Quantitative susceptibility mapping of the fear circuit: Associations with silent symptoms in relapsing-remitting multiple sclerosis

Abstract

Background: Multiple sclerosis (MS) is a long-term autoimmune inflammatory disorder that affects the central nervous system leading to neurodegeneration, and can involve a variety of symptoms. These symptoms can include fatigue, anxiety, depression, and cognitive decline, which may be silent. The objective of this study was to explore changes in brain iron deposition in people with relapsing-remitting MS (pw-RRMS) compared to healthy controls (HCs), with a particular focus on regions of fear circuit. Additionally, the study aimed to evaluate relationship between iron deposition in these areas and clinical measurements. Methods: Pw-RRMS and HCs participants underwent brain MRI scans using quantitative susceptibility mapping (QSM) to assess iron deposition in the fear circuit between the two groups. The study analyzed correlations between brain susceptibility changes and clinical measurements. Results: We recruited 35 pw-RRMS (mean age = 46.7 ± 11 years; median EDSS = 2.5) and 18 HCs (mean age = 40.6 ± 17.8 years). Our research revealed significant increases in QSM signals relating to iron deposition in pw-RRMS compared to HCs, whole fear circuit (β = 5.82, p < 0.001), caudate (β = 21.48, p < 0.001), and putamen (β = 17.53, p = 0.03), showing the greatest difference. The whole fear circuit and particularly the caudate are strongly associated with fatigue in pw-RRMS. QSM values in the anterior cingulate cortex significantly differed between pw-RRMS with normal and abnormal depression scores (p = 0.007). Conclusions: These results strengthen the relationship between increased iron deposition in fear circuit regions and specific silent symptoms in pw-RRMS. However, further studies are required to confirm these findings and clarify the implications of iron accumulation in MS pathophysiology.

Introduction

Multiple sclerosis (MS) is a chronic autoimmune disorder characterized by inflammation, demyelination, and neurodegeneration in the central nervous system (CNS). Fatigue is a common and debilitating silent symptom reported by almost 80% of people with MS (pwMS).^1,2 It is considered to be the most troublesome symptom, and often leads to a reduction in work capacity and social isolation.^2,3 Fatigue is closely linked with depression and anxiety, which are also commonly found in MS and are often described as neuropsychiatric conditions. ⁴ These symptoms have an impact on overall health and quality of life. However, the pathophysiology of these silent symptoms in pwMS is not yet fully understood. Research suggests that structural damage to the cerebral gray and white matter in the MS brain could be substantially involved.^3,5,6 Advanced magnetic resonance imaging (MRI) has been used to target many areas of the brain in order to investigate the neural mechanisms of silent symptoms in MS. A connection between fatigue and the deterioration of axons in different white matter networks and brain circuits has been found by fractional anisotropy (FA) and axial diffusivity (AD), which are metrics used in diffusion MRI (dMRI). ⁷ Functional MRI (fMRI) has revealed that fatigued pwMS have stronger functional intercortical and subcortical connections, and weaker connectivity between the cortex and subcortical areas, in comparison to healthy controls (HCs). ⁸ A different study using resting-state fMRI (rs-fMRI) revealed that higher default mode network (DMN) connectivity was associated with symptoms of fatigue and depression. ⁹ A recent study revealed that only the left dorsal prefrontal cortex in pwMS was affected, and thinning observed in this region was associated with increased anxiety levels within anxiety-related networks connected to atrophied areas. ¹⁰

Iron is vital to the CNS and has various neurophysiological roles, for example, in neurotransmitter synthesis, in myelin production, and in the functioning of oligodendrocytes. ¹¹ The mitochondria of oligodendrocytes require iron for the production of adenosine triphosphate and for the synthesis of myelin components. ¹¹ When oligodendrocytes are injured, or when there is a loss of myelin—both hallmarks of active MS lesions—iron accumulates in the extracellular spaces. ¹² Additionally, factors such as oxidative stress, glutamate excitotoxicity, and protein misfolding can cause axonal degeneration and neurotoxicity, leading to increased iron levels in the MS brain tissues. ¹³ The dysregulation of iron homeostasis can thus contribute to MS neurodegenerative processes.

Iron deposition in certain regions of the brain has been linked to some silent symptoms of MS. Iron rim lesions often associated with progressive disease are more common in pwMS who have a higher level of disability^14–16 and those who experience significant fatigue. ¹⁵ Excess iron in the deep gray matter (DGM) has been shown to be linked to physical disability and cognitive impairment in pwMS. ¹⁷ Also, in other neurodegenerative conditions like Parkinson disease (PD), iron deposition has been associated with anxiety. Studies in animals have led to the proposition of a neural network called the “fear circuit” in which the amygdala plays a central role.^18–21 In the human brain, the amygdala serves as a mediator that connects external stimuli to the cognitive and behavioral responses related to anxiety. ¹⁰ Dysfunction of the fear circuit has been identified as the neural mechanism behind anxiety in both PD patients^22–25 and generalized anxiety disorder patients. ²⁶ The fear circuit comprises the amygdala, medial prefrontal cortex (mPFC), anterior cingulate cortex (ACC), caudate, putamen, hippocampus, thalamus, and insula.^18,27,28

Quantitative susceptibility mapping (QSM) is a novel postprocessing imaging modality that allows for detailed MRI quantification of tissue magnetic susceptibility using gradient-recalled-echo (GRE) MRI techniques. Within the human body, elements with paramagnetic properties, such as iron, exert a strong influence on magnetic fields. Conversely, diamagnetic sources, such as calcium and myelin, have only slight effects on magnetic fields. QSM patterns in MS lesions exhibit high magnetic susceptibility associated with iron accumulation, primarily at lesion rims.^13,29,30 These findings not only suggest a strong correlation between iron deposition and chronic inflammation associated with disease progression, but also highlight the potential of QSM in detecting these changes.

The distinct effect of iron buildup in the fear circuit and its relation to silent MS symptoms, such as fatigue, anxiety, and depression, has not been thoroughly investigated. Our study aims to address this knowledge gap by using magnetic susceptibility as an indicator of iron levels.

The objective of the present study was to investigate differences in regional magnetic susceptibility within the fear circuit in both pwMS and HCs, and then to establish whether a correlation between iron levels in the fear circuit and the clinical scores of silent MS symptoms exists in pwMS. Additionally, this study aimed to explore whether magnetic susceptibility values in specific regions of the fear circuit could be utilized as biomarkers for assessing the severity of these symptoms.

Methods

Participants

We recruited people with relapsing-remitting MS (pw-RRMS) who met the 2010 and 2017 McDonald criteria^31,32 for diagnosis, depending on their disease duration, and had stable disease. Inclusion criteria specified that all participants should be aged between 18 and 65 years, had been on stable MS treatment for at least 8 weeks prior to enrollment with no relapse in the past 6 months, had received an Expanded Disability Status Scale (EDSS) ³³ score of less than six within the last 12 months, and had exhibited a Fatigue Severity Scale (FSS) ³⁴ score of four or higher. We included pw-RRMS with an EDSS score of less than six to focus on those in the earlier to mid-stages of MS, ensuring they were ambulatory and did not have severe physical disability prohibiting scanning. Individuals with an FSS score of four or higher were included as part of the baseline data of an ongoing longitudinal study to focus on those experiencing significant fatigue, the most common silent symptom and often linked to anxiety and depression. HCs without neurological diseases and with normal neurological examinations were included to identify changes specifically associated with MS.

Clinical assessments

In this study, we conducted a thorough clinical assessment of the pw-RRMS using standardized instruments to evaluate different aspects of their condition. To measure the impact of fatigue, we used FSS, a self-reported measure that assesses the extent of fatigue-related impairment in daily activities. The commonly used cutoff score for indicating significant fatigue on the FSS is four or higher. ³⁵

We also used the Hospital Anxiety and Depression Scale (HADS) ³⁶ to screen for prevalent neuropsychiatric conditions, such as anxiety and depression. This scale has two subscales, each with seven items, and provides insights into the emotional well-being of the participants.

A cutoff score of eight or higher indicates significance for both the anxiety and depression subscales. This threshold was selected based on previous validation studies, which demonstrated a sensitivity of 88.5% and specificity of 80.7% for the anxiety subscale and a sensitivity of 90% and specificity of 87.3% for the depression subscale in detecting generalized anxiety disorder and major depression in pwMS.^37,38

To objectively measure the level of disability, we administered the EDSS, which allows for a detailed assessment of physical functions, including mobility and coordination.

MRI protocol parameters

All participants were scanned using a 3.0 T MRI scanner (Prisma, Siemens Healthcare, Erlangen, Germany; VE11 C) using a 64-channel head/neck coil. 3D T1-MPRAGE (magnetization-prepared rapid acquisition with echo gradient) scans were obtained using the following parameters: repetition time (TR) = 2500 ms; echo time (TE) = 2.22 ms; field of view (FOV) = 256 × 256 mm²; voxel size = 0.8 × 0.8 × 0.8 mm³; GRAPPA acceleration factor = 2; 208 slices; and acquisition time = 6:54 min. For assessment of MS lesion load, a T2-FLAIR (fluid-attenuated inversion recovery) sequence was acquired using the following parameters: TR = 5000 ms; TE = 386 ms; TI = 1800 ms; FOV = 256 × 256 mm²; voxel size = 0.8 × 0.8 × 0.8 mm³; GRAPPA acceleration factor = 3; 176 slices; and acquisition time = 5:02 min.

The QSM acquisition sequence was a 3D GRE multi-echo with true axial orientation, (Figure 1), with the following scanning parameters: TR = 39 ms, 7 monopolar echoes acquired at TE1:∆TE:TE7 = 5.84: 4.79:34.58 ms, flow compensation for the first echo in the readout (AP) and “slice” encoding (FH) direction, flip angle = 15°, pixel bandwidth = 220 Hz, elliptical k-space shutter, FOV = 240 × 180 mm²; 176 slices, resulting in isotropic voxels of size 1 mm³, with GRAPPA acceleration factor = 3 in the phase encoding (RL) direction. The scan duration was 8 min and 44 s.

Figure 1.

Three orthogonal planes of 3D T1-MPRAGE images displaying the 3D slice volume of the multi-echo GRE in the true axial orientation for reproducibility.

Image analysis

The QSM reconstruction from the GRE phase and magnitude data was implemented in MATLAB R2020a (Mathworks, Natick, MA, USA). Rough brain mask extraction was performed using FSL’s BET magnitude image of the first echo. Initially, the phase images (of each echo) were scaled between −π and +π. A phase offset correction was performed using the first two echoes followed by Laplacian phase unwrapping. ³⁹ A field map was then calculated using linear fitting with a zero intercept on a per voxel basis over all 7 phase images. The brain mask was also improved by using a threshold value of 20 on the fitting residual map after performing a Gaussian smoothing. The background field was removed using the V-SHARP ³⁹ method with a spherical mean value size of 25. STI Suite software (v3.0) ⁴⁰ was used for QSM reconstruction by performing dipole inversion using the iLSQR technique (pad sizes of 10 and 50 iterations). Susceptibility values were expressed in parts per billion (ppb). Our study utilized the FreeSurfer software suite 7.2.0 for brain segmentation and spatial normalization between anatomical T1W, QSM and FLAIR images. T1W images were segmented into several cortical and subcortical regions. The QSM images were rigidly normalized to the same space (BBRegister) using the magnitude image. In order to investigate the magnetic susceptibility as a potential biomarker for assessing the silent MS symptoms, we specifically identified regions of interest (ROIs) within the fear circuit. These ROIs encompassed the amygdala, mPFC, ACC, caudate, putamen, hippocampus, thalamus, and insula. To determine the mean QSM values within these ROIs, we averaged the values from both unilateral hemispheres. Figure 2 shows QSM images and corresponding T1w MPRAGE for three regions of the fear circuit.

Figure 2.

A and B show quantitative susceptibility maps for thalamus (red) caudate (orange) and putamen (blue) of a 44-year-old female RRMS versus 40-year-old female HC, while C and D show their corresponding T1-MPRAGE images.

Statistical analysis

The statistical analysis was performed using SPSS 28 (IBM Corp. IBM SPSS Statistics, Armonk, NY) and R (R Core Team (2023), the R Foundation for Statistical Computing, Vienna, Austria). Descriptive statistics were calculated for demographic and clinical variables, including mean and standard deviation. Inferential statistics were performed using an independent samples t test, where appropriate. We used general linear models (GLMs), adjusted for age and sex as covariates to test for statistical significance of the susceptibility values among the study groups in both the right and left hemispheres for each ROI and separately. The β-coefficients, standard errors, t values, and p values were extracted from the model. The Holm-Bonferroni method was used to corrected p-values for multiple comparisons, maintaining the family-wise error rate. A p-value less than 0.05 were considered statistically significant.

Linear regression models were used to explore correlations between QSM values in ROIs and the following clinical measures: FSS, anxiety, depression and EDSS. We used the β-coefficients to quantify the size of the effect and to interpret the data.

Additionally, RRMS participants were divided into “normal” and “abnormal” subgroups based on specific cutoff scores for HADS anxiety and depression. The differences of QSM values between these subgroups were assessed by Mann-Whitney test.

Results

Participant characteristics

The study cohort comprised 35 pw-RRMS and 18 HCs. The average age of the MS group was 46.7 years (standard deviation (SD) = 11.0) with female participants predominating at 82.8%. In contrast, the HCs had an average age of 40.6 years (SD = 17.8) and an equal sex distribution (50% female, 50% male). Independent sample t-tests revealed no significant difference in age between the pw-RRMS and HCs groups (p = 0.12). However, sex distribution was found to be significantly different between the groups (p = 0.02). The clinical assessment of the pw-RRMS group showed a median EDSS score of 2.5 (range 0–4.5) and the following mean scores: FSS, 5.6 (SD ± 0.94); disease duration (DD), 12.2 years (SD ± 10.2); depression score, 6.3 (SD ± 3.6); and anxiety score, 6.7 (SD = 3.8). The mean time since the last clinical relapse was 3.9 years (SD = 3.7). The demographic data and clinical assessment scores are summarized in Table 1.

Table 1.

Demographic data and clinical assessment scores.

Participant characteristics	pwMS	Healthy controls	p Value
No. of participants	35 RRMS	18	-
Mean age (SD) (yr)	46.7 (11.0)	40.6 (17.8)	0.12
Sex (no.) (%)
Female	29 (82.8%)	9 (50%)	0.02
Male	6 (17.1%)	9 (50%)
Mean DD (SD) (yr)	12.2 (10.2)	-	-
Median EDSS score (range)	2.5 (0–4.5)	-	-
Mean last time of clinical relapse (SD) (yr)	3.9 (3.7)	-	-
Mean FSS (SD)	5.6 (0.9)	-	-
Mean HADS Anxiety (SD)	6.7 (3.8)	-	-
Frequency of abnormal (moderate-high) anxiety (≥8)	14 (40%)	-	-
Mean HADS-Depression (SD)	6.3 (3.6)	-	-
Frequency of abnormal (moderate-high) depression (≥8)	13 (37.1) %	-	-

Abbreviations: pwMS; people with multiple sclerosis; RRMS: relapsing-remitting multiple sclerosis, SD; standard deviation, DD; disease duration, EDSS; Expanded Disability Status Scale. FSS; Fatigue Severity Scale, HADS; Hospital Anxiety and Depression Scale.

Comparisons of susceptibility values between pw-RRMS and HCs

Our study demonstrated that QSM signals in pw-RRMS were significantly higher than in HCs in the whole fear circuit (β = 5.82, p < 0.0008), but in particular in the caudate (β = 21.48, p < 0.009), putamen (β = 17.53, p = 0.03), and a trend also in mPFC (β = 2.24, p = 0.06). Differences in other regions were not significant. (see Figure 3 and Table 2).

Figure 3.

Boxplot of QSM values in fear circuit regions for pw-RRMS and healthy controls. All p values are adjusted for age, and sex, and multiple comparisons using Holm-Bonferroni correction method. Abbreviations: fear circuit (FC), medial prefrontal cortex (mPFC), and anterior cingulate cortex (ACC). *p < 0.05, **p < 0.01, ***p < 0.001, ppb: part per billion.

Table 2.

Summary of GLM results of QSM in fear circuit regions.

Region	β-coefficients	St. error	t-statistic	p value
Whole FC	5.82	1.41	4.1	0.0008
Amygdala	5.55	3.32	1.66	0.5
mPFC	2.24	0.89	2.49	0.06
ACC	2.82	1.88	1.49	0.4
Caudate	21.4	4.36	4.92	0.0009
Putamen	17.5	6.04	2.90	0.03
Hippocampus	2.05	2.07	0.99	0.3
Thalamus	−2.20	.17	−1.01	0.6
Insula	−1.80	2.07	0.99	0.4

All p values are adjusted for age, and sex, and multiple comparisons using Holm-Bonferroni correction method. Abbreviations: medial prefrontal cortex (mPFC), anterior cingulate cortex (ACC), fear circuit (FC).

Comparisons of magnetic susceptibility values in the Rt and Lt fear circuit ROIs between pw-RRMS versus HCs

The right hemisphere analysis showed significant increases in the caudate (β = 20.60, p < 0.004), and amygdala (β = 9.35, p = 0.012). The left hemisphere displayed similar but nonsignificant trends in most ROIs except in the caudate (β = 18.02, p = 0.002) with a trend in the putamen (β = 15.30, p = 0.06). Overall, there were significant increases in QSM signals for the whole fear circuit in both right (β = 4.72, p = 0.014) and left (β = 3.59, p = 0.01) hemispheres, as shown in Table 3.

Table 3.

Summary of significant GLM results for the QSM of the Rt and Lt fear circuit regions.

Region	β-coefficients	St. error	t-statistic	p value
Whole FC_Lt	3.59	1.46	2.46	0.01
Whole FC_Rt	4.72	1.65	2.84	0.014
Amygdala_Rt	9.35	2.90	3.21	0.012
Caudate_Lt	18.02	4.57	3.93	0.002
Caudate_Rt	20.60	5.60	3.67	0.004
Putamen_Lt	15.30	6.27	2.43	0.06

All p values are adjusted for age, sex, and multiple comparisons using Holm-Bonferroni correction method. The FC all were adjusted by factor 2 and other regions by factor of 8 separately for Rt and Lt. Abbreviations: fear circuit (FC)Right (Rt), left (Lt), medial prefrontal cortex (mPFC), anterior cingulate cortex (ACC).

Correlations between magnetic susceptibility of fear circuit regions in the pw-RRMS and clinical scores

Based on the results above, we focused on examining changes in QSM within regions of the fear circuit involved in pw-RRMS. Specifically, we investigated the whole fear circuit, putamen, caudate, and mPFC to determine the correlation between QSM values in these regions and clinical measurements. Our findings revealed significant positive correlations between fatigue scores and QSM in the whole fear circuit (β = 2. 4 p = 0.04) and the caudate (β = 7.64 p = 0.03) (Table 4).

Table 4.

Summary of four significant GLM results regions of QSM in RRMS correlations with clinical assessments.

Clinical assessments	Region	β-coefficients	St. Error	t-statistic	p Value
FSS	mPFC	1.66	0.61	2.71	0.14
FSS	Caudate	7.64	2.28	3.34	0.034
FSS	Putamen	6.67	3.14	2.12	0.5
FSS	Whole FC	2.4	0.74	3.22	0.043

All p values are adjusted for age, and sex, and multiple comparisons using Holm-Bonferroni correction method Abbreviations: Fatigue Severity Scale (FSS); Medial prefrontal cortex (mPFC); Anterior cingulate cortex (ACC); Expanded Disability Status Scale (EDSS), fear circuit (FC).

Differences susceptibility values between anxiety and depression pw-RRMS subgroups

The study found that 21 participants (60%) had values within normal range of the HADS, while 14 participants (40%) experienced abnormal anxiety (moderate to high, HADS Anxiety ≥8). Regarding depression, 22 participants (62.9%) had normal depression scores (HADS-Depression ≤7), and 13 participants (37.1%) demonstrated abnormal depression levels (moderate to high, HADS-Depression ≥8).

Notably, there were statistically significant differences in QSM values in the ACC between pw-RRMS with normal and abnormal depression scores (U = 64, p = 0.007). In contrast, QSM values did not significantly differ between subgroups based on anxiety scores. (Tables 1S and 2S in the supplemental section).

Discussion

This study aimed to explore the correlation between magnetic susceptibility in fear circuit regions and the silent symptoms of MS. Firstly, we compared magnetic susceptibilities in the fear circuit regions of individuals with RRMS and those of HCs. We then sought to identify any correlations between magnetic susceptibility values in fear circuit regions and the severity of fatigue, anxiety, depression, and disability measures in individuals with RRMS.

Overall, the study revealed that there were increased magnetic susceptibilities in most regions of the fear circuit in pw-RRMS as compared to HCs. The whole fear circuit, caudate, putamen, amygdala, and mPFC in particular had higher QSM signals in the RRMS than in the HCs. However, decreased magnetic susceptibility signals were observed in the thalamus and insula of pw-RRMS. Our study also revealed that magnetic susceptibilities in the whole fear circuit and caudate were strongly correlated with severity of fatigue in pw-RRMS. To the best of our knowledge, this is the first time that the relationship between iron in the fear circuit and MS silent symptoms has been studied.

Abnormal iron accumulation in MS

The presence of excessive iron in MS brain tissue can result in the production of reactive oxygen species, which can lead to oxidative stress, mitochondrial dysfunction, and neuronal damage. ¹² Iron buildup has been identified as a feature of chronic active MS inflammatory lesions. Microglia and macrophages uptake react to the accumulating iron, which is visible as the paramagnetic rim of MS lesions in phase-sensitive sequences, such as QSM or susceptibility-weighted imaging,^41,42 and this subsequently leads to a proinflammatory status of microglia and macrophages and drives neuroinflammatory cascades.^29,43,44 The existence of a relationship between iron imbalance and MS pathology in the DGM areas of the MS brain ¹⁷ has been confirmed by MS postmortem histopathology studies.^45,46

MS-related silent symptoms and brain regions

Fatigue and disability in pwMS have been extensively studied, with authors variously reporting that changes on a macro/microstructural level, changes to brain volume, and changes in the functional connectivity network within the frontal, parietal, temporal, thalamic, amygdala, and basal ganglia (caudate and putamen) regions are involved in the mechanisms underlying fatigue and disability.^3,47,48 Previous quantitative MRI studies have reported changes in the cortical and subcortical GM, providing evidence for a potential link between neurodegeneration in the striatal-thalamic-frontal regions and fatigue.^3,5,47,48 MS-related fatigue, anxiety, depression, and disability are a direct consequence of disturbed connectivity within the caudate, putamen, amygdala, thalamus, hippocampus, and frontal neocortical regions of pwMS. ⁴⁹ Various cross-sectional and longitudinal MS studies have reported heterogeneity in the iron content of DGM.^50–52

Our findings align with the results of existing studies, in that we observed increased brain iron in the average of the whole fear circuit, particularly in the caudate and putamen,^51,53 while the iron content in the thalamus was found to be lower.^53,54 The caudate, putamen, thalamus, and neocortex play essential roles in motor and cognitive functions and speedy information processing and are often affected in pwMS.^55,56 Iron overload in these regions can cause disrupted connectivity, which has been reported in MS, and which leads to cognitive decline and fatigue, and possibly contributes to other MS symptoms. Beta (β) coefficient analysis of magnetic susceptibility revealed a significant effect size, particularly in the caudate (β = 21.48) and putamen (β = 17.53), suggesting that these parameters could be used as potential biomarkers for predicting MS progression over time.

Various studies on MS have noted a decrease in iron levels in the thalamus,^57–59 and this study made similar observations. The thalamus is a complex anatomical region that contains a large density of oligodendrocytes, which consequently normally store iron. ⁵⁹ High levels of myelination distinguish the thalamus from other DGM regions and could thus make it more susceptible to iron depletion during the process of demyelination that occurs during MS disease progression.^60,61 Schweser et al. (2018) observed a decrease in iron in the thalamic region in RRMS and secondary progressive MS but not in clinically isolated syndrome patients, which indicated that iron depletion is significantly associated with MS disease duration. ⁵⁹

Asymmetric brain hemisphere magnetic susceptibility

An asymmetry was observed between the magnetic susceptibility of the brain hemispheres in this study. For our pw-RRMS cohort, a greater magnetic susceptibility was observed in the right hemisphere fear circuit than in the left hemisphere fear circuit, whereas in the HCs cohort, there were no significant differences between the hemispheres. Greater magnetic susceptibility effect sizes were observed in the bilateral caudate, left putamen, and right amygdala. Also, the right amygdala in pw-RRMS exhibited a greater level of iron deposition than did the left amygdala, which may indicate a pathological change in the brain. Iron overload in the amygdala could influence emotional regulation and stress responses, ⁶² which might be related to anxiety and depression. The cause of this hemispheric asymmetry in magnetic susceptibility remains uncertain, yet it aligns with findings from other studies.^63,64 Previous research on normal adults has identified an asymmetrical distribution of brain iron, predominantly in the left hemisphere, which might have implications for dopaminergic pathways and motor lateralization. ⁶³ In contrast, studies on Alzheimer’s disease have noted a higher magnetic susceptibility in the right hemisphere compared to the left, further illustrating the complexity of cerebral iron distribution.^64,65 The variability in findings regarding the lateralization of magnetic susceptibility across studies can be attributed to several factors. These include differences in MRI acquisition parameters—such as imaging techniques, magnet strength, and head positioning—as well as variations in subject characteristics such as age, sex, brain atrophy, tissue relaxation properties, and systemic health conditions.^66,67

MS silent symptoms correlation with magnetic susceptibility

Significant correlations between fatigue and high levels of magnetic susceptibility in the whole fear circuit and particularly in the caudate were observed. These findings are in agreement with previous neuroimaging studies of the right hemisphere of the MS brain undertaken in relation to fatigue. For example, a higher RD and lower FA were reported in the right temporal cortex of a high-fatigue pwMS group. ⁶⁸ In addition, fatigue in pwMS has been linked with lower FA in the right anterior thalamic radiation region. ⁶⁹

Anxiety and depression in pwMS have not been extensively explored using MR neuroimaging, suggesting a gap exists in the research for this condition. A small number of studies on MS lesions have concluded that there is no evidence of a relationship between anxiety and observable structural abnormalities in MRI brain tissues. ⁷⁰ However, a recent study using 7T MRI revealed that the presence of lesions in the limbic system might be associated with anxiety in pwMS as assessed by HADS. ⁷¹ However, that study involved a small sample size, thus its findings can serve only as preliminary evidence. ⁷¹

We observed a significant difference in the depression subgroup of pw-RRMS with magnetic susceptibility values in ACC, while there were no significant differences in the anxiety subgroup in the magnetic susceptibility values of fear circuit regions. There are several possible explanations for this lack of statistical significance. Firstly, the mean scores for anxiety and depression were close to the clinical cutoff scores, which implies that even though participants had significant fatigue, the majority had values within normal range for the other symptoms. Secondly, disease-modifying therapy had been effective in stabilizing disease, perhaps particularly in relation to symptoms such as anxiety and depression. In fact, mean duration since last clinical relapse in our pw-RRMS cohort was approximately 4 years.

Fear circuit and iron relation

Iron abnormalities in fear circuit regions have been linked to anxiety disorders in studies on animals and on neurogenerative disorders. In a mouse model, the transport of iron along neural pathways—specifically from the ventral hippocampus to the mPFC—was demonstrated to exert a modulatory effect on anxiety-related behaviors. ⁶² This finding suggests a significant role for iron in the fear circuit, with the consequence that changes in the distribution of iron in the brain can have a significant influence on emotional regulation and stress response mechanisms. ⁶²

A recent study investigated differences in brain iron deposition in fear circuit regions between PD patients with and without anxiety. ²² QSM analysis revealed that PD patients with anxiety had significantly greater iron deposition in several brain regions, which include the mPFC, ACC, hippocampus, precuneus and angular cortex than those without. The QSM values in some of these regions were positively correlated with anxiety scores, specifically those in the mPFC (r = 0.255, p = 0.04), ACC (r = 0.381, p < 0.01), and hippocampus (r = 0.496, p < 0.01). ²² An rs-fMRI study on patients with generalized anxiety disorder revealed that overactivation of the hippocampus and thalamus, both of which are part of the fear circuit, was associated with somatic symptoms. ²⁶

In addition, a recent study revealed that pwMS with fatigue and anxiety had larger caudate volumes and a thinner left parietal cortex than those without these symptoms. ⁷² A reduction in susceptibility in the thalamus has been associated with an increased duration of MS disease and an increase in the severity of disability. ⁵⁹ It was reported by Elkady et al., that a reduction in thalamic susceptibility could be a consequence of iron depletion in oligodendrocytes and demyelination.^59,73

In our study, we observed significant differences in the QSM of the ACC between pw-RRMS with abnormal depression scores and those with normal scores. This finding aligns with Wang et al. (2022) study, which showed that higher magnetic susceptibility values in the ACC and other brain regions are linked to the severity of depression in late-life depression. ⁷⁴ The involvement of the ACC in emotional regulation and cognitive processing suggests that abnormal iron deposition in this region may contribute to the development of depression in MS. This is consistent with mechanisms observed in other neurological disorders associated with depression, where dysregulated iron metabolism is implicated in mood disturbances.

Moreover, a study using QSM in patients with depressive disorders over a range of severity revealed that the susceptibility value in the bilateral putamen of those with major depression was significantly higher than in the control group or in those with only mild-to-moderate depression. ⁷⁵ Additionally, patients with major depressive disorders had considerably lower susceptibility values in the thalamus than those in the control group. This study concluded that brain iron deposition could be associated with depression, and that iron deposition patterns could be a biomarker for the pathophysiological mechanisms underlying depression. ⁷⁵ The association between depression levels and hippocampal QSM signals suggests a potential relationship between iron and MS symptoms in that region, especially given the role of the hippocampus in emotion regulation. These insights into QSM signal intensities and clinical measurements suggest that QSM measurements have the potential for use as biomarkers of disease progression.

Connectivity abnormalities within the right hippocampus, right amygdala, and frontal regions as identified using dMRI might be associated with depression in pwMS. ⁷⁶ In addition, pwMS with depression exhibited increased functional activity in the ventrolateral prefrontal cortex relative to HCs, while at the same time a lack of functional connectivity with the amygdala and mPFC in pwMS was observed. ⁷⁷ This result suggests that demyelination may be a primary contributing factor to the atypical pattern of connections between these regions. ⁷⁷ Our findings on changes related to iron distribution in fear circuit regions could provide supporting evidence for the role of iron in the development of MS silent symptoms.

Limitations

This study was restricted to a small sample size and had a cross-sectional design, which may limit the generalizability of the findings. Also, most of our pw-RRMS sample reported high levels of fatigue, which may have had an effect on the results. Despite these limitations, our study highlights the potential of QSM in this area of research. Future research should include a more varied cohort, particularly one that includes pwMS who are highly anxious and those with significant depressive symptoms and should involve longitudinal studies. Subsequent research can build on our findings to improve diagnostic and therapeutic approaches.

Conclusion

Our study explored iron deposition in the fear circuit of pw-RRMS and correlated the findings with silent symptoms. The whole fear circuit, the caudate, putamen, and mPFC showed increase magnetic susceptibility when compared to HCs, while magnetic susceptibility levels in the whole fear circuit and the caudate were highly correlated with fatigue. In addition, the thalamus showed lower magnetic susceptibility than HCs. The subgroup with higher depression level had high magnetic susceptibility in the ACC. Our results shed light on the role of iron in the fear circuit and its relationship with silent symptoms in pw-RRMS.

Ethical statement

Ethical approval

Ethics approval was obtained from Melbourne Health Human Research Ethics Committee (HREC Reference Number: HREC/46758/MH-2018). This approval is already registered to the Hunter New England HREC and the University of Newcastle’s Human Research Ethics Committee (Reference No: H-2020-0264). All participants signed a written consent form before the MRI examination.

Informed consent

A written and signed consent form was obtained from each participant for publication of their MR images.

ORCID iDs

Ibrahim Khormi https://orcid.org/0000-0002-5002-2166

Abdulaziz Alshehri https://orcid.org/0000-0003-4893-223X

Saadallah Ramadan https://orcid.org/0000-0003-3874-7866