Anxiety and depression lead to fALFF changes in migraine patients: an fMRI cross-sectional study

2023-03-17 06:24,,,,3,,3,,,
空军军医大学学报 2023年2期

, , , ,3, ,3, , ,

1Functional and Molecular Imaging Key Laboratory of Shaanxi Province, Department of Radiology, Tangdu Hospital, Air Force Medical University, Xi’an 710038, China; 2Department of Medical Imaging, Gansu Hospital of Chinese Armed Police Force, Lanzhou 730050, China; 3School of Medical Technology, Shaanxi University of Chinese Medicine, Xianyang 712046, China

[Abstract]Objective To investigate the effects of anxiety and depression on brain functional alterations in migraine patients. Methods Resting-state functional magnetic resonance imaging data were used to analyze the changes in fractional amplitude of low-frequency fluctuation (fALFF) in 46 patients with migraine without aura and 34 age- and gender-matched healthy controls. Pearson correlation analysis was used to explore the relationship between changes in fALFF and clinical characteristics. Results The fALFF value of the posterior lobe of left cerebellum increased in migraine patients with or without anxiety and depression. In patients with migraine without anxiety and depression, decreased fALFF values were found in the thalamus, temporal lobe, and frontal lobe. In patients with migraine with anxiety and depression, fALFF value in cerebellum was negatively correlated with the Montreal Cognitive Assessment scores. Conclusion The posterior lobe of left cerebellum is a continuously activated region in migraine patients. This region may be a target region for further understanding the mechanism of migraine attack.

[Key words]migraine; magnetic resonance imaging; fractional amplitude of low-frequency fluctuation; depression

1 Introduction

Migraine is believed to be a complex disorder in which nociceptive and somatosensory information is transmitted to the cortical and subcortical areas along the trigeminovascular system, bringing about diverse symptoms with the headache[1-2].

Anxiety and depression are the most common psychiatric comorbidities in migraine patients[3]. The prevalence of depression in migraine patients ranges from 8.6% to 47.9%[4]. Anxiety levels in migraine patients are two to five times higher than in the general population[5]. Depression is often comorbid with anxiety disorders in migraine patients, placing a heavy burden on patients[6]. Understanding the functional alteration of migraine is a clear goal for further research on pathophysiological mechanisms and treatment of migraine[7].

Resting-state functional magnetic resonance imaging (fMRI) is capable of measuring each voxel frequency of the whole brain by observing signals during the rest period[8]. As a widely used metric of fMRI, the amplitude of low-frequency fluctuation (ALFF) measures the significant temporal correlation of dominant in the low-frequency range (<0.1 Hz)[9]. However, it is sensitive to biological artifacts, such as respiratory and cardiac noise. The fractional ALFF (fALFF) measures the ALFF at each voxel divided by a signal power across the whole frequency spectrum[10]. It can more effectively reflect the spontaneous neural function of the brain in resting state.

We aimed to apply fALFF to evaluate the amplitude difference of the whole brain in the interictal period of migraine patients, patients with migraine without anxiety and depression (MOAD), patients with migraine with anxiety and depression (MAD), and healthy controls (HCs).

2 Methods

2.1 Participants

Based on the International Classification of Headache Disorders, 3rd edition[11], this study included 46 participants diagnosed with migraine without aura by neurological physicians. Thirty-four demographically matched HCs were recruited through community advertising from May 2016 to June 2022. Participants were excluded if taking any prophylactic headache medicine three weeks before the scanning, and they did not experience any headaches during the scanning.

2.2 Ethics and consents

This study was approved by Medical Ethics Committee of Tangdu Hospital (TDLL-2015133). All participants wrote the informed consent form and fully comprehended the fMRI scanning procedure.

2.3 Clinical characteristics

The demographic characteristics of participants included age, gender, duration, frequency, and disease course. We assessed the anxiety and depression levels of the participants using Self-Rating Depression Scale (SDS)[12]and Self-Rating Anxiety Scale (SAS)[13]. The intensity scores (0-10) of the headache were measured by Visual Analogue Scale (VAS). The mental states of the participants were assessed by Montreal Cognitive Assessment (MoCA).

2.4 Data acquisition

Imaging data were acquired by an 8-channel head coil array, the 3.0 Tesla MRI system (MR750; GE Healthcare, Milwaukee, WI, USA). The structural images were acquired with the sagittal three-dimensional brain volume sequence and followed with these parameters: echo time (TE)=3.2 ms, repetition time (TR)=8.2 ms, interval time (TI)=450 ms, flip angle (FA)=12°, field of view (FOV)=256 mm×256 mm, matrix=256×192, voxel size=1 mm×1 mm×1 mm, and slice number=188. We utilized the gradient-recalled echo-planar imaging sequence to collect resting state blood oxygenation level-dependent images (36 contiguous slices with a slice thickness of 3 mm; TR=2 000 ms; TE=40 ms; FA=90°; FOV=240 mm×240 mm; interslice gaps=0 mm; matrix=64×64; in-plane spatial resolution=3 mm×3 mm; total volumes=185)[14].

2.5 fMRI data preprocessing

Preprocessing of functional images followed these steps: slice timing, realignment, coregistration, and smoothing, which were performed in the toolbox of Data Processing Assistant for Resting-State fMRI (DPARSF). The first five volumes of fMRI images were discarded to allow signal equilibration and to compensate for transient scanner instability noise. Images were corrected for slice-timing and head motion. Realignment was conducted to correct the motion between volumes. Head motion parameters were performed by estimating the translation in six directions and the angular rotation on each axis for each volume. The thresholds of head motion were defined as translational or rotational motion parameters less than 2 mm or 2°. Then, the transformed structure images, firstly co-registered with the mean functional image, were segmented and normalized to the Montreal Neurologic Institute space using the diffeomorphic anatomical registration through the exponentiated Lie algebra technique. Finally, all images were spatially smoothed with a Gaussian kernel of 6 mm full-width at half maximum.

2.6 fALFF analysis

Following the preprocessing, we performed the fALFF analysis using the DPARSF software. The time series of each voxel of whole brain functional images were transformed into a frequency domain without band-pass filtering. The square root was calculated at each frequency domain, and the average square root was acquired in the range of 0.01-0.08 Hz per voxel. The fALFF was the ratio of the frequency domain in the low-frequency range to the power spectrum in the entire frequency range. Lastly, the fALFF maps of each participant were obtained through the fALFF of each voxel divided by the global mean fALFF value.

2.7 Subgroup analysis

To investigate the influence of anxiety and depression on migraine, we divided migraine patients into MAD group and MOAD group based on the standardized scores of SAS (>50) and SDS (>53).

2.8 Statistics analysis

The whole brain fALFF was compared among migraine patients (MAD and MOAD patients) and HCs by two-samplet-test in DPARSF software, with age, gender, and headache location as covariates. The Gaussian Random Field theory was used to conduct statistical comparison, setting the significance threshold at voxel-levelP<0.01 and cluster-levelP<0.05. Pearson correlation analysis was conducted between the activated brain areas and clinical variables in migraine patients.

3 Results

3.1 Demographics and clinical characteristics

Forty-six migraine patients and 34 HCs were recruited for this study. No difference was found in age and gender distribution between the two groups (Table 1).

Table 1 Demographics and clinical characteristics among migraine patients (MOAD and MAD patients) and HCs

There was also no difference in gender distribution and age among the MAD patients, MOAD patients, and HCs.

3.2 Alterations of fALFF values

The left cerebellum posterior lobe of migraine patients was consistently activated. Compared with HCs, migraine patients showed increased fALFF values in the bilateral cerebellum posterior lobe and the left temporal lobe, and decreased values in right thalamus, right temporal lobe, and bilateral frontal lobe (Figure 1).

a: The maps of the fALFF changes for three groups; b: The axial, sagittal, and coronal images. Blue color shows brain region with decreased fALFF values; red color shows increased regions. MOAD: migraine without anxiety and depression; MAD: migraine with anxiety and depression.Figure 1 The fALFF alterations among migraine patients, MOAD patients, and MAD patients

The MOAD patients showed enhanced fALFF in the left cerebellum posterior lobe, and decreased fALFF in the right insula, thalamus, and left frontal lobe. The patients accompanied by anxiety or depression showed increased fALFF values in the left cerebellum, and no decreased regions were found (Table 2).

Table 2 The alterations of fALFF among subgroups of migraine patients and HCs by resting-state fMRI

3.3 Correlations between abnormal fALFF values and clinical variables

In MAD patients, increased fALFF in the left cerebellum lobe was negatively correlated with MoCA scores (r=-0.585,P=0.014, Figure 2).

MoCA: Montreal Cognitive Assessment; fALFF: fractional amplitude of low-frequency fluctuation.Figure 2 Correlation between fALFF of left cerebellum and MoCA scores

As fALFF values increased, the scores of migraine patients decreased, which indicated that the cognition function of MAD patients decreased as the fALFF values increased. No significant correlations were found between fALFF and headache duration, frequency, and intensity.

4 Discussion

This study aimed to reveal functional changes in migraine with or without anxiety and depression. Compared with HCs, migraine patients with or without anxiety and depression showed significant alterations in fALFF values in the left cerebellum posterior lobe. At the same time, migraine and MOAD patients showed decreased fALFF values in the right insula, right thalamus, and left frontal lobe. The increased fALFF values of MAD patients were negatively correlated with MoCA scores.

Compared with MOAD patients, MAD patients did not show fALFF changes in other brain regions, excluding the left cerebellum. The decreased fALFF in right insula, thalamus, and left frontal lobe in MOAD patients were the relay pathway of pain and projecting regions of many cortically mediate symptoms in migraine, including allodynia, phonophobia, photophobia, and osmophobia[2]. Therefore, we suggested that anxiety and depression might enhance the functional changes in the cerebellum of migraine patients.

The underlying neuroimaging alterations in migraine patients have previously been studied using fMRI. The hypothalamus, pons, sensorimotor cortex, and visual cortex are the common ascending pain and central trigeminovascular pathways, which show hyperexcitability during the preictal and postictal headache of migraine[15-16].

We found that the left cerebellum was simultaneously activated in all three groups of migraine patients in the interictal period of headache. The cerebellum has been involved in various motor control and coordination functions. However, more recently, it has been suggested that it also played a role in no-motor functions, including pain processing and cognition[17]. In an animal model of migraine, the cerebellum showed enhanced functional connectivity to the insula[18]. Furthermore, chronic neuropathic pain also showed that the cerebellum was activated in the enhanced oscillation of gamma in the resting state, which can be used as a characteristic marker[19]. The evidence might suggest that the activated cerebellum is the consequence of repeated attacks of headache[20]. However, the specific function of the cerebellum during pain processing and its role in pain disorder are not clear.

The previous studies have indicated that the cerebellum connecting with the affective-limbic network was related to the severity of depression and anxiety symptoms[21]. However, the present studies identified that the cerebellum was the consistently activated region in migraine patients with or without anxiety and depression.

Meanwhile, the fALFF changes in the cerebellum might lead to alteration of cognition function, especially the negative correlation with MoCA scores. It may be a potential region to further extend the mechanism of migraine and psychiatric disorders.

Although there were no functional changes in MAD patients in right thalamus and left frontal lobe, these regions were the pivotal hub of the pain processing pathway[22]and higher-order cognitive functions[23]. The decreased fALFF values in migraine patients and MOAD patients demonstrated that the cerebral functional disorder of migraine patients was complex and susceptible[8].

In addition to the involvement of the left cerebellum, the dysfunctional right thalamus and left frontal lobe contribute to the pathophysiological mechanisms of migraine. These changes should be concerned in both diagnosis and prognosis.

5 Limitations

There are a few limitations that should be considered in the results of our study. First, the pathological characteristic of diseases, i.e., migraine, depression, or anxiety, may result in different functional alterations of the brain. Due to the small sample size and strict correction threshold, MAD patients showed only cerebellum activity. More researches with larger sample sizes are needed to confirm this. Second, the single measured metric of fMRI was used in this study, lacking cross-validation evidence from other indicators to the results. Third, it is a necessity to confirm the results by multi-center data and other subtypes of migraine.

Conflict of interest: The authors declare that they have no conflicts of interest.

Acknowledgment: Our gratitude will go to Profs. LI Jinlian, CHEN Liangwei, and ZHU Junling from the Department of Radiology of Tangdu Hospital for their constructive comments on this study.

Data availability: The datasets analyzed in the study are available from the corresponding author upon reasonable request.