Although our bodies stay stubbornly stuck in real time, our minds can flit between the past and future and jump large stretches of time in just a moment. Such feats rely on the brain’s ability to continuously store information as it happens while also retrieving dramatically condensed versions of past events. Until now, scientists weren't sure how the brain simultaneously handles these competing tasks.
Researchers from The University of Texas at Austin found evidence that in the brain’s spatial system this balancing act is accomplished via dueling electrical frequencies. Results from their study in rats suggest the hippocampus, an area crucial for memory formation, rapidly switches between the two frequencies to concurrently process the current surroundings and serve up orientation clues encoded in prior experiences. “The hippocampus has to have a way for keeping what’s actually happening and being encoded into new memory storage from interfering with recall or retrieval of previously stored memories,” explains U.T. Austin neuroscientist Laura Colgin, the study’s senior author. Her findings may have implications for the treatment of schizophrenia, and they also offer clues to another mental mystery—how the brain manages to replay a daylong memory in mere seconds.
Dueling brain waves
In the new study, published last week in the journal Neuron, Colgin’s team recorded electrical activity in a type of hippocampal cells called “place cells.” Place-cell activation corresponds to specific locations in space. As a rat navigates a maze, researchers can tell by which place cells are firing where the rat is in the maze—or what part of the maze the rat is thinking of.
Like all of the brain’s neurons, place cells produce electrical signals that oscillate in waves. In particular, past research suggests that when place cells encode and compress spatial memories they produce theta waves, which operate on a relatively slow, long-wave frequency. But these theta oscillations do not work alone. They also contain shorter and more frequent gamma rhythms nested within them like folded accordion bellows.
The gamma oscillations contribute to memory compression, explains Brandeis University neuroscientist John Lisman, an expert on the theta–gamma code who was not involved in the current study. As each wave of electrical activity pops up at the gamma frequency, it conveys new information nuggets to the interacting theta wave. One overarching theta wave sees several gamma–encoded memory cues, which effectively form a compressed highlights reel relative to the longer theta wave.
In a study published in Nature in 2009 Colgin and her colleagues described an additional level of complexity in these theta–gamma interactions in the rat hippocampus, demonstrating that the gamma waves oscillate at different frequencies depending on the task at hand. When the hippocampus communicated with a brain area that relays as-it-happens sensory information from the outside world, for example, the team saw theta signals supported by so-called “fast” gamma rhythms oscillating at 60 to 100 hertz frequencies. A second, previously unappreciated set of “slow” gamma rhythms—electrical waves in the 25 to 55 hertz range—seemed to be interacting with theta waves when the hippocampus swapped messages with another part of the brain that replays memories and plans movements through space and time, Colgin explains.
Those results hinted that fast gamma rhythms might be transmitting immediate information about the environment whereas slow gamma rhythms may shuttle information related to memory retrieval.
Clues from place cells
In their current analysis, Colgin and her colleagues found new, more robust evidence that fast gamma rhythms are indeed responsible for coding new information based on an animal’s current experiences. After recording electrical signals from hippocampal place cells in seven rats as they negotiated a short linear track over three 10-minute sessions each day, the team looked at how theta and gamma waves coincided with each rat’s actual position on the track.
When the place-cell activity matched a rat’s current location on the track, the researchers found that theta sequences interacted with the shorter wave, fast gamma signals already suspected of dealing with in-the-moment spatial information. But slow gamma waves replaced fast ones when place-cell activity represented locations ahead of the rat’s current position—perhaps reflecting the animal’s memory of the upcoming route and anticipation of the track ahead. “The idea is that the animal is actually retrieving the representation of that location before they get there,” Colgin explains.
The new results are powerful evidence that the different frequency brain waves keep incoming information and memory retrieval separate—which has implications for human conditions. If the slow gamma frequency really does keep real or imagined remembrances from interfering with new information coding and vice versa, it is conceivable that the two brain frequencies may get mixed up in conditions such as schizophrenia, Colgin says. Indeed, researchers have detected diminished slow gamma synchrony between the hippocampus and other brain regions in an animal model of the disease, boosting that theory. Future therapies could try to help increase gamma synchrony and keep thoughts separate from new sensory information—although how such a feat could be accomplished remains unknown.
How memories are compressed
In the new study the researchers also made a second discovery, which may be a clue about how the brain compresses memories. Using place-cell patterns unraveled from the theta sequences, the researchers saw a jump in the amount of track being represented per millisecond when rats were using slow gamma rhythm, even though the such rhythm produces fewer new waves of electricity in any given stretch of time than the higher frequency fast gamma rhythm.
Based on how quickly the rats seemed to anticipate upcoming sections of track, the researchers speculate that a single slow gamma wave must contain more than one piece of information, implying another level of compression within an already compressed theta–gamma code. This additional degree of compression could explain how we are able to replay memories of minutes’ or hours’ worth of activity in mere seconds.
Lisman is unconvinced of the additional-compression interpretation, although he praised Colgin and her team for uncovering functional roles for the slow gamma frequency in the hippocampus. To accomplish the ultrafast coding necessary for each gamma wave to contain more than one piece of information, he explains, neurons would have to differentiate between bits of information appearing just a few milliseconds apart—faster than current biophysical estimates say is possible.
Loren Frank, a neuroscience researcher with the University of California, San Francisco, who studies spatial coding in the hippocampus but was not involved in the study, was less skeptical of the authors' interpretation, saying it “makes a great deal of sense.”
“It says the things associated with memory may be going on very, very quickly,” he says, noting that the electrical signals making up each slow gamma signal could represent multiple levels of cellular organization capable of seriously speedy coding. “I was surprised to see the results,” Frank concedes, “but I don't think there's any reason to think the brain can't do things like that.”