The sound of memory
What the hippocampus knows about things you hear
You remember things that happen to you, right? Episodes of your life play out before your mind’s eye. Just like a soap or sitcom, they have colour, sound, movement; they have a location — they happen somewhere. Some even have an accompanying laugh track from the audience, like when you shook the ketchup bottle without first holding down the lid.
Neuroscientists, in a typical demonstration of the biting wit we’re renowned for, call this memory for epsiodes of your life… episodic memory. And you thought being a neuroscientist was hard. We know the hippocampus is crucial for episodic memory. Most famously from the case of Henry Molaison, who had operations to remove his hippocampus on both sides of his brain to cure his extreme epilepsy. It worked: but he could not form a memory of anything that happened to him.
Weirdly, while we know hippocampus is essential for making memories of things that happen to us, neuroscience mainly studies what its neurons know about space (as in, where you are in the world and where things are in relation to you — not planets, stars, and aliens). Now a new study has opened a whole new chapter: the hippocampus knows about sound.
Making an episodic memory needs two ingredients. First, your brain needs to represent all the things that are happening right now, the episode. Then it needs to store them, make the memory. Yet our detailed knowledge of the neurons in the hippocampus barely mentions either how it represents or makes a memory of all the things around us, of the sounds, colours, and smells. No, our understanding of what the neurons in the hippocampus do is based almost entirely on the study of how they represent the position you’re in, right now.
If this were the only thing the hippocampus did then the only things we would remember would be the locations in which stuff happened. Anecdotes would consist entirely of “I arrived at the house, then I went into the main room, walked around there for five hours, and then I woke up on a sofa. What a party!” Yikes. So do the neurons in the hippocampus know about stuff around us — the colour, sounds, and smells? Or is the hippocampus the world’s most obessesive (real) estate agent, laser-focussed on location, location, location?
A new study by David Tank’s lab (carried out by Dmitrly Aronov and Rhino Nevers) tackled this question. They simply asked: do neurons in the hippocampus know about other things we experience? In this case: do hippocampus neurons know about the experience of pushing a lever to change a sound? They trained a bunch of rats to push a lever while a single, continuous tone sounded. The pitch of the tone increased for as long as the rats held down the lever. Why would a rat want to do such a thing? Because if the rat moved the tone up close to a specific pitch, it got rewarded with a drink. So thirsty rats learnt to push the lever just long enough to get the tone’s pitch were it needed to be. Better still, once they understood this basic set-up, when the target pitch was changed they adapted quickly, learning the new target. (And by changing how fast the pitch changed, the experimenters could show that the rats learnt the pitch itself, not how long to hold down the lever for). Rats are smart.
Once the rats had this down cold, the experimenters started recording from groups of neurons in their hippocampus. And they found a menagerie of neurons. Some neurons fired only at the start of the lever press. Some neurons fired only when the reward was delivered. Quite a lot of neurons fired only for a small range of pitches. So each represented a “sound field”, a particular “place” in the set of all possible pitches. And as the rat increased the pitch of the tone, so each neuron fired in turn, as the pitch passed through its sound field.
Great, so hippocampus knows about pitch. So what? Actually, not just the pitch. The neurons were representing the experience of the pitch changing. The experimenters showed this in a simple, elegant way: they played the rats the same tone and increased its pitch, with no lever or reward — some rats just sitting about, listening to the dullest music of all time. The exact same neurons responded with? Nothing. So the neurons were not just responding to a particular pitch (like they would do in, say, auditory cortex). They did not represent just any sound, but the personal experience of the sound in that particular moment, a sound that was relevant to the animal in that moment.
A straightforward answer, for once. Do neurons in hippocampus represent experiences other then “where you are”? Yes: they also represent “the sound is changing”.
This raised an immediate question: how does the representation of pitch relate to how these same neurons represent location? Does hippocampus have one set of neurons for place and one for sound? Or do these all get mixed together into one set of neurons? So the experimenters took the same rats, put them in a big box they could run around in, and recorded from the same neurons. The answer was: um. Some neurons that represented sound also represented specific locations, like the corner of the box. Some represented only location. Some only sound. And some, none of the above. There are not then different dedicated sets of neurons for sounds and for places in the hippocampus. No, all the information is gathered together, just as you’d need to form an episode of your life (if your life was pushing a lever in a small box. Which it might be. Let us know in the comments…).
(And a thought for the neuro-hardcore. Place cells remap when you put a rat in the same box, but in a different room. They change which locations in the box they represent. So what would happen if the pitch-finding task was reversed? Imagine if, after training on the low-to-high pitch version, the task switched to a high-to-low pitch version. Would the same cells fire in reverse order, as the pitch goes down through their “sound fields”? Or would the sound fields change to different pitches — a remapping — as the context is now different?)
In many respects the results of this study are not surprising. We know hippocampus does memory. Many had assumed that location was just one aspect of memory (its dominance in neuroscience explained by the fact that it is an easily accessible aspect of memory, as we know how to relate the firing of neurons to the location of the animal). Indeed, we’ve long known that a neuron’s firing in a specific location could be changed just by changing the smell of the arena, or even the current rules of the maze. Why would hippocampus neurons need to do that if they only represented the place, and nothing else? So it always seemed likely that hippocampus would represent more than just location.
What this new study does is give an extra, hard shove away from the idea that the hippocampus is about space, to the idea that space is just one aspect of an experience, of an episode of life to be recorded. This study immediately opens up a range of new things to test about forming memories of episodes in your life. What exactly does the hippocampus represent? Smells? This seems likely, as we know how a whiff of strong cheese immediately whips your mind back to that experimental burger you once made with stilton and chèvre (seriously, what were you thinking?) (And not least because the bit of cortex that deals with smell has privileged direct access to the neurons in hippocampus). What about touch? Taste? Colour? To form a memory of an episode of your life, surely it needs to know about all these things. Surely to vividly remember that party last year, your hippocampus needed to know about the thumping music, the taste of the prosecco, the smell of the pizza, the feel of the fluffy sofa cushions as you sank into oblivion?
But the brain is constantly surprising us. Things that we think must be true about how neurons work often end up trashed by reality. So it is not a foregone conclusion that the hippocampus must represent all the sights, sounds, and smells of the world. And we know less about the brain than you think. No, less than that. A bit less than that too. About as much as you know about how to build the screen you’re looking at, when starting from the raw materials dug out of the ground (“we need some silicon, and some metals, and do some melting, and something-magic-with-LCDs-and-sensors-and-processors, and voila! Screen!”. Neuroscience has the exact equivalent: “we have some neurons, and some connections between them, and they fire, and something-magic-happens, and voila! Memory”!). All we’ve talked about here is the representation of what is happening right now. How does this become a memory of an episode in your life? We don’t know. But when we find out, our hippocampus will remember that experience for us.
Want more? Follow us at The Spike
Twitter: @markdhumphries
