Research interests

The main research focus of the Rupprecht group is to understand learning and plasticity, in particular from the perspective of single neurons. Several projects are ongoing or under planning that are stepping stones towards this goal.

Behavioral timescale synaptic plasticity

Behavioral timescale synaptic plasticity (BTSP)  is an intriguing single-cell plasticity rule that has first been observed in the pyramidal cells of hippocampal CA1. The plasticity-inducing so-called plateau potentials are accompanied by bursts and dendritic calcium spikes. We would like to better understand the relevance, role and cellular underpinnings of such BTSP events.

Closed-loop experiments

One major limitation of artificial networks is that they are typically not embedded in the natural world. For animals, in contrast, it is crucial that they act upon the world and receive direct feedback about their action in order to be able to learn. We therefore aim to include such feedback loops as essential ingredients in our experimental designs. Mice will interact with virtual environments through their motor actions or their neuronal activity.

Self-organization of neural networks

Similar to animals, single cells need to receive feedback about their own actions. Along this line of thought, we aim to interpret neurons as control units that emit actions (action potentials) and receive feedback about these actions. This viewpoint will be used to conceptualize and model experimental results. In addition, we are also aiming to build models of self-organized neuronal networks in the future.

Calcium signals in astrocytes

Astrocytes are non-neuronal cells in the brain that react to all kinds of inputs and have been involved in neuronal plasticity. We have previously described how astrocytes integrate information about past events, and how calcium signals related to salient past events propagate to the astrocytic soma. We are interested in better understanding these observations and their role for neuronal circuits.

Calcium imaging analysis and spike inference

Calcium imaging is often used as a proxy readout for neuronal action potentials. At the same time, calcium is also an important signaling molecule involved in neuronal plasticity. We have previously worked on measuring and modeling the relationship between recorded calcium signals and action potentials. As a result, we have set up a large database of simultaneous calcium imaging and cell-attached recordings and have developed a deep learning-based method to predict spike rates from unseen calcium imaging data (CASCADE). We are continually working on expanding and improving these approaches.