 Approximately seven years ago, I experienced an event which, by all means, is the most relevant experience in my life. And that is the birth of my daughter. As a neuroscientist, I'm interested what changes in the brain happen given such strong experiences. What we know so far is that when we experience something very relevant, there are a subset of neurons in our brains which get activated and by activation of this subset of neurons, our brains sort of represent that experience in a way they translate the outer world into the inner code of the brain. It has also been shown that those neurons can help to recall a memory of that experience far in the future. In fact, people working with mice have demonstrated that they can not only label those neurons which have been activated by an experience but they can also reactivate them to artificially recall that experience or even silence them to have the animals somehow forget a given experience. Those neurons are also called angram neurons. Angram is a jargon term that basically refers to the fact that these neurons are a physical trace of a memory. As you can imagine, those neurons are very important. However, nobody really knows how these neurons are chosen. I have millions of neurons in my brain and nobody really knows what are the factors that decide which neurons are active today to represent this experience and will help me to recall it in the times to come. So what are the specific, either cellular or circuit mechanisms that force a neuron to become part of an angra? Neurons communicate to each other very intensively and they communicate to each other through synapses. Synapses basically means a connection and neurons do not work in isolation. They are embedded in, if you will, a social network. So there is a lot of neurons which communicate to each neuron and each neuron communicate to a lot of neurons. Pretty much like us in a social network, what a neuron says or its activity pattern heavily depends on what is being told or what information it finds on the field, what information it's told by other neurons. So our hypothesis was that maybe the deciding factor for a neuron to take part into an angra today in this place is just how much communications or how many synapses it receives from other neurons that project to it. We another developed a method to image dendritic spines and hence synapses in the brain of live animals and this allows us to track the dynamics of those synapses through time. Importantly, we could also label angra neurons and so we could image the synapses of angra neurons and at the same time we could image the synapses of neighboring neurons to those which by chance didn't become angra neurons. This is how it works. We image animals for approximately one week to establish the baseline dynamics of their synapses. Then we had the animals experience a very salient experience. We put them in a novel environment with a novel food, some toys and we let them have fun for a night. And at the same time we injected a drug that would allow us to label the angra neurons active during this night. For the following week we kept imaging the synapses of the neurons. Now, however, we could distinguish between neurons which became angra neurons and neurons which didn't and so we could compare the dynamics of angra neurons and non-angra neurons. Interestingly, we could also, if you will, travel back in time and as we have imaged this synapses in vivo, we could go back to the moment in which the neurons were not yet angra neurons and then compare the synapses of a neuron before and after it became an angra neuron. I'd like to highlight two main findings of our work. Thanks to the fact that we could actually image angra neurons and their connections in live animals, we could compare connectivity of angra neurons before they became angra neurons and after they became angra neurons. What we found is that the main difference between angra and non-angra neurons is actually before angra neurons become angra neurons. We found that neurons with a more stable connectivity tended to become angra neurons more often or more efficiently than neurons with a less stable connectivity. To go back to the social network analogy, for each neuron it was very important not only how many friends would talk to it, but also for how long each of the friends was able to talk to it. The second finding relates to the ability of animals to learn and recall. People have been studying the relationship between structural connectivity and the ability to learn and recall for many years now, mostly in another brain area, which is called the neocortex. Now, as we can image the hippocampus, we can also imagine the neocortex, which is called the neocortex. Now, as we can image the hippocampus, we can also draw a connection between stability of connectivity and the ability of animals to learn and recall. What people have found in the neocortex is the more stable the connectivity, the more the animals are able to learn and recall for a long time. The hippocampus, to our somewhat surprise, we find the opposite. Animals with more dynamic connectivity, so with less stable connectivity, tend to be better at recalling a task which is dependent on the function of the hippocampus. That seems to be a little bit counter-intuitive at first, but maybe an analogy with the computer can help. In a computer, the hard drive, it's a permanent storage of information, and in a permanent storage, you want to be able to write information and you want this written information to stay stable as long as you can. This, if you will, by and large, corresponds to the neocortex of an animal and of our brain. The hippocampus, however, seems to have a different function. The hippocampus is supposed to store memories only for a short period of time, and during this time, the function of the hippocampus is, among other things, to transfer this information to the neocortex. In this framework, one can imagine how an improved ability to read and write information and to erase it when it's not needed anymore might be beneficial for the hippocampus. And this is what we think we see in our animals. Animals with higher dynamics are able to write faster and, at the same time, erase faster information, and thus, maybe for this reason, they're also able to recall better memories. So the first finding is a basic as well as a clinical relevance. The basic side of things, it's obvious. So far, not much was known about the mechanism that will lead to the formation of angrioms. And for the first time, we could propose that the intrinsic stability of connectivity or a temporary stabilization of connectivity in a subset of cells can be a deciding factor for this subset of cells to become angriom neurons. The clinical relevance of this, it's a little bit more for the future. There are some diseases that lead to impairment in the ability of patients, of people, to form and recall new memories. And understanding better the basic mechanism that lead to the formation of the neurons, which are involved in this process, might lead to device better treatments for these patients. The second finding has a broad relevance for the way we understand the brain works. So far, it has been known that the neocortex acts as a long-term storage for memory, while the hippocampus acts as a shorter term storage for memory and information. Now, we are starting understanding what are the cell biological processes that lead to that function. Namely, one of them can be the stability of the connectivity, the differential stability of connectivity of these two regions. In this work, we could show a strong correlation between stability of connectivity of cells on the one hand and the probability for these cells to become angriom neurons. What we want to do in the future is to demonstrate causality. That is, if we manipulate the stability of the connectivity of a single cell, can we force this cell to become part of an angriom? The other thing we want to do is to study at the single cell level how patterns of connectivity lead to patterns of activity. That is, to image in a single cell in the brain of a live animal its connectivity and its activity patterns, and to try to establish a qualitative and quantitative relationship between connectivity pattern and activity pattern at the single cell level.