Modelling of monkey's motor cortical signals by an extended adaptive Liquid State Machine: an integrated procedure from model, identification, experiments, data fitting, to validation

In this paper, the LSM model is upgraded to enable it to the modelling of motor cortical signals, in which liquid states are no longer the spikes but the analogue potentials sampled from the neurons in the circuit and the readout layer is the standard multi-layer neural network with supervised learning algorithm. The input signals are spikes distilled from the monkey's cortex and output are the move directions of the trajectories of its right wrist. The results of the modelling process shows this LSM can be set up a good model with acceptable precision for a wide range of applications.