Progress Report 2010: Memory and Forgetting
The 2010 Progress Report on Brain Science

by Sandra A. Swanson

January, 2010

During the past several decades, researchers have worked toward a more detailed understanding of the brain structures that support memory. Their quest has revealed several of the components that control the storage of memories in brain circuits. Research in 2009 emphasized the molecular underpinnings that play a critical role.

One effort highlighted how the enzyme PKM zeta helps in sustaining long-term memories—and how the inhibition of that enzyme can rapidly erase memories.

Other research demonstrated that the protein CREB can help pinpoint the neurons linked to a particular memory. Findings with a different protein, alpha-CaM kinase II, demonstrated the chemical’s ability to erase both short-term and long-term fear memories in a very targeted fashion. Short-term memory (also called working memory) refers to information stored temporarily. Long-term memory occurs when short-term memories are encoded and stored in a more stable form, so the memory can be retrieved several weeks (and not just several minutes) after a learning experience.

A Molecular Emphasis in Neuroscience

In 1953 a patient known as H.M. unwittingly helped launch the modern era of memory research. To alleviate his severe epilepsy, doctors removed much of H.M.’s medial temporal lobe and part of his hippocampus. The surgery successfully halted the seizures. But it had an unexpected outcome too: It dramatically impaired his ability to form new memories.

H.M.’s memory loss set the stage for significant research gains. In years of subsequent study, H.M. and other patients gave scientists a better understanding of which brain regions orchestrate the conversion of short-term memory to long-term.

Recently, molecular neuroscience has gained prominence in the quest for a more precise understanding of the biological schematic of memory storage. This is no small task. In a 1998 Neuron article, researchers Brenda Milner, Larry Squire and Eric Kandel wrote: “In all the fields of all of science, the problems of cognitive neuroscience—the problems of perception, action, memory, attention and consciousness on an intellectually satisfying biological level, offer the most difficult and greatest challenge for the next millennium."1

More than a decade ago, research had already begun to illuminate the vital role of specific proteins in memory storage. One example is cyclic-AMP response element-binding protein, or CREB. In both fruit flies and sea slugs, early studies identified CREB’s role in converting short-term memories to long-term ones. By the mid-1990s, studies had already suggested a basic difference in the molecular mechanisms of short-term versus long-term memories: The latter required new protein synthesis, while the former did not.

New Research Facilitates Forgetting

Most people would like to boost the amount of information they can retain. But for some, the key to improved quality of life rests with the selective erasure of memory. The purpose is not to blot out trivial event recollections—like an awkward blind date—but to alleviate the debilitating effects of post-traumatic stress disorder, phobias and other memory-related clinical conditions.

To that end, recent research suggests that fear memories can be rapidly erased and that specific proteins have significant powers to abolish them. The results help shed light on the underlying mechanisms that govern these memories.

Progress Report 2010: Ch. 5, Fig. 2

Joe Tsien and colleague conditioned fear in mice, and then successfully deleted the fear memories by overexpressing the protein alpha-CaM kinase II. (Courtesy of Joe Tsien / Medical College of Georgia)

In October 2008, neurobiologist Joe Tsien and his colleagues published a paper that demonstrated the selective deletion of fear memories in mice.2 During the embryonic stage, the animals were injected with a DNA molecule that caused their brains (when fully developed) to constantly overexpress a protein called alpha-CaM kinase II. But the protein’s activity was controlled by carefully timed injections of an inhibitor, so the researchers could choose when it would be overexpressed.

They placed the animals in a chamber where the mice heard a tone and then received a mild shock. That conditioned the mice to fear both the chamber and the tone. Later, the rodents were placed in a different chamber—but before the tone was played again, the researchers caused the overexpression of alpha-CaM kinase II in the animals’ brains. This time, the mice did not fear the tone; they seemed to have no memory of it as a precursor to painful electric shocks. The researchers found that when alpha-CaM kinase II was overexpressed in the animals during memory recall, it could erase both short-term and long-term fear memories.

One of the most promising results was the targeted nature of this memory deletion. Tsien and his colleagues found that when alpha-CaM kinase II was overexpressed, the memory being retrieved was the only one affected—other fear memories in the mice remained intact. The selective erasure of a fear memory also opened the door to better understanding of such memories. That was the case with research conducted by the University of Toronto’s Sheena Josselyn and her colleagues. Published in March 2009, the research focused on CREB.3 This particular protein served as a marker to help researchers address a long-standing challenge: how to identify the neurons that support a particular memory. Instead of gathering in tidy, easy-to-spot bundles, the neurons linked to a specific memory tend to be scattered throughout a brain region.

Josselyn and her colleagues addressed this challenge with an experiment that trained mice to fear a tone. Josselyn’s previous work had suggested that when auditory fear memories form, they tend to recruit amygdala neurons with high levels of CREB. Building on that, the researchers wondered how fear memories would function if those CREB-rich neurons died. In this case, the animals trained to fear the tone were genetically engineered mice, with CREB-rich neurons that could be killed by a diphtheria toxin.

When a random assortment of neurons—not just CREB-rich ones—were eliminated, the mice still feared the tone. But after receiving the toxin that deletes neurons with high CREB levels, the animals stopped fearing the tone. That effect lasted over the twelve days of the experiment, suggesting permanent memory erasure. This didn’t impair the animals’ overall capacity for learning, however. They continued to encode new memories after the toxin destroyed specific CREB-rich neurons.

Some recent work with human subjects has focused on erasure as well—not of a memory but of its emotional component. Merel Kindt and her colleagues at the University of Amsterdam published a paper on this topic in March 2009.4

Previous studies pioneered by James McGaugh and colleagues at the University of California-Irvine had demonstrated that memories can be artificially altered when they are recalled, or remembered by administering a substance called propanolol. This alteration is possible through a process called reconsolidation—when a memory is retrieved, it temporarily loses stability and can be strengthened or weakened. Kindt and her colleagues studied how reconsolidation might be affected by a beta-blocker called propranolol (Inderal). Approved by the FDA as a blood-pressure drug, Inderal has also been used by musicians and other performers to combat stage fright.

The Dutch researchers didn’t force test subjects to perform onstage, but they did create a fear memory by showing them photographs of spiders and then administering a mild electric shock. After subjects had been conditioned to associate spider images with shocks, half of them received a dose of propranolol. Then all of the subjects were exposed again to the spider photos and the shock, reactivating the fear memory. The result: Subjects who received propranolol showed a loss of fear response the next day. The drug dulled the emotional component but did not delete the memory of the experience. The subjects’ “declarative” memory, which encompasses facts and events, remained intact.

The propranolol blocks adrenaline receptors concentrated in the amygdala, where fear memories are believed to be stored. By interfering with the reconsolidation of fear memories, additional treatment options may emerge for patients with post-traumatic stress disorder and other fear-related conditions.

Progress Report 2010: Ch. 5, Fig.3 A

Graphs a, c, and e show actual fear response over the three days of the experiment, while graphs b, d, and f show fear expectancy over the same period. Three types of stimuli were presented: fear condition (CS1= slide + shock), control (CS2 = slide only), and noise alone (NA = background noise present at all times). Day 1 (acquisition): Participants were shown two different frightening slides of spiders. One image (CS1) was always followed by a shock, while the second (CS2) was not. In noise alone, only background noise was presented. Participants were asked to learn how to predict when they would be shocked.

 Progress Report 2010: Ch. 5, Fig. 3B

Day 2 (extinction): Participants were given either propanolol or placebo, and were then shown the slides again. CS1-R (graphs a – d) represents the reactivation of the fear memory before propanolol or placebo was administered, whereas those shown in graphs e – f did not have the memory reactivated. Day 3 (test): After sufficient time for propanolol to be washed from the system, participants were again shown the slides. The lack of startle response by those given propanolol before reactivation of the memory (graph c) shows that the drug affected memory reconsolidation. (Copyright Nature Neuroscience 2009 12:256-258)

Memory’s Building Blocks

For decades, researchers have recognized the instability of newly formed memories—that is, short-term memories are particularly vulnerable to change and can be easily weakened. Still, the molecular and cellular underpinnings of that malleable state have remained unknown. One recent example of progress in this area is Joe Tsien’s work with the protein alpha-CaM kinase II in mice.5 Having shown earlier that overexpression of the protein could delete an established fear memory, Tsien wanted to know if elevated levels of the protein could derail short-term memory. That was the case when researchers boosted alpha-CaM kinase II activity in mice within ten minutes of engaging the animals in a learning activity—it stunted short-term memory formation. Researchers found that the timing is critical. When the same alpha-CaM kinase II] alteration took place fifteen minutes after the learning activity, it did not result in a disruption of short-term memory.

In the realm of long-term memory, one molecule that has generated significant interest is PKM zeta. This enzyme became prominent in 2006, when Todd Sacktor and his colleagues at SUNY Downstate Medical Center published a paper suggesting that PKM zeta was needed to maintain long-term memories.6

“It used to be thought that long-term memory was due to structural changes in the brain that were permanent because they were structural,” said Sacktor. “The idea was that once you make a synapse, that’s it—you can forget about the memory maintenance part.”

In contrast, Sacktor’s work shows that PKM zeta is an integral part of long-term memories’ molecular upkeep. In 2009 he reinforced his previous PKM zeta research with a study that inhibited the enzyme and then assessed the effect on long-term memory.7 Sacktor and his colleagues set this up by creating a taste aversion in rats. They exposed them to a new taste—such as saccharin—and then followed it with a lithium dose that sickened the animals. Not surprisingly, the rats avoided drinking from water bottles with saccharin—that is, until they were injected with a PKM zeta inhibitor, called ZIP. After the rats received a dose of ZIP, their memory of the taste aversion was rapidly erased. The researchers were able to delete three-month-old memories in the rats, but found that inhibiting PKM zeta had no effect on shortterm memory. This bolstered previous work, published by Sacktor and his colleagues in 2007, in which the inhibition of PKM zeta wiped out rats’ taste-aversion memories several weeks after they initially formed.8

Progress Report 2010: Ch. 5, Fig. 4

Using a compound called ZIP to inhibit the memory-maintenance protein PKM zeta, scientists caused rats conditioned with a taste aversion to forget the conditioned stimulus. Rats were trained in a single conditioning session and then given ZIP either 3 or 7 days later, or trained in two sessions one day apart, and given ZIP 25 days later. In all cases, the rats infused with ZIP (black) forgot the conditioning, as indicated by their low level of aversion when compared with rats not injected with ZIP (gray). (Copyright Science 2009)

PKM zeta poses many unanswered questions for researchers. One example is its possible connection with the memory-loss process in Alzheimer’s. One study located the enzyme within the tangles found in the brains of Alzheimer’s patients.9

Sacktor posed a broader question: “Is it possible to enhance people’s memories by giving a drug that will increase the synthesis of PKM zeta?” He said it could be, citing other researchers who found that, in fruit flies at least, an influx of PKM zeta can convert shortterm memories into long-term ones.10

Improving Memory for Better Health—or a Competitive Edge

Aside from the implications for treating certain disorders, neuropharmacology products that affect memory and cognition are also sought by healthy people looking to improve performance at work or in school.

An online poll, published in April 2008, hints at the desire to find a shortcut to that intellectual edge, particularly in competitive fields. Conducted by Nature, the poll invited academics and scientists to disclose whether they had sought a memory boost via drugs approved for treatment of narcolepsy and ADHD. Of the 1,400 respondents, one in five reported taking methylphenidate (Ritalin), modafinil (Provigil) or beta-blockers to improve memory and concentration.11

Anjan Chatterjee, neurology professor at the University of Pennsylvania, anticipated this trend in 2004 when he published an article titled “Cosmetic Neurology” in a science journal, asserting that government regulation of the development of cognitive enhancers seemed unlikely.12 He expanded on that discussion in 2009, arguing in a British Medical Journal article that it is unacceptable for people to take methylphenidate (Ritalin) for performance-enhancement reasons.13

Aside from ethical concerns, Chatterjee wrote that the most obvious objection to this use of methylphenidate is that “the cognitive benefits are minimal and the medical risks are not.” He noted that the U.S. Food and Drug Administration gave methylphenidate the most alarming of possible health warnings because of its high potential for abuse as well as risks of sudden death and serious cardiovascular complications. Medical side effects aren’t the only potential problem, he noted: “There are also possible cognitive trade-offs. For example, greater focus from long term use of methylphenidate could plausibly produce a loss in creativity, which generally requires a loosening of mental boundaries. Such trade-offs are rarely considered or investigated.”

Another treatment that may one day be sought for memory enhancement purposes is deep brain stimulation, which involves implanted electrodes in the brain (see chapter 2, “Deep Brain Stimulation”). Used most commonly for Parkinson’s disease, this treatment is also being studied for conditions such as intractable depression, cluster headaches and phantom limb pain. It may also have the potential to enhance the existing memory circuits of early-stage Alzheimer’s patients—or give a memory boost to otherwise healthy individuals.

A glimpse of that potential appeared in the Annals of Neurology in January 2008, in a documented case at a Canadian hospital. An obese man sought DBS as a possible treatment to curb his appetite. With the electrodes stimulating his hypothalamus, the man’s working memory showed significant improvement—and his IQ increased by nine points.14

Changing a patient’s diet could be a less invasive approach to boosting memory, according to a recent study in the Proceedings of the National Academy of Sciences. In Germany, a group of healthy adults ages fifty to eighty demonstrated a 20 percent improvement in verbal memory scores after reducing their calorie intake by 30 percent during a three-month period.15

The Pursuit of Plasticity

During the past decade, assumptions about the brain’s ability to rewire neurons’ connections in response to experience have changed dramatically. Synaptic plasticity was once primarily associated with youth. But research in the late twentieth century—including a 1999 study of neuron growth in the hippocampus of adult monkeys—marked the start of a shift in that scientific outlook.16 Now studies have shown that plasticity can extend well beyond childhood, if only in a relatively diminished capacity.

To recruit plasticity to help prevent age-related cognitive decline, researchers must first piece together the structural and functional changes behind this basic framework of memory formation. A molecule called myosin Vb may be indispensable to that framework, according to a study published in October 2008. Researchers observed that the myosin molecule in a rodent’s hypothalamus facilitated the movement of new receptors, which in turn strengthened synaptic connections. When researchers blocked myosin, it prevented the addition of new receptors. This molecule could represent a new target for the treatment of diseases involving synaptic abnormalities, such as Alzheimer’s or autism.17

By studying postnatal mice, researchers have also identified a protein that triggers plasticity in visual systems. Published in August 2008, the research highlighted the role of orthodenticle homeobox 2, or Otx2. This protein facilitates the maturation of parvalbumin cells, located in the visual cortex, which help rewire the brain in response to visual input. Evolutionary biology shows that Otx2 developed from a protein in fruit flies called orthodenticle, which helps determine head development and was first described by researchers R. Finkelstein and N. Perrimon.18 Otx2 is a protein that facilitates the maturation of parvalbumin-expressing cells. Located in the visual cortex, these neurons help rewire the brain in response to visual input.

One surprising result from this study: Otx2 is synthesized by the retina, then migrates to the cortex. Essentially, the eye is dictating plasticity timing for the brain. The study’s senior investigator, Takao Hensch of Children’s Boston Hospital, speculates that the visual system may not be the only sensory system that includes a molecular plasticity trigger like Otx2. If researchers can one day control the timing of plasticity, they could address a range of needs, such as learning a language or recovering from a stroke.19


The story of memory research has been long and now turns out to be wide as well, involving such disorders as post-traumatic stress disorder, which concerns the persistence of unwanted memories, and Alzheimer’s, in which memory deficiency is a symptom of the problem.

For those patients, memory research has dramatic potential to improve quality of life. To that end, scientists continue to work toward bridging the gap between laboratory results and treatments intended for clinical applications. But that’s not the only challenge. Scientists—and society at large—must resolve ethical questions that accompany the manipulation of memory. If we do find a drug that can give memory a turbo boost, how should it be regulated—and will it carve out an intellectual divide that creates unfair advantages?

To optimize the way memory functions, scientists must continue to map out its inner workings at the molecular level. Recent research underscores the role of specific proteins for preserving memory. In addition, scientists have shown that rapid erasure of memory may someday become a reality.

As the next chapter of memory research unfolds, scientists like Joe Tsien expect it will be action-packed. “For the next five years, you will really see an explosion of our understanding in terms of the general organizing principles of memory,” said Tsien. “Once you understand that, then you can have a whole new way of looking at all sorts of memory disorders. The electrical patterns in the population network of these neurons—that’s where I think the interesting things are; that’s where the important work lies ahead.”