A Full Brain 3D-Magnetic Resonance Spectroscopic Imaging Study of Migraine and Post-Traumatic Headache

Abstract

Magnetic resonance spectroscopic imaging provides a means to quantify brain metabolites including N-Acetylaspartate (NAA), choline (Cho), creatine (Cr), myoinositol (Ins), and glutamate and glutamine (Glx). The goal of this work was to investigate the metabolite differences between participants with migraine, participants with acute post-traumatic headache (PTH) with migraine-like phenotype, and healthy controls. For spectroscopy acquisitions, a 3D echo-planar spectroscopic imaging sequence was used with full brain coverage. 3D metabolite maps for NAA, Cho, Cr, Ins, and Glx were compared between participants with migraine (n = 12), participants with acute PTH (n = 10), and healthy controls (n = 11) using a full factorial design with significance defined as p < 0.05 with family-wise error corrections for multiple comparisons. The migraine group showed increased Cho and Ins in the right hippocampus and increased Ins in the bilateral parahippocampal regions, left inferior temporal, and bilateral fusiform areas relative to healthy controls. Relative to healthy controls, the acute PTH cohort had decreased NAA in the right precuneus, increased Glx in the right lingual and the right calcarine gyrus, and increased Ins in the left amygdala. Relative to individuals with migraine, those with acute PTH had higher Glx in the right calcarine gyrus, decreased Glx in the left insula, decreased Ins in the left fusiform gyrus, and increased NAA in the left frontal inferior area. The metabolite differences between migraine, PTH, and healthy controls observed in this study could provide insights into the mechanisms and consequences of migraine and PTH.

Introduction

Migraine and post-traumatic headache (PTH) share similarities in phenotype.^1–4 In fact, the majority of individuals with PTH who are seen in clinical settings have migraine-like symptoms.^1–4 Research imaging of studies demonstrates some differences in brain structure and function between migraine and PTH.^1,5–12 Using magnetic resonance spectroscopic imaging (MRSI), regional changes in brain metabolites have been reported in those with migraine as well as in those with PTH^11–21; however, similarities and differences in brain metabolites between individuals with migraine and those with acute PTH with migraine-like symptoms are unknown.

Prior studies using noninvasive MRSI have been limited to assessing brain metabolites of specific regions using a single voxel region-of-interest approach.^22,23 Here, we use a 3D MRSI sequence for the whole-brain assessment of brain metabolites including N-Acetylaspartate (NAA), glutamate (Glu), myoinositol (Ins), creatine (Cr), and choline (Cho) believed to be reflective of markers for neurons, neurotransmitters, astrocytes, cell energy, and membrane turnover, respectively.^23–25

The goal of this study was to investigate metabolite differences between participants with migraine, participants with acute PTH, and healthy controls. We hypothesized that there are metabolite differences between participants with PTH and migraine and between PTH and migraine groups compared with healthy controls.

Methods

Subject enrollment and characteristics

The study was approved by the Mayo Clinic Institutional Review Board. All participants were between the ages of 19 and 64 years and provided written informed consent prior to their participation. Potential participants were excluded if they were pregnant, had contraindications to MRI, or had a history of abnormal MRI. Individuals with migraine were excluded if they had a history of neurological disease other than migraine or a history of TBI. Participants with PTH were excluded if they had a history of moderate or severe TBI. Healthy controls with tension-type headaches on 3 or more days per month were excluded. Participants with migraine met the diagnostic criteria for migraine in accordance with the International Classification of Headache Disorders, 3^rd edition (ICHD 3).²⁶ The participants with PTH met the diagnostic criteria for acute PTH attributable to mild traumatic brain injury (mTBI) in accordance with the ICHD-3 criteria.²⁶ Individuals with PTH were enrolled between 0 and 59 days post-mTBI. Prior to imaging, all participants completed a structured interview that collected information about their headaches, including questions about average headache intensity and headache frequency. For patients with PTH, headache frequency was calculated as the percentage of days with headache since mTBI. Participants with migraine and PTH were asked about their average headache intensity, which was recorded as mild = 1; moderate = 2; or severe = 3. Imaging data were collected over a 2-year period between 2020 and 2022.

Imaging acquisition parameters

All imaging was conducted on a single 3 Tesla Siemens scanner at Mayo Clinic Arizona (Siemens Magnetom Skyra, Erlangen, Germany) using a 20-channel head and neck coil. Anatomical T₁-weighted images were acquired using a magnetization prepared rapid image acquisition gradient echo sequence with 1-mm isotropic resolution with repetition time (TR) = 2400 ms; echo time (TE) = 3.03 ms; flip angle = 8 degree covering, X = 160 mm, Y = 256 mm, and Z = 256 mm. For spectroscopy acquisitions, a spectroscopic volumetric spin echo (3D EPSI) sequence was used with full brain coverage (N_x = 64, N_y = 64, N_z = 32) with resolution 4.4 × 4.4 × 6.3 mm, scan time = 17 min, and TE/TR = 30 ms/1551 ms. The sequence included suppression of lipids using an inversion nulling preparation with an inversion time of 198 ms. A saturation band was placed behind the eye to suppress the signal from the sinus and the orbits. The slab was angled to cover the orbits in both dimensions. The sequence consisted of an interleaved acquisition of a spin echo excitation of the metabolite signal and of a gradient echo excitation for the water reference signal.¹¹ No manual shimming was performed. MRSI acquisition was performed in the same orientation as the T₁-weighted image.

MRSI analysis

The MRSI water reference and T₁-weighted images were preprocessed with MIDAS magnetic resonance spectroscopy software v2.36 (http://mrir.med.miami.edu:8000/midas/) and IDL 8.8 using a standard pipeline proposed by Maudsley et al. 2006.^27–29 Metabolite images were smoothed and interpolated to 64 × 64 × 32 voxels. Post-processing included the calculation of MRSI voxel content based on the T₁-weighted image. T₁-weighted images were used to segment gray matter, white matter, and cerebrospinal fluid and used as reference images to which all other images were realigned. MRSI maps were normalized to 2-mm isotropic Montreal Neurological Institute (MNI) space using SPM12 and smoothed with a 6-mm kernel. Quality maps represented numbers between 0 (low quality) and 4 (high quality). They were generated using the MIDAS software function “Quality Maps V2.0” (referred to as Q) with a maximum line width of 13 Hz, Creatine Cramer-Rao Lower Bound (CRLB) of 40%, and default parameters.^28,29 The Q-maps from each participant were used to generate an average quality map. Only voxels with a Q-map > 2.5 averaged across all participants were included in the analysis. The 3D metabolite maps were assessed for glutamate and glutamine (Glx), NAA, Ins, Cho, and tCr in absolute concentrations.

A full factorial design was implemented in SPM12 with two factors. The group factor consisted of three levels (healthy control/PTH/migraine), and the metabolite factor consisted of five levels corresponding to the five detectable metabolites (Glx, NAA, Ins, Cho, Cr). Both factors included ANCOVA by factor regressors while correcting for sex and age. Initially, differences were compared between all pair-wise groups (PTH-healthy control, PTH-migraine, migraine-healthy control) across all metabolites using an F-statistic to allowing for detection of both positive and negative differences. Each pair-wise group was further subdivided for each metabolite using the F-statistic to identify the contributions from each metabolite. Significance of the results was assessed at p < 0.05 family-wise error corrected for multiple comparisons. Clusters were localized using the Automated Anatomical Labelling atlas (AALv3.1)³⁰ and the Johns Hopkins University (https://identifiers.org/neurovault.collection:264) for gray matter and white matter regions, respectively. Direction of group differences for each metabolite was determined using post hoc t-tests.

Results

Twelve participants with migraine (11 female/1 male; mean age = 34, SD = 7), 10 participants with acute PTH with migraine phenotype (6 female/4 male; mean age = 55, SD = 12), and 11 healthy controls (8 female/3 male, mean age = 43, SD = 15) were included. All participants with PTH had a migraine-like phenotype, meaning that they met the A–D criterion for migraine within the ICHD-3 diagnostic criteria, with the exception that headaches could last longer than 72 h. Five participants with acute PTH had migraine-like headaches with aura, and six participants with migraine had aura. For patients with acute PTH, four had one mTBI, one had two mTBIs, two had six, and three had seven mTBIs. Three participants with acute PTH had migraine prior to having an mTBI.

The average Beck Depression Inventory scores for participants with migraine, participants with acute PTH, and healthy controls indicated minimal to mild mood disturbance. Table 1 summarizes the clinical variables and patient demographics. Figure 1 shows the average Q-map > 2.5 within all participants superimposed on the MNI 152 brain template.

FIG. 1.

Average quality map with volume >2.5 within all participants is shown in gray superimposed on Montreal Neurological Institute (MNI) 152 brain template. Any region outside of gray was not of sufficient quality to be included in the analysis.

Table 1.

Participant Demographics, Clinical Variables, and Phenotypes

	PTH	MIG	HC
Age (years)	55 ± 12	34 ± 7	43 ± 15
Sex (male/female)	4/6	1/11	3/8
Headache frequency (%)	73 ± 29	61 ± 35	n/a
Headache severity	2 ± 1	2 ± 1	n/a
Number of mTBI	4 ± 3	n/a	n/a
Time since mTBI (days)	33 ± 15	n/a	n/a
Beck Depression Inventory	11 ± 8	11 ± 7	4 ± 4

Headache frequency for participants with migraine is reported as the percent number of headache days per month.

Headache frequency for participants with PTH is reported as the percent number of headache days since mTBI.

Time since mTBI is the number of days between the date of mild traumatic brain injury and the date of imaging.

Values reported as mean ± standard deviation.

PTH, post-traumatic headache; MIG, migraine; HC, healthy controls; mTBI, mild traumatic brain injury.

Good-quality spectra were selected in cortical and deep brain structures. The significant group effects are shown in Table 2 and Figure 2. The significant group differences for each metabolite are shown in Table 3.

FIG. 2.

Group differences of F-statistic with p < 0.05 family-wise error (FWE) corrected for multiple comparisons. HC-PTH (red) and HC-MIG (blue). Note: No PTH-MIG present on the slices provided.

Table 2.

The Effect of Pooled Metabolites on Each Group (F-Statistic) from the Full Factorial Analysis Correcting for Age and Sex

Group comparisons	p_FWE	Region	Nvox	x	y	z
HC-MIG	0.03	Occipital mid L	26	−31	−89	21
HC-MIG	0.05	Angular R	1	37	−56	52
HC-MIG	0.05	Parietal inf R	1	39	−55	52
HC-MIG	0.05	Occipital mid L	1	−36	−85	24

HC-PTH	0.002	Body of corpus callosum (WM) / Cingulate mid R	299	10	−22	28
HC-PTH	0.006	Temporal sup R	168	62	−15	0
HC-PTH	0.01	Postcentral L/ Supramarginal L	104	−63	−17	17
HC-PTH	0.04	Calcarine R	7	16	−98	−2
HC-PTH	0.04	Angular L / Occipital mid L	12	−42	−73	35
HC-PTH	0.04	Supramarginal R	10	60	−40	33
HC-PTH	0.02	Vermis	44	2	−46	−30
HC-PTH	0.02	Body of corpus callosum (WM)	52	−5	−21	26
HC-PTH	0.04	Temporal sup R	7	65	−14	8
HC-PTH	0.02	Cingulate post R/Hippocampus R	55	17	−39	15
HC-PTH	0.05	Temporal sup R	1	65	−9	8
HC-PTH	0.05	Temporal sup R	1	66	−21	9
HC-PTH	0.05	Supramarginal R	2	59	−40	35

PTH-MIG	0.001	Paracentral lobule R	449	17	−35	52
PTH-MIG	0.01	Body of corpus callosum (WM) / Cingulate mid R	88	7	−19	29
PTH-MIG	0.01	Cingulate mid L	129	−6	−40	52
PTH-MIG	0.03	Occipital inf R	17	37	−77	−10
PTH-MIG	0.05	Inf frontal gyrus, triangular part L	1	−47	30	12

Only significant regions with group effect with p < 0.05 family-wise error correction for multiple comparisons are shown. The p value of each cluster (P_FWE) is provided. The normalized coordinates [x,y,z] in MNI (Montreal Neurological Institute) space of the cluster centers, cluster volume (Nvox), and the corresponding anatomical location are provided.

Mid, middle; Sup, superior; Inf, inferior; Post, posterior; L, left; R, right; WM, white matter; PTH, post-traumatic headache; MIG, migraine; HC, healthy controls.

Table 3.

Group Differences in Individual Metabolites Between Post-Traumatic Headache (PTH), Migraine (MIG), and Healthy Controls (HC), Measured by F-Statistic with p < 0.05 (Family-Wise Correction for Multiple Comparisons)

Metabolite	p_FWE	Group differences	Location	Nvox	MNI
Metabolite	p_FWE	Group differences	Location	Nvox	X	Y	Z
GLX	0.007	HC < PTH	Calcarine R	154	15	−99	−1
GLX	0.03	HC < PTH	Lingual R	16	9	−74	0
NAA	0.004	HC > PTH	Precuneus R	239	27	−47	6
Ins	0.04	HC < PTH	Amygdala L	5	−15	−4	−12
Ins	0.001	HC < MIG	Fusiform L, Temporal inferior L, Parahippocampal L	380	−30	−19	−33
Ins	0.01	HC < MIG	Parahippocampal R, Hippocampus R	98	35	−19	−24
Ins	0.01	HC < MIG	Fusiform R, Parahippocampal R	41	23	−4	−40
Cho	0.04	HC < MIG	Hippocampus R	6	38	−20	−20
GLX	0.001	PTH > MIG	Calcarine R	480	15	−99	0
GLX	0.03	PTH < MIG	Insula L	20	−34	−13	5
NAA	0.044	PTH > MIG	Inferior frontal gyrus, triangular part L	3	−47	31	7
Ins	0.01	PTH < MIG	Fusiform L	111	−32	−19	−33

The p value of each cluster (P_FWE) is provided. The normalized coordinates [x,y,z] in MNI (Montreal Neurological Institute) space of the cluster centers, cluster volume (Nvox), and the corresponding anatomical location are provided.

Glx, glutamate / glutamine complex; NAA, N-Acetylaspartate; Ins, myoinositol; Cho, choline; L, left, R, right; PTH, post-traumatic headache.

Group difference with pooled metabolites

Group differences of pooled metabolites were found between all groups, as shown in Table 2. Differences between healthy controls and participants with PTH were identified in 13 regions, including body of corpus callosum, left occipital, right posterior cingulate, bilateral supramarginal, and right temporal areas. Group differences between participants with migraine and healthy controls were found in the left occipital and right parietal lobes. Group differences between participants with PTH and migraine were found in the bilateral cingulate, right occipital, and left frontal lobe.

Group differences for each metabolite

Migraine vs healthy controls

The migraine group showed increased Cho in the right hippocampus and increased Ins in the right hippocampus, bilateral parahippocampal gyri, left inferior temporal, and bilateral fusiform areas relative to healthy controls.

PTH vs healthy controls

Relative to healthy controls, the acute PTH cohort had decreased NAA in the right precuneus, increased Glx in the right lingual and the right calcarine gyrus, and increased Ins in the left amygdala.

PTH vs migraine

We observed higher Glx in the right calcarine gyrus and decreased Glx in the left insula in participants with PTH relative to participants with migraine. Less Ins was found in the left fusiform gyrus in the PTH cohort relative to the migraine group. Decreased NAA was observed in the migraine group relative to the PTH group in the left frontal inferior area.

Discussion

The goal of this study was to investigate metabolite differences between participants with migraine, participants with acute PTH with migraine phenotype, and healthy controls. Initially, we interrogated group differences between all three groups for all metabolites. Results demonstrated differences between groups for all metabolites, including differences in metabolite concentrations in occipital, cingulate, supramarginal, and temporal regions. These results indicate that a combination of metabolites could be responsible for those group differences rather than a single metabolite. The results are consistent with those of Becerra et al., who reported a complex of metabolite differences in occipital regions of patients with episodic migraine relative to healthy controls, suggesting differences in brain chemistry.¹⁶ The authors proposed a link between these metabolite differences and biological processes that underlie brain hyperexcitability in migraine. In our case, occipital differences were identified between all three groups, and cingulate differences existed between participants with PTH and migraine and between participants with PTH and healthy controls. The supramarginal area was identified as significantly different between PTH and healthy controls. The right supramarginal region is of particular interest, as this area has previously been identified as one with potential involvement in PTH.^1,31 Supramarginal gyrus structural differences have been identified in patients with TBI with or without PTH.³² Alterations in T2*, indicative of iron accumulation, have been reported in the supramarginal area of participants with PTH. This further suggests that the atypical structure of this region is noteworthy in those with PTH.³³ Theoretically, the supramarginal gyrus might play a role in the integration of multisensory symptoms that are common among those with PTH. These group findings can inform hypothesis-driven region of interest placement for future spectroscopy studies in migraine or PTH that will include larger sample sizes and longitudinal design.

As a second part of our analysis, we looked at individual metabolite differences among the three groups. Participants with acute PTH showed significantly less NAA relative to healthy controls in the right precuneus. The right precuneus was among the few areas identified to be structurally different between PTH and healthy controls and not between migraine and healthy controls, suggestive of differences in the underlying pathophysiology of PTH and migraine.¹ The precuneus is a functional core of the default mode network.³⁴ Prior imaging studies have demonstrated default mode network abnormalities in mTBI and PTH.^35,36

Reduction of NAA has been previously reported in TBI. Friedman et al. reported reduced NAA in occipital white and gray matter in participants with mTBI, suggestive of neuronal injury and inflammation,³⁷ and Govindaraju et al. found widespread metabolite changes, including but not limited to reduced NAA/Cr and NAA/Cho in participants with mTBI.^11,12,21 Sarmento et al. showed reduced NAA in frontal and parietal lobes in individuals with PTH following mTBI.¹⁹

We observed significant increases in Glx in PTH participants with migraine-like phenotype relative to healthy controls in the lingual area and in the right calcarine gyrus. Increased Glx concentrations in acute TBI (within 2–3 days after injury) have been shown to be sensitive early injury markers and predictors of poor long-term outcomes.³⁸ Higher Glx levels are observed in severe TBI and less so in mild TBI.^39–42 These literature findings show that severe TBI causes profound and long-lasting changes in Glx suggestive of excitotoxicity post-injury. There is paucity of studies reporting whole brain changes in Glx concentrations in the post-traumatic period in participants with mTBI.

The findings of this study suggest that there are metabolite differences in Glx, NAA, and Ins between participants with PTH with migraine-like phenotype and participants with migraine. The differences in metabolite concentrations between the migraine and PTH groups consisted of increased Ins in the left fusiform gyrus, increased Glx in the left insula, decreased NAA in the inferior frontal gyrus in the migraine group, and increased Glx in the right calcarine gyrus in the PTH group. Glx is associated with energy metabolism and neurotransmission.⁴³ The observed significant differences in Glx between migraine and PTH groups signify that the metabolite function of participants with migraine is likely different than that of participants with PTH.

Future studies aimed at assessing metabolite ratios and longitudinally designed studies including larger cohorts are needed to assess the utility of MRSI for identifying potential biomarkers for PTH and migraine.

This study had some limitations. The 3D MRSI method provides voxel-by-voxel analysis of the whole brain and will output a quality spectrum for every voxel with a selected quality factor (in our case 2.5). However, it is important to note that for studies that focus on smaller regions of interest (such as thalamic subfields), a single-voxel MRS technique may be preferred, which can yield higher quality spectra with more averages. Additionally, due to signal drop-off, the 3D approach may not completely cover some occipital and frontal regions, which could be relevant in the investigation of migraine neuropathology. The choice of a quality map as a surrogate marker for metabolite quality was used rather than defining a Cramer–Rao lower bound of each metabolite for each participant. Although the analysis provided considerable full brain coverage compared with single-voxel techniques, frontal regions close to the skull were not of sufficient quality for reliable metabolite estimation. The metabolite estimation procedure did not correct for partial volume effects due to relaxation differences in gray matter and white matter. The choice of spectral quality of Q > 2.5 (out of 4) from all subjects has not been validated. Increased variability in the estimation could result in type 2 errors where metabolite differences could be missed. Another limitation of single-voxel MRS or spatially averaging 3D-MRSI is the selection bias in the choice of placement of the voxel. This limitation is overcome by using a 3D approach to MRS analysis. Another limitation of the study was the small sample size and the lack of detailed clinical information, including disability and headache impact assessments. Although migraine and PTH cohorts included male and female participants, the cohorts were too small to explore sex-specific differences in metabolite profiles. Future studies including larger cohorts may consider stratifying analyses by sex and age groups, allowing for a more detailed exploration of how these factors independently and interactively influence metabolite profiles. The temporal aspects of mTBI and recovery were not incorporated into this study. It is not known whether the metabolite concentrations immediately following TBI could be a contributing factor to the likelihood of developing acute and persistent PTH. Future work will plan to address the longitudinal changes in metabolite concentrations post mTBI.

Conclusion

We found metabolite differences between participants with migraine, participants with PTH with migraine-like phenotype, and healthy controls. The migraine group had different metabolite alterations than the PTH group, which suggests that even though they have similar phenotypes, the two groups exhibit differences in their underlying pathophysiology. Identifying the metabolite similarities and differences between migraine and PTH may play an important role in developing disease biomarkers, determining mechanisms underlying these two headache types, and identifying imaging methods for differentiating them.

Footnotes

Authors’ Contributions

S.N. was involved in data analysis and writing the article. B.W.C. and Y.Z. were involved in the sequence preparation and installation. Y.Z. helped with data acquisition and analysis. B.W.C. helped with article preparation and editing. G.D. assisted with article editing. T.J.S. provided access to patients, concept design, and assisted with article writing and editing. All coauthors agree with the contents and verify that it reflects sound research methodology. All coauthors have thoroughly reviewed and assisted in the editing of this article.

Funding Information

This study is funded by the Roubos Family Foundation and the National Institutes of Health, National Institute of Neurological Disorders and Stroke, Award Number 1R61NS113315-01.

Author Disclosure Statement

Within the prior 24 months, T.J.S. has received compensation for consulting with AbbVie, Allergan, Amgen, Axsome, Biodelivery Science, Biohaven, Collegium, Eli Lilly, Linpharma, Lundbeck, Satsuma, Scilex, and Theranica and royalties from UpToDate. He has stock options in Aural Analytics and Nocira. He has received research funding from the American Heart Association, Amgen, Henry Jackson Foundation, National Institutes of Health, Patient Centered Outcomes Research Institute, Spark Neuro, and United States Department of Defense. C.D.C. has received research funding from the American Heart Association, National Institutes of Health, and United States Department of Defense. B.W.C., S.N., G.D., and Y.Z. have no competing financial interests.

Abbreviations Used

References

Schwedt

, Chong

, Peplinski

, et al. Persistent post-traumatic headache vs. migraine: an MRI study demonstrating differences in brain structure. J Headache Pain, 2017; 18(1):87–100.

Eskridge

, Macera

, Galarneau

, et al. Injuries from combat explosions in Iraq: injury type, location, and severity. Injury, 2012; 43(10):1678–1682.

Erickson

. Treatment outcomes of chronic post-traumatic headaches after mild head trauma in US soldiers: an observational study. Headache, 2011; 51(6):932–944.

Theeler

, Erickson

. Mild head trauma and chronic headaches in returning US soldiers. Headache, 2009; 49(4):529–534.

Chong

, Berisha

, Ross

, et al. Distinguishing persistent post-traumatic headache from migraine: classification based on clinical symptoms and brain structural MRI data. Cephalalgia, 2021; 41(8):943–955.

Chong

, Peplinski

, Berisha

, et al. Differences in fibertract profiles between patients with migraine and those with persistent post-traumatic headache. Cephalalgia, 2019; 39(9):1121–1133.

Dumkrieger

, Chong

, Ross

, et al. The value of brain MRI functional connectivity data in a machine learning classifier for distinguishing migraine from persistent post-traumatic headache. Front Pain Res (Lausanne), 2022; 3:1012831.

Dumkrieger

, Chong

, Ross

, et al. Static and dynamic functional connectivity differences between migraine and persistent post-traumatic headache: a resting-state magnetic resonance imaging study. Cephalalgia, 2019; 39(11):1366–1381.

Howard

, Dumkrieger

, Chong

, et al. Symptoms of autonomic dysfunction among those with persistent posttraumatic headache attributed to mild traumatic brain injury: a comparison to migraine and healthy controls. Headache, 2018; 58(9):1397–1407.

10.

Kim

, Chong

, Dumkrieger

, et al. Clinical correlates of insomnia in patients with persistent post-traumatic headache compared with migraine. J Headache Pain, 2020; 21(1):33–36.

11.

Govind

, Gold

, Kaliannan

, et al. Whole-brain proton MR spectroscopic imaging of mild-to-moderate traumatic brain injury and correlation with neuropsychological deficits. J Neurotrauma, 2010; 27(3):483–496.

12.

Govindaraju

, Gauger

, Manley

, et al. Volumetric proton spectroscopic imaging of mild traumatic brain injury. AJNR Am J Neuroradiol, 2004; 25(5):730–737.

13.

Abad

, Rosenberg

, Roussel

, et al. Metabolic assessment of a migraine model using relaxation-enhanced (1) H spectroscopy at ultrahigh field. Magn Reson Med, 2018; 79(3):1266–1275.

14.

Aguila

, Lagopoulos

, Leaver

, et al. Elevated levels of GABA+ in migraine detected using (1) H-MRS. NMR Biomed, 2015; 28(7):890–897.

15.

Barbiroli

, Montagna

, Cortelli

, et al. Complicated migraine studied by phosphorus magnetic resonance spectroscopy. Cephalalgia, 1990; 10(5):263–272.

16.

Becerra

, Veggeberg

, Prescot

, et al. A ‘complex’ of brain metabolites distinguish altered chemistry in the cingulate cortex of episodic migraine patients. Neuroimage Clin, 2016; 11:588–594.

17.

Bell

, Stokoe

, Khaira

, et al. GABA and glutamate in pediatric migraine. Pain, 2021; 162(1):300–308.

18.

Dichgans

, Herzog

, Freilinger

, et al. 1H-MRS alterations in the cerebellum of patients with familial hemiplegic migraine type 1. Neurology, 2005; 64(4):608–613.

19.

Sarmento

, Moreira

, Brito

, et al. Proton spectroscopy in patients with post-traumatic headache attributed to mild head injury. Headache, 2009; 49(9):1345–1352.

20.

Maudsley

, Domenig

, Govind

, et al. Mapping of brain metabolite distributions by volumetric proton MR spectroscopic imaging (MRSI). Magn Reson Med, 2009; 61(3):548–559.

21.

Maudsley

, Govind

, Saigal

, et al. Longitudinal MR spectroscopy shows altered metabolism in traumatic brain injury. J Neuroimaging, 2017; 27(6):562–569.

22.

Zhu

, Barker

. MR spectroscopy and spectroscopic imaging of the brain. Methods Mol Biol, 2011; 711:203–226.

23.

Mountford

, Stanwell

, Lin

, et al. Neurospectroscopy: the past, present and future. Chem Rev, 2010; 110(5):3060–3086.

24.

Benarroch

. N-acetylaspartate and N-acetylaspartylglutamate: neurobiology and clinical significance. Neurology, 2008; 70(16):1353–1357.

25.

Delcker

, Turowski

, Mihm

, et al. Proton MR spectroscopy of neurometabolites in hepatic encephalopathy during L-ornithine-L-aspartate treatment–results of a pilot study. Metab Brain Dis, 2002; 17(2):103–111.

26.

Anonymous. Headache Classification Committee of the International Headache Society (IHS) The International Classification of Headache Disorders, 3rd edition. Cephalalgia, 2018; 38(1):1–211.

27.

Maudsley

, Darkazanli

, Alger

, et al. Comprehensive processing, display and analysis for in vivo MR spectroscopic imaging. NMR Biomed, 2006; 19(4):492–503.

28.

Ebel

, Maudsley

. Detection and correction of frequency instabilities for volumetric 1H echo-planar spectroscopic imaging. Magn Reson Med, 2005; 53(2):465–469.

29.

Ebel

, Maudsley

, Weiner

, et al. Achieving sufficient spectral bandwidth for volumetric 1H echo-planar spectroscopic imaging at 4 Tesla. Magn Reson Med, 2005; 54(3):697–701.

30.

Rolls

, Huang

, Lin

, et al. Automated anatomical labelling atlas 3. Neuroimage, 2020; 206:116189–116189.

31.

Lamm

, Decety

, Singer

. Meta-analytic evidence for common and distinct neural networks associated with directly experienced pain and empathy for pain. Neuroimage, 2011; 54(3):2492–2502.

32.

Burrowes

SAB

, Rhodes

, Meeker

, et al. Decreased grey matter volume in mTBI patients with post-traumatic headache compared to headache-free mTBI patients and healthy controls: a longitudinal MRI study. Brain Imaging Behav, 2020; 14(5):1651–1659.

33.

Nikolova

, Schwedt

, Li

, et al. T2* reduction in patients with acute post-traumatic headache. Cephalalgia, 2022; 42(4–5):357–365.

34.

Utevsky

, Smith

, Huettel

. Precuneus is a functional core of the default-mode network. J Neurosci, 2014; 34(3):932–940.

35.

Chong

, Schwedt

, Hougaard

. Brain functional connectivity in headache disorders: a narrative review of MRI investigations. J Cereb Blood Flow Metab, 2019; 39(4):650–669.

36.

, Lu

, Shang

, et al. Altered static and dynamic functional network connectivity in post-traumatic headache. J Headache Pain, 2021; 22(1):137–137.

37.

Friedman

, Brooks

, Jung

, et al. Proton MR spectroscopic findings correspond to neuropsychological function in traumatic brain injury. AJNR Am J Neuroradiol, 1998; 19(10):1879–1885.

38.

Shutter

, Tong

, Holshouser

. Proton MRS in acute traumatic brain injury: role for glutamate/glutamine and choline for outcome prediction. J Neurotrauma, 2004; 21(12):1693–1705.

39.

Hinzman

, Thomas

, Burmeister

, et al. Diffuse brain injury elevates tonic glutamate levels and potassium-evoked glutamate release in discrete brain regions at two days post-injury: an enzyme-based microelectrode array study. J Neurotrauma, 2010; 27(5):889–899.

40.

Faden

, Demediuk

, Panter

, et al. The role of excitatory amino acids and NMDA receptors in traumatic brain injury. Science, 1989; 244(4906):798–800.

41.

Globus

, Alonso

, Dietrich

, et al. Glutamate release and free radical production following brain injury: effects of posttraumatic hypothermia. J Neurochem, 1995; 65(4):1704–1711.

42.

Amorini

, Lazzarino

, Di Pietro

, et al. Severity of experimental traumatic brain injury modulates changes in concentrations of cerebral free amino acids. J Cell Mol Med, 2017; 21(3):530–542.

43.

Escartin

, Valette

, Lebon

, et al. Neuron-astrocyte interactions in the regulation of brain energy metabolism: a focus on NMR spectroscopy. J Neurochem, 2006; 99(2):393–401.