Friedemann Zenke (left) and Johannes Felsenberg (right).

June 24, 2019

In conversation with our two new group leaders

Johannes Felsenberg and Friedemann Zenke both joined the FMI as group leaders in June. While Johannes focuses on neural circuit mechanisms of memory re-evaluation, Friedemann is a computational neuroscientist interested in learning and the principles of neural computation. In this interview, the two new group leaders tell us more about their research.

You both joined the FMI from the University of Oxford. Did you know each other before?
Johannes: Yes indeed, we were both postdocs at the Centre for Neural Circuits and Behaviour (CNCB) at the University of Oxford – an institute whose interests overlap with those of the FMI neuroscience community – and we’ve known each other for about three years. Workwise, we share a strong interest in memory-related processes. However, it’s a coincidence that we both applied at the same time for a group leader position at the FMI, and it’s great that we both got a job here!

Why did you each decide to apply for a group leader position at the FMI?
Johannes: The FMI is outstanding in the field of neural circuits, learning and memory, olfactory processing, etc. So I’d read about the research of FMI scientists and seen presentations by some group leaders in Oxford. I knew that the FMI is an excellent place to do the kind of research I’m interested in. Moreover, Basel is an attractive location – which also matters when you’re looking for a new position. FMI was at the top of my list, and I’m very glad to be here now!
Friedemann: If computational neuroscience is to further our understanding of the brain, it’s important that the modeler and the bench scientist learn to speak a common language. Only then is it possible for synergies to arise. I’ve always wanted to set up a group at an interdisciplinary neuroscience institute which provides the potential for such synergies. In joining the FMI, this dream is now coming true.

Johannes, could you briefly summarize your research interests?
With my group, I want to understand how memory can be re-evaluated when an animal finds out that what it initially learned is not true anymore, and what this means for the underlying circuits and the places where the memory is stored. We know that learned information can be modified in two ways – firstly, by adding new information, a new and parallel memory, which supplements the old one; this process is called memory extinction. It’s an area that I’ve studied extensively and which will remain a focus of my research. Secondly, it can be modified by updating the original memory through a phenomenon known as memory reconsolidation – here, the memory is switched back from a stable to a vulnerable form, which can be updated and then needs to be stabilized again. This process is poorly understood, and I’ll be investigating it here.
We are working with fruit flies (Drosophila melanogaster); their numerically simple and accessible brain represents an ideal system for understanding fundamental neural operations of memory re evaluation and reconsolidation. Flies can learn to seek or avoid odors previously associated with reward or punishment, and we have a good understanding of where memories are located in the brain. Genetic tools in the fly allow us to manipulate and monitor the neural circuitry involved in information storage, down to single-neuron resolution. Applying these tools will enable us to elucidate the circuit mechanisms underlying memory re-evaluation.

Johannes, what is the long-term goal driving this research?
I want to understand how computations in neural circuits give rise to behavior, in particular how this works in the fly brain. However – like most of my colleagues I guess – I’m interested in general principles; I would like to find something that is of very general truth, e.g. the logic of a circuit motif, which holds true for most animals and could therefore guide research in other model systems. In fact, what I’m trying to do in flies – changing a memory – is something a psychotherapist is trying to do with patients suffering from maladaptive memories such as drastic fear memories. So it would be a major achievement if my fly research could help, at some level, to inform strategies for addressing human conditions.

Friedemann, could you also please briefly summarize your research interests?
First of all, I’m a computational neuroscientist. That means I use theoretical models to study how the neural networks in our brains perform computations. Like many of my experimental colleagues at the FMI, I’m interested in how neurons work together at the circuit level. One key determinant of how neural networks process information is their synaptic connectivity. Importantly, synapses are plastic, and we use this plasticity to learn and to store memories. My research focuses on plasticity and how it shapes network function. I’m interested in how plasticity, which occurs at individual synapses, is orchestrated among many neurons, and how this sophisticated interplay of different plasticity mechanisms ultimately forms a functional circuit at the network level. Specifically, the goal is to develop plasticity models which give rise to functional circuits in the computer, and then compare our insights with my colleagues’ biological findings, to see both the commonalities and what doesn’t fit. By seeing how our models break and then iteratively refining them, we can steadily improve our understanding.

Friedemann, how exactly would you explain the relevance of the theoretical neuroscience which you do and how it can complement the research of experimental scientists?
Modern neuroscience comprises approaches ranging from the molecular, through the cellular, all the way to the behavioral level. In addition, technical advances now allow us to study neural circuits in unprecedented detail. Together, these approaches provide us with a detailed picture of the often high-dimensional and strongly nonlinear circuits of the brain. To deal with this complexity, theory and computational modeling are indispensable tools that help us understand what these circuits do, why they operate in a certain way, and how they do so. By developing conceptual models at multiple levels of abstraction, theoretical work tries to build bridges between the different observational levels and help to identify unifying principles. Ultimately, it’s important that we connect our models with experimental network science, as performed at the FMI and elsewhere. To that end, I’m excited about the possibility of engaging in collaborations where theory and bench work are conducted side by side. I firmly believe that interdisciplinarity is key to unraveling the mysteries of the brain.

How many group members have you managed to recruit so far?
Friedemann: I’ve already hired two PhD students through the FMI PhD program who seem to be fantastic. I would like to hire 1-2 additional PhD students in the next round of the program. It’s important for my group to build up a critical mass of people to allow for fruitful scientific discussions within the team. The students I’m looking for typically have a quantitative background in mathematics, physics or computer science, with a keen interest in neurobiology; when they graduate, they ought to be fluent in the languages of both neurobiology and theoretical neuroscience.
Johannes: I’