
==== Front
Neuroimage Clin
Neuroimage Clin
NeuroImage : Clinical
2213-1582
Elsevier

S2213-1582(24)00091-3
10.1016/j.nicl.2024.103652
103652
Regular Article
Chronic hypercortisolism disrupts the principal functional gradient in Cushing’s disease: A multi-scale connectomics and transcriptomics study
Shang Guosong ab1
Zhou Tao ac1
Yu Xiaoteng de1
Yan Xinyuan f
He Kunyu ab
Liu Bin ab
Feng Zhebin g
Xu Junpeng ab
Zhang Yanyang sjwkzyy@163.com
ac⁎
Yu Xinguang yuxinguang_301@163.com
abc⁎
a Department of Neurosurgery, The First Medical Centre of Chinese PLA General Hospital, Beijing, China
b Chinese PLA Medical School, Beijing, China
c Neurosurgery Institute, Chinese PLA General Hospital, Beijing, China
d Department of Urology, Peking University First Hospital, Beijing, China
e Institute of Urology, Peking University, National Urological Cancer Center, Beijing, China
f Department of Psychiatry, University of Minnesota Medical School, Minneapolis, MN, USA
g Department of Neurosurgery, PLA 942 Hospital, Yinchuan, Ningxia, China
⁎ Corresponding authors at: Department of Neurosurgery, Chinese PLA General Hospital, 28 Fuxing Road, Haidian District, Beijing 100853, China. sjwkzyy@163.comyuxinguang_301@163.com
1 These authors equally contributed to this work.

10 8 2024
2024
10 8 2024
43 10365212 5 2024
22 7 2024
6 8 2024
© 2024 The Author(s)
2024
https://creativecommons.org/licenses/by-nc-nd/4.0/ This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
Highlights

• Chronic hypercortisolism induces convergent alterations along the principal gradient.

• The aberrant principal gradient spans multiple cognitive functional domains.

• Elevated cortisol levels may exacerbate abnormalities in the principal gradient.

• Synaptic dysfunction underlies principal gradient abnormalities.

Cushing’s disease (CD) represents a state of cortisol excess, serving as a model to investigate the effects of prolonged hypercortisolism on functional brain. Potential alterations in the functional connectome of the brain may explain frequently reported cognitive deficits and affective disorders in CD patients. This study aims to elucidate the effects of chronic hypercortisolism on the principal functional gradient, which represents a hierarchical architecture with gradual transitions across cognitive processes, by integrating connectomics and transcriptomics approaches. Utilizing resting-state functional magnetic resonance imaging data from 140 participants (86 CD patients, 54 healthy controls) recruited at a single center, we explored the alterations in the principal gradient in CD patients. Further, we thoroughly explored the underlying associative mechanisms of the observed characteristic alterations with cognitive function domains, biological attributes, and neuropsychiatric representations, as well as gene expression profiles. Compared to healthy controls, CD patients demonstrated changes in connectome patterns in both primary and higher-order networks, exhibiting an overall converged trend along the principal gradient axis. The gradient values in CD patients’ right prefrontal cortex and bilateral sensorimotor cortices exhibited a significant correlation with cortisol levels. Moreover, the cortical regions showing gradient alterations were principally associated with sensory information processing and higher-cognitive functions, as well as correlated with the gene expression patterns which involved synaptic components and function. The findings suggest that converged alterations in the principal gradient in CD patients may mediate the relationship between hypercortisolism and cognitive impairments, potentially involving genes regulating synaptic components and function.

Keywords

Cushing’s disease
Resting-state fMRI
Principal gradient
Brain networks
Gene expression
==== Body
pmc1 Introduction

Cushing's disease (CD), also known as adrenocorticotropic hormone (ACTH) – secreting pituitary adenoma, is characterized by chronic exposure to excess endogenous cortisol, providing a natural model for investigating the functional brain alterations under the effects of prolonged cortisol exposure (Liu et al., 2022, Papakokkinou and Ragnarsson, 2023). Cognitive dysfunctions and affective disorders are crucial clinical features in CD patients. Despite successful treatment interventions, these manifestations may persist, thereby substantially influencing the patients' quality of life (Ahn et al., 2021, Piasecka et al., 2020a, Webb and Valassi, 2022). Functional connectome disruptions serve as important neuroimaging markers for cognitive and emotional processing abnormalities in these patients, with significant implications for clinical treatment prospects (Brown et al., 2014, Jin et al., 2014, Li et al., 2022, Seeley et al., 2007, Stomby et al., 2019, Van Der Werff et al., 2015, Wang et al., 2019). However, previous studies on CD utilizing resting-state fMRI (rs-fMR) have primarily concentrated on functional connectome analyses that parcel the cortex into distinct networks of strongly interlinked regions, seeking to reveal differences between CD patients and healthy controls (HCs), with the aim to elucidate the effects of chronic hypercortisolism on functional brain. In fact, during cognitive processing from simple to complex, the brain does not operate as individual isolated regions or as a collection of fragmented networks. Instead, the brain operates as a highly integrated system that exhibits a hierarchical architecture with gradual transitions across cognitive processes (Margulies et al., 2016). The hierarchical architecture represents a fundamental organizational principle for brain which enables the efficient coding and information integration, with gradual shifts in functional connectivity profiles across multiple dimensions (Hong et al., 2020). One key feature dimension is the principal functional connectome gradient, representing the shift from unimodal to transmodal regions that accounts for the largest variation in brain functional connectivity (Margulies et al., 2016). Unimodal areas, including primary visual, somatosensory, and auditory cortices are anchored at one extreme of this gradient. The other extreme resides transmodal association areas of the default mode network (DMN). Situated between the two extremes are multimodal integration regions. This hierarchical arrangement potentially corresponds to the stratified structure of information processing, facilitating the progressive transformation of sensory inputs into higher-order abstract cognitive representations (Margulies et al., 2016). Moreover, the principal gradient demonstrates robust concordance with microstructural characteristics, semantic processing capabilities, and the temporal hierarchy inherent in the human brain (Huntenburg et al., 2018). This striking alignment underscores the fundamental importance of the principal gradient for the organizational principles of human cortical networks.

The abnormalities in the principal gradient of psychiatric disorders such as depression, schizophrenia, and autism are well-established (Dong et al., 2023, Hong et al., 2019, Xia et al., 2022, Xia et al., 2022). However, the specific effects of chronic hypercortisolism on the principal gradient have yet to be thoroughly investigated. Notably, prior studies in the functional connectome of CD have primarily emphasized the identification of aberrant topologies, while neglecting the underlying mechanisms that potentially contribute to these deviations. Recent evidences show the principal gradient mirrors topographic gene expression patterns, implicating its potential role in bridging cognitive functions and gene regulation (Hawrylycz et al., 2012, Huntenburg et al., 2018). In summary, investigating the principal gradient in CD patients can not only enhance our comprehensive understanding of macro-level characteristics of brain functional connectome under chronic hypercortisolism, but also provide insights into potential micro-level mechanisms regulating these characteristics. Motivated by theoretical foundations, this study examined whether and how alterations occur along the principal gradient in CD patients, as well as the underlying mechanisms and clinical implications of these alterations. The primary research design and analytical procedure are schematically depicted in Fig. 1.Fig. 1 Study design and analysis methodology. Gradient calculation. Based on the extracted time courses of rs-fMRI data from 140 participants, we constructed functional connectivity matrices at the voxel level (16482 × 16482 nodes). Subsequently, normalized angle matrices were derived by calculating cosine similarities and applying normalization procedures. The dimensionality reduction algorithm named 'diffusion map embedding' was utilized to capture various gradient components from matrices. Comparative analysis. After screening the explain ratios of multiple components, the principal gradients were identified for further investigation. A comprehensive statistical comparison was then conducted between the CD and HC groups to pinpoint any abnormal brain regions in CD patients. Correlation analysis. We further conducted analyses to elucidate the intricate relationships between the observed characteristic alterations of principal gradient and meta-analytic cognitive functions, clinical biochemical indices, neuropsychiatric scales, and gene expression profiles in CD patients.

2 Methods

2.1 Participants

A cohort of 141 participants was recruited a single center from May 2017 to June 2023, including 86 consecutive patients diagnosed with active CD scheduled for transsphenoidal surgery, as well as 55 HCs without any history of neuropsychiatric illness or cognitive deficits. According to the clinical practice guideline (Fleseriu et al., 2021, Nieman et al., 2015), the diagnosis of active CD was conducted by experienced endocrinologists and further confirmed through postsurgical pathology. HCs were recruited from the local community, and rigorous screening was conducted to ensure their absence of any history of psychopathological abnormalities. Additionally, all participants were confirmed to be right-handed. Vision and auditory sensation were also assessed and found to be normal. The study was approved by the Ethics Committee of the Chinese PLA General Hospital (S2021-677-01), and written informed consent was obtained from each participant.

2.2 Evaluation of neuroendocrine, neuropsychological and neuropsychiatric measures

We measured the levels of 24-hour urinary free cortisol (24 h-UFC, nmol/24 h), serum cortisol (nmol/L) and ACTH (pmol/L) at three specific time points (0 a.m., 8 a.m., and 4 p.m.) in all active CD patients. Additionally, serum cortisol and ACTH levels at 8 a.m., as well as 24 h-UFC, were also measured in HCs. Moreover, all participants underwent a comprehensive neuropsychological assessment, which encompassed the administration of the following: Self-Rating Depression Scale (SDS) (Zung, 1965), Self-Rating Anxiety Scale (SAS) (Zung, 1971), Mini-Mental State Examination (MMSE) (Folstein et al., 1975), Montreal Cognitive Assessment (MoCA) (Nasreddine et al., 2005). In addition, Cushing's Quality-of-Life (CushingQoL) (Webb et al., 2008) questionnaire was employed to evaluate the health-related quality of life of CD patients. Demographic and clinical variables, including age, years of education, neuroendocrine measures, and neuropsychological scores, were compared between the two groups using two-sample t-tests. Gender distribution was assessed using a two-tailed Pearson's chi-square test.

2.3 Image acquisition and data processing

Both structural and functional MRI data were acquired using a 3.0-Tesla MR system equipped with an 8-channel head coil. High-resolution structural 3D T1-weighted images were obtained using a sagittal Fast Spoiled Gradient-Echo (FSPGR) sequence. The acquisition parameters included a repetition time (TR) of 6.7 ms, echo time (TE) of 2.9 ms, flip angle of 7°, field of view (FOV) of 250 × 250 mm2, 192 slices, and a voxel size of 1 × 1 × 1 mm3 (without gaps). Functional imaging utilized an echo-planar imaging (EPI) sequence with a TR of 2000 ms, TE of 30 ms, flip angle of 90°, slice thickness/gap of 3.5 mm/0.5 mm, 36 slices, FOV of 224 × 224 mm2, voxel size of 3.5 × 3.5 × 3.5 mm3, and a total of 240 volumes. Soft earplugs were used to minimize scanner noise, and head motion was constrained with foam padding. During the functional scan, participants were instructed to keep their eyes closed and remain awake.

The rs-fMRI imaging preprocessing was performed with Statistical Parametric Mapping (SPM12) and Data Processing Assistant for Resting-State fMRI (DPABI) (Ashburner, 2012, Yan et al., 2016). Initially, the first ten time points of the functional images were discarded to avoid initial steady-state issues. Then, slice-timing correction was implemented to address timing differences among slices during scan acquisitions, and head motion correction was also conducted through spatial realignment of volumes to the first image volume. Subsequently, functional images were co-registered to each participant’s segmented gray matter T1 image and then spatially normalized to Montreal Neurological Institute space (MNI). Additionally, the normalized images were resampled to achieve 3-mm isotropic voxels and further smoothed using a 6 mm full-width at half maximum Gaussian kernel. The smoothed data were linearly detrended, and several confounding covariates, including the global signal, white matter signal, cerebrospinal fluid signal, and 24 motion parameters were regressed from the data. After that, temporal bandpass filtering (0.01–0.08 Hz) were carried out to reduce the effects of low-frequency drift and high-frequency physiological noise. Finally, a ‘scrubbing’ procedure was performed on the individual preprocessed data to eliminate artifacts from head motion. To reduce computational burdens, the preprocessed images were resampled to obtain a uniform 4 mm isotropic resolution using the Yeo 7 network template (Thomas Yeo et al., 2011).

The raw and normalized structural and functional images were subjected to screening, during which images showing excessive head motion with a translation of more than 2 mm or rotation exceeding 2° in any direction were excluded. Additionally, subjects with motion (mean FD Jenkinson) exceeding the threshold of 0.2 were also excluded from further analysis. After screening, a total of 86 subjects with CD (86/86) and 54 HCs (54/55) were ultimately included in the study.

2.4 Principal gradient analysis

For each participant, the voxel-wise functional connectivity matrices (16,482 × 16,482 nodes) were initially by calculating the Pearson's correlation between the time courses of each pair of nodes. The top 10 % of weighted connections per row were retained before calculating cosine similarity between all node pairs (Huo et al., 2022, Xia et al., 2022, Xia et al., 2022). The resultant similarity matrix was then normalized into an angle matrix to preclude negative values (Larivière et al., 2020, Paquola et al., 2019). We used diffusion map embedding, a nonlinear dimensionality reduction algorithm, to capture gradient components that account for the connectome variance (Hong et al., 2020, Hong et al., 2019). To establish comparability of the gradient components across individuals, we created a group-level gradient template based on an average connectivity matrix computed from the entire cohort of patients and controls. This template was then used to iteratively align individual gradient components across all subjects, with the process repeated 100 times (Dong et al., 2023, Hong et al., 2019, Langs et al., 2015).. Furthermore, the identified aligned gradient components ranked in descending order based on their explained ratios of connectome variance. Given the central focus of this study on the principal gradient, which exhibits a strong association with neuronal microstructure and cognitive functions (Huntenburg et al., 2018, Margulies et al., 2016), the gradient component demonstrating the highest explained variance in connectome patterns was selected as the focal point for further analysis in each participant.

Global gradient metrics, including the explain ratio, range, and spatial variation, were calculated to assess the extent to which the gradient accounts for variance, score differences between extreme gradient values, and the standard deviation of gradient values (Huo et al., 2022, Xia et al., 2022, Xia et al., 2022). Between-group comparisons (CD vs HC) were conducted using a general linear model (GLM, significance threshold of p < 0.05). At the regional level, the case-control differences in the principal gradient were evaluated using the two-sample t-test module of DPABI. Statistical significance was determined for the regional gradient value maps at the voxel level with a threshold of p < 0.001, followed by applying Gaussian random field (GRF) correction at a cluster level of p < 0.05. Furthermore, false discovery rate (FDR) correction and permutation test (5,000 times) were used to validate the reliability of the positive results. Network differences were obtained by mapping the Z-values of each voxel within identified clusters that exhibited statistical significance. Additionally, mean gradient values mapped to each functional network were compared between the two groups using a GLM. P-values from these comparisons were then adjusted via FDR correction to account for multiple testing (significance threshold of p < 0.05, qFDR < 0.05). Age, gender, and head motion were defined as confounding variables and controlled for in statistical process described above to eliminate their potential impact on the results.

2.5 Correlation analysis of CD-related gradient alterations and cognitive terms

The ‘Image Decoder’ function, leveraging the meta-analysis capabilities of the Neurosynth database (https://www.neurosynth.org/), facilitates the decoding of neuroimaging data to reveal patterns of brain activity linked to distinct cognitive, emotional, or perceptual processes. This function was employed to scrutinize alterations in CD gradient maps, extracting pertinent feature terms and their corresponding ‘r’ values associated with specific local brain regions. To assess the significance of correlation coefficients for each cognitive term, we conducted permutation tests (10,000 iterations) with spatial autocorrelation correction using null models generated via the random shuffling of Z-maps (Burt et al., 2020). The resulting surrogate Z-maps underwent thresholding (voxel-level p < 0.001, cluster-level GRF-corrected p < 0.05) and categorization into regions where CD > HC and CD < HC (M. Xia et al., 2022, Xia et al., 2022). Spatial correlations were then calculated between each cognitive term's meta-analytic map and the surrogate map. This process was repeated 10,000 times to create a null distribution, from which p-values for each cognitive term's correlation coefficient were derived. FDR correction was applied to account for multiple comparisons. Subsequently, correlation-based filtering identified the top 25 cognition-related terms for visualization with Word Claud.

2.6 Correlation analysis of CD-related global gradient metrics and regional gradient values with clinical indices

Initially, partial analyses were conducted to investigate the relationships between gradient global metrics (explain ratio, range, standard variation) and clinical indices (serum cortisol, 24 h-UFC, ACTH, SDS, SAS, MoCA, MMSE, Cushing QoL, illness duration) within the CD group. Moreover, we extracted mean gradient values in brain regions affected by CD-related gradient alterations and conducted partial correlation analyses to examine the association with clinical indices. Controlling for confounding variables (age, gender, and head motion), partial correlations were considered significant at a threshold of p < 0.05, qFDR < 0.05.

2.7 Correlation analysis of CD-related gradient alterations and gene expression profiles

Microarray gene expression data were obtained from the Allen Human Brain Atlas (AHBA, https://human.brain-map.org/), which provides whole-brain transcriptional profiles from six adult human donors aged 24–57 years (Hawrylycz et al., 2012). The preprocessing procedure adhered to the recommended workflow for integrating AHBA and neuroimaging data (Arnatkeviciute et al., 2019). To address incomplete coverage and incompatible templating of the original AHBA dataset at voxel level, we implemented the analysis approach of Xia et al. (Xia et al., 2022, Xia et al., 2022, Xia et al., 2022, Xia et al., 2022) using their openly available, preprocessed microarray data encompassing 284 cortical regions, with each regions containing the expression of 10,027 genes Supplementary methods. Gene expression values were then averaged across all samples and subjects to generate a gene expression map (284 regions × 10027 genes). Finally, we aligned these data to our gray matter mask template, which removed regions with <50 % gray matter coverage.

Pearson correlation analysis was used to examine the relationship between gene expression maps and the principal gradient Z-map derived from between-group regional gradient value comparisons. The significance of the correlation coefficients was evaluated through 10,000 permutations alongside spatial autocorrelation corrections using null models generated via the random shuffling of Z-maps (Burt et al., 2020; M. Xia et al., 2022, Xia et al., 2022). Genes exhibiting significant correlations underwent screening and ranking based on the significance levels of their correlation coefficients. Subsequently, Gene ontology (GO) enrichment analysis was then performed on the ranked gene lists (ascending order and descending order) using GOrilla (https://cbl-gorilla.cs.technion.ac.il/), applying a p-value threshold of 10-5 with multiple testing correction (qFDR<0.05) to identify enriched biological process (BP), molecular function (MF), and cellular component terms (CC).

3 Results

3.1 Demographic, endocrinological, and neuropsychological results

A case-control study finally enrolled 86 CD patients (78 females, 8 males; mean age 39.66 ± 11.18 years) and 54 HCs (51 females, 3 males; mean age 34.53 ± 10.70 years). Demographic analysis revealed a comparable gender ratio and education level between groups. However, CD patients were significantly older than HCs (t = 2.683, p = 0.008). In neuropsychological assessments, CD patients scored lower on the MMSE and MoCA (all p < 0.01), indicating cognitive impairments. Depression and anxiety symptoms were more prevalent in CD patients, as evidenced by higher scores of SDS and SAS (all p < 0.01). Endocrine abnormalities also characterized CD patients, including elevated 8 a.m. cortisol, 24 h- UFC, and ACTH levels compared to controls (all p < 0.01). For detailed information, please refer to Table 1.Table 1 Demographic and clinical indices of the participants.

	CD (n = 86)	HC (n = 54)	t/χ2	p	
Age (years)	39.66 ± 11.18	34.53 ± 10.70	2.68	a8.00 × 10-3	
Gender (male/female)	8/78	3/51	0.69	0.41	
Education (years)	11.80 ± 4.26	11.80 ± 3.10	−0.06	0.95	
Illness duration (month)	41.34 ± 46.60	−	−	−	
Neuropsychological Tests					
 SDS	43.38 ± 10.92	27.06 ± 4.38	12.37	a3.09 × 10-23	
 SAS	42.66 ± 11.67	26.94 ± 4.42	11.27	a1.84 × 10-20	
 MoCA	22.74 ± 4.03	27.80 ± 1.70	−10.25	a3.04 × 10-18	
 MMSE	27.83 ± 2.28	29.30 ± 0.92	−5.33	a4.52 × 10-4	
 Cushing QoL	37.04 ± 9.70	–	–	–	
Endocrinological Tests					
 Serum cortisol(nmol/L)					
  0 a.m.	575.98 ± 222.20	–	–	–	
  8 a.m.	705.23 ± 262.49	355.40 ± 107.47	10.98	a6.03 × 10-20	
  4 p.m.	647.00 ± 262.36	–	–	–	
 ACTH (pmol/L)					
  0 a.m.	15.41 ± 9.86	–	–	–	
  8 a.m.	19.72 ± 14.18	4.93 ± 3.05	9.33	a3.70 × 10-15	
  4 p.m.	19.62 ± 13.28	–			
24 h-UFC/(nmol/24 h)	2113.84 ± 1329.69	242.70 ± 116.43	12.97	a4.78 × 10-22	
Note:a p < 0.05. Data are presented as mean ± standard deviation. The significant age difference would not affect the group differences in other indices (all p < 0.05). Abbreviations:CD Cushing's disease, HC healthy control, SDS self-rating depression scale, SAS self-rating anxiety scale, MoCA Montreal cognitive assessment, MMSE mini-mental state examination, Cushing QoL Cushing quality of life scale, ACTH adrenocorticotropic hormone, 24 h-UFC 24-hour urinary free cortisol.

3.2 Altered principal gradient in CD patients

Both CD and HC groups exhibited a principal gradient extending hierarchically from primary sensory to transmodal regions, aligning with previous findings (Fig. 2a, b). However, the principal gradient accounted for significantly less variance in functional connectome in the CD group compared to the HC group (8.7 ± 2.7 % vs 9.9 ± 1.8 %, t = −2.79, p = 0.0061) (Fig. 2c, Fig. S1). Additionally, the gradient range and spatial variation were also remarkably lower in the CD group (all p < 0.01). These global gradient metric abnormalities revealed deficient hierarchical organization of the principal gradient in CD patients. We further examined regional-level alterations in CD group compared to HC group (Fig. 2d, Table S1). The results demonstrated compressed variation along the principal gradient axis in CD patients (voxel-level p < 0.001, GRF-corrected p < 0.05), with increased gradient values primarily distributed in the sensorimotor and visual cortices. Decreased gradient values were mainly localized in the bilateral cingulate cortex and right prefrontal cortex (r-PFC) in CD patients. Similar results were obtained after FDR correction and permutation test 5 000 times for statistical correction (Fig. S2). Mapping of regions with significant gradient differences to large-scale cortical networks revealed a differential distribution pattern (Fig. 2e, f). The visual network (VIS) (48.7 %) and somatomotor network (SMN) (22.9 %) collectively accounted for 71.6 % of areas exhibiting gradient increases, whereas the remaining 28.4 % were distributed across other networks. In contrast, the DMN overlapped with 57.7 % of areas with decreased values, followed by the ventral attention network (VAN) (21.0 %) and frontoparietal network (FPN) (9.8 %). In addition, the mean gradient values mapped to the functional networks were also compared between the two groups, revealing statistical differences across networks (all p < 0.01, qFDR < 0.01) except the DAN (t = 0.28, p = 0.78) (Table S2).Fig. 2 Comparison and statistical analysis of principal gradient mapping between CD and HC groups. (a). The principal gradient demonstrated a spectrum from unimodal (dark blue) to transmodal (sienna) regions in both the CD group and HC group. (b). As depicted in the global and network-based histograms, CD group exhibited altered convergent representation along the principal gradient compared to HC group. (c). The CD group exhibited decreased global gradient metrics compared to controls, including explained variance, range, and spatial variation (**p < 0.01). (d). Voxel-wise intergroup statistical comparisons showed enhanced and reduced activations in the CD group along the principal gradient, mapped in orange and blue, respectively. (e). Significant intergroup statistical differences were mapped into a voxel-wise scatterplot constructed from mean gradient values of the HC and CD groups. Specifically, orange points in the left lower quadrant indicated elevated gradient values of CD in negative regions associated with primary brain areas, while blue points in the right upper quadrant suggested reduced gradient values and involvement of higher-order regions. (f). Mapping revealed a differential distribution pattern across cortical networks. The VIS and SMN accounted for 71.6 % of areas with increased gradient values, while the DMN, VAN and FPN overlapped with 88.5 % of areas with decreased gradient values. VIS visual network, SMN sensorimotor network, DAN dorsal attention network, VAN ventral attention network, LIB limbic network, FPN frontoparietal network, DMN default mode network.

3.3 Relation to cognitive function domains

A Neurosnyth database meta-analysis elucidated distinct cognitive domains implicated in the altered gradient patterns observed in CD patients (Fig. 3a, Table S3). Regions exhibiting increased gradient values participated primarily in lower-order cognition, such as “visual”, “spatial attention”, “location”. Relatively, areas with decreased gradients were associated with abstract facets of higher-order cognition like “referential”, “theory of mind”, “thinking”.Fig. 3 Aberrant principal gradient patterns associated with cognitive function domains, clinical indices, and gene expression in CD patients. (a). According to the word cloud graph generated from the Neurosnyth database meta-analysis, regions exhibiting increased gradient values participated primarily in lower-order cognition. Relatively, areas with decreased gradient values were more associated with abstract facets of higher-order cognition (deeper color and larger size represent a larger r value.). (b). The gradient values observed in the r-PFC and bilateral sensorimotor cortices of CD patients exhibited a significant correlation with cortisol levels. The cortical localization of the region is depicted by the brain map located above. (c). A total of 3215 genes exhibited significant correlations with the principal gradient Z-statistic values. The genes were sorted in ascending and descending order according to their correlation coefficients to prepare for subsequent gene enrichment analysis.

3.4 Relation to clinical features in CD patients

We analyzed the correlation between clinical indices of CD and gradient values in brain regions showing group differences (Fig. 3b, Fig. S3). In CD patients, global gradient metrics did not exhibit significant associations with clinical indices. However, the gradient values of the r-PFC were negatively correlated with 24 h-UFC (r = −0.29, p = 0.007, qFDR=0.025). Meanwhile, bilateral sensorimotor cortical gradient values exhibited positive correlations with serum cortisol at 0 a.m. (left r = 0.28, right r = 0.31, all p < 0.05, qFDR<0.05). Furthermore, the gradient values of the right sensorimotor cortex were also positively associated with 24 h-UFC and serum cortisol at 4 p.m. (r = 0.29, r = 0.31, all p < 0.05, qFDR<0.05). MoCA and SDS scores demonstrated negative correlations with sensorimotor cortex gradient values, though these relationships did not remain significant after multiple comparisons correction (r = −0.28, r = −0.24, all p < 0.05, qFDR > 0.05).

3.5 Relation to gene expression profiles

A total of 3215 genes across 281 brain regions exhibited significant correlations with the principal gradient Z-map (Fig. 3c, Table S4). Gene set enrichment analysis revealed numerous enrichments related to synaptic components and function within the BP, MF, and CC ontologies amongst genes ranked in ascending order of correlation coefficients (Fig. 4, Table S5). However, no significant enrichments were observed in the genes ranked in descending order. Overall, the enrichment analysis of associated genes primarily implicated synaptic components and function, specifically encompassing synaptic signal transduction and regulation within BP, neurotransmitter receptor and ion channel activities among MF, and presynaptic membrane components in CC (Fig. S4).Fig. 4 Gene enrichment results were visualized as bubble plots of related GO terms. Gene ontology enrichment analysis revealed the genes exhibiting expression correlation with principal gradient changes were significantly enriched in terms associated with biological process (BP), molecular function (MF), and cellular component (CC). The term “neurotransmitter receptor activity” exhibited the highest statistical significance, while the term “transmitter-gated ion channel activity involved in regulation of postsynaptic membrane potential” obtained the most robust enrichment scores. Taken together, the enrichment results were closely associated with synaptic components and function.

4 Discussion

4.1 Convergent representation of principal gradient in CD patients

This study systematically investigated the principal gradient within the hierarchical architecture of the brain in CD patients, provided novel insights from multiple analytical perspectives. Compared to HCs, the reduced explanatory capacity of the principal gradient for variance in functional connectome in CD patients indicates a disruption in the brain's normal hierarchical organization of information processing. These widespread functional connectome disturbances across multiple brain regions in CD patients, emphasizing the considerable diversity and complexity of the connectome changes, which may underpin the reduced capacity of the principal gradient extracted from CD patients to capture the predominant patterns of whole-brain functional connectome adequately. In addition, our study also found the range of principal gradient values narrowed with decreased spatial variation in CD patients compared to HCs, suggesting a closer distance between the two poles along the principal gradient axis, possibly indicating overall compression and weakened internal heterogeneity of the principal gradient, exhibiting a convergent change.

Mapping gradients in functional connectome offers insight into the topological embedding of brain networks and the possible links to cognitive processes of varying complexity (Huntenburg et al., 2018, Margulies et al., 2016). Thus, alterations in the magnitude of gradient values likely reflect shifts in the relative embedding of diverse networks across the functional connectome hierarchy (Dong et al., 2023). Such hierarchical reorganization may impair the integration and segregation of brain functions, subsequently disrupting cognitive activities (Hong et al., 2019, Margulies et al., 2016; M. Xia et al., 2022, Xia et al., 2022). In CD patients, elevated gradient values were chiefly detected in primary networks (SMN, VN) compared to HCs, signifying greater similarity with downstream networks that potentially imply aberrant operation as these networks take on non-native roles in processing external stimuli. Such cross-domain functional enhancement could impair the more specialized filtering capacities of primary networks, enabling noise to permeate information flow to subsequent networks. Conversely, reduced gradient values within VAN, FPN and DMN imply these higher-order networks are becoming more functionally aligned with primary networks, a phenomenon that may stem from advanced network endeavoring to offset abnormalities in primary network processing and consequently impairing their inherent functionality. Collectively, we hypothesize that alterations in the principal gradient pattern in CD patients lead to convergence of the hierarchical architecture disrupting the gap between distinct networks and consequently impairing cognitive processes.

4.2 Principal gradient alterations link to cognitive-emotional impairments in CD patients

Based on Neurosynth meta-analyses, regions with increased gradients were predominantly involved in sensory processing and certain higher-order cognitive functions, while regions with decreased gradients localized almost entirely within areas related to higher-order cognition in CD patients. Several cognitive functions have been validated as deficient in CD patients, such as memory, attention, reasoning, visuospatial processing, concept formation, and executive functioning (Piasecka et al., 2020b). While abnormalities in visuospatial processing implicating primary networks such as the occipital lobe align with our findings in CD patients, these primary network perturbations remain peripheral within the current research landscape (Forget et al., 2000, Jiang et al., 2017). In contrast, disruption of higher-order networks involved in cognitive and emotional processing, including the DMN, FPN, and LIB, represents the critical characterization of CD patients (Li et al., 2022, Liu et al., 2022, Papakokkinou and Ragnarsson, 2023, Stomby et al., 2019, Wang et al., 2019). Abnormalities among these networks require intense research focus due to their criticality in driving the cognitive and affective deficits to this disorder. The DMN plays avital role in maintaining normal cognition, because of its “stimulus independence” and “content heterogeneity”, which allow internally-driven thinking across domains without external stimulus (van den Heuvel and Sporns, 2013). DMN activity is content-independent, serving as an integrative hub that abstracts common information patterns across cortical regions. The principal gradient offers a coherent explanation for the characteristics of the DMN – its farthest position from primary modal areas allows functional independence, while equidistant connections to each modality facilitate the integration of diverse sensory information, embodying the network's heterogeneity (Margulies et al., 2016). Alterations in the position of DMN relative to primary modalities along the principal axis undermined its independence, increasing susceptibility to external input and potentially hindering integration and abstraction of complex information, explaining the deficits in higher-order cognitive functions in CD patients.

Affective symptoms (e.g., depression, anxiety) represent a primary clinical manifestation in CD patients (Piasecka et al., 2020b, Pivonello et al., 2015, Valassi et al., 2017). Dysfunction within the DMN has been strongly implicated in these affective disturbances (Broyd et al., 2009, Yan et al., 2019). A recent study of functional connectome gradients in major depressive disorder showed reduced DMN gradient values, aligning with our current findings (M. Xia et al., 2022, Xia et al., 2022). Moreover, brain maps of the principal gradient significantly predicted changes in Hamilton Depression Rating Scale scores in patients with MDD following 8 weeks of antidepressant therapy, with the DMN making the greatest contribution. In summary, these results suggest that the abnormal position of the DMN within the principal gradient could explain the emotional symptoms of CD patients.

4.3 Elevated cortisol levels exacerbate abnormalities in the principal gradient of CD

Hypercortisolism exerts deleterious effects on the central nervous system (Hansson et al., 2006, Hansson et al., 2000, Jeanneteau et al., 2008), potentially causing CD-associated cognitive impairments through aberrant functional connectome (Li et al., 2022, Wang et al., 2019). Our study found that with the elevation of cortisol, divergences in principal gradients are amplified between CD patients and HCs. Specifically, gradient values increase in bilateral sensorimotor cortices while decreasing in the r-PFC. The principal gradient constitutes key component that characterize the functional connectome patterns during complex cognitive processing. Utilizing principal gradient alternations as a biomarker reveals novel insights into the neural mechanisms linking hypercortisolism to cognitive impairments in CD. Our multimodal methodology elucidates the potential role of cortisol in disrupting functional connectome between large-scale brain networks associated with cognitive functions. This means that targeting cortisol levels remains an important treatment strategy for improving cognitive function and emotional processing in CD patients, rather than relying solely on antipsychotic treatment.

4.4 Synaptic gene enrichment linked to abnormalities in the principal gradient of CD

Gene enrichment analysis revealed expression genes correlated with principal gradient alternations were highly enriched for BP, MF, and CC related to synaptic components and function. BP singled out genes governing synaptic signal transduction, underscoring their regulatory roles in mediating signal conduction and preserving synaptic functional integrity. MF analysis highlighted enrichment of diverse synaptic receptors and channels, implicating these as key signaling molecules on the synaptic membrane that critically mediate input and output. CC enrichment revealed genes encoding synaptic scaffolding proteins, highlighting their essentiality in constructing presynaptic compartments, stabilizing the synaptic interface, and furnishing requisite infrastructure for neurotransmission. The convergence of enrichment across these three domains provides compelling evidence that alterations in synaptic components and function may represent a pivotal molecular substrate underlying the aberrant gradient changes observed in CD patients.

Prolonged cortisol overexposure induces deleterious structural and functional alterations in neural synapses, causing impairments to cognitive and emotional processing (Popoli et al., 2011, Tata et al., 2006, Tata and Anderson, 2010). Glucocorticoids exert effects in different brain regions by binding to glucocorticoid receptors (GRs) and mineralocorticoid receptors (MRs) with differing affinities (Chen et al., 2013, Kalafatakis et al., 2019). In addition, some membrane-associated receptors, including G protein-coupled receptors, may also participate in mediating certain glucocorticoid effects (Johnson et al., 2005, Karst et al., 2005, Orchinik et al., 1992). Studies on the synaptic impacts of prolonged glucocorticoid overexposure and chronic stress have principally utilized animal models, concentrating heavily on the PFC and hippocampus – brain regions with MRs expression, elevated glucocorticoid sensitivity, and roles in complex cognition (Popoli et al., 2011). Chronic stress exposure leads to reductions in hippocampal synapse number and volume, and downregulation of postsynaptic glutamate receptors and elevated presynaptic glutamate levels in the PFC (Cerqueira et al., 2007, Goldwater et al., 2009, Popoli et al., 2011, Yuen et al., 2012). These stress-induced synaptic modifications in the hippocampus and PFC remodel synaptic architecture and function. Moreover, a brain functional network-based genomic study provides further evidence that synaptic function is an integral component in the maintenance of brain network activity (Richiardi et al., 2015). Since principal gradient reflect the hierarchical organizational relationships within brain networks, these results lend collateral support to the reliability of our conclusion that chronic hypercortisolism disrupts synaptic components and function, leading to alternations in principal gradient in CD patients.

4.5 Limitations

The principal gradient examined in this study were obtained at the voxel level. Although gradient patterns show consistency across resolutions, validating discoveries at the vertex level remains essential to demonstrate generalizability. In addition, the application of a cortex-parcellated, voxel-level Yeo 7 network template enabled the capture of the principal gradient governing whole-brain functional network connectivity, promoting a unified explanatory framework. However, the exclusive cortical focus of this approach inevitably overlooked subcortical and cerebellar regions. Since these overlooked areas are also closely correlated with brain functional abnormalities in CD patients, future efforts will shift to examinations of hippocampal and cerebellar focal regions to unravel localized gradient features.

5 Conclusions

Treating Cushing diseases as a pathological model of glucocorticoid excess, our study investigated the effects of chronic hypercortisolemia on the functional connectome from a novel perspective integrating macro and micro levels. Converged aberrant alterations in the hierarchical organization of brain networks were revealed at the macroscale level by disruptions to the principal gradient in CD patients. Regions exhibiting gradient abnormalities may mediate associations between cortisol levels and cognitive deficits. At the microscale level, the corresponding gene expression profiles implicated dysregulation of synaptic components and function as a potential driving mechanism underlying glucocorticoid-induced alterations of the principal gradient in brain networks and resultant cognitive and emotional impairments in CD. Overall, by linking macroscale imaging phenotypes to microscale genomic pathways, this work enables novel perspectives to advance intricate pathophysiological processes underlying CD, and facilitates collaborative and integrative research across disciplines.

6 Ethical compliance statement

All reported procedures were conducted within routine clinical care and in accordance with the Declaration of Helsinki. Informed written consent to report and publish collected patient data was obtained. The study was approved by the Ethics Committee of the Chinese PLA General Hospital (S2021-677-01).

7 Funding sources

Funding for this study was provided by the National Natural Science Foundation of China (Grant Nos. 82001798 to Xinguang Yu; Grant Nos. 81871087 to Yanyang Zhang) and the Young Talent Project of Chinese PLA General Hospital (Grant Nos. 20230403 to Yanyang Zhang). All funding sources had no further role in study design; in the collection, analysis and interpretation of data; in the writing of the report; and in the decision to submit the paper for publication.

CRediT authorship contribution statement

Guosong Shang: Writing – review & editing, Writing – original draft, Visualization, Software, Methodology, Investigation, Formal analysis. Tao Zhou: Writing – review & editing, Writing – original draft, Resources, Methodology, Formal analysis. Xiaoteng Yu: Writing – review & editing, Writing – original draft, Software, Formal analysis, Data curation. Xinyuan Yan: Writing – review & editing, Methodology, Formal analysis. Kunyu He: Writing – review & editing, Investigation. Bin Liu: Writing – review & editing, Investigation. Zhebin Feng: Writing – review & editing, Investigation. Junpeng Xu: Writing – review & editing, Investigation. Yanyang Zhang: Writing – review & editing, Supervision, Funding acquisition, Conceptualization. Xinguang Yu: Writing – review & editing, Project administration, Funding acquisition, Conceptualization.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A Supplementary data

The following are the Supplementary data to this article:Supplementary Data 1

Supplementary Data 2

Data availability

Gradient and gene analysis were obtained from the available GitHub repository associated with previous work by Mingrui Xia et.al. (http://github.com/mingruixia/MDD_ConnectomeGradient). Figures were generated using relevant Word cloud (1.9.2) packages in Python and Brain Net Viewer packages in MATLAB ((R2022a)), ChiPlot (https://www.chiplot.online/). The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Acknowledgments

We would like to acknowledge Mingrui Xia et al. for making available open-source code and adapted intermediate data, which were critical to the successful completion of our research.

Appendix A Supplementary data to this article can be found online at https://doi.org/10.1016/j.nicl.2024.103652.
==== Refs
References

Ahn C.H. Kim J.H. Park M.Y. Kim S.W. Epidemiology and comorbidity of adrenal Cushing syndrome: A nationwide cohort study J. Clin. Endocrinol. Metab. 106 2021 e1362 e1372 10.1210/clinem/dgaa752 33075802
Arnatkeviciute A. Fulcher B.D. Fornito A. A practical guide to linking brain-wide gene expression and neuroimaging data Neuroimage 189 2019 353 367 10.1016/j.neuroimage.2019.01.011 30648605
Ashburner J. SPM: a history Neuroimage 62 2012 791 800 10.1016/j.neuroimage.2011.10.025 22023741
Brown V.M. LaBar K.S. Haswell C.C. Gold A.L. Workgroup M.-A. McCarthy G. Morey R.A. Altered resting-state functional connectivity of basolateral and centromedial amygdala complexes in posttraumatic stress disorder Neuropsychopharmacology 39 2014 351 359 10.1038/npp.2013.197 23929546
Broyd S.J. Demanuele C. Debener S. Helps S.K. James C.J. Sonuga-Barke E.J.S. Default-mode brain dysfunction in mental disorders: a systematic review Neurosci. Biobehav. Rev. 33 2009 279 296 10.1016/j.neubiorev.2008.09.002 18824195
Burt J.B. Helmer M. Shinn M. Anticevic A. Murray J.D. Generative modeling of brain maps with spatial autocorrelation Neuroimage 220 2020 117038 10.1016/j.neuroimage.2020.117038
Cerqueira J.J. Mailliet F. Almeida O.F.X. Jay T.M. Sousa N. The prefrontal cortex as a key target of the maladaptive response to stress J. Neurosci. 27 2007 2781 2787 10.1523/JNEUROSCI.4372-06.2007 17360899
Chen Y. Li Y. Chen X. Sun Q. Neuropsychiatric disorders and cognitive dysfunction in patients with Cushing’s disease Chin. Med. J. (Engl.) 126 2013 3156 3160 23981629
Dong D. Yao D. Wang Y. Hong S.-J. Genon S. Xin F. Jung K. He H. Chang X. Duan M. Bernhardt B.C. Margulies D.S. Sepulcre J. Eickhoff S.B. Luo C. Compressed sensorimotor-to-transmodal hierarchical organization in schizophrenia Psychol. Med. 53 2023 771 784 10.1017/S0033291721002129 34100349
Fleseriu M. Auchus R. Bancos I. Ben-Shlomo A. Bertherat J. Biermasz N.R. Boguszewski C.L. Bronstein M.D. Buchfelder M. Carmichael J.D. Casanueva F.F. Castinetti F. Chanson P. Findling J. Gadelha M. Geer E.B. Giustina A. Grossman A. Gurnell M. Ho K. Ioachimescu A.G. Kaiser U.B. Karavitaki N. Katznelson L. Kelly D.F. Lacroix A. McCormack A. Melmed S. Molitch M. Mortini P. Newell-Price J. Nieman L. Pereira A.M. Petersenn S. Pivonello R. Raff H. Reincke M. Salvatori R. Scaroni C. Shimon I. Stratakis C.A. Swearingen B. Tabarin A. Takahashi Y. Theodoropoulou M. Tsagarakis S. Valassi E. Varlamov E.V. Vila G. Wass J. Webb S.M. Zatelli M.C. Biller B.M.K. Consensus on diagnosis and management of Cushing’s disease: a guideline update Lancet Diabetes Endocrinol. 9 2021 847 875 10.1016/S2213-8587(21)00235-7 34687601
Folstein M.F. Folstein S.E. McHugh P.R. “Mini-mental state”. A practical method for grading the cognitive state of patients for the clinician J. Psychiatr. Res. 12 1975 189 198 10.1016/0022-3956(75)90026-6 1202204
Forget H. Lacroix A. Somma M. Cohen H. Cognitive decline in patients with Cushing’s syndrome J. Int. Neuropsychol. Soc. 6 2000 20 29 10.1017/s1355617700611037 10761364
Goldwater D.S. Pavlides C. Hunter R.G. Bloss E.B. Hof P.R. McEwen B.S. Morrison J.H. Structural and functional alterations to rat medial prefrontal cortex following chronic restraint stress and recovery Neuroscience 164 2009 798 808 10.1016/j.neuroscience.2009.08.053 19723561
Hansson A.C. Cintra A. Belluardo N. Sommer W. Bhatnagar M. Bader M. Ganten D. Fuxe K. Gluco- and mineralocorticoid receptor-mediated regulation of neurotrophic factor gene expression in the dorsal hippocampus and the neocortex of the rat Eur. J. Neurosci. 12 2000 2918 2934 10.1046/j.1460-9568.2000.00185.x 10971634
Hansson A.C. Sommer W.H. Metsis M. Stromberg I. Agnati L.F. Fuxe K. Corticosterone actions on the hippocampal brain-derived neurotrophic factor expression are mediated by exon IV promoter J. Neuroendocrinol. 18 2006 104 114 10.1111/j.1365-2826.2005.01390.x 16420279
Hawrylycz M.J. Lein E.S. Guillozet-Bongaarts A.L. Shen E.H. Ng L. Miller J.A. Van De Lagemaat L.N. Smith K.A. Ebbert A. Riley Z.L. Abajian C. Beckmann C.F. Bernard A. Bertagnolli D. Boe A.F. Cartagena P.M. Chakravarty M.M. Chapin M. Chong J. Dalley R.A. Daly B.D. Dang C. Datta S. Dee N. Dolbeare T.A. Faber V. Feng D. Fowler D.R. Goldy J. Gregor B.W. Haradon Z. Haynor D.R. Hohmann J.G. Horvath S. Howard R.E. Jeromin A. Jochim J.M. Kinnunen M. Lau C. Lazarz E.T. Lee C. Lemon T.A. Li L. Li Y. Morris J.A. Overly C.C. Parker P.D. Parry S.E. Reding M. Royall J.J. Schulkin J. Sequeira P.A. Slaughterbeck C.R. Smith S.C. Sodt A.J. Sunkin S.M. Swanson B.E. Vawter M.P. Williams D. Wohnoutka P. Zielke H.R. Geschwind D.H. Hof P.R. Smith S.M. Koch C. Grant S.G.N. Jones A.R. An anatomically comprehensive atlas of the adult human brain transcriptome Nature 489 2012 391 399 10.1038/nature11405 22996553
Hong S.-J. Vos de Wael R. Bethlehem R.A.I. Lariviere S. Paquola C. Valk S.L. Milham M.P. Di Martino A. Margulies D.S. Smallwood J. Bernhardt B.C. Atypical functional connectome hierarchy in autism Nat. Commun. 10 2019 1022 10.1038/s41467-019-08944-1 30833582
Hong S.-J. Xu T. Nikolaidis A. Smallwood J. Margulies D.S. Bernhardt B. Vogelstein J. Milham M.P. Toward a connectivity gradient-based framework for reproducible biomarker discovery Neuroimage 223 2020 117322 10.1016/j.neuroimage.2020.117322
Huntenburg J.M. Bazin P.-L. Margulies D.S. Large-scale gradients in human cortical organization Trends Cogn. Sci. 22 2018 21 31 10.1016/j.tics.2017.11.002 29203085
Huo T. Xia Y. Zhuang K. Chen Q. Sun J. Yang W. Qiu J. Linking functional connectome gradient to individual creativity Cereb. Cortex 32 2022 5273 5284 10.1093/cercor/bhac013 35136988
Jeanneteau F. Garabedian M.J. Chao M.V. Activation of Trk neurotrophin receptors by glucocorticoids provides a neuroprotective effect Proc. Natl. Acad. Sci. U.S.A. 105 2008 4862 4867 10.1073/pnas.0709102105 18347336
Jiang H. He N.-Y. Sun Y.-H. Jian F.-F. Bian L.-G. Shen J.-K. Yan F.-H. Pan S.-J. Sun Q.-F. Altered spontaneous brain activity in Cushing’s disease: a resting-state functional MRI study Clin. Endocrinol. 86 2017 367 376 10.1111/cen.13277
Jin C. Qi R. Yin Y. Hu X. Duan L. Xu Q. Zhang Z. Zhong Y. Feng B. Xiang H. Gong Q. Liu Y. Lu G. Li L. Abnormalities in whole-brain functional connectivity observed in treatment-naive post-traumatic stress disorder patients following an earthquake Psychol. Med. 44 2014 1927 1936 10.1017/S003329171300250X 24168716
Johnson L.R. Farb C. Morrison J.H. McEwen B.S. LeDoux J.E. Localization of glucocorticoid receptors at postsynaptic membranes in the lateral amygdala Neuroscience 136 2005 289 299 10.1016/j.neuroscience.2005.06.050 16181741
Kalafatakis K. Giannakeas N. Lightman S.L. Charalampopoulos I. Russell G.M. Tsipouras M. Tzallas A. Utilization of the allen gene expression atlas to gain further insight into glucocorticoid physiology in the adult mouse brain Neurosci. Lett. 706 2019 194 200 10.1016/j.neulet.2019.05.020 31100428
Karst H. Berger S. Turiault M. Tronche F. Schütz G. Joëls M. Mineralocorticoid receptors are indispensable for nongenomic modulation of hippocampal glutamate transmission by corticosterone Proc. Natl. Acad. Sci. U.S.A. 102 2005 19204 19207 10.1073/pnas.0507572102 16361444
Langs G. Golland P. Ghosh S.S. Predicting activation across individuals with resting-state functional connectivity based multi-atlas label fusion Med. Image Comput. Comput. Assist. Interv. 9350 2015 313 320 10.1007/978-3-319-24571-3_38 26855977
Larivière S. Vos de Wael R. Hong S.-J. Paquola C. Tavakol S. Lowe A.J. Schrader D.V. Bernhardt B.C. Multiscale structure-function gradients in the neonatal connectome Cereb. Cortex 30 2020 47 58 10.1093/cercor/bhz069 31220215
Li C. Zhang Y. Wang W. Zhou T. Yu X. Tao H. Altered hippocampal volume and functional connectivity in patients with Cushing’s disease Brain Behav. 12 2022 e2507 10.1002/brb3.2507
Liu Y.-F. Pan L. Feng M. Structural and functional brain alterations in Cushing’s disease: A narrative review Front. Neuroendocrinol. 67 2022 101033 10.1016/j.yfrne.2022.101033
Margulies D.S. Ghosh S.S. Goulas A. Falkiewicz M. Huntenburg J.M. Langs G. Bezgin G. Eickhoff S.B. Castellanos F.X. Petrides M. Jefferies E. Smallwood J. Situating the default-mode network along a principal gradient of macroscale cortical organization Proc. Natl. Acad. Sci. U.S.A. 113 2016 12574 12579 10.1073/pnas.1608282113 27791099
Nasreddine Z.S. Phillips N.A. Bédirian V. Charbonneau S. Whitehead V. Collin I. Cummings J.L. Chertkow H. The Montreal Cognitive Assessment, MoCA: a brief screening tool for mild cognitive impairment J. Am. Geriatr. Soc. 53 2005 695 699 10.1111/j.1532-5415.2005.53221.x 15817019
Nieman L.K. Biller B.M.K. Findling J.W. Murad M.H. Newell-Price J. Savage M.O. Tabarin A. Endocrine Society Treatment of Cushing’s syndrome: an Endocrine Society Clinical Practice Guideline J. Clin. Endocrinol. Metab. 100 2015 2807 2831 10.1210/jc.2015-1818 26222757
Orchinik M. Murray T.F. Franklin P.H. Moore F.L. Guanyl nucleotides modulate binding to steroid receptors in neuronal membranes Proc. Natl. Acad. Sci. U. S. A. 89 1992 3830 3834 10.1073/pnas.89.9.3830 1570300
Papakokkinou E. Ragnarsson O. Functional brain alterations in Cushing’s syndrome Front. Endocrinol. 14 2023 1163482 10.3389/fendo.2023.1163482
Paquola C. Vos De Wael R. Wagstyl K. Bethlehem R.A.I. Hong S.-J. Seidlitz J. Bullmore E.T. Evans A.C. Misic B. Margulies D.S. Smallwood J. Bernhardt B.C. Microstructural and functional gradients are increasingly dissociated in transmodal cortices PLoS Biol. 17 2019 e3000284 10.1371/journal.pbio.3000284
Piasecka M. Papakokkinou E. Valassi E. Santos A. Webb S.M. de Vries F. Pereira A.M. Ragnarsson O. Psychiatric and neurocognitive consequences of endogenous hypercortisolism J. Intern. Med. 288 2020 168 182 10.1111/joim.13056 32181937
Piasecka M. Papakokkinou E. Valassi E. Santos A. Webb S.M. Vries F. Pereira A.M. Ragnarsson O. Psychiatric and neurocognitive consequences of endogenous hypercortisolism J. Intern. Med. 288 2020 168 182 10.1111/joim.13056 32181937
Pivonello R. Simeoli C. De Martino M.C. Cozzolino A. De Leo M. Iacuaniello D. Pivonello C. Negri M. Pellecchia M.T. Iasevoli F. Colao A. Neuropsychiatric disorders in Cushing’s syndrome Front. Neurosci. 9 2015 10.3389/fnins.2015.00129
Popoli M. Yan Z. McEwen B. Sanacora G. The stressed synapse: the impact of stress and glucocorticoids on glutamate transmission Nat. Rev. Neurosci. 13 2011 22 37 10.1038/nrn3138 22127301
Richiardi, J., Altmann, A., Milazzo, A.-C., Chang, C., Chakravarty, M.M., Banaschewski, T., Barker, G.J., Bokde, A.L.W., Bromberg, U., Büchel, C., Conrod, P., Fauth-Bühler, M., Flor, H., Frouin, V., Gallinat, J., Garavan, H., Gowland, P., Heinz, A., Lemaître, H., Mann, K.F., Martinot, Jean-Luc, Nees, F., Paus, Tomáš, Pausova, Z., Rietschel, M., Robbins, T.W., Smolka, M.N., Spanagel, R., Ströhle, A., Schumann, G., Hawrylycz, M., Poline, J.-B., Greicius, M.D., IMAGEN consortium, Albrecht, L., Andrew, C., Arroyo, M., Artiges, E., Aydin, S., Bach, C., Banaschewski, T., Barbot, A., Barker, G., Boddaert, N., Bokde, A., Bricaud, Z., Bromberg, U., Bruehl, R., Büchel, C., Cachia, A., Cattrell, A., Conrod, P., Constant, P., Dalley, J., Decideur, B., Desrivieres, S., Fadai, T., Flor, H., Frouin, V., Gallinat, J., Garavan, H., Briand, F.G., Gowland, P., Heinrichs, B., Heinz, A., Heym, N., Hübner, T., Ireland, J., Ittermann, B., Jia, T., Lathrop, M., Lanzerath, D., Lawrence, C., Lemaitre, H., Lüdemann, K., Macare, C., Mallik, C., Mangin, J.-F., Mann, K., Martinot, Jean- Luc, Mennigen, E., Mesquita De Carvahlo, F., Mignon, X., Miranda, R., Müller, K., Nees, F., Nymberg, C., Paillere, M.-L., Paus, Tomas, Pausova, Z., Poline, J.-B., Poustka, L., Rapp, M., Robert, G., Reuter, J., Rietschel, M., Ripke, S., Robbins, T., Rodehacke, S., Rogers, J., Romanowski, A., Ruggeri, B., Schmäl, C., Schmidt, D., Schneider, S., Schumann, M., Schubert, F., Schwartz, Y., Smolka, M., Sommer, W., Spanagel, R., Speiser, C., Spranger, T., Stedman, A., Steiner, S., Stephens, D., Strache, N., Ströhle, A., Struve, M., Subramaniam, N., Topper, L., Whelan, R., Williams, S., Yacubian, J., Zilbovicius, M., Wong, C.P., Lubbe, S., Martinez-Medina, L., Fernandes, A., Tahmasebi, A., 2015. Correlated gene expression supports synchronous activity in brain networks. Science 348, 1241–1244. https://doi.org/10.1126/science.1255905.
Seeley W.W. Menon V. Schatzberg A.F. Keller J. Glover G.H. Kenna H. Reiss A.L. Greicius M.D. Dissociable intrinsic connectivity networks for salience processing and executive control J. Neurosci. 27 2007 2349 2356 10.1523/JNEUROSCI.5587-06.2007 17329432
Stomby A. Salami A. Dahlqvist P. Evang J.A. Ryberg M. Bollerslev J. Olsson T. Johannsson G. Ragnarsson O. Elevated resting-state connectivity in the medial temporal lobe and the prefrontal cortex among patients with Cushing’s syndrome in remission Eur. J. Endocrinol. 180 2019 329 338 10.1530/EJE-19-0028 30939453
Tata D.A. Anderson B.J. The effects of chronic glucocorticoid exposure on dendritic length, synapse numbers and glial volume in animal models: Implications for hippocampal volume reductions in depression Physiol. Behav. 99 2010 186 193 10.1016/j.physbeh.2009.09.008 19786041
Tata D.A. Marciano V.A. Anderson B.J. Synapse loss from chronically elevated glucocorticoids: Relationship to neuropil volume and cell number in hippocampal area CA3 J. Comp. Neurol. 498 2006 363 374 10.1002/cne.21071 16871536
Thomas Yeo B.T. Krienen F.M. Sepulcre J. Sabuncu M.R. Lashkari D. Hollinshead M. Roffman J.L. Smoller J.W. Zöllei L. Polimeni J.R. Fischl B. Liu H. Buckner R.L. The organization of the human cerebral cortex estimated by intrinsic functional connectivity J. Neurophysiol. 106 2011 1125 1165 10.1152/jn.00338.2011 21653723
Valassi E. Crespo I. Keevil B.G. Aulinas A. Urgell E. Santos A. Trainer P.J. Webb S.M. Affective alterations in patients with Cushing’s syndrome in remission are associated with decreased BDNF and cortisone levels Eur. J. Endocrinol. 176 2017 221 231 10.1530/EJE-16-0779 27932530
van den Heuvel M.P. Sporns O. Network hubs in the human brain Trends Cogn. Sci. 17 2013 683 696 10.1016/j.tics.2013.09.012 24231140
Van Der Werff S.J.A. Pannekoek J.N. Andela C.D. Meijer O.C. Van Buchem M.A. Rombouts S.A.R.B. Van Der Mast R.C. Biermasz N.R. Pereira A.M. Van Der Wee N.J.A. Resting-state functional connectivity in patients with long-term remission of Cushing’s disease Neuropsychopharmacology 40 2015 1888 1898 10.1038/npp.2015.38 25652248
Wang X. Zhou T. Wang P. Zhang L. Feng S. Meng X. Yu X. Zhang Y. Dysregulation of resting-state functional connectivity in patients with Cushing’s disease Neuroradiology 61 2019 911 920 10.1007/s00234-019-02223-y 31101946
Webb S.M. Badia X. Barahona M.J. Colao A. Strasburger C.J. Tabarin A. van Aken M.O. Pivonello R. Stalla G. Lamberts S.W.J. Glusman J.E. Evaluation of health-related quality of life in patients with Cushing’s syndrome with a new questionnaire Eur. J. Endocrinol. 158 2008 623 630 10.1530/EJE-07-0762 18426820
Webb S.M. Valassi E. Quality of life impairment after a diagnosis of Cushing’s syndrome Pituitary 25 2022 768 771 10.1007/s11102-022-01245-9 35767164
Xia, M., Liu, J., Mechelli, A., Sun, X., Ma, Q., Wang, X., Wei, D., Chen, Y., Liu, B., Huang, C.-C., Zheng, Y., Wu, Y., Chen, T., Cheng, Y., Xu, X., Gong, Q., Si, T., Qiu, S., Lin, C.-P., Cheng, J., Tang, Y., Wang, F., Qiu, J., Xie, P., Li, L., DIDA-MDD Working Group, He, Y., 2022. Connectome gradient dysfunction in major depression and its association with gene expression profiles and treatment outcomes. Mol Psychiatry 27, 1384–1393. https://doi.org/10.1038/s41380-022-01519-5.
Xia Y. Xia M. Liu J. Liao X. Lei T. Liang X. Zhao T. Shi Z. Sun L. Chen X. Men W. Wang Y. Pan Z. Luo J. Peng S. Chen M. Hao L. Tan S. Gao J.-H. Qin S. Gong G. Tao S. Dong Q. He Y. Development of functional connectome gradients during childhood and adolescence Sci. Bull. (Beijing) 67 2022 1049 1061 10.1016/j.scib.2022.01.002 36546249
Yan C.-G. Chen X. Li L. Castellanos F.X. Bai T.-J. Bo Q.-J. Cao J. Chen G.-M. Chen N.-X. Chen W. Cheng C. Cheng Y.-Q. Cui X.-L. Duan J. Fang Y.-R. Gong Q.-Y. Guo W.-B. Hou Z.-H. Hu L. Kuang L. Li F. Li K.-M. Li T. Liu Y.-S. Liu Z.-N. Long Y.-C. Luo Q.-H. Meng H.-Q. Peng D.-H. Qiu H.-T. Qiu J. Shen Y.-D. Shi Y.-S. Wang C.-Y. Wang F. Wang K. Wang L. Wang X. Wang Y. Wu X.-P. Wu X.-R. Xie C.-M. Xie G.-R. Xie H.-Y. Xie P. Xu X.-F. Yang H. Yang J. Yao J.-S. Yao S.-Q. Yin Y.-Y. Yuan Y.-G. Zhang A.-X. Zhang H. Zhang K.-R. Zhang L. Zhang Z.-J. Zhou R.-B. Zhou Y.-T. Zhu J.-J. Zou C.-J. Si T.-M. Zuo X.-N. Zhao J.-P. Zang Y.-F. Reduced default mode network functional connectivity in patients with recurrent major depressive disorder Proc. Natl. Acad. Sci. U. S. A. 116 2019 9078 9083 10.1073/pnas.1900390116 30979801
Yuen E.Y. Wei J. Liu W. Zhong P. Yan Z. Repeated stress causes cognitive impairment by suppressing glutamate receptor expression and function in prefrontal cortex Neuron 73 2012 962 977 10.1016/j.neuron.2011.12.033IF: 16.2 Q1 22405206
Zung W.W. A self-rating depression scale Arch. Gen. Psychiatry 12 1965 63 70 10.1001/archpsyc.1965.01720310065008 14221692
Zung W.W. A rating instrument for anxiety disorders Psychosomatics 12 1971 371 379 10.1016/S0033-3182(71)71479-0 5172928
Yan CG, Wang XD, Zuo XN, Zang YF. DPABI: Data Processing & Analysis for (Resting-State) Brain Imaging. Neuroinformatics. 2016 Jul;14(3):339-51. doi: 10.1007/s12021-016-9299-4IF: 2.7 Q2 . PMID: 27075850.
