How does a brain hold on to a memory? There’s evidence for a number of processes, from potentially transient changes in gene expression, through long-term changes in DNA packaging, and up to alterations of the connections among cells. Complicating matters further, none of these processes is mutually exclusive, so all of them might be involved in one context or another.
That complexity makes one of this week’s headline stories—”Memory Transferred between Snails,” to use one example—a bit surprising. If it were that easy, doesn’t it imply memories have to be relatively simple?
The researchers behind the headlines did something impressive, but it certainly wasn’t transferring a memory as we typically think of it. As we’ll explain here, the work tells us something about one element of memory, but it probably won’t end the debate about which processes let us recall familiar faces and places.
The persistence of sea slugs
To begin with, much of the coverage has claimed that memories were transferred between two snails. This is wrong; is a sea hare, a group within the sea slugs. Thanks to the pioneering work of Nobel Laureate Eric Kandel, these sea slugs have been used extensively to study memory. That’s not because you can train them to turn left at the third coral; instead, imprints of past events can influence some of their current instinctual reactions.
Specifically, these animals have a siphon that they can extend out from a protective pouch. Poke it, and they’ll quickly withdraw it again. If the animals were subject to other stresses—one option is to implant electrodes elsewhere and give them a bit of a shock—this reflex could be sped up and the siphon would get withdrawn for longer periods. This effect worked even after the shocks were stopped, indicating that formed an associative memory. Because its nervous system isn’t as complex as a mammal’s, researchers were able to use this to get some indications of how memories can form at the level of individual neurons.
It’s this system that researchers at UCLA used to test the possibility of a memory transfer.
Their approach was relatively crude. They performed the sensitization training as described above so that they had a group of animals with an enhanced siphon response (and a matched control group). Once trained, the relevant areas of the animals’ nervous systems were dissected out, and the researchers isolated the RNA from these cells.
That RNA is a complicated mix of material that helps catalyze key reactions, RNAs that encode proteins, and RNAs that are involved in regulating gene activity. So it’s important to note that the authors don’t entirely know what this batch of RNA is providing.
Thanks for the memories
But when the RNA is injected into an animal that hasn’t been sensitized, it changes them. Following injection into the same area of the nervous system that the RNAs originated, the animal behaved as if it had been sensitized. RNA from unsensitized animals didn’t have this effect, indicating that it’s not something caused by the injection itself. The injection also didn’t place anything particular inside cells; the nerve cells in the area must be picking up some material on their own.
Whatever is happening, it seems to involve changes in how the alter their DNA through epigenetics. In some animals, the researchers mixed the RNA with an inhibitor that blocks a chemical modification of DNA bases. When the inhibitor was present, the RNA had no effect on the animals it was injected into. So if the animals can’t change their DNA, they won’t be able to take in the new memory.
The authors also dissected out both sensory and motor neurons from unsensitized animals and exposed these to the RNA. In these experiments, sensory neurons became sensitized and fired for longer in response to a stimulus. The motor neurons weren’t changed by being exposed to the RNA. So the effect doesn’t seem to involve a general increase in the activity of neurons.
In part because of that specificity to sensory neurons, the authors argue that they’ve physically transferred a memory from one animal to another. And if we’re to define memory as simply meaning “do the two animals act as if they’ve had the same experience?” then a memory was transferred.
But of course it’s a little more complex than that. Increasing a neuron’s response based on a history of activity is something that’s extremely useful, and it’s done in a variety of contexts, not just in memory. And you would expect that nerve cells with different functions, like the sensory and motor neurons tested here, might handle controlling that response in different ways. Looked at a different way, the authors showed that they could transfer a heightened state of activity from one set of neurons to another—very interesting but far less provocative.
And it’s also important to note that a heightened response in an instinctual reaction isn’t what most people view as a typical memory. Things like knowing where to turn while driving or recalling a past meal involve multiple systems within our brain and probably require a storage system that’s significantly more complex than the one shown here. So it’s important to recognize that this work doesn’t mean we’re anywhere close to transferring anything like those memories.
None of this is meant to detract from this research. Within the definition of memory that the authors are working with, they’re almost certainly right that they’ve transferred it, and they have shown that it can be done by an unexpected vehicle (the RNA found in nerve cells). But this is just a small step in what has already been a long process of untangling how sophisticated memories are stored in something as complex as the human brain.