In the spring of 2019, neuroscientist Heather Cameron set up a simple experiment. She and her colleagues put an adult rat in the middle of a plastic box with a water bottle at one end. They waited until the rat started drinking and then made a startling noise to see how the animal would respond. The team did this repeatedly with regular rats and with animals that were genetically altered so that they couldn’t make new neurons in their hippocampuses, a brain region involved in learning and memory. When the animals heard the noise, those that could make new hippocampal neurons immediately stopped slurping water and looked around, but the animals lacking hippocampal neurogenesis kept drinking. When the team ran the experiment without the water bottle, both sets of rats looked around right away to figure out where the sound was coming from. Rats that couldn’t make new neurons seemed to have trouble shifting their attention from one task to another, the researchers concluded.
Aging humans, in whom neurogenesis is thought to decline, often have trouble remembering details that distinguish similar experiences.
“It’s a very surprising result,” says Cameron, who works at the National Institute of Mental Health (NIMH) in Bethesda, Maryland. Researchers studying neurogenesis in the adult hippocampus typically conduct experiments in which animals have had extensive training in a task, such as in a water maze, or have experienced repetitive foot shocks, she explains. In her experiments, the rats were just drinking water. “It seemed like there would be no reason that the hippocampus should have any role,” she says. Yet in animals engineered to lack hippocampal neurogenesis, “the effects are pretty big.”
The study joins a growing body of work that challenges the decades-old notion that the primary role of new neurons within the adult hippocampus is in learning and memory. More recently, experiments have tied neurogenesis to forgetting, one possible way to ensure the brain doesn’t become overloaded with information it doesn’t need, and to anxiety, depression, stress, and, as Cameron’s work suggests, attention. Now, neuro-scientists are rethinking the role that new neurons, and the hippocampus as a whole, play in the brain.
Related health conditions
Most of the research into neurogenesis involves boosting or inhibiting animals’ generation of new neurons, then training animals on a complex memory task such as finding a treat in a maze, and later retesting the animals. Decreasing neurogenesis tends to hamper the animals’ ability to remember.
Alzheimer’s disease, Parkinson’s disease
Training mice or rats on a memory task before manipulating neurogenesis has also been found to affect the strength of the trained memory. Boosting neurogenesis reduced the memory’s strength, perhaps an extreme form of forgetting that at normal levels avoids the remembering of unnecessary details.
Alzheimer’s disease and other forms of dementia
Research has linked decreased neurogenesis with more anxious and depressive behaviors in mice. Stress can reduce neurogenesis, ultimately leading mice to be more anxious in future stressful situations.
PTSD, anxiety, depression
Research has linked decreased neurogenesis with trouble switching focus.
The memory link
The first hint that adult animal brains may make new neurons appeared in the early 1960s, when MIT neurobiologist Joseph Altman used radioactive labeling to track the proliferation of nerve cells in adult rats brains. Other data published in the 1970s and 1980s supported the conclusion, and in the 1990s, Fred “Rusty” Gage and his colleagues at the Salk Institute in La Jolla, California, used an artificial nucleotide called bromodeoxyuridine (BrdU) to tag new neurons born in the brains of adult rats and humans. Around the same time, Elizabeth Gould of Princeton University and her collaborators showed that adult marmoset monkeys made new neurons in their hippocampuses, specifically in an area called the dentate gyrus. While some researchers questioned the strength of the evidence supporting the existence of adult neurogenesis, most of the field began to shift from studying whether adult animal brains make new neurons to what role those cells might play.
See “Brain Gain”
In 2011, René Hen at Columbia University and colleagues created a line of transgenic mice in which neurons generated by neuro-genesis survived longer than in wildtype mice. This boosted the overall numbers of new neurons in the animals’ brains. The team then tested the modified mice’s cognitive abilities. Boosting
numbers of newly born neurons didn’t improve the mice’s performances in water mazes or avoidance tasks compared with control mice. But it did seem to help them distinguish between two events that were extremely similar. Mice with more new neurons didn’t freeze as long as normal mice when put into a box that was similar to but not exactly the same as one in which they’d experienced a foot shock in earlier training runs.
These results dovetailed with others coming out at the time, particularly those showing that aging humans, in whom neurogenesis is thought to decline, often have trouble remembering details that distinguish similar experiences, what researchers call pattern separation. “The line of thinking is that the memories that are most likely to be impacted by neurogenesis are memories that are really similar to each other,” says Sarah Parylak, a staff scientist in Gage’s lab at the Salk Institute.
As insights into pattern separation emerged, scientists were beginning to track the integration of new rodent neurons into existing neural networks. This research showed that new neurons born in the dentate gyrus had to compete with mature neurons for connections to neurons in the entorhinal cortex (EC), a region of the brain with widespread neural networks that play roles in memory, navigation, and the perception of time. (See “Memories of Time” on page 32.) Based on detailed anatomical images, new dentate gyrus neurons in rodents appeared to tap into preexisting synapses between dentate gyrus neurons and EC neurons before creating their own links to EC neurons.
To continue exploring the relationship between old and new neurons, a group led by the Harvard Stem Cell Institute’s Amar Sahay, who had worked with Hen on the team’s 2011 study, wiped out synapses in the dentate gyruses of mice. The researchers overexpressed the cell death–inducing protein Krüppel-like factor 9 in young adult, middle-aged, and old mice to destroy neuronal dendritic spines, tiny protrusions that link up to protrusions of other neurons, in the brain region. Those lost connections led to increased integration of newly made neurons, especially in the two older groups, which outperformed age-matched, untreated mice in pattern-separation tasks. Adult-born dentate gyrus neurons decrease the likelihood of reactivation of those old neurons, Sahay and colleagues concluded, preventing the memories from being confused.
Parylak compares this situation to going to the same restaurant after it has changed ownership. In her neighborhood in San Diego, there’s one location where she’s dined a few times when the restaurant was serving different cuisine. It’s the same location, and the building retains many of the same features, “so the experiences would be easy to mix up,” she says, but she can tell them apart, possibly because of neurogenesis’s role in pattern separation. This might even hold true for going to the same restaurant on different occasions, even if it served the same food.
That’s still speculative at this point. Researchers haven’t been able to watch neurogenesis in action in a living human brain, and it’s not at all clear if the same thing is going on there as in the mouse brains they have observed. While many scientists now agree that neurogenesis does occur in adult human brains, there is little consensus about what it actually does. In addition to the work supporting a role for new neurons in pattern separation, researchers have accumulated evidence that it may be more important for forgetting than it is for remembering.
How Adult-Born Neurons Integrate into the Brain
In recent years, images and videos taken with state-of-the-art microscopy techniques have shown that new neurons in the dentate gyrus of the hippocampus go through a series of changes as they link up to existing networks in the brain.
© lisa clark
A neural stem cell divides to generate a new neuron (green).
© lisa clark
As the new neuron grows, it rotates from a horizontal to a vertical position and connects to an interneuron (yellow) in a space called the hilus that sits within the curve of the dentate gyrus. The young neuron also starts making connections with well-established dentate gyrus neurons (blue) as well as neurons in the hippocampus (red).
© lisa clark
Once connections are formed, mature neurons send signals into the new neuron, and the cell starts firing off more of its own signals. At around four weeks of age, the adult-born neuron gets hyperexcited, sending electrical signals much more often than its well-established neuronal neighbors do.
© lisa clark
As the new neuron connects with still more neurons, interneurons in the hilus start to send it signals to tamp down its activity.
The importance of forgetting
It seems counterintuitive for neurogenesis to play a role in both remembering and forgetting, but work by Paul Frankland of the Hospital for Sick Children Research Institute in Toronto suggests it is possible. In 2014, his team showed that when mice made more new neurons than normal, they were more forgetful. He and his colleagues had mice run on wheels to boost levels of neurogenesis, then trained the animals on a learning task. As expected, they did better than control mice who hadn’t exercised. (See “How Exercise Reprograms the Brain,” The Scientist, October 2018.) In other animals, the researchers boosted neurogenesis after the mice learned information thought to be stored, at least in the short term, in the hippocampus. “When we did that, what we found was quite surprising,” Frankland says. “We found a big reduction in memory strength.”
His team was puzzled by the result. Adding to the confusion, the researchers had observed a larger effect in memory impairment with mice that learned, then exercised, than they had seen in memory improvement when the mice ran first and then learned. As he dug into the literature, Frankland realized the effect was what other neuroscientists had called forgetting. He found many theoretical papers based on computational modeling that argued that as new neurons integrate into a circuit, the patterns of connections in the circuit change, and if information is stored in those patterns of connections, that information may be lost. (See “Memory Munchers” on page 21.)
The notion surprised other neuroscientists, mainly because up to that point they’d had two assumptions related to neurogenesis and forgetting. The first was that generating new neurons in a normal animal should be good for memory. The second was that forgetting was bad. The first assumption is still true, Frankland says, but the second is not. “Many people think of forgetting as some sort of failure in our memory systems,” he explains. Yet in healthy brains there’s tons of forgetting happening all of the time. “And, in fact, it’s important for memory function,” Frankland says. “It would actually be disadvantageous to remember everything we do.”
Experiments have tied neurogenesis to forgetting, anxiety, depression, stress, and attention.
Parylak says this idea of forgetting “certainly has provoked a lot of discussion.” It’s unclear, for example, whether the mice in Frankland’s experiments are forgetting, or if they are identifying a repeat event as something novel. This is the point, she explains, where doing neurogenesis research in humans would be beneficial. “You could ask a person if they’d actually forgotten or if they are making some kind of extreme discrimination.”
Despite the questions regarding the results, Frankland and his colleagues continued their work, testing mice’s forgetfulness with all types of memories, and more recently they asked whether the forgetting effect jeopardized old and new memories alike. In experiments, his team gave mice a foot shock, then boosted hippocampal neurogenesis (with exercise or a genetic tweak to neural progenitor cells), and put the mice in the same container they’d been shocked in. With another group of mice, the researchers waited nearly a month after the foot shock before boosting neurogenesis and putting the mice back in the container. Boosting the number of new neurons, the team found, only weakened the newly made memory, but not one that had been around for a while. “This makes a lot of sense,” Frankland says. “As our memories of everyday events gradually get consolidated, they become less and less dependent on the hippocampus,” and more dependent on another brain region: the cortex. This suggests that remote memories are less sensitive to changes in hippocampal neurogenesis levels.
The hippocampus tracks what’s happened to you, Frankland says. “Much of that’s forgotten because much of it is inconsequential. But every now and then something interesting seems to happen,” and it’s these eventful memories that seem to get “backed up” in other areas of the brain.
How adult-born neurons function in a circuit
Researchers think neurogenesis helps the brain distinguish between two very similar objects or events, a phenomenon called pattern separation. According to one hypothesis, new neurons’ excitability in response to novel objects diminishes the response of established neurons in the dentate gyrus to incoming stimuli, helping to create a separate circuit for the new, but similar, memory.
© lisa clark
At NIMH, one of Cameron’s first studies looking at the effects of neurogenesis tested the relationship between new neuronal growth and stress. She uncovered the connection studying mice that couldn’t make new neurons and recording how they behaved in an open environment with food at the center. Just like mice that could still make new neurons, the neuro-genesis-deficient mice were hesitant to go get the food in the open space, but eventually they did. However, when the animals that couldn’t make new neurons were stressed before being put into the open space, they were extremely cautious and anxious, whereas normal mice didn’t behave any differently when stressed.
Cameron realized that the generation of new neurons also plays a role in the brain separate from the learning and memory functions for which there was growing evidence. In her experiments, “we were looking for memory effects and looked for quite a while without finding anything and then stumbled onto this stress effect,” she says.
The cells in the hippocampus are densely packed with receptors for stress hormones. One type of hormone in particular, glucocorticoids, is thought to inhibit neurogenesis, and decreased neurogenesis has been associated with depression and anxiety behaviors in rodents. But there wasn’t a direct link between the experience of stress and the development of these behaviors. So Cameron and her colleagues set up an experiment to test the connection.
When the team blocked neurogenesis in adult mice and then restrained the animals to moderately stress them, their elevated glucocorticoid levels were slow to recover compared with mice that had normal neurogenesis. The stressed mice that could not generate new neurons also acted oddly in behavioral tests: they avoided food when put in a new environment, became immobile and increasingly distressed when forced to swim, and drank less sugary water than normal mice when it was offered to them, suggesting they don’t work as hard as normal mice to experience pleasure. Impaired adult neurogenesis, the experiments showed, played a direct role in developing symptoms of depression, Cameron says.
The notion that neurogenesis and stress might be tied directly to our mental states led Cameron to look back into the literature, where she found many suggestions that the hippocampus plays a role in emotion, in addition to learning and memory. Even Altman, who unexpectedly identified neurogenesis in adult rodents in the 1960s, and colleagues suggested as much in the 1970s. Yet the argument has only appeared sporadically in the literature since then. “Stress is complicated,” Cameron says; it’s hard to know exactly how stressful experiences affect neurogenesis or how the generation of new neurons will influence an animal’s response to stress. Some types of stress can decrease neurogenesis while others, such as certain forms of intermittent stress, can increase new neuronal growth. Last year, Cameron and colleagues found that generating new neurons helps rats used to model post-traumatic stress disorder recover from acute and prolonged periods of stress.
Neurogenesis appears to play a role in both remembering and forgetting.
Her work has also linked neurogenesis to other characteristics of rodent behavior, including attention and sociability. In 2016, with Gould at Princeton and a few other collaborators, she published work suggesting that new neurons are indeed tied to social behavior. The team created a hierarchy among rats, and then deconstructed those social ranks by removing the dominant male. When the researchers sacrificed the animals and counted new neurons in their brains, the rats from deconstructed hierarchies had fewer new neurons than those from control cages with stable ranks. Rats with uncertain hierarchies and fewer new neurons didn’t show any signs of anxiety or reduced cognition, but they weren’t as inclined as control animals to spend time with new rats put into their quarters, preferring to stick with the animals they knew. When given a drug—oxytocin—to boost neurogenesis, they once again began exploring and spending time with new rats that entered their cages.
The study from Cameron’s lab on rats’ ability to shift their attention grew out of the researchers’ work on stress, in which they observed that rodents sometimes couldn’t switch from one task to the next. Turning again to the literature, Cameron found a study from 1969 that seemed to suggest that neurogenesis might affect this task-switching behavior. Her team set up the water bottle experiments to see how well rats shifted attention. Inhibiting neurogenesis in the adult mice led to a 50 percent decrease in their ability to switch their focus from drinking to searching for the source of the sound.
“This paper is very interesting,” says J. Tiago Gonçalves, a neuroscientist at Albert Einstein College of Medicine in New York who studies neurogenesis but was not involved in the study. It could explain the findings seen in some behavioral tasks and the incongruences between findings from different behavioral tasks, he writes in an email to The Scientist. Of course, follow-up work is needed, he adds.
Cameron argues that shifting attention may be yet another behavior in which the hippocampus plays an essential role but that researchers have been overlooking. And there may be an unexplored link between making new neurons and autism or other attention disorders, she says. Children with autism often have trouble shifting their attention from one image to the next in behavioral tests unless the original image is removed.
It’s becoming clear, Cameron continues, that neurogenesis has many functions in the adult brain, some that are very distinct from learning and memory. In tasks requiring attention, though, there is a tie to memory, she notes. “If you’re not paying attention to things, you will not remember them.”
Do new neurons appear anywhere else in the brain?
Many, though not all, neuroscientists agree that there’s ongoing neurogenesis in the hippocampus of most mammals, including humans. In rodents and many other animals, neurogenesis has also been observed in the olfactory bulbs. Whether newly generated neurons show up anywhere else in the brain is more controversial.
There had been hints of new neurons showing up in the striatum of primates in the early 2000s. In 2005, Heather Cameron of the National Institute of Mental Health and colleagues corroborated those findings, showing evidence of newly made neurons in the rat neocortex, a region of the brain involved in spatial reasoning, language, movement, and cognition, and in the striatum, a region of the brain involved in planning movements and reacting to rewards, as well as self-control and flexible thinking (J Cell Biol, 168:415–27). Nearly a decade later, using nuclear-bomb-test-derived carbon-14 isotopes to identify when nerve cells were born, Jonas Frisén of the Karolinska Institute in Stockholm and colleagues examined the brains of postmortem adult humans and confirmed that new neurons existed in the striatum (Cell, 156:1072–83, 2014).
“Those results are great,” Cameron says. They support her idea that there are different types of neurons being born in the brain throughout life. “The problem is they’re very small cells, they’re very scattered, and there’re very few of them. So they’re very tough to see and very tough to study.”