Multitasking is no problem for these brain cells: Individual brain cells can ID both cars and cats

June 9, 2010
In previous studies showed that when monkeys are trained to categorize animals by distinguishing cats from dogs, some neurons in the prefrontal cortex become tuned to the concept of “cat” while others respond to the idea of “dog.” This new study trained the monkeys to perform two different categorization tasks — distinguishing cats from dogs and sedans from sports cars.

( -- Researchers at MIT's Picower Institute for Learning and Memory found that single brain cells, if confronted with a difficult task, can identify objects as dissimilar as sports cars and dogs.

In humans and other primates, the prefrontal cortex is the seat of high-level functions such as learning, decision making, and planning. Neuroscientists have long wondered whether neurons in that part of the brain are specialized for one type of task or if they are “generalists” — that is, able to participate in many functions. A new study from MIT’s Picower Institute for Learning and Memory comes down in favor of the generalist theory.

MIT professor Earl Miller and others in his lab showed that when they trained monkeys to perform two different categorization tasks, about half of the neurons involved could switch between the two. The findings, reported in the June 10 issue of the journal Neuron, suggest that neurons of the prefrontal cortex have a much greater ability to adapt to different cognitive demands than neurons in other parts of the brain. These results support ideas about the malleability of neurons — nervous-system cells that process and transmit information — that Miller first proposed a decade ago.

Miller, the Picower Professor of Neuroscience in MIT’s Department of Brain and Cognitive Sciences, says he’s not surprised by the findings. “We have a lot of mental flexibility,” he says. “We can change our topic of conversation, we can change what we’re thinking about. Some part of the brain has to have that flexibility on a neural level.”

Listening to single neurons

Most neuroscientists who study brain activity in monkeys train the animals on only one task, so until now it had been impossible to reveal whether single neurons in the prefrontal cortex could be involved in more than one job.

In previous studies, Miller has shown that when monkeys are trained to categorize animals by distinguishing cats from dogs, some neurons in the prefrontal cortex become tuned to the concept of “cat” while others respond to the idea of “dog.”

This time, Miller, postdoctoral fellow Jason Cromer, and research scientist Jefferson Roy trained the monkeys to perform two different categorization tasks — distinguishing cats from dogs and sedans from sports cars. They recorded activity from about 500 neurons in the monkeys’ prefrontal cortex as the animals switched back and forth between the tasks.

Although they found that some neurons were more attuned to car images and others to animal images, they also identified many neurons that were active during both tasks. In fact, these “multitasking” neurons were best at making correct identifications in both categories.

The findings suggest that neurons in the prefrontal cortex have a unique ability to adapt to different tasks, says Miller. In other parts of the brain, earlier research has shown, most neurons are highly specialized. Neurons in the visual cortex, for example, are programmed to respond to very specific inputs, such as a vertical line or a certain color. Some have even been shown to fire only in response to one particular face.

“Our results suggest that the prefrontal cortex is different from the sensory cortex and the motor cortex. It’s highly plastic,” says Miller. “That’s important, because it means the human brain has the capacity to absorb a lot of information.”

The Neuron study focused on two categorization tasks, but Miller hopes to run another study in which the monkeys learn a third task involving some other cognitive function. That could give another hint about how much information our brains can handle, says David Freedman, an assistant professor of neurobiology at the University of Chicago.

“We’re very good at learning dozens, hundreds, or even thousands of categories,” he says. “You wonder if there is some limit, or would these neurons be as flexible as we are as observers?”

Freedman says he would also be interested to see whether the same prefrontal-cortex neurons can multitask between activities that involve different kinds of sensory inputs — for example, a visual task and an auditory task.


Meanwhile, Miller has a study under way that he believes could demonstrate a biological basis for the impaired categorization ability often seen in people with autism. Autistic children often have a hard time understanding that two slightly different objects — for example, a red toothbrush and a blue toothbrush — both belong to the same category.

Miller theorizes that an evolutionarily older part of the brain, known as the basal ganglia, gathers information about new objects, and the prefrontal cortex learns how to categorize them. “The basal ganglia learn the pieces of the puzzle, and the prefrontal cortex puts the pieces together,” he says.

In his current study, Miller is monitoring brain activity in monkeys as they learn a categorization task. He expects to find a sharp peak in prefrontal-cortex activity at the moment when the monkeys learn that certain objects belong to the same category.

Eventually, he hopes to show that in autism, the balance between those two brain regions is thrown off: there is either too much activity in the basal ganglia or not enough in the prefrontal cortex.

Explore further: Turning Sensation into Perception

More information: "Representation of Multiple, Independent Categories in the Primate Prefrontal Cortex," Jason A. Cromer, Jefferson E. Roy, and Earl K. Miller. Neuron. 10 June 2010.

Related Stories

Turning Sensation into Perception

November 6, 2005

Perceiving a simple touch may depend as much on memory, attention, and expectation as on the stimulus itself, according to new research from Howard Hughes Medical Institute (HHMI) international research scholar Ranulfo Romo ...

Picower research finds unexpected activity in visual cortex

March 16, 2006

For years, neural activity in the brain's visual cortex was thought to have only one job: to create visual perceptions. A new study by researchers at MIT's Picower Institute for Learning and Memory shows that visual cortical ...

The Brain's Executive Is An 'Event Planner'

May 18, 2006

Studies in which monkeys were asked to manipulate computer cursors for fruit juice rewards have revealed that the brain's "executive" center plans behaviors not by specifying movements required for given actions, but rather ...

Long-distance brain waves focus attention (w/Video)

May 28, 2009

( -- Just as our world buzzes with distractions -- from phone calls to e-mails to tweets -- the neurons in our brain are bombarded with messages. Research has shown that when we pay attention, some of these neurons ...

Why we learn more from our successes than our failures

July 29, 2009

( -- If you've ever felt doomed to repeat your mistakes, researchers at MIT's Picower Institute for Learning and Memory may have explained why: Brain cells may only learn from experience when we do something right ...

Recommended for you

How the finch changes its tune

August 3, 2015

Like top musicians, songbirds train from a young age to weed out errors and trim variability from their songs, ultimately becoming consistent and reliable performers. But as with human musicians, even the best are not machines. ...

Machine Translates Thoughts into Speech in Real Time

December 21, 2009

( -- By implanting an electrode into the brain of a person with locked-in syndrome, scientists have demonstrated how to wirelessly transmit neural signals to a speech synthesizer. The "thought-to-speech" process ...


Adjust slider to filter visible comments by rank

Display comments: newest first

5 / 5 (2) Jun 10, 2010
The brain shows every indication that it indexes 'most recently used' perceptions. Thus the sports car and dog both have this recency in common, and that is why the same neuron fires ie they are both recognised as recent perceptual phenomena.

I'm sure everyone has experience this for themselves ~ when, for instance, you relearn something, say you meet an old friend and talk about the old days, those memories are indexed for recency and will be easier to recall for some period of time. The common element of those memories is that they are now marked recent and probably indexed for quick recall (when a recency neuron fires the quick index, where so called 'mother cells' may be found, is checked for a match).

That is a far more plausible explanation than the idea that a single neuron can recognise anything, let alone phenomena as complex as dogs vs cats or styles of motor car bodies.
5 / 5 (1) Jun 10, 2010
Absolutely right Robert,

I too seriously doubt a single neuron is capable of such large, complex tasks. How much information is sent to a single neuron? Certainly not an entire image of a dog or cat. Clearly, these "dog only" neurons are simply bits of the knowledge of dogs... whether each of these "dog only" neurons are where dog data is stored or are part of the recently indexed data, clearly, each individual neuron only knows a tiny piece of a dog.
not rated yet Jun 11, 2010
Thanks, Sharpy,
at least one of my 'recognition' neurons became excited at your familiar voice though the name CSharpner does not ring any bells :)

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.