Penn Researchers Discover New Law Guiding the Way Humans Perceive the World
Laws of perception explain why people see the world the way they do. Alan Stocker of the University of Pennsylvania and his former graduate student Xue-Xin Wei, now a postdoc at Columbia University, have discovered a new such law, one of only a handful in existence.
The researchers confirmed a link between how sensitive people are to changes occurring in front of them, what’s called the discrimination threshold, and the perceptual bias, or the amount perception deviates from reality. They published their findings in the Proceedings of the National Academy of Sciences.
To understand the new connection, Stocker suggests considering a wall that’s uniformly lit. An experimenter changes the brightness of the light ever so slightly, then a little more and a little more. “The discrimination threshold tells us how fine the experimenter can actually change the light levels such that we still cannot detect it,” said Stocker, an associate professor of psychology. Once we can detect it, we’ve crossed the discrimination threshold boundary.
He offered as another example the leaning Tower of Pisa: If the building continues tilting at a rate so infinitesimally slow that while it’s happening it’s not visible to the human eye, then at some point it will become possible to tell that the building tilts at a different angle from where it started. That point is the discrimination threshold.
“People have done this for many different perceptual variables,” Stocker said. “It could be color. It could be auditory, the pitch of a tone, the intensity of a tone. It could be the weight in my hand. How much could it change so that I actually notice the difference?”
Perceptual bias distinguishes what people believe they are experiencing from the truth of what’s happening. Optical illusions are the classic example, he said. “Our perceptual sense fools us into seeing something that’s actually not there. This fooling, this deviation from the reality, we call this perceptual bias.”
Until now, there had been no direct relationship between these two measures. So the researchers returned to a theory of perception they had developed several years ago.
For humans to perceive a stimulus, it must go through a two-step process. In the first, called encoding, information from the senses get filtered and represented to the brain. “The second step, the decoding, is taking that signal and combining it with what we know about the world, intelligence, prior beliefs that get used to interpret this encoded signal in a way that makes sense,” Stocker said.
It’s well-established that human understanding of how the world works shapes the decoding step, but Stocker and Wei theorized that this might also influence the encoding. Once this connection became clear, the researchers made the leap to link discrimination threshold and perceptual bias.
To explain, Stocker brings up the idea of an uncommon color. “If there is a color we rarely see, then our brain would not dedicate many neurons to actually represent that color,” he said. “It’s so rare that it’s not worth it, so to speak. We have a limited amount of neural resources. We would rather dedicate the neurons to those colors that are more frequent. It’s a tradeoff.”
In other words, the brain will store the color blue and perceive it more easily than, for example, YInMn blue, a new hue discovered in June 2016 that looks similar to cobalt blue.
There aren’t practical applications for the new perception law yet. Stocker said he thinks that could change once more is known about perception in someone with Parkinson’s disease or autism. Does the encoding-decoding relationship remain intact? Is the reduction in prior knowledge a cognitive deficit or something else?
Answering these “might tell us about how these processes are neurally implemented,” he said.
But that’s research down the line. For now, this research provides a better picture of just how important people’s previous experiences are to shaping their perception of the world.
Funding for the research came from the Office of Naval Research, supported by grant number N000141110744.