Abnormal functional connectivity of the intrinsic networks in adolescent bipolar I versus bipolar II disorder

doi:10.21203/rs.3.rs-1544068/v1

Download PDF

Research Article

Abnormal functional connectivity of the intrinsic networks in adolescent bipolar I versus bipolar II disorder

https://doi.org/10.21203/rs.3.rs-1544068/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 31 May, 2024

Read the published version in Psychiatry Research: Neuroimaging →

Version 1

posted

You are reading this latest preprint version

Background: The symptoms of pediatric bipolar disorder (PBD)-I and PBD-II differ, but accurate identification at an early stage is difficult and may prevent effective treatment of this disorder. Therefore, it is urgent to elucidate a biological marker based on objective imaging indicators to help distinguish the two. Therefore, this research aims to compare the functional connectivity between PBD-I patient and PBD-II patient in different brain networks.

Methods: Our study enrolled 31 PBD-I and 23 PBD-II patients from 12 to 17 years of age. They were analyzed by resting state-functional connectivity through Independent component analysis（ICA）.

Results: We found differences between PBD-I and PBD-II in functional connectivity of the default network, frontoparietal network, salience network and limbic system. In addition, the clinical features, cognitive functions are associated with the functional connectivity of the intrinsic networks in PBD-I and PBD-II separately.

Conclusion: This research is the first to find differences in functional connectivity between PBD-I and PBD-II, suggesting that abnormality of the functional connectivity within large networks may be biomarkers that help differentiate PBD-I from PBD-II in the future.

fMRI

pediatric bipolar disorder

functional connectivity，resting state

default network

frontoparietal network

salience network，limbic system.

Pediatric bipolar disorder (PBD) is presented by recurrent depressive, manic, and hypomanic episodes^[1]. The prevalence of PBD in the adolescent population is 1%^[1]. PBD differs from adult bipolar disorder (BD) in that its symptoms are more atypical^[2]. Because of the atypical symptoms of PBD, the differential diagnosis of PBD-I and PBD-II is more difficult. The presence or absence of manic episode is a major feature for distinguishing PBD-I from PBD-II disorders^[1]. It is very important to identify PBD-I versus PBD-II biomarkers. Because PBD-I may begin with depressive episode or hypomanic episode in the early emotional episode, it is very easy to misdiagnose at this time without manic episode^[2]. However, PBD-I and PBD-II are different in medication therapy and risk prevention^[3], so it is very important to correctly identify them with biomarkers in early stage. Some researches have showed differences in biological mechanisms compared BD-I with BD-II ^{[3, 4]}. Psychosis^[3], neurotoxic processes^[5], neurochemical^[6] and cognitive dysfunction^[7] are greater common in BD-I than in BD-II. The Magnetic resonance imaging (MRI) researches have showed that there are difference of cortical-subcortical system compared BD-I with BD-II in volumes^[8–10], thickness^{[8, 11]}, white matter lesions^[9], metabolism^[12] and functional connectivity^[13]. However, studies on MRI differences between PBD-I and PBD-II are obviously lacking.

The emergence of studies have found that researching the functional connectivity of bipolar disorder can clarify the neuropathological mechanism of this disorder. Resting state functional magnetic resonance imaging (fMRI) has been explored to explore bipolar disorder^{[13, 14]}. The statistical dependence of blood oxygen level dependent (BOLD) sign among the two neural regions was hypothesized as a functional connectivity^[15]. The latest meta-analysis in 2021 found alteration in functional connectivity of the fronto-parietal network (FPN), default network (DMN), salience network(SN)and limbic system in bipolar disorder^[14]. Previous study using resting state-fMRI and ICA has found aberrations in the intrinsic connectivity neural circuits mentioned above in BD^[16–19]. To date, few resting state fMRI studies have been conducted specifically in PBD. Our earlier study has adopted the Independent Components Analysis (ICA) to find discrepancy in resting state DMN functional connectivity in PBD^[20].

Aberrant functional connectivity of the DMN indicates the shortage of integration of internal thoughts, which may show decreased cognitive flexibility and impaired self-monitoring process of BD^[14]. Functional connectivity abnormalities in DMN may provide evidence for clinical symptoms observed in BD, such as working memory dysfunction and rumination^[14]. Effective emotional regulation depends on limbic system^[21]. Dysfunctional connectivity of the limbic system is related to the severity of manic mood^[21]. And SN involved in goal-directed adjustment of external stimulus and internal thought, its dysfunctional connectivity shows imbalanced communication between the neural networks^[22]. BD patients will consider emotional information as external stimulation rather than emotion itself, which further explains the emotional dysregulation in BD^[22]. Abnormal functional connectivity within FPN is also commonly found in BD patients^[23]. The abnormal functional connectivity of the FPN in BD subjects may be the basis of impulsivity and attention deficit in BD^{[14, 23]}. In recent years, many studies have found that PBD is similar to adult BD in functional connectivity of the DMN, FPN, SN, and limbic system^[24–26].

However, no studies have analyzed the functional connectivity in the DMN, FPN, SN, and limbic system compared PBD-I with PBD-II patients. The clinical manifestations of PBD-I were found to be more severe than PBD-II in terms of manic symptoms and cognitive dysfunction^[27], but whether the functional connectivity of the neural circuits serves as the neural underpinning is not clear and need to be explored. To detect whole-brain connectivity mode, we used the model-free methods to explore connectivity patterns without defining prior seed regions. Comparing with the seed-based approach, the model-free approach aims to find widespread mode of connectivity in the whole brain. The ICA-based approach is probably the most generally used and has been demonstrated a high degree of consistency ^[28].

In this research, our aim was to compare differences in functional connectivity comparing PBD-I with PBD-II. Our hypothesis was that there might be differences in functional connectivity of the DMN, FPN, SN and limbic system compared PBD-I with PBD-II. Furthermore, we examined the correlation among manic symptoms, depressive symptoms, sleep quality, suicide risk, cognitive function and the functional connectivity of the intrinsic networks, to see if differences in fMRI indicators were related to clinical symptoms and cognitive impairment. To our knowledge, this was the first functional connectivity research to assess wide neural circuits compared PBD-I with PBD-II.

2.1 Participants

We collected 31 PBD-I and 23 PBD-II adolescents in the outpatient service of pediatric psychiatry, Xiangya Second Hospital, China. Xiangya Hospital Research Ethics Committee approved this research. The process of this study obeyed the 1964 Helsinki ethical statement. Each subject and their parents agreed to join in the research and signed informed consent.

This research used the DSM-5 criteria to diagnose PBD-I and PBD-II patients. Exclusion criteria as following: (a) autism spectrum disorder; (b) Wechsler intelligence total score ≤ 80; (c) schizophrenia spectrum disorder; (c) mental disorders due to physical diseases; (d) other specific/unspecific bipolar disorder; (e)in pregnancy; (f) Psychoactive drug use in the two months.

2.2. Emotional and cognitive features

2.2.1. Mood assessments and Demographics

Structured interviews were conducted with PBD patients by the K-SADS-PL^[29]. Two subtypes of PBD were included type I and type II. The key difference comparing PBD-I with PBD-II was that there was no full-blown manic episode in PBD-II. The consistency between psychiatrists was proved to be good. The clinical features of all the PBD patients was assessing using Wechsler Abbreviated Scale of Intelligence (WASI) ^[30], the Mood and feelings Questionnaire (MFQ) ^[32], Young Mania Rating Scale (YMRS)^[31], Beck Suicide Ideation Scale (BSIS)^[34] an Pittsburgh Sleep Quality Index (PSQI)^[33]. We used the Edinburgh Inventory to determine handedness^[35]. All interviews and assessments were conducted on the right day of the MRI acquisition to improve accuracy.

2.2.2 Cognitive Function Test

To measure the patient's cognitive function, cognitive test including: Stroop color-word test (SCWT), digit span test (DST), trail making test (TMT), and visual reproduction immediate recall test (VR I). These tests were operated by trained psychiatrists. The brief descriptions of the cognitive tests were following:

(1) Stroop Color-word Test (SCWT)

This SCWT ^[36] could measure participants’ attention and response inhibition. It contained three subtasks, such as reading characters (SCWT-A), reading colour (SCWT-B), and reading colour disturbance (SCWT-C). SCWT-A included 100 words which participants read in 45 s. SCWT-B contained 100 symbols that participants named color correctly in 45 s. SCWT-C included 100 colored words that subjects named color within 45 s.

(2) Trail Making Test (TMT)

TMT consisted of TMT-A and TMT-B^[37]. The first test demanded the participants to link the scattered numbers in order (1–25), and the second test asked the participants to connect a line sequentially through the 13 numbers (from 1 to 13) and 12 English alphabet (from A to L) alternately. TMT scores were the whole times for participants to accomplish these tasks. TMT-A measured processing speed and attention, as well as TMT-B measured cognitive flexibility^[37].

(3) Digit Span Subtest (DST)

The participants were required to recite the assigned number forwards (DST-A) and to recite the assigned number backwards (DST-B). The scores were based on the longest sequence of numbers recited. DST-A could measure attention level and DST-B could assess ability of working memory ^[30].

(4) Visual Reproduction Immediate test (VR I)

This test was operated to measure visual memory. Evaluators showed patients three geometric images, one page each time. At first, participants were asked to view each picture for 10 seconds. Then, the participants were asked to draw the image and details as correctly as possible according to the memory^[38].

2.3 MRI Acquisition

MRI images were collected by a 3.0-Tesla Siemens scanner. Functional image was collected by a single-shot gradient echo-echo imaging sequence. The parameters is: echo time (TE) = 30 ms, repetition time (TR) = 2000 ms, slice thickness = 4 mm, slices = 30, field of view (FOV) = 240×240 mm², gap = 0.4 mm, matrix = 64×64, flip angle = 90°.

The acquisition of structural images was used a 3-dimensional T1-weighted sequence. The following parameters: TE = 2.03ms, TR = 2300 ms, matrix = 256×256, slices = 176, slice thickness = 1.0 mm, FOV = 256×256 mm2, flip angle = 9°. Overall, 250 volumes were acquired in 500s. During the acquisition, all patients were demanded to hold still, relax, close their eyes, not think special. We used the foam padding to reduce head motion.

2.4 Data Preprocessing

This study used Statistical Parametric Mapping software SPM12 to preprocess MRI Data. For head motion correction, slice timing and realignment were used. Any patients whose head rotation more than 1.0°or motion more than 1.0 mm was excluded.

We applied the standard Montreal Neurological Institute (MNI) template within SPM12 to normalize MRI images. This research used the voxel size of 3 mm×3 mm×3 mm to resample. For minimizing spatial noise, the normalized data were smoothed by convolution with the isotropic Gaussian kernel (8 mm full-width at half-maximum). At every process, the output was proved, and data were excluded if their quality did not meet standard.

2.5 Group ICA and recognition of the intrinsic networks

The fMRI Toolbox Spatial ICA was performed to determine independent components (ICs) for all subjects (GIFT, version 4.0b). We measure the number of ICs using the dimension estimation by the minimum length criteria^[39]. In the two groups, the numbers of ICs were 40. We could make sure that the neural circuits shared the semblable spatial mode in the two groups.

This study used normative means of denying factitious ICA neural circuits. We visually identified neural circuits to exclude those obviously showing artifacts. Then the default mode network (DMN), fronto-parietal network (FPN), salience network (SN) and limbic system of the two groups were chosen according to the biggest spatial correlation with a previous pattern^[40].

2.6 Statistical Analysis

Clinical features, demographics and cognitive function were statistically analyzed by SPSS version 22.0. FMRI images were calculated by SPM12. To explore the intrinsic networks differences between PBD-I and PBD-II, two-sample T test was calculated of the DMN, FPN, SN and limbic system in the voxel-by-voxel pattern [P<0.05, using family-wise error (FWE) correction]. Head motion, gender and age were performed as covariates in our statistical analysis. Additionally, correlation between variations in DMN, FPN, SN, limbic system z values and clinical/cognitive scores were performed by Spearman’s rank correlation within both PBD-I and PBD-II. Correlation analysis was calculated by SPSS 22.0 (threshold, P<0.05, uncorrected).

3.1 Subjects’ Demographics, Clinical Features and Cognitive Functions

The demographics, symptomatic characteristics and cognitive function were described at Table 1. There was no obvious differences in age (t = 0.86, p = 0.40), education level (t = 0.75, p = 0.46), gender (chi-square = 1.00, p = 0.32 ), intelligence (t =-1.06, p = 0.30), BSIS( t = 0.40, p = 0.70), PSQI(t = 0.46 ,p = 0.65 ) ,MFQ (t =-0.41 ,p = 0.68 ), YMRS (t = 1.70, p = 0.10 ), age of onset (t =-0.08, p = 0.94), with psychotic symptoms (chi-square = 0.04, p = 0.85), onset frequency (t =-0.48, p = 0.64), and familial psychiatric history (chisquare = 0.08, p = 0.78).

Table 1

Clinical characteristics and cognitive function of PBD-I and PBD-II
	PBD-I Mean (SD)	PBD-II Mean (SD)	t/\(\)²	P value
Age	15.16 (1.95)	14.74 (1.54)	0.857	0.395
Gender (male/female)	15/16	10/13	1.000	0.317
Education (years)	8.19 (1.94)	7.83 (1.56)	0.747	0.458
IQ	101.68 (13.13)	105.35 (11.85)	-1.058	0.295
BSIS	3.10 (4.57)	2.64 (3.81)	0.386	0.701
PSQI	6.61 (3.67)	6.14 (3.73)	0.463	0.646
MFQ	13.42 (12.97)	14.87 (12.60)	-0.411	0.683
YMRS	17.00 (14.73)	10.74 (12.31)	1.699	0.095
Course of disease (months)	19.65 (13.32)	11.78 (13.67)	2.12	0.039
Age of onset (years)	13.61 (1.71)	13.65 (2.01)	-0.077	0.939
Onset frequency	3.71 (2.27)	4.36 (7.17)	-0.477	0.635
Familial BD history (yes/no)	6/25	7/16	0.077	0.782
Psychotic symptoms (yes/no)	15/16	14/9	0.034	0.853
Medication, n (%)	23 (74.2%)	15 (65.2%)	1.684	0.194
SCWT-A	53.29 (15.16)	51.59 (14.04)	0.414	0.680
SCWT-B	66.97 (19.02)	67.82 (20.87)	-0.154	0.878
SCWT-C	31.10 (8.54)	28.95 (8.28)	0.911	0.367
TMT-A	39.06 (13.69)	38.36 (9.40)	0.208	0.836
TMT-B	98.03 (45.44)	83.95 (25.52)	1.312	0.196
VR I	9.71 (2.69)	9.32 (4.47)	0.367	0.716
DST-A	8.23 (1.43)	8.55 (1.54)	-0.778	0.440
DST-B	4.65 (1.50)	4.91 (2.02)	-0.547	0.587

In the PBD-I group, subjects were taking valproate (n = 15), lithium (n = 10), antipsychotics (n = 22), and antidepressants (n = 1). PBD-II subjects were taking valproate (n = 10), lithium (n = 7), antidepressants (n = 3) and antipsychotics (n = 14). There was no obvious difference in medication between PBD-I and PBD-II (chisquare = 1.68, p = 0.19).

No significant differences were exhibited in the Stroop Color-word test (SCWT)-A (t = 0.41, p = 0.68 ), SCWT-B (t =-0.15, p = 0.88), SCWT-C (t = 0.91, p = 0.37), Trail making tes(TMT)-A (t = 0.21, p = 0.84), TMT-B (t = 1.31, p = 0.20), VR I (t = 0.37, p = 0.72), DST-A (t =-0.78, p = 0.44) and DST-B (t =-0.55, p = 0.59)(see Table 1).

3.2 Between-Group ICA Analyses

The findings showed significant differences in the default network (DMN), fronto-parietal network (FPN), salience network (SN) and limbic system compared PBD-I with PBD-II (p < 0.05, FWE corrected). Intra-network functional connectivity compared PBD-I with PBD-II are shown in Table 2, as following:

Table 2

Different functional connectivity in the intrinsic networks comparing PBD-I with PBD-II
PBD-I vs PBD-II
DMN				FPN				SN				Limbic system
Brain regions	MNI x, y, z	Peak t value	Volume	Brain regions	MNI x, y, z	Peak t value	Volume	Brain regions	MNI x, y, z	Peak t value	Volume	Brain regions	MNI x, y, z	Peak t value	Volume
Increased connectivity				Increased connectivity				Increased connectivity				Increased connectivity
PCC (L/R)	-5,-47,23/ 5,-42,28	4.30/ 5.90	20/ 28	MTG (R)	52,-57,15	7.00	80	AIn (L/R)	-35,16,-9 37,10,-6	7.70 5.17	530/ 500	PoG (L)	-58,-7,28	3.55	380
AnG (L/R)	-45,-63,42 49,-57,39	8.40/ 5.45	25/ 21	dlSFG (R)	16,65,10	5.41	180	IFG (L/R)	-50,34,1 50,34,-1	9.47 8.76	530/ 500	MFOr (L)	-23,35,-18	4.02	356
				FOr (L/R)	-37,50,-12 33,54,-12	8.10/ 10.43	97/ 223					IFOr (R)	21,30,-21	3.90	240
												PHG (L/R)	-26,-1,-33/ 25,-8,-30	3.78/ 5.18	140/ 270
Decreased connectivity				Decreased connectivity				Decreased connectivity				Decreased connectivity
SPG (L/R)	-29,-59,52 29,-61,52	-5.6/ -4.5	19/ 18	SPG (R)	30,-56,55	-8.56	530	ACC (L/R)	-6,34,15 6,25,24	-4.20 -6.44	500/ 530	olf (L/R)	-5,15,-16 5,14,-16	-6.22 -4.84	67/ 52
PrCu (L/R)	-9,-54,34/ 7,-56,31	-5.45/ -6.61	21/ 28	dlSFG (L)	-20,64.9	-11.97	115	MCC (L/R)	-7,10,49 7,12,35	-5.67 -6.12	500/ 530	GR (L/R)	-8,13,-17/ 6,45,-14	-3.23/ -3.09	113/ 144
Abbreviations: PBD, pediatric bipolar disorder; MNI, the Montreal Neurological Institute; L, left; R, right; PCC, posterior cingulate gyrus; AnG angular gyrus; SPG, superior parietal gyrus; PrCu, precuneus; MTG, middle temporal gyrus; dlSFG, dorsolateral superior frontal gyrus; FOr, orbitofrontal gyrus; AIn, anterior insula; IFG, inferior frontal gyrus; ACC, anterior cingulate gyrus; MCC middle cingulate gyrus; PoG, posterior central gyrus; MFOr, middle orbitofrontal gyrus; IFOr, inferior orbitofrontal gyrus; PHG, parahippocampal gyrus; Olf, olfactory cortex; GR, rectus gyrus.

DMN: Compared with PBD-II, PBD-I subjects presented greater connectivity in the posterior cingulate gyrus and bilateral angular gyrus, while reduced connectivity in the bilateral superior parietal gyrus and bilateral precuneus (see Fig. 1).
FPN: Compared to PBD-II, PBD-I presented increased connectivity in the right middle temporal gyrus, right dorsolateral prefrontal gyrus and bilateral orbitofrontal gyrus, meanwhile reduced connectivity in the right superior parietal gyrus, left dorsolateral prefrontal gyrus (see Fig. 2).
SN: Compared to PBD-II, PBD-I showed greater connectivity in the bilateral anterior insula and bilateral inferior frontal gyrus, while reduced connectivity in the bilateral anterior cingulate gyrus and bilateral middle cingulate gyrus (see Fig. 3).
Limbic system: Compared to PBD-II, PBD-I exhibited greater connectivity in the left postcentral gyrus, left middle orbitofrontal cortex, right inferior orbitofrontal cortex, and bilateral parahippocampal gyrus, meanwhile reduced connectivity in the bilateral olfactory cortices and bilateral rectus gyrus (see Fig. 4).

3.3 Correlation Analysis

The correlation between clinical characteristics, cognitive function and intrinsic network connectivity were as following (p<0.05):

Correlation in DMN connectivity. In PBD-I group, connectivity of the posterior cingulate cortex was positively related with forward order recite (DST-A)(r = 0.49,p = 0.009), and the left angular gyrus was positively related to reverse order recite (DST-B) (r = 0.42,p = 0.030). In PBD-II group, the connectivity of the posterior cingulate cortex was negatively related to suicidal ideation (BSIS) (r=-0.69,p = 0.003), and the right precuneus was negatively associated with depression (MFQ) (r=-0.53,p = 0.017) (see Fig. 1C).
Correlation in FPN connectivity. No correlation was exhibited between FPN connectivity and mood symptoms in PBD-I and PBD-II. In PBD-I, visual memory (VR I)(r = 0.48,p = 0.007), working memory (DST-B)(r = 0.37,p = 0.044)was found to be positively associated with right superior parietal gyrus. In PBD-II group, attention (SCWT-A/B) (r = 0.64, p = 0.005/ r = 0.53, p = 0.029), response inhibition (SCWT-C)(r = 0.65,p = 0.005)and cognitive flexibility (TMT-B) was positively related to the left orbitofrontal cortex(r = 0.62,p = 0.007), as well as visual memory (VR I) was negatively related to the right dorsolateral prefrontal cortex(r=-0.54,p = 0.026)(see Fig. 2C).
Correlation in SN connectivity. In PBD-I group, the connectivity of the right anterior insula was correlated with attention (Stroop-A) (r=-0.44,p = 0.016), processing speed (TMT-A) (r = 0.54,p = 0.002) and working memory (DST-B) (r=-0.42,p = 0.029). In PBD-II, correlation in SN connectivity as following: 1.Depression (MFQ) (r=-0.48,p = 0.030)was negatively correlated with the left anterior insula; 2.Attention (Stroop-A, Stroop-B) (r=-0.61,p = 0.009; r=-0.52,p = 0.034), response inhibition (Stroop-C) (r=-0.50,p = 0.040)and cognitive flexibility (TMT-B) (r = 0.50,p = 0.042)was correlated with bilateral anterior insula; 3.Attention (Stroop-B) (r = 0.53,p = 0.028)and visual memory (VR I) (r = 0.58,p = 0.028) was positively related to the left inferior frontal gyrus; 4. Attention and processing speed (TMT-A) was negatively correlated with the middle cingulate gyrus (r=-0.51,p = 0.037) (see Fig. 3C).
Correlation in limbic system connectivity. In the PBD-I group, sleep quality (PSQI) was positively correlated with the left postcentral gyrus (r = 0.40,p = 0.028). Visual memory (VR I) (r=-0.56,p = 0.002) and working memory (DST-B) (r=-0.42,p = 0.039)were negatively correlated with the right parahippocampal gyrus, as well as cognitive flexibility (TMT-B) (r=-0.41,p = 0.035),working memory (DST-A) (r = 0.47,p = 0.014)were associated with the left middle orbitofrontal gyrus in PBD-I. Within PBD-II group, sleep quality (PSQI) was positively correlated with the bilateral parahippocampal gyrus (r = 0.53,p = 0.028), as well as manic symptom (YMRS) was positively associated with the left middle orbitofrontal gyrus(r = 0.68,p = 0.008), right inferior orbitofrontal gyrus(r = 0.56,p = 0.039)and bilateral rectus gyrus connectivity(r = 0.58,p = 0.029). Additionally, response inhibition (Stroop-C) was negatively correlated with the left parahippocampal gyrus in PBD-II (r=-0.67,p = 0.003) (see Fig. 4C).

BD-II was used to consider as briefly a ‘softer’ pattern of BD-I, thus supposing that the identical neuropathological dysfunction underlying BD-I and BD-II. It's just that the degree of the dysfunctions can vary. Recently, this perspective has been challenged through clinical and neurobiological findings. Clinical researches presented that course and severity of symptoms in BD-II were at least as severe as in BD-I ^{[27, 41]}. The differences in resting state fMRI comparing BD-I with BD-II were insufficient, particularly in pediatric bipolar disorder (PBD). This research focused on the difference of the functional connectivity in resting state fMRI comparing PBD-I with PBD-II.

In general, Our findings were consistent with prior results of BD, with differences between BD-I and BD-II^[13]. PBD-I showed more abnormal functional connectivity within default mode network (DMN), fronto-parietal network (FPN), salience network (SN), and limbic system compared with PBD-II. According to meta-analysis^{[14, 42]}, abnormal functional connectivity of the intrinsic networks has been found in adult BD related to healthy control. Besides, we also found that functional connectivity in intrinsic networks were correlated with emotional symptoms and cognitive impairments.

Default mode network (DMN)

Our study found that functional connectivity of posterior cingulate gyrus and bilateral angular gyrus were enhanced, while functional connectivity of bilateral superior parietal cortex and bilateral precuneus were weakened when PBD-I compared with PBD-II. In PBD-I, the posterior cingulate cortex was positively associated with forward order digital recite, and the left angular gyrus was positively correlated with reverse order digital recite. In PBD-II, the posterior cingulate cortex was negatively associated with suicidal ideation, and the right precuneus was negatively associated with depression.

In the latest meta-analysis, the patients of BD exhibited changed functional connectivity in the DMN^[14]. The altered functional connectivity within the DMN presented the shortage of integration of inner activity, might show aberrant communication in external cognitive flexibility and internal self-monitoring function^{[22, 43]}. Our previous fMRI research showed weakened functional connectivity of the DMN in the PBD^[44].

Working memory and self-reflection processing task-fMRI research in BD-I showed abnormal activation and functional connectivity of the precuneus, which supported the precuneus could be the pathophysiologic mechanism of BD^{[45, 46]}. Precuneus has been shown to involve self-referential processing and depressive rumination^[47]. This study exhibited difference of the precuneus connectivity comparing PBD-I with PBD-II, and which was correlated with depression only in PBD-II. This result indicated that PBD-I versus PBD-II had distinction in the functional connectivity of the precuneus.

Disrupted functional connectivity of the posterior cingulate cortex (PCC) was likely to support the evidence for working memory and cognitive function impairment in BD^[14]. Our study supported this theory and found that decreased functional connectivity in PCC when PBD-I compared with PBD-II. In PBD-I, PCC was positively associated with attention, and the left angular gyrus was positively correlated with working memory. PCC is also a key nodes of the DMN, the activity of which is involved in emotional processing^[48], self-evaluation^[48] and suicidal ideation^[49] in bipolar disorder. Our study verified this theory and found that PCC connectivity was associated with the suicidal ideation of PBD-II. Overall, our result suggested that difference in DMN connectivity between PBD-I and PBD-II was obvious, which was associated with depressive symptoms, suicidal ideation and cognitive function.

Fronto-parietal network (FPN)

We found both hyper- (FPN seeds and regions of the right middle temporal cortex, right dorsolateral prefrontal gyrus and bilateral orbitofrontal cortex) and hypo-connectivity (FPN region of the right superior parietal gyrus, left dorsolateral prefrontal gyrus) within the FPN while PBD-I compared with PBD-II. The FPN was contributed to varied cognitive function, such as working memory, attention and response inhibition ^{[23, 50–52]}. In the correlation analysis of this study, FPN functional connectivity was associated with attention, working memory, cognitive flexibility and response inhibition in PBD-I and PBD-II respectively. Altered resting state-functional connectivity in the FPN are common findings in cognitive dysfunction of BD^{[23, 50–52]}. Our finding presented further evidence for the connectivity difference comparing PBD-I with PBD-II in the FPN.

Salience network (SN)

Salience network (SN) is considered to a core hub in exploring salient stimuli, and modulate the attention and working memory resources for that^{[53, 54]}. The crucial view is the SN mediate dynamic communication between other neural circuits referring to internally oriented attention and externally oriented cognition^{[53, 55]}. This study showed both hyper- (SN seeds and regions of bilateral anterior insula and bilateral inferior frontal gyrus) and hypo-connectivity (SN seeds and region of bilateral anterior cingulate gyrus and middle cingulate gyrus) within the SN compared PBD-I with PBD-II.

The inferior frontal cortex, anterior insula, and anterior cingulate cortex are significant node of a "salience network" which helps extract the most relevant information in the stimuli to regulate behavior^[53]. Taken together, the observed hyper-functional connectivity in anterior insula and inferior frontal gyrus between the PBD-I and PBD-II might show the imbalanced communication between cognitive and affective network^[53]. This study also exhibited hypo-connectivity within the anterior cingulate cortex and bilateral middle cingulate cortex comparing PBD-I with PBD-II. Anterior cingulate gyrus involves in conducting salience, which present disrupted coordination with neural circuits involved in internal oriented thought and external goal-oriented control^[56]. This study provided a further explanation for more dysfunction of the cognitive function and emotion regulation in contrast PBD-I with PBD-II^{[57, 58]}.

It is worth emphasizing that anterior insula is crucial in the regulating negative emotion^[59]. This study supported this perspective, and we found functional connectivity of the left anterior insula was related to depression in PBD-II. Besides, anterior insula, as a core node of SN, plays an important effect in high-level cognitive and attentional process^[56]. In this research, the functional connectivity of the anterior insula was correlated with attention in both PBD-I and PBD-II. Additionally, the left inferior frontal gyrus was correlated with attention and visual memory, and the middle cingulate gyrus was correlated with attention and processing speed in PBD-II. In general, PBD-I versus PBD-II were found to differ in functional connectivity of the salience network, and which was associated with mood and cognitive dysfunction.

Limbic system

A meta-analysis study found that prefrontal-limbic system dysfunction was an important pathological mechanism of PBD^[21]. This study exhibited that PBD-I had different functional connectivity of the limbic system compared with PBD-II (enhanced functional connectivity in the left postcentral cortex, left middle orbitofrontal cortex, right inferior orbitofrontal cortex, and bilateral parahippocampal gyrus; weakened functional connectivity in the bilateral olfactory cortices and bilateral rectus gyrus). In previous comparative studies of BD-I and BD-II, distinction were reported in the structure^[8],functional connectivity and anatomical connectivity^[13] of the limbic system. Our study confirmed that PBD-I versus PBD-II had altered functional connectivity in the limbic system as those in adult patients ^[13]. It could be seen that limbic systems had significant differences based on functional connectivity comparing PBD-I with PBD-II.

Additionally, we found an association between functional connectivity of the left postcentral cortex and sleep quality in PBD-I; and the functional connectivity of the bilateral parahippocampal gyrus of PBD-II was correlated with sleep quality. It has been proposed that the postcentral gyrus and parahippocampal gyrus in the limbic system participate in the regulation of sleep in adolescents^[60]. Clinical observation has proved that children with severe insomnia due to brain surgery injury to the limbic system^[61]. Therefore, abnormal functional connectivity of the limbic system in PBD may be the mechanism of the common sleep problems in PBD patients.

Besides, the functional connectivity in the left middle orbitofrontal gyrus, right inferior orbitofrontal cortex and bilateral rectus cortex were correlated with manic symptoms. Manic symptoms are the core emotional symptoms of PBD-I and the key difference between PBD-I and PBD-II (only hypomanic episodes occur in PBD-II)^[27]. The orbitofrontal gyrus in the frontal-limbic system is known to be referred to emotional processing^[62]. The fact that the symptoms of mania was correlated with the functional connectivity in the orbitofrontal gyrus suggested that the possible mechanism underlying the difference in manic mood comparing PBD-I with PBD-II.

In the correlation of cognitive function and functional connectivity, PBD-I found that the right parahippocampal gyrus was correlated with visual memory and working memory; and the left orbitofrontal gyrus was correlated with cognitive flexibility, working memory and attention. PBD-II only found that the left parahippocampal gyrus was associated with response inhibition. The parahippocampal gyrus and the orbitofrontal gyrus in the limbic system is involved in memory processing, and the damage of which structure can cause the abnormality of emotion and memory^[63]. Our result suggested that differences in the functional connectivity of the limbic system between PBD-I and PBD-II are associated with cognitive impairment, especially memory dysfunction. The difference of functional connectivity in the limbic system may provide new evidence for the pathological mechanism that PBD-I has more disorganized cognitive function and emotion than PBD-II^[27].

Some limitation of our research should be noticed. Firstly, our sample size was not large. If verified in greater PBD sample, this result could propose as the biomarker for the differential diagnosis of PBD-I and PBD-II. Secondly, major PBD subjects were using medication. Prior researches have presented that medication neutralized the aberrant functional connectivity of BD^[64]. Although this research showed changed functional connectivity in PBD-I compared with PBD-II, it still needed to be detected how psychoactive medication impacted the functional connectivity of PBD-I and PBD-II.

In summary, this study showed that abnormal functional connectivity in the DMN, FPN, salience network and limbic system when PBD-I compared with PBD-II. Besides, associations between clinical features, cognitive function and changed functional connectivity in the intrinsic networks were presented in PBD subtypes. This study, from the first finding to explore difference of functional connectivity comparing PBD-I with PBD-II, proposed the possibility that altered functional connectivity in the DMN, FPN, salience network and limbic system could exhibit a different neural mechanism between PBD-I and PBD-II.

Funding: This research was funded by the Youth Science Foundation of Xiangya Hospital:2020Q20, Natural Science Foundation of Jiangsu Province, No.BK20191369, the National Natural Science Foundation of China（No. 81871344）and Higher Education Institutions “Qing Lan Project” in the Jiangsu Province.
No conflict exists: None of the authors have a conflict of interest to declare.
Consent to Publish：All the co-authors, including Qian Xiao, Gui Zhang, Yuan Zhong have consented to publish this paper.
Ethical approval: All research process obeyed the standards of the 1964 Helsinki Declaration. The Research Ethics Committee of Xiangya Hospital approved our research.
Consent to Participate: All adolescent and their parent consent to take part in the research and signed written consent.
Contribution：The main contributors are Qian Xiao, Gui Zhang, Yuan Zhong. The contributions of Xiao Qian's are the enrollment of subjects, MRI analysis and paper writing. The contribution of Gui Zhang is the analysis of imaging information and statistical analysis. The contribution of Yuan Zhong is revision of this manuscript.
Acknowledgement: No acknowledgement！
Availability of data and materials: Data and materials are available.

Johnson, J. G., Cohen, P., & Brook, J. S. (2000). Associations between bipolar disorder and other psychiatric disorders during adolescence and early adulthood: a community-based longitudinal investigation[J]. Am J Psychiatry, 157(10), 1679–1681
Perlis, R. H., Miyahara, S., Marangell, L. B., et al. (2004). Long-term implications of early onset in bipolar disorder: data from the first 1000 participants in the systematic treatment enhancement program for bipolar disorder (STEP-BD)[J]. Biol Psychiatry, 55(9), 875–881
Parker, G., & Fletcher, K. (2014). Differentiating bipolar I and II disorders and the likely contribution of DSM-5 classification to their cleavage[J]. J Affect Disord, 152–154, 57–64
Phillips, M. L., & Kupfer, D. J. (2013). Bipolar disorder diagnosis: challenges and future directions[J]. Lancet, 381(9878), 1663–1671
Uemura, T., Green, M., Corson, T. W., et al. (2011). Bcl-2 SNP rs956572 associates with disrupted intracellular calcium homeostasis in bipolar I disorder[J]. Bipolar Disord, 13(1), 41–51
Atagun, M. I., Sikoglu, E. M., Can, S. S., et al. (2018). Neurochemical differences between bipolar disorder type I and II in superior temporal cortices: A proton magnetic resonance spectroscopy study[J]. J Affect Disord, 235, 15–19
Bora, E., Yucel, M., Pantelis, C., et al. (2011). Meta-analytic review of neurocognition in bipolar II disorder[J]. Acta Psychiatr Scand, 123(3), 165–174
Toma, S., Islam, A. H., Metcalfe, A., et al. (2019). Cortical Volume and Thickness Across Bipolar Disorder Subtypes in Adolescents: A Preliminary Study[J]. J Child Adolesc Psychopharmacol, 29(2), 141–151
Ambrosi, E., Rossi-Espagnet, M. C., Kotzalidis, G. D., et al. (2013). Structural brain alterations in bipolar disorder II: a combined voxel-based morphometry (VBM) and diffusion tensor imaging (DTI) study[J]. J Affect Disord, 150(2), 610–615
Abe, C., Ekman, C. J., Sellgren, C., et al. (2016). Cortical thickness, volume and surface area in patients with bipolar disorder types I and II[J]. J Psychiatry Neurosci, 41(4), 240–250
Abe, C., Rolstad, S., Petrovic, P., et al. (2018). Bipolar disorder type I and II show distinct relationships between cortical thickness and executive function[J]. Acta Psychiatr Scand, 138(4), 325–335
Li, C. T., Hsieh, J. C., Wang, S. J., et al. (2012). Differential relations between fronto-limbic metabolism and executive function in patients with remitted bipolar I and bipolar II disorder[J]. Bipolar Disord, 14(8), 831–842
Caseras, X., Murphy, K., Lawrence, N. S., et al. (2015). Emotion regulation deficits in euthymic bipolar I versus bipolar II disorder: a functional and diffusion-tensor imaging study[J]. Bipolar Disord, 17(5), 461–470
Gong, J., Wang, J., Chen, P., et al. (2021). Large-scale network abnormality in bipolar disorder: A multimodal meta-analysis of resting-state functional and structural magnetic resonance imaging studies[J]. J Affect Disord, 292, 9–20
Hori, Y., Schaeffer, D. J., Gilbert, K. M., et al. (2020). Comparison of resting-state functional connectivity in marmosets with tracer-based cellular connectivity[J]. Neuroimage, 204, 116241
Thomas, S. A., Christensen, R. E., Schettini, E., et al. (2019). Preliminary analysis of resting state functional connectivity in young adults with subtypes of bipolar disorder[J]. J Affect Disord, 246, 716–726
Wang, J., Wang, Y., Wu, X., et al. (2020). Shared and specific functional connectivity alterations in unmedicated bipolar and major depressive disorders based on the triple-network model[J]. Brain Imaging Behav, 14(1), 186–199
Rai, S., Griffiths, K. R., Breukelaar, I. A., et al. (2021). Default-mode and fronto-parietal network connectivity during rest distinguishes asymptomatic patients with bipolar disorder and major depressive disorder[J]. Transl Psychiatry, 11(1), 547
Vargas, C., Lopez-Jaramillo, C., & Vieta, E. (2013). A systematic literature review of resting state network–functional MRI in bipolar disorder[J]. J Affect Disord, 150(3), 727–735
Zhong, Y., Wang, C., Gao, W., et al. (2019). Aberrant Resting-State Functional Connectivity in the Default Mode Network in Pediatric Bipolar Disorder Patients with and without Psychotic Symptoms[J]. Neurosci Bull, 35(4), 581–590
Lee, M. S., Anumagalla, P., Talluri, P., et al. (2019). Attentional engagement increases inferior frontal gyrus activity and mutes limbic activity in pediatric bipolar disorder: Meta-analyses of fMRI studies[J]. Prog Neuropsychopharmacol Biol Psychiatry, 91, 14–19
Gong, J., Chen, G., Jia, Y., et al. (2019). Disrupted functional connectivity within the default mode network and salience network in unmedicated bipolar II disorder[J]. Prog Neuropsychopharmacol Biol Psychiatry, 88, 11–18
Oertel-Knochel, V., Reinke, B., Matura, S., et al. (2015). Functional connectivity pattern during rest within the episodic memory network in association with episodic memory performance in bipolar disorder[J]. Psychiatry Res, 231(2), 141–150
Cao, W., Chen, H., Jiao, Q., et al.. Altered Functional Integration in the Salience and Default Mode Networks in Euthymic Pediatric Bipolar Disorder[J]. Neural Plast, 2020,2020:5853701.
Guo, Y., Wang, J., Jiao, Q., et al. (2021). Altered spatiotemporal consistency of corticolimbic circuitry in euthymic pediatric bipolar disorder[J]. Brain Imaging Behav, 15(3), 1290–1299
Guo, Y. B., Gao, W. J., Long, Z. L., et al. (2021). Shared and specific patterns of structural and functional thalamo-frontal disturbances in manic and euthymic pediatric bipolar disorder[J]. Brain Imaging Behav, 15(5), 2671–2680
Baek, J. H., Park, D. Y., Choi, J., et al. (2011). Differences between bipolar I and bipolar II disorders in clinical features, comorbidity, and family history[J]. J Affect Disord, 131(1–3), 59–67
Ma, Y., & MacDonald, I. A. (2021). Impact of Independent Component Analysis Dimensionality on the Test-Retest Reliability of Resting-State Functional Connectivity[J]. Brain Connect, 11(10), 875–886
Kaufman, J., Birmaher, B., Brent, D., et al. (1997). Schedule for Affective Disorders and Schizophrenia for School-Age Children-Present and Lifetime Version (K-SADS-PL): initial reliability and validity data[J]. J Am Acad Child Adolesc Psychiatry, 36(7), 980–988
Wechsler, D. (1999). Wechsler abbreviated scale of intelligence[M]. San Antonio: The Psychological Corporation
Young, R. C., Biggs, J. T., Ziegler, V. E., et al. (1978). A rating scale for mania: reliability, validity and sensitivity[J]. Br J Psychiatry, 133, 429–435
Wood, A., Kroll, L., Moore, A., et al. (1995). Properties of the mood and feelings questionnaire in adolescent psychiatric outpatients: a research note[J]. J Child Psychol Psychiatry, 36(2), 327–334
Smyth, C. (2003). The Pittsburgh Sleep Quality Index[J]. Dermatol Nurs, 15(2), 195–196
Beck, A. T., Kovacs, M., & Weissman, A. (1979). Assessment of suicidal intention: the Scale for Suicide Ideation[J]. J Consult Clin Psychol, 47(2), 343–352
Oldfield, R. C. (1971). The assessment and analysis of handedness: the Edinburgh inventory[J]. Neuropsychologia, 9(1), 97–113
Scarpina, F., & Tagini, S. (2017). The Stroop Color and Word Test[J]. Front Psychol, 8, 557
Demakis, G. J. (2004). Frontal lobe damage and tests of executive processing: a meta-analysis of the category test, stroop test, and trail-making test[J]. J Clin Exp Neuropsychol, 26(3), 441–450
Welch, L. W., Nimmerrichter, A., Gilliland, R., et al. (1997). "Wineglass" confabulations among brain-damaged alcoholics on the Wechsler Memory Scale-Revised visual reproduction subtest[J]. Cortex, 33(3), 543–551
Jafri, M. J., Pearlson, G. D., Stevens, M., et al. (2008). A method for functional network connectivity among spatially independent resting-state components in schizophrenia[J]. Neuroimage, 39(4), 1666–1681
Liao, W., Chen, H., Feng, Y., et al. (2010). Selective aberrant functional connectivity of resting state networks in social anxiety disorder[J]. Neuroimage, 52(4), 1549–1558
Merikangas, K. R., Jin, R., He, J. P., et al. (2011). Prevalence and correlates of bipolar spectrum disorder in the world mental health survey initiative[J]. Arch Gen Psychiatry, 68(3), 241–251
Gong, J., Wang, J., Qiu, S., et al. (2020). Common and distinct patterns of intrinsic brain activity alterations in major depression and bipolar disorder: voxel-based meta-analysis[J]. Transl Psychiatry, 10(1), 353
Menon, V. (2011). Large-scale brain networks and psychopathology: a unifying triple network model[J]. Trends Cogn Sci, 15(10), 483–506
Zhong, Y., Wang, C., Gao, W., et al. (2019). Aberrant Resting-State Functional Connectivity in the Default Mode Network in Pediatric Bipolar Disorder Patients with and without Psychotic Symptoms[J]. Neurosci Bull, 35(4), 581–590
Zhang, L., Opmeer, E. M., Ruhe, H. G., et al. (2015). Brain activation during self- and other-reflection in bipolar disorder with a history of psychosis: Comparison to schizophrenia[J]. Neuroimage Clin, 8, 202–209
Wu, G., Wang, Y., Mwansisya, T. E., et al. (2014). Effective connectivity of the posterior cingulate and medial prefrontal cortices relates to working memory impairment in schizophrenic and bipolar patients[J]. Schizophr Res, 158(1–3), 85–90
Zhou, H. X., Chen, X., Shen, Y. Q., et al. (2020). Rumination and the default mode network: Meta-analysis of brain imaging studies and implications for depression[J]. Neuroimage, 206, 116287
Buckner, R. L., & Vincent, J. L. (2007). Unrest at rest: default activity and spontaneous network correlations[J]. Neuroimage, 37(4), 1091–1096
Gong, J., Chen, G., Zhou, M., et al. (2020). Characteristics of temporal dynamics of intrinsic brain activity in unmedicated bipolar disorder with suicidality[J]. Aust N Z J Psychiatry, 54(11), 1115–1124
Altinay, M. I., Hulvershorn, L. A., Karne, H., et al. (2016). Differential Resting-State Functional Connectivity of Striatal Subregions in Bipolar Depression and Hypomania[J]. Brain Connect, 6(3), 255–265
Chen, G., Zhao, L., Jia, Y., et al. (2019). Abnormal cerebellum-DMN regions connectivity in unmedicated bipolar II disorder[J]. J Affect Disord, 243, 441–447
Luo, X., Chen, G., Jia, Y., et al. (2018). Disrupted Cerebellar Connectivity With the Central Executive Network and the Default-Mode Network in Unmedicated Bipolar II Disorder[J]. Front Psychiatry, 9, 705
Menon, V., & Uddin, L. Q. (2010). Saliency, switching, attention and control: a network model of insula function[J]. Brain Struct Funct, 214(5–6), 655–667
Seeley, W. W., Menon, V., Schatzberg, A. F., et al. (2007). Dissociable intrinsic connectivity networks for salience processing and executive control[J]. J Neurosci, 27(9), 2349–2356
Garcia-Garcia, I., Jurado, M. A., Garolera, M., et al. (2013). Alterations of the salience network in obesity: a resting-state fMRI study[J]. Hum Brain Mapp, 34(11), 2786–2797
Menon, V. (2011). Large-scale brain networks and psychopathology: a unifying triple network model[J]. Trends Cogn Sci, 15(10), 483–506
Lima, I., Peckham, A. D., & Johnson, S. L. (2018). Cognitive deficits in bipolar disorders: Implications for emotion[J]. Clin Psychol Rev, 59, 126–136
Lois, G., Gerchen, M. F., Kirsch, P., et al. (2017). Large-scale network functional interactions during distraction and reappraisal in remitted bipolar and unipolar patients[J]. Bipolar Disord, 19(6), 487–495
Phillips, M. L., Drevets, W. C., Rauch, S. L., et al. (2003). Neurobiology of emotion perception I: The neural basis of normal emotion perception[J]. Biol Psychiatry, 54(5), 504–514
Jamieson, D., Broadhouse, K. M., Lagopoulos, J., et al. (2020). Investigating the links between adolescent sleep deprivation, fronto-limbic connectivity and the Onset of Mental Disorders: a review of the literature[J]. Sleep Med, 66, 61–67
Rosen, G. M., Bendel, A. E., Neglia, J. P., et al. (2003). Sleep in children with neoplasms of the central nervous system: case review of 14 children[J]. Pediatrics, 112(1 Pt 1), e46–e54
Cooch, N. K., Stalnaker, T. A., Wied, H. M., et al. (2015). Orbitofrontal lesions eliminate signalling of biological significance in cue-responsive ventral striatal neurons[J]. Nat Commun, 6, 7195
Teghil, A., Bonavita, A., Guariglia, C., et al. (2021). Commonalities and specificities between environmental navigation and autobiographical memory: A synthesis and a theoretical perspective[J]. Neurosci Biobehav Rev, 127, 928–945
Phillips, M. L., Travis, M. J., Fagiolini, A., et al. (2008). Medication effects in neuroimaging studies of bipolar disorder[J]. Am J Psychiatry, 165(3), 313–320

Download PDF

Journal Publication

published 31 May, 2024

Read the published version in Psychiatry Research: Neuroimaging →

Version 1

posted

You are reading this latest preprint version

Abnormal functional connectivity of the intrinsic networks in adolescent bipolar I versus bipolar II disorder

Status:

Journal Publication

Version 1

Abstract

Figures

1 Introduction

2 Materials And Methods

2.1 Participants

2.2. Emotional and cognitive features

2.2.1. Mood assessments and Demographics

2.2.2 Cognitive Function Test

2.3 MRI Acquisition

2.4 Data Preprocessing

2.5 Group ICA and recognition of the intrinsic networks

2.6 Statistical Analysis

3 Result

3.1 Subjects’ Demographics, Clinical Features and Cognitive Functions

3.2 Between-Group ICA Analyses

3.3 Correlation Analysis

4 Discussion

Declarations

References

Status:

Journal Publication

Version 1