Researchers are figuring out how we make memories — and how to study better

Psychologists and neuroscientists have learned a lot about the brain and how its stormy electrical activity yields human experience, but there are still far more questions than answers about how we encode memories and retrieve them.

One type of long-term memory that researchers are especially interested in is called episodic memory. A hazy recollection of blowing out candles at your fifth birthday party, the last conversation you had with a loved one, and that scandalous story a co-worker shared with you two months ago are all episodic memories.

"This kind of memory is really what defines us as individual human beings because that's our personal experience in life," Rutishauser says. These are the memories you draw on when reminiscing about — or ruminating on — the past.

Episodic memory is a key part of the human experience, but neuroscientists are just beginning to understand the broad outlines of how they’re recorded, stored, and recalled.

"We actually know very little about how this happens in the human body, or any [other animal's] brain for that matter. Very little is known about this."

The brain and the mind seem to work very differently

Researchers have figured out a few important things about episodic memories.

For one, they’re encoded by a huge number of minuscule physical changes in individual brain cells and how certain groups of neurons relate to each other. “The connections between the relevant neurons [are] strengthened, and those changes can potentially last forever," he says.

"What's thought to change in your brain when you look at something you have never seen before is [in] your synapses," the tiny gap between neurons where electrical impulses and specialized chemicals carry messages from one neuron to the next.

"What it means in physical terms is that the synapses that connect two different neurons are actually physically modified. They either get bigger or additional ion channels are inserted," changing the way neurotransmitters flow across the synapse. "There's an actual physical change."

A second major conclusion is that how those physical changes happen is quite different from how we experience the events of our lives.

"We know that memory is not continuous. It's chunked [into episodes]", he says.

This means that the way your brain records its memory of a two-hour movie (in several discrete packets) is very different from the conscious experience of sitting and watching for two hours (as one long flow of time).

What neuroscientists don’t know is how the brain knows when to close out one episode and open up a new one.

“What defines when a new chunk starts and when an old one ends?" Rutishauser asks.

He says the evidence is scant because most studies define the chunk of memory to be remembered before the research participants ever walk into the lab.

“To date, the large majority of studies on how memory works are based on showing discrete stimuli on the screen,” he says. That usually means research participants look at a blank screen and then see a picture appear and then disappear.

Rutishauser and colleagues recently devised a different kind of study to get a clearer view of how this process of chunking works.

Using a series of carefully crafted video clips and a pool of special research participants, the investigators discovered that certain neurons segment memory by quickly detecting and marking "discontinuities in the narrative,” he says.

A clinical collaboration made this work possible

Neuroscientists have an alphabet soup of methods for monitoring what’s happening inside of living brains. Electroencephalography (EEG), positron emission tomography (PET), magnetic resonance imaging (MRI), and functional magnetic resonance imaging (fMRI) are the most common ways of looking into a brain from the outside.

What’s relatively new is the ability to listen to specific neurons from inside the brain of a living person. The method is called single-neuron recording, and it uses incredibly small wires to record electrical activity as it happens in the brain.

In studies on non-humans, this is an unremarkable technique. It’s rarely used in people because of the risk inherent in embedding wires directly alongside neurons deep in the brain.

But there is an exception. Review boards allow single-neuron recording, “when it has to be done anyway for clinical reasons,” says Rutishauser.

His group, which includes researchers at the University of Toronto and Children's Boston Hospital, is one of the ten or fifteen around the world who are doing this kind of work, he says.

To learn about how the brain decides when to start a new chunk of episodic memory, the researchers worked with patients with a form of focal epilepsy that’s resistant to treatment with drugs.

For those patients, the best treatment is brain surgery. Doctors can cure the condition by cutting out the part of the brain where seizures are starting.

To figure out which part of the brain should be removed, doctors temporarily embed a series of electrodes into those patients’ brains. When the next seizure comes, doctors, "can see on which of these wires it came [to] first, and second, and third and how [the seizure] spread" through the brain, he says.

It’s sort of similar to the way seismologists monitor earthquakes and nuclear tests.

Patients usually spend two or three weeks in a hospital bed, hooked up to machines, while they wait for the next seizure to come. That’s when the participants took part in this study.

"We really should learn as much as we can from them having to undergo this [procedure],” he says.

Special neurons divide the chunks of memory

The researchers kept track as the participants watched short video clips and made memories about what they saw.

"We can see individual pulses of a brain cell," he says.

Those signals — action potentials, in the language of science — are the binary messages that individual neurons use to communicate with one another. By themselves, those signals carry very little information. But together, across all 86 billion neurons in the brain, they are what make it possible to perceive, think, remember, breathe, and many other things.

Tracking them is easier said than done, though. The brain is a noisy place, and Rutishauser's team relies on algorithms and a lot of data processing to separate signal from noise.

"The experience of listening to a neuron is kind of like trying to hone in and listen to an extremely weak radio station. It's almost buried into noise. [You're] trying to move the antenna a little and dial a little bit" to get a clear signal."

The instrument is so sensitive that getting the right signal can be impossible. Sometimes "there are wires in the walls of the hospital or a machine in the floor above" that causes too much electrical noise to hear the individual neuron.

The researchers were able to get usable data from 19 of the 20 subjects in the study. (Read IE’s coverage of the study here.)

They discovered that it’s the content of the story that drives the formation of new chunks of memory.

"What we found is that there are cells in our memory system in the hippocampus that mark cognitive boundaries," he says.

When these cells activate, a boundary between memories has formed. The boundaries that correspond to small changes in narrative, like a group of people moving from one location to another, are different from the boundaries that separate memories of two completely different stories.

"Episodic memory has this very narrative characteristic," with memories typically defined by where, what, and when something happened. Those are, of course, key aspects of any story (or science news article).

The boundaries seem to be formed when the where, what, or when changes.Specific cells the researchers identified in the electrical recording are “segmenting the memories," he says.

If you want to remember something, find a new spot

What use is this knowledge?

There won’t be an immediate clinical use. This is basic research without a specific application in mind. But that doesn’t mean it won’t have a concrete effect on people’s lives.

"One of the reasons we really don't have any good treatment for memory disorders is that we don't understand well enough how it works," he says.

That understanding is essential to developing effective treatments for all kinds of neurological diseases.

The findings may also offer an explanation of why memories seem so strongly tied to place. If a new environment causes these neurons to trigger chunks, then someone trying to remember something new can benefit from moving to a location where they haven’t spent a lot of time.

“If you study in a new place, where you have never been before, instead of on your couch where everything is familiar, you will create a much stronger memory of the material.”