.