In vision, information is a two-way street


Researchers in Portugal are untangling the complex connections that inform visual perception. Tom Keeble reports.


The complexities of visual perception are gradually coming into focus.
The complexities of visual perception are gradually coming into focus.
Marianna Armata / Getty Images

Neuroscientists have established that the brain makes predictions about what it’s going to see.

In a paper published in the journal Nature Neuroscience, scientists in Lisbon, Portugal, shed new light on the one of the most-studied, yet most enduring mysteries of the brain – our highly accurate visual systems.

Visual information arrives and travels from your eye into the primary visual cortex at the back of the brain, before moving forward into higher parts of the cortex for further signal processing, extracting key features such as edges and motion.

The higher brain areas also send signals to a tangle of connections that feed information back to the lower visual cortex. The new work expands on the role of these feedback connections, which until now have confounded scientists.

A team led by Leopoldo Petreanu from the Champalimaud Centre for the Unknown confirmed that cells in the higher visual system of mice send information back to the lower one.

The process maintains the visual system’s so-called retinotopic map, created by retinal neurons sending information into the cortex. However, the feedback connections also encoded information from more distant locations in the visual space – a novel finding which the authors predict will subsequently be shown in higher primates and other mammals.

Tiago Marques, first co-author of the study along with Julia Nguyen, says, “We found that there is a beautiful organisation, where feedback connections target specific neurons in lower structures depending on the signals they carry.”

The scientists hypothesise that the feedback connections provide extra information, in real time, to the lower visual cortex, to more fully inform the brain about the context in which it is seeing something, allowing it to distinguish, for instance, between a red apple and a cricket ball.

This discovery motivated the researchers to delve deeper into what other types of information the feedback connections might be sending to the primary visual cortex. They wondered whether they might help primary visual cortical neurons detect object edges, or predict an object’s direction of motion.

Using experiments that involved imaging the activity of the feedback connections, the team determined that they dampen down the activity of neighbouring primary neurons, to give the brain an even sharper picture of what was happening in the visual environment.

Furthermore, the feedback connections fired in the appropriate pattern as the original stimuli was being observed by the mouse – in effect, the clairvoyant brain predicted what the eyes would see even as the primary visual cortex was receiving the information.

“We believe this set of feedback connections learn what to expect from the world and then use this knowledge to shape incoming visual information,” explains Petreanu.

“In the world, objects are defined by continuous lines, not scattered dots, and moving objects tend to maintain their trajectory, not move around randomly. So feedback connections try to accentuate these particular features that they have learned to anticipate."

“This novel work addresses many of the great mysteries of the brain: how feedback information is organised and the functional consequences of this organisation,” comments Lucy Palmer of the Neural Networks Laboratory at the Florey Institute of Neuroscience and Mental Health in Melbourne, Australia.

“It helps us in our understanding of how feedback information is organised in the visual cortex, and the role it plays in our ability to see.”

Tom Keeble has a PhD in developmental neuroscience and currently works as a biomedical communicator at the Florey Institute of Neuroscience and Mental Health in Melbourne, Australia.
  1. https://doi.org/10.1038/s41593-018-0135-z
  2. https://doi.org/10.1038/s41593-018-0135-z
Latest Stories
MoreMore Articles