An interview with Dr Christopher Zimmerman, Princeton Neuroscience Institute, conducted by Alison Cranage

Most of us have had the experience – hopefully singular, not multiple – of being put off a particular food after it made us ill. The subsequent dislike and avoidance of that food can last a long time. Personally, I didn’t like or eat hazelnut chocolate spread for 20 years after eating too much of it and then being sick when I was about 10. Or, perhaps you avoid a specific restaurant – if you were ill after eating there, you’re probably never going to go back.

It’s a sensible approach, evolutionarily. But how does the brain know what it was that made you sick? How do we make the connection, many hours later, between something we ate and illness? This type of learning isn’t restricted to humans – it’s an advantage for any animal to learn what is safe to eat, and what isn’t.

Dr Christopher Zimmerman, a postdoctoral researcher at the Princeton Neuroscience Institute, is fascinated by body-to-brain communication and how it influences nearly every aspect of our behaviour. 

His research currently focuses on the signals our brains get from eating and drinking, but he is also interested in signals from other parts of the body that may influence our behaviour, mood, or even choice of partner. 

Chris recently spoke at SWC as an ENSS winner, and in this Q&A he discusses his work on how our bodies and our brains connect.

What got you interested in studying body-brain communication, and how we learn from it?

A few reasons really motivate my interest in understanding how we learn from feedback from our bodies. First, it has important implications for conditions like eating disorders and obesity, which often involve learned preferences or aversions to food that influence what and how much we eat. Another reason is that it impacts basic aspects of everyday human experience. Like many neuroscientists, I want to know why we behave the way we do, and especially how our bodies contribute to that. 

But the biggest reason that led me to study this specific question is the unique set of time scales over which we learn from our bodies. While most forms of learning that scientists study in the lab – think of Pavlov and his dogs – involve cues and outcomes that are never separated by more than a few seconds, feedback from the body can arise minutes or even hours after a meal but still cause us to learn extremely specific associations. This creates a really fun set of challenges for the brain, and I want to know how the brain has evolved ways of overcoming those challenges.

What are the gaps in understanding that you're trying to address with your work when it comes to feedback from the gut after eating?

We know a lot about the pathways that take taste and odour information from our mouths and noses, and illness or nutrient signals from our gut, and send those to the brain. 

What we don't know is how those signals then come together in the brain to form a memory that's specific for the association between a particular flavour, or a particular context – like a restaurant – and the feeling of food poisoning. The gap that I'd like to address is understanding this memory formation and learning process.

Could you tell us about your recent findings and how the brain learns to dislike a certain flavour after illness?

The question that this project addressed is how can we learn an aversion towards some food we ate in a meal, if the feedback signal that causes us to change our preferences doesn't happen for a really long time, like feeling ill when we get food poisoning.

Our main finding is that illness signals from the gut are able to reactivate and strengthen the representation of flavours from a recent meal in a region of the brain called the amygdala. This gut-driven change in the way the amygdala represents a flavour is important for the aversive learning process.

How does the brain link food consumption with the delayed signal? 

One striking finding from this project is that signals from the gut are able to selectively turn back on a flavour representation in the brain, even after a long delay after a meal ends. But, all of the other things that are encoded in this part of the brain don’t seem to be reactivated. Where does this selectivity come from?

We don't have a concrete answer for how these delayed reactivations happen, and there are different potential models that could explain it. One possibility that I think is very interesting is that during a meal when you encounter a food that you think may be more likely to make you sick – like novel foods for the mice in our experiment – the set of cells that encode that flavour may be imprinted, biochemically, in a way that makes it easier for the gut to engage with them later on. 

This might then increase the overall excitability of those cells, or even specifically increase their responsiveness to illness inputs. These biochemical changes would need to last for a long time after a meal, so that the set of flavour-coding cells remains imprinted as being more excitable and all the other cells kind of fade away and go back to their normal excitability. 

We have some evidence for this model, but there are a lot of details that we still need to figure out. 

Your findings are in mice – do you think the same mechanism operates in the human brain?

There’s good reason to think it does, because humans have the same brain structures – like the central amygdala – and these structures are implicated in learning in humans as well. 

At the same time, this behavioural phenomenon is very robust across different animal species, including humans. Many of us have experienced food poisoning and then disliked and avoided a flavour for a long time afterwards. 

In mice, I focus on their ability to easily learn aversions to novel foods after food poisoning, but not to familiar foods that they’ve already learned are safe. Novelty is an incredibly salient trigger for learning in lab mice, probably because they haven’t had many different things to eat in their lives. 

In people, it's likely more complicated. We have different priors about which foods might make us sick. We may think there might be bacteria growing on potentially contaminated vegetables sometimes, or oysters or seafood may be spoiled – things like that. And we also form richer memories – it's more complicated than just disliking a food itself. For example, we also learn to avoid the restaurant that served us that food – maybe we’re more likely to blame the chef or the restaurant than the taste of the actual food item.

So I think that the system as a whole will be much richer in humans, but I do expect that the core circuit and core mechanism that we are studying will be conserved, although we don't know that for sure yet.

How distinct is the representation in the amygdala of a certain flavour? If you’ve learnt to avoid a grape-flavoured drink, will you also dislike grapes?

We don't have a complete answer to that question yet, at least at the level of the amygdala neurons. Behaviourally, animals will generalise over similar flavours and similar concentrations of a flavour to the one that made them sick.

That sort of generalisation could arise because the representations in the amygdala are more similar for foods that share a flavour profile or are chemically related. This is something I'm excited to work on in the future, because we don’t yet understand the dimensionality or structure of how different flavours are encoded in this part of the brain, which could have important implications for how we learn.

For example, can all of the different flavours we might encounter in the world be discriminated from neural activity in the amygdala? Or is that something that is happening upstream, for example in the cortex. 

In addition to knowing how different tastes are encoded in the cortex and the amygdala, I also want to understand how learning changes the relationship between these flavour-coding amygdala neurons and their main output pathway – called the BNST – because that region shows some very interesting differences in activity before and after illness. 

Neuroscience is at an exciting point, with the development of Neuropixels probes and other tools, where we can now record from all of these regions simultaneously in animals in a way that lets us ask these questions. This would have been impossible 5 or 10 years ago.

Is this mechanism involved in other kinds of illnesses, not just food poisoning, but if you have an allergic reaction or the flu?

There are a lot of ways you can feel ill. You might have a virus in your stomach that causes vomiting or diarrhoea. You might have a bacterial infection or have eaten something that you're allergic to, both of which will engage your immune system. Or you might be pregnant and have a very different hormonal or physiological state in your body that leads to morning sickness. 

We don’t fully understand how different forms of illness in the gut impact the brain yet, but my working model is that all of those ways of feeling ill will converge in the brain into a singular representation of illness.

That unified illness signal can then drive learning aversions to foods, and also control lots of other behavioural and physiological responses that are common to many forms of illness.

However, we also know that there must be changes within the brain that are different depending on the kind of illness you've had. We know this because after learning, re-encountering a flavour that was associated with a food allergy or a bacterial infection can turn on the immune system pre-emptively. So somewhere, the brain has access to this knowledge. 

These illness-specific responses for the most part remain a black box right now. We know behaviourally or physiologically that different forms of illness can have different effects on learning and on behaviour.

But we don't know much about how that happens in the brain. Is there a food or flavour avoidance system, which is what I've been describing, which then also turns on the immune system? Or are they completely separate systems?

What other kinds of body-to-brain communication are you interested in?

Part of what makes body-brain communication incredibly fascinating to me is how much still remains unknown. Many organ systems send sensory signals to our brains, and the GI tract certainly isn’t the only one that can create long-lasting changes in our behaviour. But how other organ systems might influence learning isn’t as well studied.

I think the most profound example of this is the reproductive tract. These organs play a clear and obvious role in mating behaviour, as well as in other aspects of people’s lives like menstruation, menopause and childbirth. 

But we don't know much about the nature of sensory signals from the reproductive tract or the cells that send them to the brain. 

This means that there’s a big opportunity to understand the origins of these signals, when they arise during mating, and how they change the way the brain represents potential partners. As well as how they contribute to other aspects of physiology.

Do you think reproductive organs play a role in choosing a partner?

There are two particular situations where sensory feedback from the reproductive organs probably plays a role in choosing a partner.

One is sexual satiety – the reduction in sex drive after successful mating. This is a phenomenon seen in most mammals that have been studied. And, in rodents, for example, this requires feedback from the reproductive tract, especially in females. 

Second, reproductive signals may contribute to the formation of long-term pair bonds in some species. Prairie voles for example, make partner choices that last a very long time and reject other potential partners – a rodent model of monogamy. We don't have a complete understanding of the cues that are important for pair bonding in voles, but we know that sex is a key component. And that makes me suspect that feedback from the reproductive tract during mating might also play a role in pair bonding.

So it will be exciting to tease apart how signals from the reproductive tract influence the way an animal sees potential partners, and decides whether or not to initiate mating or even to form a lifelong bond, and especially how that might vary across species.

Biography

Dr. Zimmerman is currently a postdoctoral fellow in Dr. Ilana Witten's lab at the Princeton Neuroscience Institute. He was previously a PhD student in Dr. Zachary Knight's lab in the Department of Physiology at UCSF. He develops new experimental and computational tools for studying how the brain and body communicate, with a focus on the role of body–brain interactions in learning.