News Release

When believing becomes seeing, or how the brain learns to fill in the blanks

Boston University psychologists find the downside to perceptual learning

Peer-Reviewed Publication

Boston University

(Boston) -- Our ability to learn to see things that may be new or unfamiliar to us is a plus; it allows us to adapt to changes in our surrounding environment. A big benefit is that it allows us to learn to do new tasks, such as becoming skilled at seeing the mere suggestion of a tumor on a mammogram. Learning to increase our sensitivity to a visual stimulus also seems to come at a cost, according to new research by Takeo Watanabe, an associate professor and director of the Vision Science Laboratory in Boston University's Department of Psychology.

It seems when learning to see things that are there, we also learn to see things that aren't.

"It's a manifestation of overlearning," says Watanabe, "such as when we find a man's face on Mars' surface or in a forest or on a cloud. We've overlearned human faces so we see them where they aren't."

This double-edged result has to do with a learning characteristic known as plasticity, the brain's ability to remold its responses because of its task-specific information repeatedly presented to it through its visual sensory system.

Watanabe and his team of researchers report their findings in this week's issue of the Proceedings of the National Academy of Sciences. The research was funded by grants from the National Institutes of Health, the Japan Society for the Promotion of Science, and the Human Frontier Science Program.

It has long been thought that only the "higher" level processes of the brain retained plasticity as we aged; that "lower" level processes become stable or "hard-wired" shortly after birth. This stability is considered to be a protective function, preventing the brain from basically relearning everything all the time, something that could make a walk in the park an insurmountable challenge.

In this plasticity–stability construct, a radiologist in the prime of life can learn to detect seemingly undetectable images of tumors on an X-ray image. In comparison, our ability to detect and "process" visual cues related to the direction and relative speed of moving objects in our immediate environment develops and becomes "fixed" in the first six or so months of life.

In their complex study design, Watanabe's team exposed groups of participants to a three-phased experiment structured to evaluate perceptual learning. In before- and after-test phases that bracketed an eight-day training phase, participants were assessed for changes in their ability to detect movement, direction, and the presence of imperceptible dots moving coherently in given directions about a fixed point on a computer screen.

In the in-between training phase, researchers exposed participants to screens on which imperceptible dots moved in one of several directions in the region behind a fixed point. On the fixed point, any of eight letters could appear in a rapid, serial manner. Some letters, known as target letters, were overlapped briefly by dots moving in one direction (called a paired direction). The researchers asked participants to identify which letters were the target letters.

When the researchers compared data from the before- and after-test sessions, they found that subliminal coherent motion improved participants' ability to detect the direction of motion paired with target letters. Surprisingly, the researchers also found that this improved performance occurred even when there were no dots traveling on the screen. Moreover, participants who were more likely to report seeing coherently moving dots when none were present were also more likely to report the movement as a paired direction.

The participants, the team concluded, had developed a perceptual bias -- they had learned to see the stimuli even when they were not there. "This perceptual bias," writes the team, "is an important demonstration of how perceptual learning can indeed have a cost. Adult sensory plasticity can be dangerous not only because it can lead to perceptual errors, but also because it can result in misjudgments of old perceptions. Sensory learning, which updates the sensory system to better perceive the current environment, may contribute to the formation of errors of memory."


The Department of Psychology at Boston University offers programs of research in three general areas: brain, behavior, and cognition; clinical psychology; or human development. The department's Vision Sciences Laboratory takes an interdisplinary approach to studying vision, embracing researchers who specialize in motion perception, attention, perceptual learning, categorization learning, motion-form interaction, neon-color spreading, and neural modeling.

Boston University, the fourth-largest independent university in the nation, has an enrollment of more than 30,000 in its 17 schools and colleges.

Disclaimer: AAAS and EurekAlert! are not responsible for the accuracy of news releases posted to EurekAlert! by contributing institutions or for the use of any information through the EurekAlert system.