Neuromorphic Processing for Optical Microbead Arrays: Dimensionality Reduction and Contrast Enhancement

Abstract

Abstract-This paper presents a neuromorphic approach for sensor-based machine olfaction that combines a portable chemical detection system based on microbead array technology with a biologically inspired model of signal processing in the olfactory bulb. The sensor array contains hundreds of microbeads coated with solvatochromic dyes adsorbed in, or covalently attached on, the matrix of various microspheres. When exposed to odors, each bead sensor responds with corresponding intensity changes, spectral shifts, and time-dependent variations associated with the fluorescent sensors. The bead array responses are subsequently processed using a model of olfactory circuits that capture the following two functions: chemotopic convergence of receptor neurons and center on-off surround lateral interactions. The first circuit performs dimensionality reduction, transforming the high-dimensional microbead array response into an organized spatial pattern (i.e., an odor image). The second circuit enhances the contrast of these spatial patterns, improving the separability of odors. The model is validated on an experimental dataset containing the responses of a large array of microbead sensors to five different analytes. Our results indicate that the model is able to significantly improve the separability between odor patterns, compared to that available from the raw sensor response.

Principles of odor coding in vertebrates and artificial chemosensory systems

“…The sum of signals from sensor replicas would correspond, in zeroth approximation, to the summing of receptor neuron inputs to glomeruli, although the processing of input signals in glomeruli is more sophisticated than summing or averaging. Data obtained from optical sensors based on functionalized fiber-optic tips (177,1148) were used to assemble a sort of OB modeled by a self-organizing map neural network (1149). In contrast to other neural network models, the self-organizing map creates a sort of spatially organized representation of input signals allowing for an ordered arrangement of input signals (1150).…”

“…Self-organizing maps are here used to cluster sensors rather than samples. The trained selforganizing map becomes equivalent to a glomerular map of the OB, where each node of the network receives signals from sensors that have similar responses since they are supposed to be based on the same indicator (1149).…”

Mimicking Biological Design and Computing Principles in Artificial Olfaction

“…Raman and co-workers investigated spiking as well as firing-rate models ,, of the olfactory bulb circuits for odor identity and intensity encoding. Input data from a pair of temperature-modulated metal-oxide sensors were first reformatted using a self-organizing model of chemotopic convergence of receptor neurons onto a lattice of olfactory bulb neurons to create ordered spatial maps that decoupled odor identity and intensity.…”

Towards biological plausibility of electronic noses: A spiking neural network based approach for tea odour classification

“…Recent efforts towards mimicking biological neuron type computation and information transfer has led to the development of neural networks such as the spiking neural network (SNNs) [17,18]. Researchers have used these SNNs to successfully develop bio-mimicking pattern recognition schemes for the ENs [6,9,[19][20][21][22][23][24][25][26][27][28][29]. Along these lines, we use the SNN for odour discrimination over the temporal spike codes.…”