Research at MIT provides new insight into how the brain adapts to new stimuli while maintaining a fixed number of cells. The work, described in the May 3 issue of Nature, also helps explain optical illusions.
It turns out that some neurons in the primary visual cortex are quick-change artists. Neurons located in certain places have the ability to take on briefly the function of neighboring cells and then switch back to their own jobs, said Valentin Dragoi, a postdoctoral fellow at the Center for Learning and Memory and one of the authors of the paper.
While the brain's cells were thought to all work the same way, "we show for the first time that neurons have different properties depending on where they are in the network," Dr. Dragoi said. This has "deep connotations for understanding the structure and function of the adult brain," said co-author Mriganka Sur, head of the Department of Brain and Cognitive Sciences and the Sherman Fairchild Professor of Neuroscience.
This newfound ability of certain neurons explains the visual phenomenon called the tilt aftereffect. After staring at a set of tilted lines, when you look at a vertical line, it seems tilted in the opposite direction.
No one knew whether this and other optical illusions were caused by the eye or the brain. Drs. Dragoi, Sur and co-author Casto Rivadulla, previously a visiting scientist at MIT and now in Spain, demonstrate that in this case, the neurons responsible for recognizing line segments that lean to the right have briefly taken over the job of neighboring neurons that recognize line segments thatlean to the left.
"This is the first time that anyone in this field has been able to show how local differences in the way the map of cells that make up the brain is constituted have consequences for how the brain works," Professor Sur said.
To form pictures of objects, our brains have cells that are experts at recognizing lines oriented around the compass. Some neurons "see" horizontal line segments, some respond to vertical line segments and intermediate kinds respond to line segments tilted at other angles.
Using optical imaging techniques that show how cells in the primary visual cortex that respond to tilted lines are grouped, Dr. Dragoi mapped the neurons by function. Regions of cortex with cells that all have the same angle preference lie next to regions with a slightly different preference. Where the different types of neurons come together, the regions radiate out in the shape of a pinwheel.
After being "tuned" during development, these orientation preferences were thought to be hard-wired in the adult brain.
Yet cells are dependent on input. Dr. Dragoi showed that neurons have different properties depending on where in the map they're located and what kind of input they get from their neighbors.
Neurons at or near pinwheel centers receive strong local inputs from neurons of all orientations. Those in more homogeneous areas, called iso-orientation domains, receive only one kind of input.
The researchers found that neurons with more diverse neighboring cells have a broader capacity to change and adapt than those with neighbors who are all just like themselves. The iso-orientation domains are zones of stability; pinwheels are the agents of change.
"We know that what we have just heard or just felt will influence what we next hear or feel, at least for a short period of time," Professor Sur said. Apparently, the same is true for vision. When a visual neuron receives input in the form of a stimulus -- say, staring at a pattern -- it changes slightly and then reverts to its previous state.
This makes sense, Dr. Dragoi pointed out, because our brains must constantly be ready for the next stimulus, yet must keep a stable picture of the way the world usually works. "If the brain changed every time we looked at something, nothing would look the same twice. Some parts of the brain change and some stay constant to provide a dynamic view of the world while maintaining a fixed frame of reference," he said.
This work is funded by the National Institutes of Health and by a Merck fellowship.
A version of this article appeared in MIT Tech Talk on May 9, 2001.