Objective PSI algorithm is applied to EEG data of healthy and epileptic people to analyze the flow direction and intensity of information between lead，which intends to reference for clinical medical research epilepsy disease. Methods First of all,on the basis of phase-slope index theoretical method,causal relation of flow direction of information between different brain regions was studied. Then we took an appropriate threshold (2 was chosen in this paper) as an effective result. After that the paper analyzed the latest health and epileptic EEG data from the military general hospital and examined whether exist a significant difference between health data and epileptic EEG data. Results There was not any preferred direction in the right-left flow and there was a clear direction from front to back at the alpha frequency. Under the broadband method (all frequency),there was no statistical significance in the difference about the percentage of effective results between the data of the healthy and epileptic people. Under the narrow band method (alpha frequency),significant results percentage obtained from healthy data was higher than that from epileptic patient data. Conclusions The significant results percentage based on phase-slope index method can disclose statistical significance in the difference among the two signals,which has good reference for epilepsy data analysis of clinical medicine.