Exploring the Causal Relationships Between Brain Functional Networks and Psychiatric Disorders: A Mendelian Randomization Approach

doi:10.21203/rs.3.rs-5290369/v1

Download PDF

Research Article

Exploring the Causal Relationships Between Brain Functional Networks and Psychiatric Disorders: A Mendelian Randomization Approach

https://doi.org/10.21203/rs.3.rs-5290369/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Background:

Numerous studies have reported brain functional network impairments in individuals with psychiatric disorders; however, the causal relationships between the two remain unclear. We aimed to investigate the potential causal relationships between resting-state functional magnetic resonance imaging (rsfMRI) phenotypes and psychiatric disorders via Mendelian randomization (MR) analysis.

Method:

Employing a bidirectional two-sample MR analysis approach, this study assessed the associations between 191 rsfMRI phenotypes and 9 psychiatric disorders. Genetic variations were utilized as instrumental variables, ensuring the minimization of confounding factors in accordance with Mendel's laws of inheritance. Causal inferences were drawn by selecting genetic variants that were directly associated with the exposure variables and excluding those that might influence outcomes via alternative pathways. The study employed various statistical methods, including inverse variance weighting, the weighted median, and the MR Egger method, to evaluate causal relationships and adjusted for false discovery rates among outcomes.

Results:

The study identified significant causal associations between 21 rsfMRI phenotypes and five psychiatric disorders. For instance, in anxiety disorders, increased neural activity intensity in the parietal, frontal, and temporal lobes, along with enhanced functional connectivity between the attention, central executive, and default mode networks, are significantly associated with an increased risk of anxiety disorders. With respect to dementia, increased activity in the frontal lobe region was associated with a higher risk of dementia, and increased functional connectivity between the salience network and the central executive network was also linked to an increased risk of dementia.

Conclusion:

The findings of this study support the causal relationships between rsfMRI and psychiatric disorders, offering new insights for future prevention and treatment strategies.

rsfMRI

Psychiatric Disorders

Causal Inference

Mendelian Randomization

Brain Functional Networks

Psychiatric health disorders encompass a wide range of conditions that impact a person's emotions, thoughts, and actions. The spectrum of psychiatric disorders is broad, with anxiety disorders, obsessive‒compulsive disorder, and sleep disorders under its umbrella. A recent analysis in the Netherlands, spanning from 2007–2009 and then from 2019–2022, revealed a worrying upward trend in the occurrence of such disorders, with a particularly sharp increase noted among the younger demographic (aged 18–34) [1]. Furthermore, statistics from the World Health Organization continue to highlight the persistently high rates of these disorders within specific communities, imposing a significant toll on affected individuals and societies alike [2]. Psychological factors, environmental influences, trauma, and changes in brain structure and function may be involved in the occurrence of psychiatric disorders. Understanding the mechanisms of psychiatric disorders will help develop more effective prevention and treatment strategies, and provide more comprehensive care and support to patients.

The network structure and information communication within the human brain are crucial for our behavior and cognitive functions. Even when not performing specific tasks or receiving external stimuli (i.e., during a resting state), the brain remains active, exhibiting intrinsic functional patterns [3]. rsfMRI technology observes the spontaneous activity of the brain in this resting state by monitoring changes in blood oxygen level-dependent signals [4]. This technology not only offers critical insights into the functional organization and network communication of the brain but also plays a key role in revealing the neurobiological mechanisms of brain diseases.

Numerous studies have shown that dysfunction in brain functional networks plays a significant role in psychiatric illnesses. For example, compared with healthy subjects, patients with schizophrenia show reduced connectivity between the thalamus and cerebellum, but stronger connectivity with sensory regions, suggesting that thalamic connectivity dysfunction may be a core neurobiological feature of schizophrenia [5, 6]. Striatal connectivity dysfunction is also associated with schizophrenia [7]. Similarly, a rsfMRI study revealed increased thalamus‒sensory region connectivity in schizophrenia patients, whereas patients with bipolar disorder presented increased thalamus‒sensorimotor connectivity [8]. In adolescents with major depressive disorder, reduced functional connectivity between the amygdala and the hippocampus, parahippocampus and brainstem has been observed, with this reduced connectivity often leading to stronger negative emotions such as depression, irritability, and fatigue [9, 10]. A meta-analysis revealed that cognitive impairment in Parkinson’s disease patients is related to reduced connectivity in brain networks associated with cognition. MR analysis revealed causal relationships between brain functional networks and 12 psychiatric illnesses [11, 12]. These findings suggest a complex relationship between brain network dysfunction and various psychiatric disorders. Although evidence supports the association between these brain network dysfunctions and psychiatric illnesses, the causal relationships require further exploration. The application of methods such as MR may help clarify the causal relationships between psychiatric disorders and brain functional networks, providing a foundation for future mechanistic research.

In this study, we applied a bidirectional two-sample MR analysis method to investigate the potential causal relationships between 191 rsfMRI phenotypes and 9 psychiatric disorders. These findings provide strong evidence for understanding the causal relationship between brain functional abnormalities and psychiatric disorders.

Study design

Two-sample MR analysis was employed to assess the causal relationships between 191 rsfMRI phenotypes and 9 psychiatric disorders. MR infers causality on the basis of Mendel's laws of inheritance, allowing for the minimization of confounding factors [13, 14]. MR leverages genetic variants as proxies, or instrumental variables (IVs) for exposure risk factors, necessitating the selection of valid IVs that adhere to the three fundamental assumptions essential for inferring causality. First, the genetic variants should have a direct correlation with the exposure variables, indicating their capacity to modulate these characteristics and, by extension, potentially influence the progression of the disease. Second, there should be no underlying confounding factors between the genetic variants, exposure, and outcome, ensuring that the selected IVs are independent. Finally, the impact of the genetic variants on the outcome must be mediated exclusively through the exposure, ruling out the potential for influencing the outcome via alternative pathways.

Selection of IVs

To ensure the accuracy of the causal relationship, a range of criteria were applied to the selection of IVs. All included single nucleotide polymorphisms (SNPs) had to meet the genome-wide significance threshold, with a p-value less than 5E-08, to ensure a strong association with both the exposure and the outcome [15]. Linkage disequilibrium was controlled by setting limits for R² and kilobase pairs (kb), ensuring that the selected IVs were independent. Specifically, the R² threshold was set at 0.001, and kb was set at 10,00. Additionally, the strength of the IVs was evaluated by calculating the F-statistic, with an F value greater than 10 considered strong instruments, and an F value less than 10 classified as weak instruments and thus excluded [15, 16]. The formula for R² is R² = β²/ (β² + se² × N), where β represents the genetic effect size from the exposure data, se represents the standard error of the effect size, and N represents the sample size. The formula for the F- statistic is: F = R² (N − 2) / (1 − R²), where R² is the proportion of variance in the phenotype explained by the genetic variants, and N represents the sample size.

Statistical analysis

We performed this work via R 4.4.1 software on Windows. The main R packages used were twoSampleMR (v.0.6.6) and Mendelian randomization (v.0.10.0), and methods such as the inverse variance weighted (IVW), weighted median, weighted mode, simple mode, and MR Egger were employed. IVW was used to combine the causal effects of different genetic variants (e.g., SNPs). It allows proper weighting on the basis of the statistical precision of different SNPs, making it highly effective in assessing the impact of various genetic variants on the causal relationship between a given exposure and outcome. Therefore, we used the IVW method to determine the primary results [17]. The weighted median method estimates causal effects by assigning weights to genetic variants on the basis of their statistical precision and calculating the weighted median of effect sizes, which is particularly useful when genetic variants might affect multiple traits [18]. The weighted mode method assumes that the largest group of variants is valid and estimates causal effects by identifying and weighting variants with similar effects, which is suitable for cases where there are multiple effect estimates [19]. The simple mode method estimates causal effects by clustering the effect estimates of genetic variants and selecting the cluster with the highest mode. The MR Egger method is used to assess whether genetic variants exhibit horizontal pleiotropy relative to the mean outcome, which is determined by evaluating the intercept. A statistically significant intercept (p- value < 0.05) suggests the possibility of significant horizontal pleiotropy [20, 21], and such data were excluded. In addition, Cochran’s Q test was applied to determine heterogeneity between the variables, where a P- value < 0.05 indicated heterogeneity, and such data were excluded [22]. We performed false discovery rate (FDR) correction to control the proportion of false positives in multiple hypothesis testing, and data with an FDR P-value (PFDR) > 0.05 were excluded. Compared with the Bonferroni correction, the FDR is generally more powerful, as it maintains the ability to detect true effects while controlling the false discovery rate. In the forward MR analysis, rsfMRI was considered the exposure variable, and psychiatric disorders were treated as the outcome variables. In the reverse MR, the roles of rsfMRI and psychiatric disorders were reversed. Figure 1 shows the workflow of the MR analysis.

Data sources

In this study, we utilized rsfMRI data provided in the literature [23]. This study conducted an association analysis between brain activity patterns and genetic variants in participants from the UK Biobank (UKB). Given that the genetic influence on brain functional networks is typically smaller than that on brain structure, 191 patterns significantly associated with genetic variants (significance level of 5E-08) were selected from 1,777 different brain activity patterns. These patterns include 75 indicators reflecting the intensity of spontaneous brain activity in regions (nodes), 111 indicators describing functional connectivity between brain regions (edges), and 5 indicators describing overall brain functional connectivity (edges). These indicators cover several key brain networks, such as the Salience Network (SN), Default Mode Network (DMN), and Central Executive Network (CEN).

The data on psychiatric disorders were sourced from the FinnGen database, a large biobank project in Finland. The goal of this project is to enhance the understanding of various diseases by analyzing and researching the genetic and health data of large populations. FinnGen facilitates large-scale genome-wide association studies (GWASs) and other types of genetic research. We obtained GWASs for psychiatric disorders from the FinnGen database’s official website. Some psychiatric disorders with smaller sample sizes, such as obsessive‒compulsive disorder, anorexia nervosa, and social phobia, were not included in the study. Ultimately, we selected nine diseases for MR analysis: sleep apnea, dementia, mood disorders, sleep disorders, anxiety disorders, migraines, suicide or self-harm, personality disorders, and epilepsy. Supplemental table 1 summarizes the sample sizes for these diseases.

Forward MR analysis

In forward MR analysis, we identified 21 rsfMRI phenotypes that may have a causal relationship with 5 psychiatric disorders, with all PFDRs being less than 0.05.

The effects of rsfMRI traits on anxiety disorders

Anxiety disorders encompass a wide range of conditions, including generalized anxiety disorder, panic disorder, social anxiety disorder, specific phobias, agoraphobia, etc., and are characterized by persistent, excessive worry and fear [24]. Our study identified two rsfMRI phenotypes with causal relationships with anxiety disorders, as shown in Fig. 2. We found that the neural activity intensity in the parietal, frontal, and temporal regions was positively correlated with anxiety disorders. An increase of 1 standard deviation (s.d.) in functional connectivity between the attention, CEN, and DMN networks was associated with a 17% increase in the risk of anxiety disorders (IVW OR = 1.17, 95% CI: 1.07–1.28, P = 2.84E-04). Similarly, the neural activity intensity in the calcarine, lingual, and cuneus regions was also positively correlated with anxiety disorders. An increase of 1 s.d. in amplitude features of the visual network (VN) was associated with a 13% increase in the risk of anxiety disorders (IVW OR = 1.13, 95% CI: 1.06–1.21, P = 2.03E-04).

The effects of rsfMRI traits on dementia

Dementia is a syndrome characterized by cognitive decline, typically representing a collection of symptoms caused by various conditions rather than a specific disease [25]. In the forward MR analysis, using dementia as the outcome variable, we identified two rsfMRI phenotypes that had a causal relationship with dementia (Fig. 3). The study revealed that increased activity in the frontal region increased the risk of dementia, and a 1 s.d. increase in functional connectivity between the SN and CEN was associated with a 12% increase in dementia risk (IVW OR = 1.12, 95% CI: 1.05–1.20, P = 2.84E-04). Increased activity in the cuneus and occipital regions was closely related to the development of dementia, with a 1 s.d. increase in amplitude features of the VN associated with a 24% increase in the risk of dementia (IVW OR = 1.24, 95% CI: 1.10–1.39, P = 2.03E-04).

The effects of rsfMRI traits on mood disorders

Mood disorders, also known as affective disorders, are a group of psychiatric illnesses characterized by significant and persistent changes in mood or emotional state [26]. These disorders not only affect the emotional state of patients but are also often accompanied by changes in cognition, behavior, and physiological functions. Our study identified two rsfMRI phenotypes that have a positive causal relationship with mood disorders (Fig. 4). For example, neural activity in the angular gyrus, temporal regions, and precuneus was positively correlated with mood disorders; with a 1 s.d. increase in DMN and CEN functional connectivity, the risk of mood disorders increased by 21% (IVW OR = 1.21, 95% CI: 1.10–1.34, P = 1.04E-04). Additionally, neural activity in the postcentral and precentral regions was positively correlated with mood disorders, affecting mainly the motor network, and a 1 s.d. increase in motor network functional connectivity increased the risk of mood disorders by 11% (IVW OR = 1.11, 95% CI: 1.05–1.18, P = 2.17E-04).

The effects of rsfMRI traits on migraine

Migraine is a common chronic neurovascular disorder characterized by recurrent, pulsating, moderate-to-severe headaches. Migraine has a high prevalence, affecting more than one billion people worldwide [27]. Our study identified several rsfMRI phenotypes that have a positive causal relationship with migraine (Fig. 5). For example, neural activity in the parietal region was positively correlated with migraine, and a 1 s.d. increase in CEN and attention network functional connectivity increased the risk of migraine by 17% (IVW OR = 1.17, 95% CI: 1.10–1.26, P = 5.29E-06). Neural activity in the temporal and occipital regions was positively correlated with migraine, and a 1 s.d. increase in functional connectivity between the attention network and VN increased the risk of migraine by 16% (IVW OR = 1.16, 95% CI: 1.05–1.27, P = 2.17E-03). Neural activity in the cerebellum and temporal regions was also positively correlated with migraine, and a 1 s.d. increase in subcortical-cerebellum network functional connectivity increased the risk of migraine by 16% (IVW OR = 1.16, 95% CI: 1.05–1.28, P = 3.70E-03).

The effects of rsfMRI traits on sleep apnoea

Sleep apnoea syndrome is a common sleep disorder characterized by recurrent episodes of apnea and hypopnea during sleep, often accompanied by snoring, excessive daytime sleepiness, and other symptoms [28]. Through MR analysis, we identified five rsfMRI phenotypes that have a causal relationship with sleep apnoea (Fig. 6). For example, neural activity in the paracentral and postcentral regions was negatively correlated with sleep apnoea, and a 1 s.d. increase in motor network functional connectivity reduced the risk of sleep apnoea by 7% (IVW OR = 0.93, 95% CI: 0.89–0.97, P = 1.06E-03). Neural activity in the occipital region was negatively correlated with sleep apnoea, and a 1 s.d. increase in VN functional connectivity reduced the risk of sleep apnoea by 16% (IVW OR = 0.84, 95% CI: 0.78–0.92, P = 8.46E–05). Neural activity in the frontal region was negatively correlated with sleep apnoea, and a 1 s.d. increase in limbic network functional connectivity reduced the risk of sleep apnoea by 12% (IVW OR = 0.88, 95% CI: 0.82–0.95, P = 8.10E-04).

Reverse MR analysis

We conducted a reverse MR analysis, where psychiatric disorders were the exposure variables and rsfMRI phenotypes were the outcome variables, to observe the effect of psychiatric disorders on brain functional networks. The study revealed that dementia reduced the connectivity between the DMN, CEN, and motor networks (IVW OR = 0.96, 95% CI: 0.93–0.98, P = 8.38E-04, Fig. 7).

This study utilized the MR method to explore the potential causal relationships between brain functional networks and psychiatric disorders on the basis of the association between rsfMRI phenotypes and psychiatric illnesses. The results revealed significant associations between neural activity in multiple brain regions and specific psychiatric disorders, offering new perspectives on the neurobiological mechanisms underlying psychiatric illnesses.

Our findings related to anxiety disorders are consistent with those of previous studies. Functional abnormalities in the frontal lobe, particularly the prefrontal cortex, may be associated with the development and maintenance of anxiety disorders. Studies have shown that the prefrontal cortex plays a central role in cognitive control, emotional regulation, and threat assessment and that abnormalities in these functions may lead to anxiety symptoms [29]. The temporal lobe, especially the amygdala, is a key region for processing emotional responses, particularly those related to fear and anxiety. The hippocampus is also involved in emotional regulation and is closely connected to the amygdala, playing a role in generating anxiety-related emotions [30, 31]. The parietal lobe and lingual gyrus are also thought to play a role in processing anxiety-related bodily sensations [32, 33]. Patients with anxiety disorders frequently demonstrate abnormal functional connectivity within the DMN, especially between the prefrontal cortex and hippocampus, which may underlie symptoms such as rumination and excessive worry [34, 35]. Additionally, anxiety patients often exhibit weakened CEN function, particularly in the connections between the prefrontal and parietal cortices, potentially explaining their difficulties with emotional regulation and executive function tasks [36]. The SN in individuals with anxiety disorders tends to show hyperactivation, particularly in the anterior cingulate cortex and insula, leading to heightened sensitivity to negative stimuli and increased emotional reactivity [37]. Moreover, abnormalities in the VN have been linked to anxiety, possibly due to overprocessing of threatening visual stimuli in these patients [38].

In dementia research, we observed that increased functional connectivity in the frontal regions was linked to a greater risk of developing dementia. One study on brain connectivity in Alzheimer's disease (AD) patients revealed that those in the early stages of the disease presented unusually high activity in specific brain areas, such as the prefrontal cortex, which might reflect a compensatory mechanism in response to early neurodegenerative changes [39]. In the early phases of dementia, particularly in patients with frontotemporal dementia and AD, the SN is often compromised. The SN plays a crucial role in detecting important external stimuli and internal sensations relevant to behavior, as well as in shifting between the DMN and the CEN. This switching process is vital for maintaining cognitive flexibility and adapting to changes in the environment [40]. As dementia progresses, the function of the CEN becomes significantly impaired. The decline in CEN connectivity is strongly associated with overall cognitive deterioration in AD, manifesting as pronounced deficits in attention, problem-solving, working memory, and other executive functions [41]. The VN, which mainly involves the occipital lobe, is critical for processing and interpreting visual information. Research has suggested that alterations in VN function in dementia may be associated with cognitive decline and visual perception difficulties, particularly in conditions such as AD [42].

Our study revealed that abnormal activity in several brain regions plays a key role in the development and progression of mood disorders. These regions are critical for emotional regulation, and structural or functional abnormalities may be part of the pathophysiological basis of these conditions. For example, the precuneus, located in the parietal lobe, is involved in self-referential processing, episodic memory, and consciousness. In mood disorders, the precuneus may contribute to self-reflective thoughts and emotional experiences, and abnormalities in its structural or functional connectivity could be linked to depressive symptoms [43]. Additionally, the angular gyrus, which is primarily responsible for language, cognitive integration, and emotional processing, may also play a role in mood disorders. An fMRI study revealed that angular gyrus activity, which is related to abnormalities in negative emotion processing, emotional regulation, and self-referential information processing, was reduced in patients with depression. Changes in angular gyrus function may lead to emotional and cognitive impairments in depressed patients [44]. Changes in the functional connections of the motor network, DMN, and CEN all have adverse effects on mood disorders. The CEN is closely associated with cognitive control and working memory in emotional regulation. Studies have shown that in patients with mood disorders, such as depression, CEN functional connectivity is often weakened, which impairs patient’s ability to regulate emotions effectively. Additionally, the dysregulation of interactions between the CEN and DMN makes it difficult for patients to shift from self-reflection to more effective cognitive control, exacerbating the symptoms of mood disorders [45, 46].

In the study of migraine, we identified multiple brain functional networks associated with an increased risk of migraine. In migraine patients, abnormal SN function is thought to be related to pain perception and emotional and autonomic dysfunction. Studies have shown that the SN is hyperactive during migraine attacks, which may promote hypersensitivity to pain and excessive focus on headaches [47]. The DMN is associated primarily with self-reflection, mind wandering, and emotional regulation. Abnormal functional connectivity in the DMN has been found in migraine patients, particularly in those with chronic migraines. These changes in connectivity may cause patients to experience anxiety and negative emotions even between migraine attacks, potentially influencing both the frequency and intensity of migraines [48]. The CEN, which is associated with executive function, decision-making, and cognitive control, has been shown to exhibit abnormal functioning in migraine patients, particularly in relation to pain management and cognitive tasks. This may explain common symptoms such as difficulty concentrating, memory lapses, and impaired decision-making [49]. The VN plays a significant role in migraines, especially those accompanied by aura. Research has demonstrated that migraine patients exhibit abnormal functional connectivity and altered metabolic activity in the VN during the interictal phase, with these abnormalities being particularly pronounced in individuals experiencing visual aura [50]. Such dysfunctions are likely linked to visual perception disturbances, photophobia, and aura symptoms such as flashes of light or blind spots. Additionally, the subcortical-cerebellum network is crucial for pain modulation and multisensory integration, and plays a significant role in migraine pathophysiology. While the cerebellum was once primarily associated with motor control, it is now recognized for its role in cognitive and sensory processing, particularly in pain-related pathways. Its connections to the trigeminal pain pathway and subcortical structures such as the thalamus and basal ganglia suggest its involvement in regulating the pain experiences of migraine sufferers [51].

Our study revealed that neural activity in certain brain regions is associated with a reduced risk of sleep apnoea. Sleep apnoea patients, particularly those with obstructive sleep apnea and central sleep apnea, often exhibit impaired frontal lobe function. This is closely related to structural changes in the brain caused by sleep disruption and hypoxemia. Studies have shown that these patients experience a reduction in frontal gray matter volume, which may affect their cognitive abilities and emotional regulation [52]. The paracentral and postcentral lobes are responsible for processing sensory information and motor control. In patients with sleep apnoea, the somatosensory cortex may be affected by sleep disruption and hypoxia, leading to impairments in motor control and sensory processing [53]. The cingulate gyrus plays an important role in attention, emotional processing, and cognitive control. Sleep apnoea, particularly central sleep apnoea, may lead to dysfunction in this region, affecting patients' attention and emotional management abilities [54]. Although the occipital lobe is primarily responsible for visual processing, sleep apnoea may indirectly affect its function through broader impacts on brain function. Patients with sleep apnoea may experience impaired visual processing abilities, particularly in visual-spatial tasks [55].

Notably, our work is based on the FinnGen database, which primarily involves a European population, and therefore needs to be validated in other populations to confirm its generalizability and applicability. Additionally, since the MR method relies on the selection of genetic IVs, careful consideration of the strength of genetic IVs and potential confounders is necessary. Future research could further explore how neural activity in specific brain regions influences the occurrence and development of psychiatric disorders through molecular and cellular mechanisms.

In conclusion, the findings of this study provide a theoretical basis for the development of biomarkers based on brain functional networks, which may aid in the early diagnosis and prognosis of psychiatric disorders. Our results highlight the potential role of brain functional networks in the occurrence of psychiatric disorders and provide new insights for future prevention and treatment strategies. Future research can utilize these findings to develop new therapeutic approaches, such as noninvasive brain stimulation techniques targeting specific brain networks, to improve the clinical symptoms and quality of life of patients with psychiatric disorder.

rsfMRI

resting-state functional magnetic resonance imaging

Mendelian randomization

IVs

instrumental variables

SNPs

single nucleotide polymorphisms

kilobase pairs

IVW

Inverse Variance Weighted

FDR

False Discovery Rate

PFDR

FDR P-value

UKB

UK Biobank

Salience Network

DMN

Default Mode Network

CEN

Central Executive Network

GWAS

genome-wide association studies

Visual Network

Alzheimer's Disease

Ethics approval and consent to participate

The GWAS data we used is publicly available, and ethics approval and consent to participate have already been obtained.

Consent for publication

Not applicable

Data Availability

All GWAS statistics in this study are publicly available from the FinnGen R10 version. Here are the links to the summary statistics for each disorder:

Anxiety disorders: https://storage.googleapis.com/finngen-public-data-r10/summary_stats/finngen_R10_KRA_PSY_ANXIETY_EXMORE.gz

Dementia: https://storage.googleapis.com/finngen-public-data-r10/summary_stats/finngen_R10_F5_DEMENTIA.gz

Mood disorders: https://storage.googleapis.com/finngen-public-data-r10/summary_stats/finngen_R10_F5_MOOD.gz

Migraine: https://storage.googleapis.com/finngen-public-data-r10/summary_stats/finngen_R10_G6_MIGRAINE.gz

Sleep apnoea: https://storage.googleapis.com/finngen-public-data-r10/summary_stats/finngen_R10_G6_SLEEPAPNO_INCLAVO.gz

Sleep disorders: https://storage.googleapis.com/finngen-public-data-r10/summary_stats/finngen_R10_SLEEP.gz

Suicide or other Intentional self-harm: https://storage.googleapis.com/finngen-public-data-r10/summary_stats/finngen_R10_VWXY20_SUICI_OTHER_INTENTI_SELF_H.gz

Personality disorders: https://storage.googleapis.com/finngen-public-data-r10/summary_stats/finngen_R10_KRA_PSY_PERSON_EXMORE.gz

Epilepsy: https://storage.googleapis.com/finngen-public-data-r10/summary_stats/finngen_R10_G6_EPLEPSY.gz

The rsfMRI datasets can be accessed via Zenodo at the link https://zenodo.org/record/5775047 (as referenced in citation 23). The original data accompanying this paper are also provided.

Competing interests

The authors declare no competing interests.

Funding

This study did not receive support from any institution or project.

Author contributions

Q.X. and Y.S. contributed equally to the work presented in this paper and were considered co-first authors.

W.W. was the mastermind behind the study, orchestrating its design and overseeing its progress. Q. X. and Y.S. took on the data analysis and initial manuscript drafting. J.S. handled the code adjustments and provided technical support. W. W. also took charge of the manuscript revisions and finalization. Each author played a role in refining the manuscript, offering critical feedback, and ultimately endorsing the finished product.

Acknowledgements

We thank J.S. for his assistance with technical support and guidance in code writing.

Supplementary Information

Supplementary File Contents: R language code；Psychiatric disorder GWAS summary statistics (Table S1)；Information of instrumental variables (IVs) for significant exposure-outcome pairs in forward and reverse Mendelian Randomization (MR) analyses (Tables S2 and S3)；Significant results of forward and reverse MR (Tables S4 and S5); STROBE-MR checklist; Summary of sensitivity analysis results (such as forest, scatter and leave‒one‒out analysis) and phenotypes (Table S6).

Clinical trial number: not applicable.

Ten Have M, Tuithof M, van Dorsselaer S, et al. Prevalence and trends of common mental disorders from 2007–2009 to 2019–2022: results from the Netherlands Mental Health Survey and Incidence Studies (NEMESIS), including comparison of prevalence rates before vs. during the COVID-19 pandemic. World Psychiatry. 2023;22(2):275–85. 10.1002/wps.21087.
Charlson F, van Ommeren M, Flaxman A, et al. New WHO prevalence estimates of mental disorders in conflict settings: a systematic review and meta-analysis. Lancet. 2019;394(10194):240–8. 10.1016/S0140-6736(19)30934-1.
Zhao B, Li T, Smith SM, et al. Common variants contribute to intrinsic human brain functional networks. Nat Genet. 2022;54(4):508–17. 10.1038/s41588-022-01039-6.
Lee MH, Smyser CD, Shimony JS. Resting-state fMRI: a review of methods and clinical applications. AJNR Am J Neuroradiol. 2013;34(10):1866–72. 10.3174/ajnr.A3263.
Ferri J, Ford JM, Roach BJ, et al. Resting-state thalamic dysconnectivity in schizophrenia and relationships with symptoms. Psychol Med. 2018;48(15):2492–9. 10.1017/S003329171800003X.
Anticevic A, Haut K, Murray JD, et al. Association of Thalamic Dysconnectivity and Conversion to Psychosis in Youth and Young Adults at Elevated Clinical Risk. JAMA Psychiatry. 2015;72(9):882–91. 10.1001/jamapsychiatry.2015.0566.
Horga G, Cassidy CM, Xu X, et al. Dopamine-Related Disruption of Functional Topography of Striatal Connections in Unmedicated Patients With Schizophrenia. JAMA Psychiatry. 2016;73(8):862–70. 10.1001/jamapsychiatry.2016.0178.
Skåtun KC, Kaufmann T, Brandt CL, et al. Thalamo-cortical functional connectivity in schizophrenia and bipolar disorder. Brain Imaging Behav. 2018;12(3):640–52. 10.1007/s11682-017-9714-y.
Cullen KR, Westlund MK, Klimes-Dougan B, et al. Abnormal amygdala resting-state functional connectivity in adolescent depression. JAMA Psychiatry. 2014;71(10):1138–47. 10.1001/jamapsychiatry.2014.1087.
Connolly CG, Ho TC, Blom EH, et al. Resting-state functional connectivity of the amygdala and longitudinal changes in depression severity in adolescent depression. J Affect Disord. 2017;207:86–94. 10.1016/j.jad.2016.09.026.
Wolters AF, van de Weijer SC, Leentjens AF, et al. Resting-state fMRI in Parkinson's disease patients with cognitive impairment: A meta-analysis. Parkinsonism Relat Disord. 2019;62:16–27. 10.1016/j.parkreldis.2018.12.016.
Mu C, Dang X, Luo XJ. Mendelian randomization analyses reveal causal relationships between brain functional networks and risk of psychiatric disorders. Nat Hum Behav. 2024;8(7):1417–28. 10.1038/s41562-024-01879-8.
Holmes MV, Ala-Korpela M, Smith GD. Mendelian randomization in cardiometabolic disease: challenges in evaluating causality. Nat Rev Cardiol. 2017;14(10):577–90. 10.1038/nrcardio.2017.78.
Larsson SC, Butterworth AS, Burgess S. Mendelian randomization for cardiovascular diseases: principles and applications. Eur Heart J. 2023;44(47):4913–24. 10.1093/eurheartj/ehad736.
Tang C, Lei X, Ding Y, et al. Causal relationship between immune cells and neurodegenerative diseases: a two-sample Mendelian randomisation study. Front Immunol. 2024;15:1339649. 10.3389/fimmu.2024.1339649.
Gu J, Qiao Y, Cong S. Causal role of immune cells on risk of Parkinson's disease: a Mendelian randomization study. Front Aging Neurosci. 2024;16:1368374. 10.3389/fnagi.2024.1368374.
Mounier N, Kutalik Z. Bias correction for inverse variance weighting Mendelian randomization. Genet Epidemiol. 2023;47:314–31. 10.1002/gepi.22393.
Bowden J, Davey Smith G, Haycock PC, et al. Consistent Estimation in Mendelian Randomization with Some Invalid Instruments Using a Weighted Median Estimator. Genet Epidemiol. 2016;40(4):304–14. 10.1002/gepi.21965.
Burgess S, Foley CN, Allara E, et al. A robust and efficient method for Mendelian randomization with hundreds of genetic variants. Nat Commun. 2020;11(1):376. 10.1038/s41467-019-14156-4.
Burgess S, Thompson SG. Interpreting findings from Mendelian randomization using the MR-Egger method. Eur J Epidemiol. 2017;32(5):377–89. 10.1007/s10654-017-0255-x.
Cho Y, Haycock PC, Sanderson E, et al. Exploiting horizontal pleiotropy to search for causal pathways within a Mendelian randomization framework. Nat Commun. 2020;11(1):1010. 10.1038/s41467-020-14452-4.
Kulinskaya E, Dollinger MB, Bjørkestøl K. On the moments of Cochran's Q statistic under the null hypothesis, with application to the meta-analysis of risk difference. Res Synth Methods. 2020;11(6):920. 10.1002/jrsm.1446.
Zhao B, Li T, Smith SM et al. Common variants contribute to intrinsic human brain functional networks. Nat Genet, 54(4), 508–17. https://doi.org/10.1038/s41588-022-01039-6
Penninx BW, Pine DS, Holmes EA, et al. Anxiety disorders. Lancet. 2021;397(10277):914–27. 10.1016/S0140-6736(21)00359-7.
Hafiz R, Alajlani L, Ali A, et al. The Latest Advances in the Diagnosis and Treatment of Dementia. Cureus. 2023;15(12):e50522. 10.7759/cureus.50522.
Datta S, Suryadevara U, Cheong J, Minneapolis. Minn). 2021;27(6):1712–37. 10.1212/CON.0000000000001051.
Pleș H, Florian IA, Timis TL, et al. Migraine: Advances in the Pathogenesis and Treatment. Neurol Int. 2023;15(3):1052–105. 10.3390/neurointl15030067.
Lee JJ, Sundar KM. Evaluation and Management of Adults with Obstructive Sleep Apnea Syndrome. Lung. 2021;199(2):87–101. 10.1007/s00408-021-00426-w.
Bishop SJ. Neurocognitive mechanisms of anxiety: an integrative account. Trends Cogn Sci. 2007;11(7):307–16. 10.1016/j.tics.2007.05.008.
Fetterhoff D, Costa M, Hellerstedt R, et al. Neuronal population representation of human emotional memory. Cell Rep. 2024;43(4):114071. 10.1016/j.celrep.2024.114071.
Kenwood MM, Kalin NH, Barbas H. The prefrontal cortex, pathological anxiety, and anxiety disorders. Neuropsychopharmacology. 2022;47(1):260–75. 10.1038/s41386-021-01109-z.
Lyamzin D, Benucci A. The mouse posterior parietal cortex: Anatomy and functions. Neurosci Res. 2019;140:14–22. 10.1016/j.neures.2018.10.008.
Couvy-Duchesne B, Strike LT, de Zubicaray GI et al. Lingual Gyrus Surface Area Is Associated with Anxiety-Depression Severity in Young Adults: A Genetic Clustering Approach. eNeuro. 2018;5(1):ENEURO.0153-17.2017. 10.1523/ENEURO.0153-17.2017
Tao Y, Liu B, Zhang X, et al. The Structural Connectivity Pattern of the Default Mode Network and Its Association with Memory and Anxiety. Front Neuroanat. 2015;9:152. 10.3389/fnana.2015.00152.
Chen X, Song X, Öngür D, et al. Association of default-mode network neurotransmitters and inter-network functional connectivity in first episode psychosis. Neuropsychopharmacology. 2023;48(5):781–8. 10.1038/s41386-023-01546-y.
Sylvester CM, Corbetta M, Raichle ME, et al. Functional network dysfunction in anxiety and anxiety disorders. Trends Neurosci. 2012;35(9):527–35. 10.1016/j.tins.2012.04.012.
Robinson OJ, Pike AC, Cornwell B, et al. The translational neural circuitry of anxiety. J Neurol Neurosurg Psychiatry. 2019;90(12):1353–60. 10.1136/jnnp-2019-321400.
Kveraga K. Visual Perception of Threat: Structure, Dynamics, and Individual Differences. In: Wang JZ, Adams RB, editors. Modeling Visual Aesthetics, Emotion, and Artistic Style. Cham: Springer; 2024. 10.1007/978-3-031-50269-9_5.
Jones DT, Machulda MM, Vemuri P, et al. Age-related changes in the default mode network are more advanced in Alzheimer disease. Neurology. 2011;77(16):1524–31. 10.1212/WNL.0b013e318233b33d.
Seeley WW, Zhou J, Kim EJ. Frontotemporal dementia: what can the behavioral variant teach us about human brain organization? Neuroscientist. 2012;18(4):373–85. 10.1177/1073858411410354.
Brier MR, Thomas JB, Snyder AZ, Benzinger TL, Zhang D, Raichle ME, Holtzman DM, Morris JC, Ances BM. Loss of intranetwork and internetwork resting state functional connections with Alzheimer's disease progression. J Neurosci. 2012;32(26):8890–9. 10.1523/JNEUROSCI.5698-11.2012.
Lehmann M, Crutch SJ, Ridgway GR, et al. Cortical thickness and voxel-based morphometry in posterior cortical atrophy and typical Alzheimer's disease. Neurobiol Aging. 2011;32(8):1466–76. 10.1016/j.neurobiolaging.2009.08.017.
Dadario NB, Sughrue ME. The functional role of the precuneus. Brain. 2023;146(9):3598–607. 10.1093/brain/awad181.
Kang L, Wang W, Zhang N, et al. Anhedonia and dysregulation of an angular gyrus-centred and dynamic functional network in adolescent-onset depression. J Affect Disord. 2023;324:82–91. 10.1016/j.jad.2022.12.057.
Bertocci MA, Afriyie-Agyemang Y, Rozovsky R, et al. Altered patterns of central executive, default mode and salience network activity and connectivity are associated with current and future depression risk in two independent young adult samples. Mol Psychiatry. 2023;28:1046–56. 10.1038/s41380-022-01899-8.
Blume J, Dhanasekara CS, Kahathuduwa CN, et al. Central Executive and Default Mode Networks: An Appraisal of Executive Function and Social Skill Brain-Behavior Correlates in Youth with Autism Spectrum Disorder. J Autism Dev Disord. 2024;54:1882–96. 10.1007/s10803-023-05961-4.
Xue T, Yuan K, Zhao L, et al. Intrinsic brain network abnormalities in migraines without aura revealed in resting-state fMRI. PLoS ONE. 2012;7(12):e52927. 10.1371/journal.pone.0052927.
Hubbard CS, Khan SA, Keaser ML, et al. Altered brain network integrity in migraine without aura. Cephalalgia. 2014;34(5):407–16. 10.1177/0333102413517014.
Coppola G, Di Renzo A, Tinelli E, et al. Thalamo-cortical network activity during spontaneous migraine attacks. Neurology. 2016;87(20):2154–60. 10.1212/WNL.0000000000003327.
Marzoli SB, Criscuoli A. The role of visual system in migraine. Neurol Sci. 2017;38(Suppl 1):99–102. 10.1007/s10072-017-2890-0.
Matoso A, Fouto AR, Esteves I, et al. Involvement of the cerebellum in structural connectivity enhancement in episodic migraine. J Headache Pain. 2024;25(1):154. 10.1186/s10194-024-01854-8.
Alchanatis M, Deligiorgis N, Zias N, et al. Frontal brain lobe impairment in obstructive sleep apnoea: a proton MR spectroscopy study. Eur Respir J. 2004;24(6):980–6. 10.1183/09031936.04.00127603.
Liu X, Ma Y, Ouyang R, et al. The relationship between inflammation and neurocognitive dysfunction in obstructive sleep apnea syndrome. J Neuroinflammation. 2020;17(1):229. 10.1186/s12974-020-01905-2.
Weihs A, Frenzel S, Grabe HJ. The Link Between Obstructive Sleep Apnoea and Neurodegeneration and Cognition. Curr Sleep Med Rep. 2021;7:87–96. 10.1007/s40675-021-00210-5.
Bao J, Zhao Z, Qin S, et al. Elucidating the association of obstructive sleep apnea with brain structure and cognitive performance. BMC Psychiatry. 2024;24(1):338. 10.1186/s12888-024-05789-x.

No competing interests reported.

Download PDF

Version 1

posted

You are reading this latest preprint version

Exploring the Causal Relationships Between Brain Functional Networks and Psychiatric Disorders: A Mendelian Randomization Approach

Status:

Version 1

Abstract

Background:

Method:

Results:

Conclusion:

Figures

Background

Methods

Study design

Selection of IVs

Statistical analysis

Data sources

Results

Forward MR analysis

The effects of rsfMRI traits on anxiety disorders

The effects of rsfMRI traits on dementia

The effects of rsfMRI traits on mood disorders

The effects of rsfMRI traits on migraine

The effects of rsfMRI traits on sleep apnoea

Reverse MR analysis

Discussion

Abbreviations

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1