Identification of Common Thalamocortical Dysconnectivity in Four Major Psychiatric Disorders

Abstract

Background

Recent genetic and imaging analyses of large datasets suggested that common biological substrates exist across psychiatric diagnoses. Functional connectivity (FC) abnormalities of thalamocortical circuits were consistently found in patients with schizophrenia but have been less studied in other major psychiatric disorders. This study aimed to examine thalamocortical FC in 4 major psychiatric disorders to identify the common connectivity abnormalities across major psychiatric disorders.

Methods

This study recruited 100 patients with schizophrenia, 100 patients with bipolar I disorder, 88 patients with bipolar II disorder, 100 patients with major depressive disorder, and 160 healthy controls (HCs). Each participant underwent resting functional magnetic resonance imaging. The thalamus was used to derive FC maps, and group comparisons were made between each patient group and HCs using an independent-sample t test. Conjunction analysis was used to identify the common thalamocortical abnormalities among these 4 psychiatric disorders.

Results

The 4 groups of patients shared a similar pattern of thalamocortical dysconnectivity characterized by a decrease in thalamocortical FC with the dorsal anterior cingulate, anterior prefrontal cortex and inferior parietal cortex. The groups also showed an increase in FC with the postcentral gyrus, precentral gyrus, superior temporal cortex, and lateral occipital areas. Further network analysis demonstrated that the frontoparietal regions showing hypoconnectivity belonged to the salience network.

Conclusion

Our findings provide FC evidence that supports the common network hypothesis by identifying common thalamocortical dysconnectivities across 4 major psychiatric disorders. The network analysis also supports the cardinal role of salience network abnormalities in major psychiatric disorders.

Introduction

Current diagnostic systems classify psychiatric disorders into different categories according to the set of symptoms. However, recent genetic and imaging analyses of large datasets suggested that major psychiatric disorders may share a common neural substrate. Two important genetic studies^1,2 attempted to identify specific variants shared between 5 psychiatric disorders (autism spectrum disorder, attention deficit hyperactivity disorder, bipolar disorder [BD], major depressive disorder [MDD], and schizophrenia [SZ]) using the dataset from the Psychiatric Genomics Consortium. The results suggested a shared genetic etiology for psychiatric disorders and encouraged the investigation of common neural substrates for related disorders. Several meta-analyses of magnetic resonance imaging (MRI) findings have attempted to identify the common neural substrates. An investigation based on a meta-analysis of 193 voxel-based morphometry (VBM) studies comprising 15 892 individuals across 6 diverse diagnostic groups (SZ, BD, depression, addiction, obsessive-compulsive disorder, and anxiety) with matched control groups found that gray matter (GM) loss converged across diagnoses in 3 regions: the dorsal anterior cingulate cortex (dACC) and the bilateral insula.³ These 3 regions were suggested to form a tightly interconnected network, the salience network, by functional connectivity (FC) analysis. Similar findings were also found by a meta-analysis of a recent functional MRI (fMRI) study.⁴ This concordance provides an organizing model that emphasizes the importance of shared neural substrates across psychopathologies, despite the likelihood of diverse etiologies.

One important method for exploring the neural abnormalities of various psychiatric disorders is to evaluate abnormalities in functional connectomics by resting-state functional connectivity MRI.^5,6 Although 1 meta-analysis of published MRI data from more than 20 000 subjects and 26 different brain disorders found that the common lesions across all brain disorders were more likely to be located in hubs of the normal brain connectome,⁷ the hypothesis that patients with various psychiatric disorders may share common network abnormalities was not directly investigated. In this study, we sought to test the hypothesis that cross-diagnostic connectivity abnormalities exist by investigating the FC of the thalamus. There are several rationales for focusing on thalamocortical connectivities. First, previous attempts to explain the diverse symptoms of various psychiatric disorders with a single underlying mechanism or structure have focused on the thalamus, which has been suggested to play an important role in sensory relay as well as cognitive monitoring.⁸ Second, several important FC studies have shown a consistent pattern of aberrant corticothalamic connections in SZ, which is characterized by increased thalamocortical FC with the primary sensorimotor and decreased FC with the prefrontal cortex.^9–12 Thalamocortical dysconnectivity is present in both chronic and early stages of psychosis.¹³ Increased primary sensorimotor connectivity was also found in high-risk participants and especially in those transitioning to psychosis.¹⁴ Furthermore, ketamine-induced psychosis in healthy participants also induced a similar hypoconnectivity of the primary sensorimotor cortex.¹⁰ Third, a similar pattern of thalamocortical FC abnormalities was also found in other studies of patients with BD^10,15 and MDD.^16–18 Although these findings suggest the possibility of shared thalamocortical connectivity abnormalities in various psychiatric disorders, previous studies rarely recruited patients diagnosed with all these major psychiatric disorders in the same study, and it is unclear if these disorders share a similar pattern of thalamocortical dysconnectivity.

This study aimed to evaluate thalamocortical connectivity in 4 major psychiatric disorders, namely, SZ, BD-I, BD-II, and MDD, and to test the hypothesis that common thalamocortical FC abnormalities exist that extend across diagnoses. We used conjunction analysis to identify the pattern of common thalamocortical FC abnormalities in these 4 groups of patients. On the basis of the meta-analysis of VBM findings³ and our previous FC studies of salience network dysfunction in SZ,^19,20 we also hypothesized that the dACC is a key structure for analysis and that several regions involved in hypoconnectivity belong to the dACC-centered salience network.

Materials and Methods

Participants

The groups included 100 patients with SZ, 100 patients with BD-I, 88 patients with BD-II, 100 patients with MDD, and 100 healthy controls (HCs). Patients with a diagnosis of schizoaffective disorders were not included in this study. All of the participants were recruited from outpatient and inpatient units of the Taipei Veterans General Hospital in Taiwan (table 1). Structured clinical interviews based on the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition,²¹ confirmed the diagnoses. The patients were also evaluated using the Mini International Neuropsychiatric Inventory Plus (MINI).²² The participants were screened to exclude those with the following conditions: (1) substance abuse or dependence issues over the preceding 6 months; (2) a history of head injuries that resulted in sustained loss of consciousness, cognitive sequelae, or both; and (3) neurological illnesses or any other disorder that affects cerebral metabolism. HCs were recruited via advertisements. An experienced psychiatrist used the MINI to screen and exclude the candidates with major psychiatric illnesses. In addition, candidates with a history of first-degree relatives with axis-I disorders, including SZ, MDD, and BD, were excluded. The clinical status of the patients with SZ was characterized using the Positive and Negative Syndrome Scale (PANSS).²³ The clinical assessments of the BD-I, BD-II, and MDD patients included the Young Mania Rating Scale and Montgomery–Åsberg Depression Rating Scale, but only a certain number of these patients received complete ratings (table 1). These patients were receiving treatment with various atypical antipsychotics, antidepressants, and mood stabilizers before participating in the experiment (detailed in supplementary table S1).

Table 1.

Demographic Data of Participants in This Study

	SZ	BD-I	BD-II	MDD	HC	P
	n = 100	n = 100	n = 88	n = 100	n = 160
Sex (M/F)	50/50	44/56	29/59	37/63	80/80	.04
Age (mean ± SD)	35.5 ± 9.2	39.5 ± 10.5	39.8 ± 11.3	39.0 ± 11.4	34.3 ± 9.3	<.001
Education (years)	13.3 ± 2.6	14.2 ± 3.8	14.5 ± 3.5	13.5 ± 3.4	15.3 ± 1.7	<.001
Age at onset	22.7 ± 7.1	25.5 ± 9.5	26.6 ± 11.4	31.1 ± 11.7		<.001
Length of illness	12.8 ± 9.2	13.6 ± 9.1	12.9 ± 10.7	7.8 ± 7.1		<.001
		n = 49	n = 42	n = 75
MADRS		7.5 ± 9.1	13.1 ± 11.0	26.7 ± 9.5
YMRS		3.2 ± 4.5	2.6 ± 3.9	2.0 ± 2.0
	n = 100
PANSS total	63.1 ± 19.0
Positive	14.2 ± 5.4
Negative	17.0 ± 5.6
General psychopathology	32.2 ± 10.2

Note: SZ, schizophrenia; BD-I, bipolar I disorder; BD-II, bipolar II disorder; MDD, major depressive disorder; HC, healthy control; MADRS, Montgomery–Åsberg Depression Rating Scale; YMRS, Young Mania Rating Scale; PANSS, Positive And Negative Syndrome Scale for Schizophrenia; SD, standard deviation.

MRI Image Acquisition

MRI images were acquired using a 3.0 Tesla GE Discovery 750 whole-body high-speed imaging device with an 8-channel high-resolution brain coil. Head stabilization was achieved with cushioning, and all the participants wore earplugs (29 dB rating) to attenuate the noise. Automated shimming procedures were performed, and scout images were obtained. The resting-state functional images were collected using a gradient echo T2* weighted sequence (repetition time [TR]/echo time [TE]/Flip = 2500 ms/30 ms/90°). Forty-seven contiguous horizontal slices parallel to the inter-commissural plane (voxel size: 3.5 × 3.5 × 3.5 mm) were acquired and interleaved. These slices covered the cerebellum of each participant. During the functional scans, the participants were instructed to remain awake with their eyes open (each scan lasted 8 min and 24 s across 200 time points). In addition, a high-resolution structural image was acquired in the sagittal plane using a high-resolution sequence (TR = 2530 ms, echo spacing = 7.25 ms, TE = 3 ms, flip angle = 7°) and an isotropic 1 mm voxel (field of view 256 × 256).

Quality Check of the Functional Image

The signal-to-noise ratio (SNR) was computed by obtaining the mean signal and standard deviation for a given slice across the Blood Oxygen-Level Dependent run while excluding all non-brain voxels). We provide the SNR for all patients in supplementary table S2. There was no significant difference among the participant groups. Regarding head motion during image acquisition, we used the method of scrubbing within regression (spike regression) suggested by a previous study²⁴ to minimize the effect of head motion on the measurement of FC. This method identifies “bad” time points using a threshold of framewise displacement > 0.2 mm as well as 1 back and 2 forward neighbors (as reported in Power et al²⁵ and then models each “bad” time point as a separate regressor in the regression models.^26,27 We also calculated the number of contaminated volumes in each group, and there was no significant difference between the groups (supplementary table S2).

Analysis of Resting-State FC

FC Preprocessing.

All preprocessing was performed using the Data Processing Assistant for Resting-State fMRI (DPARSF²⁸; http://www.restfmri.net), which is based on Statistical Parametric Mapping (SPM12; http://www.fil.ion.ucl.ac.uk/spm) and Resting-State fMRI Data Analysis Toolkit (REST²⁹; http://www.restfmri.net). The functional scans received slice-timing correction and the time series of images for each subject was realigned using a 6-parameter (rigid body) linear transformation with a 2-pass procedure. Individual structural images were co-registered to the mean functional image after realignment using a 6 degrees-of-freedom linear transformation without resampling. Although the transformed structural images were then segmented into GM, white matter (WM), and cerebrospinal fluid30 all the voxels within the brain were included for FC analysis. Additional preprocessing steps, described in previous reports,³¹ were used to prepare the data for FC analysis: (1) spatial smoothing using a Gaussian kernel (6 mm full width at half maximum), (2) temporal filtering (0.009 Hz < f < 0.08 Hz), and (3) removal of spurious or nonspecific sources of variance by regression of the following variables: (a) 6 head motion parameters and autoregressive models of motion—6 head motion parameters, 6 head motion parameters one time point before, and the 12 corresponding squared items³² (Friston 24-parameter model); (b) the mean whole-brain signal; (c) the mean signal within the lateral ventricles; and (d) the mean signal within a deep WM region of interest (ROI). As we mentioned in the “Quality Check of the Functional Image” section, the regressors by the method of scrubbing within regression (spike regression) were also included to minimize the effect of head motion on the measurement of FC. The regression of each of these signals was computed simultaneously, and the residual time course was retained for the correlation analysis.

FC Analyses.

We adopted 1 ROI that covered the bilateral thalamus according to the atlas (figure 1a). The FC maps of the thalamus were identified based on correlations of low-frequency fMRI fluctuations with the ROIs. Fisher’s r-to-z transformation was used to convert correlation maps into z maps.³¹ The z-transformed maps of these participants were compared between the patient groups and the controls by an independent sample t test using age, sex, and education as the covariates of no interest. We used an uncorrected threshold of P < .001 for the initial voxel-wise comparisons. To correct for multiple comparisons, a Monte Carlo simulation with 10 000 times was performed by AlphaSim of AFNI 18.1.18.³³ Only the clusters with a significance threshold of P < .05 at the cluster level (a minimum cluster size of 36 in this study) were selected for conjunction analysis. We adopted a conjunction analysis to find the common areas that showed a significant FC difference in thalamocortical connectivity in all 4 groups of psychiatric patients. With regard to these common areas, we did a correlation analysis to understand the effect of age, duration of illness, and symptoms on FC of these regions by Pearson correlation coefficient. The analysis was exploratory and the structures with P < .05 (uncorrected for multiple comparison) were reported. We also performed a supplementary analysis to explore the difference in thalamocortical FC between the patients in each diagnostic group. The clusters with a significance threshold of P < .05 at the cluster level were reported.

Fig. 1.

(a) The a priori region of interest (ROI) in the thalamus used to drive the thalamocortical connectivity in this study. (b) The thalamocortical dysconnectivity in patients with SZ, BD-I, BD-II, and MDD. The maps were derived from between-group comparisons of each group of patients and healthy controls demonstrated with the threshold of P < .001 (uncorrected) at the voxel level. SZ, schizophrenia; BD-I, bipolar I disorder; BD-II, bipolar II disorder; MDD, major depressive disorder, HC, healthy control; L, left; R, right; A, anterior.

Results

The demographic data and clinical ratings of the participants were shown in table 1. The age, sex, and education level were not fully matched in the sample and were added as covariates of no interest during between-group comparisons of thalamocortical connectivity.

Compared to HCs, the 4 groups of patients showed a similar pattern of thalamocortical FC abnormalities, characterized by a decrease in FC with the dACC, posterior cingulate gyrus, frontal poles, and inferior parietal cortex and an increase in FC with the postcentral gyrus, precentral gyrus, superior temporal cortex, and lateral occipital areas (figure 1b). The conjunction analysis further identified the common cortical structures showing significant hypo- or hyperconnectivity in all 4 groups of patients (table 2, figure 2a). The effect sizes of each comparison were shown in supplementary table S3 and supplementary figure S1.

Table 2.

The Structures Showing Significant Functional Connectivity Abnormalities in the Conjunction Analysis

FC Change	Cluster	MNI Coordinate			Harvard-Oxford Cortical
↑↓	Size	x	y	z	Structural Atlas
↓	913	2	23	39	R. cingulate gyrus, anterior division
↓	208	2	−27	25	R. cingulate gyrus, posterior division
↓	519	38	48	24	R. frontal pole
↓	136	−41	47	21	L. frontal pole
↓	81	51	−51	54	R. supramarginal gyrus, posterior division
↑	1529	−43	−25	52	L. postcentral gyrus
↑	1408	48	−20	50	R. Postcentral Gyrus
↑	1003	−45	−24	11	L. Heschl’s gyrus (includes H1 and H2)
↑	621	41	−17	13	R. Heschl’s gyrus (includes H1 and H2)
↑	1332	3	−28	64	R. precentral gyrus
↑	1034	61	−15	−5	R. superior temporal gyrus
↑	857	−58	−13	−7	L. superior temporal gyrus
↑	810	49	−63	−4	R. lateral occipital cortex
↑	480	−47	−71	2	L. lateral occipital cortex

Note: MNI, Montreal Neurological Institute; FC, functional connectivity; L, left; R, right.

Fig. 2.

(a) The conjunction map of thalamocortical dysconnectivity in 4 groups of patients. The common structures showing a decrease in FC include the dorsal anterior cingulate, middle frontal gyrus, and inferior parietal cortex. The common regions showing an increase in FC include the primary and supplementary motor cortex, superior and middle temporal cortex, and lateral occipital areas. (b) The common structures showing a decrease in FC overlapped with the FC map of the dACC, suggesting that a thalamocortical disconnection within the dACC-centered salience network. dACC, dorsal anterior cingulate cortex; FC, functional connectivity; L, left; R, right.

Correlation analyses were performed to investigate the effect of age and duration of illness on the thalamocortical FC in these structures within each diagnostic group (detailed results in supplementary tables S4 and S5, and supplementary figure S2). The age was negatively correlated with the thalamocortical FC of the dACC in BD-I patients (r = −0.32, P = .001) and the thalamocortical FC of left postcentral gyrus in BD-I (r = −0.27, P = .006) and BD-II (r = −0.22, P = .037). The duration of illness was negatively correlated with thalamocortical FC of the dACC in BD-I (r = −0.22, P = .023). In BD-II, the duration of illness was also negatively correlated with thalamocortical FC of left superior temporal cortex (r = −0.27, P = .009) and posterior cingulate gyrus (r = 0.30, P = .004). In summary, the age and duration of illness had a more significant effect on the thalamocortical FC in BD-I and BD-II patients.

We also explored the correlation between FC abnormalities and symptom severity in the group of patients with SZ (detailed results in supplementary table S6). The most significant finding was that the thalamocortical FC of left superior temporal cortex was positively correlated with total PANSS scores (r = 0.32, P = .002), positive subscores (r = 0.309, P = .002), negative symptoms subscores (r = 0.327, P = .001), and general pathology subscores (r = 0.29, P = .008). Because the thalamocortical FC of left superior temporal cortex was significantly higher in SZ, the findings suggested that higher clinical severity was associated with more severe thalamocortical FC abnormalities of left superior temporal cortex in SZ.

The detailed results of thalamocortical FC difference between each diagnostic group were shown in table 3. There was no significant FC difference between SZ and BD. SZ had a significantly higher thalamocortical FC in precentral gyrus than BD-II and MDD. BD-I had a significantly higher thalamocortical FC in lateral occipital cortex than BD-II and a significantly higher FC in postcentral gyrus than MDD. The major difference between BD-II and MDD was lower thalamocortical FC in the middle frontal gyrus in patients with BD-II. In summary, SZ and BD-I patients showed a similar level of thalamocortical abnormalities and more severe FC abnormalities than BD-II and MDD in several common cortical structures showing significant hypo- or hyperconnectivity identified in the conjunction analysis.

Table 3.

The Structures Showing Significant Functional Connectivity Differences Between the 4 Groups of Patients

		FC Change	Cluster	MNI Coordinate			Harvard-Oxford Cortical
Comparisons		↑↓	Size	x	y	z	Structural Atlas
SZ vs	BD-I	↑	—	—	—	—
		↓	—	—	—	—
	BD-II	↑	122	−30	−14	52	L. precentral gyrus
		↑	56	−24	14	4	L. putamen
		↑	68	−18	38	−6	L. frontal medial cortex
		↓	663	6	−36	−6	R. lingual gyrus
	MDD	↑	158	−32	−14	52	L. precentral gyrus
		↑	61	40	−18	44	R. postcentral gyrus
		↑	55	24	−8	38	R. precentral gyrus
		↑	60	−50	−38	54	L. supramarginal gyrus
		↓	176	−4	−30	−12	L. brain-stem
		↓	70	−12	−18	8	L. thalamus
		↓	60	16	−18	8	R. thalamus
		↓	207	34	−62	2	R. lateral occipital cortex
BD-I vs	BD-II	↑	78	−58	−64	18	L. lateral occipital cortex
		↑	86	62	−40	20	R. supramarginal gyrus
		↓	64	10	−64	−54	R. cerebellar VIIIb
		↓	105	−6	−8	8	L. thalamus
	MDD	↑	54	56	−6	34	R. postcentral gyrus
		↓	55	16	−28	18	R. thalamus
BD-II vs	MDD	↑	103	−36	−70	36	L. lateral occipital cortex
		↓	677	−22	38	12	L. middle frontal gyrus
		↓	221	40	−92	0	R. occipital pole
		↓	158	60	−38	24	R. supramarginal gyrus

Note: SZ, schizophrenia; BD-I, bipolar I disorder; BD-II, bipolar II disorder; MDD, major depressive disorder; FC, functional connectivity; L, left; R, right.

Discussion

In this study, we investigated the FC of the thalamus to identify the common FC abnormalities in 4 major psychiatric disorders. Consistent with our hypothesis, we found that the 4 disorders shared the same pattern of thalamocortical dysconnectivities, characterized by a decrease in FC in several frontal and parietal regions and an increase in FC with the postcentral gyrus, precentral gyrus, temporal, and occipital regions. The pattern of FC abnormalities is largely consistent with previous FC studies evaluating SZ.

One major pattern of thalamocortical dysconnectivity was hypoconnectivity in several frontal and parietal regions in the 4 major psychiatric disorders. Although previous FC studies have also reported a decreased FC with inferior parietal regions,^11,15 previous FC studies of SZ focused on the pattern of thalamo-frontal hypoconnectivity. However, the possibility that these frontoparietal regions belong to the same network was not further explored. We used the ROI in the dACC derived from the conjunction analysis to calculate the anatomical component of the salience network and found that the cortical component of the salience network overlapped well with the regions of the bilateral middle frontal gyrus, inferior parietal cortex in the conjunction analysis (figure 2b). Our previous studies also found a cortico-striatal and corticothalamic disconnection within the salience network in SZ.^19,20 The current study extends these findings and provides evidence for an important role of salience network dysfunction in other major psychiatric disorders.

Our findings support the cardinal role of the salience network in major psychiatric disorders. The salience network is one of the 3 important cognitive-related large-scale networks defined with the development of resting fMRI studies: the default mode network, the salience network, and the cognitive control network.^34–36 These 3 networks were proposed to play an important role in neuropsychiatry in the tri-network model of neuropsychiatric illness.³⁷ In this model, deficits in access, engagement, and disengagement in large-scale neurocognitive networks are suggested to play a prominent role in various neuropsychiatric disorders. The model further emphasizes the surprising parallels that are beginning to emerge across psychiatric and neurological disorders. Although the default mode network was consistently found to be involved in Alzheimer’s dementia, our results suggest that the salience network plays a more important role in the 4 examined adult-onset psychiatric disorders. This finding is also in accordance with the meta-analysis suggesting the dACC-centered network may be the core common deficit in psychiatric disorders.

Another major finding of this study was the increase in FC with the postcentral gyrus, precentral gyrus, temporal, and occipital regions, which are important for auditory and visual processing. The finding is similar to those in FC studies using the thalamus as the ROI in patients with SZ.¹⁰ The relatively small portion of the temporal and occipital regions showing hyperconnectivity may explain why the findings of temporal and occipital regions were not demonstrated in previous FC studies using large cortical ROIs.⁹ However, increased corticothalamic FC with the temporal cortex¹² and occipital regions¹¹ was also reported in several previous FC studies. Among these regions, the postcentral gyrus and precentral gyrus had the largest cluster size and was better investigated in previous studies. We suggest that hyperexcitability of the primary motor cortex demonstrated by transcranial magnetic stimulation may contribute to the FC findings.³⁸ A recent review of cortical excitability studies in various psychiatric disorders suggested a general alteration in motor cortical inhibition in mental illness, rather than a disease-specific change.³⁹ A meta-analysis also demonstrated that inhibitory deficits in the motor cortex are a ubiquitous finding across obsessive-compulsive disorder, MDD, and SZ.⁴⁰ Deficits in γ-aminobutyric acid inhibitory neurotransmission were implicated in the inhibitory deficits of various psychiatric disorders.

Several methodological limitations and alternative conceptualizations of our findings merit consideration. First, our patient groups had chronic exposure to various psychotropic medications. Previous studies have evaluated the effect of antipsychotics,^17,41,42 antidepressants,^43–45 lithium,⁴⁶ and mood stabilizers⁴⁷ on FC of the brain, and we cannot exclude the possibility that the effects of medications contributed to our findings. However, previous FC studies of high-risk patients with SZ have suggested that thalamocortical connectivity occurs without the use of antipsychotics.¹⁴ We performed additional analyses with the use of antipsychotics, antidepressants, and mood stabilizers as covariates, and the results were not significantly changed after adding these covariates. Second, the demographic properties of participants were not fully matched in this study. Although age, sex, and education level were added as covariates of no interest during between-group comparisons, their effects on FC may not be the same for different diagnostic groups and they may confound our findings in this study. Third, we adopted the procedure of global signal regression for image preprocessing because previous studies suggested that the procedure effectively removed noise related to motion or physiological signals⁴⁸ and also increased the spatial specificity in seed-based functional connectivity analysis.⁴⁹ However, it was also proposed that the procedure may induce artificial anticorrelation between different brain regions.⁵⁰ We did an analysis without the use of global signal regression and found that the findings related to prefrontal cortex were absent. The result is consistent with a recent analysis suggesting that some case–control difference emerged only after global signal regression was adopted.⁵¹ Therefore, we should be more cautious in interpreting our results with global signal regression. Forth, we did not analyze the correlations between cognitive function and thalamocortical dysconnectivity because not all of our participants received full cognitive and functional assessments. The relationship between thalamocortical dysconnectivity and cognitive impairments in these 4 groups of patients was not clear in this study and required further investigations. At last, this study focused on the FC of the thalamus because previous studies of thalamocortical FC in SZ showed the most consistent results, and other regions were not explored. It is possible that the 4 diagnostic groups also share FC abnormalities in other neural structures.

In conclusion, this study investigated the thalamocortical FC of 4 major psychiatric disorders and identified the common FC abnormalities across diagnoses. Our findings provide FC evidence to support the common network hypothesis and support the cardinal role of salience network abnormalities in major psychiatric disorders.

Funding

Taipei Veterans General Hospital (V99C1-040, V101C1-159, V104C-039); National Science Council (NSC99-2628-B-010-021-MY2).