NIPS researchers studied the neural mechanism that permits familiar pictures to be seen even when they are blurred.

Feting Blue

Neuronal mechanisms that enable us to recognize familiar pictures, even if they are blurry, have been studied by NIPS researchers. Rats undertaking a visual orientation discrimination test showed an increase in neurons responding to low-contrast visual stimuli rather than high-contrast visual stimuli after repeated exposures.

Incorrect-choice trials elicited more activity from these neurons than correct-choice trials. These neurons well represented Low-contrast stimuli. Thus, the preference for low contrast in V1 action may lead to better low-contrast visual discrimination.

Objects may alter their appearance at any time. For example, in low-light conditions, the contrast between things becomes less noticeable, making it harder to tell them apart.

On the other hand, the brain can recognize individual things even if they are blurry because it has been exposed to them many times before. We still don't know precisely how low-contrast everyday items are perceived.

It has long been assumed that the visual responses in the primary visual cortex (V1), the part of the brain that processes the essential visual information, directly reflect the intensity of external stimuli. Since high contrast images generate powerful reactions, the opposite is true.

Repetitive exposure to low-contrast stimuli causes an increase in V1 neurons that selectively react to low-contrast stimuli in rats. Compared to high-contrast visual stimuli, low-contrast visual stimuli elicit more significant responses in these neurons. Neurons that favor low contrast are more active when rats correctly identify a known low-contrast item.

Science Advances initially demonstrated that V1 prefers low-contrast visual information, and this preference is increased in an experience-dependent way. A familiar item may be seen even though it is blurry because of this technique.

Flexible information representation may allow everyday items with any contrast to be seen the same way, Kimura claims. "Our brain can adapt that allows us to feel, even if we may not be conscious of it. By combining both high and low contrast-preferring neurons into an artificial neural network model, researchers believe they may accurately replicate the human experience."


The contribution of low contrast–preferring neurons to information representation in the primary visual cortex after learning published in science advance

SCIENCE ADVANCES•26 Nov 2021•Vol 7, Issue 48•DOI: 10.1126/sciadv.abj9976

Comments / 0

Published by

some lines on happiness


More from Feting Blue

Comments / 0