Event Abstract Back to Event Sparse and Invariant Representations of Multi-component Odors in the Mushroom Body How does the brain form singular and invariant percepts from complex stimuli? We address this question in the insect brain by examining processing in the antennal lobe (AL) and the mushroom body (MB), the first and second neuropils of the olfactory pathway. The MB is a structure analogous to the vertebrate olfactory cortex and plays a key role in the formation and recall of olfactory memories. Its intrinsic neurons are 50,000 Kenyon cells (KCs), which receive direct input from 830 projection neurons (PNs) of the AL. Odor representations differ dramatically in the AL and MB. Representations are distributed and dynamic in the AL, but are sparse, brief and synthetic in the MB. Ensemble PN responses can be described geometrically by stimulus-specific trajectories reflecting the state of the AL network. How do these trajectories change with changes in the stimulus? To answer this question, we first probed the PNs with stimuli changing progressively from one to the other. This produced a progressive rather than abrupt transformation from one odor-specific trajectory to another, suggesting that the PN network may be able to optimize the use of its encoding space. It has been shown that families of concentration-specific trajectories form odor-specific manifolds, (Stopfer et al., 2003). We find that the PN trajectories can form manifolds along various stimulus parameters. For example, by keeping the ratio of the binary mixture the same, we can form ratio-invariant manifolds. By combining odors progressively, we can form manifolds across odor mixtures. Next, we systematically increased the complexity of the mixture by combining odors in various ways to make mixtures with increasing number of components (1-8). We estimated the PN trajectories to mixtures from those evoked by their components using a linear model, and examined the deviation between these estimated trajectories and the experimentally observed trajectories. We found that linear predictions are reasonably good for binary mixtures, but that they worsen rapidly as more components were added. PN trajectories for single odors (e.g., A) were very different from those produced when 3-7 other odors were added (e.g., ABCDW). Surprisingly, we found that many KCs could detect single components (e.g., A) contained in a mixture (e.g., AB, ABC, - ABCDWXYZ). How is this possible, when PN population output seems so different across mixture conditions? The intuition is that individual KCs sample only subspaces of the entire PN space, and within these subspaces, representations of components and mixtures overlap. To test this intuition, we developed a simple KC model, to be tested with experimental PN data as input. The model consists of 4 features derived from experimental findings: (i) random 50% PN-KC connectivity, (ii) delayed feed-forward inhibition to mimic Lateral Horn interneurons, (iii) non-linear EPSP amplification and sharpening, and (iv) adaptive KC spike-threshold for gain control. We find that model KCs match the probability of KC responses and ensemble PSTHs, and in some instances, individual model KCs perform odor segmentation just as real KCs do. This model thus seem sufficient to explain high-level computations that arise in the MB. Conference: Computational and systems neuroscience 2009, Salt Lake City, UT, United States, 26 Feb - 3 Mar, 2009. Presentation Type: Poster Presentation Topic: Poster Presentations Citation: (2009). Sparse and Invariant Representations of Multi-component Odors in the Mushroom Body. Front. Syst. Neurosci. Conference Abstract: Computational and systems neuroscience 2009. doi: 10.3389/conf.neuro.06.2009.03.108 Copyright: The abstracts in this collection have not been subject to any Frontiers peer review or checks, and are not endorsed by Frontiers. They are made available through the Frontiers publishing platform as a service to conference organizers and presenters. The copyright in the individual abstracts is owned by the author of each abstract or his/her employer unless otherwise stated. Each abstract, as well as the collection of abstracts, are published under a Creative Commons CC-BY 4.0 (attribution) licence (https://creativecommons.org/licenses/by/4.0/) and may thus be reproduced, translated, adapted and be the subject of derivative works provided the authors and Frontiers are attributed. For Frontiers’ terms and conditions please see https://www.frontiersin.org/legal/terms-and-conditions. Received: 02 Feb 2009; Published Online: 02 Feb 2009. Login Required This action requires you to be registered with Frontiers and logged in. To register or login click here. Abstract Info Abstract The Authors in Frontiers Google Google Scholar PubMed Related Article in Frontiers Google Scholar PubMed Abstract Close Back to top Javascript is disabled. Please enable Javascript in your browser settings in order to see all the content on this page.