Transnosographic features suggest that suicidal behavior has its own physiopathology and should be considered a specific clinical entity. According to the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5),1 suicidal behavior could be a separate diagnosis, emphasizing the relevance of identifying specific biomarkers.2
Identifying the neural substrates of suicidal behavior is key to understanding suicide etiology and might help to reduce suicide rates among psychiatric patients by promoting the development of novel therapeutic strategies based on behavioral neuroscience and brain stimulation.
Neuroimaging studies have shown structural and functional brain abnormalities in the prefrontal cortex, insula, and striatal regions and also in the connections between these brain areas.3 Disturbances in the structure and function of brain regions can lead to brain network alterations, including the ability of brain regions to coordinate their activity in a system. Resting-state functional magnetic resonance imaging (rs-fMRI) has been used to explore the intrinsic activity or resting state brain networks. In their review, Serafini et al4 showed that vulnerability to suicidal behavior might be associated with abnormalities of the frontolimbic and frontoparietal-cerebellar pathways and with negative blood-oxygenation-level–dependent (BOLD) responses in the anterior regions of the default mode network. Moreover, a recent electroencephalography-based study reported that the regulation of intrinsic brain networks is altered in patients with depression and history of suicidal attempt in the last 6 months compared with healthy controls (HCs).5 However, differences in acquisition/analysis methods and in population characteristics among fMRI studies limit the finding interpretability. Some studies focused on adolescents,6–9 and others included adults.10–14 Importantly, “suicidality” has been operationalized differently among studies (eg, suicidal ideation, history of suicide attempts, or composite assessment).8,15
Suicidal ideation and behaviors are still considered a symptom or a consequence of a concomitant psychiatric disorder, mainly major depressive disorder (MDD).16 However, rs-fMRI studies showed that in patients with MDD, resting-state functional connectivity (rs-FC) of the left amygdala-right insula/left superior orbitofrontal area and right amygdala-left middle temporal area was higher in depressed suicide attempters (SAs) than in controls.12 Such enhanced rs-FC in limbic regions may contribute to emotional dysregulation. Moreover, the functional network connectivity between insular-default mode network and insular-cerebellum was decreased in SAs, reflecting impaired internal processing.11 In addition, the amplitude of low-frequency fluctuation (ALFF) values were significantly higher in the right superior and left middle temporal and occipital regions and significantly lower in the left superior and middle frontal gyri in SAs than in controls.6,17 Lastly, regional homogeneity (ReHo) in the right posterior cingulate cortex was increased in SAs compared with nonattempters.18 It is important to disentangle the neural underpinnings of suicidal ideation and attempt.19,20 Indeed, connectivity patterns between cognitive control network and salience and default mode networks were different in SAs and suicide ideators.19 A study showed that ALFF value changes (decreased in the frontoparietal network and increased in hippocampus) could be used to distinguish SAs from suicide ideators, psychiatric controls, and HCs.20
Two meta-analyses reported that the blood concentrations of the proinflammatory cytokines interleukin (IL)-1β and IL-6 are significantly increased and those of the anti-inflammatory IL-2 are decreased in SAs compared with nonsuicidal patients and HCs.21,22 Similarly, C-reactive protein (CRP) concentration in blood was higher in SAs than in depressed nonattempters, independently of the interval after the suicide attempt.23 Interestingly, CRP basal level has been associated with rs-FC in the default mode network, including in the ventromedial prefrontal cortex,24–26 that in turn was correlated with anhedonia severity in patients with MDD.25 Moreover, IL-6 covaried with connectivity of the medial prefrontal cortex in HCs27 and in patients with bipolar disorder.28 However, data on associations between cytokines and rs-fMRI metrics in patients with MDD are still lacking. Moreover, it is important to investigate homogeneous populations while controlling for the presence of major depression to better characterize the association between suicidal, resting-state activity, and inflammatory status.29
Therefore, the aims of our study were (1) to identify neural markers in a homogeneous sample of SAs using rs-fMRI and (2) to find possible associations between rs-fMRI metrics/connectivity, suicidal phenotype, and peripheral inflammation markers in patients with MDD. We hypothesized that between-group differences of intrinsic activity in prefrontal areas may be independent of the suicidal ideation intensity and may correlate with peripheral proinflammatory cytokines.
METHODS
Participants
The study sample, recruited between February 1 and November 30, 2017, included 20 HCs (individuals without any history of psychiatric disorders according to the DSM-5 criteria and no lifetime history of suicide attempt) and 42 patients with MDD [21 SAs with history of suicide attempt in the last 8 days and 21 affective controls (ACs) without lifetime history of suicide attempt].
HCs were recruited through advertisement. Patients (SAs and ACs) were recruited among the inpatients of the Department of Emergency Psychiatry and Acute Care, Montpellier Academic Hospital, France. All participants were right-handed, and Caucasian adults aged between 18 and 70 years. All patients (ACs and SAs) had a current depressive episode according to the DSM-5 criteria. Exclusion criteria were current neurological or inflammatory disease; CRP >10 mg/L; current antibiotic or anti-inflammatory treatment; history of head trauma with loss of consciousness; current substance misuse, excluding tobacco; lifetime history of manic or hypomanic episode, schizophrenia, or schizoaffective disorder according to the DSM-5; pregnancy and breastfeeding; and contraindication to magnetic resonance imaging.
The study was conducted in accordance with the Consolidated Standards of Reporting Trials ethical guidelines. All participants gave their written informed consent. The study was approved by Montpellier University Hospital (CPP Sud Méditerranée IV) ethics committee (N°ID-RCB: 2016-01267-44 (CPP Sud Méditerranée IV) and registered in ClinicalTrials.gov (NCT03052855). Participants received a financial compensation of 50 euros.
Clinical Assessment
Participants underwent a standardized interview led by a psychiatrist to determine the depression score and to assess psychopathology according to the DSM-5 criteria using the Mini-International Neuropsychiatric Interview, version 7.0.0. Depressive symptomatology was assessed using the clinician-rated Inventory of Depressive Symptomatology,30 current psychological pain using a 10- point Likert scale,31 suicidal ideation intensity in the week before inclusion using the Columbia-Suicide Severity Rating Scale (C-SSRS),32 and suicidal intent with the Suicidal Intent Scale.33 The number of past suicide attempts and the lethality of the last attempt were recorded. All psychotropic treatments (name and daily dose) in the last 24 hours before MRI were recorded, and the medication load was calculated for each patient.34
Laboratory Analyses
Blood samples were collected after overnight fasting, between 7:00 and 9:00 AM, the day after admission, in 5- mL ethylenediaminetetraacetic acid–coated tubes (BD Vacutainer, Franklin Lakes, NJ). High-sensitivity CRP in serum was measured with the immunoturbidimetric method and a cobas 8000 modular analyzer (Roche Diagnostic, Meylan, France). Reagents and calibrators were used according to the manufacturer’s guidelines with analytic measuring ranges set at 0.3–350 mg/L.
Plasma was separated by centrifugation at 3,000 ×g at room temperature for 20 minutes and stored at −80°C for batch analysis by multiplex or enzyme-linked immunosorbent assay of IL-1β, IL-6, and IL-2, and tumor necrosis factor-α (TNF-α). Analytes with >30% missing data were excluded from the analysis (IL-1β and IL-2).
Brain MRI
Brain MRI was performed using a 3T MRI scanner with a 32-channel head coil (MAGNETOM Skyra, Siemens Healthcare, Erlangen, Germany) at the I2FH platform, Neuroradiology Department, Montpellier Hospital.
A high-resolution 3DT1 magnetization-prepared rapid gradient-echo sequence was acquired with the following parameters: repetition time (TR)/echo time (TE) = 2,300/2.900 ms, time of inversion = 900 ms, flip angle = 9°, voxel size = 1.2 ×1 ×1 mm3, and 176 sagittal slices. During rs-fMRI gradient echo-echo planar image acquisition (TR/TE = 3,000/30 ms, flip angle = 80°, voxel size = 3.4 × 3.4 × 3.4 mm3, 42 transverse slices, 200 volumes measures, and acquisition time = 10 min), participants were instructed to keep the eyes closed, lie still, and not think about anything in particular.
MRI Data Preprocessing
Imaging data were preprocessed using MATLAB (The MathWorks Inc, MA) and SPM12 (http://www.fil.ion.ucl.ac.uk/spm/software/spm12/). All images were reoriented using the anterior commissure as reference. Standard Diffeomorphic Anatomical Registration Through Exponentiated Lie Algebra (DARTEL) segmentation was performed on T1 images to extract gray matter, white matter, and cerebrospinal fluid posterior probability distribution. A dedicated DARTEL template was then used to normalize all T1 images (Ashburner).35
Resting-state images were processed using a standard processing pipeline that included signal heterogeneity correction using ANTS NI4TK, slice timing, motion correction, coregistration on the T1 images, normalization (based on the T1 images), and spatial smoothing (full width at half maximum = 6 mm).
Resting-State Metrics
ALFF, ReHo, and voxel-mirrored homotopic connectivity (VMHC) were estimated to evaluate local alterations of signal fluctuations.
VMHC is considered one of the most salient features of the brain intrinsic functional architecture36 because it assesses the functional homotopy (ie, interhemispheric rs-FC) by quantifying the functional connectivity between a voxel and its mirrored counterpart in the contralateral hemisphere. ALFF, ReHo, and VMHC maps were prepared with the REST toolbox (Resting State fMRI Data Analysis Toolkit V1.8, REST-Group, http://www.restfmri.net). Functional images were detrended and filtered (0.01–0.08 Hz). Smoothing was performed after ReHo score calculation to avoid artificially increasing the similarity between neighboring voxels. Maps were z-scored. The average gray matter ALFF, ReHo, and VMHC values were extracted from 116 regions using the Automated Anatomical Labelling (AAL) parcellation atlas (AAL 116 toolbox37). Based on our previous fMRI studies38–40 and literature data,3 analyses focused on regions implicated in suicidal vulnerability (Supplementary Figure 1): (1) orbitofrontal cortex (OFC), particularly the inferior frontal orbital gyrus, the middle frontal orbital gyrus, and the medial frontal orbital gyrus; (2) ventrolateral prefrontal cortex (VLPFC), particularly the pars opercularis of the inferior frontal gyrus (IFG); (3) dorsal prefrontal cortex; particularly the middle frontal gyrus; (4) anterior cingulate cortex (ACC); and (5) insula. The corresponding regions in the AAL atlas included 7 regions in each hemisphere.
Resting-State Functional Connectivity
A region of interest–to–region of interest (ROI-to ROI[tnq_query9])–based analysis was performed using the CONN toolbox (version 21.a, www.nitrc.org/projects/conn) based on the chosen a priori regions previously described with the AAL atlas. Additional preprocessing included denoising with voxel-wise removal of linear trends over each patient’s rs-fMRI dataset and temporal low-pass filtering (0.009 Hz < f < 0.08 Hz). The BOLD signal of white matter, cerebrospinal fluid, and movement parameters were used as covariates to remove unwanted physiologic and motion artifact effects. First-level analyses used a weighted general linear model to estimate the Fisher transformed bivariate correlation coefficients between ROIs (gray matter only). Second-level, random-effects analyses were carried out for the rs-FC values on a ROI to-ROI basis.
Statistical Analyses
Statistical analyses were performed with R version 4.1.2 (R Foundation for Statistical Computing, Vienna, Austria). A general linear model was used to model each regional measure as an outcome, while adjusting for age, sex, and education level. To identify the main group effects, a type II analysis of variance was performed with the fitted model. A false discovery rate (FDR) procedure was used for each set of rs-fMRI metrics and for rs-FC to correct for multiple comparisons, and post hoc analyses were performed for regions with a FDR P < .05. Exploratory correlation analyses were performed to examine possible associations between rs-fMRI metrics with a main group effect and clinical variables (depression, psychological pain intensity, and suicidal ideation intensity) in all patients as well as suicidal intent and lethality of last suicide attempt in SAs. Exploratory correlation analyses were also performed to identify possible associations between rs-fMRI metrics and peripheral inflammation markers in all patients (adjustment for sex and age). All correlation analyses were followed by multiple comparison correction (FDR).
RESULTS
Description of the Sample
The whole sample included 46 (74%) women, and the mean age was 44 (SD = 14) years. The mean IDSC score was 35 (SD = 25) (ie, moderate depressive symptomatology), without between-group differences. The mean C-SSRS score (suicidal ideation intensity) was 11 (SD = 10) and was higher in SAs than in ACs. The mean CRP concentration was 2.7 (SD = 3.4) mg/L and was higher in SAs than in HCs. The sociodemographic and clinical characteristics of the different groups are presented in Table 1.
Main Group Effect on rs-fMRI
Overall, a main group effect on ALFF values was observed in the left and right pars opercularis of IFG (F = 6.75, P = .033 and F = 5.66, P = .040, respectively). After post hoc correction for multiple testing, the ALFF values in the pars opercularis of IFG were lower in SAs than in controls (ACs and HCs) (Figure 1). This result remained significant after adjustment for suicidal ideation intensity and CRP concentration (not shown). There was no main group effect on ReHo and VMHC values.
Overall, main group effects on rs-FC were mainly observed between ROIs of the OFC and ACC and within the OFC. After post hoc correction for multiple testing, the rs-FC values of bilateral ACC and left pars orbitalis of the IFG and middle frontal orbital gyrus were higher in SAs compared to ACs and HCs. The rs-FC values of the left middle frontal orbital gyrus and right inferior frontal orbital gyrus were higher in patients (SAs and ACs) than in HCs. The rs-FC values of inferior frontal orbital gyrus, left ACC, left middle frontal orbital gyrus, and right ACC were lower in ACs than in SAs and HCs (Figure 2).
Correlations Between rs-fMRI, Clinical Measures, and Inflammatory Markers in All Patients (ACs and SAs)
The correlations between suicidal ideation intensity and all rs-fMRI metrics in the pars opercularis (except for the ALFF and ReHo of the left side) were significant after FDR correction (Table 2). Moreover, the correlation between psychological pain and rs-FC of left middle frontal gyrus and left insula remained significant after FDR correction (Supplementary Tables 1–4). No correlation between peripheral inflammation markers (CRP, IL-6, and TNF-α), rs-fMRI metrics (Table 3), and rs-FC (Supplementary Tables 6 and 7) remained significant after FDR correction in patients (ACs and SAs).
DISCUSSION
The main result of this study is the association between resting-state activity of the pars opercularis of the IFG and suicide attempt and suicidal ideation in patients with MDD. Specifically, the difference in ALFF values for the left pars opercularis of the IFG between SAs and controls remained significant even after controlling for suicidal ideation, despite the negative correlation between ALFF values in the same region and suicidal ideation intensity. This should encourage us to thoroughly investigate the hypothesis that the neurobiology underpinnings of suicidal ideation and act overlap but are not identical.41 Our results are in accordance with previous studies showing lower ALFF values in the VLPFC in SAs compared with depressed patients20,42 as well as decreased IFG functional connectivity in SAs compared with HCs43 and in patients at high compared with low suicidal risk.15 As ALFF is directly correlated with task-based BOLD responses,44 our results may indicate altered functioning of the pars opercularis associated with impaired verbal fluency45 and response inhibition,46 both involved in the suicidal process.47,48 Indeed, individuals who are less able to verbalize emotions or distress are more likely to use nonverbal communication, such as self-harm.49 Moreover, impaired response inhibition in patients with left IFG damage is a significant predictor of future suicidal ideation or act.47 Our results strengthen the hypothesis that altered activity of the pars opercularis of the IFG may exacerbate suicidal ideation and act, possibly through its association with language impairment and its involvement in the circuits mediating cognitive control.
Our second main result is that connectivity between ACC and orbital parts of IFG and middle frontal gyrus was higher in SAs than in controls (ACs and HCs). The lower rs-FC in ACs than in HCs suggests that there is not a continuum between HCs, ACs, and SAs, but rather a specific pathophysiology of suicidal vulnerability which does not overlap with depression vulnerability. The ACC receives information from the OFC and is implicated in learning associations between actions and reward vs punishment outcomes.50 It has been shown that the rs-FC between these regions is altered in patients with depression and suicidal ideation.51 Interestingly, studies in HCs found that the rs-FC of OFC and ACC was positively associated with risk taking (ie, the tendency to take risks when making decisions or aim for goals)52 and risk preference (ie, the preference for risky choices over safe alternatives) with a mediating effect of self-control.53 Our results add to the hypothesis of the involvement of a dysfunctional frontocingulate connectivity that may underlie altered decision-making toward risk,54 impulsivity, and self-control55 in SAs.
Our findings strengthen the relevance of meditation based psychotherapies for SAs. Meditation modifies55 the activity of prefrontal brain regions,56 as indicated by the ALFF increase during meditation compared with resting.57 Moreover, emotional regulation, in the framework of mindfulness training, has been linked to corticolimbic modulation.56 It has been shown that mindfulness training decreases the association between depressive symptomatology and suicidal ideation in patients with history of suicidal depression.58 Moreover, in SAs, acceptance and commitment therapy is more effective than relaxation in reducing the intensity of suicidal ideation, even after 3 months of follow-up.59 Meditation-based psychotherapies also improve decision making60 that is altered in SAs due to altered VLPFC functioning.40
The analysis of peripheral inflammation markers showed that high CRP concentration was associated with history of suicidal acts, as previously reported.23 However, CRP was not associated with rs-fMRI metrics or rs-FC in patients (SAs and ACs). In the present analysis, CRP was considered a continuous variable and not a categorical variable (cutoff >3 mg/L) as done by Aruldass et al24 who found a link between rs-FC and high CRP. The correlation between proinflammatory cytokine concentration and the rs-fMRI metrics did not remain after FDR correction. Of note, previous studies reported a negative correlation between IL-6 and functional connectivity in insula-involving networks.24,61 Untargeted proteomic approaches to study inflammatory markers might be more suitable to identify possible associations with rs-fMRI metrics in patients with depression. In addition, future studies should focus on the hypothalamic-pituitary-adrenal axis because it may exert a modulatory influence on suicide risk through its dysregulation and its association with inflammatory pathways and neuroplasticity.62
Our study has several limitations. First, the sample size was small. However, our main results are mainly consistent with literature data. Second, the potential association between rs-fMRI, lethality of last suicidal act, and number of past suicide attempts could not be assessed due to the variable distribution. Third, in this study, only few factors involved in the inflammatory pathway were investigated. The association between rs-fMRI metrics and other peripheral inflammatory markers cannot be excluded and should be investigated in larger samples. Fourth, body mass index (BMI) was not considered in our analyses despite its association with inflammatory status because the small sample size did not allow including many covariables. However, BMI values were not different between groups. Fifth, a cross sectional methodology was used, and specific causal links could not be investigated. However, SAs with a recent history of suicide attempt (in the last 8 days) were recruited in order to increase the sample homogeneity and to emphasize the impact of the suicide attempt rather than depression. Indeed, most previous studies included depressed subjects with lifetime history of SA, limiting the possibility to determine whether and to what extent inflammatory markers are related to current depression or to a previous suicidal act. Nevertheless, the selection of such a specific sample may limit the generalizability of our results. Sixth, a ROI approach was chosen. Nevertheless, many statistical tests were carried out by including clinical and biological peripheral markers. Therefore, multiple testing corrections were performed to reduce the risk of false-positive results.
To conclude, our study showed that the VLPFC resting-state activity was altered in SAs and was influenced by the suicidal ideation intensity. Moreover, SA was associated with altered rs-FC between ACC and OFC, which may contribute to choose risky outcomes. These findings support the relevance of meditation based therapy for these patients. It is also important to identify specific subgroups of suicidal patients who might have increased vulnerability to inflammation that influences cerebral activity and who might benefit from anti-inflammatory agents. Finally, rs-fMRI metrics could be candidate biomarkers to detect patients at risk in order to improve suicide prevention.
Article Information
Published Online: June 26, 2024. https://doi.org/10.4088/JCP.23m15148
© 2024 Physicians Postgraduate Press, Inc.
Submitted: October 18, 2023; accepted April 3, 2024.
To Cite: Olié E, Clain G, Malestroit M, et al. Association of suicidal status, inflammation markers, and resting-state functional activity and connectivity in patients with major depressive disorder. J Clin Psychiatry. 2024;85(3):23m15148.
Author Affiliations: IGF, University of Montpellier, CNRS, INSERM, Montpellier, France (Olié, Courtet); Department of Emergency Psychiatry and Acute Care, Lapeyronie Hospital, CHU Montpellier, Montpellier, France (Olié, Clain, Malestroit, Courtet); Department of Neuroradiology, Gui de Chauliac Hospital and University of Montpellier, Montpellier, France (Le Bars); Institut d’Imagerie Fonctionnelle Humaine, I2FH, Department of Neuroradiology, Gui de Chauliac Hospital and University of Montpellier, Montpellier, France (Deverdun, Le Bars); Clinic La Lironde, Clinea, Montpellier, France (Fiedos); Université Jean Monnet Saint-Étienne, Mines Saint Étienne, INSERM, SAINBIOSE U1059, Saint-Étienne, France (Cognasse); Etablissement Français du Sang Auvergne-Rhône-Alpes, Saint-Étienne, France (Cognasse).
Corresponding Author: Emilie Olié, MD, PhD, Department of Emergency Psychiatry and Acute Care Lapeyronie Hospital, 371 Ave Doyen Gaston Giraud 34295 Montpellier cedex 5, France (e-olie@chu-montpellier.fr).
Relevant Financial Relationships: None.
Funding/Support: This study was funded by AOI Montpellier (2015) and the Etablissement Français du Sang (EFS).
Role of the Sponsors: The sponsors had no role in the conduct and publication of the study.
Author Contributions: Fiedos, Olié, and Courtet designed the study. Fiedos collected the data. Clain, Deverdun, and Malestroit performed the clinical and fMRI analyses. Cognasse, Olié, and Le Bars supervised the analyses. Olié wrote the first draft of the manuscript. Fiedos, Le Bars, Clain, Cognasse, Malestroit, Deverdun, and Courtet commented on the draft. All authors contributed to and have approved the final manuscript.
Acknowledgments: We thank Claire Belloc, MsC, Department of Medical Information, CHU Montpellier, Montpellier, France, and Myriam Benramdane, MsC, Department of Emergency Psychiatry and Acute Care, Lapeyronie Hospital, CHU Montpellier, Montpellier, France, for data management. We thank Elisabetta Andermascher (medical writer) for her careful reading.
We thank Charles Antoine Arthaud, MsC, Etablissement Français du Sang, Auvergne Rhône-Alpes & SAINBIOSE,U1059, France; Marie Ange Eyraud, MsC, Etablissement Français du Sang, Auvergne Rhône-Alpes & SAINBIOSE,U1059, France; and Amélie Prier, MsC, Etablissement Français du Sang, Auvergne Rhône-Alpes & SAINBIOSE, U1059, France, for technical support throughout our investigations. We thank Dr Hind Hamzeh-Cognasse, PhD, Université Jean Monnet Saint-Étienne, Mines Saint-Étienne, INSERM, SAINBIOSE U1059, Saint-Étienne, for her participation in biological analyses. All named individuals have no conflicts of interest to declare.
Supplementary Material: Available at Psychiatrist.com.
Clinical Points
- Resting-state activity of the ventrolateral prefrontal cortex was altered in suicide attempters (SAs), possibly contributing to language impairment and poor decision making and cognitive control.
- High C-reactive protein concentration was associated with history of suicidal acts, but none of the tested peripheral inflammatory markers (C-reactive protein, interleukin [IL]-1β, IL-2, and IL-6, and tumor necrosis factor-α) was associated with resting-state functional magnetic resonance imaging metrics.
- The relevance of meditation-based psychotherapies in recent suicide attempters needs to be thoroughly studied because these approaches modify the resting-state activity of prefrontal brain regions and reduce suicidal ideation.
The post Association of Suicidal Status, Inflammation Markers, and Resting-State Functional Activity and Connectivity in Patients With Major Depressive Disorder appeared first on Psychiatrist.com.