Investigation Of Odor-evoked Responses In Rat’s Olfactory Bulb In Vivo And Research On Bioelectronic Nose

Olfactory system has exquisite capability to discriminate odors, especially in speed and accuracy of odor recognition, which is much better than current electronic nose based on chemical sensor array. However, there are still many challenges remained in applying biological olfaction to industrial odor detection. Odor molecules interact with olfactory receptor neurons (ORNs) in nasal nostrils. During the process, chemical signals are changed into electric signals, and transmitted to the relay station of odor information processing called olfactory bulb (OB). After coding process of odor information in OB, the olfactory cortex receives the signal from OB and perceives odors. OB is considered as primary cortex of olfaction, which plays an important role in odor perception. Understanding the mechanism of odor perception in OB is worthwhile for both odor information processing and industrial odor detection.Currently, patch clamp recording technique and optical imaging technique are primary measurement methods for odor-evoked response recording of OB cells. Constrained by recording few cells synchronously, patch clamp technique cannot provide coherent activities of large numbers of neurons. While optical imaging technique has barriers in image resolution due to background noise. Therefore, with the advantage of long-term, synchronous multi-site measurement, implanted microelectrode arrays have been used for electrophysiological recording of OB. This technique benefits for understanding of cell information processing, furthermore, helps for accomplishing design of bio-enose. Based on the electrophysiological recordings, the paper established and simulated multiple models of olfactory cells and network including ORN, mitral cell, and olfactory neuronal network. In addition, the paper described the platform built for odor stimulation and odor-evoked response recording of OB cells in vivo based on implanted sensors. Using this platform, We investigated the temporal firing patterns of mitral cell ensemble. Combining biological method with engineering method, we established bio-enose system for odor recognition.The major contents and contributions of this thesis are as follows:1. Modeling and simulating neurons and neuron network of olfactory system. Based on electrophysiological data of voltage-gated channels, we established ORN model and simulated the action potential evoked by current pulse stimulation. In addition, we simulated the model of mitral cell and neuron network of olfactory system, and optimized parameters in models according to experimental data.2. Establishing research platform for study on odor-evoked responses of anesthetized free-breathing rats. The platform included odor delivery module, respiratory signal recording module, electrophysiological signal recording module, electrode fabrication module and data analysis module. Procedures of surgery and electrode insertion were illustrated in detail. The recording site of electrode array was examined by post-mortem histological reconstruction of the electrode track in OB slice.3. Uncovering rapid odor perception of anesthetized free-breathing rat in OB under short-term odor stimulation and context-based odor discrimination by mitral cell group. Utilizing matrix paradigm of cell group firing, we drew response curve of mitral cell ensemble in feature space which indicates that odor perception was achieved within the first breathing cycle after short-term odor stimulation. Features of odor-evoked response were extracted from mitral cells which were synchronously recorded with similar firing rate histogram. The result showed that five odors can be classified by principal component analysis.4. Designing bioelectronic nose system based on field potential of mitral cell layer in OB. Field potential is stable and easy to acquire. Besides, power spectrum distributions have difference under different odor stimuli. We applied K-nearest neighboring classification method for spectrograms of four odors stimuli by multitaper spectrum estimation. The classification accuracy of four odors reached77.4%in gamma band (40-120Hz).