New research in sea slugs sparks controversy about the way that memories are stored

Can slugs teach us about human memory?

Hannah Thomasy


University of Washington

Transferring memories from one mind to another seems like something out of science fiction. But biologists from UCLA have recently found that memory transfer is in fact possible—at least in sea slugs.

In the study, researchers in Professor David Glanzman’s lab examined memory in sea slugs using something called the siphon withdrawal reflex. The siphon is a sort of fleshy tube that sea slugs and other mollusks use to propel water in and out of their bodies. When the siphon is touched, the sea slug ordinarily withdraws it into its body as a protective reflex. The scientists “trained” the sea slugs, giving them a series of light shocks which caused them to become more sensitive to future stimuli. Then the researchers tested the slugs' siphon withdrawal reflex 48 hours later. Sea slugs that had been shocked kept their siphons withdrawn for a significantly longer period of time than untrained sea slugs. This suggests that the conditioned sea slugs had formed a memory of their shocks. So far, nothing unusual. 

But then researchers removed RNA from the neurons of these trained sea slugs, and injected it into a new group—into sea slugs that hadn't ever been shocked. Weirdly, even these otherwise naive sea slugs withdrew their siphons for much longer than normal when they were poked. In other words, they acted like they remembered being shocked, even though the shocks didn't actually happen to them.

These findings have been extremely controversial in the neuroscience community because they challenge the way that most researchers understand memory. For decades, scientists have believed that memories are stored in the brain’s synaptic connections. In other words, changing the connections between neurons (or groups of neurons) is what allows us to form and store memories.

But Glanzman’s research on sea slugs turns that idea upside-down. Glanzman says that this study suggests that memory is stored within the neurons themselves, and that the synaptic changes are less important than previously believed. “According to my hypothesis,” he says, “synapses are required to express the memory—but not to permanently store it. The metaphor I like to use is that of the musical knowledge possessed by a skilled concert pianist. The pianist's knowledge of how to play Mozart, for example, is not located in her hands, but without her hands she cannot express that knowledge; in other words, there will be no music.” In other words, Glanzman is suggesting that the sea slugs need synapses to express their memory of being shocked—but that the memory is stored in the neurons themselves, not the synapses. 

A smart sea slug

Chad King / NOAA MBNMS

Other researchers aren’t so sure about this. Indeed, this research bears more than a passing similarity to the now-discredited work that James McConnell performed in the 1960s. In these experiments, McConnell took RNA from trained flatworms (which, like sea slugs, are extremely simple creatures) and injected it into untrained flatworms. The second set of worms then allegedly behaved as though they had received training. While a few other researchers managed to replicate these results, the majority could not. McConnell was eventually discredited and closed his lab in 1971.

Even if these results can be replicated, other memory researchers say that the changes in sea slug behavior don't constitute true memory, but rather a simple increase in sensitivity to stimuli. "Saying 'memory' was transferred is an overstatement," says Professor Wayne Sossin, who studies learning and memory at the Montreal Neurological Institute. "The kinds of memory that most people think of as 'memories' require a network of synaptic changes, similar to the mechanisms used in deep learning and AI. There is no evidence that this could be encoded by small RNAs or peptides that can be transferred, and I would argue that it is not conceivable that it could be encoded by small RNAs or peptides that can be transferred."

Although Glanzman’s hypothesis that memory is stored completely within neurons is highly controversial, the idea that changes within neurons are important for memory isn’t completely without precedent. Other researchers have shown that epigenetic changes, those that occur within cells, are also involved in learning and memory. Think, for a moment, of our genome as a library: In our library, our DNA is the books—the actual information that’s being stored. To express a gene, we take a book off the shelf so that our bodies can translate it into RNA, which is eventually used as a template to make proteins. The information that’s in the book (the DNA), determines the type of protein we get. But the library also needs shelves to hold and organize these books. That's the role of the epigenome. Changing the shelves doesn’t change the information in the book, but it can make the books easier or harder to get to. If we change shelves such that a specific book (or piece of DNA) is harder to get to, we’ll end up with less of the protein that that piece of DNA codes for.

Studies have shown that learning induces epigenetic changes in the brain, and that blocking these epigenetic changes can block memory formation. However, these studies don’t mean that synapses aren’t also important. Researchers think that these epigenetic changes may help neurons alter their synapses—which in turn leads to the formation and storage of memories.

Glanzman’s study did find that when epigenetic changes were blocked using a special drug, the RNA injection didn’t have any effect on the sea slugs. Glanzman concluded that the RNA injection was inducing epigenetic changes, which then caused the changes in behavior.

Glanzman has high hopes for the implications of his research. He says, “I anticipate that our approach will eventually lead to novel treatments for disorders of human memory, such as Alzheimer’s disease and PTSD.”

Samantha Yammine / Yasunori Kano

Getting from sea slugs (which have about 10,000 neurons) to humans (who have about 100 billion neurons) is a pretty big jump. To bridge the gap, Glanzman says his lab is collaborating with Professor Michael Fanselow, from UCLA’s psychology department, to see if the same type of memory transfer can be performed in rats.

It’s fascinating that even though there have been tens of thousands of scientific papers written on memory, there is still some debate about the basic principles of how memories are stored. It is truly a testament to the mind-boggling complexity of the human brain. Just how many mysteries are waiting to be solved inside the three-pound blob of tissue we each carry around inside our skulls?