Tim C. Pearce

Handbook of Machine Olfaction

Odors are sensations that occur when compounds (called odorants) stimulate receptors located in the olfactory epithelium at the roof of the nasal cavity. Odorants are hydrophobic, volatile compounds with a molecular weight of less than 300 daltons. Humans can recognize and distinguish up to 10 000 different substances on the basis of their odor quality. Odorant receptors (ORs) in the nasal cavity detect and discriminate among these thousands of diverse chemical ligands. An individual odorant can bind to multiple receptor types, and structurally different odorants can bind to a single receptor. Specific patterns of activation generate signals that allow us to discriminate between the vast number of distinct smells. The physicochemical attributes of odorants that induce specific odor sensations are not well understood. The genes that code for ORs have been cloned, and results from cloning studies indicate that ORs are members of a superfamily of hundreds of different G-protein-coupled receptors that possess seven transmembrane domains. A complete knowledge of structureodor relationships in olfaction awaits the three-dimensional analysis of this large family of ORs. Ultimately, simultaneous knowledge of the three-dimensional structure of ORs as well as odorants will allow us to develop a pattern recognition paradigm that can predict odor quality.

Electronic nose for monitoring the flavour of beers

The flavour of a beer is determined mainly by its taste and smell, which is generated by about 700 key volatile and non-volatile compounds. Beer flavour is traditionally measured through the use of a combination of conventional analytical tools (e.g., gas chromatography) and organoleptic profiling panels. These methods are not only expensive and time-consuming but also inexact due to a lack of either sensitivity or quantitative information. In this paper an electronic instrument is described that has been designed to measure the odour of beers and supplement or even replace existing analytical methods. The instrument consists of an array of up to 12 conducting polymers, each of which has an electrical resistance that has partial sensitivity to the headspace of beer. The signals from the sensor array are then conditioned by suitable interface circuitry and processed using a chemometric or neural classifier. The results of the application of multivariate statistical techniques are given. The instrument, or electronic nose, is capable of discriminating between various commercial beers and, more significantly, between standard and artificially-tainted beers. An industrial version of this instrument is now undergoing trials in a brewery.

An artificial moth: Chemical source localization using a robot based neuronal model of moth optomotor anemotactic search

Robots have been used to model nature, while nature in turn can contribute to the real-world artifacts we construct. One particular domain of interest is chemical search where a number of efforts are underway to construct mobile chemical search and localization systems. We report on a project that aims at constructing such a system based on our understanding of the pheromone communication system of the moth. Based on an overview of the peripheral processing of chemical cues by the moth and its role in the organization of behavior we emphasize the multimodal aspects of chemical search, i.e. optomotor anemotactic chemical search. We present a model of this behavior that we test in combination with a novel thin metal oxide sensor and custom build mobile robots. We show that the sensor is able to detect the odor cue, ethanol, under varying flow conditions. Subsequently we show that the standard model of insect chemical search, consisting of a surge and cast phases, provides for robust search and localization performance. The same holds when it is augmented with an optomotor collision avoidance model based on the Lobula Giant Movement Detector (LGMD) neuron of the locust. We compare our results to others who have used the moth as inspiration for the construction of odor robots.

Robust stimulus encoding in olfactory processing: hyperacuity and efficient signal transmission

We investigate how efficient signal transmission and reconstruction can be achieved within the olfactory system. We consider a theoretical model of signal integration within the olfactory pathway that derives from its convergent architecture and results in increased sensitivity to chemical stimuli between the first and second stages of the system. This phenomenon of signal integration in the olfactory system is formalised as an instance of hyperacuity. By exploiting a large population of chemically sensitive microbeads, we demonstrate how such a signal integration technique can lead to real gains in sensitivity in machine olfaction. In a separate computational model of the early olfactory pathway that is driven by real-world chemosensor input, we investigate how spike-based signal and graded-potential signalling compares for supporting the accuracy of reconstruction of the chemical stimulus at later stages of neuronal processing.

Programmable Logic Construction Kits for Hyper-Real-Time Neuronal Modeling

Programmable logic designs are presented that achieve exact integration of leaky integrate-and-fire soma and dynamical synapse neuronal models and incorporate spike-time dependent plasticity and axonal delays. Highly accurate numerical performance has been achieved by modifying simpler forward-Euler-based circuitry requiring minimal circuit allocation, which, as we show, behaves equivalently to exact integration. These designs have been implemented and simulated at the behavioral and physical device levels, demonstrating close agreement with both numerical and analytical results. By exploiting finely grained parallelism and single clock cycle numerical iteration, these designs achieve simulation speeds at least five orders of magnitude faster than the nervous system, termed here hyper-real-time operation, when deployed on commercially available field-programmable gate array (FPGA) devices. Taken together, our designs form a programmable logic construction kit of commonly used neuronal model elements that supports the building of large and complex architectures of spiking neuron networks for real-time neuromorphic implementation, neurophysiological interfacing, or efficient parameter space investigations.

Combined smart chemFET/resistive sensor array

Here we describe a novel CMOS compatible gas sensor array based on a combined resistive/chemFET sensor cell. We have fabricated an array of 70 sensors with integrated drive, gain and baseline removal circuitry using an AMS 0.6 /spl mu/m CMOS process. The sensing materials are carbon black/polymer composite (CB) thin films, which have been previously reported to have good vapour-sensing properties. Different CB films have been deposited onto the sensor array and have been shown to respond differently to volatile organic compounds. This combined sensing element both reduces silicon area and, more importantly, measures different physical properties of the same gas sensitive material improving discrimination and giving more insight into the sensing mechanism.

Fisher information and optimal odor sensors

Abstract We discuss how the Fisher information matrix (FIM) may be used as part of an optimization procedure for selecting odor sensors within a population so as to maximize the accuracy with which the overall sensory system may estimate the stimulus. While the same approach may be equally applied to any sensory system that exploits a population coding of the stimulus in order to optimize its performance, we demonstrate how this technique may be used to analyze the performance of both biological and artificial olfactory systems. Thus one application of this method is the optimal design of arrays of artificial olfactory sensors.

Towards an artificial olfactory mucosa for improved odour classification

Here, we report on a biologically inspired analytical system that represents a new concept in the field of machine olfaction. Specifically, this paper describes the design and fabrication of a novel sensor system, based upon the principle of ‘nasal chromatography’, which emulates the human olfactory mucosa. Our approach exploits the physical positioning of a series of broadly tuned sensors (equivalent to the olfactory epithelium) along the length of a planar chromatographic channel (analogous to the thin mucus coating of the nasal cavity) from which we extract both spatial (response magnitude) and temporal (retentive delay) sensor signals. Our study demonstrates that this artificial mucosa is capable of generating both spatial and temporal signals which, when combined, create a novel spatio-temporal representation of an odour. We believe that such a system not only offers improved odour discrimination over a sensor array-based electronic nose, but also shorter analysis times than conventional gas chromatographic techniques.

Stimulus encoding during the early stages of olfactory processing: A modeling study using an artificial olfactory system

Abstract This paper addresses the issue of how efficient stimulus encoding may be carried out within the early stages of the olfactory system—in particular how a rate-coding scheme compares to the direct transmission of graded potentials in terms of the accuracy of the estimate that an ideal observer may make about the stimulus. We make use of a spiking neuronal model of the early stages of the olfactory system that is driven by fluorescent microbead chemosensors in order to compare these two coding schemes. Our results indicate how the charging time-constants present at the first stages of neuronal information processing within the olfactory bulb directly affects its ability to accurately reconstruct the stimulus.

Non-linear blend coding in the moth antennal lobe emerges from random glomerular networks

Neural responses to odor blends often exhibit non-linear interactions to blend components. The first olfactory processing center in insects, the antennal lobe (AL), exhibits a complex network connectivity. We attempt to determine if non-linear blend interactions can arise purely as a function of the AL network connectivity itself, without necessitating additional factors such as competitive ligand binding at the periphery or intrinsic cellular properties. To assess this, we compared blend interactions among responses from single neurons recorded intracellularly in the AL of the moth Manduca sexta with those generated using a population-based computational model constructed from the morphologically based connectivity pattern of projection neurons (PNs) and local interneurons (LNs) with randomized connection probabilities from which we excluded detailed intrinsic neuronal properties. The model accurately predicted most of the proportions of blend interaction types observed in the physiological data. Our simulations also indicate that input from LNs is important in establishing both the type of blend interaction and the nature of the neuronal response (excitation or inhibition) exhibited by AL neurons. For LNs, the only input that significantly impacted the blend interaction type was received from other LNs, while for PNs the input from olfactory sensory neurons and other PNs contributed agonistically with the LN input to shape the AL output. Our results demonstrate that non-linear blend interactions can be a natural consequence of AL connectivity, and highlight the importance of lateral inhibition as a key feature of blend coding to be addressed in future experimental and computational studies.