Imaging and coding in the olfactory system.

Functional imaging methods permit analysis of neuronal systems in which activity is broadly distributed in time and space. In the olfactory system the dimensions that describe odorant stimuli in "odorant space" are still poorly defined. One way of trying to characterize the attributes of this space is to examine the ways in which its dimensions are encoded by the neurons and circuits making up the system and to compare these responses with physical-chemical attributes of the stimuli and with the output behavior of the animal. For documenting distributed events as they occur, imaging methods are among the few tools available. We are still in the early stages of this analysis; however, a number of recent studies have contributed new information to our understanding of the odorant coding problem. This paper describes imaging results in the context of other data that have contributed to our understanding of how odors are encoded by the peripheral olfactory pathway.

Relationships between odor-elicited oscillations in the salamander olfactory epithelium and olfactory bulb.

Oscillations in neuronal population activity, or the synchronous neuronal spiking that underlies them, are thought to play a functional role in sensory processing in the CNS. In the olfactory system, stimulus-induced oscillations are observed both in central processing areas and in the peripheral receptor epithelium. To examine the relationship between these peripheral and central oscillations, we recorded local field potentials simultaneously from the olfactory epithelium and olfactory bulb in tiger salamanders (Ambystoma tigrinum). Stimulus-induced oscillations recorded at these two sites were matched in frequency and slowed concurrently over the time course of the response, suggesting that the oscillations share a common source or are modulated together. Both the power and duration of oscillations increased over a range of amyl acetate concentrations from 2.5 x 10(-2) to 1 x 10(-1) dilution of saturated vapor, but peak frequency was not affected. The frequency of the oscillation did vary with different odorant compounds in both olfactory epithelium and bulb (OE and OB): amyl acetate, ethyl fenchol and d-carvone elicited oscillations of significantly different frequencies, and there was no difference in OE and OB oscillation frequencies. No change in the power or frequency of OE oscillations was observed after sectioning the olfactory nerve, indicating that the OE oscillations have a peripheral source. Finally, application of 1.0 and 10 microM tetrodotoxin to the epithelium blocked OE oscillations in a dose-dependent and reversible manner, suggesting that peripheral olfactory oscillations are related to receptor neuron spiking.

Convergent, self-encoded bead sensor arrays in the design of an artificial nose.

We report a new approach to designing an artificial nose based on high-density optical arrays that directly incorporate a number of structural and operational features of the olfactory system. The arrays are comprised of thousands of microsphere (bead) sensors, each belonging to a discrete class, randomly dispersed across the face of an etched optical imaging fiber. Beads are recognized and classified after array assembly by their unique, "self-encoded" response pattern to a selected vapor pulse. The high degree of redundancy built into the array parallels that found in nature and affords new opportunities for chemical-sensor signal amplification. Since each bead is independently addressable through its own light channel, it is possible to combine responses from same-type beads randomly distributed throughout the array in a manner reminiscent of the sensory-neuron convergence observed in the mammalian olfactory system. Signal-to-noise improvements of approximately n1/2 have been achieved using this method.

Characterizing complex chemosensors: information-theoretic analysis of olfactory systems.

The mechanisms that underlie a wine lover's ability to identify a favorite vintage and a dog's ability to track the scent of a lost child are still deep mysteries. Our understanding of these olfactory phenomena is confounded by the difficulty encountered when attempting to identify the parameters that define odor stimuli, by the broad tuning and variability of neurons in the olfactory pathway,and by the distributed nature of olfactory encoding. These issues pertain to both biological systems and to newly developed 'artificial noses' that seek to mimic these natural processes. Information theory, which quantifies explicitly the extent to which the state of one system (for example, the universe of all odors) relates to the state of another (for example, the responses of an odor-sensing device),can serve as a basis for analysing both natural and engineered odor sensors. This analytical approach can be used to explore the problems of defining stimulus dimensions, assessing strategies of neuronal processing, and examining the properties of biological systems that emerge from interactions among their complex components. It can also serve to optimize the design of artificial olfactory devices for a variety of applications, which include process control, medical diagnostics and the detection of explosives.

An olfactory neuronal network for vapor recognition in an artificial nose.

Odorant sensitivity and discrimination in the olfactory system appear to involve extensive neural processing of the primary sensory inputs from the olfactory epithelium. To test formally the functional consequences of such processing, we implemented in an artificial chemosensing system a new analytical approach that is based directly on neural circuits of the vertebrate olfactory system. An array of fiber-optic chemosensors, constructed with response properties similar to those of olfactory sensory neurons, provide time-varying inputs to a computer simulation of the olfactory bulb (OB). The OB simulation produces spatiotemporal patterns of neuronal firing that vary with vapor type. These patterns are then recognized by a delay line neural network (DLNN). In the final output of these two processing steps, vapor identity is encoded by the spatial patterning of activity across units in the DLNN, and vapor intensity is encoded by response latency. The OB-DLNN combination thus separates identity and intensity information into two distinct codes carried by the same output units, enabling discrimination among organic vapors over a range of input signal intensities. In addition to providing a well-defined system for investigating olfactory information processing, this biologically based neuronal network performs better than standard feed-forward neural networks in discriminating vapors when small amounts of training data are used.

Current trends in 'artificial-nose' technology.

Basic principles derived from biological olfaction, such as combining semiselective sensor arrays with pattern recognition, have been used to develop instrumentation capable of broad-band chemical detection and quantification. Commercially available instruments are useful in areas including quality control in the food, beverage and fragrance industries, environmental monitoring, chemical-purity and -mixture analysis, and medical diagnostics. Ongoing research is aimed at the development of more-advanced instruments that are smaller, cheaper, faster and more stable and reliable. These second-generation instruments are likely to find an increasing number of applications, including the on-line monitoring of fermentation and other bioprocesses.

Odorant response properties of convergent olfactory receptor neurons.

Information about odorant stimuli is thought to be represented in spatial and temporal patterns of activity across neurons in the olfactory epithelium and the olfactory bulb (OB). Previous studies suggest that olfactory receptor neurons (ORNs) distributed in the nasal cavity project to localized regions in the glomerular layer of the OB. However, the functional significance of this convergence is not yet known, and in no studies have the odorant response properties of individual ORNs projecting to defined OB regions been measured directly. We have retrogradely labeled mouse ORNs connecting to different glomeruli in the dorsal OB and tested single cells for responses to odorants using fura-2 calcium imaging. ORNs that project to clusters of dorsomedial (DM) glomeruli exhibit different odorant response profiles from those that project to dorsolateral (DL) glomeruli. DL-projecting ORNs showed responses to compounds with widely different structures, including carvone, eugenol, cinnamaldehyde, and acetophenone. In contrast, DM-projecting neurons exhibited responses to a more structurally restricted set of compounds and responded preferentially to organic acids. These data demonstrate that ORN afferents segregate by odorant responsiveness and that the homogeneity of ORN and glomerular input varies with different OB regions. The data also demonstrate that a subpopulation of ORNs projecting to DM glomeruli is functionally similar.

Olfactory processing: a time and place for everything.

The behavioral effects of pharmacologically desynchronizing neuronal firing in the brain of the honeybee provide new evidence that the oscillatory synchronization of neuronal activity plays an important role in fine olfactory discrimination.

Rapid classical conditioning of odor response in a physiological model for olfactory research, the tiger salamander.

In recent years there have been a number of important advances in the understanding of cellular mechanisms related to olfactory function. Hypotheses regarding the complex relationships among odorant structure, physiological activity and behavioral outcome generated by these findings, however, remain largely untested due to a paucity of psychophysical data on stimulus discrimination in the same experimental species. Comparisons between behavioral and physiological responses are essential for elucidating the critical aspects of stimulus coding in sensory systems. We have developed a method for generating psychophysical data in one of the primary model species used in olfactory research, the tiger salamander, Ambystoma tigrinum. These psychophysical experiments are carried out under the same conditions as physiological experiments in our laboratory. Using classical conditioning, individual salamanders are trained over a period of 2-3 h to show skin potential responses to odor and not air. Failure to train using backward pairing demonstrates that the response is not due to sensitization or pseudoconditioning. The conditioned response is mediated by the olfactory pathway, as it is blocked by olfactory nerve section. We show that salamanders detect three odorants that are commonly used stimuli in physiological experiments (butyl alcohol, butyl acetate and amyl acetate), but cannot detect a fourth common experimental odorant, camphor. This method should be a powerful tool for studying olfactory information processing by providing data on discriminability of stimuli used in salamander physiological studies.

New structure, the "olfactory pit," in human olfactory mucosa.

A whole-mount immunocytochemical method was devised to study the olfactory receptor neurons on the surface of the human olfactory mucosal sheet. Antibodies to neuron-specific tubulin and/or microtubule-associated protein 5 and phosphorylated neurofilament protein were used. Specimens taken at autopsy from 56 patients ranging in age from 2 days to 92 years revealed a structure not previously described, an olfactory pit. Round or oval openings with a diameter of 50 to 500 microns were observed on the surface of the olfactory epithelium in the whole-mount specimen. The morphology, number, and distribution of these openings varied among the different individuals. A detailed analysis of these structures was carried out by rehydrating and sectioning the whole-mount specimens. The olfactory pit (OP) is a blind pouch lined with olfactory epithelium (OE), which appears as an invagination of OE into the connective tissue, with a depth varying between 150 and 200 microns. In some sections through an OP, a thick axon bundle emerging from the bottom of the pouch was visible. The extension and termination of this axon bundle in the central nervous system has not been explored. We have found OPs in monkey olfactory mucosa, but none in rodents. The function of the pit specialization is unclear, but it appears to be a feature of normal, young epithelium. The configuration of the blind pouch may prolong odorant association with the olfactory receptor neurons, or the OP may contain specialized neurons that have not yet been recognized by morphological, biochemical, or functional techniques.

Generating Sensor Diversity through Combinatorial Polymer Synthesis.

A new approach for rapid, simple generation of uniquely responding sensors for use in polymer-based sensor arrays has been developed. Polymerization reactions between different combinations of two starting materials have been found to lead to many new, unique sensors with responses not simply related to the proportion of the starting materials. This approach is demonstrated in two ways: (a) the use of discrete polymer sensing cones each comprised of a specific monomer combination and (b) the fabrication of a gradient sensor, containing all combinations between the starting and ending monomer concentrations. Gradient sensors were fabricated using two different binary monomer systems, with both systems showing regions of broadly diverse fluorescence responses to organic vapor pulses.

A chemical-detecting system based on a cross-reactive optical sensor array.

The vertebrate olfactory system has long been recognized for its extraordinary sensitivity and selectivity for odours. Chemical sensors have been developed recently that are based on analogous distributed sensing properties, but although an association between artificial devices and the olfactory system has been made explicit in some previous studies, none has incorporated comparable mechanisms into the mode of detection. Here we describe a multi-analyte fibre-optic sensor modelled directly on the olfactory system, in the sense that complex, time-dependent signals from an array of sensors provide a 'signature' of each analyte. In our system, polymer-immobilized dye molecules on the fibre tips give different fluorescent response patterns (including spectral shifts, intensity changes, spectral shape variations and temporal responses) on exposure to organic vapours, depending on the physical and chemical nature (for example, polarity, shape and size) of both the vapour and the polymer. We use video images of temporal responses of the multi-fibre tip as the input signals to train a neural network for vapour recognition. The system is able to identify individual vapours at different concentrations with great accuracy. 'Artificial noses' such as this should have wide potential application, most notably in environmental and medical monitoring.

Cell death in the midbrain of the murine mutation weaver.

The midbrain of the adult homozygous weaver (wv/wv) mouse is notable for a reduction in the numbers of dopamine-containing cells in the substantia nigra (A9) and the retrorubral nucleus (A8). We have determined that the reduction in tyrosine hydroxylase (TH)-positive neurons in the ventral midbrain of the weaver is attributable to the loss of neurons after postnatal day 7 (P7). Because the number and spatial distribution of TH-positive mesencephalic neurons in wv/wv, heterozygous weavers (+/wv), and wild-type mice are not significantly different on P7, we conclude that the early developmental steps of proliferation and migration have taken place normally in the mutant. Although numbers and distribution of cells are normal in the wv/wv on P7, the appearance of the TH-stained ventral midbrain is abnormal because of the paucity of TH-stained dendritic processes. The ventrally extending TH-positive dendrites are largely absent in the young wv/wv. The wv/wv also can be distinguished from both homozygous normal (+/+) and wv/wv littermates on P7 based on the appearance of dendrites that are more numerous than in the wv/wv but thin, disorganized, and sparse compared with +/+. Most cell death seems to take place in wv/wv before P21. However, at least one subset of dopamine-containing neurons disappears later. The zone of densely packed TH-positive neurons in the substantia nigra that is likely to be the origin of innervation to striosomes in the caudoputamen disappears between P21 and adulthood. Despite the early pathology evident in the mesencephalic dopamine-producing neurons of the +/wv, no evidence for cell death was observed there even in the oldest +/wv weavers studied.

Rapid analyte recognition in a device based on optical sensors and the olfactory system.

We report here the development of a new vapor sensing device that is designed as an array of optically based chemosensors providing input to a pattern recognition system incorporating artificial neural networks. Distributed sensors providing inputs to an integrative circuit is a principle derived from studies of the vertebrate olfactory system. In the present device, primary chemosensing input is provided by an array of fiber-optic sensors. The individual fiber sensors, which are broadly yet differentially responsive, were constructed by immobilizing molecules of the fluorescent indicator dye Nile Red in polymer matrices of varying polarity, hydrophobicity, pore size, elasticity, and swelling tendency, creating unique sensing regions that interact differently with vapor molecules. The fluorescent signals obtained from each fiber sensor in response to 2-s applications of different analyte vapors have unique temporal characteristics. Using signals from the fiber array as inputs, artificial neural networks were trained to identify both single analytes and binary mixtures, as well as relative concentrations. Networks trained with integrated response data from the array or with temporal data from a single fiber made numerous errors in analyte identification across concentrations. However, when trained with temporal information from the fiber array, networks using "name" or "characteristic" output codes performed well in identifying test analytes.

Olfactory marker protein mRNA is found in axons of olfactory receptor neurons.

The separation between the cell bodies of olfactory receptor neurons in the nasal cavity and their axon terminals in the olfactory bulb make them attractive for studying axonal transport. Although high molecular weight RNAs are generally believed to be excluded from axons of mature neurons, we demonstrate here that mRNA for olfactory marker protein (OMP), an abundant cytoplasmic protein selectively expressed in mature receptor cells, is present in rodent olfactory receptor axons. OMP RNA was detected by in situ hybridization at the light microscope level in axons and in terminals. By nuclease protection, the level of OMP RNA in the olfactory bulb was 5-10% of that in the olfactory epithelium where the cell bodies reside. In contrast to axonally transported vasopressin and oxytocin mRNAs, which are deficient in their 3' polyA tails, axonal OMP RNA fractionated as polyA+. OMP RNA was lost from axons and terminals after deafferentation, suggesting that OMP RNA was synthesized in receptor cell bodies in the epithelium and was transported into axons and terminals in the olfactory bulb. RNA for G(olf), a G-protein highly expressed in dendrites of mature olfactory receptor neurons, was not detected in the olfactory bulb. We hypothesize that the immature nature of the cytoskeleton and, specifically, the lack of tightly bundled microtubules allows transport of particular mRNAs in olfactory receptor axons.

Salamander olfactory bulb neuronal activity observed by video rate, voltage-sensitive dye imaging. III. Spatial and temporal properties of responses evoked by odorant stimulation.

1. Activity patterns across and within the laminae of the olfactory bulb were analyzed by imaging voltage-sensitive dye responses during odorant stimulation of all or part of the ventral olfactory mucosa. 2. The time course of the signals was generally characterized by a brief, small hyperpolarization, followed by a period of depolarization, and then a longer-lasting hyperpolarization similar to that seen with electric stimulation but with longer durations. 3. The activity was distributed nonhomogeneously across the bulbar laminae in the form of spatially segregated clusters having bandlike appearances. Clusters were observed with three monomolecular odorants, amyl acetate, ethyl-n-butyrate, and limonene, and with the complex odor of meal worms. Although response patterns to different odorants overlapped, they also showed differences in overall distribution. 4. Delivery of high odorant concentrations increased the size of the activated areas and accentuated the degree of response pattern overlap among different odorants. The general properties of the response patterns generated by each odorant were, however, similar at different odorant concentrations and in each of the animals tested. 5. The spatial and temporal distributions of the bulbar responses were somewhat similar regardless of whether the odorants were applied to local epithelial regions via punctate stimulation or to the entire mucosa. Certain regions did, however, have lower thresholds than others for eliciting bulbar activity in response to particular odorants. 6. Odorants applied to regions of the epithelium outside the areas of maximum sensitivity elicited odorant-related activity patterns with depolarizing and hyperpolarizing components similar to those seen with overall stimulation, but only if higher concentrations were used. Activation of distributed odorant sensitivities presumably gave rise to these patterns. 7. These data suggest that subsets of odorant receptor types are found in different areas of the olfactory epithelium, and demonstrate that there is widespread distribution across the epithelium of receptors sensitive to particular odorants. On the basis of the structure of these epithelial fields and the bulb response patterns that they relate to, these findings also provide evidence for complex spatial relationships between the olfactory epithelium and bulb. 8. The findings from this study suggest that representation of odor information in the salamander olfactory bulb does not occur by activation of a few selective bulbar regions, each related to a different odorant species. Instead, large regions of bulbar circuitry are involved in which molecular epitopes may be the unit of representation. Incorporation of new data presented here into a hypothesis of odor coding is discussed.

Salamander olfactory bulb neuronal activity observed by video rate, voltage-sensitive dye imaging. I. Characterization of the recording system.

1. In this paper we describe properties of a video imaging system used to acquire voltage-sensitive dye fluorescence signals from the salamander olfactory bulb. Sources of noise in these signals were evaluated in preparations stained with the potentiometric probe RH-414. These were compared with noise levels in signals obtained from a light-emitting diode array designed to stimulate the experimental conditions with light levels similar to those seen in the salamander bulb recordings. These experiments define a number of determinants of video image quality to standardize optical voltage measurements in the salamander olfactory bulb. 2. Images were acquired at video rates using a Newvicon camera in a standard upright microscope and digitized with an eight-bit video frame grabber. 3. Sources of noise related to camera sensitivity, stability of illumination, and mechanical vibration were characterized. Camera dark noise was less than the pixel variability due to photon noise at the camera faceplate. This pixel noise was the limiting factor for discriminating the spatial and temporal properties of the optical responses. 4. No significant noise was found to be related to image digitization, transmission, or readout by the eight-bit frame grabber. Mechanical vibration, light stability, and other sources of noise could be controlled in vitro. In this condition, voltage-sensitive dye signal noise was similar to that in stimulated experiments using the light-emitting diode array. Higher levels of noise were found in vivo; some of this was reduced by sychronizing frame acquisition to the heartbeat. 5. On the basis of photodiode and video measurements, voltage-sensitive dye responses in the salamander olfactory bulb typically fell between 0.75% and 2.5% fractional change of background fluorescence. By appropriately adjusting the video signals before analog-to-digital conversion, we could detect fractional changes of < 0.5%. 6. Both response averaging and low-bandpass spatial filtering improved the signal-to-noise ratios of the images. For small numbers of averaged runs, the best improvement was obtained by low-bandpass spatial filtering. 7. Acquisition of high-spatial resolution video images permitted the use of low-bandpass spatial filters to suppress pixel noise. The degree of spatial enhancement depended on the relationship between the size of the structures of interest, pixel density, and the properties of the convolution filter kernel. This method avoided exposure of the preparation to prolonged illumination and the necessity of applying the large numbers of repeated stimuli required for averaging.

Salamander olfactory bulb neuronal activity observed by video rate, voltage-sensitive dye imaging. II. Spatial and temporal properties of responses evoked by electric stimulation.

1. Video imaging of changes in voltage-sensitive dye (VSD) fluorescence was used to analyze spatial and temporal properties of activity patterns in the in vivo salamander olfactory bulb and primordium piriform cortex after electric stimulation. Distribution of activity among and within the neuronal layers was analyzed after orthodromic stimulation of the whole olfactory nerve (ON), isolated fascicles, or local epithelial sites, and after antidromic stimulation of the medial olfactory tract (OT). 2. Optical signals propagated through the bulbar layers with a sequence that correlates with electrophysiological responses. After orthodromic stimulation, VSD responses started in the glomerular layer, spread to the deeper laminae, and, after reaching the region of mitral/tufted somata, were observed as a brief burst of activity in the OT. Compound action potentials in the ON were associated with short-duration, rapidly depolarizing optical responses in the ON layer. Responses in glomerular layer and external plexiform layer (EPL) first showed in some recordings a brief, small-amplitude hyperpolarization, followed by a period of depolarization, followed by a second, longer-lasting hyperpolarization. The periods of optical hyperpolarization could be related to events observed in intracellular mitral/tufted cell recordings. 3. With shocks delivered to the entire ON, depolarizing responses were nonhomogeneously distributed, appearing as multiple foci or bands of activity. Spatial patterns within each bulbar layer had poorly defined borders. Sites showing short-latency responses were often those with the largest and longest-lasting activity. 4. Increasing the intensity of stimulation to the ON enhanced the size and duration of the depolarizing and hyperpolarizing responses. The short-latency, early hyperpolarization was best seen with low-intensity, peripherally placed stimuli. 5. ON stimulation also elicited activity in the contralateral bulb. Activity started at the innermost layers and spread in patches to regions of the EPL just beneath the glomeruli. These had durations similar to ipsilateral responses, but longer latencies. A period of early hyperpolarization, longer than that on the ipsilateral side, was followed by prolonged depolarization and then by a second, later hyperpolarization. 6. Antidromic stimuli applied to the OT evoked optical responses consisting of a period of depolarization followed by hyperpolarization, similar to the components elicited by orthodromic stimuli. These responses had short time courses, began in the deeper layers, and spread to the superficial region of the bulb usually without reaching the glomerular region. 7. Punctate stimulation of the mucosa or nerve elicited depolarizing and hyperpolarizing events that depended on the stimulation site.(ABSTRACT TRUNCATED AT 400 WORDS)

Lineage analysis of the olfactory epithelium using a replication-incompetent retrovirus.

We have analysed the lineage of olfactory receptor neurons using a replication-incompetent retrovirus injected beneath the olfactory epithelium of young rats. There are two major types of clusters of infected cells seen at 5--40 days after infection: (i) horizontal basal cells (HBCs); (ii) variable numbers of globose basal cells (GBCs), and immature and mature sensory neurons. Olfactory nerve lesion increased the frequency of the globose/sensory neuron clusters, as well as the number of cells/cluster, but did not change the number of HBC clusters or cells/cluster. No clusters contained sustentacular cells. These data indicate that, at least in young rats: (i) HBCs are not precursors of olfactory neurons; (ii) there is a lineage path from GBCs to mature neurons; and (iii) sustentacular cells arise from a separate lineage.