Circuits - Computation - Models

Circuits - Computation - Models

In our department, we are interested in how the brain computes. In particular, we want to understand neural information processing at the level of individual neurons and small neural circuits. As an example for neural computation, we study visual course control in the fruit fly Drosophila. This is a tractable system for the following reasons: It involves computations of modest complexity, these computations are implemented in circuits that contain a rather limited number of neurons (typically 10-20), and each of these neurons can be genetically targeted allowing manipulation and recording of its activity.

For a comprehensive picture of information processing in the visual system, we apply electrophysiology and 2-photon Calcium imaging to characterize the visual response properties of the various neurons involved, as well as genetic activation and inactivation to probe for their functional role within the circuit and in behavior. Computational modeling allows us to confirm our findings in a theoretical framework and make predictions for future experiments.


Directional motion information is used for visual course control, prey capture or escape from predators and is therefore of extreme importance for every sighted animal. However, the direction in which something is moving is not explicitly represented at the level of the photoreceptors, but rather needs to be computed by downstream neural circuits. We study this important and most basic neural computation in the fly visual system. more
Visual feedback arising from self-motion, termed optic flow, provides animals with a rich source of information regarding their own rotational and translational movement components as well as the spatial structure of the environment. In this project, we aim to elucidate how specific optic flow fields are detected by wide-field motion-sensitive neurons. Furthermore, we investigate the roles of local and global motion detection for course control in behaving flies. more
We investigate how the direction-selective T4/T5 neurons of the Drosophila visual system acquire their properties during development. We focus on the developmental programs endowing the different subtypes of T4/T5 neurons with different dendritic morphologies, which are thought to confer them with distinct directional tuning properties. more

Experimental Approaches

We apply Drosophila genetics to express arbitrary proteins in neurons of interest. For instance, cellular markers enable detailed anatomical analyses. As another example, biosensors convert neural activity into fluorescence signals, which can be detected by 2-photon imaging in three spatial dimensions. Last, in combination with physiological and behavioral readouts, the controlled expression of ion channels, toxins and other effectors interfering with neuronal physiology are indispensable to probe the roles of neurons in the computation of interest. more
We exploit the powerful genetic toolkit of Drosophila, as well as state-of-the-art transcriptomic technologies, in order to elucidate the molecular programs underlying the development and function of visual motion circuits. more
Information flow in the brain is represented by electrical activity of nerve cells. To capture this activity, we perform electrophysiological recordings via electrodes from identified visual neurons such as wide-field motion-sensitive cells in the lobula plate of the fly. By combining this technique with genetic silencing or optogenetic stimulation of upstream neural elements, we investigate the functional role and connectivity of specific cell types involved in motion detection. more
As an alternative to electrophysiology, we express protein biosensors that report neuronal activity via fluorescence. We use 2-photon microscopy to confine the necessary excitation light to a small moving voxel in the brain, thereby scanning neural activity in three dimensions without stimulation of photoreceptors. With this technique, we study the visual responses of multiple small, columnar neurons which are inaccessible to electrophysiological recordings. more
To explore the transformation from visual input to behavioral output we quantify turning and other parameters in tethered walking and flying flies. This approach affords high experimental control, for instance over the presented visual stimulus. In conjunction with selective neuronal activity manipulations, we use this approach in two ways: 1) to effectively probe the requirement of columnar non-direction-selective neurons for visual computations, and 2) to study the behavioral roles of higher-order optic flow detectors for walking and flight behavior. more
To obtain a more naturalistic view on motion vision we study course control in both freely walking and flying flies. To this end we take videos of flies and develop algorithms to automatically extract features such as position, orientation and posture. This approach is combined with visual stimulation and genetic manipulations of selected visual neurons to infer their roles in naturalistic behavior. more
In our studies of the motion detection circuit, modeling represents an integral part, accompanying our experiments on a daily basis. Modeling helps to check whether our intuitive explanation of a certain result is correct, what to expect in a subsequent experiment given that explanation and to define what we do not yet understand. Our models are based at several levels of biological complexity. more
Go to Editor View