Uncovering circuits for stimulus-driven valence integration and action selection

• Applicant: Dr. Anna-Maria Jürgensen
• Subject Area: Experimental and Theoretical Network Neuroscience
• Term: since 2025
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 552223197

Goal-oriented, adaptive behaviour is essential in complex, dynamic environments. It helps animals avoid threats and seek out advantageous circumstances by selecting behaviour towards avoiding or approaching innately valent, inherently meaningful sensory stimuli. Additionally, the valence of initially neutral stimuli can be learned or updated upon positive or negative experience. Goal-directed behaviour in each moment depends on the dynamic integration of innate and learned valences. In the insect brain, this integration occurs in the convergence of two pathways, encoding innate and learned valences, the lateral horn and the mushroom body, respectively. Past research has targeted the detection and encoding of sensory stimuli and the representation of their innate or learned valences. Especially the acquisition of learned valence and its effect on behaviour have been investigated extensively. However, insights into interaction and integration with innate valences are sparse. Beyond the scope of previous theoretical and experimental work, I will implement the first computational whole-brain model to dynamically investigate processing from sensory input to behaviour to disentangle the contributions of learned and innate valence to behaviour in each moment. The Drosophila larva is the ideal model for this approach due to its small network size, available connectome, and comprehensible behavioural repertoire. I will derive the network architecture of the entire brain from the connectome. The free model parameters will be fit using Calcium imaging data that determines the response of key neurons in the valence integration circuit to sensory stimuli before, during, and after learning. This will yield a comprehensive computational model that I will use to make predictions about the functional role of the individual elements of the interaction and convergence circuitry of the lateral horn and the mushroom body for valence integration. I will obtain such predictions in simulation experiments by manipulating individual neurons or connections directly. Finally, I will conduct action selection experiments that feature camera tracking of freely behaving larvae. Their approach or avoidance behaviour in response to sensory input will be quantified before, during, and after learning as a proxy for innate and learned valence. I will optogenetically activate or inactivate key brain neurons to test the model predictions. This multidisciplinary approach will exploit the complementary benefits of model and animal experiments to first isolate separate circuit mechanisms for valence integration and then verify their predicted effects in naturalistic animal experiments. In conclusion, this Walter Benjamin Fellowship will identify how stored and new information about sensory stimuli interact to influence animal behaviour.

• DFG Programme: WBP Fellowship
• International Connection: United Kingdom
• Host: Professorin Marta Zlatic, Ph.D.