PRINCETON, NJ – We've all been there: one bad oyster ruins seafood forever. Now, Princeton neuroscientists have pinpointed the exact "memory hub" in the brain responsible for these powerful food aversions in mice.
The new results reveal how "one-shot learning", where a single experience creates lasting memories, unfolds in rodents, which might shed light on how similar memories form in people, such as when a single traumatic event leads to PTSD. As such, work on food poisoning in mice may help potentially inform future clinical treatments in humans.
The findings were published on April 2 in the journal Nature.
Many people can recall a vivid food poisoning experience that's led them to avoid certain foods that made them ill. Christopher Zimmerman, Ph.D. has heard many such stories.
"I haven't had food poisoning in a while, but now whenever I talk to people at meetings, I hear all about their food poisoning experiences," said Zimmerman, lead author of the new paper, and postdoctoral fellow at the Princeton Neuroscience Institute (PNI).
Though the experience is common enough, the mystery that has long puzzled researchers was the time gap. Unlike touching a hot stove and feeling immediate pain, food poisoning involves a significant delay between eating contaminated food and getting sick – what Zimmerman calls the "meal-to-malaise" delay.
Working in the lab of Ilana Witten, Ph.D. , professor of neuroscience at PNI, Zimmerman first started to tease out the brain mechanisms behind learning to avoid sickening food by asking mice to try a new flavor they had never encountered before: grape Kool-Aid.
"It's a better model for how we actually learn," Zimmerman said. "Normally, scientists in the field will use sugar alone, but that's not a normal flavor that you would encounter in a meal. Kool-Aid, while it's still not typical, is a little bit closer since it has more dimensions to its flavor profile."
Mice learned that poking their nose in a special area of their cage would deliver a drop of Kool-Aid. Thirty minutes later after enjoying their first taste of the purple beverage, the mice received a one-time injection causing temporary food poisoning-like illness.
Unsurprisingly, when offered a choice two days later, the mice strongly avoided the once-appealing Kool-Aid and preferred to just drink water. What did stick out to Zimmerman and Witten, however, was where in the brain this juice/illness association was found: the central amygdala.
"If you look across the entire brain, at where novel versus familiar flavors are represented, the amygdala turns out to be a really interesting place because it's preferentially activated by novel flavors at every stage in learning," Zimmerman said. "It's active when the mouse is drinking, when the mouse is feeling sick later, and then when the mouse retrieves that negative memory days later."
The central amygdala, a small group of cells towards the bottom of the brain involved in emotion and fear learning, also processes a lot of information from the environment, including smells and tastes.
Zimmerman's results are the first to show how critical the central amygdala is at every step along the way of learning and were striking even with the very first experiment.
"I remember making the plot from the first animal and sharing it on Slack with Ilana," Zimmerman said. "She was at my desk a minute later to talk about how exciting this is."
Now that they knew where aversive flavor memories are formed, the team then traced how illness signals from the gut reach the brain. Based on hints from previous research, they identified specialized hindbrain cells containing a specific protein (CGRP) that directly connect to the central amygdala. Stimulating these cells 30 minutes after a mouse's Kool-Aid experience created the same aversion as actual food poisoning.
They also found that feeling sick caused the Kool-Aid-activated neurons to reactivate.
"It was as if the mice were thinking back and remembering the prior experience that caused them to later feel sick," Witten said. "It was very cool to see this unfolding at the level of individual neurons."
The team believes novel flavors may "tag" certain brain cells to remain sensitive to illness signals for hours after eating, allowing those cells to be specifically reactivated by sickness, and therefore connect cause and effect despite the time delay.
This research opens new pathways for understanding how the brain forms connections between distant events – with implications beyond understanding how bad shellfish memories are cemented.
"Often when we learn in the real world, there's a long delay between whatever choice we've made and the outcome. But that's not typically studied in the lab, so we don't really understand the neural mechanisms that support this kind of long delay learning," Zimmerman said. "Our hope is that these findings will provide a framework for thinking about how the brain might leverage memory recall to solve this learning problem in other situations."
Funding for the study was provided by the National Institutes of Health (K99-DA059957, U19-NS104648, U19-NS123716, DP1-MH136573, RF1-MH128776, P50-MH136296), the Howard Hughes Medical Institute, the Simons Collaboration on the Global Brain, the Helen Hay Whitney Foundation, and the Brain Research Foundation.
CITATION: "A neural mechanism for learning from delayed postingestive feedback," Christopher A. Zimmerman, Scott S. Bolkan, Alejandro Pan-Vazquez, Bichan Wu, Emma F. Keppler, Jordan B. Meares-Garcia, Eartha Mae Guthman, Robert N. Fetcho, Brenna McMannon, Junuk Lee, Austin T. Hoag, Laura A. Lynch, Sanjeev R. Janarthanan, Juan F. López Luna, Adrian G. Bondy, Annegret L. Falkner, Samuel S.-H. Wang, Ilana B. Witten. Nature, April 02, 2025. DOI: 10.1038/s41586-025-08828-z
