Previously, I wrote about new SFN data on the role for newborn neurons in regulating emotion. The second half of the SFN meeting rounded out the story because the bulk of the functional presentations focussed on the role of new neurons in that other, classic function of the hippocampus: memory. Spanning synaptic plasticity, circuit function, and then linking it all to behavior, we have quite a complete story here.

SYNAPTIC PLASTICITY IN YOUNG NEURONS

Every time I get worked up about all various neurogenesis findings I think about one acronym that returns me to a state of inner peace: ACSF-LTP. Yes, I plagiarized that last line from my previous post. We all know about LTP right? The ability of synapses to strengthen their connections in response to activity? It has been used for decades as a physiological model of memory formation. It’s pretty well accepted that newborn neuron ACSF-LTP is a unique form of LTP – one that is insensitive to GABAergic inhibition (hence “Artificial Cerebro Spinal Fluid” LTP, in contrast to LTP that also requires inhibition of GABA neurotransmission), one that requires a the NR2B subunit of the NMDA receptor, and one that is induced more easily than that of mature neurons. ACSF-LTP has quite a history:

1. It was first shown by Sabrina Wang in Martin Wojtowicz’s lab back in 2000. The first demonstration of a unique functional role for new neurons. That work was done in single neurons from rat, by patch clamping.
2. I followed up on that work in the Wojtowicz lab and found that this LTP could be observed in field recordings, i.e. this LTP stood out amongst the background of activity from all granule neurons when you stimulated their synaptic inputs. In fact, it was the only type of LTP observed in the absence of GABA blockers (making it pretty easy to identify and measure, even by novice electrophysiologists such as myself).
3. Ge et al (2007) confirmed these findings in single neurons in mice and characterized the “critical period”, showing that only immature neurons displayed this plasticity, and not older adult-born neurons.
4. Wang et. al (2008) did a nice trick and increased neurogenesis with fluoxetine and in turn increased ACSF-LTP. Irradiation abolished both.
5. Garthe et al (2009) used a chemical method for stopping neurogenesis. ACSF-LTP was gone.
6. Sahay et al (2011) also increased neurogenesis and therefore ACSF-LTP, via a novel transgenic method.

AND NOW, I saw a poster by Kheirback et. al, who now used a completely novel method to confirm the existence of this special form of plasticity in new neurons. They used Nestin CreER and a floxed NR2B to delete the NR2B subunit specifically from young neurons. This didn’t kill neurons, it didn’t affect plasticity in mature neurons, but it wiped out ACSF-LTP. This is not entirely surprising because the NR2B subunit is known to endow new neuron with their enhanced plasticity. It just was a novel, elegant and specific way to demonstrate it.

Whereas the behavioral data on new neurons is less advanced and more variable, the fact that ACSF-LTP has been demonstrated in different species (mouse and rat), is absent after using different methods to reduce neurogenesis (irradiation, chemical, transgenic), is enhanced after using different methods for increasing neurogenesis (fluoxetine, transgenic) arguably makes it the strongest support for a real, significant function for hippocampal neurogenesis.

Going back to these NR2B-deficient new neurons, these mice also had behavioral effects: they had impaired ability to behaviorally distinguish contexts during fear conditioning, reduced object exploration, and no preference for a novel object over a familiar object. There was also a tendency, albeit less robust, for these mice to be more innately anxious and fearful.

OPTOGENETICS

ACSF-LTP is so well documented because the anatomy and physiology of the inputs to granule neurons is not super complex. In contrast, if there’s one brain region that could really benefit from optogenetics, it’s the dentate gyrus output onto CA3. The connectivity is such that it is practically impossible to stimulate only new neurons with conventional techniques and record from their postsynaptic targets. At least the ones that are further away, like pyramidal neurons. This can be overcome by shining light on the dentate gyrus and ensuring that only new neurons are activated, through selective expression of channelrhodopsin or its variants. This is exactly what Gu et al did. They characterized, for the first time, mossy fiber LTP in new neurons and found that 4-week-old neurons have greater LTP than either 3 or 8-week-old cells. Using archaerhodopsin to inhibit new neurons they also found that these 4-week-old cells are especially important for memory retrieval in fear conditioning and water maze paradigms. The deficits were not huge, but they were consistent. Considering that they used viral methods to express archaerhodopsin, which probably only infects a small proportion of the total new neuron population, the effect these new neurons are having on behavior is quite remarkable.

SO WE HAVE SYNAPSES AND BEHAVIOR, BUT WHAT’S IN THE MIDDLE?

The most commonly-proposed, but least understood, cognitive function for new neurons is pattern separation. Originally a computational term, it refers to the process by which neurons take similar inputs (spatiotemporal patterns of incoming synaptic activity) and maximize their differences. When brains do this well we are able to distinguish very similar experiences in our memory. When it fails you spend 30 minutes looking for your car before you remember that you parked in a different lot. The discrimination of safe from dangerous contexts during fear conditioning is used more and more as a behavioral test of pattern separation. While it may depend on pattern separation by new neurons, this has never been shown. To get at this, Niibori et al. used the catFISH method to see what is happening in CA3 neuronal ensembles when mice don’t have neurogenesis. catFISH uses the spatiotemporal patterns of activity-dependent gene expression to identify neuronal populations that were activated by different experiences. It’s been shown that when rodents are put into 2 different contexts, distinct ensembles of CA3 neurons are activated by the 2 experiences. The memory traces are kept distinct in order to not confuse the experiences. If new neurons are indeed performing a pattern separation function then you’d predict that, in their absence, the same population of CA3 neurons is activated by the 2 experiences. And this is exactly what Niibori et al found. Furthermore, confirming that new neurons are not just performing this separation function, but are actually required to behaviorally distinguish the 2 environments, mice lacking neurogenesis showed similar levels of fear behavior in both a context that was paired with shock as well as a context that had never been paired with shock. And so the hypothesized role for new neurons in pattern separation just got a whole lot stronger.

WHERE ARE WE AT?

There’s only so much functional work going on out there. Some behavior, less electrophysiology, a new method for measuring circuit properties. Your study of a phenomenon feels complete when you’ve seen presentations that cover function at these different levels of analysis. When they replicate and then build upon previous findings it’s even more of a confirmation that progress is being made. I have to say I think it was a pretty good meeting for adult neurogenesis.

“Random” roundup because any posts linking to articles or ideas I’ve recently found noteworthy will never occur on a regular basis (as others manage to do – I applaud you) but only when enough interesting material has accrued and I have a spare moment. The links will, however, not be random. For example, you can expect many links to point to articles on adult neurogenesis or hippocampal function but will likely find few links directing you to photos of puppy dogs.

Dopaminergic Modulation of Cortical Inputs during Maturation of Adult-Born Dentate Granule Cells. A pretty thorough examination of dopaminergic modulation of synaptic transmission and synaptic plasticity in the dentate gyrus. Dopamine reduced synaptic transmission in both immature and mature granule neurons, but through different receptor subtypes. Additionally, dopamine reduced long-term plasticity in immature neurons but not mature neurons. Given the suggestion that dopamine could gate the entry of information into long-term memory, these findings suggest young and old neurons could have quite different behavioral functions.

Mu Y, Zhao C, & Gage FH (2011). Dopaminergic Modulation of Cortical Inputs during Maturation of Adult-Born Dentate Granule Cells. The Journal of neuroscience : the official journal of the Society for Neuroscience, 31 (11), 4113-23 PMID: 21411652

————————-

Lidocaine attenuates anisomycin-induced amnesia and release of norepinephrine in the amygdala. Memory consolidation is the phenomenon by which memories are encoded through enduring structural changes in the brain and is often demonstrated by showing that memory loss occurs when you inhibit protein synthesis around the time of learning. This paper shows that one of the most commonly-used protein synthesis inhibitors, anisomycin, leads to increased norepinephrine release in the amygdala which could, by itself, impair memory.  The interesting final experiment showed that the effects of anisomycin on memory and norepiniphrine were reduced when the amygdala was totally shut down with lidocaine.

Sadowski RN, Canal CE, & Gold PE (2011). Lidocaine attenuates anisomycin-induced amnesia and release of norepinephrine in the amygdala. Neurobiology of learning and memory PMID: 21453778

————————-

Evidence for the Re-Enactment of a Recently Learned Behavior during Sleepwalking. I’ve written a number of times about how neuronal firing patterns observed during waking experience are replayed during sleep, and could therefore reflect consolidation of memory and even dream content. Of course no one knows what rats are experiencing during sleep or whether they dream like us. To get around this problem, these authors trained sleepwalkers on a motor task with very defined arm movements and then examined sleepwalking behavior on the following night. Indeed, a video shows one subject who wakes up the following night and, for a few seconds, seems to be performing the same stereotyped task movements. Only one subject but tantalizing evidence and a cool experimental strategy nonetheless.

Oudiette D, Constantinescu I, Leclair-Visonneau L, Vidailhet M, Schwartz S, & Arnulf I (2011). Evidence for the Re-Enactment of a Recently Learned Behavior during Sleepwalking. PloS one, 6 (3) PMID: 21445313

————————-

Increasing adult hippocampal neurogenesis is sufficient to improve pattern separation. One of the biggest questions in the neurogenesis field is whether adult-born neurons are important for behavior. Usually this is tested by examining behavior in animals that lack adult neurogenesis but many studies have correlated increased neurogenesis in enriched or athletic animals with “improved” behavior (smarter, less depressed etc). Of course, the major confound is that enrichment and exercise do many other things besides increasing neurogenesis. To get around this Sahay et al. made a mouse in which neurogenesis could be specifically increased in adulthood. These mice were better at discriminating between related contexts and, after exercise, showed much greater exploratory activity in an open field.

Sahay A, Scobie KN, Hill AS, O’Carroll CM, Kheirbek MA, Burghardt NS, Fenton AA, Dranovsky A, & Hen R (2011). Increasing adult hippocampal neurogenesis is sufficient to improve pattern separation. Nature PMID: 21460835

————————-

Necessity of Hippocampal Neurogenesis for the Therapeutic Action of Antidepressants in Adult Nonhuman Primates. This study potentially bridges a big big gap by extending the role of adult neurogenesis in the antidepressant response from rodents all the way to monkeys. Chronic stress induced anhedonic and subordinate behaviors and these effects could be reversed with fluoxetine, but not in irradiated monkeys that had reduced neurogenesis. Could someone follow this up with a transgenic model?

Perera, T., Dwork, A., Keegan, K., Thirumangalakudi, L., Lipira, C., Joyce, N., Lange, C., Higley, J., Rosoklija, G., Hen, R., Sackeim, H., & Coplan, J. (2011). Necessity of Hippocampal Neurogenesis for the Therapeutic Action of Antidepressants in Adult Nonhuman Primates PLoS ONE, 6 (4) DOI: 10.1371/journal.pone.0017600

————————-

Systemic 5-bromo-2-deoxyuridine induces conditioned flavor aversion and c-Fos in the visceral neuraxis. OH NOOO! Rats don’t like BrdU! These authors show that pairing a BrdU injection with exposure to a sweet palatable drink causes rats to avoid that drink in the future. It also leads to a mildly elevated stress response and elevated c-fos expression in areas of the brain that represent viscera, consistent with the possibility that BrdU could be exerting unpleasant effects in the gut, where there is a lot of cell division. The authors conclude that the effects on behavior in subsequent days and weeks are probably minimal (phew!), but I’d certainly keep these data in mind when considering injecting BrdU around the time of behavioral testing.

Kimbrough A, Kwon B, Eckel LA, & Houpt TA (2011). Systemic 5-bromo-2-deoxyuridine induces conditioned flavor aversion and c-Fos in the visceral neuraxis. Learning & memory (Cold Spring Harbor, N.Y.), 18 (5), 292-5 PMID: 21498563

————————-

Compensatory network changes in the dentate gyrus restore long-term potentiation following ablation of neurogenesis in young-adult mice. In an interesting study of plasticity following neurogenesis reduction, these authors find that LTP was dramatically reduced after arresting neurogenesis, but only transiently. LTP recovered within weeks, possibly because of compensatory reductions in inhibition and enhanced survival of neurons born before neurogenesis ablation. Hat tip to Sil for this one.

Singer BH, Gamelli AE, Fuller CL, Temme SJ, Parent JM, & Murphy GG (2011). Compensatory network changes in the dentate gyrus restore long-term potentiation following ablation of neurogenesis in young-adult mice. Proceedings of the National Academy of Sciences of the United States of America, 108 (13), 5437-42 PMID: 21402918

————————-

That’s it.

It’s often thought that things change as you move into parts of the brain that represent more complex and abstract concepts. For example, in the hippocampus, many neurons receive the same inputs so it’s generally assumed that different neurons are equally capable of representing a given piece of information. While wiring between neurons must play a role in determining which neurons are activated, the diffuseness of the wiring means that related information need not be stored in spatially neighboring neurons as in the sensory regions of neocortex. Indeed, if you look at hippocampal neurons activated by a given experience they don’t appear to have any particular spatial arrangement but are randomly distributed, anatomically. Alternatively, it could be that certain hippocampal neurons are hard-wired to respond to specific stimuli, it’s just that we don’t understand the wiring.

I’ve mentioned before (here and here) how anatomical patterns of activity in the hippocampus are not always so random – in the dentate gyrus the same neurons are often repeatedly activated and by very different experiences. Furthermore, half of the dentate gyrus (the infrapyramidal blade) never seems to be noticeably active, period. But anatomical biases have been reported outside of the dentate gyrus too. Hampson and Deadwyler showed that spatial and nonspatial information is segregated in distinct septotemporal regions of CA1/CA3. Also, Nakamura et al. have suggested that CA1 neurons that represent a given spatial environment are more likely to be spatially clustered together.

While these studies suggest there may be a hard-wired anatomical pattern by which information is represented in regions such as the hippocampus, we really have have no idea how that pattern might be established. I was therefore intrigued to see a couple papers shed new light on this issue. One is a recent paper by Yassin et al. who used a Fos-GFP mouse to identify and record from neurons recently activated by behavioral experience. Fos is an immediate-early gene that is upregulated in neurons that are involved in learning and so, in this mouse, those neurons fluoresced green and could be examined electrophysiologically. They found that the Fos-GFP neurons fired at higher rates than neighboring neurons that were not expressing GFP and that they tended to be more connected to one another (and thus they were dubbed Facebook neurons), suggesting that there may be a subset of neurons that is preselected to be involved in representing experiences (perhaps not unlike the population of highly-active dentate gyrus neurons). There is a bit of a chicken and egg problem here, because we don’t know if the GFP+ neurons always fire at higher rates (and are hard-wired to be more involved in representing experience) or if they only fire at higher rates because they were recently activated (i.e. behavior-induced plasticity changed them). Intriguing nonetheless and a good approach for future studies I think.

The other study is pretty revolutionary I think and also has to do with predetermined, hard-wired patterns of neuronal activity. One of the exciting developments of the last 15 years has been the finding that patterns of neuronal activity are replayed during sleep. It is thought that this “replay” is the physiological correlate of memory consolidation, i.e. the rehearsal of recent experience and integration of that new information into the brain’s circuitry. Now, Dragoi and Tonegawa have found that the patterns of neuronal activity, seen as a mouse explores a novel environment, can also be seen during rest/sleep episodes before the mouse has ever been in that environment. Essentially, they discovered that the brain has created a representation (or at least a fraction) of an experience that has not even happened yet. They call the phenomenon “preplay”.

The preplay phenomenon does fit with previous data. The Mosers, in their News and Views piece on this study, note that “…place cells continue to fire in regular sequences when an animal’s position is fixed, for example, when a rat is running in a wheel. Moreover, rat pups exploring an open space for the first time show adult-like place cell sequences, which indicates that path sequences are hard-wired in the synaptic connection matrix by either genetic programs or early experience.” Also relevant is the finding from John Guzowski’s lab showing that very brief experiences (perhaps too brief to be even remembered) are capable of inducing transcription of the plasticity-related gene, Arc, in a full complement of CA3 neurons. In contrast, CA1 neurons were only fully activated after multiple experiences over multiple days, suggesting less of a role for hard-wiring and more of a role for plasticity and learning in shaping neural representations in this region.

Why preplay? One interesting hypothesis is that the hippocampus is needed to imagine the future (a reasonable role for a structure responsible for remembering the past). Could preplay be an attempt to predict future experience? Or might a shared pattern of activity simply be a way to bind together two events and create a coherent history? Don’t worry – I’m sure that, as we speak, there are rodents with implanted electrode arrays running around, working hard, to give us the answer.

Yassin L, Benedetti BL, Jouhanneau JS, Wen JA, Poulet JF, & Barth AL (2010). An embedded subnetwork of highly active neurons in the neocortex. Neuron, 68 (6), 1043-50 PMID: 21172607
Dragoi G, & Tonegawa S (2011). Preplay of future place cell sequences by hippocampal cellular assemblies. Nature, 469 (7330), 397-401 PMID: 21179088

I’m kidding, of course, about the predictions.  But I do think “pattern separation” is starting to interest a lot of people and that it would be worthwhile to step back and consider how we got here (and where we are).

In the past year or so there were two very interesting papers by the Bussey and Gage groups relating adult hippocampal neurogenesis to pattern separation.  One paper (Clelland et al., 2009) is notable because, in my opinion, it is one of the strongest papers to date linking adult hippocampal neurogenesis to a psychological process.  The paper rises above previous work because it uses two independent behavioral tasks to assess a single psychological construct.  Compare this to the usual single-task approach—for instance, the work demonstrating that blocking adult hippocampal neurogenesis impairs contextual fear conditioning (… to pick an example at random).  The fear conditioning finding, by itself, gives us relatively little insight into the underlying psychological mechanism: is it a failure to remember the shock context, a failure to associate a context with an aversive outcome, a failure to discriminate the shock contexts from other contexts?  One way to generate more specific mechanistic information is to use multiple tasks that vary in their behavioral requirements but share a common psychological requirement.  This is the approach used by Clelland et al., who demonstrated the blocking neurogenesis impairs the ability to discriminate nearby (but not far-apart) spatial locations. The paper used two tasks that differed substantially in their behavioral requirements: One task was an operant location discrimination in mice; the other was a spatial version of delayed-non-match-to-place in rats.  The fact that both tasks detected a similar effect of arresting neurogenesis constitutes pretty strong evidence that adult hippocampal neurogenesis plays a role in fine spatial discriminations.  (The other paper showed that spatial discrimination performance is enhanced by exercise, which potently stimulates neuronal proliferation in the hippocampus.)

So there you have it: the function of adult hippocampal neurogenesis is pattern separation.  But how much does this really tell us about the psychology? If you think about it, almost any psychological process constitutes a form of pattern separation.  Take a simple classical conditioning experiment: a tone is paired with food.  To learn the association, the subject must isolate the tone from among the blooming, buzzing confusion of other sensory stimuli present during the trial.  The pattern of sensory stimulation provided by the tone must be separated from the sound of the whirring exhaust fan, the odor of fresh bedding, the sight of the illuminated houselight, etc.  This is pattern separation.  Take another example: probabilistic category learning (e.g., the weather prediction task).  The subject learns to predict an outcome using an assortment of multidimensional stimuli, with each stimulus providing only partial information about the outcome.  With practice, the subject learns to separate the complex patterns of stimuli into different categories.  This is pattern separation.  Except the hippocampus (and, by inference, adult hippocampal neurogenesis) is not very important for either of these tasks.

The term “pattern separation,” as people are currently using it, comes from the computational literature, where it refers to an operation in which overlapping patterns of input stimulation are transformed into non- (or less-) overlapping patterns of output stimulation. Through pattern separation, stimuli that activate a common population of neurons are made to activate separate populations of neurons.  According to several models (e.g., O’Reilly & McClelland, 1994), this operation occurs in the dentate gyrus.

So how does this computational pattern separation map onto behavior?  Ray Kesner’s group has been investigating this question for a while, and they have shown that lesions of the dentate gyrus impair fine spatial discriminations but do not impair other types of discrimination, such as temporal order or reward magnitude discriminations.  Work by the Leutgebs and Mosers confirms that the dentate granule cells are very sensitive to small spatial changes (read more about this here). But dentate gyrus pattern separation is not strictly spatial.  A recent human imaging study (Bakker et al., 2008) showed that the dentate gyrus (and possibly CA3) are activated when people see objects that are very similar to –but not the same as– objects they’ve seen before.

More generally, we know that the dentate gyrus receives input from most or all sensory modalities, suggesting that the dentate gyrus does not operate solely on spatial information.  So, while it seems clear that adult hippocampal neurogenesis has an important role in spatial pattern separation, some key questions remain.  Is dentate gyrus involved in pattern separation across multiple perceptual modalities?  If so, what defines dentate gyrus pattern separation relative to the other types of pattern separation in the brain? And does adult hippocampal neurogenesis contribute to all types of dentate pattern separation or to only a subset?

For answers, stay tuned to the next 369,999,999 papers.

1. A group of scientists reduce neurogenesis and report a memory deficit.
2. A second group repeats the experiment, with only a few minor differences in protocol, and fails to find a memory deficit.
3. A third group, using the same species as the first group but a protocol more similar to the second group, replicates the original finding but only when the experiment is performed on Wednesdays.
4. Faith is restored.
5. Five groups report no such neurogenesis-dependent memory deficit.
6. It is reported that developmental exposure to strontium reduces adult neurogenesis by 40% AND produces the much sought after memory deficit. In a technical tour de force follow-up experiment, artisanal cheeses restore neurogenesis and reverse the memory deficits. Causation is established.
7. BDNF.
8. Everyone proclaims the role of neurogenesis in memory and is totally confused at the same time.
9. Someone systematically examines all of the variables in the memory test to determine whether or not the whole thing is a hoax and they should just change careers**.
10. We have never gotten this far.

Even at level 8, the neurogenesis-fear conditioning story was one of the more convincing arguments of new neuron functionality. With this study by Drew et al. we may soon be jumping for joy as we appear to be graduating to level 9.

The contribution of adult neurogenesis to contextual fear conditioning was greatest when mice were only given a brief training experience – mice lacking adult neurogenesis showed reduced fear of a context where they previously received a single footshock during a brief (3 min) exploration session. With longer exposures to the context, or additional footshocks, neurogenesis-deficient mice showed normal memory. This finding could be explained by the fact that young neurons have a lower threshold for synaptic plasticity, allowing them to encode fleeting experiences that would be forgotten if left to mature neurons.

So, brief training protocols may now likely be my first choice, at least when using mice. In fact, the only times I have observed contextual fear memory deficits in mice has been after brief training protocols almost identical to those used by Drew et al. So we just might have taken a big step forward. If not, check back in 5 years for my revised “How progress is made” list.

*or any other field for that matter

**this is not entirely a joke because, in this case, it both 1) appears to not be a hoax, and 2) marks the launch of the next phase of Michael Drew’s career (congrats)

Reference
Drew MR, Denny CA, & Hen R (2010). Arrest of adult hippocampal neurogenesis in mice impairs single- but not multiple-trial contextual fear conditioning. Behavioral neuroscience, 124 (4), 446-54 PMID: 20695644

The main experiment is illustrated on the left. Groups of rats were either exposed to 4 different contexts (A,B,C,D) or a single context (D) over the course of several months (TRAINING EXPOSURE). Then rats were either exposed to all 4 contexts or the single context over the course of half an hour (TEST) and expression of the immediate-early gene, Arc, was used to identify “activated” neurons. The idea is that, if subsets of granule neurons encode memories at different times in an animal’s life, then recalling those distant memories at test may re-activate those granule neurons. Alme et al. provide two hypotheses that their experiment will test and I illustrate them below with black-eyed peas.

1) The selective tuning hypothesis.

According to the selective tuning hypothesis, different populations of newborn granule neurons (i.e. those in their critical period) encode the contexts explored at each point in an animal’s life. As these cells mature they become less active and no longer encode new information, causing subsequent experiences to be encoded in newly-added populations of granule neurons (top half of above figure; red cells for A, blue cells for B, green cells for C*). However, even when old and unexcitable, granule neurons that were originally involved in encoding an experience (back when they were immature), are reactivated during recall of that experience. In other words, the neurons are selectively tuned to memories for events that happened at a specific point in time. Thus, during the test phase, when group #1 rats are exposed to contexts A, B, C, and D, a relatively large proportion of granule neurons will be activated.

In contrast, group #3 (bottom half) will also experience all 4 contexts during the test but would be expected to have fewer activated neurons because the contexts are novel and will therefore all be encoded, for the first time, in the same cohort of granule neurons (those that are in their critical period at the time of testing – shown as orange cells).

2) The retirement hypothesis.

Under the retirement hypothesis, again, only immature neurons are involved in memory encoding. However, unlike the selective tuning hypothesis, as neurons mature they become unexcitable to the point that they are never reactivated, even when recalling the memories that were formed during their critical period (top half of figure). So, during the testing phase when memories for A, B, C, and D are recalled, it is only the currently-immature population of neurons that is activated (the orange cells) – a much smaller population of granule neurons is activated compared to the selective tuning hypothesis. Additionally, one would expect that testing in A, B, C, D would activate the same number of neurons regardless of whether or not a rat had previously experienced the contexts (compare top and bottom halves of figure) because, at the time of test, neurons that may have previously been involved in encoding these contexts are no longer functional and it is only the currently-immature population of neurons that is activated.

What did they find?

The data generally aligned with the predictions of the retirement hypothesis: the same number of cells were activated regardless of whether rats experienced ABCD for the first time (group #3) or were re-exploring familiar contexts (group #1). So, it would appear that recalling distant memories does not reactivate additional subsets of granule neurons that would have been involved in the original encoding of those experiences. So, the selective tuning hypothesis is out.

Exploring four contexts at test activated twice as many neurons as exploration of only a single context. So, different experiences are capable of activating different populations of granule neurons, but to a much smaller extent than in other regions of the hippocampus (CA1, CA3). Alme et al. do a bunch of math to show that the probability that a granule neuron is activated in multiple contexts is about 40 times more than would be expected by chance, if random granule neurons were activated during different experiences.

What does this mean for my dentate gyrus?

Does this mean that these silent granule neurons were once active, but have succumbed to old age? Perhaps, but it hinges on one assumption that is not yet settled – that young granule neurons are preferentially involved in storing/processing information. Since no one has ever directly compared activation of young, adult-born neurons with activation of mature, perinatal-born neurons (with a birthdating marker such as BrdU), it remains possible that many of the activated cells in this study are perinatal-born, mature granule neurons (which Alme et al. acknowledge).

Here are some of the studies, and their limitations, that address the issue of whether young, adult-born neurons are more “activatable” than mature neurons:

1. Many electrophysiological studies have shown that immature neurons have enhanced plasticity and excitability. But does this mean they’re more likely to be active during behavior? The young neurons that are more likely to undergo LTP also have fewer, maybe weaker, synapses than mature neurons. So, maybe they’re even harder to activate. I don’t know.
2. Tashiro 2007 nicely demonstrated that environmental enrichment increased neurogenesis and the total number of young activated neurons. However, the proportion of young neurons that were activated was not different from the general population.
3. Studies that compare activation of BrdU+ cells to activation of the general population (e.g. cells stained for DAPI or NeuN) could introduce sampling biases due to the different methods used to quantify immature neurons and other/mature neurons. For example, Kee 2007 found preferential activation of BrdU+ adult-born neurons relative to NeuN+ cells but newer work from the same group, comparing CldU+ perinatal-born cells with IdU+ adult-born cells found no differences.
4. A similar sampling bias could explain the enhanced activation of young BrdU+ cells relative to the general population in the study by Ramirez-Amaya 2006. Then again, Ramirez-Amaya examined rats and my own (indirect) data from rats also suggests adult-born neurons may indeed be more “activatable”.

In short, the verdict is still out on what would seem like a simple question to answer – does behavioral stimulation preferentially activate young neurons relative to mature neurons? A definitive answer to this question will shed a lot of light on the function of adult neurogenesis and the dentate gyrus. In either case, thanks to Alme et al. for being the first to address the potential role granule neurons might play in encoding time, even if the only definitive conclusion is that most granule neurons appear to be doing absolutely nothing. Clearly there is a long way to go before we understand exactly what the dentate gyrus is doing.

*I only had so many beans and so many colored markers, so I apologize for not illustrating context D, but I think you get the idea.

Reference

Alme, C., Buzzetti, R., Marrone, D., Leutgeb, J., Chawla, M., Schaner, M., Bohanick, J., Khoboko, T., Leutgeb, S., Moser, E., Moser, M., McNaughton, B., & Barnes, C. (2010). Hippocampal granule cells opt for early retirement Hippocampus DOI: 10.1002/hipo.20810

A fundamental property of the hippocampus is its ability to rapidly encode memories while simultaneously keeping them distinct. Recording from hippocampal neurons one can clearly see that different populations of neurons are active as a rat explores two environments. This is thought to be one mechanism by which information is kept distinct in the brain.

For the last 15-20 years it has been thought that the dentate gyrus (DG), a major subfield of the hippocampus, serves to take small changes in incoming sensory information and orthogonalize them (i.e. make them more different). This idea was built in part on the fact that there are many more DG neurons than upstream cortical neurons. Thus, the DG could use completely different populations of neurons to represent different sets of incoming information and then pass on these representations to CA3, which may bind them into coherent events/memories (the interconnectedness of CA3 neurons, via “recurrent collatorals”, is thought to be a mechanism by which the different components of a memory are bound together).

However, a “problem” arose when Leutgeb et al. found that it is always the same population of dentate granule neurons (~1% of the total population) that are active as an animal explores different environments, even very different ones. This was a bit of a surprise. Still consistent with the proposed role of the DG in orthogonalizing information, however, was the fact that the DG neurons fired (i.e. generated action potentials, which transmit information from neuron to neuron) at different rates/frequencies in the different environments. Thus, changes in sensory information were represented by changes in patterns of activity within the same population of cells, not by recruiting different populations of cells. This is but one study – the question of how the DG encodes and extracts information is far from settled (e.g. what are the other 99% of granule neurons doing? Surely there is a situation in which they are active, no?). But the findings were robust and raise many questions, namely: How does the same population of DG neurons activate different populations of downstream CA3 neurons, during different experiences?

Until now I had been in denial, fixated on trying to understand what types of behavioral experiences might activate different populations of dentate gyrus neurons. But maybe now it’s time to face the data.

The consensus, both in vitro (e.g. here and here) and in vivo (here), seems to be that if DG neurons are sufficiently active they can reliably activate CA3 neurons. Can a single population of DG neurons account for the amount of CA3 activity seen in the behaving animal? Well, 1% activation of the total DG population (1 million neurons) is 10 000 DG neurons. Each DG neuron contacts about 10 CA3 neurons. So if all active DG neurons activated all their downstream targets, then you’d expect about 100 000 active CA3 neurons – a third of the population. Indeed, about 20% of CA3 neurons are active when a rat explores a novel environment. So it’s possible. But it’s probably unlikely.

One reason it’s unlikely is that it doesn’t explain how different populations of CA3 neurons are activated by different experiences if it is the same population of DG neurons that are always driving them. In other words, since DG neurons are relatively hard-wired to CA3 neurons, how could a given DG neuron activate a CA3 neuron under some conditions and not others? One answer is that maybe it doesn’t – quite a while ago, McNaughton et al. showed that, even when the DG is lesioned, CA3 neurons are still able to selectively encode spatial locations as a rat traverses the environment, probably due to direct inputs from the cortex. And so perhaps the primary function of the DG is not to selectively activate different CA3 populations. However, the DG could certainly shape activity within CA3 or insert unique information into the CA3 network. How?

One possible mechanism, which may be dead obvious to electrophysiologists, is frequency itself. Leutgeb et al. found that frequency of activity is how DG neurons encode information and so frequency of activity may also be the way DG neurons transmit information to CA3 during different experiences.

It has been known for some time now that the output of DG neurons, the mossy fiber axons, show extraordinary frequency-dependent synaptic facilitation. Basically, as a DG neuron fires more action potentials over shorter periods of time, the amount of neurotransmitter it releases onto CA3 neurons increases (thereby increasing the likelihood a CA3 neuron will in turn fire action potentials and be recruited to participate in memory encoding). This means that at low firing rates, a DG neuron will activate some CA3 neurons and, at higher firing rates, it will recruit different or at least additional CA3 neurons.

Wouldn’t this cause a problem where, as DG firing rates increase, it is not different populations of CA3 neurons that become activated, but more populations? Well, it is known that some DG neurons increase their activity, and others decrease their activity, as an animal has different experiences, so the net activity in CA3 could remain constant, while still activating different CA3 populations. However, the DG-CA3 circuitry is certainly complicated enough to allow for other mechanisms. For example, while the dentate gyrus projects to CA3, and it is connections between these hippocampal regions that are thought to encode memories, DG neurons actually contact more inhibitory interneurons than CA3 neurons. Furthermore, there is a wide variety of synaptic connections between DG neurons and interneurons and these connections can be made weaker or stronger in a state- and frequency-dependent manner. Suffice it to say, by firing at different frequencies, it is plausible that a given DG neuron could activate different populations of interneurons, which in turn could inhibit different populations of downstream CA3 neurons, making them less likely to participate in memory encoding.

This ties in loosely to a peculiarity of the dentate gyrus that, until now, has just been a source of pretty histological images (to me) – the variability of calbindin expression in dentate gyrus neurons. Calbindin is a protein that binds calcium, it acts as a buffer, and gives DG neurons their property of facilitation (briefly: A single action potential in a DG neuron will travel down the axon and trigger the opening of calcium channels in the synaptic terminal at a CA3 neuron. Calcium is necessary for neurotransmitter release and subsequent activation of the CA3 neuron. Calbindin will bind this small amount of calcium, thereby preventing neurotransmitter release and CA3 activation. However, as the number and frequency of action potentials increases, calbindin will fail to effectively “mop up” the extra calcium and neurotransmission will proceed.). If you look at the picture at the top of this post, you can see that the amount of calbindin varies greatly in DG neurons. Immature DG neurons, identified by PSA-NCAM expression, are devoid of calbindin (arrows point to clear examples) and even when they are quite mature (10w of age) 40% will still be devoid of calbindin (see my data in this montage). Lastly, calbindin expression can be modified by experience. So the variable and modifiable expression of calbindin might be yet another mechanism by which DG neurons are capable of shaping activity in CA3 neurons. Indeed, at least one study, from Robert Sapolsky’s lab, has shown that genetically altering calbindin expression in the dentate gyrus dramatically influences DG-CA3 physiology and impairs memory.

Thanks to A.P. for posing the question.

Reference

Leutgeb JK, Leutgeb S, Moser MB, & Moser EI (2007). Pattern separation in the dentate gyrus and CA3 of the hippocampus. Science, 315 (5814), 961-6 PMID: 17303747

One positive thing that came out of looking at these very immature neurons was that I got the chance to see several examples of pyknotic (dying) cells. Older, adult-born neurons also die, particularly after an experience (see here and here), but it’s infrequent and hard to visualize. However, a relatively large proportion of new neurons die within a few days of their birth making them easier to find – the cluster of cells shown below is an example that caught my attention.

You can clearly see two BrdU-labeled cells (in green; marked with arrowheads) that also express doublecortin (DCX; red). The blue stain, Hoechst, stains DNA allowing for the visualization of all cell nuclei. Collectively, these 3 stains tell us that the cells are 1-day-old (because BrdU was injected 1 day before brains were collected), that they’re neurons (because they express the immature marker DCX) and that they’re dying (because BrdU and Hoechst both label DNA and show that the DNA is condensed in a ball, as is typically seen when cells undergo pyknosis). The arrow points to a lucky, neighboring neuron that is not dying.

Why were these two cells born if they’re only going to die 24 hours later?

I can understand the speculation that neural activity influences the survival of more mature neurons in a “use it or lose it” manner – essentially, if a memory is stored in a young neuron there must be a mechanism to ensure that the neuron, and therefore the memory, survives. But is it possible that a similar mechanism also influences the survival of very immature neurons? It’s hard to imagine, since very young neurons do not have synapses and cannot participate in memory processing/storage. Consistent with this idea, Tashiro has shown that NMDA receptors (a synaptic ingredient essential for many forms of memory) regulate the survival of 2-3 week-old neurons, which are just beginning to form synapses, but not younger neurons that have not yet formed synapses. However, the possibility remains that learning could do something to these 1-day-old neurons – e.g. epigenetically imprint them – so that they have some sort of cellular memory that causes them to subsequently participate in certain behaviors but not others. Since information is typically thought to be stored at synapses, I can’t imagine that these memories could be terribly specific but they could bias a young neuron to be more involved in a general class of behavior (e.g. spatial memory vs. stress) that is associated with certain broad differences in activity (e.g. firing patterns, neuromodulators, hormones). It would be really cool if someone shows this.

I’ve always enjoyed making lists. As a kid I can remember writing lists of rhyming words, lists of all the Ocean Pacific clothes I owned, lists of all the people I knew. Many years later, I hope I’ve now made a list that is actually useful.

Adult neurogenesis is now undisputed. Pretty much on a weekly basis there is a new paper that examines both levels of adult hippocampal neurogenesis and behavior, attempting to draw a functional connection. The good news is that the argument for a behavioral function for adult neurogenesis continues to get stronger. The bad news is that there’s a massive pileup of data, and it’s becoming hard to filter through the relevant studies – first you have to find them amongst the 1000+ studies of adult neurogenesis. Then you have to read them. What behaviors are examined? Is there an effect of reducing or enhancing neurogenesis? What method is used to manipulate neurogenesis? What do other studies find that performed a similar analysis?

In this spreadsheet I’ve tried to provide summary answers to these questions. The data can be sorted by the type of behavior examined (e.g. depressive behaviors, memory etc), how neurogenesis was manipulated (e.g. via irradiation, transgenic tools or exogenous factors like anti-mitotic drugs), and behavioral effect.

It should be noted that I essentially took authors’ claims at face value and nothing here should be blindly accepted as evidence for or against a behavioral function for neurogenesis – read the papers! Task, neuronal age, and other methods should all be considered. Also, at this time, I have only entered data for a fraction of the studies, namely those that have claimed to use a technique specific for reducing neurogenesis. In reality, no such technique exists and I’d like to enter the same data for all studies that correlate neurogenesis with behavior, even those that have manipulated neurogenesis using methods that have widespread effects in the nervous system (e.g. exercise, enriched environment).

Kitamura et al. (2009) Adult Neurogenesis Modulates the Hippocampus-Dependent Period of Associative Fear Memory. Cell. 139:814-827.

It’s great to see this paper finally in print. At SFN 2008 the authors had a poster that generated a lot of excitement, at least in our circles.  And the poster was quite a sight: there was such a profusion of data that the poster poured off the easel, nearly reaching the floor. With 27 (!) supplemental figures in the final article, one has to wonder if this is the final straw that led to this article.

The authors use an ingenious approach to address an idea that has been floating around for a while: that adult neurogenesis regulates memory turnover in the hippocampus. The hippocampus appears to have a temporary role in memory storage — for instance, lesions to the hippocampus often impair recall of recent memories, but have little or no effect on remote memories. One interpretation of this result is that memories are stored in the hippocampus for only a short time and then transferred to other brain structures (likely neocortex) in a process called systems consolidation. This may then free up space in the hippocampus for new memories. Neurogenesis could be a mechanism for this process: for instance, the addition of new neurons to the dentate gyrus of the hippocampus might destabilize old memories and simultaneously provide a fresh substrate for new memories.

LTP lasts longer in irradiated rats (red circles) that lack adult neurogenesis.

Kitamura et al performed a truly heroic number of experiments to provide evidence that adult neurogenesis is involved in memory turnover in the hippocampus. First, they examined whether neurogenesis is important for the persistence of hippocampal long-term potentiation (LTP) in rats. LTP is an activity-dependent increase in the strength of synaptic connections between neurons, and is studied as a cellular model of memory.

They showed that blocking adult neurogenesis via irradiation extends the duration of hippocampal LTP: LTP lasts less than 2 weeks in control rats but lasts more than 3 weeks in irradiated rats (left). Thus, at a cellular/synaptic level, it would seem that neurogenesis causes memories to be cleared from the hippocampus.

Inactivating the hippocampus with TTX impairs memory in irradiated mice (red circles).

Then, in a series of fear conditioning experiments, the authors demonstrated that adult neurogenesis is in fact involved in the clearance of hippocampal memories in mice. In these experiments neurogenesis was blocked with irradiation or a genetic method prior to conditioning. On the surface, both control and neurogenesis-arrested mice acquired fear conditioning normally, and both groups remembered for at least a month. However, a critical difference emerged when the hippocampus was inactivated during remote recall testing: one month after training, hippocampus inactivation failed to impair recall in control mice, replicating a classic effect and suggesting the memory had been transferred to other brain regions. In contrast, in neurogenesis-arrested mice, hippocampal inactivation did impair recall of the remote memory, indicating that the memory still resided in the hippocampus. Importantly, hippocampal inactivation did impair recall in controls 1 day after conditioning, when memories had not yet been consolidated into the neocortex. Collectively, these findings suggest that adult neurogenesis contributes to the consolidation of memory to extra-hippocampal structures.

The authors then go ahead and answer some of the exciting new questions raised by these findings: For example, what happens if, instead of having reduced neurogenesis, you have increased levels of neurogenesis? Well, consistent with their hypothesis, mice that were given access to running wheels had increased neurogenesis (as expected) but as soon as 7 days after conditioning their memory was no longer hippocampal-dependent, unlike controls. Thus, increasing neurogenesis accelerated memory consolidation.

Another interesting finding was that the hippocampal-dependence of memory was prolonged when mice were irradiated only 11 days before conditioning. Since 11-day-old neurons do not even have functional (excitatory) synapses this suggests that it may not be functional young neurons present at the time of learning, but instead the synaptic integration of new neurons after learning that facilitates clearance of memory from the hippocampus.

Why else is this paper cool? Well, to date, adult neurogenesis has been implicated in many different types of behavior but it has been hard to identify common threads – this paper makes a new prediction that neurogenesis is specifically involved in consolidation. It also reconciles a discrepancy in the field – whether or not young neurons are involved in long-term memory. Some studies have found that reducing neurogenesis impairs long-term memory (Snyder 2005, Jessberger, 2009), but others have not observed this effect (Saxe, 2006). Kitamura’s data shows that even if long-term memory appears normal at a behavioral level, reducing adult neurogenesis will have altered the neural substrate of the memory. And so could these memories really be equivalent? This will be an exciting avenue for future studies to explore: how is hippocampus-dependent long-term memory (in neurogenesis-arrested animals) different from hippocampus-independent long-term memory?

It will also be interesting to further explore how arresting neurogenesis affects memory acquisition. If clearing memories from the hippocampus makes room for new memories, then shouldn’t the acquisition of new memories be hindered when clearance stops? Kitamura’s failure to observe significant impairments in memory acquisition may mean that neurogenesis is involved less in memory clearance than in reorganizing memory networks to include extra-hippocampal structures. That is, instead of simply expunging hippocampal memories, new neurons may facilitate the generation of additional memory representations in extra-hippocampal structures.

Another interesting issue is how the apparent give-and-take relationship between hippocampus and neocortex (presumably) arises. The data seem to suggest that the presence of a hippocampal memory representation prevents a cortical memory representation from arising. It’s as if the memory is a discrete entity that can be passed from hippocampus to cortex: as long as the memory remains in the hippocampus, the cortex is deprived. It would be great to know how –and why– this happens.

Kitamura T, Saitoh Y, Takashima N, Murayama A, Niibori Y, Ageta H, Sekiguchi M, Sugiyama H, & Inokuchi K (2009). Adult neurogenesis modulates the hippocampus-dependent period of associative fear memory. Cell, 139 (4), 814-27 PMID: 19914173