No Self-organizing Neural Computing Theory And Application

As a new neuro-transmitter, Nitric Oxide (NO) plays a very important role in modulation of synaptic activity. It is the ability of NO to diffuse isotropically in aqueous and lipid environments that has suggested a radically different form of signaling in which the transmitter acts four-dimensionally in space and time, affecting volumes of the brain containing many neurons and synapses. In contrast, the traditional diffusion is localized in pre-synapses and post-synapses. So from 1980's on, the function of NO causes the extensive concern from international and domestic research institutions.In this paper a spatial and temporal dynamical model of the NO diffusion within the brain is described firstly. On theory, based on the latest research of the international neuro-physiologist and psychologist on NO, We detail several important results highlighting the role of source structure in the diffusion process, such as point source, hollow sphere source and irregular structure. Second, we have had a advanced research on Self-Organizing Feature Maps. And some qualities of SOFM have been analyzed. According to the application of such processes to artificial network used for robot control, called GasNets, the system has been introduced. Evidence is presented that such network is more amenable to the evolutionary computation approach. Finally we combine the NO diffusional model with the SOFM, and present novel NO-SOFM neural networks. According to the article, some discuss and simulation on Image(data) fusion based on SOFM.