A Robust Feedforward Model of the Olfactory System

Most natural odors have sparse molecular composition. This makes the principles of compressed sensing potentially relevant to the structure of the olfactory code. Yet, the largely feedforward organization of the olfactory system precludes reconstruction using standard compressed sensing algorithms. To resolve this problem, recent theoretical work has shown that signal reconstruction could take place as a result of a low dimensional dynamical system converging to one of its attractor states. However, the dynamical aspects of optimization slowed down odor recognition and were also found to be susceptible to noise. Here we describe a feedforward model of the olfactory system that achieves both strong compression and fast reconstruction that is also robust to noise. A key feature of the proposed model is a specific relationship between how odors are represented at the glomeruli stage, which corresponds to a compression, and the connections from glomeruli to third-order neurons (neurons in the olfactory cortex of vertebrates or Kenyon cells in the mushroom body of insects), which in the model corresponds to reconstruction. We show that should this specific relationship hold true, the reconstruction will be both fast and robust to noise, and in particular to the false activation of glomeruli. The predicted connectivity rate from glomeruli to third-order neurons can be tested experimentally.     

Genetic and functional subdivision of the Drosophila antennal lobe

Olfactory systems confer the recognition and discrimination of a large number of structurally distinct odor molecules. Recent molecular analysis of odorant receptor (OR) genes and circuits has led to a model of odor coding in which a population of olfactory sensory neurons (OSNs) expressing a single OR converges upon a unique olfactory glomerulus. Activation of the OR can thus be read out by the activation of its cognate glomerulus. Drosophila is a powerful system in which to test this model because the entire repertoire of 62 ORs can be manipulated genetically. However, a complete understanding of how fly olfactory circuits are organized is lacking. Here, we present a nearly complete map of OR projections from OSNs to the antennal lobe (AL) in the fly brain. Four populations of OSNs coexpress two ORs along with Or83b, and a fifth expresses one OR and one gustatory receptor (GR) along with Or83b. One glomerulus receives coconvergent input from two separate populations of OSNs. Three ORs label sexually dimorphic glomeruli implicated in sexual courtship and are thus candidate Drosophila pheromone receptors. This olfactory sensory map provides an experimental framework for relating ORs to glomeruli and ultimately behavior.     

Information processing in the olfactory systems of insects and vertebrates

Insects and vertebrates separately evolved remarkably similar mechanisms to process olfactory information. Odors are sampled by huge numbers of receptor neurons, which converge type-wise upon a much smaller number of principal neurons within glomeruli. There, odor information is transformed by inhibitory interneuron-mediated, cross-glomerular circuit interactions that impose slow temporal structures and fast oscillations onto the firing patterns of principal neurons. The transformations appear to improve signal-to-noise characteristics, define odor categories, achieve precise odor identification, extract invariant features, and begin the process of sparsening the neural representations of odors for efficient discrimination, memorization, and recognition.     

Olfactory neurons expressing identified receptor genes project to subsets of glomeruli within the antennal lobe of Drosophila melanogaster

We have used green fluorescent protein to trace the projection patterns of olfactory neurons expressing identified candidate odorant receptors to the brain of Drosophila. At the periphery, receptor expression correlates with specific sense-organ subtype, independent of location on the antennal surface. The majority of neurons expressing a given receptor converge onto one or two major glomeruli as described previously. However, we detected a few additional glomeruli, which are less intensely innervated and also tend to be somewhat variable. This means that functionally similar olfactory neurons connect to small subsets of glomeruli rather than to a single glomerulus as believed previously. This finding has important implications for our understanding of odor coding and the generation of olfactory behavior.     

Physiology and glomerular projections of olfactory receptor neurons on the antenna of female Heliothis virescens (Lepidoptera: Noctuidae) responsive to behaviorally relevant odors

The neurophysiology and antennal lobe projections of olfactory receptor neurons housed within short trichoid sensilla of female Heliothis virescens F. (Noctuidae: Lepidoptera) were investigated using a combination of cut-sensillum recording and cobalt-lysine staining techniques. Behaviorally relevant odorants, including intra- and inter-sexual pheromonal compounds, plant and floral volatiles were selected for testing sensillar responses. A total of 184 sensilla were categorized into 25 possible sensillar types based on odor responses and sensitivity. Sensilla exhibited both narrow (responding to few odors) and broad (responding to many odors) response spectra. Sixty-six percent of the sensilla identified were stimulated by conspecific odors; in particular, major components of the male H. virescens hairpencil pheromone (hexadecanyl acetate and octadecanyl acetate) and a minor component of the female sex pheromone, (Z)-9-tetradecenal. Following characterization of the responses, olfactory receptor neurons within individual sensilla were stained with cobalt lysine (N=39) and traced to individual glomeruli in the antennal lobe. Olfactory receptor neurons with specific responses to (Z)-9-tetradecenal, a female H. virescens sex pheromone component, projected to the female-specific central large female glomerulus (cLFG) and other glomeruli. Terminal arborizations from sensillar types containing olfactory receptor neurons sensitive to male hairpencil components and plant volatiles were also localized to distinct glomerular locations. This information provides insight into the representation of behaviorally relevant odorants in the female moth olfactory system. Electronic Supplementary Material Supplementary material is available for this article at and is accessible for authorized users.     

Rat olfactory bulb mitral cells receive sparse glomerular inputs

Center-surround receptive fields are a fundamental unit of brain organization. It has been proposed that olfactory bulb mitral cells exhibit this functional circuitry, with excitation from one glomerulus and inhibition from a broad field of glomeruli within reach of the lateral dendrites. We investigated this hypothesis using a combination of in vivo intrinsic imaging, single-unit recording, and a large panel of odors. Assuming a broad inhibitory field, a mitral cell would be influenced by >100 contiguous glomeruli and should respond to many odors. Instead, the observed response rate was an order of magnitude lower. A quantitative model indicates that mitral cell responses can be explained by just a handful of glomeruli. These glomeruli are spatially dispersed on the bulb and represent a broad range of odor sensitivities. We conclude that mitral cells do not have center-surround receptive fields. Instead, each mitral cell performs a specific computation combining a small and diverse set of glomerular inputs.     

Maturation of odor representation in the honeybee antennal lobe

The antennal lobe (AL) is the first center for processing odors in the insect brain, as is the olfactory bulb (OB) in vertebrates. Both the AL and the OB have a characteristic glomerular structure; odors sensed by olfactory receptor neurons are represented by patterns of glomerular activity. Little is known about when and how an odor begins to be perceived in a developing brain. We address this question by using calcium imaging to monitor odor-evoked neural activity in the ALs of bees of different ages. We find that odor-evoked neural activity already occurs in the ALs of bees as young as 1 or 2 days. In young bees, the responses to odors are relatively weak and restricted to a small number of glomeruli. However, different odors already evoke responses in different combinations of glomeruli. In mature bees, the responses are stronger and are evident in more glomeruli, but continue to have distinct odor-dependent signatures. Our findings indicate that the specific glomerular patterns for odors are conserved during the development, and that odor representations are fully developed in the AL during the first 2 weeks following emergence.     

Normal glomerular organization of the antennal lobes is not necessary for odor-modulated flight in female moths

A prominent hypothesis for the function of the glomerular structures in the primary olfactory neuropil of many groups of vertebrate and invertebrate animals is that they enable the processing and coding of information about the chemical compounds that compose complex odors. Previous studies have indicated that various degrees of glomerulus formation in the antennal lobes of the brain of the moth Manduca sexta can be effected by reducing the number of olfactory sensory axons that grow from the antenna into the antennal lobe during metamorphosis. To test the hypothesis that the presence of glomerular structure is necessary to process and identify odors, we substantially reduced, by surgery, the number of antennal segments in developing moths and upon metamorphosis we observed and quantified behavioral responses known to be elicited by odors. Intact and lesioned adult female moths were challenged to fly upwind to the source of an attractive host-plant odor in a wind tunnel. Some of the moths that had developed with reduced olfactory input flew upwind to the odor source. The flight behavior of these individuals was similar to the odor-mediated flight typically observed in moths that had developed normally. Histological analysis of the moths antennal lobes revealed that the lobes of more than half of the respondents that had been lesioned during development lacked normal glomerular organization. The neuropil of these abnormally developed antennal lobes was mostly aglomerular, but with a few isolated, clearly abnormal glomerulus-like structures. This suggests either that even a few abnormal glomeruli are sufficient to mediate this specific behavior or that canonical glomerular organization per se is not necessary for this odor-mediated behavior.     

Glomerular maps without cellular redundancy at successive levels of the Drosophila larval olfactory circuit

BACKGROUND: Drosophila larvae possess only 21 odorant-receptor neurons (ORNs), whereas adults have 1,300. Does this suggest that the larval olfactory system is built according to a different design than its adult counterpart, or is it just a miniature version thereof? RESULTS: By genetically labeling single neurons with FLP-out and MARCM techniques, we analyze the connectivity of the larval olfactory circuit. We show that each of the 21 ORNs is unique and projects to one of 21 morphologically identifiable antennal-lobe glomeruli. Each glomerulus seems to be innervated by a single projection neuron. Each projection neuron sends its axon to one or two of about 28 glomeruli in the mushroom-body calyx. We have discovered at least seven types of projection neurons that stereotypically link an identified antennal-lobe glomerulus with an identified calycal glomerulus and thus create an olfactory map in a higher brain center. CONCLUSIONS: The basic design of the larval olfactory system is similar to the adult one. However, ORNs and projection neurons lack cellular redundancy and do not exhibit any convergent or divergent connectivity; 21 ORNs confront essentially similar numbers of antennal-lobe glomeruli, projection neurons, and calycal glomeruli. Hence, we propose the Drosophila larva as an "elementary" olfactory model system.     

Excitatory interactions between olfactory processing channels in the Drosophila antennal lobe

Each odorant receptor gene defines a unique type of olfactory receptor neuron (ORN) and a corresponding type of second-order neuron. Because each odor can activate multiple ORN types, information must ultimately be integrated across these processing channels to form a unified percept. Here, we show that, in Drosophila, integration begins at the level of second-order projection neurons (PNs). We genetically silence all the ORNs that normally express a particular odorant receptor and find that PNs postsynaptic to the silent glomerulus receive substantial lateral excitatory input from other glomeruli. Genetically confining odor-evoked ORN input to just one glomerulus reveals that most PNs postsynaptic to other glomeruli receive indirect excitatory input from the single ORN type that is active. Lateral connections between identified glomeruli vary in strength, and this pattern of connections is stereotyped across flies. Thus, a dense network of lateral connections distributes odor-evoked excitation between channels in the first brain region of the olfactory processing stream.     

Olfactory maps and odor images

The understanding of neuronal processing of olfactory stimuli has been furthered by genetic studies and specialized imaging of particular neuronal populations. Selective optical imaging of odor-induced presynaptic and postsynaptic glomerular activity in the olfactory bulb/antennal lobe has visualized odorant-responsive receptor repertoires and shown a more confined odor image at the level of projection neurons compared to their olfactory receptor neuron input. Genetic tracing of projection neurons connected to particular glomeruli has revealed a somewhat dispersed spatial map of termination areas for these neurons both in insects and in vertebrates. Modifications of the glomerular odor map have resulted in altered percepts of the corresponding odors.     

The olfactory glomerulus: A cortical module with specific functions

The axons of many olfactory receptor cells converge on an individual glomerulus in the olfactory bulb, where they make contacts with the distal dendrites of mitral and tufted cells. Each glomerulus is targeted by olfactory receptor neurons expressing a single type of olfactory receptor protein. The glomerulus provides a unique model in which the function of a cortical module can be unambiguously established. Here we review the increasing evidence that a key functional operation of the glomerulus is to act as a signal-to-noise enhancing device in the processing of sensory input and that this function is critical across vertebrate and invertebrate species for the ability to detect specific odor stimuli within “noisy” odor environments and to carry out discriminations between odor molecules that are structurally closely related.     

昆虫气味认知的嗅觉神经结构及分子机制

大多数昆虫主要通过气味认知感知外界环境的变化,维持生命活动。探究昆虫气味认知的嗅觉系统神经结构及分子机制,对于完善气味认知神经生物学理论及利用其原理进行仿生学研究等有重要的科学意义。近年,关于昆虫气味认知科学研究有了很大的进展。本文从昆虫神经生物学的视角详细综述了近年关于昆虫气味认知的嗅觉神经结构、分子机制及气味信号的神经传导途径等方面的基本理论及最新研究成果。综述结果显示:昆虫对气味的认知是通过嗅觉神经系统的触角感器、触角叶(AL)、蕈形体(MB)等脑内多层信号处理神经结构来实现的。当外界气味分子进入触角感器内后,由感器内特定的气味识别蛋白(OBP)将气味分子运载到达嗅觉感受神经元(ORN)树突膜上的受体位点,气味分子与表达特定气味的受体(OR)结合产生电信号,并以动作电位的形式通过ORN的轴突传到脑内的触角叶。在触角叶经过嗅觉纤维球对气味信息选择性加工处理,再由投射神经元(PNs)将初步的识别和分类的气味信息传到蕈形体和外侧角(LH)等神经中枢,实现对气味的识别和认知。虽然,近年昆虫气味认知神经生物学的研究有了很大的进步,但是,我们认为目前的研究成果还不能完全阐明昆虫气味认知的神经机制,还有很多问题,例如,触角叶上众多的嗅觉纤维球是如何对嗅觉感受神经元传入的气味信息进行编码处理的?等有待进一步深入研究。为了搞清这些疑难问题,我们认为需要提高现有的实验技术水平,加强电生理学和分子神经生物学相结合的实验研究,从分子水平探究气味认知的神经机制可能是未来研究的热点。     

The molecular basis of odor coding in the Drosophila larva

We have analyzed the molecular basis of odor coding in the Drosophila larva. A subset of Or genes is found to be expressed in larval olfactory receptor neurons (ORNs). Using an in vivo expression system and electrophysiology, we demonstrate that these genes encode functional odor receptors and determine their response spectra with 27 odors. The receptors vary in their breadth of tuning, exhibit both excitation and inhibition, and show different onset and termination kinetics. An individual receptor appears to transmit signals via a single ORN to a single glomerulus in the larval antennal lobe. We provide a spatial map of odor information in the larval brain and find that ORNs with related functional specificity map to related spatial positions. The results show how one family of receptors underlies odor coding in two markedly different olfactory systems; they also provide a molecular mechanism to explain longstanding observations of larval odor discrimination.     

Integration of chemosensory pathways in the Drosophila second-order olfactory centers

BACKGROUND: Behavioral responses to odorants require neurons of the higher olfactory centers to integrate signals detected by different chemosensory neurons. Recent studies revealed stereotypic arborizations of second-order olfactory neurons from the primary olfactory center to the secondary centers, but how third-order neurons read this odor map remained unknown. RESULTS: Using the Drosophila brain as a model system, we analyzed the connectivity patterns between second-order and third-order olfactory neurons. We first isolated three common projection zones in the two secondary centers, the mushroom body (MB) and the lateral horn (LH). Each zone receives converged information via second-order neurons from particular subgroups of antennal-lobe glomeruli. In the MB, third-order neurons extend their dendrites across various combinations of these zones, and axons of this heterogeneous population of neurons converge in the output region of the MB. In contrast, arborizations of the third-order neurons in the LH are constrained within a zone. Moreover, different zones of the LH are linked with different brain areas and form preferential associations between distinct subsets of antennal-lobe glomeruli and higher brain regions. CONCLUSIONS: MB is known to be an indispensable site for olfactory learning and memory, whereas LH function is reported to be sufficient for mediating direct nonassociative responses to odors. The structural organization of second-order and third-order neurons suggests that MB is capable of integrating a wide range of odorant information across glomeruli, whereas relatively little integration between different subsets of the olfactory signal repertoire is likely to occur in the LH.     

Emerging principles of molecular signal processing by mitral/tufted cells in the olfactory bulb

The olfactory system shares many principles of functional organization with other sensory systems, but differs in that the sensory input is in the form of molecular information carried in odor molecules. Current studies are providing new insights into how this information is processed. In analogy with the spatial receptive fields of visual neurons, the molecular receptive range of olfactory cells is defined as the range of odor molecules that will affect the firing of that cell. Olfactory receptor molecules belong to a large gene family; it is hypothesized that individual receptor molecule may have relatively broad molecular receptive ranges, and that an individual receptor cell need therefore express only one or a few different types of receptors to cover a broad range. Mitral/tufted cells have narrower molecular receptive ranges, comprising molecules with related structures (odotopes). This is believed to reflect processing through the olfactory glomeruli, each glomerulus acting as a convergence center for related inputs. Varying overlapping specificities of receptor cells, glomeruli and mitral/tufted cells appear to provide the basis for discrimination of odor molecules, in analogy with discrimination of color in the visual systems.     

Friends and foes from an ant brain's point of view--neuronal correlates of colony odors in a social insect

Background

Successful cooperation depends on reliable identification of friends and foes. Social insects discriminate colony members (nestmates/friends) from foreign workers (non-nestmates/foes) by colony-specific, multi-component colony odors. Traditionally, complex processing in the brain has been regarded as crucial for colony recognition. Odor information is represented as spatial patterns of activity and processed in the primary olfactory neuropile, the antennal lobe (AL) of insects, which is analogous to the vertebrate olfactory bulb. Correlative evidence indicates that the spatial activity patterns reflect odor-quality, i.e., how an odor is perceived. For colony odors, alternatively, a sensory filter in the peripheral nervous system was suggested, causing specific anosmia to nestmate colony odors. Here, we investigate neuronal correlates of colony odors in the brain of a social insect to directly test whether they are anosmic to nestmate colony odors and whether spatial activity patterns in the AL can predict how odor qualities like “friend” and “foe” are attributed to colony odors.

Methodology/Principal Findings

Using ant dummies that mimic natural conditions, we presented colony odors and investigated their neuronal representation in the ant Camponotus floridanus. Nestmate and non-nestmate colony odors elicited neuronal activity: In the periphery, we recorded sensory responses of olfactory receptor neurons (electroantennography), and in the brain, we measured colony odor specific spatial activity patterns in the AL (calcium imaging). Surprisingly, upon repeated stimulation with the same colony odor, spatial activity patterns were variable, and as variable as activity patterns elicited by different colony odors.

Conclusions

Ants are not anosmic to nestmate colony odors. However, spatial activity patterns in the AL alone do not provide sufficient information for colony odor discrimination and this finding challenges the current notion of how odor quality is coded. Our result illustrates the enormous challenge for the nervous system to classify multi-component odors and indicates that other neuronal parameters, e.g., precise timing of neuronal activity, are likely necessary for attribution of odor quality to multi-component odors.     

Transmission of olfactory information between three populations of neurons in the antennal lobe of the fly

Three classes of neurons form synapses in the antennal lobe of Drosophila, the insect counterpart of the vertebrate olfactory bulb: olfactory receptor neurons, projection neurons, and inhibitory local interneurons. We have targeted a genetically encoded optical reporter of synaptic transmission to each of these classes of neurons and visualized population responses to natural odors. The activation of an odor-specific ensemble of olfactory receptor neurons leads to the activation of a symmetric ensemble of projection neurons across the glomerular synaptic relay. Virtually all excited glomeruli receive inhibitory input from local interneurons. The extent, odor specificity, and partly interglomerular origin of this input suggest that inhibitory circuits assemble combinatorially during odor presentations. These circuits may serve as dynamic templates that extract higher order features from afferent activity patterns.     

Strong single-fiber sensory inputs to olfactory cortex: implications for olfactory coding

Olfactory information is first encoded in a combinatorial fashion by olfactory bulb glomeruli, which individually represent distinct chemical features of odors. This information is then transmitted to piriform (olfactory) cortex, via axons of olfactory bulb mitral and tufted (M/T) cells, where it is presumed to form the odor percept. However, mechanisms governing the integration of sensory information in mammalian olfactory cortex are unclear. Here we show that single M/T cells can make powerful connections with cortical pyramidal cells, and coincident input from few M/T cells is sufficient to elicit spike output. These findings suggest that odor coding is broad and distributed in olfactory cortex.