Modeling the Brain on ‘Replay’: A Q&A with Attila Losonczy

KIBS researchers aim to crack the code of the mammalian brain, starting with one of its memory networks. Neuroscientist Attila Losonczy discusses the ambitious plan and why it has received the support of President Obama’s BRAIN Initiative.

Neuroscientists can describe the way a brain cell works in exquisite detail, but they still can’t explain how those cells act together in networks to create our thoughts and actions. They’ve chipped away at the problem, mapping some neural networks in invertebrates, such as the one flies use to detect motion, but not in vertebrates, never mind in mammals.

Attila Losonczy, Kavli Institute for Brain Science at Columbia University, as well as Columbia’s Zuckerman Mind Brain Behavior Institute.
Attila Losonczy, Kavli Institute for Brain Science at Columbia University and Columbia’s Zuckerman Mind Brain Behavior Institute.

Now, a quartet of brain researchers, including Attila Losonczy of the Kavli Institute for Brain Science at Columbia University in New York City, as well as Columbia’s Zuckerman Mind Brain Behavior Institute, aims to change that. Working with mice, the group plans to describe for the first time how information is coded and flows through circuits of neurons to form memories that last a lifetime.

More specifically, they want to understand how cells in one brain region, the CA3 region of the hippocampus, generate waves of electrical activity that replay stored memories. These waves, called “sharp-wave-ripples” are a form of mental replay that is essential to making and retrieving long-term memories.

Over the next three years, they will use a suite of new technologies to create a full-scale computational model of CA3—with all of its parts, connections and emergent properties—to explain how it helps perform one of the brain’s most important cognitive functions.

The idea is audacious and attainable enough to have received funding from President Obama’s Brain Research through Advancing Innovative Neurotechnologies (BRAIN) Initiative, the decade-plus mission to unlock the mysteries of the brain by spurring the developing of new tools and technologies. (Today, the BRAIN Initiative awarded $40 million to researchers, the first round of what eventually may total $4.5 billion in spending.)

The project is a collaboration between Ivan Soltesz from the University of California, Irvine, Gyorgy Buzsaki from New York University (who is also a leader of the Neurodata Without Borders initiative), John Lisman from Brandeis University and Losonczy.

The Kavli Foundation spoke to Losonczy about why the time is right for this project and what success would mean for our understanding of memory.


THE KAVLI FOUNDATION: What are the big questions about the brain that you are trying to answer?

ATTILA LOSONCZY: I am interested in the neural networks that lie in the hippocampus, a cortical brain region that is critically important for “episodic memory,” our ability to learn and remember sequences of events. How do the various cell types in the hippocampus interact to perform this highly important cognitive function? How do they generate memories of our lifetime experiences and allow us to remember those experiences for a long time?

TKF: The main goal of the proposal you’ve just had funded is to make a detailed model of a part of the hippocampus that generates so-called “sharp-wave-ripples (SWRs).” What makes these waves so interesting?

LOSONCZY: They’re of major interest because they represent a replay of a memory, and that replay is thought to be necessary for us to make lasting memories, a process called consolidation.

During the last two decades or so, SWRs have been one of the major focuses of hippocampal memory research. The original observation linked those events to sleep. SWRs are very prominent and abundant during sleep, which suggests that sleep is critical for memory consolidation.

More recently, SWRs have been observed during wakefulness as well. During times when an animal is quiet but awake, there are short bursts of SWR activity. This suggests that SWRs are involved in on-the-fly strengthening of memory traces. In our research, we can take advantage of these “awake” waves to analyze their underlying circuitry.

TKF: No one has modeled a neural network in vertebrates before. What makes it so difficult? And why do you think this network in the hippocampus is a good starting point?

The brain replays stored memories while we sleep to make them lasting. New research aims to model this replay network in mice. (Courtesy: A. Losonczy)
The brain replays stored memories while we sleep to make them lasting. New research aims to model this replay network in mice. (Courtesy: A. Losonczy)

LOSONCZY: A major goal in neuroscience is to understand how networks of brain cells produce cognitive functions, such as memory. That is a difficult enterprise. Even the simplest networks contain an abundance of cell types, each with particular properties. How then can one understand the emergent properties of these cells and their connections? Work on simple invertebrate networks provides some guidance but there have so far been no attempts to meet these challenging criteria in the study of any region of the mammalian brain.

We think that the neural circuitry that lies in the CA3 region of the mammalian hippocampus—circuitry that underlies SWRs—is an excellent target for this first attempt. Here, we have a very tight link between a circuit and a behavior. And we can build on previous work establishing functional connectivity between the various cell types in the CA3.

TKF: What is your lab’s contribution going to be?

LOSONCZY: So my lab is developing brain imaging tools that we can use to look at the activity of hundreds of cells simultaneously in awake, behaving animals. We can analyze which cells are active during SWRs and in which order.

TKF: The BRAIN Initiative is focused on enhancing the technological capabilities of brain researchers. What new tools do you need to make this project work?

LOSONCZY: It is a great and persistent need in neuroscience to get detailed knowledge of connectivity—how different cell types connect to each other and the numbers of synapses. Having this kind of anatomical information helps us build and constrain our models. But until now, the methods used to study connectivity have been low-throughput. So we’ve proposed to develop a new method to rapidly quantify the synaptic connections between cells types using light microscopy instead.

TKF: How else does this project meet the goals of the BRAIN Initiative?

LOSONCZY: Our project was designed to integrate a whole range of experimental approaches to understand the function of an important brain network. Another requirement was to cross disciplinary boundaries. We’ve brought together a bunch of leaders in the field of hippocampal research, working at different levels, from the computational level down to single cells. We already have frequent interaction but this opportunity under the BRAIN Initiative really helped to bring us together.

TKF: Will being part of the BRAIN Initiative have an impact on your students—our future neuroscientists?

LOSONCZY: I’m really hoping for that. This is a great privilege for my collaborators and me and for my students. Through this collaboration, they are going to be exposed to the knowledge of our collaborators. They will also see how the BRAIN Initiative’s global aim of "understanding the brain” really materializes.

TKF: If this project succeeds, what impact could it have on the field?

LOSONCZY: We expect that the results will provide transformative insights into the mechanism of network computation in vertebrates in general, and mechanisms of episodic memory formation and storage in the mammalian brain in particular.

—Lindsay Borthwick, September, 2014