ARTICLE IN BRIEF
In an emerging view of the hippocampal role in memory, based on animal studies, the hippocampus serves not just as a temporary storage place for memories destined to be lodged elsewhere, but instead is central to our ability to envision and plan for the future.
NEW ORLEANS—A rat sits in the corner of a maze, waiting to begin its search for a food reward. As it waits, the rat's hippocampal neurons, each corresponding to a spot in the maze, light up in sequence, tracing a path that, for now, exists only in its brain. Then off the rat goes, following the path it had just imagined.
That simple yet remarkable scenario illustrates the new model of hippocampal function emerging from the work of David Foster, PhD, and other neuroscientists exploring the links between memory, planning, and imagination within the resting brain. In this view, the hippocampus serves not just as a temporary storage place for memories destined to be lodged elsewhere, but instead is central to our ability to envision and plan for the future. It is an integral part of brain circuits that not only continuously replay previous experiences, but rehearse novel activities, mapping them out in time and space.
Dr. Foster, who is assistant professor of neuroscience at Johns Hopkins University in Baltimore, outlined this new view in a plenary lecture here at the AAN annual meeting earlier this year.
The classical picture of hippocampal function came from patients like “H.M.,” who famously had almost complete anterograde amnesia following surgery that destroyed most of his hippocampus. “So people began to think of the hippocampus as a temporary storage device for memory,” Dr. Foster said. But a growing body of work has begun to sketch out “a much more flexible notion of what the hippocampus is doing.”
That picture has been informed by experiments in rats, exploring the role of the hippocampus in mapping the local environment. Excitatory hippocampal neurons will fire as the animal navigates around, with individual neurons responding to different locations. The earliest studies showed that the same neurons that fired while the animal was navigating fired again in sequence while it was sleeping, leading to the idea of replay during sleep as a tool for memory consolidation. Further work showed that replay can happen not only during sleep, but also during rest: “The rat has only to pause for about a second and the hippocampus starts immediately spitting out these sequences,” Dr. Foster said.
But closer study revealed these were not merely step-for-step recapitulations of the original action, like a videotape rewound and played again. Instead, “they can be more abstract than the original experience that presumably gave rise to them,” he said, with sequences played at times in reverse. These make sense in terms of a model Dr. Foster and others propose that replay functions in part to consolidate reward-based behavior, such as maze navigation. Running the sequence backward may focus attention on the places closest to the reward, where the rat was just before finding the food. “You start with the reward position, since this is the most important position, and then learn about all the positions that lead up to that.”
But the most recent published results from several labs have led to “a new view of what this activity is,” Dr. Foster said. For example, if an animal approaches a choice point in a maze, “the hippocampal representation of position can move ahead of where the animal is, exploring now one, then the other, possibility facing him in the future.” That is, the firing patterns in the hippocampus reflect not where the animal has been, but where it might want to go.
“It has also been shown that replay can consist of sequences that were never actually experienced,” and can instead “be constructed from individual pieces that have been experienced, again suggesting that the hippocampus is not just storing and spitting out information,” but is reconstructing it in a new way. Resting in his corner, spinning out sequences of possible movements, “it looks a lot like the animal is thinking.”
This emerging view of the hippocampus aligns with other developments in understanding human brain behavior, derived from functional neuroimaging studies of the resting brain. Discovered by accident, the so-called “default network” is a pattern of activity that takes over when the subject is not actively engaged in other brain tasks. The network includes the hippocampus, the prefrontal cortex and parietal areas, as well as the posterior cingulate and retrosplenial cortex, Dr. Foster explained. The hypothesis is that activity in this network subserves both remembering and planning, and perhaps more speculatively, imagining and daydreaming.
Imagination is a new frontier for neuroscientists, but as they turn their attention there, they are finding preliminary evidence of impairments to this vital ability in patients with hippocampal damage. One recent paper described how such patients, in addition to their amnesia, were unable to describe an imaginary experience of being at the seaside, beyond the most basic elements of water and sand. Other studies have suggested that damage to the default network may underlie the impairment in imagining the perspective of others that characterizes autism. Dr. Foster sees potential in exploring hippocampal activity patterns in animal models of disease.
“Ultimately, we would like to be able to relate what we see in the animals to patients,” he said, and to better understand how the hippocampus is involved in both normal thought and cognitive disorders.
“One of the many exciting implications of this research is that it begins to unlock fundamental patterns of neural behavior of memory formation and spatial navigation,” said Jeffrey Ellenbogen, MD, chief of the Division of Sleep Medicine at Massachusetts General Hospital and assistant professor of neurology at Harvard Medical School in Boston. “By knowing these properties of the brain, we are closer to being able to enhance learning, memory, planning and navigation through engineering techniques.” He also noted that discoveries about the replay phenomenon have the potential to help better understand the functions of sleep.
An important question, he said, is whether replay mechanisms break down in Alzheimer's disease, and if so, whether that decline might permit earlier diagnosis, or monitoring success of treatments. “In the long term, knowing these fundamental properties of memory systems might enable diagnostic strategies and treatment techniques centered around patients with cognitive difficulties,” he said.
• Foster DJ, Wilson MA. Reverse replay of behavioural sequences in hippocampal place cells during the awake state. Nature 2006;440(7084):680–683. E-pub 2006 Feb 12.
• Foster DJ, Knierim JJ. Sequence learning and the role of the hippocampus in rodent navigation. Curr Opin Neurobiol 2012;22(2):294–300. E-pub 2012 Jan 7.