Implementations of context-dependent predictions in the early olfactory system

• Applicant: Professor Dr. Wolfgang Kelsch
• Subject Area: Experimental and Theoretical Network Neuroscience ; Cognitive, Systems and Behavioural Neurobiology
• Term: since 2026
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 572462528

Olfactory perception is shaped not only by bottom-up sensory input but also by top-down modulation from cortical areas, allowing early sensory regions to integrate contextual and predictive information. This proposal investigates how two reciprocally connected olfactory cortices contribute to context-dependent prediction coding in the main olfactory bulb. Preliminary data using a novel Markov chain-based conditioning paradigm and related experiments show distinct top-down signatures from the two cortices that modulate olfactory bulb activity during learned olfactory sequences. Critically, we will exploit a strict computational definition of context, in which contextual cues do not acquire associative strength themselves but modulate stimulus–outcome relationships, allowing precise analysis of context-specific learning mechanisms. We aim to dissect how corticobulbar projections convey contextual, predictive, and error signals that influence olfactory bulb encoding. We will combine multi-site single-unit population recordings, fiber photometry, and optogenetic manipulations in behaving mice to understand the circuit logic underlying context-dependent learning. Further, we will model the learning process using biologically informed reinforcement learning algorithms. The project will benefit from the proposed Mercator Fellow. We will work with her on advanced methods for cell assembly detection and inter-regional coordination analysis. This work will provide mechanistic insights into how contextual modulation emerges in early sensory processing and how the brain implements flexible, predictive representations of the environment.

• DFG Programme: Research Grants