How the Brain Learns to See

June 9, 2005

Most of us don’t have much trouble recognizing what we see. Whether it is a face in a crowd, a bird in a tree, or papers on a desk, our brains expertly distinguish the target from the clutter. It is a simple skill most of us take for granted, but object recognition is not hard-wired. As we navigate our environment, the brain’s visual centers continually reorganize themselves, classify novel features, and learn to pick out important objects from the background. Just how the human brain does this is not well understood, but new research by Zoe Kourtzi and colleagues may have uncovered some important clues.

To investigate how the human brain learns to separate targets (signal) from noise Kourtzi et al. showed subjects pictures of novel shapes embedded in a cluttered background and asked the subjects to determine whether or not the shapes were symmetrical. The researchers recorded the subjects’ responses while using functional magnetic resonance imaging (fMRI) to measure neuronal activity in brain regions associated with visual processing. Each subject was tested using two sets of novel shapes: high-salience shapes (shapes easily distinguished from the background), and low-salience shapes (shapes camouflaged by the background). After the initial testing, the subjects were trained to recognize a subset of the new shapes from each group, and then re-tested.

Visual input is thought to go through a hierarchy of processing centers that transform retinal images into complex objects and scenes. Kourtzi et al. recorded responses from both early (V1, V2, Vp, and V4) and late (lateral occipital cortex) stages of visual analysis in 26 subjects. The authors found that subjects demonstrated an increased number of correct responses for shapes they encountered during the training sessions, regardless of the type of background the shapes were presented on. By contrast, the fMRI responses differed dramatically, depending on whether the surroundings made the shapes easy or difficult to detect. Low-salience shapes triggered an increased fMRI response across all brain regions following training; high-salience shapes precipitated a decrease in fMRI response in the regions of the lateral occipital cortex, but produced no change in any of the early visual areas (V1, V2, Vp, and V4).

These results demonstrate that the ability to learn to detect novel shapes is independent of the degree of difficulty, but suggest that the brain employs different mechanisms of perceptual learning depending on whether the objects stand out from their surroundings, or are obscured by them. Learning to detect highly camouflaged shapes results in increased brain activity levels that are presumed to reflect an increase in signal processing at the level of both the early visual areas and higher levels of cortical analysis. On the other hand, the reduction of neural activity that occurs during learning of more distinctive shapes likely reflects efficient neural coding of the critical features for their recognition at later stages of visual analysis.

According to Kourtzi and her colleagues, their results provide evidence that the visual brain is capable of tailoring the mechanism of perception to best suit the task. When the signal is weak—as in the case of viewing camouflaged targets—learning amplifies neural responses to the target shapes and drowns out the noise. But when the signal is strong—as in the case of viewing easily distinguishable, highly salient targets—neural activity in the visual cortex is reduced, possibly because training engages smaller populations of neurons that respond much more selectively to distinctive features of the stimulus.

In other words, all visual stimuli are not treated equally, and with just cause: the brain’s unique ability to treat ambiguous signals differently than robust ones likely allows it to optimize neural coding, and in doing so, learn to increase detection of a broad spectrum of visual signals.

Citation: (2005) Cutting through the Clutter: How the Brain Learns to See. PLoS Biol 3(7): e256

Source: PLoS Biology

Explore further: For worms, positive thinking is the key to finding food

Related Stories

For worms, positive thinking is the key to finding food

September 15, 2015

Caenorhabditis elegans, a tiny roundworm, spends much of its lifetime searching for soil bacteria to eat. This humble creature possesses 302 neurons, which may not seem like a lot compared to the billions of nerve cells that ...

An insect eye for drones

August 12, 2015

Is it possible to catch a fly? Small insects seem to possess a sixth sense that allows them to dodge any threats. Yet there is no magic trick, but only a compound eye that is an organ of vision made of thousands of ommatidia. ...

What neuroscience can learn from computer science

August 10, 2015

What do computers and brains have in common? Computers are made to solve the same problems that brains solve. Computers, however, rely on a drastically different hardware, which makes them good at different kinds of problem ...

Recommended for you

Ancient genome from Africa sequenced for the first time

October 8, 2015

The first ancient human genome from Africa to be sequenced has revealed that a wave of migration back into Africa from Western Eurasia around 3,000 years ago was up to twice as significant as previously thought, and affected ...


Please sign in to add a comment. Registration is free, and takes less than a minute. Read more

Click here to reset your password.
Sign in to get notified via email when new comments are made.