CAMBRIDGE, Mass.–In the moments after lightning streaks through the sky, we wait for the clap of thunder that experience has told us is likely to follow. In a finding that may have implications for treating Alzheimer’s disease, researchers at the RIKEN-MIT Center for Neural Circuit Genetics in the Picower Institute for Learning and Memory at MIT report in the Dec. 9 issue of Science that they have identified for the first time the part of the brain responsible for that delayed association.

The entorhinal cortex, or EC, is one of the first brain areas affected in Alzheimer’s disease.  Interestingly,early onset Alzheimer’s affects performance in memory tasks with a delay between learning and recalling items. “Our findings provide new insights into
how patients with Alzheimer’s disease develop deficits in working memory with consequent failure of the formation of episodic memory,” said study co-author Junghyup Suh, research scientist at the Picower Institute. “ In addition, by identifying the circuits responsible for the cognitive deficits in human patients with the disease, we begin to lay a potential framework for selective therapeutic intervention in Alzheimer’s disease.”

Anticipating thunder after seeing lightning or a bee sting after hearing a buzzing insect is called temporal association because the first sensory experience is separated from its associated experience by a short gap in time.  

RIKEN-MIT Center researchers found that inputs from a part of the brain called the EC layer III to the hippocampus, the seahorse-shaped part of the brain responsible for memory formation and retrieval, are crucial for temporal associations. “This study shows for the first time that the EC is important for processing non-spatial information such as a time element of episodic memory,” Suh said. The research was conducted in the laboratory of Susumu Tonegawa, Picower Professor of Biology and Neuroscience at the Picower Institute. In addition to Suh, study authors include Picower Institute postdoctoral associate Alexander J. Rivest; Toshiaki Nakashiba, research scientist; Takashi Tominaga of Tokushima Bunri University in Japan; and Tonegawa.

The EC acts as an interface between the hippocampus and the neocortex. Previous research outside of MIT had found that different kinds of cells in the EC are tied to where an animal is in space, suggesting the EC relays spatial information to the hippocampus. In addition, studies have shown that certain cells in the EC can fire continuously for tens of seconds; this type of persistent firing can help maintain sensory information (such as buzzing and lightning) through time in absence of sensory inputs.

“This study is among the first to examine the interactions between the hippocampus and its adjacent cortical areas in cognitive processes using genetic tools with great temporal and spatial specificity,” Suh said. “It also opens the door to future research with regard to how the hippocampus and EC communicate, process information and guide behaviors.”

This study is supported by the National Institutes of Health and the RIKEN Brain Science Institute.

Written by: Deborah Halber, MIT’s Picower Institute for Learning and Memory

CAMBRIDGE, Mass. – Neuroscientists at MIT’s Picower Institute of Learning and Memory have uncovered why relatively minor details of an episode are sometimes inexplicably linked to long-term memories. The work is slated to appear in the Jan. 13 issue of Neuron.

“Our finding explains, at least partially, why seemingly irrelevant information like the color of the shirt of an important person is remembered as vividly as more significant information such as the person’s impressive remark when you recall an episode of meeting this person,” said co-author Susumu Tonegawa, Picower Professor of Biology and Neuroscience and director of the RIKEN-MIT Center for Neural Circuit Genetics.

The data also showed that irrelevant information that follows the relevant event rather than precedes it is more likely to be integrated into long-term memory.

Shaping a memory
One theory holds that memory traces or fragments are distributed throughout the brain as biophysical or biochemical changes called engrams. The exact mechanism underlying engrams is not well understood.

MIT neuroscientists Arvind Govindarajan, assistant director of the RIKEN-MIT Center for Neural Circuit Genetics; Picower Institute postdoctoral associate Inbal Israely; and technical associate Shu-Ying Huang; and Tonegawa looked at single neurons to explore how memories are created and stored in the brain.

Previous research has focused on the role of synapses—the connections through which neurons communicate. An individual synapse is thought to be the minimum unit necessary to establish a memory engram.

Instead of looking at individual synapses, the MIT study explored neurons’ branch-like networks of dendrites and the multiple synapses within them.

Boosting the signal
Neurons sprout dendrites that transmit incoming electrochemical stimulation to the trunk-like cell body. Synapses located at various points act as signal amplifiers for the dendrites, which play a critical role in integrating synaptic inputs and determining the extent to which the neuron acts on incoming signals.

In response to external stimuli, dendritic spines in the cerebral cortex undergo structural remodeling, getting larger in response to repeated activity within the brain. This remodeling is thought to underlie learning and memory.

The MIT researchers found that a memory of a seemingly irrelevant detail—the kind of detail that would normally be relegated to a short-term memory–may accompany a long-term memory if two synapses on a single dendritic arbor are stimulated within an hour and a half of each other.

“A synapse that received a weak stimulation, the kind that would normally accompany a short-term memory, will express a correlate of a long-term memory if two synapses on a single dendritic branch were involved in a similar time frame,” Govindarajan said.

This occurs because the weakly stimulated synapse can steal or hitchhike on a set of proteins synthesized at or near the strongly stimulated synapse. These proteins are necessary for the enlargement of a dendritic spine that allows the establishment of a long-term memory.

“Not all irrelevant information is recalled, because some of it did not stimulate the synapses of the dendritic branch that happens to contain the strongly stimulated synapse,” Israely said.

Source: “The dendritic branch is the preferred integrative unit for protein synthesis-dependent LTP,” by Arvind Govindarajan, Inbal Israely, Shu-Ying Huang and Susumu Tonegawa. Neuron, Jan. 13, 2011.

Funding: RIKEN, the Howard Hughes Medical Institute and the National Institutes of Health.

Written by: Deborah Halber, MIT’s Picower Institute for Learning and Memory

To view the Japanese version of the press release, click here.

CAMBRIDGE, Mass. — Researchers at MIT’s Picower Institute for Learning and Memory report for the first time how animals’ knowledge obtained through past experiences can subconsciously influence their behavior in new situations.

The work, which sheds light on how our past experiences inform our future choices, will be reported on Dec. 22 in an advance online publication of Nature.

Previous work has shown that when a mouse explores a new space, neurons in its hippocampus, the center of learning and memory, fire sequentially like gunpowder igniting a makeshift fuse. Individual neurons called place cells fire in a specific pattern that mirrors the animal’s movement through space. By looking at the time-specific patterns and sequences recorded from the firing cells, researchers can tell which part of the maze the animal was running at the time.

In the current work, research scientist George Dragoi and Susumu Tonegawa, Picower Professor of Biology and Neuroscience and director of the RIKEN-MIT Center for Neural Circuit Genetics, found that some of the sequences of place cells in mice’ brains that fired during a novel spatial experience such as running a new maze had already occurred while the animals rested before the experience.

“These findings explain at the neuronal circuit level the phenomenon through which prior knowledge influences our decisions when we encounter a new situation,” Dragoi said. “This explains in part why different individuals form different representations and respond differently when faced with the same situation.”

When a mouse pauses and rests while running a maze, it mentally replays its experience. Its neurons fire in the same pattern of activity that occurred while it was running. Unlike this version of mental replay, the phenomenon found by the MIT researchers is called preplay. It occurred before the animal even started the new maze.

“These results suggest that internal neuronal dynamics during resting organize cells within the hippocampus into time-based sequences that help encode a related experience occurring in the future,” Tonegawa said.

“Previous work largely ignored internal neuronal activities representing prior knowledge that occurred before a new event, space or situation. Our work shows that an individual’s access to prior knowledge can help predict a response to a new but similar experience,” he said.

Source: “Preplay of future place cell sequences by hippocampal cellular assemblies,” by George Dragoi and Susumu Tonegawa. Nature, Dec. 22, 2010.

Funding: National Institutes of Health

Written by: Deborah Halber, MIT’s Picower Institute for Learning and Memory

To view the Japanese version of the press release, click here.

CAMBRIDGE, Mass. — Neuroscientists at MIT and Tsinghua University in Beijing show that increasing brain magnesium with a new compound enhanced learning abilities, working memory, and short- and long-term memory in rats. The dietary supplement also boosted older rats’ ability to perform a variety of learning tests.

Magnesium, an essential element, is found in dark, leafy vegetables such as spinach and in some fruits. Those who get less than 400 milligrams daily are at risk for allergies, asthma and heart disease, among other conditions. In 2004, Guosong Liu and colleagues at MIT discovered that magnesium might have a positive influence on learning and memory. They followed up by developing a new magnesium compound — magnesium-L-threonate (MgT) — that is more effective than conventional oral supplements at boosting magnesium in the brain, and tested it on rats.

“We found that elevation of brain magnesium led to significant enhancement of spatial and associative memory in both young and aged rats,” said Liu, now director of the Center for Learning and Memory at Tsinghua University. “If MgT is shown to be safe and effective in humans, these results may have a significant impact on public health.” Liu is cofounder of Magceutics, a California-based company developing drugs for prevention and treatment of age-dependent memory decline and Alzheimer’s disease.

“Half the population of the industrialized countries has a magnesium deficit, which increases with aging. If normal or even higher levels of magnesium can be maintained, we may be able to significantly slow age-related loss of cognitive function and perhaps prevent or treat diseases that affect cognitive function,” Liu said.

How they did it: To understand the molecular mechanisms underlying this MgT-induced memory enhancement, the researchers studied the changes induced in functional and structural properties of synapses. They found that in young and aged rats, MgT increased plasticity among synapses, the connections among neurons, and boosted the density of synapses in the hippocampus, a critical brain region for learning and memory.

Susumu Tonegawa at MIT’s Picower Institute for Learning and Memory helped carry out the initial behavioral experiments that showed that magnesium boosted memory in aged rats. Min Zhou’s laboratory at the University of Toronto helped demonstrate the enhancement of synaptic plasticity in magnesium-treated rats.

Next steps: This study not only highlights the importance of a diet with sufficient daily magnesium, but also suggests the usefulness of magnesium-based treatments for aging-associated memory decline, Tonegawa said. Clinical studies in Beijing are now investigating the relationship between body magnesium status and cognitive functions in older humans and Alzheimer’s patients.

Source: “Enhancement of Learning and Memory by Elevating Brain Magnesium,” Inna Slutsky, Nashat Abumaria, Long-Jun Wu, Chao Huang, Ling Zhang, Bo Li, Xiang Zhao, Arvind Govindarajan, Ming-Gao Zhao, Min Zhuo, Susumu Tonegawa, and Guosong Liu in Neuron, published Jan. 28, 2010.

Funding: This work was supported by grants from the National Institutes of Health, the National Basic Research Program of China, the National Natural Science Foundation of China and the National High Technology Research and Development Program of China.

Researchers at the Picower Institute for Learning and Memory at MIT report in the Jan. 24 online edition of Science that they have created a way to see, for the first time, the effect of blocking and unblocking a single neural circuit in a living animal.

This revolutionary method allowed Susumu Tonegawa, Picower Professor of Biology and Neuroscience, and colleagues to see how bypassing a major memory-forming circuit in the brain affected learning and memory in mice.

“Our data strongly suggest that the hippocampal neural pathway called the tri-synaptic pathway, or TSP, plays a crucial role in quickly forming memories when encountering new events and episodes in day-to-day life,” Tonegawa said. “Our results indicate that the decline of these abilities, such as that which accompanies neurodegenerative diseases and normal aging in humans, is likely to be due, at least in part, to the malfunctioning of this circuit.”

Combining several cutting-edge genetic engineering techniques, Tonegawa’s laboratory invented a method called doxycycline-inhibited circuit exocytosis-knockdown, or DICE-K–an acronym that also reflects Tonegawa’s admiration of ace Boston Red Sox pitcher Daisuke Matsuzaka. DICE-K allows researchers for the first time to induce and reverse a blockade of synaptic transmission in specific neural circuits in the hippocampus.

“The brain is the most complex machine ever assembled on this planet,” Tonegawa said. “Our cognitive abilities and behaviors are based on tens of thousands of molecules that compose several billion neurons, as well as how those neurons are connected.

“One effective way to understand how this immensely complex cellular network works in a major form of cognition like memory is to intervene in the specific neural circuit suspected to be involved,” he said.

Computing memories

The hippocampus, a seahorse-shaped brain region, plays a part in memory and spatial navigation. In Alzheimer’s disease, the hippocampus is one of the first regions to suffer damage; memory problems and disorientation are among the disease’s first symptoms.

The hippocampus is made up of several regions–CA1, CA3 and the dentate gyrus–that are wired up with distinct pathways.

The MIT study sought to determine how the interactions between neural pathways and the hippocampal regions affect learning and memory tasks.

Imagine that the three hippocampal regions are computers, and neural pathways are the conduits through which the computers get data from all over the brain. The computers perform different tasks, so the types of data processing will depend on which conduits the data travels through.

The hippocampus has two major, parallel information-carrying routes: the tri-synaptic pathway (TSP) and the shorter monosynaptic pathway (MSP). The TSP includes data processing from all three hippocampal regions, whereas the MSP skips through most of them.

Using DICE-K, the researchers were surprised to find that mice in which the major TSP pathway was shut down could still learn to navigate a maze. The shorter MSP pathway was sufficient for the job.

However, the maze is a task that is slowly learned over many repeated trials. When the mice were tested with a different task in a new environment that required rapid learning and memory formation, the researchers found that the mice with TSP shut down could not perform the task. Thus, the TSP pathway is required for animals to quickly acquire memories in a new environment. “This kind of learning results in the most sophisticated form of memory that makes animals more intelligent and is known to decline with age,” Tonegawa said.

In addition to Tonegawa, a Howard Hughes Medical Institute investigator, authors include Picower Institute research scientist Toshiaki Nakashiba; postdoctoral associate Jennie Z. Young; research scientist Thomas J. McHugh; and HHMI staff affiliate Derek L. Buhl.

This work is supported by the National Institutes of Health and the RIKEN Brain Science Institute.

A version of this article appeared in MIT Tech Talk on January 30, 2008 (download PDF).

Researchers at the Picower Institute for Learning and Memory at MIT have, for the first time, reversed symptoms of mental retardation and autism in mice.

The work will be reported in the online early edition of the Proceedings of the National Academy of Sciences the week of June 25-29.

The mice were genetically manipulated to model Fragile X Syndrome (FXS), the leading inherited cause of mental retardation and the most common genetic cause of autism. The condition, tied to a mutated X chromosome gene called fragile X mental retardation 1 (FMR1) gene, causes mild learning disabilities to severe autism.

According to the Centers for Disease Control, FXS affects one in 4,000 males and one in 6,000 females of all races and ethnic groups. The prevalence of autism ranges from one in 500 to one in 166 children. There is no effective treatment for FXS and other types of autism.

“Our study suggests that inhibiting a certain enzyme in the brain could be an effective therapy for countering the debilitating symptoms of FXS in children, and possibly in autistic kids as well,” said co-author Mansuo L. Hayashi, a former Picower Institute postdoctoral fellow currently at Merck Research Laboratories in Boston.

The study identifies a key enzyme-a chemical reaction-inducing protein-as a possible target for an FXS drug. The enzyme, called p21-activated kinase, or PAK, affects the number, size and shape of connections between neurons in the brain.

Halting PAK’s enzymatic activity reversed the structural abnormality of neuronal connections found in the FXS mice, said co-author Susumu Tonegawa, 1987 Nobel laureate and Picower Professor of Biology and Neuroscience. “Strikingly, PAK inhibition also restored electrical communication between neurons in the brains of the FXS mice, correcting their behavioral abnormalities in the process,” he said.

There are known chemical compounds that inhibit the enzymatic activity of PAK. These compounds or versions of them may be useful in the future development of drugs for treating FXS, he said.

“These are intriguing findings because the expression of the gene that inhibits PAK occurs in the third week after birth, which means that the neuronal abnormalities in the fragile X mouse are reversed after they appear,” said Eric Klann, a professor at New York University’s Center for Neural Science. “This is very exciting because it suggests that PAK inhibitors could be used for therapeutic purposes to reverse already established mental impairments in fragile X children.”

Restoring neuronal connections

Tonegawa, Hayashi, MIT graduate student Bridget M. Dolan of the Department of Biology and colleagues study the molecules that govern the formation of neuronal connections in the brain. They explore how abnormalities in these molecules could interfere with an animal’s behavior.

In the brain, small protrusions called dendritic spines on the branch-like dendrites of one neuron receive chemical signals from other neurons and communicate them to the main cell body. The numbers and shapes of dendritic spines are key to normal brain function.

FXS patients have higher numbers of dendritic spines in their brains, but each spine is longer and thinner, and transmits weaker electric signals, than those in non-affected individuals. When the enzymatic activity of PAK was inhibited in the FXS mice, abnormalities in their spine number and structure-as well as the weaker electrical communication between their neurons-were reversed.

Reversing behavioral symptoms

The FXS mice exhibited symptoms similar to those in FXS patients. These included hyperactivity; purposeless, repetitive movements reminiscent of autistic people; attention deficits and difficulty with learning and memory tasks.

“These behavioral abnormalities are ameliorated, partially or fully, by inhibiting the enzymatic activity of PAK,” Tonegawa said. “Notably, due to an elegant genetic manipulation method employed by the Picower Institute researchers, PAK inhibition in the FXS mice did not take place until a few weeks after appearance of disease symptoms. This implies that future treatment may still be effective even after symptoms are already pronounced.”

“While future studies will be necessary to further characterize the precise molecular nature of the interaction between PAK and FMR1, our findings clearly demonstrate that PAK inhibition can counteract several key cellular and behavioral symptoms of FXS,” the authors noted.

In addition to Tonegawa, a Howard Hughes Medical Institute investigator, Hayashi and Dolan, authors include colleagues at the National Institute of Mental Health and Neurosciences; the Tata Institute of Fundamental Research in India; and Seoul National University in Korea.

This work was supported by the FRAXA Foundation, the Simons Foundation, the Wellcome Trust and the National Institutes of Health.

Neuroscientists at the Picower Institute for Learning and Memory at MIT report in the June 7 early online edition of Science that they have identified for the first time a neuronal mechanism that helps us rapidly distinguish similar, yet distinct, places. The discovery helps explain the sensation of deja vu.

The work could lead to treatments for memory-related disorders, as well as for the confusion and disorientation that plague elderly individuals who have trouble distinguishing between separate but similar places and experiences.

Forming memories of places and contexts in which episodes occur engages a part of the brain called the hippocampus. Study co-author Susumu Tonegawa, Picower Professor of Biology and Neuroscience, and colleagues have been exploring how each of the three hippocampal subregions–the dentate gyrus, CA1 and CA3–contribute to different aspects of learning and memory.

Tonegawa, a Howard Hughes Medical Institute investigator and a frequent world traveler, described his own occasional experience of finding the airport in a new city uncannily familiar. This occurs, he said, because of the similarity of the modules–gates, chairs, ticket counters–that comprise airports worldwide. It is only by seeking out unique cues that the specific airport can be identified, he said. “In this study, we have revealed that learning in the dentate gyrus is crucial in rapidly recognizing and amplifying the small differences that make each place unique,” Tonegawa said.

In addition to Tonegawa, authors include Picower Institute research scientist Thomas J. McHugh; former MIT postdoctoral associate Matthew W. Jones; Matthew A. Wilson, Picower Scholar and Professor; and colleagues from the University of California at Los Angeles and Beth Israel Deaconess Medical Center in Boston.

Overlapping blueprints

In this study, the researchers used a genetically altered mouse to pinpoint how the dentate gyrus contributes to the kind of pattern separation involved in telling the difference between new and old spaces.

Researchers believe that a set of neurons called place cells fire to provide a sort of blueprint for any new space we encounter. The next time we see the space, those same neurons fire. Thus we know when we’ve been somewhere before and don’t have to relearn our way around familiar turf.

But if we enter a space very similar to one we have seen before, a new but overlapping set of neurons creates the blueprint. When there is enough overlap between the two sets, we experience an eerie feeling of deja  vu–a French phrase that literally means, “already seen.”

As we age, or as neurodegenerative disease such as Alzheimer’s advances, it becomes difficult to form unique memories for similar yet distinct places and experiences, leading to the confusion that afflicts some elderly individuals.

Forgetting fear

In experiments with mice genetically engineered to lack a certain gene in the dentate gyrus, Tonegawa and colleagues pinpointed the signaling pathway underlying the recall of specific places.

Different sets of mice were placed in two similar chambers, one of which gave them a mild foot shock. After three days, the mice began to freeze in fear in both chambers, even the one in which they had never been harmed.

Within two weeks, the normal mice learned to associate only one chamber with the foot shocks while recognizing the second as safe. The genetically engineered mice “had a significant but transient deficit in their ability to distinguish similar contexts,” McHugh said. “This study shows that plasticity–the ability to change in response to experience–in the dentate gyrus contributes to spatial learning and fine-tuning pattern separation.”

This work was supported by the National Institute for Mental Health and the National Institutes of Health.

Gene mutations governing a key brain enzyme make people susceptible to schizophrenia and may be targeted in future treatments for the psychiatric illness, according to MIT and Japanese researchers.

The work, by scientists from MIT’s Picower Institute for Learning and Memory and Japan’s RIKEN Brain Science Institute, will be reported in the early online edition of the Proceedings of the National Academy of Sciences on Feb. 20.

According to the National Institute for Mental Health, an estimated 51 million people worldwide suffer from schizophrenia. Although 80 percent of schizophrenia cases appear to be inherited, the specific genetic components underlying individuals’ susceptibility and pathology are largely unknown.

By studying genetically engineered mice and the genetic makeup of schizophrenic individuals, the MIT and Japanese scientists pinpointed the PPP3CC gene and other genes in the early growth response (EGR) gene family (specifically, EGR3) as likely suspects for causing the disease.

These genes are critical in the signaling pathway for the brain enzyme calcineurin. Calcineurin is prevalent in the central nervous system, where it plays a role in many neuronal functions whose disturbances would play into the disorganized thinking, attention deficits, memory and language problems that characterize schizophrenia.

The researchers confirmed that the PPP3CC gene is involved in diagnosed schizophrenia in Caucasian, African-American and Japanese individuals. EGR3 involvement was confirmed through a separate test.

“These data suggest that the brain signals governed by calcineurin stand at a convergent point of the molecular disease pathology of schizophrenia, and the involvement of the EGR genes reinforces this,” said co-author Takeo Yoshikawa of the RIKEN Brain Science Institute. This knowledge could lead to new schizophrenia therapeutics targeting the calcineurin system, he said.

“This study provides genetic and biological evidence that PPP3CC and EGR3, both constituents of the calcineurin signaling pathway, may independently elicit increased risk for schizophrenia,” said co-author Susumu Tonegawa, Picower Professor of Biology and Neuroscience at MIT. “These findings raised a novel and potentially important role for EGR genes in schizophrenia pathogenesis.”

In addition to Yoshikawa and Tonegawa, authors are Kazuo Yamada, Yoshimi Iwayama, Tetsuo Ohnishi, Hisako Ohba, Tomoko Toyota and Jun Aruga of RIKEN Brain Sciences Institute; David J. Gerber of the Howard Hughes Medical Institute and the RIKEN-MIT Neuroscience Research Center; and Yoshio Minabe of Kanazawa University School of Medicine in Japan.

This work is supported by the RIKEN Brain Science Institute and other agencies and institutes.

A version of this article appeared in MIT Tech Talk on Feb. 28, 2007 (download PDF).

Experts have long suspected that part of the process of turning fleeting short-term memories into lasting long-term memories occurs during sleep. Now, researchers at the RIKEN-MIT Center for Neural Circuit Genetics of MIT’s Picower Institute for Learning and Memory have shown that mice prevented from “replaying” their waking experiences while asleep do not remember them as well as mice who are able to perform this function.

The work, which has a profound implication in the century-old search for the purpose of sleep, will be reported in the June 25 issue of Neuron.

It is widely believed that memories of events and spaces are stored briefly in the hippocampus before they are consolidated in the neocortex for permanent storage. The seahorse-shaped hippocampus is thought to play a key role in learning and memory, but the precise circuits and mechanisms involved are not well understood.

“Our work demonstrates the molecular link between post-experience sleep and the establishment of long-term memory of that experience,” said Susumu Tonegawa, the Picower Professor of Biology and Neuroscience at MIT and lead author of the study. “Ours is the first study to demonstrate this link between memory replay and memory consolidation. The sleeping brain must replay experiences like video clips before they are transformed from short-term into long-term memories.”

The researchers looked at a circuit within the hippocampus known as the trisynaptic pathway, in which neuronal information passes through the hippocampus’ three main substructures before moving on. “We demonstrated that this pathway is crucial for the transformation of a recent memory, formed within a day, to a remote memory that still exists at least six weeks later,” Tonegawa said.

Creating a strain of engineered mice in which a change of diet shuts down trisynaptic circuits, the researchers implanted electrodes that monitored the activities of the animals’ hippocampal cells as the animals ran a maze and then slept.

Not-so-instant replay

While they were still awake and running, the mice formed within their brains a pattern of place cells, or neurons that were firing in recognition of the maze the mice had learned to negotiate. During their post-run sleep, particularly during a deep sleep phase called slow-wave, the specific sequence of place cells that fired during the run was “replayed” in a similar sequence.

In human studies testing the role of slow-wave sleep in memory consolidation, the group that napped after memorizing word pairs such as “fruit-banana” and “tool-pliers,” was able to recall a greater number of word pairs than those who did not nap.

This replay during sleep had been speculated, but has never been demonstrated, to be important for converting the recent memory stored in the hippocampus to a more permanent memory stored in the neocortex. “We have demonstrated that in the mutant mice in which the trisynaptic pathway is blocked, this replay process during the slow-wave sleep is impaired.” Tonegawa said. The animals were able to form long-term memories of the maze only when their trisynaptic pathways were functioning after the formation of the short-term memory.

“Our conclusion is that the trisynaptic pathway-mediated replay of the hippocampal memory sequence during sleep plays a crucial role in the formation of a long-term memory,” he said.

In addition to Tonegawa, authors are Picower Institute research scientist Toshiaki Nakashiba, Picower Institute postdoctoral associate Derek L. Buhl and Picower Institute research scientist Thomas J. McHugh.

This work was supported by the National Institutes of Health and Otsuka Pharmaceutical Development & Commercialization Inc. based in Tokyo.

MIT researchers have provided the first two-pronged evidence–based on both behavior and physiology–that a specific juncture in the memory center of the brain is crucial for rapid learning.

The work, presented Oct. 18 at a meeting of the Society for Neuroscience in Atlanta, helps explain how injury or Alzheimer’s disease result in loss of the ability to form new memories of facts and events.

The researchers, led by Thomas J. McHugh, research scientist at the Picower Institute for Learning and Memory, engineered a mouse lacking a receptor for a key neurotransmitter in the dentate gyrus. This serrated strip of gray matter is wrapped around and within the seahorse-shaped hippocampus, which is crucial in memory formation. Information arriving at the hippocampus first travels through the dentate gyrus.

“While it has long been known that damage to this region of the hippocampus affects short-term memory formation, little is understood about how each type of neuron-to-neuron connection contributes to memory in this circuit,” McHugh said.

The researchers observed the behavior of the genetically manipulated mice and measured their neuronal activity. They found that neurons at a key juncture in the dentate gyrus that receives new input from other parts of the brain help mice recognize and remember new environments.

The mice without neurotransmitter receptors at this juncture “learned normally when trained slowly with hours or days between trials, but showed learning deficits when challenged to learn the same tasks quickly, with only minutes between trials,” McHugh said. The finding shows that synapses–the connections among neurons–at the dentate gyrus are critical for rapid learning.

“This advance in the understanding of how the hippocampal circuit functions suggests possible therapeutic targets in diseases that lead to memory deficits,” McHugh said.

McHugh’s MIT colleagues on the work are Matthew Wilson, Picower Scholar and professor of neuroscience; Susumu Tonegawa, Picower Professor of Biology and Neuroscience and director of the Picower Institute; and Matthew W. Jones, a former Picower postdoctoral associate now at the University of Bristol.

This work was supported by the National Institutes of Health and MIT-RIKEN.

A version of this article appeared in MIT Tech Talk on October 25, 2006 (download PDF).