Novel Neuronal Ensembles Encoding Analysis Method Based on Rough Set Theory

Relations hip between neuronal ensemble activities and limb movement is the key issue in brain-machine interface research. Several mathematical methods have been used to extract movement information when performing specific tasks. But how to simply and effectively encode the information underlying the neuronal ensemble activities is still a big challenge for researchers because the neural network is regarded as a non-linear time-varying system. Rough set theory is a new mathematical tool to deal with vagueness and uncertainty. It can analyze the facts hidden in the data without any additional knowledge about the data and has been applied in many domains, such as medicine, telecommunication, pattern recognition, control theory, etc. In this paper, the neuronal ensemble activities and forelimb movements were simultaneously recorded when rats were performing "extend-flex" tasks. An encoding analysis method based on rough set theory was established to reveal how the forelimb movements were represented by neuronal firing patterns. Firstly, the information about the neuronal ensemble activities and forelimb movements were stored in a table. Secondly, the redundant information was deleted by simplifying the table. Finally, the useful encoding information rule acquisition ability of rough set theory was mined out by using the rule acquisition ability of rough set theory. The acquired rules were understandable. In our experiments on the data sets of rats in "extend-flex" tasks, the performance of the method can be over 90%. Results show that the neuronal ensemble activities in primary motor cortex of the rats were closely related with the modes of forelimb movements. Experimental results also show that rough set theory can be used as a novel neuronal ensembles encoding analysis method.