Prognostic Value Of Resting EEG Complexity In Patients With Disorders Of Consciousness

BackgroundSevere brain injury often leads to disorders of consciousness(DOC).These patients have a complex and critical condition,and the majority of them have dismal prognoses,which places a significant load on the family and society.Early,reliable,and sensitive prediction of outcomes can help clinicians and patients’ families make timely and appropriate decisions regarding treatment and care,while taking ethical,therapeutic,and financial issues into account.Currently,a variety of techniques are frequently employed to determine the prognosis of patients with disorders of consciousness,including bedside behavioral assessment,electroencephalogram,functional magnetic resonance imaging,positron emission tomography and others.However,due to their subjectivity and difficulty in acquiring at the bedside,these methodologies have considerable limitations in terms of practical application.Therefore,more accurate and sensitive prediction techniques are needed.Researchers have recently measured the complexity of brain signals using nonlinear dynamic factors like permutation entropy and found that the complexity of brain activity in patients with disorders of consciousness is significantly lower than that in healthy controls.The brain’s activity is more complicated and its functionality may be more comprehensive when the entropy value is higher.However,the current relevant research is still insufficient,and it is necessary to further explore the relationship between brain complexity changes and the prognosis of patients with disorders of consciousness,in order to provide a more accurate prediction method to assess the prognosis of patients with disorders of consciousness.ObjectiveThe resting-state electroencephalogram(EEG)of patients with disorders of consciousness was examined for changes in brain complexity using the permutation entropy(PE)method,and risk variables associated with their prognosis were identified using clinical traits and traditional EEG markers.Methods1.Retrospective enrollment was used to split the patients with coma,Unresponsive wakefulness syndrome(UWS),and minimally conscious state(MCS)who were admitted to the Xijing Hospital’s Neurology Intensive Care Unit between January 2020 and June 2022 into two groups.In the initial phase of the trial,62 individuals who were in comas were included.Twenty-six(59.1%)patients with VS/UWS and eighteen(40.9%)with MCS were among the 44 patients with prolonged disorders of consciousness(p Do C)who were included in Part II.2.Within 72 hours of admission,baseline demographic information,clinical traits,electrophysiological markers,and resting-state EEG recordings lasting at least an hour were gathered.PE was further quantified to evaluate the complexity of the brain.3.After onset,all patients were monitored for a total of six months and assessed on the Glasgow Outcome Scale(GOS).The prognosis of the patients was divided into two categories: GOS 1-2 represented a poor prognosis and GOS 3-5 represented a favorable prognosis.4.By using the t test,U test,or chi-square test,the baseline data and clinical traits of patients in various prognostic groups were compared.In order to further investigate the independent risk factors affecting patients’ prognoses,multivariate logistic regression analysis was utilized.To assess the utility of each indicator in predicting the prognosis of patients with disorders of consciousness,the ROC curve was created,and the sensitivity,specificity,and accuracy of each index in predicting the prognosis were computed.Results1.62 patients with coma were included in the study’s initial phase.42 patients(67.7%)were male,and the median age was 59[44,67] years.32 patients(51.6%)had a favorable prognosis during the follow-up period,while 30 patients(48.4%)had a poor prognosis.Age,duration of the disturbance of consciousness,etiology,Glasgow Coma Scale(GCS)score upon admission,and sleep spindles were not significantly different between the two groups,according to a univariate analysis.The patients with favorable prognosis had a greater percentage of male patients(p=0.011)and a higher Full Outline of Un Responsiveness(FOUR)score(p=0.028)in comparison to the patients with poor prognosis.Additionally,when EEG markers were compared,patients with favorable prognoses showed higher percentages of present EEG reactivity(p=0.011)and higher percentages of EEG activity.In coma patients,the Synek scale I-II was independently associated with a favorable prognosis(p=0.024,OR =3.766,95% CI: 1.195-11.868;multivariate logistic regression analysis.PE was used to compare the EEG complexity between the groups with favorable and poor prognosis,and PE maps for each frequency band were created for the two groups.The findings demonstrated that the PE values of the patients with favorable prognosis were higher in the alpha,beta,and gamma frequency bands as compared to the patients with poor prognosis.Univariate analysis revealed a significant difference in PE in the gamma band(p<0.001),and after controlling for gender and age,the difference remained statistically significant.The ROC curve showed that PE in gamma frequency band had a high predictive efficiency for the prognosis of coma patients,with AUC= 0.815,95% CI: 0.704-0.927,sensitivity 88%,specificity 70%,accuracy 79%,which were higher than those of Synek scale(AUC=0.674,95% CI: 0.514-0.834,sensitivity 70%,accuracy 79%).,sensitivity 78%,specificity 57%,accuracy 68%).In addition,the combination of Synek scale and gamma band PE can significantly improve the prediction performance,AUC=0.847,95%CI: 0.748-0.946.2.44 individuals with p Do C were enrolled in the second phase of the trial,including 26(59.1%)UWS patients and 18(40.9%)MCS patients.The median Coma Recovery Scale— Revised(CRS-R)score for the patients was5 [4,10] and their mean age was 44.61±13.26 years.Of the total patients,29(or 65.9%)were male.19 patients(43.2%)had a favorable prognosis during the follow-up period,while 25 patients(56.8%)had a poor prognosis.The results of univariate analysis showed that there was no significant difference between the two groups with different prognosis in gender,age,length of awareness disturbance,etiology,sleep spindles,or Synek scale.Patients with favorable prognosis had greater proportions of MCS(p=0.028)and higher CRS-R scores after admission(p=0.005)when compared to patients with poor prognosis.In addition,in comparison with EEG markers,patients with favorable prognosis had a higher proportion of present EEG reactivity(p=0.011).Multivariate logistic regression analysis showed that higher CRS-R score(p=0.005,OR=1.335,95% CI: 1.089-1.636)was an independent predictor of favorable prognosis in patients with prolonged disorders of consciousness.PE was used to compare the EEG complexity between the groups with favorable and poor prognoses,and PE maps for each frequency band were created for the two groups.The findings demonstrated that the PE values of the patients with favorable prognosis were higher in the alpha,beta,and gamma frequency bands as compared to the patients with poor prognosis.Univariate analysis revealed a significant difference in PE in the gamma band(p<0.001),and after controlling for gender and age,the difference remained statistically significant.ROC curve indicated that beta band PE had high prognostic efficacy in coma patients,AUC=0.801,95%CI: 0.665-0.939,sensitivity 68%,specificity 88%,accuracy 80%,higher than CRS-R score(AUC=0.739,95%CI: 0.589-0.889,sensitivity 68%,specificity 80%,accuracy 72%).In addition,the combination of CRS-R score and beta band PE significantly improved the prediction efficiency,with AUC=0.851,95%CI: 0.735-0.996.Conclusions1.According to this study,PE can be a useful indicator for assessing the prognosis of patients with impairments of consciousness.A higher PE indicates more complicated and unexpected brain activity and a higher possibility of a favorable prognosis.2.In patients with coma,PE in the gamma band and Synek scale may be employed as separate predictors of early prognosis,and that PE in the gamma band had superior predictive accuracy than Synek grade.In patients with prolonged disorders of consciousness,PE in beta band and CRS-R score may be employed as independent indicators of early prognosis,and the predictive efficiency of these two variables was also demonstrated.