Performance of odorants inference depending on functional and spatial neuron grouping

초록

This paper presents performance of olfactory neural decoding depending on functional and spatial neural selection. Multi-channel extra-cellular single-unit recording were done by micro-wire electrodes implanted in the mitra/tufted cell layers of the main olfactory bulb (MOB) of anesthetized rates to obtain neural responses to various odors. All neurons are classified according to significant differences response to using t-test (P<0.01). Odor can be robustly inferred by using subpopulations of neurons. The results indicate that the performance of odor inference is highly dependent on the neural selection. Also, we compared different statistical methods correlations of olfactory decoding accuracy. The results show that t-test is better criterion of assessment of neuron's capacity than others. © 2011 IEEE.