| Objective: To investigate the correlation between electroencephalogram (EEG) and cognitive rating scale in patients with cognitive impairment after stroke, and to analyze its clinical significance. Methods: A total of 75 stroke patients were selected and screened using the simple mental state Examination Scale (MMSE). They were randomly divided into cognitive impairment group (PSCI group) and non-cognitive impairment group (non-PSCI group), with 30 cases in each group and 30 healthy subjects served as control group. EEG, Montreal Cognitive Scale (MoCA) and Loewenston Occupational Therapy Cognition Rating Scale (LOTCA) were performed in all three groups. Spearman level correlation was used to analyze the correlation between EEG characteristics and MoCA and LOTCA, and the results of the three groups were compared. Results: The scores of MMSE, MoCA and LOTCA cognitive scale in PSCI group were significantly lower than those in non-PSCI group and healthy control group (P<0.05), while those in non-PSCI group were significantly lower than those in healthy control group (P<0.05). Two EEG indices of αAP and αRP were lower in PSCI group than in non-PSCI group and healthy control group (P<0.05), and lower in non-PSCI group than in healthy control group (P<0.05). DTABR index in healthy control group was significantly lower than in non-PSCI group and PSCI group (P<0.05), and that in non-PSCI group was significantly lower than in PSCI group (P<0.05). αAP was positively correlated with MoCA and LOTCA (r=0.734, r=0.922, P<0.05). αRP was positively correlated with MoCA and LOTCA (r=0.575, r=0.630, P<0.05). DTABR was negatively correlated with MoCA and LOTCA (r=-0.569, r=-0.614, P<0.05). Conclusion: EEG data based on frequency domain power spectrum analysis can be used as an assessment method of cognitive function after stroke, and they are correlated with the cognitive assessment scale. The combination of EEG data and the cognitive assessment scale can evaluate the existence of cognitive impairment more objectively and comprehensively. |
