NEUROPHYSIOLOGICAL CHARACTERISTICS OF BIPOLAR DISORDER BY RESTING EEG MICROSTATE ANALYSIS

Abstract Background Behind the difficulty in diagnosing bipolar disorder (BD) is the reality that it takes several years after the initial consultation before the diagnosis is confirmed. In this context, it is essential to establish biomarkers that contribute to the diagnosis of BD. In recent years, electroencephalography (EEG) microstates (MS) have garnered great attention as a useful tool to elucidate whole-brain neural network activity due to their high reproducibility. Therefore, EEG-MS is expected to be an aid in the diagnosis of psychiatric disorders, including BD. Specifically, MS-E is associated with the processing of internal and emotional information as well as cognitive functions. Aims & Objectives In this study, we aimed to identify EEG-MS biomarker of BD by comparison with that of healthy controls (HC). Method We included 36 individuals with BD (age: 47.7±15.7, %female: 38.9%) and 72 HC (age: 47.5±15.4, %female: 38.9%). Using a 64-channel high-density EEG system, we recorded resting-state EEG for 5 minutes with participants eyes closed. EEG-MS were classified into five states from A to E using the k- means clustering method. Then, three temporal indices were calculated for each MS: duration, occurrence, and coverage. Two-way repeated-measures analysis of variance (ANOVA) was conducted for the three temporal indices of each MS. MS classes and three indices served as within-subject factors while participant groups were treated as a between-subject factor. For MS indices for which the interaction was significant, an independent t-test was performed as a post-hoc analysis to test for differences between the BD and HC groups. Results The ANOVA showed significant interactions between the groups, MS classes, and temporal indices (F(1,8)=3.91, p<0.001). Subsequent ANOVAs indicated significant interactions across the two groups in duration (F(1,4)=2.412, p=0.049), occurrence (F(1,4)=3.861, p=0.004), and coverage (F(1,4)=2.924, p=0.021), as well as in MS-D (F(1,2)=3.304, p=0.039) and MS-E (F(1,2)=6.759, p=0.001). Finally, post-hoc independent t-tests for MS-D and MS-E revealed that in MS-D, the BD group showed longer values in duration (t(106)=3.967, p<0.001) and coverage (t(106)=2.904, p=0.004) compared to the HC group. In contrast, in MS-E, the BD group showed lower values in occurrence (t(106)=-2.600, p=0.011) and coverage (t(106)=-2.382, p=0.019) compared to the HC group. Discussion & Conclusion This study demonstrated that the BD group exhibited a higher occurrence and longer duration for MS-D compared to the HC group. Conversely, regarding MS-E, the BD group showed a lower occurrence and decreased coverage than the HC group. Based on the physiological significance of MS-E, there may be a potential decline in cognitive function in the BD group compared to the HC group.