Analog Circuit Implementation Research Of The Neural Network For Touch Sensitivity In Caenorhabditis Elegans

C.elegans as an important model organism has been widely applied to genetics, cell biology, neuroscience and other fields of study. In the research of C.elegans, the neural network of the c.elegans is an important research direction. Because of simple structure and rich functional behavioral, the neural network of the c.elegans has been an important platform for research of machine intelligence.The research of the neural network of the c.elegans mainly divided into theoretical research, applied research and implementation technology research and so on. At present, the realization of the neural network of the c.elegans is mainly through software simulation. Compared with software implementation, hardware implementation of neural network can bring the advantages of neural network such as parallel information processing and distributed storage to the fullest. It has more realistic significance for the practical study of neural network. Therefore, this paper tries to achieve the neural network for touch sensitivity in c.elegans and simulate the withdrawal functional behavior of c.elegans based on a hardware implementation method—analog circuit. Our main work as followings:(1) Introduced the theory and research status of the neural network of the c.elegans; Studied the implementation technique of the neural network detailed; Understanded the constitution of the neural network model; Mainly introduced the neural network for touch sensitivity in c.elegans and it’s common network model.(2) Implemented the circuit of the neural network for touch sensitivity in c.elegans. At first, we respectively implemented the synaptic circuits, sigmoid function circuit which own to neuron in a modular way and used circuit simulator software HSPICE to verified the validity of the circuit. Then, implemented the circuit of the neural network for touch sensitivity in c.elegans by the circuit module.(3) Simulated the withdrawal functional behavior of c.elegans based on the circuit of the neural network for touch sensitivity in c.elegans that implemented by hardware. The validity of the circuit has verified at the same time. In order to obtain optimization parameters of the circuit, a multi-strategy particle swarm optimization algorithm based on chaos mutation(CMMSPSO) was proposed, which was used to optimize the neural network for touch sensitivity in c.elegans. The simulation result based on HSPICE verified the validity of the circuit and showed that the circuit of the neural network for touch sensitivity in c.elegans implemented by hardware could realized the withdrawal functional behavior of c.elegans.