 Extracting Behaviorally Relevant Traits from Natural Stimuli: Benefits of Combinatorial Representations at the Accessory Olfactory BulbAnat Kahan and Yoram Ben-Shaul PLOS Computational Biology, 2016, vol. 12, issue 3, 1-25 Abstract: For many animals, chemosensation is essential for guiding social behavior. However, because multiple factors can modulate levels of individual chemical cues, deriving information about other individuals via natural chemical stimuli involves considerable challenges. How social information is extracted despite these sources of variability is poorly understood. The vomeronasal system provides an excellent opportunity to study this topic due to its role in detecting socially relevant traits. Here, we focus on two such traits: a female mouse’s strain and reproductive state. In particular, we measure stimulus-induced neuronal activity in the accessory olfactory bulb (AOB) in response to various dilutions of urine, vaginal secretions, and saliva, from estrus and non-estrus female mice from two different strains. We first show that all tested secretions provide information about a female’s receptivity and genotype. Next, we investigate how these traits can be decoded from neuronal activity despite multiple sources of variability. We show that individual neurons are limited in their capacity to allow trait classification across multiple sources of variability. However, simple linear classifiers sampling neuronal activity from small neuronal ensembles can provide a substantial improvement over that attained with individual units. Furthermore, we show that some traits are more efficiently detected than others, and that particular secretions may be optimized for conveying information about specific traits. Across all tested stimulus sources, discrimination between strains is more accurate than discrimination of receptivity, and detection of receptivity is more accurate with vaginal secretions than with urine. Our findings highlight the challenges of chemosensory processing of natural stimuli, and suggest that downstream readout stages decode multiple behaviorally relevant traits by sampling information from distinct but overlapping populations of AOB neurons.Author Summary: Across the animal kingdom, chemical senses play a central role in guiding social behaviors by conveying information about particular behaviorally relevant traits. However, decoding these traits from profiles of chemical cues is challenging since cue levels are modulated by multiple factors. Here, we investigate how the mouse vomeronasal system, a chemosensory system important for processing social information, detects behaviorally relevant traits from natural stimuli. We focus on detection of a female’s genetic background (a model for individuality) and estrus-state (a measure of sexual receptivity) by neurons in the first vomeronasal brain relay, the accessory olfactory bulb (AOB). We show that information about both genetic background and receptivity can be obtained from various stimulus sources: urine, vaginal secretions, and saliva. Importantly, while individual AOB neurons can only provide limited decoding ability of these traits, simple networks sampling AOB neuronal ensembles provide considerable improvement. Our analyses highlight an overlooked challenge associated with chemosensory processing and suggest how it can be overcome by downstream neurons that read information from multiple AOB neurons. Date: 2016 Downloads: (external link) Related works: Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text Persistent link: https://EconPapers.repec.org/RePEc:plo:pcbi00:1004798 DOI: 10.1371/journal.pcbi.1004798 Access Statistics for this article More articles in PLOS Computational Biology from Public Library of Science |
Is your work missing from RePEc? Here is how to contribute.
Questions or problems? Check the EconPapers FAQ or send mail to .
EconPapers is hosted by the
School of Business at Örebro University.