| Cognition is closely related to human life.Previous studies have found that cognition is affected by many factors,resulting in individual difference in cognitive ability or even functional abnormalities,while cognitive impairment is a common symptom of depression.Electroencephalogram(EEG)is one of the commonly used neuroimaging techniques to investigate cognition,but it is still a problem to obtain stable biomarkers to measure individualized cognition and objectively reflect the cognitive impairment of depression.Based on EEG signals collected in the working memory(WM)task,this paper aims to find objective EEG indicators that are relevant to individualized WM ability and WM impairment in depression.It mainly includes:1.The spatiotemporal dynamic metrics concerning the individual difference of WM were explored in sensor level.By the approach of co-activation patterns(CAP),dynamic functional connectivity(d FC)networks(CAP1～7)in the theta frequency band seeded with Fz electrode were extracted from 145 healthy subjects.Scalp topology results showed stable frontal midline theta-band activation,and a splitting phenomenon of the frontal-parietal network.In addition,the results of Pearson correlation analysis showed that between-subjects variability in WM was associated with temporal dynamic parameters of specific activated states,especially the activation patterns of electrodes in midline and right temporal.In conclusion,the dynamic neural activity of theta frequency band under WM task can be used as an indicator to measure individual difference in WM capacity.2.Forty-one major depressive disorder(MDD)patients and forty-one age-,sex-and education-matched healthy controls(HE)were recruited in this experiment.At the EEG source level,the CAP analysis was performed for the bilateral dorsolateral prefrontal cortex(DLPFC),which were set as a priori brain region,to search for abnormal d FC metrics in theta frequency band closely related to WM impairment in MDD.The independent samples t-test results showed that compared with HE,MDD reduced the ability to detect target(P < 0.001)with prolonged response time(P < 0.01)when performing the WM task.Temporal ratio(TR)of CAP1 of the bilateral DLPFC,which represents the whole-brain interaction process,was significantly decreased(P < 0.01),while the transfer probability(TP)between different CAP states was significantly increased(P < 0.01).Besides,in Pearson correlation analysis,it was found that the TR of CAP1 in the left DLPFC was negatively correlated with the severity of depression(r =-0.436,P = 0.004),while in partial correlation analysis,which was controlled with depression severity,it was found that the TR of CAP1 in the right DLPFC was negatively correlated with the MDD’s WM response time(r =-0.475,P = 0.002).The machine learning model results of binary logistic regression analysis(AUC = 0.899,P < 0.001)and support vector machine(mean accuracy: 0.82,sensitivity: 0.81,specificity: 0.83)showed that integrating the above behavioral and dynamic network features can identify MDD effectively.Further analysis of the CAP spatial pattern information using the "overall dominant CAP-set" indicator revealed that the "overall dominant CAP-set" of the bilateral DLPFC of MDD contained more CAP states and spatial consistency among CAPs was reduced.Therefore,abnormal dynamic functional network activity in the theta band of the bilateral dorsolateral prefrontal cortex is one of the potential causes of WM impairment of MDD patients.3.Based on the changes of d FC metrics before and after single repetitive transcranial magnetic stimulation(r TMS),the prediction of individual effects of r TMS treatment was explored.Twelve MDD patients who received r TMS treatment were recruited.At the same time,in order to figure out whether a single r TMS stimulation had different effects on MDD patients and healthy controls(HE),twelve age-,sex-,education-matched HE who experienced a single r TMS stimulation were also included in the study.Based on the previous two parts of the work,the left DLPFC,which is also the stimulation target of r TMS,was setted as the seed point to extract the d FC network metrics in the theta-band co-activation patterns of MDD and HE before and after a single r TMS,and quantitated the difference before and after stimulation as dynamic activity change.The results of the paired t-test revealed that the WM response time of MDD patients was significantly shortened after r TMS treatment.The nonparametric test results showed that,compared with healthy controls,there was no difference in the behavioral data changes of MDD patients before and after a single r TMS stimulation.However,the change trend of transfer probability between different CAPs was opposite to that of healthy controls and there was significant difference(P < 0.05).Further Spearman correlation analysis showed that the change in transfer probability before and after a single r TMS stimulation in MDD was correlated with the decrease of WM response time after r TMS treatment(r = 0.636,P = 0.036).Thus,single r TMS had different aftereffect on the theta-band dynamic functional network activity in the left DLPFC of healthy and MDD patients,and the quantitative parameters of this aftereffect also showed potential in predicting the treatment of WM impairment in MDD by r TMS.In summary,our study explored the dynamic brain activity in the theta frequency band associated with individual difference of WM and WM impairment in depression.This dynamic activity was not only associated with widespread individual difference in WM ability,but was also a potential marker of neural activity for abnormal WM in MDD patients and could be used to predict remission of this abnormality in r TMS treatment. |
PDF Full Text Request