How the Brain Decides What To Memorize

Electrical ripples in the resting brain mark memories for storage, confirming advice about the importance of rest. Djorgy Buzsaky first started messing around with waves when he was in middle school.

In his childhood home in Hungary, he assembled a radio receiver, tuned it to various electromagnetic frequencies, and used the radio transmitter to communicate with strangers from the Faroe Islands to Jordan. Some of these conversations he remembers better than others, just as people remember only certain events from their past.

Now a professor of neuroscience at New York University, Buzsaki moved from radio waves to brain waves and wondered: how does the brain decide what to memorize? By studying the electrical circuits in the brain, Buzsaki seeks to understand how experiences are represented and stored as memories.

Known as "sharp wave pulsations," these waves, resulting from the firing of many thousands of neurons within milliseconds of each other, are "like fireworks in the brain," says Wannan Yang, a doctoral student in Buzsaki's lab who led the new work, which was published in the journal Science in March 2024. They are triggered when the mammals' brains are at rest - during breaks between tasks or during sleep.

It was already known that sharp wave ripples are involved in the consolidation of memories or their storage. The new study shows that they are also involved in their selection, indicating the importance of these waves in the process of long-term memory formation. This finding also provides a neurological rationale for why rest and sleep are important for information retention.

The brain at rest and awake seems to run different programs: "If you are constantly asleep, you will not have memories formed. If you are constantly awake, they won't form either. If you only run one algorithm, you will never learn anything," Buzsaki says. "You have to have breaks."

Buzsaki will never forget the first time he detected a sharp wave ripple. It was 1981, and he was a postdoc at the University of Western Ontario, listening through a loudspeaker to the brain activity of rodents. In nine years of research, he had grown accustomed to the rhythmic, melodic vibrations that awake animals made as they explored their environment. He wasn't prepared for the sudden "bong" that came from the speaker when the rodents were asleep.

What's going on. The rodents' brains in the awake state generated electrical activity that fluctuated at a constant rate. However, when they were under anesthesia, their brains seemed to irregularly produce much faster impulses.

Fast waves had been observed by other researchers Cornelius Vanderwolf described irregular waves in 1969, and Nobel Prize-winning neuroscientist John O'Keefe coined the term "ripples" to describe them in the 1970s. But it wasn't until Buzsaki saw them firsthand that he became obsessed.

Over the next decade, he spent much of his time in the lab trying to characterize these electrical bursts. In the late 1980s, researchers discovered that it was possible to get neurons to make stronger connections related to learning and memory by artificially stimulating them to make rapid bursts. For Buzsaki, these bursts were very similar to his waves. In 1989, he first hypothesized that sharp wave ripples might be part of a mechanism for forming and anchoring memories in the brain.

"He had the idea that this was not some noise activity, but something that had to do with the brain," says Michael Zugaro, a neuroscientist at the College de France who worked as a postdoc in Buzsaki's lab in 2002. "It was a foretaste of future discoveries because very little was known at the time."

In the 1990s and early 2000s, researchers took advantage of increased computing power and new tools that allowed them to record the electrical activity of more than 100 neurons simultaneously to better characterize the sharp wave ripples.

Buzsaki and other scientists have found that these pulsations appear to replicate the activity of an animal's brain, such as when running through a maze, but the pulsations oscillate 10 to 20 times faster than the original signals. According to Hiroaki Norimoto, a professor of neuroscience at Nagoya University in Japan, one 2002 paper that "made sharp wave pulsations very famous" showed that these pulsations reactivate a sequence of neuronal activity.

In 2009 and 2010, two papers, including one led by Zugaro, showed that sharp wave pulsations are involved in anchoring memories that persist for long periods of time. When researchers suppressed or disrupted these pulsations, rats performed worse on memory tasks. More recent studies have shown that lengthening or creating more ripples improves the rats' memory.

It became clear that the pulsations were being played repeatedly to solidify the memory. "The brain rehearses," says Lila Davachi, a professor of psychology at Columbia University. "Even in waking moments, your brain continues to rehearse and replay the past."

Imagine that an experience is a "melody on the piano," says Daniel Bendor, a neuroscientist at University College London. A particular sequence of neurons is triggered to record the experience, much like a pianist striking a particular sequence of keys.

Then, during sleep, the hippocampus repeats this sequence - but faster and potentially hundreds or thousands of times. The frantic, sharp wave pulsations travel from the hippocampus, which is the brain's staging area for "episodic memories" of specific experiences, to the cortex, which is involved in storing long-term memory.

However, no one has been able to explain why the pulsations spread when the animal is awake and resting. "They must serve some other purpose," Bendor recalls. Scientists had many ideas. Some hypothesized that the pulsations during wakefulness helped with planning or decision-making. Others speculated that they somehow alter or redistribute memories.

Another idea proposed by several groups was that playback during wakefulness and playback during sleep are closely related and may be the mechanism by which the brain chooses which experiences to remember.

At New York University, a room with infrared lighting held boxes of resting and sleeping mice. In the next room were mazes handmade of plastic and duct tape. One by one, the mice were placed in the mazes. They ran through it wearing electrodes that recorded the activity of about 500 hippocampal neurons, and learned that for passing certain routes they would be rewarded with water.

While exploring the maze, the mice took short breaks to rest or tidy up. And after the tests were over, they were returned to their home cage for a nap. The researchers continued to record their brain activity while they slept.

Yang analyzed the data by determining which neurons triggered during different trials. She saw many variations: some neurons triggered during early trials, while others triggered during later trials. Sometimes they triggered at different rates. This suggests that the brain captures the animal's experiences differently during individual trials. Notably, some trials were followed by bursts of sharp wave pulsations, while others were not.

She then compared the brain activity recorded while the mice were traversing the maze with corresponding ripples that appeared later. The trials that repeated more frequently when the mice were resting were the same trials that repeated when they were asleep. And those trials that were not reproduced when the mice were awake were not reproduced during sleep either.

"Perhaps the ripples during wakefulness are the memory tags that consolidate certain impressions for long-term storage," Yang says. "In contrast, those that are not marked are not reproduced during sleep and are forgotten."

In December 2023, a research team led by Bendor at University College London published findings in the journal Nature Communications that foreshadowed Yang and Buzsaki's findings. They, too, found that sharp wave pulsations occurring in mice in the awake and asleep states seem to label experiences for memorization. However, their analysis averaged the number of different experiences, an approach that was less accurate than Yang and Buzsaki's.

The key innovation of the NYU team was the inclusion of a time element in the analysis, which distinguishes similar memories from each other. The mice ran through the same mazes, but the researchers were able to distinguish between blocks of trials at the neuronal level, something that had never been achieved before.

Brain patterns mark "something closer to the event and less like general knowledge," says Loren Frank, a neuroscientist at the University of California, San Francisco, who was not involved in the study.

"They show that the brain may be creating some kind of time code to distinguish between different memories occurring in the same place," said Freya Olafsdottir, a neuroscientist at Radboud University who was not involved in the work.

Shantanu Jadhav, a neuroscientist at Brandeis University, praised the study. However, he hopes to see a continuation of the experiment that includes a behavioral test. Demonstrating that the animal forgets or remembers certain trial blocks would be "real evidence that this is a tagging mechanism."

The study leaves unanswered a burning question: Why is one experience chosen over another? The new work suggests how the brain marks certain experiences for memorization. But it can't say how the brain decides what's worth memorizing.

Sometimes what people remember seems random or unimportant, and certainly different from what they would choose if given a choice. It feels like the brain is prioritizing things based on "determined importance," Frank says. Since research shows that emotional or novel experiences are remembered better, it's possible that internal fluctuations in arousal or levels of neuromodulators such as dopamine or adrenaline and other chemicals that affect neurons ultimately choose the experience.

Jadhav supported this thought by saying: "The internal state of the body may influence an experience to be encoded and retained more effectively." But it's not known what makes one experience more likely to be retained than others. And in the case of Yang and Buzsaki's study, it's unclear why a mouse memorizes one experience better than another.

Buzsaki continues to study the role of sharp wave ripples in the hippocampus, although his team is also interested in the potential applications of these observations. For example, it's possible that scientists may be able to disrupt the ripples as part of a treatment for conditions such as post-traumatic stress disorder, in which people remember certain experiences too vividly.

"There's an uncomplicated result here - erase the sharp waves and forget what you experienced," he said. But for now, Buzsaki will continue to monitor these powerful brainwaves to learn more about why people remember what they remember.