Disorder and the neural representation of complex odors: smelling in the real world

Abstract

Animals smelling in the real world use a small number of receptors to sense a vast number of natural molecular mixtures, and proceed to learn arbitrary associations between odors and valences. Here, we propose a new interpretation of how the architecture of olfactory circuits is adapted to meet these immense complementary challenges. First, the diffuse binding of receptors to many molecules compresses a vast odor space into a tiny receptor space, while preserving similarity. Next, lateral interactions "densify" and decorrelate the response, enhancing robustness to noise. Finally, disordered projections from the periphery to the central brain reconfigure the densely packed information into a format suitable for flexible learning of associations and valences. We test our theory empirically using data from Drosophila. Our theory suggests that the neural processing of olfactory information differs from the other senses in its fundamental use of disorder.Animals sense and respond to volatile molecules that carry messages from and about the world. Some kinds of olfactory behaviors require sensing of particular molecules such as pheromones. These molecules and the receptors that bind to them have likely co-evolved over long periods of time to ensure precise and specific binding. However, to be useful as a general purpose tool for interaction with a diverse and changing world, the olfactory system should be prepared to sense and process any volatile molecule. There are a very large number of such monomolecular odorants (perhaps billions [1]), far more than the number of receptor types available to bind these odorants. Humans and mice, for instance, have just ∼ 300 and ∼ 1000 functional olfactory receptor types, respectively. Yet, animals may be able to discriminate between orders of magnitude more odors than the number of receptor types (a high estimate is given in [2], but see [3]).At an abstract level, the early stage of the olfactory system faces the immense challenge of embedding a very high-dimensional input space (the space of odor molecules) into a low-dimensional space of sensors (the response space of olfactory receptors). This embedding must preserve similarity between different odors well enough to permit the judgements of sameness and difference that are crucial for behavior. Furthermore, experiments [4] suggest that this odor representation is reorganized in higher brain regions to be enormously flexible, allowing learning of nearly arbitrary associations between valences and different groups of odors. Here, we propose * co-first author † lead author a new theoretical framework (Fig. 1), and provide empirical evidence, suggesting that the olfactory system powerfully exploits physiological and structural disorder at different stages of processing to meet these two complementary challenges: (i) compression of a vast odor space into a tiny receptor space, and (ii) reorganization of the information to allow flexible learning. To perform effectively within its design constraints, a sensory system must exploit st...

Adaptation of olfactory receptor abundances for efficient coding

“…For example, the distribution of odorants could be modeled using a Gaussian mixture, rather than the normal distribution used in this paper to enable analytic calculations. Each Gaussian in the mixture would model a different odor object in the environment, more closely approximating the sparse nature of olfactory scenes discussed in, e.g., [32].…”

“…This can be thought of as a maximum-entropy approximation of the true distribution of odorant concentrations, constrained by the environmental means and covariances. This simple environmental model misses some sparse structure that is typical in olfactory scenes [51,32]. Nevertheless, approximating natural distributions with Gaussians is common in the efficient-coding literature, and often captures enough detail to be predictive [2,5,52,13].…”

“…In contrast to previous work applying efficient-coding ideas to the olfactory system [29,30,31,32], here we take the receptor-odorant affinities to be fixed quantities and do not attempt to explain their distribution or their evolution and diversity across species. Instead, we focus on the complementary question of the optimal way in which the olfactory system can use the available receptor genes.…”

The optimal odor-receptor interaction network is sparse in olfactory systems: Compressed sensing by nonlinear neurons with a finite dynamic range

“…For example, Zhang and Sharpee proposed a fast reconstruction algorithm in a simplified setup with binary ORNs and binary odor mixtures without concentration information [43]. In another work, Krishnamurphy et al studied how the overall "hour-glass" (compression followed by decompression) structure of the olfactory circuit can facilitate olfactory association and learning, again with the assumption that ORN responses to odor mixtures are linear [44]. Following ideas in CS theory, Singh et al recently proposed a fast olfactory decoding algorithm that might be implemented in the downstream olfactory system [45].…”

Front-end Weber-Fechner gain control enhances the fidelity of combinatorial odor coding

“…In antennal lobe (AL) glomeruli, mutual lateral inhibition normalizes population response, reducing the dependency of activity patterns on odor concentration [21,22]. Further downstream, sparse connectivity to the mushroom body (MB) helps maintain neural representations of odors, and facilitates compressed sensing and associative learning schemes [23][24][25][26]. Finally, temporal features of neural responses contribute to concentration-invariant representations of odor identity [27][28][29][30].…”