| Objective Mild cognitive impairment(MCI)is a cognitive impairment syndrome between normal ageing and Alzheimer's disease(AD).Working memory dysfunction is the main clinical manifestation of MCI.It is important to investigate working memory dysfunction of MCI.At present,research on brain network is the hotspot in the field of neuroscience.From the view of brain network,the aim of this paper is to investigate the effective connectivity pattern among electroencephalographs(EEGs)data by Dynamic Causal Modeling(DCM)analysis.EEGs were recorded from MCI patients,while they performed a working memory task.This may provide a new perspective for exploring the working memory dysfunction mechanism of MCI.Methods 1.Experimental data: Behavioral data and 34-channel EEGs were recorded from 15 MCI patients,15 normal elder(Elderly group)and 15 normal youth(Youth group)while they performed a visual working memory task.2.Behavioral data analysis: Calculate the average reaction time and accuracy of MCI group,Elderly group and Youth group under working memory task.3.Data preprocessing: 50Hz-frequency interference,baseline drift,electrooculogram interference and electromyogram interference were removed from the recorded original EEGs.4.Time-frequency analysis: EEG spectrum distribution of the preprocessed EEGs was calculated by short-time Fourier transform(STFT)to analyze the principal frequency band and principal period of working memory.5.Source analysis: Distributed linear model method was applied to determine the activation brain areas of MCI group,Elderly group and Youth group working memory EEGs in the principal frequency bands and principal periods.6.DCM analysis: According to the activation brain areas of working memory which determined by the source analysis,the DCM network was constructed.Exceedance Probability(EP)of DCM network connections of the three groups were calculated to analysis the effective connectivity of working memory.Exceedance Probability in different direction connection families(forward,backward and lateral)and the left-right connection families in the DCM network of the three groups were calculated,and the changes of working memory information and the deviation of the brain in MCI group were compared and analyzed.7.Statistical analysis: One-way analysis of variance(ANOVA)method was used to test the experimental data in SPSS 22.0 statistical analysis software.Results 1.Behavioral resultsMean reaction time of MCI,Elderly and Youth group were 1387.8±80.0ms,1262.1±69.9ms,912.7±66.4ms,respectively.The reaction time of MCI group was significantly higher than that of Elderly group(P <0.01)and Youth group(P <0.001).Mean accuracy of MCI,Elderly and Youth group were 73.38±1.10%,82.58±1.42%,92.53±1.49%,respectively.The accuracy of MCI group was significantly lower than that of Elderly group(P <0.01)and Youth group(P <0.001).2.EEG spectrum of working memory EEGsThe EEG spectrum of working memory EEGs in three groups was mainly concentrated in theta band(4-8 Hz).Compared with Elderly and Youth groups,the spectrum of theta band was lower in MCI group.The average time-frequency representations showed that the spectrum was concentrated in the delay period of working memory for all groups.3.Activated regions of working memoryThe activated regions of MCI group were bilateral middleline-frontal cortex(MFC),the activated regions of Elderly group were bilateral dorsolateral prefrontal cortex(DLPFC)and bilateral MFC,and the activated regions of Youth group were bilateral DLPFC,bilateral MFC and bilateral parietal cortex(PAR).The activated regions of Youth group were analyzed as network nodes for DCM analysis.4.DCM analysis results(1)Exceedance probability of DCM network sub-connectionsThe exceedance probability from bilateral PAR to ipsilateral DLPFC(M7,M13)and MFC(M9,M15)were bigger than 5% in MCI group,and the exceedance probability from left PAR to left MFC(M9)was the biggest,while the exceedance probability of other connections were close to zero.The exceedance probability from bilateral PAR to ipsilateral DLPFC(M7,M13),bilateral PAR to ipsilateral MFC(M9,M15),left DLPFC to left MFC(M3),left DLPFC to left PAR(M8),left MFC to left PAR(M10)and right PAR to left PAR(M12)were bigger than 5% in Elderly group,and the exceedance probability from right PAR to right DLPFC(M13)was biggest,while the exceedance probability of other connections were close to zero.The exceedance probability from bilateral PAR to ipsilateral DLPFC(M7,M13)and MFC(M9,M15)were bigger than 5% in Youth group,and the exceedance probability from left PAR to left MFC(M9)was biggest,while the exceedance probability of other connections were close to zero.The total Exceedance probability of effective connections of MCI group was significantly lower than that of Elderly group(P <0.05)and Youth group(P <0.01).(2)Exceedance probability in different direction familiesThe exceedance probability of forward family(EP_forward)was higher than backward family(EP_backward)and lateral family(EP_lateral)in all groups(P<0.01).Compared with Elderly group(0.4626±0.0492),the EP_forward of MCI group(0.6113±0.0193)was higher significantly(P<0.01).The EP_backward of MCI group(0.1942±0.0085)was lower than Elderly group(0.2926±0.0440)(P<0.01).There was no difference between MCI and Elderly group in EP_lateral.The EP_forward,EP_backward and EP_lateral were not different between MCI and Youth group.(3)Exceedance probability of left and right familyThere was no significant difference between the exceedance probability of left family(EP_left)and right family(EP_right)both in MCI group and Elderly group.EP_left was higher than EP_right in Youth groups(P<0.001).There was no significant difference between EP_left of MCI group(0.5671 ± 0.0751)and EP_left of Elderly group(0.5259 ± 0.0592);EP_left both in MCI group and Elderly group was less significantly than Youth group.Conclusion 1.Behavioral results show that the mean reaction time of working memory in MCI group was significantly increased,and the accuracy was significantly decreased.According to the EEG spectrum,theta band is determined as the feature frequency band of working memory,and the feature period of working memory is the delay stage.The activity of working memory EEGs in theta band of MCI group was decreased.2.The results of source analysis showed that the activated brain regions of working memory in MCI group were MFC,and the activated brain regions of the Elderly group were DLPFC and MFC.The activated brain regions of the Youth group were DLPFC,MFC and PAR.The activated brain regions of working memory of MCI group decreased than that of Elderly group and Youth group.3.The results of the DCM analysis show that the contribution of effective connections in the network of MCI group less than that of Elderly group and Youth group.4.The mainly information flow of working memory in delay stage of MCI group and Youth group was from parietal cortex to prefrontal cortex,and the mainly information flow of working memory in delay stage of Elderly group was from parietal cortex to prefrontal cortex,as well as from frontal cortex to parietal cortex,suggesting that normal elderly without working memory dysfunction may be related to the increasing of information flow of working memory in elderly,but MCI patients did not form this compensation mechanism. |
PDF Full Text Request