Manipulating Memory: Brian Wiltgen
Researchers generally believe that memories are encoded as vast networks of neurons become active in specific sequences. Remembering those events appears to involve the reactivation of those networks of neurons. In recently published papers, Wiltgen and his colleagues have provided strong evidence supporting this theory of memory, and have mapped out neural pathways in a small brain structure involved in memory.
Wiltgen’s research involves combining cutting-edge scientific techniques and recently developed tools to examine the activity in networks of neurons while animals are learning, and later remembering, experiences. Wiltgen emphasizes the importance of “tool makers” in his field, but describes his own role as that of a “tool-user.”
Tools and techniques
In order to understand his work, one must first review his techniques and tools. The first main technique is neuron labelling or tagging. This process involves genetically-engineered animals and Green Fluorescing Proteins (GFPs), which are used to label neurons in the animals.
When neurons become active in a genetically engineered animal, they express proteins, such as cFos. Other researchers have created mice that have a second protein, tTA, attached to the naturally-occurring cFos. The tTA protein is useful because it can be regulated with dietary supplements, allowing researchers to control when cFos-tTA proteins can be expressed.
Green Fluorescing Proteins reflect a fluorescent light. The first GFP was extracted from the jellyfish Aequorea victoria, and many variants have been found in nature or synthesized. The “tool-makers” Wiltgen extolled have also created mice with GFPs attached to the proteins expressed by active neurons, while others created GFPs that degrade more slowly than early variants. Wiltgen bred mice that express a long-lasting GFP with mice that have the cFos-tTA regulation gene. This allows him to turn off the expression of GFP with food supplements, and to measure the GFP in neurons weeks after it is expressed.
Wiltgen’s second main technique is optogenetics, so-called because it involves using lights and genetics to control activity in neurons. The activity of neurons depends on electrically charged ions passing into and out of the cells through ion channels. Researchers have created mice that have light-sensitive ion channels in their neurons. Optogenetic techniques can be used to activate those neurons, or prevent them from activating by opening and closing these ion channels.
“We want to understand the rules that govern how the brain encodes, stores, and retrieves memories.”
Equally critical to his research, Wiltgen says, is the diligent work of graduate students and post-doctoral collaborators, who must tediously count the neurons that are labelled with the GFP—a process that can take weeks for one group of animals, even when examining only a tiny sub-region within a small structure in the brain.
In a recent study, Wiltgen and his colleagues provided strong evidence for the general theory that remembering a prior event involves the reactivation of the neurons that were active when the event was originally encoded—that is, when the memory was formed.
In this study, Wiltgen first let animals create a memory by training the animals to be fearful of a specific environment. Importantly, Wiltgen trained the animals at a time when the GFP protein was being expressed in active neurons. In the time before training, the animals were on a supplement that prevented expression of the GFP. After training, Wiltgen then added the food supplement again, so that GFP would not continue to be expressed in neurons that became active later. Only neurons that were active when the animals were learning were labelled with the GFP, which Wiltgen was able to measure weeks later.
In this way, Wiltgen’s team could verify that the animals had actually created a memory by examining whether they exhibit more fearful behaviors (freezing in place, for example) in the training environment than another, similar environment. After a few weeks, Wiltgen re-tested their memory, identifying neurons that were active during the memory test. The long-lasting GFP, which should only be found in neurons that were active during the learning phase, would glow green. The other neuron label would glow red, and should only be found in neurons that were active during the testing phase.
The animals remembered the prior experience when put into the learning environment, as evidenced by their freezing in place. When Wiltgen’s team examined which neurons were active during learning and later during remembering, they found significant overlap between the cells labelled in green and in red.
In general, about 10 percent of the neurons that were active during learning will be active anytime you measure (Wiltgen considers this “background brain activity” that doesn’t necessarily support memory.) But when he examined the activity of neurons in animals whose behavior suggested they were actually remembering the prior experience, around 40 percent of the neurons that were active during learning were reactivated during the test phase.
This data is simply correlational, but it tells us that when we remember something, a lot of the neurons that were active when we learned the event are active again.
Taking this correlational evidence a step further, Wiltgen has used optogenetics to show that preventing the reactivation of neurons can suppress the retrieval of a memory. After training animals and labelling the neurons that were active when the animals were creating a memory, Wiltgen’s team tested their memory while using optogenetic techniques to inactivate neurons. Tests showed that selectively preventing the neurons that were active during learning from becoming active during the test prevented the animals from remembering the prior experience (as suggested by their behavior).
Wiltgen and his collaborators have shown that the same neurons that encode a memory are reactivated when the animal is retrieving the memory. His team has also effectively blocked memory retrieval, showing that an animal cannot remember a prior experience if the neurons that were active during learning are prevented from reactivating. Thus, we can be confident that remembering an event involves the reactivation of networks of neurons that were active during learning of that event.
In another line of work, Wiltgen has mapped out networks of neurons in the hippocampus, a brain structure that is critically important for memory, in order to understand their separate functions.
Research in humans suggests the hippocampus is necessary to recall past experiences in detail, and there are a variety of theories describing how the circuits of neurons in the hippocampus support different memory functions. Wiltgen uses neuron-labelling techniques to examine the networks of connected neurons in this important structure, in order to gain insight into the neural circuitry that supports human memory.
Wiltgen and his colleagues have shown that remembering a context, such as a spatial location, involves one specific circuit of neurons in the hippocampus. This research provides evidence that the neural circuits involved in memory for a spatial context are separate from the circuits that supports memory for individual objects found in that spatial context. This work aligns well with a theory of human memory proposed by other researchers at UC Davis, which proposes that the hippocampus is involved in binding memories for objects with memories for spatial contexts by coordinating the activity of two separate neural circuits. Thus, Wiltgen has used studies of animals to provide evidence in support of a prevalent theory of human memory functioning.
When asked about what questions he intends to pursue in the future, Wiltgen gets even more animated than usual. There are a vast number of unanswered questions in his field, he explains, particularly around the issue of how memories are maintained between learning and remembering.
“Knowing these rules will facilitate development of treatments for neuropsychiatric conditions and degenerative diseases.”
One prevalent theory, with many variants, proposes that networks of neurons “replay” the events during certain phases of sleep. Rather than suppress or manipulate memory retrieval, Wiltgen is looking toward manipulating memory storage or consolidation, by manipulating this “replay” activity in neurons. If storing and maintaining a memory requires that the neural activity from learning be replayed while the animal is sleeping, Wiltgen wonders if preventing those neurons from reactivating during sleep will cause the memory to be forgotten.
Some theories specify that the replay activity critical for maintaining a memory happens in a specific pattern of brain activity, known as the Sharp Wave Ripple (SWR). This pattern occurs in a specific stage of sleep, known as Stage 2. Thus, it would be even better to prevent neurons from reactivating during this specific stage of sleep, rather than throughout the sleeping period.
With respect to the future of his field more generally, and the promises this work holds for humans, Wiltgen is confident the science will continue to advance at an impressive rate. He describes current techniques for manipulating memory traces as “open-loop” systems, in which researchers decide when to activate or inactivate neurons. But he predicts that, in the near future, “closed-loop” systems will be developed, in which researchers can use brain activity to determine when to manipulate neural activity—rather than deciding for themselves. For example, researchers may soon be able to use brain activity to predict an upcoming SWR, and use this predictive signal to prevent neurons from becoming active at that very moment.
When asked how his work with animals applies to human memory, Wiltgen expresses a desire to understand the rules that govern how the brain remembers. “Knowing these rules will facilitate development of treatments for neuropsychiatric conditions and degenerative diseases.”
Learn more about Dr. Brian Wiltgen and his lab at the UC Davis Center for Neuroscience.