Researchers Puzzle Out How the Brain Learns Odors

by Faith Hickman Brynie

July 14, 2009

Long-term potentiation, or LTP, the process that strengthens connections among neurons, is the neuronal basis of learning. Scientists have observed the process in many areas of the brain but, until recently, not in the brain’s olfactory bulb, the primary region that processes smell signals coming in from the nose. Now researchers have found evidence of LTP—but working in an unexpected way.

Scientists have long believed that LTP must occur in the olfactory bulb for two reasons. One is behavioral: Animals learn smells. The other is biochemical and neurological, involving the synapse, the junction via which neurons communicate.

When a nerve impulse arrives at a synapse, NMDA (N-methyl-D-aspartate) receptors open on cell membranes, beginning a series of biochemical reactions. If those reactions occur often enough, LTP occurs—the signal that is transmitted across the synapse becomes stronger.

 “We knew there were lots of NMDA receptors in the olfactory bulb,” says Ben Strowbridge, associate professor of neuroscience and physiology/biophysics at the Case Western Reserve University School of Medicine. But finding evidence of LTP in the bulb was hard. “It is very difficult to know that you are activating a specific kind of connection in the olfactory bulb,” he says. “Some brain areas have very obvious pathways going in and out, and you can just stimulate those areas. But in the olfactory bulb—it’s all just a mishmash.”

Strowbridge and graduate student Yuan Gao took slices from the olfactory bulbs of rat brains and kept the tissue alive in artificial cerebrospinal fluid. They used a two-photon laser scanning microscope, which allowed them to examine cell structure down to the micron level.

First they found and identified a particular type of olfactory cell called the granule cell. Granule cells are interneurons, which pass signals between other kinds of neurons. Then they used the laser system to position a stimulating electrode at a precise point on a single granule cell.

The targeting was so exact that Strowbridge and Gao could activate inputs very near the cell body, the same as those that come from the brain’s “higher-up” smell center, the olfactory cortex. Or they could stimulate points farther away from the cell body, simulating inputs coming from the sensory neurons of the nose.

Surprising findings

When they measured the responses of granule cells, they found changes in the strength of connections between cells. After repeated stimulation, the cells’ electrical responses grew, indicating that LTP was occurring—but in a surprising way.

Scientists had predicted that odor learning would occur in the “forward synapses,” the pathways that ascend from the nose into the olfactory bulb and then into the olfactory cortex. But Strowbridge found a massive input that travels in the opposite direction, from the olfactory cortex back down to the olfactory bulb. “Olfactory learning actually takes place as an interaction between the olfactory cortex and the olfactory bulb,” Strowbridge says.

Another surprise was that LTP occurs on the surface of an interneuron rather than on a main cell in the olfactory bulb. “We think the heart of [smell learning] is changing the synaptic strength from the cortical cell to the interneuron in the olfactory bulb, that is, the granule cell,” Strowbridge says.


What happens when LTP occurs on a granule cell? A granule cell is an inhibitor; when stimulated, it prevents other kinds of cells from firing. So, when the cortex strengthens its signal to a granule cell, inhibition of adjacent cells increases.

“This opens up a whole new window of thinking about what the olfactory bulb does, because the higher levels, the olfactory cortex, may be making a prediction about what the sensory input is rather than just listening to the downstream areas,” Strowbridge says. “[The cortex] might actively be making a prediction and then using synaptic ... changes to configure the olfactory bulb to test whether its prediction is correct.”

The research was published online May 3 and in the June print edition of Nature Neuroscience. In the research report, Strowbridge suggests that olfactory learning involves a lot more back-and-forth communication between higher and lower brain areas than researchers previously suspected.

“We are just beginning to explore the function of the feedback circuits that inform low-level parts of the brain, like the olfactory bulb, about predictions made by higher-order brain regions,” Strowbridge says.

“Recent work from Diego Restrepo’s laboratory [at the University of Colorado, Denver] has shown that the signaling responses of olfactory bulb neurons are profoundly altered when an animal learns to associate an odor with a reward,” says Michael Shipley, chairman of the department of anatomy and neurobiology at the University of Maryland School of Medicine. “The Strowbridge team now demonstrates that interconnections between the olfactory bulb and the olfactory cortex possess precisely the kinds of cellular and network mechanisms needed to support that kind of learning.”