Research on recognition of disorders of consciousness based on Markov transfer probabilities
A method for exploring low-order Markov transfer probabilities based on short stationarity microstate sequences is proposed.Based on resting-state EEG signals from 18 patients with disorders of consciousness,the microstate sequences were segmented on a time scale with stationarity(60 s),and then the low-order Markov transfer probabilities of the state transition sequences were calculated to obtain statistically different transfer pattern between the minimally conscious and vegetative states,and the pattern was cross-validated as a feature,and a classification accuracy of up to 92%was obtained.The pattern of statistical differences in patients with disorders of consciousness tended to shift toward microstates C and D,and the first-order Markov has a better classification effect than the second-order one.Experimental results provide a new method for the study of microstates in the medical field.
microstateselectroencephalography(EEG)markovstationaritytransfer probabilitiesdisorders of consciousness
国家科技期刊平台
NSTL
万方数据
维普
