Reconstruction of the neural network model of motor control for virtual C.elegans on the basis of actual organism information

Introduction: C. elegans neural network is a good sample for neural networks studies, because its structural details are completely determined. In this study, the virtual neural network of this worm that was proposed by Suzuki et al. for control of movement was reconstructed by adding newly discovered synapses for each of these network neurons. These synapses are newly discovered in the actual organism by laser ablation technique. Methods: An input weight of each neuron in this network was obtained from wormweb site. Internal state of each neuron and the whole network in this new situation were calculated after mathematical modeling. Then output of the network in this new model was simulated and drawn by Simulink software in MATLAB environment and compared with Suzuki model outputs. Results: The results show that in spite of adding new synapses to the last model, its output shape does not change and preserve its sinusoidal form, and only its frequency, phase and amplitude of internal state of whole network are changed after modification. Also, adding of these new synapses to models change silent neurons in forward motion in Suzuki model to sinusoidal internal state producing neurons in this new situation. Conclusion: These findings not only show that this neural network has high capability for production of more output patterns than previously supposed, but also illustrates the effect of addition of new synapses per each neuron and rewiring on the whole network outputs as a simulation of situation that possibly happens after neural plasticity.