Abnormal Spontaneous Brain Activity and Functional Connectivity of the Default Mode Network in Patients with Bipolar Disorder

Background: No consistent evidence on the specic brain regions is available in the default mode network (DMN), which show abnormal spontaneous activity in bipolar disorder (BD). We aim to identify this region that is particularly impaired in patients with BD by using several different indices measuring spontaneous brain activity and then investigate its functional connectivity (FC). Methods: A total of 56 patients with BD and 71 healthy controls (HC) underwent resting-state functional magnetic resonance imaging. Three commonly used functional indices were used to identify the brain region showing abnormal spontaneous brain activity in BD. Then, this region served as the seed region for resting-state FC analysis to identify its functional networks altered in BD. Results: The BD group exhibited decreased fALFF, ReHo, and DC values in the left precuneus. The BD group had decreased rsFC within the DMN, indicated by decreased resting-state FC within the left precuneus and between the left precuneus and the medial prefrontal cortex. The BD group had decreased negative connectivity between the left precuneus and the left putamen, extending to the left insula. Conclusions: The ndings provide convergent evidence for the abnormalities in the DMN of BD, particularly located in the left precuneus. Decreased FC within the DMN and the disruptive anticorrelation between the DMN and the salience network are found in BD. These ndings suggest that the DMN is a key aspect for understanding the neural basis of BD, and the altered functional patterns of DMN may be a potential candidate biomarker of BD.

depression showed decreased rsFC of the PCC with the MPFC and the precuneus 16 . Another study found that patients with unmedicated BD during depressive episode showed weaker rsFC in the left MPFC and right precuneus compared with healthy controls 19 . In addition, decreased rsFC between the DMN and the SN was reported in patients with BD, but the sample was composed of patients at different stages of BD (including manic, depressive, mixed, and euthymic phases) 20 . In these previous studies, the researchers often applied the method of independent component analysis (ICA) to identify the DMN as a whole and then analyze the rsFC of the DMN 14,[17][18][19] , or select speci c brain regions as seeds for the rsFC of the DMN on the basis of prior hypothesis 12,16,20,21 . Although promising ndings have been obtained in these previous studies, no consistent evidence is found on the speci c brain regions in the DMN, thereby showing abnormal spontaneous activity in patients with BD.
In addition to the two methods, other analytical approaches can be applied to the studies of DMN, such as fractional amplitude of low-frequency uctuation (fALFF) 22 , regional homogeneity (ReHo) 23 , and degree centrality (DC) 24 . The fALFF re ects the contribution of low-frequency signal strength in the entire frequency range; the ReHo re ects the synchronization of local neural activity, and the DC re ects the importance of a certain voxel in the whole-brain functional connectivity matrix. With a whole-brain search strategy, these indices can identify the abnormality in the brain. Using these indices, the researcher have already found decreased brain activity of DMN in patients with BD [25][26][27] . Therefore, the abnormal spontaneous brain activity in the DMN can be unbiasedly identi ed using these functional indicators. Finding the abnormal rsFC of the DMN in an unbiased approach is possible by using the regions identi ed by these functional indicators as the seed region in patients with BD. In a previous study, the researchers have found that the MPFC, an important node of the DMN, showed abnormal brain activity as indicated by the ReHo and the DC in patients with bipolar mania 28 .
We attempted to use a whole-brain search strategy to identify the regions particularly impaired in the DMN and to identify the rsFC of the DMN in the patients with BD. Speci cally, from the point of view of local spontaneous brain activity and functional network, we used multiple brain functional indices including fALFF, ReHo, and DC to unravel abnormal brain spontaneous activities in BD. Subsequently, using the cluster that showed signi cant between-group differences as a seed, rsFC analyses, which re ect functional interactions of spatially distributed brain regions, were performed to further characterize the abnormal functional connectivity of the seed region in patients with BD.

Participants
A total of 150 participants were recruited from the community (85 HCs) and the Beijing Anding Hospital (65 patients with BD); the sample is the same as in our previous publication 29 . In summary, the patients and HC were matched for age, gender, and education. Twenty-three subjects (nine patients with BD and 14 HCs) were excluded from the nal analysis due to excessive head motion (details are shown in the "fMRI data preprocessing"). All patients satis ed the diagnostic criteria of the Structured Clinical Interview for DSM-IV (SCID) 30 . Our inclusion and exclusion criteria were described in our previous study 29 . In this experiment, the Young Mania Rating Scale (YMRS) 31 scores of all patients were less than 7, thereby indicating that all patients were not in a manic mood state. We focused on patients with BD in a depressive mood state or a euthymic mood state because BD patients spend more than 80% of their time in a nonmanic mood state. The Hamilton Depression Rating Scale (HAMD) 32 was used to assess patients' depressive symptoms.
The current study has been approved by the Institutional Review Board of Brain Image Center, Beijing Normal University and Beijing Anding Hospital, Capital Medical University. Written informed consents were obtained from all subjects prior to participation in this study.

Fmri Data Acquisition
All imaging data were acquired at the Brain Imaging Center of the Beijing Normal University by using a Siemens TIM Trio 3T scanner (Siemens, Erlangen, Germany). Resting state functional images (240 volumes) were acquired initially, followed by the T1 images. During scanning, subjects were asked to try not to swallow or move their body, especially their heads. During resting state, subjects were also instructed to close their eyes, relax, and remain awake. Resting state functional images were collected using echo-planar imaging (EPI) sequence, as follows: axial scanning, repetition

Fmri Data Preprocessing
Resting state fMRI date was preprocessed in the toolbox for the Data Processing and Analysis for (Resting State) Brain Imaging (DPABI v3.1) 33 . The preprocessing steps included removing the rst ve time points, slice timing, realignment, detrending, removing the head motion effect (using Friston 24-parameter model) 34 , regressing out nuisance covariates (the white matter signals, the cerebral spinal uid signals, and the global mean signals), normalizing the MNI space (using T1 images), resampling (voxel size = 2 × 2 × 2 mm 3 ), smoothing (FWHM = 4 mm, only for the fALFF and rsFC analysis), and band-pass ltering (0.01-0.1 Hz, only for the ReHo, DC, and rsFC analysis). In addition, we used volumebased framewise displacement (FD) 35 to quantify head motion. FD re ects head motion from one volume to the next, and a volume with FD < 0.2 indicates a good time point. Mean FD is calculated by the average of the sum of the absolute values of the differentiated realignment estimates (by backwards differences) at every timepoint 36 . Subjects would be excluded if their mean FD exceeded three standard deviations beyond the mean value of the entire sample or the number of good time points was less than 120.

Fmri Data Analysis
First, three indices of the fALFF, the ReHo, and the DC were used to unravel abnormal brain activities in the patients with BD. Subsequently, using the overlapped cluster of the signi cant brain region found in the analysis as seed, rsFC analysis was performed to explore possible brain regions, which showed abnormal rsFC with the overlapped cluster in patients with BD. Finally, two additional strategies were used to reidentify the seed region and then con rmed our rsFC analysis.
fALFF, ReHo, and DC calculation. Using the preprocessed resting state fMRI data, we calculated the fALFF, ReHo, and DC values for each subject in DPABI v3.1. All values were calculated at voxel level. The fALFF value was calculated using the signal strength of the low frequency range (i.e., 0.01-0.1 Hz) to divide the detectable entire frequency range 22 . The ReHo value, also called the Kendall's coe cient of concordance (KCC), of the time series of a given voxel with its nearest neighbors (27 voxels were considered) was calculated to generate individual ReHo maps 23 . The DC value of a certain voxel was the sum of the Pearson's correlation coe cients of all possible pairs of voxels (correlation threshold Page 5/20 of r 0 was set at 0.2) 24,37 . Lastly, the fALFF, ReHo, and DC maps for each subject were converted into Z-score maps, respectively. For the ReHo and DC maps, smoothing (FWHM = 4 mm) was performed.
Functional connectivity analysis. This analysis was performed in DPABI v3.1. The overlapped brain region, which showed signi cant group differences in fALFF, ReHo, and DC values between patients with BD and HC, was used as a seed in the following functional connectivity analysis. First, time series of each voxel was extracted within the seed. Second, the mean time series of the seed was calculated by averaging each voxel's time series. Third, the Pearson correlation coe cients of the mean time series of the seed and all other voxels' time series were calculated. They were used to construct each subject's rsFC map. Finally, each subject's rsFC map was converted into Z-score map.

Statistical Analysis
Two sample T test or chi-square tests were used to compared the differences between the patients with BD and the HC in demographic (including age, gender, education, and mean FD) and clinical (i.e. HAMD scores) variables in SPSS 20.0.
For each functional index (i.e., the fALFF, the ReHo, the DC, and the rsFC), analyses of the differences between the two groups were conducted in SPM12 (Wellcome Department of Cognitive Neurology, London, UK). In these analyses, twosample T tests were used with age, gender, education, and mean FD controlled. The signi cance level was set at the combination of voxel-level uncorrected P < 0.001 and cluster-level Family Wise Error (FWE)-corrected P < 0.05. These statistical analyses were limited in a pre-made gray mask. The gray mask was calculated by the average of the gray mask segmented from each subjects' T1 images.

Con rmation Analyses Of The Rsfc
To validate the ndings obtained in the rsFC analysis, we used two strategies to con rm the ndings. (i) We used the signi cant cluster obtained in the group analyses of the fALFF and the DC, but not the ReHo (due to the negative nding in a strict statistical threshold) to calculate the overlapped cluster. The overlapped cluster was then used as a seed in the rsFC analysis. (ii). We attempted to map our interested brain regions identi ed in the analysis of the fALFF, ReHo, and DC to the Brainnetome atlas 38 . Then, we selected the corresponding brain region of the atlas as seeds for further rsFC analysis. This method reduces the subjective bias in selecting the seed region and makes our ndings more likely to be replicated by future studies. Table 1 summarizes the demographic and the clinical data of all subjects. The two groups were comparable in terms of age, gender, education, and mean FD (all P-values > 0.05). The HAMD scores of the BD group were signi cantly higher than those of the HC group (T = 7.69, P < 0.001). As for drug use, the majority of patients was treated with at least one type of psychotropic medication (Table 1).

Falff, Reho And Dc Analyses
The fALFF, ReHo, and DC map within each group are shown in Fig. 1-A. The whole-brain analysis found that the BD group showed signi cantly lower fALFF and DC in the left precuneus. As for the ReHo, though the peak voxel reached the signi cant level of voxel-level FWE-corrected P < 0.05, the largest cluster, where the peak voxel was located did not reach our signi cant level. Thus, we used a relatively less stringent criteria for our ReHo data analysis, i.e., we regarded the largest cluster as an interesting cluster (p < 0.001, cluster-level uncorrected). Then, the signi cant clusters were overlapped, and the common overlapped cluster consisted of 47 voxels, located in the left precuneus ( Fig. 1-B and Table 2).

Seed-based Rsfc Analysis
The rsFC map within each group is also shown in Fig. 2-A. Using the overlapped 47 voxels found above as the seed, we found that the BD group showed decreased rsFC of the seed with the left precuneus and the MPFC and increased rsFC (i.e., the decreased negative rsFC) of the seed with the left putamen (extended to the left insula) (Fig. 2-B and Table 3).

Con rmation Analyses Of The Rsfc
For the rst con rmation analysis, when we used the signi cant region found in the fALFF and the DC analysis to identify the overlapped region, we found a cluster with 57 voxels located in the left precuneus (Fig. 3). Using this cluster as a seed, the rsFC map within each group is shown in Fig. 4-A. We found signi cantly decreased rsFC of the seed with the left precuneus and the MPFC and increased rsFC (i.e., the decreased negative rsFC) of the seed with the left putamen ( Fig. 4-B and Table 3).
For the second con rmation analysis, we found that the signi cant clusters of every single functional index (i.e., fALFF, ReHo, and DC) and the overlapping clusters corresponded consistently to the brain regions No. 153 and 154, also named bilateral area 31 (A31) of precuneus, in the Human Brainnetome atlas (Fig. 5-A). Then, we used the sum of the brain region Nos. 153 and 154 as a seed in the following rsFC analysis. The rsFC map within each group is shown in Fig. 5-B. The whole-brain analysis found a signi cantly decreased rsFC of the seed with the MPFC (Fig. 5-C and Table 3).

Discussion
We used three main functional indices (fALFF, ReHo, and DC) to measure the spontaneous activity from the local activity, local connectivity to network-level functional connectivity. We found that the different functional indices convergently identi ed the decreased spontaneous activity in the left precuneus, a core region in the DMN, in patients with BD. Furthermore, using this overlapped region in the left precuneus as a seed for rsFC, we found the decreased rsFC within the DMN, indicated by the decreased rsFC with the left precuneus and decreased rsFC between the seed precuneus and the MPFC in patients with BD. In addition, we found decreased negative functional connectivity between the left precuneus and the core regions in the SN (i.e., the left putamen/insula), suggesting decreased anticorrelation between the DMN and the SN in patients with BD.

Altered Spontaneous Activity Of The Dmn
We used multiple functional indices and found that the patients with BD exhibited decreased fAlff, ReHo, and DC values in the left precuneus (a speci c brain region of the DMN) compared with healthy controls. The decrease in the fALFF indicates reduced local BOLD signal uctuations of the precuneus 22,39 . The decrease in the ReHo indicates decreased synchronizing ability in relevant voxels within the precuneus 23 , and the decrease in DC indicates reduced functional connections of the precuneus with other voxels in the brain 40 . Our ndings are consistent with most previous studies 26, 41,42 , which found decreased spontaneous brain activities of the left precuneus in patients with BD. However, another study, which only recruited 17 bipolar depression patients and 16 HCs, reported increased ReHo values of the left precuneus in patients 43 .
The dysfunction detected by these functional indices may be explained by the abnormal metabolism and structure in this region. One PET study found reduced metabolism of the left precuneus in the old euthymic patients with BD 44 .
Another study using cortical gray matter thickness as an indicator found cortical thinning of the left precuneus in 1837 patients with BD 45 . The precuneus is located in the posterior medial part of the parietal lobe, which is extensively connected with cortex and subcortical brain areas 46 . The precuneus is associated with emotion regulation 47 , and is especially sensitive to negative emotion. A left lateralized activation of the precuneus was found when participants viewed negative images 48 or when they were asked to increase their negative emotions when facing aversive stimuli 49 .
One study using a negative emotion task found reduced activation of the left precuneus during negative images condition in the euthymic BD-I patients, with respect to controls 50 . This nding suggested that decreased brain activity of the left precuneus in patients with BD may somehow interfere with their ability to process emotional contents. The results found in the current resting-state fMRI study highlighted the importance of the precuneus, especially its left part, in the pathogenesis of BD. Thus, the patients' left precuneus was unable to maintain normal signal uctuation, synchronizing ability, and connectivity with other brain regions.

Decreased Rsfc Within The Dmn
The followed rsFC analysis found hypoconnectivity within the left precuneus in the patients with BD. The results were consistent with previous studies 51, 52 . Wang et al. 51 used FC strength (FCS) to measure the whole-brain FCS patterns, and found that unmedicated patients with BD-II depression showed decreased FCS in the DMN (including the left precuneus). Khadka et al. 52 using ICA found that the rsFC in the posterior DMN consisting mainly of bilateral precuneus was signi cantly decreased in patients with psychotic BD-I than the HC. Combining these ndings with our results, we speculated that the decreased local spontaneous activity of the left precuneus may account for its decreased rsFC. In addition, the precuneus is a functional core of the DMN 53 . It not only shows the highest resting metabolic rate within the network 54 , but has widespread connectivity with other brain regions 46 . It suggested that the anomaly of precuneus could have an important effect on the overall functional abnormality of the DMN 46, 53 .
The patients with BD also exhibited reduced rsFC between the left precuneus and the MPFC, which is another core node of the DMN 55 . This result was repeated in our con rmation analyses. The current result was, to some extent, consistent with previous report of Wang et al. 56 , although two additional studies using smaller sample size (the number of patients with BD is 15 and 30) did not nd positive results 21,57 . In Wang's study, unmedicated patients with BD under depressive episode were included, and patients showed disrupted intramodular connectivity within the DMN. The reason for the hypoconnectivity within the DMN may be the dysfunction of neurotransmission, such as one study found that the rsFC of the core DMN (composed of the precuneus and the MPFC) is related to K + transport protein, axonal guidance protein, neuronal generation/development protein, and cell differentiation protein 58,59 . The reduced rsFC between the MPFC and the left precuneus re ected weaker synchronization of spontaneous neural activities within the DMN, which could reduce the e ciency of the information communication and integration of the network 60 . The precuneus is involved in self-centered mental imagery strategies and episodic memory retrieval 46 . The MPFC is one of the most clearly delineated region in terms of its functional roles; it serves an important role in emotional experience, social cognition, memory, and decision making 61, 62 . Some researchers suggested that the abnormal rsFC between the anterior (e.g., the MPFC and the superior frontal gyrus) and the posterior (e.g., the PCC/precuneus) components of the DMN sever as the neural basis for the defects of attention and working memory 11 . The patients with BD have been reported to have these cognitive abnormalities 63 . One study on the relationship between the DMN and clinical symptoms of BD found that the decreased rsFC within the posterior DMN in patients with BD depression might be associated with rumination 16 , which is a typical symptom for depressed patients. It is characterized as a pattern of response to pain or stress that involves repetition and passive attention to negative self-related information 64 . From this aspect, it would be reasonable to speculate that the decreased rsFC between the MPFC and the precuneus may result in the cognitive dysfunction and clinical symptoms for patients with BD. In summary, our ndings with all previous studies support that the functional dysconnectivity within the DMN may play a relatively important role in the neural basis of BD.

Decreased Anticorrelation Between Dmn And Sn
Decreased negative connectivity was found between the left precuneus and the left putamen/insula, the main nodes of the SN 2, 65 , in the patients with BD. This nding re ects the decreased anticorrelation between the DMN and the SN.
In healthy population, resting-state fMRI signals of the DMN are anticorrelated with that of task-positive networks, such as the SN 66-68 . The SN generates control signals and causally mediates the "switching" between the DMN and the central executive network 69 . This condition leads to a switch between their functions of introspective and extroversive attention orientation, and then allows individuals to stay vigilant when meeting unexpected environmental events 11 .
This anticorrelation between DMN and SN has even been proven more important than the DMN's own activity 70 . The decreased anticorrelation between these networks may result in the reduction of introspective thinking and attention de cit, which could be the potential interference source of goal-directed behavior 11 . Weaker anticorrelation between the DMN and the SN was also found in many other brain disorders 66 . Our result expanded previous ndings of disruption of the DMN-SN functional connectivity in mental disorders 20 by showing that the negative rsFC between the left precuneus and the left putamen/insula was decreased in patients with BD.

Limitations
Some limitations are as follows. First, although the patients were not in a manic mood state, our sample was still heterogeneous in the subtypes or the stages of the illness. Thus, future work should recruit patients in the same mood state (mania, depression, or euthymia) to identify whether the signi cant results were state-dependent or trait-like characters for BD. Second, most patients have taken different types or dosages of drugs. Although no consistent effect of a certain drug was found on BOLD signals, different drugs or dosages may in uence BOLD signals from different aspects 71 , thereby interfering our results. In the future, unmedicated patients with BD should be recruited or the types and dosages of the patients' drug use should be strictly controlled. Third, our study did not consider whether these changes in the brain could account for any clinical or cognitive symptoms, such as rumination, and whether these brain abnormalities have changed over time. Thus, further studies are necessary to address these issues.

Conclusions
In summary, by using multiple functional indices, we found the decreased spontaneous brain activity in DMN, particularly located in the left precuneus, and decreased functional connectivity within the DMN and the disruptive anticorrelation between the DMN and the SN in patients with BD. These ndings suggest that the DMN is a key aspect for understanding the neural basis of BD, and the altered functional patterns of DMN may be a potential candidate biomarker for diagnosis of BD.  rsFC (using the overlapped 47 voxels as a seed) map within each group (A) and between-group differences in the rsFC map (B). The panel B showed that the left precuneus and the medial prefrontal cortex whose rsFC with the left precuneus were decreased, and the left putamen whose rsFC with the left precuneus was increased in patients with BD. Figure 4 rsFC (using the overlapped 57 voxels as a seed) map within each group (A) and between-group differences in the rsFC map (B). The panel B showed that the left precuneus and the medial prefrontal cortex whose rsFC with the left precuneus were decreased, and the left putamen whose rsFC with the left precuneus was increased in patients with BD.