Analysis of combinatorial olfactory coding

信息素气味的处理是在所谓的标记线系统中进化的，在这种系统中，给定的信息素混合物的单一气味与特定的受体以及特定的下游处理路径结合。信息素气味编码问题的解决方案是根据交配位置的特定进化限制和对同种异种与密切相关物种的区分而进化出来的。在非信息素气味的情况下，嗅觉系统似乎已经进化出少量相对宽调的感受器以及肾小球，通过以组合的方式处理气味信息来编码大量短暂的气味，从而沿着一些气味维度评估气味。事实上，就其本质而言，组合编码方案比标记线方案具有更大的气味识别能力。然而，虽然气味编码的分布变得越来越清楚，但气味维度的规范仍然定义不清。R03的具体目标是开发方法来量化和分析非信息素气味在信号处理中的变化，这些非信息素气味沿着一组有限的拟议气味维度系统地变化，例如碳链长度，以及经验是否增强了这些表示的统计独特性。目标1旨在评估相关气味在时空重叠或组合代码中被编码的程度，组合代码沿着将这些气味的几何形状的相应变化编码的维度系统地变化。在这里，我们将对气味阵列的时空相关性进行统计建模，这些阵列沿着特定的几何维度系统地变化。研究还表明，学习会以某种清晰且可重复的方式改变气味编码。然而，人们对这些变化所起的作用知之甚少。一种可能的可能性是，这些变化增强了气味的辨别能力。目标2通过观察这些重叠的时空气味编码的关联性是如何随着飞蛾对食物相关气味的适应而变化的，从而进一步发展了目标1。

Processing of pheromonal odorants has evolved in what is referred to as a labeled-line system, whereby single odorants of a given pheromonal blend bind to specific receptors as well as specific down stream processing pathways. The solution of the pheromonal odor-coding problem has evolved in response to specific evolutionary constrains of mate location and discrimination of conspecifics from those of closely related species. In the case of non-pheromonal odors, olfactory systems appear to have evolved with a small number of relatively broadly tuned receptors, as well as glomeruli, to code for a large number of ephemeral odors by processing odor information in a combinatorial manner whereby odors are assessed along a number of odor dimensions. Indeed, by their very nature, combinatorial coding schemes have a far greater capacity for odor discrimination than do labeled line schemes. However, while it is becoming increasingly clear that odor codes are distributed the specification of odor dimensionality is still poorly defined. The specific aims of this R03 are designed to develop methods for quantifying and analyzing variation in signal processing of non-pheromonal odors that vary systematically along a restricted set of proposed odor dimensions, such as carbon chain length, and, whether experience enhances the statistical distinctiveness of these representations. Aim 1 is designed to assess the degrees to which related odors are coded for in spatiotemporal overlapping, or combinatorial codes that vary systematically along dimensions that codify corresponding variation in the geometry of these odors. Here we will statistically model the spatiotemporal relatedness of arrays of odors that systematically vary along specified geometric dimensions. It has also been demonstrated that learning changes the odor code in some clear and reproducible ways. Yet the role these changes play are poorly understood. One likely possibility is that these changes enhance the discriminability of odors. Aim 2 further develops Aim 1 by looking at how the relatedness of these overlapping spatiotemporal odor codes changes as moths are conditioned to related odors to food.