Abstract：To solve the continuity lack of sample entropy algorithm and the complex with slow computing speed of improved entropy algorithm for fuzzy entropy, according to sample entropy and fuzzy entropy algorithm characteristics, the idea of a string variable was introduced instead of phase space reconstruction to establish a new fuzzy entropy algorithm. To analyze the event related potential embodied cognition characteristics, 35 college students were chosen as subjects, and the sample entropy, fuzzy entropy and the new fuzzy entropy algorithm were respectively used to calculate the complexity parameter. The results show that the calculation results of new improved algorithm are consistent with sample entropy and fuzzy entropy algorithm, and the new improved algorithm can be applied to EEG feature extraction. Compared with the sample entropy algorithm, the new improved algorithm can obtain bigger variations in entropy value of event related potential(ERP) to identify subjects′ midbrain dynamism of cognitive activities more effectively. The computation time of improved algorithm with 1 000 length EEG is only about 2.1 s, which improves computation speed greatly with comparison to those of sample entropy with 3.2 s and fuzzy entropy with 10.4 s.
2015, Vol. 36 
